1 2014-05-30 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3 * config/arm/thumb2.md (*thumb2_movhi_insn): Set type of movw
5 * config/arm/vfp.md (*thumb2_movsi_vfp): Likewise.
7 2014-05-30 Richard Sandiford <rdsandiford@googlemail.com>
9 * ira.c (ira_get_dup_out_num): Check for output operands at
10 the start of the loop. Handle cases where an included alternative
11 follows an excluded one.
13 2014-05-29 Mike Stump <mikestump@comcast.net>
16 * collect2.c (maybe_run_lto_and_relink): Be sure to always run
17 post ld passes when lto is used.
19 2014-05-29 Vladimir Makarov <vmakarov@redhat.com>
21 PR rtl-optimization/61325
22 * lra-constraints.c (process_address): Rename to
24 (process_address): New function.
26 2014-05-29 Alan Lawrence <alan.lawrence@arm.com>
28 * config/aarch64/aarch64-builtins.c (aarch64_types_binopv_qualifiers,
29 TYPES_BINOPV): New static data.
30 * config/aarch64/aarch64-simd-builtins.def (im_lane_bound): New builtin.
31 * config/aarch64/aarch64-simd.md (aarch64_ext, aarch64_im_lane_boundsi):
33 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_const_1): Match
35 (aarch64_evpc_ext): New function.
37 * config/aarch64/iterators.md (UNSPEC_EXT): New enum element.
39 * config/aarch64/arm_neon.h (vext_f32, vext_f64, vext_p8, vext_p16,
40 vext_s8, vext_s16, vext_s32, vext_s64, vext_u8, vext_u16, vext_u32,
41 vext_u64, vextq_f32, vextq_f64, vextq_p8, vextq_p16, vextq_s8,
42 vextq_s16, vextq_s32, vextq_s64, vextq_u8, vextq_u16, vextq_u32,
43 vextq_u64): Replace __asm with __builtin_shuffle and im_lane_boundsi.
45 2014-05-29 Tom de Vries <tom@codesourcery.com>
47 * rtl.h (BLOCK_SYMBOL_CHECK): Use SYMBOL_REF_FLAGS.
49 2014-05-29 Richard Earnshaw <rearnsha@arm.com>
50 Richard Sandiford <rdsandiford@googlemail.com>
52 * arm/iterators.md (shiftable_ops): New code iterator.
53 (t2_binop0, arith_shift_insn): New code attributes.
54 * arm/predicates.md (shift_nomul_operator): New predicate.
55 * arm/arm.md (insn_enabled): Delete.
56 (enabled): Remove insn_enabled test.
57 (*arith_shiftsi): Delete. Replace with ...
58 (*<arith_shift_insn>_multsi): ... new pattern.
59 (*<arith_shift_insn>_shiftsi): ... new pattern.
60 * config/arm/arm.c (arm_print_operand): Handle operand format 'b'.
62 2014-05-29 Radovan Obradovic <robradovic@mips.com>
63 Tom de Vries <tom@codesourcery.com>
65 * config/mips/mips.h (POST_CALL_TMP_REG): Define.
66 * config/mips/mips.c (mips_emit_call_insn): Add POST_CALL_TMP_REG
68 (mips_split_call): Use POST_CALL_TMP_REG.
69 (TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): Redefine to true.
71 2014-05-29 Tom de Vries <tom@codesourcery.com>
73 * final.c (collect_fn_hard_reg_usage): Guard variable declaration
74 with #ifdef STACK_REGS.
76 2014-05-28 Jan Hubicka <hubicka@ucw.cz>
78 * varasm.c (get_variable_section): Walk aliases.
79 (place_block_symbol): Walk aliases.
81 2014-05-28 Tom de Vries <tom@codesourcery.com>
84 2014-05-28 Tom de Vries <tom@codesourcery.com>
86 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
87 * lra.c (initialize_lra_reg_info_element): Add init of
88 actual_call_used_reg_set field.
89 (lra): Call lra_create_live_ranges before lra_inheritance for
91 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
93 * lra-constraints.c (need_for_call_save_p): Use actual_call_used_reg_set
94 instead of call_used_reg_set for -fuse-caller-save.
95 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
97 2014-05-28 Richard Sandiford <rdsandiford@googlemail.com>
99 * doc/md.texi: Document that the % constraint character must
100 be at the beginning of the string.
101 * genoutput.c (validate_insn_alternatives): Check that '=',
102 '+' and '%' only appear at the beginning of a constraint.
103 * ira.c (commutative_constraint_p): Delete.
104 (ira_get_dup_out_num): Expect the '%' commutativity marker to be
105 at the start of the string.
106 * config/alpha/alpha.md (*movmemdi_1, *clrmemdi_1): Remove
108 * config/arm/neon.md (bicdi3_neon): Likewise.
109 * config/iq2000/iq2000.md (addsi3_internal, subsi3_internal, sgt_si)
110 (slt_si, sltu_si): Likewise.
111 * config/vax/vax.md (sbcdi3): Likewise.
112 * config/h8300/h8300.md (*cmpstz): Remove duplicate '+'.
113 * config/arc/arc.md (mulsi_600, mulsidi_600, umulsidi_600)
114 (mul64): Move '%' to beginning of constraint.
115 * config/arm/arm.md (*xordi3_insn): Likewise.
116 * config/nds32/nds32.md (add<mode>3, mulsi3, andsi3, iorsi3)
119 2014-05-28 Richard Sandiford <rdsandiford@googlemail.com>
121 * doc/md.texi: Document the restrictions on the "enabled" attribute.
123 2014-05-28 Jason Merrill <jason@redhat.com>
126 * cgraph.h (symtab_node::get_comdat_group_id): New.
127 * cgraphunit.c (analyze_functions): Call it.
128 * symtab.c (dump_symtab_node): Likewise.
129 * tree.c (decl_comdat_group_id): New.
130 * tree.h: Declare it.
131 * lto-streamer-out.c (write_symbol): Use it.
132 * trans-mem.c (ipa_tm_create_version_alias): Likewise.
134 2014-05-28 Francois-Xavier Coudert <fxcoudert@gcc.gnu.org>
137 * wide-int.cc: Do not include longlong.h when compiling with clang.
139 2014-05-28 Richard Biener <rguenther@suse.de>
141 * tree-ssa-propagate.c (add_control_edge): Print less vertical space.
142 * tree-vrp.c (extract_range_from_ssa_name): Also copy VR_UNDEFINED.
143 (vrp_visit_assignment_or_call): Print less vertical space.
144 (vrp_visit_stmt): Likewise.
145 (vrp_visit_phi_node): Likewise. For a PHI argument with
146 VR_VARYING range consider recording it as copy.
148 2014-05-28 Richard Biener <rguenther@suse.de>
151 2014-05-28 Richard Biener <rguenther@suse.de>
153 * hwint.h (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
155 2014-05-28 Bernd Edlinger <bernd.edlinger@hotmail.de>
157 * expr.c (expand_assignment): Fold the bitpos in the to_rtx if
158 sufficiently aligned and an offset is used at the same time.
159 (expand_expr_real_1): Likewise.
161 2014-05-28 Richard Biener <rguenther@suse.de>
164 * fold-const.c (fold_comparison): When folding
165 X +- C1 CMP Y +- C2 to X CMP Y +- C2 +- C1 also ensure
166 the sign of the remaining constant operand stays the same.
168 2014-05-28 Kaushik Phatak <kaushik.phatak@kpit.com>
170 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
171 __RL78_64BIT_DOUBLES__ or __RL78_32BIT_DOUBLES__.
172 (ASM_SPEC): Pass -m64bit-doubles or -m32bit-doubles on
174 (DOUBLE_TYPE_SIZE): Use 64 bit if TARGET_64BIT_DOUBLES
176 * gcc/config/rl78/rl78.opt (m64bit-doubles): New option.
177 (m32bit-doubles) Likewise.
178 * gcc/config/rl78/t-rl78: Add 64-bit-double multilib.
179 * doc/invoke.texi: Document -m32bit-doubles and -m64bit-doubles
182 2014-05-28 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
184 * configure.ac ($gcc_cv_ld_clearcap): New test.
185 * configure: Regenerate.
186 * config.in: Regenerate.
187 * config/sol2.opt (mclear-hwcap): New option.
188 * config/sol2.h (LINK_CLEARCAP_SPEC): Define.
189 * config/sol2-clearcap.map: Moved here from
190 testsuite/gcc.target/i386/clearcap.map.
191 * config/sol2-clearcapv2.map: Move here from
192 gcc.target/i386/clearcapv2.map.
193 * config/t-sol2 (install): Depend on install-clearcap-map.
194 (install-clearcap-map): New target.
195 * doc/invoke.texi (Option Summary, Solaris 2 Options): Document
198 2014-05-28 Richard Biener <rguenther@suse.de>
200 * hwint.h (*_HALF_WIDE_INT*): Move to ...
201 * wide-int.cc (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT):
202 ... here and remove the rest.
203 * hwint.h (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
205 2014-05-28 Richard Biener <rguenther@suse.de>
207 PR tree-optimization/61335
208 * tree-vrp.c (vrp_visit_phi_node): If the compare of old and
209 new range fails, drop to varying.
211 2014-05-28 Tom de Vries <tom@codesourcery.com>
213 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
214 * lra.c (initialize_lra_reg_info_element): Add init of
215 actual_call_used_reg_set field.
216 (lra): Call lra_create_live_ranges before lra_inheritance for
218 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
220 * lra-constraints.c (need_for_call_save_p): Use actual_call_used_reg_set
221 instead of call_used_reg_set for -fuse-caller-save.
222 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
224 2014-05-28 Radovan Obradovic <robradovic@mips.com>
225 Tom de Vries <tom@codesourcery.com>
227 * doc/invoke.texi (@item Optimization Options): Add -fuse-caller-save to
229 (@item -fuse-caller-save): New item.
231 2014-05-28 Radovan Obradovic <robradovic@mips.com>
232 Tom de Vries <tom@codesourcery.com>
234 * opts.c (default_options_table): Add OPT_LEVELS_2_PLUS entry with
235 OPT_fuse_caller_save.
237 2014-05-28 Radovan Obradovic <robradovic@mips.com>
238 Tom de Vries <tom@codesourcery.com>
240 * df-scan.c (df_get_call_refs): Use get_call_reg_set_usage.
241 * caller-save.c (setup_save_areas, save_call_clobbered_regs): Use
242 get_call_reg_set_usage.
243 * resource.c (mark_set_resources, mark_target_live_regs): Use
244 get_call_reg_set_usage.
245 * ira-int.h (struct ira_allocno): Add crossed_calls_clobbered_regs
247 (ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS): Define.
248 * ira-lives.c (process_bb_node_lives): Use get_call_reg_set_usage.
249 Calculate ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
250 * ira-build.c (ira_create_allocno): Init
251 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
252 (create_cap_allocno, propagate_allocno_info)
253 (propagate_some_info_from_allocno)
254 (copy_info_to_removed_store_destinations): Handle
255 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
256 * ira-costs.c (ira_tune_allocno_costs): Use
257 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS to adjust costs.
259 2014-05-28 Radovan Obradovic <robradovic@mips.com>
260 Tom de Vries <tom@codesourcery.com>
262 * cgraph.h (struct cgraph_rtl_info): Add function_used_regs
263 and function_used_regs_valid fields.
264 * final.c: Move include of hard-reg-set.h to before rtl.h to declare
265 find_all_hard_reg_sets.
266 (collect_fn_hard_reg_usage, get_call_fndecl, get_call_cgraph_rtl_info)
267 (get_call_reg_set_usage): New function.
268 (rest_of_handle_final): Use collect_fn_hard_reg_usage.
269 * regs.h (get_call_reg_set_usage): Declare.
271 2014-05-28 Georg-Johann Lay <avr@gjlay.de>
274 * config/dbx.h (License): Add Runtime Library Exception.
275 * config/newlib-stdint.h (License): Same.
276 * config/rtems.h (License): Same
277 * config/initfini-array.h (License): Same
278 * config/v850/v850.h (License): Same.
279 * config/v850/v850-opts.h (License): Same
280 * config/v850/rtems.h (License): Same.
282 2014-05-28 Georg-Johann Lay <avr@gjlay.de>
285 * doc/extend.texi (Local Labels): Note that label differences are
286 not supported for AVR.
288 2014-05-27 Andrew Pinski <apinski@cavium.com>
290 * config/aarch64/aarch64.md (stack_protect_set_<mode>):
291 Use <w> for the register in assembly template.
292 (stack_protect_test): Use the mode of operands[0] for the
294 (stack_protect_test_<mode>): Use <w> for the register
295 in assembly template.
297 2014-05-27 DJ Delorie <dj@redhat.com>
299 * config/rx/rx.c (add_vector_labels): New.
300 (rx_output_function_prologue): Call it.
301 (rx_handle_func_attribute): Don't require empty arguments.
302 (rx_handle_vector_attribute): New.
303 (rx_attribute_table): Add "vector" attribute.
304 * doc/extend.texi (interrupt, vector): Document new/changed
305 RX-specific attributes.
307 * config/rx/rx.c (rx_adjust_insn_length): Skip for non-insns.
309 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
311 * double-int.c (div_and_round_double) <ROUND_DIV_EXPR>: Use the proper
312 predicate to detect a negative quotient.
314 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
316 * fold-const.c (fold_comparison): Clean up and extend X +- C1 CMP C2
317 to X CMP C2 -+ C1 transformation to EQ_EXPR/NE_EXPR.
318 Add X - Y CMP 0 to X CMP Y transformation.
319 (fold_binary_loc) <EQ_EXPR/NE_EXPR>: Remove same transformations.
321 2014-05-27 Segher Boessenkool <segher@kernel.crashing.org>
323 * stmt.c (dump_case_nodes): Don't convert values to HOST_WIDE_INT
326 2014-05-27 Steve Ellcey <sellcey@mips.com>
328 * config/mips/mips.c: Add include of cgraph.h.
330 2014-05-27 Richard Biener <rguenther@suse.de>
332 * system.h (__STDC_FORMAT_MACROS): Define as very first thing.
334 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
337 * config/arm/arm.h (License): Add note to COPYING.RUNTIME.
338 * config/arm/arm-cores.def (License): Same.
339 * config/arm/arm-opts.h (License): Same.
340 * config/arm/aout.h (License): Same.
341 * config/arm/bpabi.h (License): Same.
342 * config/arm/elf.h (License): Same.
343 * config/arm/linux-elf.h (License): Same.
344 * config/arm/linux-gas.h (License): Same.
345 * config/arm/netbsd-elf.h (License): Same.
346 * config/arm/uclinux-eabi.h (License): Same.
347 * config/arm/uclinux-elf.h (License): Same.
348 * config/arm/vxworks.h (License): Same.
350 2014-05-27 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
352 * config/arm/neon.md (neon_bswap<mode>): New pattern.
353 * config/arm/arm.c (neon_itype): Add NEON_BSWAP.
354 (arm_init_neon_builtins): Handle NEON_BSWAP.
355 Define required type nodes.
356 (arm_expand_neon_builtin): Handle NEON_BSWAP.
357 (arm_builtin_vectorized_function): Handle BUILTIN_BSWAP builtins.
358 * config/arm/arm_neon_builtins.def (bswap): Define builtins.
359 * config/arm/iterators.md (VDQHSD): New mode iterator.
361 2014-05-27 Richard Biener <rguenther@suse.de>
363 * tree-vrp.c (vrp_evaluate_conditional_warnv_with_ops_using_ranges):
364 Try using literal operands when comparing value-ranges failed.
366 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
368 * ira.c (commutative_operand): Adjust for change to recog_data.
369 [Missing from previous commit.]
371 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
373 * system.h (TEST_BIT): New macro.
374 * recog.h (alternative_mask): New type.
375 (ALL_ALTERNATIVES, ALTERNATIVE_BIT): New macros.
376 (recog_data_d): Replace alternative_enabled_p array with
377 enabled_alternatives.
378 (target_recog): New structure.
379 (default_target_recog, this_target_recog): Declare.
380 (get_enabled_alternatives, recog_init): Likewise.
381 * recog.c (default_target_recog, this_target_recog): New variables.
382 (get_enabled_alternatives): New function.
383 (extract_insn): Use it.
384 (recog_init): New function.
385 (preprocess_constraints, constrain_operands): Adjust for change to
387 * postreload.c (reload_cse_simplify_operands): Likewise.
388 * reload.c (find_reloads): Likewise.
389 * ira-costs.c (record_reg_classes): Likewise.
390 * ira-lives.c (single_reg_class): Likewise. Fix bug in which
391 all alternatives after a disabled one would be skipped.
392 (ira_implicitly_set_insn_hard_regs): Likewise.
393 * ira.c (ira_setup_alts): Adjust for change to recog_data.
394 * lra-int.h (lra_insn_recog_data): Replace alternative_enabled_p
395 with enabled_alternatives.
396 * lra.c (free_insn_recog_data): Update accordingly.
397 (lra_update_insn_recog_data): Likewise.
398 (lra_set_insn_recog_data): Likewise. Use get_enabled_alternatives.
399 * lra-constraints.c (process_alt_operands): Likewise. Handle
400 only_alternative as part of the enabled mask.
401 * target-globals.h (this_target_recog): Declare.
402 (target_globals): Add a recog field.
403 (restore_target_globals): Restore this_target_recog.
404 * target-globals.c: Include recog.h.
405 (default_target_globals): Initialize recog field.
406 (save_target_globals): Likewise.
407 * reginfo.c (reinit_regs): Call recog_init.
408 * toplev.c (backend_init_target): Likewise.
410 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
412 * gencodes.c (main): Make LAST_INSN_CODE higher than any insn code,
413 rather than any named insn's code.
415 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
418 * config/arm/arm-opts.h (License): Add GCC Runtime Library Exception.
419 * config/arm/arm-cores.def (License): Same.
421 2014-05-26 Jan Hubicka <hubicka@ucw.cz>
423 * tree.h (decl_comdat_group): Declare.
424 * cgraph.h (symtab_in_same_comdat_p): Move offline to ...
425 * tree.c (decl_comdat_group): Here.
427 2014-05-26 Richard Sandiford <r.sandiford@uk.ibm.com>
429 PR rtl-optimization/61222
430 * combine.c (simplify_shift_const_1): When moving a PLUS outside
431 the shift, truncate the PLUS operand to the result mode.
433 2014-05-26 Uros Bizjak <ubizjak@gmail.com>
436 * config/i386/i386.c (ix86_rtx_costs)
437 <case CONST_INT, case CONST, case LABEL_REF, case SYMBOL_REF>:
440 2014-05-26 Martin Jambor <mjambor@suse.cz>
442 * ira.c (split_live_ranges_for_shrink_wrap): Remove bailout on
445 2014-05-26 Richard Biener <rguenther@suse.de>
447 * wide-int.h (wi::int_traits <long>, wi::int_traits <unsigned long>,
448 wi::int_traits <long long>, wi::int_traits <unsigned long long>):
449 Provide specializations.
450 (wi::int_traits <HOST_WIDE_INT>,
451 wi::int_traits <unsigned HOST_WIDE_INT>): Remove specializations.
453 2014-05-26 Alan Modra <amodra@gmail.com>
456 * config/rs6000/rs6000.c (rs6000_emit_set_const): Remove unneeded
457 params and return a bool. Remove dead code. Update comment.
458 Assert we have a const_int source. Remove bogus code from
459 32-bit HWI days. Move !TARGET_POWERPC64 handling, and correct
460 handling of constants > 2G and reg_equal note, from..
461 (rs6000_emit_set_long_const): ..here. Remove unneeded param and
462 return value. Update comment. If we can, use a new pseudo
463 for intermediate calculations.
464 * config/rs6000/rs6000-protos.h (rs6000_emit_set_const): Update
466 * config/rs6000/rs6000.md (movsi_internal1_single+1): Update
467 call to rs6000_emit_set_const in splitter.
468 (movdi_internal64+2, +3): Likewise.
470 2014-05-26 Richard Biener <rguenther@suse.de>
472 * system.h: Define __STDC_FORMAT_MACROS before
473 including inttypes.h.
474 * hwint.h (HOST_WIDEST_INT, HOST_BITS_PER_WIDEST_INT,
475 HOST_WIDEST_INT_PRINT, HOST_WIDEST_INT_PRINT_DEC,
476 HOST_WIDEST_INT_PRINT_DEC_C, HOST_WIDEST_INT_PRINT_UNSIGNED,
477 HOST_WIDEST_INT_PRINT_HEX, HOST_WIDEST_INT_PRINT_DOUBLE_HEX,
478 HOST_WIDEST_INT_C): Remove.
479 (PRId64, PRIi64, PRIo64, PRIu64, PRIx64, PRIX64): Define
480 if C99 inttypes.h is not available.
481 * coretypes.h (gcov_type, gcov_type_unsigned): Use [u]int64_t.
482 * gcov-io.h (gcov_type, gcov_type_unsigned): Likewise.
483 * gcov-io.c (gcov_histo_index): Drop non-64bit hwi case.
484 * cfgloop.h (struct niter_desc): Use uint64_t for niter field.
485 * bitmap.c (struct bitmap_descriptor_d): Use uint64_t for counters.
486 (struct output_info): Likewise.
487 (print_statistics): Adjust.
488 (dump_bitmap_statistics): Likewise.
489 * bt-load.c (migrate_btr_defs): Print with PRId64.
490 * cfg.c (dump_edge_info, dump_bb_info): Likewise.
491 (MAX_SAFE_MULTIPLIER): Adjust.
492 * cfghooks.c (dump_bb_for_graph): Print with PRId64.
493 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee,
494 dump_cgraph_node): Likewise.
495 * final.c (dump_basic_block_info): Likewise.
496 * gcov-dump.c (tag_counters, tag_summary, dump_working_sets): Likewise.
497 * gcov.c (format_gcov): Likewise.
498 * ipa-cp.c (good_cloning_opportunity_p): Likewise. Use int64_t
500 (get_clone_agg_value): Use HOST_WIDE_INT for offset.
501 * ipa-inline.c (compute_max_insns): Use int64_t for calcuation.
502 (inline_small_functions, dump_overall_stats, dump_inline_stats):
503 Use PRId64 for dumping.
504 * ipa-profile.c (dump_histogram, ipa_profile): Likewise.
505 * ira-color.c (struct allocno_hard_regs): Use int64_t for cost.
506 (add_allocno_hard_regs): Adjust.
507 * loop-doloop.c (doloop_modify): Print using PRId64.
508 * loop-iv.c (inverse): Compute in uint64_t.
509 (determine_max_iter, iv_number_of_iterations): Likewise.
510 * loop-unroll.c (decide_peel_completely, decide_peel_simple):
512 * lto-streamer-out.c (write_symbol): Use uint64_t.
513 * mcf.c (CAP_INFINITY): Use int64_t maximum.
514 (dump_fixup_edge, create_fixup_graph, cancel_negative_cycle,
515 find_max_flow, adjust_cfg_counts): Use int64_t and dump with PRId64.
516 * modulo-sched.c (const_iteration_count): Use int64_t.
517 (sms_schedule): Dump using PRId64.
518 * predict.c (dump_prediction): Likewise.
519 * pretty-print.h (pp_widest_integer): Remove.
520 * profile.c (get_working_sets, is_edge_inconsistent,
521 is_inconsistent, read_profile_edge_counts): Dump using PRId64.
522 * tree-pretty-print.c (pp_double_int): Remove case handling
523 HOST_BITS_PER_DOUBLE_INT == HOST_BITS_PER_WIDEST_INT.
524 * tree-ssa-math-opts.c (struct symbolic_number): Use uint64_t
526 (pass_optimize_bswap::execute): Remove restriction on hosts.
527 * tree-streamer-in.c (streamer_alloc_tree): Use HOST_WIDE_INT.
528 * tree-streamer-out.c (streamer_write_tree_header): Likewise.
529 * tree.c (widest_int_cst_value): Remove.
530 * tree.h (widest_int_cst_value): Likewise.
531 * value-prof.c (dump_histogram_value): Print using PRId64.
532 * gengtype.c (main): Also inject int64_t.
533 * ggc-page.c (struct max_alignment): Use int64_t.
534 * alloc-pool.c (struct allocation_object_def): Likewise.
535 * ira-conflicts.c (build_conflict_bit_table): Use uint64_t
537 * doc/tm.texi.in: Remove reference to HOST_WIDEST_INT.
538 * doc/tm.texi: Regenerated.
539 * gengtype-lex.l (IWORD): Handle [u]int64_t.
540 * config/sh/sh.c (expand_cbranchdi4): Use gcov_type.
541 * config/mmix/mmix-protos.h (mmix_intval, mmix_shiftable_wyde_value,
542 mmix_output_register_setting): Use [u]int64_t in prototypes.
543 * config/mmix/mmix.c (mmix_print_operand, mmix_output_register_setting,
544 mmix_shiftable_wyde_value, mmix_output_shiftvalue_op_from_str,
545 mmix_output_octa, mmix_output_shifted_value): Adjust.
546 (mmix_intval): Adjust. Remove unreachable case.
547 * config/mmix/mmix.md (*nonlocal_goto_receiver_expanded): Use int64_t.
549 2014-05-26 Richard Biener <rguenther@suse.de>
551 * configure.ac: Drop __int64 type check. Insist that we
552 found uint64_t and int64_t.
553 * hwint.h (HOST_BITS_PER___INT64): Remove.
554 (HOST_BITS_PER_WIDE_INT): Define to 64 and remove __int64 case.
555 (HOST_WIDE_INT_PRINT_*): Remove 32bit case.
556 (HOST_WIDEST_INT*): Define to HOST_WIDE_INT*.
557 (HOST_WIDEST_FAST_INT): Remove __int64 case.
558 * vmsdbg.h (struct _DST_SRC_COMMAND): Use int64_t
559 for dst_q_src_df_rms_cdt.
560 * configure: Regenerate.
561 * config.in: Likewise.
563 2014-05-26 Michael Tautschnig <mt@debian.org>
566 * doc/extend.texi (X86 Built-in Functions): Fix parameter lists of
567 __builtin_ia32_vfrczs[sd] and __builtin_ia32_mpsadbw256.
569 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
571 PR rtl-optimization/61278
572 * shrink-wrap.c (move_insn_for_shrink_wrap): Check df_live.
574 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
576 PR rtl-optimization/61220
577 Part of PR rtl-optimization/61225
578 * shrink-wrap.c (move_insn_for_shrink_wrap): Skip SP and FP adjustment
579 insn; skip split_edge for a block with only one successor.
581 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
583 * symtab.c (symtab_nonoverwritable_alias): Copy READONLY flag
586 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
588 * ipa-visibility.c (can_replace_by_local_alias_in_vtable): New function.
589 (update_vtable_references): New function.
590 (function_and_variable_visibility): Rewrite also vtable initializers.
591 * varpool.c (cgraph_variable_initializer_availability): Remove assert.
593 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
595 * ggc.h (ggc_grow): New function.
596 * ggc-none.c (ggc_grow): New function.
597 * ggc-page.c (ggc_grow): Likewise.
599 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
601 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
602 address_taken_from_non_vtable_p, comdat_can_be_unshared_p_1,
603 comdat_can_be_unshared_p, cgraph_externally_visible_p,
604 varpool_externally_visible_p, can_replace_by_local_alias,
605 update_visibility_by_resolution_info, function_and_variable_visibility,
606 pass_data_ipa_function_and_variable_visibility,
607 make_pass_ipa_function_and_variable_visibility,
608 whole_program_function_and_variable_visibility,
609 pass_data_ipa_whole_program_visibility,
610 make_pass_ipa_whole_program_visibility): Move to ipa-visibility.c
611 * cgraph.h (cgraph_local_node_p): Declare.
612 * ipa-visibility.c: New file.
613 * Makefile.in (OBJS): Add ipa-visiblity.o
615 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
617 * gimple-fold.c (can_refer_decl_in_current_unit_p): Be sure
618 that var decl is available.
620 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
622 * tree-core.h (tree_decl_with_vis): Replace comdat_group by
624 * tree.c (copy_node_stat): Be sure to not copy symtab_node pointer.
625 (find_decls_types_r): Do not walk COMDAT_GROUP.
626 * tree.h (DECL_COMDAT_GROUP): Revamp to use decl_comdat_group.
627 * varasm.c (make_decl_one_only): Use set_comdat_group;
628 create node if needed.
629 * ipa-inline-transform.c (save_inline_function_body): Update
630 way we decl->symtab mapping.
631 * symtab.c (symtab_hash, hash_node, eq_node
632 symtab_insert_node_to_hashtable): Remove.
633 (symtab_register_node): Update.
634 (symtab_unregister_node): Update.
635 (symtab_get_node): Reimplement as inline function.
636 (symtab_add_to_same_comdat_group): Update.
637 (symtab_dissolve_same_comdat_group_list): Update.
638 (dump_symtab_base): Update.
639 (verify_symtab_base): Update.
640 (symtab_make_decl_local): Update.
641 (fixup_same_cpp_alias_visibility): Update.
642 (symtab_nonoverwritable_alias): Update.
643 * cgraphclones.c (set_new_clone_decl_and_node_flags): Update.
644 * ipa.c (update_visibility_by_resolution_info): UPdate.
645 * bb-reorder.c: Include cgraph.h
646 * lto-streamer-out.c (DFS_write_tree_body, hash_tree): Do not deal
648 * ipa-comdats.c (set_comdat_group, ipa_comdats): Update.
649 * cgraph.c (cgraph_get_create_node): Update.
650 * cgraph.h (struct symtab_node): Add get_comdat_group, set_comdat_group
652 (symtab_get_node): Make inline.
653 (symtab_insert_node_to_hashtable): Remove.
654 (symtab_can_be_discarded): Update.
655 (decl_comdat_group): New function.
656 * tree-streamer-in.c (lto_input_ts_decl_with_vis_tree_pointers):
658 * lto-cgraph.c (lto_output_node, lto_output_varpool_node): Stream out
660 (read_comdat_group): New function.
661 (input_node, input_varpool_node): Use it.
662 * trans-mem.c (ipa_tm_create_version_alias): Update code creating
664 * mips.c (mips_start_unique_function): Likewise.
665 (ix86_code_end): Likewise.
666 (rs6000_code_end): Likweise.
667 * tree-streamer-out.c (DECL_COMDAT_GROUP): Do not stream comdat group.
669 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
671 * gengtype-state.c (fatal_reading_state): Bring offline.
672 * optabs.c (widening_optab_handler): Bring offline.
673 * optabs.h (widening_optab_handler): Likewise.
674 * final.c (get_attr_length_1): Likewise.
676 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
678 * sched-int.h (sd_iterator_cond): Manually tail recurse.
680 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
682 * config/rs6000/440.md (ppc440-integer): Include shift without dot.
683 (ppc440-compare): Include shift with dot.
684 * config/rs6000/e300c2c3.md (ppce300c3_iu): Include shift without dot.
685 * config/rs6000/e5500.md (e5500_sfx2): Include constant shift
687 * config/rs6000/e6500.md (e6500_sfx): Exclude constant shift
689 (e6500_sfx2): Include it.
690 * config/rs6000/rs6000.md ( *zero_extend<mode>di2_internal1,
691 *zero_extend<mode>di2_internal2, *zero_extend<mode>di2_internal3,
692 *zero_extendsidi2_lfiwzx, andsi3_mc, andsi3_nomc,
693 andsi3_internal0_nomc, extzvsi_internal, extzvdi_internal,
694 *extzvdi_internal1, *extzvdi_internal2, rotlsi3, *rotlsi3_64,
695 *rotlsi3_internal4, *rotlsi3_internal7le, *rotlsi3_internal7be,
696 *rotlsi3_internal10le, *rotlsi3_internal10be, rlwinm,
697 *lshiftrt_internal1le, *lshiftrt_internal1be,
698 *lshiftrt_internal4le, *lshiftrt_internal4be, rotldi3,
699 *rotldi3_internal4, *rotldi3_internal7le, *rotldi3_internal7be,
700 *rotldi3_internal10le, *rotldi3_internal10be,
701 *rotldi3_internal13le, *rotldi3_internal13be, *ashldi3_internal4,
702 ashldi3_internal5, *ashldi3_internal6, *ashldi3_internal7,
703 ashldi3_internal8, *ashldi3_internal9, anddi3_mc, anddi3_nomc,
704 *anddi3_internal2_mc, *anddi3_internal3_mc, and 4 anonymous
705 define_insns): Use type "shift" in the appropriate alternatives.
707 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
709 * config/rs6000/rs6000.md (type): Add "logical". Delete
711 (dot): Adjust comment.
712 (andsi3_mc, *andsi3_internal2_mc, *andsi3_internal3_mc,
713 *andsi3_internal4, *andsi3_internal5_mc, *boolsi3_internal2,
714 *boolsi3_internal3, *boolccsi3_internal2, *boolccsi3_internal3,
715 anddi3_mc, *anddi3_internal2_mc, *anddi3_internal3_mc,
716 *booldi3_internal2, *booldi3_internal3, *boolcdi3_internal2,
717 *boolcdi3_internal3, *boolccdi3_internal2, *boolccdi3_internal3,
718 *mov<mode>_internal2, and 10 anonymous define_insns): Use "logical".
719 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
721 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
722 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
723 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
724 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
725 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
726 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
727 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
728 * config/rs6000/8540.md (ppc8540_su): Adjust.
729 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
730 cell-cmp-microcoded): Adjust.
731 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
732 * config/rs6000/e500mc.md (e500mc_su): Adjust.
733 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
734 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
735 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
736 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
737 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
738 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
739 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
741 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
742 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare):
743 Adjust. Adjust comment.
744 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
745 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
747 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
749 * config/rs6000/rs6000.md (type): Add "add".
750 (*add<mode>3_internal1, addsi3_high, *add<mode>3_internal2,
751 *add<mode>3_internal3, *neg<mode>2_internal, and 5 anonymous
752 define_insns): Use it.
753 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
755 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
756 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
757 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
758 * config/rs6000/601.md (ppc601-integer): Adjust.
759 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
760 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
761 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
762 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
763 * config/rs6000/8540.md (ppc8540_su): Adjust.
764 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
765 cell-cmp-microcoded): Adjust.
766 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
767 * config/rs6000/e500mc.md (e500mc_su): Adjust.
768 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
769 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
770 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
771 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
772 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
773 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
774 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
776 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
777 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare): Adjust.
778 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
779 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
781 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
783 * config/rs6000/rs6000.md (type): Delete "var_shift_rotate",
784 "delayed_compare", "var_delayed_compare".
785 (var_shift): New attribute.
786 (cell_micro): Adjust.
787 (*andsi3_internal2_mc, *andsi3_internal3_mc, *andsi3_internal4,
788 *andsi3_internal5_mc, *extzvsi_internal1, *extzvsi_internal2,
789 rotlsi3, *rotlsi3_64, *rotlsi3_internal2, *rotlsi3_internal3,
790 *rotlsi3_internal4, *rotlsi3_internal5, *rotlsi3_internal6,
791 *rotlsi3_internal8le, *rotlsi3_internal8be, *rotlsi3_internal9le,
792 *rotlsi3_internal9be, *rotlsi3_internal10le, *rotlsi3_internal10be,
793 *rotlsi3_internal11le, *rotlsi3_internal11be, *rotlsi3_internal12le,
794 *rotlsi3_internal12be, ashlsi3, *ashlsi3_64, lshrsi3, *lshrsi3_64,
795 *lshiftrt_internal2le, *lshiftrt_internal2be, *lshiftrt_internal3le,
796 *lshiftrt_internal3be, *lshiftrt_internal5le, *lshiftrt_internal5be,
797 *lshiftrt_internal5le, *lshiftrt_internal5be, ashrsi3, *ashrsi3_64,
798 rotldi3, *rotldi3_internal2, *rotldi3_internal3, *rotldi3_internal4,
799 *rotldi3_internal5, *rotldi3_internal6, *rotldi3_internal7le,
800 *rotldi3_internal7be, *rotldi3_internal8le, *rotldi3_internal8be,
801 *rotldi3_internal9le, *rotldi3_internal9be, *rotldi3_internal10le,
802 *rotldi3_internal10be, *rotldi3_internal11le, *rotldi3_internal11be,
803 *rotldi3_internal12le, *rotldi3_internal12be, *rotldi3_internal13le,
804 *rotldi3_internal13be, *rotldi3_internal14le, *rotldi3_internal14be,
805 *rotldi3_internal15le, *rotldi3_internal15be, *ashldi3_internal1,
806 *ashldi3_internal2, *ashldi3_internal3, *lshrdi3_internal1,
807 *lshrdi3_internal2, *lshrdi3_internal3, *ashrdi3_internal1,
808 *ashrdi3_internal2, *ashrdi3_internal3, *anddi3_internal2_mc,
809 *anddi3_internal3_mc, as well as 11 anonymous define_insns): Adjust.
810 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
811 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
813 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
814 * config/rs6000/440.md (ppc440-integer): Adjust.
815 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
816 * config/rs6000/601.md (ppc601-integer, ppc601-compare): Adjust.
817 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
818 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
819 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
820 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
821 * config/rs6000/8540.md (ppc8540_su): Adjust.
822 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
823 cell-cmp-microcoded): Adjust.
824 * config/rs6000/e300c2c3.md (ppce300c3_cmp): Adjust.
825 * config/rs6000/e500mc.md (e500mc_su): Adjust.
826 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2,
827 e500mc64_delayed): Adjust.
828 * config/rs6000/e5500.md (e5500_sfx, e5500_delayed): Adjust.
829 * config/rs6000/e6500.md (e6500_sfx, e6500_delayed): Adjust.
830 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
831 * config/rs6000/power4.md (power4-integer, power4-compare): Adjust.
832 * config/rs6000/power5.md (power5-integer, power5-compare): Adjust.
833 * config/rs6000/power6.md (power6-shift, power6-var-rotate,
834 power6-delayed-compare, power6-var-delayed-compare): Adjust.
835 * config/rs6000/power7.md (power7-integer, power7-compare): Adjust.
836 * config/rs6000/power8.md (power8-1cyc, power8-compare): Adjust.
838 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
839 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
841 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
843 * config/rs6000/rs6000.md (type): Delete "idiv", "ldiv". Add "div".
844 (bits): New mode_attr.
845 (idiv_ldiv): Delete mode_attr.
846 (udiv<mode>3, *div<mode>3, div<div_extend>_<mode>): Adjust.
847 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
848 rs6000_adjust_priority, is_nonpipeline_insn,
849 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
851 * config/rs6000/40x.md (ppc403-idiv): Adjust.
852 * config/rs6000/440.md (ppc440-idiv): Adjust.
853 * config/rs6000/476.md (ppc476-idiv): Adjust.
854 * config/rs6000/601.md (ppc601-idiv): Adjust.
855 * config/rs6000/603.md (ppc603-idiv): Adjust.
856 * config/rs6000/6xx.md (ppc604-idiv, ppc620-idiv, ppc630-idiv,
857 ppc620-ldiv): Adjust.
858 * config/rs6000/7450.md (ppc7450-idiv): Adjust.
859 * config/rs6000/7xx.md (ppc750-idiv): Adjust.
860 * config/rs6000/8540.md (ppc8540_divide): Adjust.
861 * config/rs6000/a2.md (ppca2-idiv, ppca2-ldiv): Adjust.
862 * config/rs6000/cell.md (cell-idiv, cell-ldiv): Adjust.
863 * config/rs6000/e300c2c3.md (ppce300c3_divide): Adjust.
864 * config/rs6000/e500mc.md (e500mc_divide): Adjust.
865 * config/rs6000/e500mc64.md (e500mc64_divide): Adjust.
866 * config/rs6000/e5500.md (e5500_divide, e5500_divide_d): Adjust.
867 * config/rs6000/e6500.md (e6500_divide, e6500_divide_d): Adjust.
868 * config/rs6000/mpc.md (mpccore-idiv): Adjust.
869 * config/rs6000/power4.md (power4-idiv, power4-ldiv): Adjust.
870 * config/rs6000/power5.md (power5-idiv, power5-ldiv): Adjust.
871 * config/rs6000/power6.md (power6-idiv, power6-ldiv): Adjust.
872 * config/rs6000/power7.md (power7-idiv, power7-ldiv): Adjust.
873 * config/rs6000/power8.md (power8-idiv, power8-ldiv): Adjust.
874 * config/rs6000/rs64.md (rs64a-idiv, rs64a-ldiv): Adjust.
875 * config/rs6000/titan.md (titan_fxu_div): Adjust.
877 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
879 * config/rs6000/rs6000.md (type): Delete "insert_word",
880 "insert_dword". Add "insert".
881 (size): Update comment.
882 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
883 insn_must_be_first_in_group): Adjust.
884 (insvsi_internal, *insvsi_internal1, *insvsi_internal2,
885 *insvsi_internal3, *insvsi_internal4, *insvsi_internal5,
886 *insvsi_internal6, insvdi_internal): Adjust.
888 * config/rs6000/40x.md (ppc403-integer): Adjust.
889 * config/rs6000/440.md (ppc440-integer): Adjust.
890 * config/rs6000/476.md (ppc476-simple-integer): Adjust.
891 * config/rs6000/601.md (ppc601-integer): Adjust.
892 * config/rs6000/603.md (ppc603-integer): Adjust.
893 * config/rs6000/6xx.md (ppc604-integer): Adjust.
894 * config/rs6000/7450.md (ppc7450-integer): Adjust.
895 * config/rs6000/7xx.md (ppc750-integer): Adjust.
896 * config/rs6000/8540.md (ppc8540_su): Adjust.
897 * config/rs6000/cell.md (cell-integer, cell-insert): Adjust.
898 * config/rs6000/e300c2c3.md (ppce300c3_iu): Adjust.
899 * config/rs6000/e500mc.md (e500mc_su): Adjust.
900 * config/rs6000/e500mc64.md (e500mc64_su): Adjust.
901 * config/rs6000/e5500.md (e5500_sfx): Adjust.
902 * config/rs6000/e6500.md (e6500_sfx): Adjust.
903 * config/rs6000/mpc.md (mpccore-integer): Adjust.
904 * config/rs6000/power4.md (power4-integer, power4-insert): Adjust.
905 * config/rs6000/power5.md (power5-integer, power5-insert): Adjust.
906 * config/rs6000/power6.md (power6-insert, power6-insert-dword): Adjust.
907 * config/rs6000/power7.md (power7-integer): Adjust.
908 * config/rs6000/power8.md (power8-1cyc): Adjust.
909 * config/rs6000/rs64.md (rs64a-integer): Adjust.
910 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
912 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
914 * config/rs6000/rs6000.md (type): Add "mul". Delete "imul",
915 "imul2", "imul3", "lmul", "imul_compare", "lmul_compare".
916 (size): New attribute.
917 (dot): New attribute.
918 (cell_micro): Adjust.
919 (mulsi3, *mulsi3_internal1, *mulsi3_internal2, mulsidi3,
920 umulsidi3, smulsi3_highpart, umulsi3_highpart, muldi3,
921 *muldi3_internal1, *muldi3_internal2, smuldi3_highpart,
922 umuldi3_highpart): Adjust.
923 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
924 rs6000_adjust_priority, is_nonpipeline_insn,
925 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
927 * config/rs6000/40x.md (ppc403-imul, ppc405-imul, ppc405-imul2,
928 ppc405-imul3): Adjust.
929 * config/rs6000/440.md (ppc440-imul, ppc440-imul2): Adjust.
930 * config/rs6000/476.md (ppc476-imul): Adjust.
931 * config/rs6000/601.md (ppc601-imul): Adjust.
932 * config/rs6000/603.md (ppc603-imul, ppc603-imul2): Adjust.
933 * config/rs6000/6xx.md (ppc604-imul, ppc604e-imul, ppc620-imul,
934 ppc620-imul2, ppc620-imul3, ppc620-lmul): Adjust.
935 * config/rs6000/7450.md (ppc7450-imul, ppc7450-imul2): Adjust.
936 * config/rs6000/7xx.md (ppc750-imul, ppc750-imul2, ppc750-imul3):
938 * config/rs6000/8540.md (ppc8540_multiply): Adjust.
939 * config/rs6000/a2.md (ppca2-imul, ppca2-lmul): Adjust.
940 * config/rs6000/cell.md (cell-lmul, cell-lmul-cmp, cell-imul23,
942 * config/rs6000/e300c2c3.md (ppce300c3_multiply): Adjust.
943 * config/rs6000/e500mc.md (e500mc_multiply): Adjust.
944 * config/rs6000/e500mc64.md (e500mc64_multiply): Adjust.
945 * config/rs6000/e5500.md (e5500_multiply, e5500_multiply_i): Adjust.
946 * config/rs6000/e6500.md (e6500_multiply, e6500_multiply_i): Adjust.
947 * config/rs6000/mpc.md (mpccore-imul): Adjust.
948 * config/rs6000/power4.md (power4-lmul-cmp, power4-imul-cmp,
949 power4-lmul, power4-imul, power4-imul3): Adjust.
950 * config/rs6000/power5.md (power5-lmul-cmp, power5-imul-cmp,
951 power5-lmul, power5-imul, power5-imul3): Adjust.
952 * config/rs6000/power6.md (power6-lmul-cmp, power6-imul-cmp,
953 power6-lmul, power6-imul, power6-imul3): Adjust.
954 * config/rs6000/power7.md (power7-mul, power7-mul-compare): Adjust.
955 * config/rs6000/power8.md (power8-mul, power8-mul-compare): Adjust.
957 * config/rs6000/rs64.md (rs64a-imul, rs64a-imul2, rs64a-imul3,
959 * config/rs6000/titan.md (titan_imul): Adjust.
961 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
963 * config/rs6000/rs6000.md (type): Add new value "halfmul".
964 (*macchwc, *macchw, *macchwuc, *macchwu, *machhwc, *machhw,
965 *machhwuc, *machhwu, *maclhwc, *maclhw, *maclhwuc, *maclhwu,
966 *nmacchwc, *nmacchw, *nmachhwc, *nmachhw, *nmaclhwc, *nmaclhw,
967 *mulchwc, *mulchw, *mulchwuc, *mulchwu, *mulhhwc, *mulhhw,
968 *mulhhwuc, *mulhhwu, *mullhwc, *mullhw, *mullhwuc, *mullhwu): Use it.
969 * config/rs6000/40x.md (ppc405-imul3): Add type halfmul.
970 * config/rs6000/440.md (ppc440-imul2): Add type halfmul.
971 * config/rs6000/476.md (ppc476-imul): Add type halfmul.
972 * config/rs6000/titan.md: Delete nonsensical comment.
973 (titan_imul): Add type imul3.
974 (titan_mulhw): Remove type imul3; add type halfmul.
976 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
978 * config/rs6000/rs6000.md (type): Reorder, reformat.
980 2014-05-23 Martin Jambor <mjambor@suse.cz>
982 PR tree-optimization/53787
983 * params.def (PARAM_IPA_MAX_AA_STEPS): New param.
984 * ipa-prop.h (ipa_node_params): Rename uses_analysis_done to
985 analysis_done, update all uses.
986 * ipa-prop.c: Include domwalk.h
987 (param_analysis_info): Removed.
988 (param_aa_status): New type.
989 (ipa_bb_info): Likewise.
990 (func_body_info): Likewise.
991 (ipa_get_bb_info): New function.
992 (aa_overwalked): Likewise.
993 (find_dominating_aa_status): Likewise.
994 (parm_bb_aa_status_for_bb): Likewise.
995 (parm_preserved_before_stmt_p): Changed to use new param AA info.
996 (load_from_unmodified_param): Accept func_body_info as a parameter
997 instead of parms_ainfo.
998 (parm_ref_data_preserved_p): Changed to use new param AA info.
999 (parm_ref_data_pass_through_p): Likewise.
1000 (ipa_load_from_parm_agg_1): Likewise. Update callers.
1001 (compute_complex_assign_jump_func): Changed to use new param AA info.
1002 (compute_complex_ancestor_jump_func): Likewise.
1003 (ipa_compute_jump_functions_for_edge): Likewise.
1004 (ipa_compute_jump_functions): Removed.
1005 (ipa_compute_jump_functions_for_bb): New function.
1006 (ipa_analyze_indirect_call_uses): Likewise, moved variable
1008 (ipa_analyze_virtual_call_uses): Accept func_body_info instead of node
1009 and info, moved variable declarations down.
1010 (ipa_analyze_call_uses): Accept and pass on func_body_info instead of
1012 (ipa_analyze_stmt_uses): Likewise.
1013 (ipa_analyze_params_uses): Removed.
1014 (ipa_analyze_params_uses_in_bb): New function.
1015 (ipa_analyze_controlled_uses): Likewise.
1016 (free_ipa_bb_info): Likewise.
1017 (analysis_dom_walker): New class.
1018 (ipa_analyze_node): Handle node-specific forbidden analysis,
1019 initialize and free func_body_info, use dominator walker.
1020 (ipcp_modif_dom_walker): New class.
1021 (ipcp_transform_function): Create and free func_body_info, use
1022 ipcp_modif_dom_walker, moved a lot of functionality there.
1024 2014-05-23 Marek Polacek <polacek@redhat.com>
1025 Jakub Jelinek <jakub@redhat.com>
1027 * builtins.def: Change SANITIZE_FLOAT_DIVIDE to SANITIZE_NONDEFAULT.
1028 * gcc.c (sanitize_spec_function): Likewise.
1029 * convert.c (convert_to_integer): Include "ubsan.h". Add
1030 floating-point to integer instrumentation.
1031 * doc/invoke.texi: Document -fsanitize=float-cast-overflow.
1032 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_CAST and
1033 SANITIZE_NONDEFAULT.
1034 * opts.c (common_handle_option): Handle -fsanitize=float-cast-overflow.
1035 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW,
1036 BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW_ABORT): Add.
1037 * ubsan.c: Include "realmpfr.h" and "dfp.h".
1038 (get_ubsan_type_info_for_type): Handle REAL_TYPEs.
1039 (ubsan_type_descriptor): Set tkind to 0xffff for types other than
1040 float/double/long double.
1041 (ubsan_instrument_float_cast): New function.
1042 * ubsan.h (ubsan_instrument_float_cast): Declare.
1044 2014-05-23 Jiong Wang <jiong.wang@arm.com>
1046 * config/aarch64/predicates.md (aarch64_call_insn_operand): New
1048 * config/aarch64/constraints.md ("Ucs", "Usf"): New constraints.
1049 * config/aarch64/aarch64.md (*sibcall_insn, *sibcall_value_insn):
1050 Adjust for tailcalling through registers.
1051 * config/aarch64/aarch64.h (enum reg_class): New caller save
1053 (REG_CLASS_NAMES): Likewise.
1054 (REG_CLASS_CONTENTS): Likewise.
1055 * config/aarch64/aarch64.c (aarch64_function_ok_for_sibcall):
1056 Allow tailcalling without decls.
1058 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
1060 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
1061 Rewrite check for ORT_PARALLEL and ORT_COMBINED_PARALLEL.
1063 * omp-low.c (expand_omp_for_static_chunk): Rename variable si to
1064 gsi, and variables v_* to v*.
1066 2014-05-23 Eric Botcazou <ebotcazou@adacore.com>
1068 * varasm.c (output_constructor_bitfield): Fix thinkos in latest change.
1070 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
1072 * gimple.h (enum gf_mask): Add and use GF_OMP_FOR_SIMD.
1073 * omp-low.c: Update accordingly.
1075 * gimple.h (enum gf_mask): Rewrite "<< 0" shift expressions used
1076 for GF_OMP_FOR_KIND_MASK, GF_OMP_FOR_KIND_FOR,
1077 GF_OMP_FOR_KIND_DISTRIBUTE, GF_OMP_FOR_KIND_SIMD,
1078 GF_OMP_FOR_KIND_CILKSIMD, GF_OMP_TARGET_KIND_MASK,
1079 GF_OMP_TARGET_KIND_REGION, GF_OMP_TARGET_KIND_DATA,
1080 GF_OMP_TARGET_KIND_UPDATE.
1082 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
1083 Explicitly enumerate the expected region types.
1085 2014-05-23 Paul Eggert <eggert@cs.ucla.edu>
1088 * doc/extend.texi (Function Attributes): Fix __attribute__ ((malloc))
1089 documentation; the old documentation didn't clearly state the
1090 constraints on the contents of the pointed-to storage.
1092 2014-05-23 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1094 Fix bootstrap error on ia64
1095 * config/ia64/ia64.c (ia64_first_cycle_multipass_dfa_lookahead_guard):
1096 Return default value.
1098 2014-05-23 Thomas Preud'homme <thomas.preudhomme@arm.com>
1100 PR tree-optimization/54733
1101 * tree-ssa-math-opts.c (nop_stats): New "bswap_stats" structure.
1103 (find_bswap_or_nop_load): New.
1104 (find_bswap_1): Renamed to ...
1105 (find_bswap_or_nop_1): This. Also add support for memory source.
1106 (find_bswap): Renamed to ...
1107 (find_bswap_or_nop): This. Also add support for memory source and
1108 detection of bitwise operations equivalent to load in host endianness.
1109 (execute_optimize_bswap): Likewise. Also move its leading comment back
1110 in place and split statement transformation into ...
1111 (bswap_replace): This.
1113 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
1115 PR rtl-optimization/61215
1116 * lra-elelimination.c (lra_eliminate_regs_1): Don't use
1117 simplify_gen_subreg until final substitution.
1119 2014-05-23 Alan Modra <amodra@gmail.com>
1122 * config/rs6000/rs6000.c (mem_operand_gpr): Handle SImode.
1123 * config/rs6000/rs6000.md (extendsidi2_lfiwax, extendsidi2_nocell):
1124 Use "Y" constraint rather than "m".
1126 2014-05-23 Kugan Vivekanandarajah <kuganv@linaro.org>
1128 * config/aarch64/aarch64.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New
1130 * config/aarch64/aarch64-protos.h (aarch64_atomic_assign_expand_fenv):
1131 New function declaration.
1132 * config/aarch64/aarch64-builtins.c (aarch64_builtins) : Add
1133 AARCH64_BUILTIN_GET_FPCR, AARCH64_BUILTIN_SET_FPCR.
1134 AARCH64_BUILTIN_GET_FPSR and AARCH64_BUILTIN_SET_FPSR.
1135 (aarch64_init_builtins) : Initialize builtins
1136 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
1137 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
1138 (aarch64_expand_builtin) : Expand builtins __builtins_aarch64_set_fpcr
1139 __builtins_aarch64_get_fpcr, __builtins_aarch64_get_fpsr,
1140 and __builtins_aarch64_set_fpsr.
1141 (aarch64_atomic_assign_expand_fenv): New function.
1142 * config/aarch64/aarch64.md (set_fpcr): New pattern.
1143 (get_fpcr) : Likewise.
1144 (set_fpsr) : Likewise.
1145 (get_fpsr) : Likewise.
1146 (unspecv): Add UNSPECV_GET_FPCR and UNSPECV_SET_FPCR, UNSPECV_GET_FPSR
1147 and UNSPECV_SET_FPSR.
1148 * doc/extend.texi (AARCH64 Built-in Functions) : Document
1149 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
1150 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
1152 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
1154 PR rtl-optimization/60969
1155 * ira-costs.c (record_reg_classes): Process NO_REGS for matching
1156 constraints. Set up mem cost for NO_REGS case.
1158 2014-05-22 Thomas Schwinge <thomas@codesourcery.com>
1160 * builtin-types.def: Simplify examples for DEF_FUNCTION_TYPE_*.
1162 2012-05-22 Bernd Schmidt <bernds@codesourcery.com>
1164 * config/darwin.c: Include "lto-section-names.h".
1165 (LTO_SEGMENT_NAME): Don't define.
1166 * config/i386/winnt.c: Include "lto-section-names.h".
1167 * lto-streamer.c: Include "lto-section-names.h".
1168 * lto-streamer.h (LTO_SECTION_NAME_PREFIX): Don't define.
1169 * lto-wrapper.c: Include "lto-section-names.h".
1170 (LTO_SECTION_NAME_PREFIX): Don't define.
1171 * lto-section-names.h: New file.
1172 * cgraphunit.c: Include "lto-section-names.h".
1174 2014-05-22 Peter Bergner <bergner@vnet.ibm.com>
1176 * config/rs6000/htm.md (ttest): Use correct shift value to get CR0.
1178 2014-05-22 Richard Earnshaw <rearnsha@arm.com>
1181 * arm.md (arm_cmpdi_unsigned): Fix length calculation for Thumb2.
1183 2014-05-22 Nick Clifton <nickc@redhat.com>
1185 * config/msp430/msp430.h (ASM_SPEC): Add spaces after inserted options.
1187 2014-05-22 Eric Botcazou <ebotcazou@adacore.com>
1189 * tree-ssa-forwprop.c (associate_plusminus): Extend (T)(P + A) - (T)P
1190 -> (T)A transformation to integer types.
1192 2014-05-22 Teresa Johnson <tejohnson@google.com>
1194 * gcov-io.c (gcov_position): Use gcov_nonruntime_assert.
1195 (gcov_is_error): Remove gcc_assert from IN_LIBGCOV code.
1196 (gcov_rewrite): Use gcov_nonruntime_assert.
1198 (gcov_write_words): Ditto.
1199 (gcov_write_length): Ditto.
1200 (gcov_read_words): Use gcov_nonruntime_assert, and remove
1201 gcc_assert from IN_LIBGCOV code.
1202 (gcov_read_summary): Use gcov_error to flag profile corruption.
1203 (gcov_sync): Use gcov_nonruntime_assert.
1204 (gcov_seek): Remove gcc_assert from IN_LIBGCOV code.
1205 (gcov_histo_index): Use gcov_nonruntime_assert.
1206 (static void gcov_histogram_merge): Ditto.
1207 (compute_working_sets): Ditto.
1208 * gcov-io.h (gcov_nonruntime_assert): Define.
1209 (gcov_error): Define for !IN_LIBGCOV
1211 2014-05-22 Richard Biener <rguenther@suse.de>
1213 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1214 BUILT_IN_REALLOC like BUILT_IN_STRDUP.
1215 (call_may_clobber_ref_p_1): Handle BUILT_IN_REALLOC as allocation
1216 and deallocation site.
1217 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
1218 Handle BUILT_IN_REALLOC similar to BUILT_IN_STRDUP with also
1219 passing through the incoming points-to set.
1220 (handle_lhs_call): Use flags argument instead of recomputing it.
1221 (find_func_aliases_for_call): Call handle_lhs_call with proper
1224 2014-05-22 Jakub Jelinek <jakub@redhat.com>
1226 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Make sure
1227 all padding bits in REAL_VALUE_TYPE are cleared.
1229 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1231 Cleanup and improve multipass_dfa_lookahead_guard
1232 * config/i386/i386.c (core2i7_first_cycle_multipass_filter_ready_try,)
1233 (core2i7_first_cycle_multipass_begin,)
1234 (core2i7_first_cycle_multipass_issue,)
1235 (core2i7_first_cycle_multipass_backtrack): Update signature.
1236 * config/ia64/ia64.c
1237 (ia64_first_cycle_multipass_dfa_lookahead_guard_spec): Remove.
1238 (ia64_first_cycle_multipass_dfa_lookahead_guard): Update signature.
1239 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove
1241 (ia64_first_cycle_multipass_dfa_lookahead_guard): Merge logic from
1242 ia64_first_cycle_multipass_dfa_lookahead_guard_spec. Update return
1244 * config/rs6000/rs6000.c (rs6000_use_sched_lookahead_guard): Update
1246 * doc/tm.texi: Regenerate.
1248 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove.
1249 * haifa-sched.c (ready_try): Make signed to allow negative values.
1250 (rebug_ready_list_1): Update.
1251 (choose_ready): Simplify.
1252 (sched_extend_ready_list): Update.
1254 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1256 Remove IA64 speculation tweaking flags
1257 * config/ia64/ia64.c (ia64_set_sched_flags): Delete handling of
1258 speculation tuning flags.
1259 (msched-prefer-non-data-spec-insns,)
1260 (msched-prefer-non-control-spec-insns): Obsolete options.
1261 * haifa-sched.c (choose_ready): Remove handling of
1262 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
1263 * sched-int.h (enum SPEC_SCHED_FLAGS): Remove PREFER_NON_CONTROL_SPEC
1264 and PREFER_NON_DATA_SPEC.
1265 * sel-sched.c (process_spec_exprs): Remove handling of
1266 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
1268 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1270 Improve scheduling debug output
1271 * haifa-sched.c (debug_ready_list): Remove unnecessary prototype.
1272 (advance_one_cycle): Update.
1273 (schedule_insn, queue_to_ready): Add debug printouts.
1274 (debug_ready_list_1): New static function.
1275 (debug_ready_list): Update.
1276 (max_issue): Add debug printouts.
1277 (dump_insn_stream): New static function.
1278 (schedule_block): Use it. Also better indent printouts.
1280 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1282 Fix sched_insn debug counter
1283 * haifa-sched.c (schedule_insn): Update.
1284 (struct haifa_saved_data): Add nonscheduled_insns_begin.
1285 (save_backtrack_point, restore_backtrack_point): Update.
1286 (first_nonscheduled_insn): New static function.
1287 (queue_to_ready, choose_ready): Use it.
1288 (schedule_block): Init nonscheduled_insns_begin.
1289 (sched_emit_insn): Update.
1292 2014-05-22 Kugan Vivekanandarajah <kuganv@linaro.org>
1294 * config/aarch64/aarch64.c (aarch64_regno_regclass) : Change CORE_REGS
1296 (aarch64_secondary_reload) : LikeWise.
1297 (aarch64_class_max_nregs) : Remove CORE_REGS.
1298 * config/aarch64/aarch64.h (enum reg_class) : Remove CORE_REGS.
1299 (REG_CLASS_NAMES) : Likewise.
1300 (REG_CLASS_CONTENTS) : LikeWise.
1301 (INDEX_REG_CLASS) : Change CORE_REGS to GENERAL_REGS.
1303 2014-05-21 Guozhi Wei <carrot@google.com>
1306 * config/aarch64/arm_neon.h (vqdmulh_n_s16): Change the last operand's
1308 (vqdmulhq_n_s16): Likewise.
1310 2014-05-21 Segher Boessenkool <segher@kernel.crashing.org>
1312 * config/rs6000/predicates.md (update_indexed_address_mem): Delete.
1314 2014-05-21 Marek Polacek <polacek@redhat.com>
1317 * ubsan.c (is_ubsan_builtin_p): Turn assert into a condition.
1319 2014-05-21 Martin Jambor <mjambor@suse.cz>
1321 * doc/invoke.texi (Optimize Options): Document parameters
1322 ipa-cp-eval-threshold, ipa-max-agg-items, ipa-cp-loop-hint-bonus and
1323 ipa-cp-array-index-hint-bonus.
1325 2014-05-21 Mark Wielaard <mjw@redhat.com>
1328 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_type if DWARF
1329 version >= 3 or not strict DWARF.
1330 * langhooks.h (struct lang_hooks_for_types): Add
1331 enum_underlying_base_type.
1332 * langhooks.c (lhd_enum_underlying_base_type): New function.
1333 * gcc/langhooks.h (struct lang_hooks_for_types): Add
1334 enum_underlying_base_type.
1335 * langhooks-def.h (lhd_enum_underlying_base_type): New declaration.
1336 (LANG_HOOKS_ENUM_UNDERLYING_BASE_TYPE): New define.
1337 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add new lang hook.
1339 2014-05-21 Richard Biener <rguenther@suse.de>
1341 * doc/invoke.texi (-flto-partition=): Document one and none algorithms.
1343 2014-05-21 John Marino <gnugcc@marino.st>
1345 * config.gcc (*-*-dragonfly*): New target.
1346 * configure.ac: Detect dl_iterate_phdr (*freebsd*, *dragonfly*).
1347 * configure: Regenerate.
1348 * config/dragonfly-stdint.h: New.
1349 * config/dragonfly.h: New.
1350 * config/dragonfly.opt: New.
1351 * config/i386/dragonfly.h: New.
1352 * ginclude/stddef.h: Detect _PTRDIFF_T_DECLARED for DragonFly.
1354 2014-05-21 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1356 * tree.def (VOID_CST): New.
1357 * tree-core.h (TI_VOID): New.
1358 * tree.h (void_node): New.
1359 * tree.c (tree_node_structure_for_code, tree_code_size)
1360 (iterative_hash_expr): Handle VOID_CST.
1361 (build_common_tree_nodes): Initialize void_node.
1363 2014-05-21 Bernd Schmidt <bernds@codesourcery.com>
1365 * reload1.c (remove_init_insns, will_delete_init_insn_p): New static
1367 (reload, calculate_needs_all_insns, reload_as_needed): Use them.
1369 * config/bfin/bfin.c (split_load_immediate): Use gen_int_mode in a few
1372 * cfgrtl.c (cfg_layout_initialize): Weaken assert to only trigger if
1373 flag_reorder_blocks_and_partition.
1374 * hw-doloop.c (reorg_loops): Avoid reordering if that flag is set.
1376 2014-05-21 Oleg Endo <olegendo@gcc.gnu.org>
1379 * config/sh/sh.md (*addc_r_1): Rename to addc_t_r. Remove empty
1381 (*addc_r_t): Add new insn_and_split.
1383 2014-05-21 Jakub Jelinek <jakub@redhat.com>
1386 * omp-low.c (handle_simd_reference): New function.
1387 (lower_rec_input_clauses): Use it. Defer adding reference
1388 initialization even for reduction without placeholder if in simd,
1389 handle it properly later on.
1391 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1393 PR tree-optimization/60899
1394 * gimple-fold.c (can_refer_decl_in_current_unit_p): Cleanup;
1395 assume all static symbols will have definition wile parsing and
1396 check the do have definition later in compilation; check that
1397 variable referring symbol will be output before concluding that
1398 reference is safe; be conservative for referring local statics;
1399 be more precise about when comdat is output in other partition.
1401 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1404 * ipa-inline-transform.c (inline_call): Use add CALLEE_REMOVED
1406 * ipa-inline.c (inline_to_all_callers): If callee was removed; return.
1407 (ipa_inline): Loop inline_to_all_callers until no more aliases
1410 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1412 * ipa.c (ipa_discover_readonly_nonaddressable_var): Fix dumping;
1413 set writeonly flag only for vars actually written to.
1415 2014-05-20 Dehao Chen <dehao@google.com>
1417 * ipa-inline-transform.c (clone_inlined_nodes): Use min of edge count
1418 and callee count to get clone count.
1419 * tree-inline.c (expand_call_inline): Use callee count instead of bb
1422 2014-05-20 Richard Sandiford <rdsandiford@googlemail.com>
1424 PR rtl-optimization/61243
1425 * emit-rtl.c (emit_copy_of_insn_after): Copy CROSSING_JUMP_P.
1427 2014-05-20 Xinliang David Li <davidxl@google.com>
1429 * cgraphunit.c (walk_polymorphic_call_targets): Add
1430 dbgcnt and fopt-info support.
1431 * ipa-prop.c (ipa_make_edge_direct_to_target): Ditto.
1432 * ipa-devirt.c (ipa_devirt): Ditto.
1433 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Ditto.
1434 * ipa.c (walk_polymorphic_call_targets): Ditto.
1435 * gimple-fold.c (fold_gimple_assign): Ditto.
1436 (gimple_fold_call): Ditto.
1437 * dbgcnt.def: New counter.
1439 2014-05-20 DJ Delorie <dj@redhat.com>
1441 * config/msp430/msp430.md (split): Don't allow subregs when
1442 splitting SImode adds.
1443 (andneghi): Fix subtraction logic.
1444 * config/msp430/predicates.md (msp430_nonsubreg_or_imm_operand): New.
1446 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1448 * tree.h (DECL_ONE_ONLY): Return true only for externally visible
1450 * except.c (switch_to_exception_section, resolve_unique_section,
1451 get_named_text_section, default_function_rodata_section,
1452 align_variable, get_block_for_decl, default_section_type_flags):
1453 Use DECL_COMDAT_GROUP instead of DECL_ONE_ONLY.
1454 * symtab.c (symtab_add_to_same_comdat_group,
1455 symtab_make_decl_local, fixup_same_cpp_alias_visibility,
1456 symtab_nonoverwritable_alias, symtab_get_symbol_partitioning_class):
1458 * cgraphclones.c (cgraph_create_virtual_clone): Likewise.
1459 * bb-reorder.c (pass_partition_blocks::gate): Likewise.
1460 * config/c6x/c6x.c (c6x_elf_unique_section): Likewise.
1461 (c6x_function_in_section_p): Likewise.
1462 * config/darwin.c (machopic_select_section): Likewise.
1463 * config/arm/arm.c (arm_function_in_section_p): Likewise.
1464 * config/mips/mips.c (mips_function_rodata_section): Likewise.
1465 * config/mep/mep.c (mep_select_section): LIkewise.
1466 * config/i386/i386.c (x86_64_elf_unique_section): Likewise.
1468 2014-05-20 Eric Botcazou <ebotcazou@adacore.com>
1470 * tree-ssa-dom.c (hashable_expr_equal_p) <EXPR_CALL>: Also compare the
1471 EH region of calls to pure functions that can throw an exception.
1472 * tree-ssa-sccvn.c (vn_reference_eq): Remove duplicated test.
1473 (copy_reference_ops_from_call): Also copy the EH region of the call if
1474 it can throw an exception.
1476 2014-05-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1478 * simplify-rtx.c (simplify_binary_operation_1): Optimize case of
1479 nested VEC_SELECTs that are inverses of each other.
1481 2014-05-20 Richard Biener <rguenther@suse.de>
1483 * tree-ssa-sccvn.c (process_scc): Dump SCC here, when iterating,
1484 (extract_and_process_scc_for_name): not here.
1485 (cond_dom_walker::before_dom_children): Only process
1486 stmts that end the BB in interesting ways.
1487 (run_scc_vn): Mark param uses as visited.
1489 2014-05-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1491 * config/arm/arm.md (arith_shiftsi): Do not predicate for
1494 2014-05-20 Nick Clifton <nickc@redhat.com>
1496 * config/msp430/msp430.c (TARGET_GIMPLIFY_VA_ARG_EXPR): Define.
1497 (msp430_gimplify_va_arg_expr): New function.
1498 (msp430_print_operand): Handle (CONST (ZERO_EXTRACT)).
1500 * config/msp430/msp430.md (zero_extendpsisi2): Use + constraint on
1501 operand 0 in order to prevent confusion about the number of
1504 2014-05-20 Richard Biener <rguenther@suse.de>
1506 PR tree-optimization/61221
1507 * tree-ssa-pre.c (el_to_update): Remove.
1508 (eliminate_dom_walker::before_dom_children): Handle released
1509 VDEFs by value-numbering them to the associated VUSE. Update
1510 stmt immediately for substituted call address.
1511 (eliminate): Remove delayed stmt updating code.
1512 * tree-ssa-sccvn.c (vuse_ssa_val): New function valueizing
1513 possibly late re-numbered vuses.
1514 (vn_reference_lookup_2): Adjust.
1515 (vn_reference_lookup_pieces): Likewise.
1516 (vn_reference_lookup): Likewise.
1518 2014-05-20 Richard Biener <rguenther@suse.de>
1520 * config.gcc: Remove need_64bit_hwint.
1521 * configure.ac: Do not define NEED_64BIT_HOST_WIDE_INT.
1522 * hwint.h: Do not check NEED_64BIT_HOST_WIDE_INT but assume
1524 * config.in: Regenerate.
1525 * configure: Likewise.
1527 2014-05-19 David Wohlferd <dw@LimeGreenSocks.com>
1529 * doc/extend.texi: Create Label Attributes section,
1530 move all label attributes into it and reference it.
1532 2014-05-19 Richard Earnshaw <rearnsha@arm.com>
1534 * arm.c (thumb1_reorg): When scanning backwards skip anything
1535 that's not a proper insn.
1537 2014-05-19 Richard Biener <rguenther@suse.de>
1539 PR tree-optimization/61221
1540 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
1541 Do nothing for unreachable blocks.
1542 * tree-ssa-sccvn.c (cond_dom_walker::before_dom_children):
1543 Improve unreachability detection.
1545 2014-05-19 Richard Biener <rguenther@suse.de>
1547 PR tree-optimization/61209
1548 * tree-ssa-sccvn.c (visit_phi): Avoid setting expr to VN_TOP.
1550 2014-05-19 Nick Clifton <nickc@redhat.com>
1552 * except.c (init_eh): Fix computation of builtin setjmp buffer
1553 size to allow for targets where POINTER_SIZE > BITS_PER_WORD.
1555 2014-05-19 Richard Biener <rguenther@suse.de>
1557 PR tree-optimization/61184
1558 * tree-vrp.c (is_negative_overflow_infinity): Use
1559 TREE_OVERFLOW_P and do that check first.
1560 (is_positive_overflow_infinity): Likewise.
1561 (is_overflow_infinity): Likewise.
1562 (vrp_operand_equal_p): Properly treat operands with
1563 differing overflow as not equal.
1565 2014-05-19 Bernd Schmidt <bernds@codesourcery.com>
1567 * simplify-rtx.c (simplify_unary_operation_1): Use CONST_INT_P in
1568 shift simplification where it was intended.
1570 2014-05-19 Christian Bruel <christian.bruel@st.com>
1573 * config/sh/sh.md (movsf_ie): Unset fp_mode for fmov.
1575 2014-05-19 Richard Sandiford <r.sandiford@uk.ibm.com>
1578 * config/sparc/sparc.c (sparc_fold_builtin): Use widest_int rather
1581 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
1583 * reg-notes.def (CROSSING_JUMP): Likewise.
1584 * rtl.h (rtx_def): Update comment for jump flag.
1585 (CROSSING_JUMP_P): Define.
1586 * cfgcleanup.c (try_forward_edges, try_optimize_cfg): Use it instead
1587 of a REG_CROSSING_JUMP note.
1588 * cfghooks.c (tidy_fallthru_edges): Likewise.
1589 * cfgrtl.c (fixup_partition_crossing, rtl_verify_edges): Likewise.
1590 * emit-rtl.c (try_split): Likewise.
1591 * haifa-sched.c (sched_create_recovery_edges): Likewise.
1592 * ifcvt.c (find_if_case_1, find_if_case_2): Likewise.
1593 * jump.c (redirect_jump_2): Likewise.
1594 * reorg.c (follow_jumps, fill_slots_from_thread): Likewise.
1595 (relax_delay_slots): Likewise.
1596 * config/arc/arc.md (jump_i, cbranchsi4_scratch, *bbit): Likewise.
1597 (bbit_di): Likewise.
1598 * config/arc/arc.c (arc_reorg, arc_can_follow_jump): Likewise.
1599 * config/sh/sh.md (jump_compact): Likewise.
1600 * bb-reorder.c (rotate_loop): Likewise.
1601 (pass_duplicate_computed_gotos::execute): Likewise.
1602 (add_reg_crossing_jump_notes): Rename to...
1603 (update_crossing_jump_flags): ...this.
1604 (pass_partition_blocks::execute): Update accordingly.
1606 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
1608 * tree.h: Remove extraneous template <>.
1610 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1612 * ipa.c (symtab_remove_unreachable_nodes): Remove
1613 symbol from comdat group if its body was eliminated.
1614 (comdat_can_be_unshared_p_1): Static symbols can always be privatized.
1615 * symtab.c (symtab_remove_from_same_comdat_group): Break out from ...
1616 (symtab_unregister_node): ... this one.
1617 (verify_symtab_base): More strict checking of comdats.
1618 * cgraph.h (symtab_remove_from_same_comdat_group): Declare.
1620 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1622 * tree-pass.h (make_pass_ipa_comdats): New pass.
1623 * timevar.def (TV_IPA_COMDATS): New timevar.
1624 * passes.def (pass_ipa_comdats): Add.
1625 * Makefile.in (OBJS): Add ipa-comdats.o
1626 * ipa-comdats.c: New file.
1628 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1630 * ipa.c (update_visibility_by_resolution_info): New function.
1631 (function_and_variable_visibility): Use it.
1633 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1635 * cgraph.h (symtab_first_defined_symbol, symtab_next_defined_symbol):
1637 (FOR_EACH_DEFINED_SYMBOL): New macro.
1638 (varpool_first_static_initializer, varpool_next_static_initializer,
1639 varpool_first_defined_variable, varpool_next_defined_variable):
1641 (symtab_in_same_comdat_p): Correctly deal with inline functions.
1643 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1645 * ggc-page.c (ggc_handle_finalizers): Add comment.
1647 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1649 * ggc-common.c (ggc_internal_cleared_alloc): Adjust.
1650 * ggc-none.c (ggc_internal_alloc): Assert if a finalizer is passed.
1651 (ggc_internal_cleared_alloc): Likewise.
1652 * ggc-page.c (finalizer): New class.
1653 (vec_finalizer): Likewise.
1654 (globals::finalizers): New member.
1655 (globals::vec_finalizers): Likewise.
1656 (ggc_internal_alloc): Record the finalizer if any for the block being
1658 (ggc_handle_finalizers): New function.
1659 (ggc_collect): Call ggc_handle_finalizers.
1660 * ggc.h (ggc_internal_alloc): Add arguments to allow installing a
1662 (ggc_internal_cleared_alloc): Likewise.
1663 (finalize): New function.
1664 (need_finalization_p): Likewise.
1665 (ggc_alloc): Install the type's destructor as the finalizer if it
1667 (ggc_cleared_alloc): Likewise.
1668 (ggc_vec_alloc): Likewise.
1669 (ggc_cleared_vec_alloc): Likewise.
1671 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1673 * ggc.h (ggc_alloc_cleared_simd_clone_stat): Remove function.
1675 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1677 * alias.c (record_alias_subset): Adjust.
1678 * bitmap.c (bitmap_element_allocate): Likewise.
1679 (bitmap_gc_alloc_stat): Likewise.
1680 * cfg.c (init_flow): Likewise.
1681 (alloc_block): Likewise.
1682 (unchecked_make_edge): Likewise.
1683 * cfgloop.c (alloc_loop): Likewise.
1684 (flow_loops_find): Likewise.
1685 (rescan_loop_exit): Likewise.
1686 * cfgrtl.c (init_rtl_bb_info): Likewise.
1687 * cgraph.c (insert_new_cgraph_node_version): Likewise.
1688 (cgraph_allocate_node): Likewise.
1689 (cgraph_create_edge_1): Likewise.
1690 (cgraph_allocate_init_indirect_info): Likewise.
1691 * cgraphclones.c (cgraph_clone_edge): Likewise.
1692 * cgraphunit.c (add_asm_node): Likewise.
1693 (init_lowered_empty_function): Likewise.
1694 * config/aarch64/aarch64.c (aarch64_init_machine_status): Likewise.
1695 * config/alpha/alpha.c (alpha_init_machine_status): Likewise.
1696 (alpha_use_linkage): Likewise.
1697 * config/arc/arc.c (arc_init_machine_status): Likewise.
1698 * config/arm/arm.c (arm_init_machine_status): Likewise.
1699 * config/avr/avr.c (avr_init_machine_status): Likewise.
1700 * config/bfin/bfin.c (bfin_init_machine_status): Likewise.
1701 * config/c6x/c6x.c (c6x_init_machine_status): Likewise.
1702 * config/cris/cris.c (cris_init_machine_status): Likewise.
1703 * config/darwin.c (machopic_indirection_name): Likewise.
1704 (darwin_build_constant_cfstring): Likewise.
1705 (darwin_enter_string_into_cfstring_table): Likewise.
1706 * config/epiphany/epiphany.c (epiphany_init_machine_status): Likewise.
1707 * config/frv/frv.c (frv_init_machine_status): Likewise.
1708 * config/i386/i386.c (get_dllimport_decl): Likewise.
1709 (ix86_init_machine_status): Likewise.
1710 (assign_386_stack_local): Likewise.
1711 * config/i386/winnt.c (i386_pe_record_external_function): Likewise.
1712 (i386_pe_maybe_record_exported_symbol): Likewise.
1713 (i386_pe_record_stub): Likewise.
1714 * config/ia64/ia64.c (ia64_init_machine_status): Likewise.
1715 * config/iq2000/iq2000.c (iq2000_init_machine_status): Likewise.
1716 * config/m32c/m32c.c (m32c_init_machine_status): Likewise.
1717 (m32c_note_pragma_address): Likewise.
1718 * config/mep/mep.c (mep_init_machine_status): Likewise.
1719 (mep_note_pragma_flag): Likewise.
1720 * config/mips/mips.c (mflip_mips16_use_mips16_p): Likewise.
1721 (mips16_local_alias): Likewise.
1722 (mips_init_machine_status): Likewise.
1723 * config/mmix/mmix.c (mmix_init_machine_status): Likewise.
1724 * config/moxie/moxie.c (moxie_init_machine_status): Likewise.
1725 * config/msp430/msp430.c (msp430_init_machine_status): Likewise.
1726 * config/nds32/nds32.c (nds32_init_machine_status): Likewise.
1727 * config/nios2/nios2.c (nios2_init_machine_status): Likewise.
1728 * config/pa/pa.c (pa_init_machine_status): Likewise.
1729 (pa_get_deferred_plabel): Likewise.
1730 * config/rl78/rl78.c (rl78_init_machine_status): Likewise.
1731 * config/rs6000/rs6000.c (builtin_function_type): Likewise.
1732 (rs6000_init_machine_status): Likewise.
1733 (output_toc): Likewise.
1734 * config/s390/s390.c (s390_init_machine_status): Likewise.
1735 * config/score/score.c (score_output_external): Likewise.
1736 * config/sparc/sparc.c (sparc_init_machine_status): Likewise.
1737 * config/spu/spu.c (spu_init_machine_status): Likewise.
1738 * config/tilegx/tilegx.c (tilegx_init_machine_status): Likewise.
1739 * config/tilepro/tilepro.c (tilepro_init_machine_status): Likewise.
1740 * config/xtensa/xtensa.c (xtensa_init_machine_status): Likewise.
1741 * coverage.c (coverage_end_function): Likewise.
1742 * dbxout.c (dbxout_init): Likewise.
1743 * doc/gty.texi: Don't mention variable_size attribute.
1744 * dwarf2cfi.c (new_cfi): Adjust.
1745 (new_cfi_row): Likewise.
1746 (copy_cfi_row): Likewise.
1747 (create_cie_data): Likewise.
1748 * dwarf2out.c (dwarf2out_alloc_current_fde): Likewise.
1749 (new_loc_descr): Likewise.
1750 (find_AT_string_in_table): Likewise.
1751 (add_addr_table_entry): Likewise.
1752 (new_die): Likewise.
1753 (add_var_loc_to_decl): Likewise.
1754 (clone_die): Likewise.
1755 (clone_as_declaration): Likewise.
1756 (break_out_comdat_types): Likewise.
1757 (new_loc_list): Likewise.
1758 (add_loc_descr_to_each): Likewise.
1759 (add_location_or_const_value_attribute): Likewise.
1760 (add_linkage_name): Likewise.
1761 (lookup_filename): Likewise.
1762 (dwarf2out_var_location): Likewise.
1763 (new_line_info_table): Likewise.
1764 (dwarf2out_init): Likewise.
1765 (mem_loc_descriptor): Likewise.
1766 (loc_descriptor): Likewise.
1767 (add_const_value_attribute): Likewise.
1768 (tree_add_const_value_attribute): Likewise.
1769 (comp_dir_string): Likewise.
1770 (dwarf2out_vms_debug_main_pointer): Likewise.
1771 (string_cst_pool_decl): Likewise.
1772 * emit-rtl.c (set_mem_attrs): Likewise.
1773 (get_reg_attrs): Likewise.
1774 (start_sequence): Likewise.
1775 (init_emit): Likewise.
1776 (init_emit_regs): Likewise.
1777 * except.c (init_eh_for_function): Likewise.
1778 (gen_eh_region): Likewise.
1779 (gen_eh_region_catch): Likewise.
1780 (gen_eh_landing_pad): Likewise.
1781 (add_call_site): Likewise.
1782 * function.c (add_frame_space): Likewise.
1783 (insert_temp_slot_address): Likewise.
1784 (assign_stack_temp_for_type): Likewise.
1785 (get_hard_reg_initial_val): Likewise.
1786 (allocate_struct_function): Likewise.
1787 (prepare_function_start): Likewise.
1788 (types_used_by_var_decl_insert): Likewise.
1789 * gengtype.c (variable_size_p): Remove function.
1790 (enum alloc_quantity): Remove enum.
1791 (write_typed_alloc_def): Remove function.
1792 (write_typed_struct_alloc_def): Likewise.
1793 (write_typed_typedef_alloc_def): Likewise.
1794 (write_typed_alloc_defns): Likewise.
1796 * ggc-common.c (ggc_cleared_alloc_htab_ignore_args): Adjust.
1797 (ggc_cleared_alloc_ptr_array_two_args): Likewise.
1798 * ggc.h (ggc_alloc): new function.
1799 (ggc_cleared_alloc): Likewise.
1800 (ggc_vec_alloc): Template on type of vector element, and remove
1801 element size argument.
1802 (ggc_cleared_vec_alloc): Likewise.
1803 * gimple.c (gimple_build_omp_for): Adjust.
1804 (gimple_copy): Likewise.
1805 * ipa-cp.c (get_replacement_map): Likewise.
1806 (find_aggregate_values_for_callers_subset): Likewise.
1807 (known_aggs_to_agg_replacement_list): Likewise.
1808 * ipa-devirt.c (get_odr_type): Likewise.
1809 * ipa-prop.c (ipa_node_duplication_hook): Likewise.
1810 (read_agg_replacement_chain): Likewise.
1811 * loop-iv.c (get_simple_loop_desc): Likewise.
1812 * lto-cgraph.c (input_node_opt_summary): Likewise.
1813 * lto-section-in.c (lto_new_in_decl_state): Likewise.
1814 * lto-streamer-in.c (lto_input_eh_catch_list): Likewise.
1815 (input_eh_region): Likewise.
1816 (input_eh_lp): Likewise.
1817 (input_cfg): Likewise.
1818 * optabs.c (set_optab_libfunc): Likewise.
1819 (init_tree_optimization_optabs): Likewise.
1820 (set_conv_libfunc): Likewise.
1821 * passes.c (do_per_function_toporder): Likewise.
1822 * rtl.h: Don't use variable_size gty attribute.
1823 * sese.c (if_region_set_false_region): Adjust.
1824 * stringpool.c (gt_pch_save_stringpool): Likewise.
1825 * target-globals.c (save_target_globals): Likewise.
1826 * toplev.c (general_init): Likewise.
1827 * trans-mem.c (record_tm_replacement): Likewise.
1828 (split_bb_make_tm_edge): Likewise.
1829 * tree-cfg.c (move_sese_region_to_fn): Likewise.
1830 * tree-data-ref.h (lambda_vector_new): Likewise.
1831 * tree-eh.c (add_stmt_to_eh_lp_fn): Likewise.
1832 * tree-iterator.c (tsi_link_before): Likewise.
1833 (tsi_link_after): Likewise.
1834 * tree-scalar-evolution.c (new_scev_info_str): Likewise.
1835 * tree-ssa-loop-niter.c (record_estimate): Likewise.
1836 * tree-ssa-operands.c (ssa_operand_alloc): Likewise.
1837 * tree-ssa-operands.h: Don't use variable_size gty attribute.
1838 * tree-ssa.c (init_tree_ssa): Adjust.
1839 * tree-ssanames.c (set_range_info): Likewise.
1840 (get_ptr_info): Likewise.
1841 (duplicate_ssa_name_ptr_info): Likewise.
1842 (duplicate_ssa_name_range_info): Likewise.
1843 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Likewise.
1844 (unpack_ts_fixed_cst_value_fields): Likewise.
1845 * tree.c (build_fixed): Likewise.
1846 (build_real): Likewise.
1847 (build_string): Likewise.
1848 (decl_priority_info): Likewise.
1849 (decl_debug_expr_insert): Likewise.
1850 (decl_value_expr_insert): Likewise.
1851 (decl_debug_args_insert): Likewise.
1852 (type_hash_add): Likewise.
1853 (build_omp_clause): Likewise.
1854 * ubsan.c (decl_for_type_insert): Likewise.
1855 * varasm.c (get_unnamed_section): Likewise.
1856 (get_noswitch_section): Likewise.
1857 (get_section): Likewise.
1858 (get_block_for_section): Likewise.
1859 (create_block_symbol): Likewise.
1860 (build_constant_desc): Likewise.
1861 (create_constant_pool): Likewise.
1862 (force_const_mem): Likewise.
1863 (record_tm_clone_pair): Likewise.
1864 * varpool.c (varpool_create_empty_node): Likewise.
1866 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1868 * dwarf2out.c (tree_add_const_value_attribute): Call
1869 ggc_internal_cleared_alloc instead of ggc_alloc_cleared_atomic.
1870 * gengtype.c (write_typed_alloc_def): Call ggc_internal_<x>alloc
1871 instead of ggc_internal_<x>alloc_stat.
1872 * ggc-common.c (ggc_internal_cleared_alloc): Drop _stat suffix.
1873 (ggc_realloc): Likewise.
1874 * ggc-none.c (ggc_internal_alloc): Likewise.
1875 (ggc_internal_cleared_alloc): Likewise.
1876 * ggc-page.c: Likewise.
1877 * ggc.h (ggc_internal_alloc_stat): Likewise.
1878 (ggc_internal_alloc): Remove macro.
1879 (ggc_internal_cleared_alloc_stat): Drop _stat suffix.
1880 (ggc_internal_cleared_alloc): Remove macro.
1881 (GGC_RESIZEVEC): Adjust.
1882 (ggc_resizevar): Remove macro.
1883 (ggc_internal_vec_alloc_stat): Drop _stat suffix.
1884 (ggc_internal_cleared_vec_alloc_stat): Likewise.
1885 (ggc_internal_vec_cleared_alloc): Remove macro.
1886 (ggc_alloc_atomic_stat): Drop _stat suffix.
1887 (ggc_alloc_atomic): Remove macro.
1888 (ggc_alloc_cleared_atomic): Remove macro.
1889 (ggc_alloc_string_stat): Drop _stat suffix.
1890 (ggc_alloc_string): Remove macro.
1891 (ggc_alloc_rtx_def_stat): Adjust.
1892 (ggc_alloc_tree_node_stat): Likewise.
1893 (ggc_alloc_cleared_tree_node_stat): Likewise.
1894 (ggc_alloc_cleared_gimple_statement_stat): Likewise.
1895 (ggc_alloc_cleared_simd_clone_stat): Likewise.
1896 * gimple.c (gimple_build_omp_for): Likewise.
1897 (gimple_copy): Likewise.
1898 * stringpool.c (ggc_alloc_string_stat): Drop _stat suffix.
1899 * toplev.c (realloc_for_line_map): Adjust.
1900 * tree-data-ref.h (lambda_vector_new): Likewise.
1901 * tree-phinodes.c (allocate_phi_node): Likewise.
1902 * tree.c (grow_tree_vec_stat): Likewise.
1903 * vec.h (va_gc::reserve): Adjust.
1905 2014-05-17 Ajit Agarwal <ajitkum@xilinx.com>
1907 * config/microblaze/microblaze.c (break_handler): New Declaration.
1908 (microblaze_break_function_p,microblaze_is_break_handler): New.
1909 (compute_frame_size): Use microblaze_break_function_p.
1910 Add the test of break_handler.
1911 (microblaze_function_prologue) : Add the test of variable
1912 break_handler. Check the fnname by BREAK_HANDLER_NAME.
1913 (microblaze_function_epilogue) : Add the test of break_handler.
1914 (microblaze_globalize_label) : Add the test of break_handler.
1915 Check the name by BREAK_HANDLER_NAME.
1917 * config/microblaze/microblaze.h (BREAK_HANDLER_NAME): New macro
1919 * config/microblaze/microblaze.md (*<optab>,<optab>_internal): Add
1920 microblaze_is_break_handler test.
1921 (call_internal1,call_value_intern): Use microblaze_break_function_p.
1922 Use SYMBOL_REF_DECL.
1924 * config/microblaze/microblaze-protos.h
1925 (microblaze_break_function_p,microblaze_is_break_handler):
1928 * doc/extend.texi (MicroBlaze break_handler Functions): Document
1929 new MicroBlaze break_handler functions.
1931 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
1933 * doc/extend.texi (Size of an asm): Move node text according
1934 to its @menu entry position.
1936 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
1938 PR tree-optimization/61140
1939 PR tree-optimization/61150
1940 PR tree-optimization/61197
1941 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
1943 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
1945 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
1947 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
1949 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
1950 __SIZEOF_INT128__ is defined.
1952 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
1954 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
1955 (rs6000_delegitimize_address): Use it.
1957 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
1959 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
1960 inplace argument. Store the new address in the original MEM when true.
1961 * emit-rtl.c (change_address_1): Likewise.
1962 (adjust_address_1, adjust_automodify_address_1, offset_address):
1964 * rtl.h (plus_constant): Add an inplace argument.
1965 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
1966 when true. Avoid generating (plus X (const_int 0)).
1967 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
1968 in-place. Pass true to plus_constant.
1969 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
1971 2014-05-16 Dehao Chen <dehao@google.com>
1973 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
1975 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1978 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
1980 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
1982 2014-05-16 Dehao Chen <dehao@google.com>
1984 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
1985 optimize_function_for_size_p.
1986 * regs.h (REG_FREQ_FROM_BB): Likewise.
1988 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1991 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
1992 negt_reg_operand cases.
1993 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
1995 * config/sh/predicates.md (cbranch_treg_value): Simplify.
1997 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1999 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
2002 2014-05-16 David Malcolm <dmalcolm@redhat.com>
2005 2014-04-29 David Malcolm <dmalcolm@redhat.com>
2007 * tree-cfg.c (dump_function_to_file): Dump the return type of
2008 functions, in a line to itself before the function body, mimicking
2009 the layout of a C function.
2011 2014-05-16 Dehao Chen <dehao@google.com>
2013 * cfghooks.c (make_forwarder_block): Use direct computation to
2014 get fall-through edge's count and frequency.
2016 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
2018 * config/arc/arc.c (arc_init): Fix typo in error message.
2019 * config/i386/i386.c (ix86_expand_builtin): Likewise.
2020 (split_stack_prologue_scratch_regno): Likewise.
2021 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
2022 word from error message.
2024 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
2026 * ira-costs.c: Fix typo in comment.
2028 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
2030 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
2032 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
2034 * varpool.c (dump_varpool_node): Dump write-only flag.
2035 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
2037 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
2038 write-only variables.
2039 * ipa.c (process_references): New function.
2040 (set_readonly_bit): New function.
2041 (set_writeonly_bit): New function.
2042 (clear_addressable_bit): New function.
2043 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
2044 fix handling of aliases.
2045 * cgraph.h (struct varpool_node): Add writeonly flag.
2047 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
2049 PR rtl-optimization/60969
2050 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
2051 Calculate costs for this case.
2053 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
2055 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
2056 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
2058 2014-05-16 Richard Biener <rguenther@suse.de>
2060 PR tree-optimization/61194
2061 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
2062 bool patterns ending in a COND_EXPR.
2064 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2066 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
2068 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2070 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
2071 where we were unable to cost an RTX.
2073 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2075 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
2078 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2079 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2081 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
2083 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2084 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2086 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
2087 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
2089 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2090 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2092 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
2095 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2096 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2098 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
2101 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2102 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2104 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
2105 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
2107 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2108 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2110 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
2113 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2114 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2116 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
2117 ZERO_EXTEND and SIGN_EXTEND better.
2119 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2120 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2122 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
2125 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2126 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2128 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
2129 costs when costing loads and stores to memory.
2131 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2132 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
2134 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
2137 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2139 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
2141 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2142 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2144 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
2146 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
2148 (aarch64_rtx_mult_cost): New.
2149 (aarch64_rtx_costs): Use it, refactor as appropriate.
2151 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2152 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2154 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
2155 emit instructions, return number of instructions which would
2157 (aarch64_add_constant): Update call to aarch64_build_constant.
2158 (aarch64_output_mi_thunk): Likewise.
2159 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
2162 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2164 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
2165 (TARGET_RTX_COSTS): Call it.
2167 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2169 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
2170 (cortexa57_vector_cost): Likewise.
2171 (cortexa57_tunings): Use them.
2173 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2175 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
2176 (cpu_addrcost_table): Use it.
2177 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
2178 (aarch64_address_cost): Rewrite using aarch64_classify_address,
2181 2014-05-16 Richard Biener <rguenther@suse.de>
2183 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
2184 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
2185 (visit_phi): Ignore edges marked as not executable.
2186 (class cond_dom_walker): New.
2187 (cond_dom_walker::before_dom_children): Value-number
2188 control statements and mark successor edges as not
2189 executable if possible.
2190 (run_scc_vn): First walk all control statements in
2191 dominator order, marking edges as not executable.
2192 * tree-inline.c (copy_edges_for_bb): Be not confused
2193 about random edge flags.
2195 2014-05-16 Richard Biener <rguenther@suse.de>
2197 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
2199 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
2202 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
2203 (__TM_simple_begin): Use it.
2204 (__TM_begin): Likewise.
2206 2014-05-15 Martin Jambor <mjambor@suse.cz>
2209 * ipa-prop.c (update_indirect_edges_after_inlining): Check
2210 type_preserved flag when the indirect edge is polymorphic.
2212 2014-05-15 Martin Jambor <mjambor@suse.cz>
2214 PR tree-optimization/61090
2215 * tree-sra.c (sra_modify_expr): Pass the current gsi to
2216 build_ref_for_model.
2218 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2220 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
2221 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
2223 2014-05-15 Jakub Jelinek <jakub@redhat.com>
2225 PR tree-optimization/61158
2226 * fold-const.c (fold_binary_loc): If X is zero-extended and
2227 shiftc >= prec, make sure zerobits is all ones instead of
2228 invoking undefined behavior.
2230 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2232 * regcprop.h: New file.
2233 * regcprop.c (skip_debug_insn_p): New decl.
2234 (replace_oldest_value_reg): Check skip_debug_insn_p.
2235 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
2236 * shrink-wrap.c: Include regcprop.h.
2237 (prepare_shrink_wrap): Call
2238 copyprop_hardreg_forward_bb_without_debug_insn.
2240 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2242 * shrink-wrap.h: Update comment.
2243 * shrink-wrap.c: Update comment.
2244 (next_block_for_reg): Rename to live_edge_for_reg.
2245 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
2246 (move_insn_for_shrink_wrap): Split live_edge.
2247 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
2249 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
2251 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
2253 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
2254 * config/sparc/sparc.md (fptype_ut699): New attribute.
2255 (in_branch_delay): Return false if -mfix-ut699 is specified and
2256 fptype_ut699 is set to single.
2257 (truncdfsf2): Add fptype_ut699 attribute.
2258 (fix_truncdfsi2): Likewise.
2259 (floatsisf2): Change fptype attribute.
2260 (fix_truncsfsi2): Likewise.
2261 (negtf2_notv9): Delete.
2262 (negtf2_v9): Likewise.
2263 (negtf2_hq): New instruction.
2264 (negtf2): New instruction and splitter.
2265 (negdf2_notv9): Rewrite.
2266 (abstf2_notv9): Delete.
2267 (abstf2_hq_v9): Likewise.
2268 (abstf2_v9): Likewise.
2269 (abstf2_hq): New instruction.
2270 (abstf2): New instruction and splitter.
2271 (absdf2_notv9): Rewrite.
2273 2014-05-14 Cary Coutant <ccoutant@google.com>
2276 * opts.c (common_handle_option): Don't special-case "-g".
2277 (set_debug_level): Default to at least level 2 with "-g".
2279 2014-05-14 DJ Delorie <dj@redhat.com>
2281 * config/msp430/msp430.c (msp430_builtin): Add
2282 MSP430_BUILTIN_DELAY_CYCLES.
2283 (msp430_init_builtins): Register void __delay_cycles(long long).
2284 (msp430_builtin_decl): Add it.
2285 (cg_magic_constant): New.
2286 (msp430_expand_delay_cycles): New.
2287 (msp430_expand_builtin): Call it.
2288 (msp430_print_operand_raw): Change integer printing from "int" to
2290 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
2291 (delay_cycles_start): New.
2292 (delay_cycles_end): New.
2293 (delay_cycles_32): New.
2294 (delay_cycles_32x): New.
2295 (delay_cycles_16): New.
2296 (delay_cycles_16x): New.
2297 (delay_cycles_2): New.
2298 (delay_cycles_1): New.
2299 * doc/extend.texi: Document __delay_cycles().
2301 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
2303 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
2304 length attribute computation.
2306 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
2309 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
2311 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
2314 * config/sparc/sparc.md: Fix types of low and high in DI constant
2315 splitter. Use gen_int_mode in some other splitters.
2317 2014-05-14 Martin Jambor <mjambor@suse.cz>
2320 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
2322 2014-05-14 James Norris <jnorris@codesourcery.com>
2324 * omp-low.c (expand_parallel_call): Remove shadow variable.
2325 (expand_omp_taskreg): Likewise.
2327 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
2329 * common/config/i386/i386-common.c
2330 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
2331 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
2332 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
2333 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
2334 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
2335 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
2336 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
2337 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
2338 xsavecintrin.h, xsavesintrin.h.
2339 (x86_64-*-*): Ditto.
2340 * config/i386/clflushoptintrin.h: New.
2341 * config/i386/xsavecintrin.h: Ditto.
2342 * config/i386/xsavesintrin.h: Ditto.
2343 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
2344 (bit_XSAVES): Ditto.
2345 (bit_XSAVES): Ditto.
2346 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
2347 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
2349 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
2350 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
2351 OPTION_MASK_ISA_XSAVES.
2352 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
2354 (PTA_CLFLUSHOPT) Define.
2355 (PTA_XSAVEC): Ditto.
2356 (PTA_XSAVES): Ditto.
2357 (ix86_option_override_internal): Handle new options.
2358 (ix86_valid_target_attribute_inner_p): Ditto.
2359 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
2360 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
2361 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
2362 (bdesc_special_args): Add __builtin_ia32_xsaves,
2363 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
2364 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
2365 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
2366 (ix86_expand_builtin): Handle new builtins.
2367 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
2368 (TARGET_CLFLUSHOPT_P): Ditto.
2369 (TARGET_XSAVEC): Ditto.
2370 (TARGET_XSAVEC_P): Ditto.
2371 (TARGET_XSAVES): Ditto.
2372 (TARGET_XSAVES_P): Ditto.
2373 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
2374 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
2375 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
2377 (ANY_XRSTOR64): Ditto.
2379 (xrstor): Change into <xrstor>.
2380 (xrstor_rex64): Change into <xrstor>_rex64.
2381 (xrstor64): Change into <xrstor>64
2383 * config/i386/i386.opt (mclflushopt): New.
2386 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
2388 * doc/invoke.texi: Document new options.
2390 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
2392 PR rtl-optimization/60866
2393 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
2394 Default it to -1. Pass it down to init_simplejump_data.
2395 (init_simplejump_data): New parameter old_seqno. Pass it down
2396 to get_seqno_for_a_jump.
2397 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
2398 initializing new jump seqno as a last resort. Add comment.
2399 (sel_redirect_edge_and_branch): Save old seqno of the conditional
2400 jump and pass it down to sel_init_new_insn.
2401 (sel_redirect_edge_and_branch_force): Likewise.
2403 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
2405 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
2406 shifted values to avoid build warning.
2408 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
2410 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
2411 * cfgrtl.c (rtl_merge_blocks): Fix comment.
2412 (cfg_layout_merge_blocks): Likewise.
2413 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
2415 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
2417 PR rtl-optimization/60901
2418 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
2419 bb predecessor belongs to the same scheduling region. Adjust comment.
2421 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
2423 * doc/sourcebuild.texi: (dfp_hw): Document.
2424 (p8vector_hw): Likewise.
2425 (powerpc_eabi_ok): Likewise.
2426 (powerpc_elfv2): Likewise.
2427 (powerpc_htm_ok): Likewise.
2428 (ppc_recip_hw): Likewise.
2431 2014-05-13 Cary Coutant <ccoutant@google.com>
2433 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
2435 2014-05-13 David Malcolm <dmalcolm@redhat.com>
2437 * gengtype-parse.c (require3): Eliminate in favor of...
2439 (require_template_declaration): Update to support optional single *
2442 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
2443 (create_user_defined_type): Handle a single level of explicit
2444 pointerness within template arguments.
2445 (struct write_types_data): Add field "kind".
2446 (filter_type_name): Handle "*" character.
2447 (write_user_func_for_structure_ptr): Require a write_types_data
2448 rather than just a prefix string, so that we can look up the kind
2449 of the wtd and use it as an index into wrote_user_func_for_ptr,
2450 ensuring that such functions are written at most once. Support
2451 subclasses by invoking the marking function of the ultimate base class.
2452 (write_user_func_for_structure_body): Require a write_types_data
2453 rather than just a prefix string, so that we can pass this to
2454 write_user_func_for_structure_ptr.
2455 (write_func_for_structure): Likewise.
2456 (ggc_wtd): Add initializer of new "kind" field.
2457 (pch_wtd): Likewise.
2459 * gengtype.h (enum write_types_kinds): New.
2460 (struct type): Add field wrote_user_func_for_ptr to the "s"
2463 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
2465 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
2466 instead of const_binop.
2467 (fold_binary_loc): Likewise.
2469 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
2471 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
2472 calculation to match get_ref_base_and_extent.
2474 2014-05-13 Catherine Moore <clm@codesourcery.com>
2475 Sandra Loosemore <sandra@codesourcery.com>
2477 * configure.ac: Fix assembly for explicit JALR relocation check.
2478 * configure: Regenerate.
2480 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2482 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
2483 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
2484 Remove associated type declarations and initialisations.
2485 (arm_expand_neon_builtin): Likewise.
2486 (neon_emit_pair_result_insn): Delete.
2487 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
2488 * config/arm/neon.md (neon_vtrn<mode>): Delete.
2489 (neon_vzip<mode>): Likewise.
2490 (neon_vuzp<mode>): Likewise.
2492 2014-05-13 Richard Biener <rguenther@suse.de>
2495 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
2496 it needs revisiting whether the call still may be tail-called.
2498 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2500 * rtl.def (SYMBOL_REF): Remove middle "0" field.
2501 * rtl.h (block_symbol): Reduce number of fields to 2.
2502 (rtx_def): Add u2.symbol_ref_flags.
2503 (SYMBOL_REF_FLAGS): Use it.
2504 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
2505 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
2506 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
2507 Lower index of SYMBOL_REF_DATA.
2508 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
2509 Print SYMBOL_REF_FLAGS at the same time.
2510 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
2512 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2514 * rtl.def (VAR_LOCATION): Remove "i" field.
2515 * rtl.h (rtx_def): Add u2.var_location_status.
2516 (PAT_VAR_LOCATION_STATUS): Use it.
2517 (gen_rtx_VAR_LOCATION): Declare.
2518 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
2519 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
2520 * var-tracking.c (emit_note_insn_var_location): Remove casts.
2522 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2524 * rtl.def (scratch): Fix outdated comment and remove "0" field.
2525 * gengtype.c (adjust_field_rtx_def): Update accordingly.
2527 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2529 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
2530 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
2531 * rtl.h (rtx_def): Add insn_uid to u2 field.
2532 (RTX_FLAG_CHECK8): Delete in favor of...
2533 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
2534 (INSN_DELETED_P): Update accordingly.
2535 (INSN_UID): Use u2.insn_uid.
2536 (INSN_CHAIN_CODE_P): Define.
2537 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
2538 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
2539 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
2540 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
2541 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
2542 indices accordingly.
2543 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
2544 Update indices for insn-chain rtxes.
2545 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
2546 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
2547 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
2548 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
2549 * combine.c (try_combine): Likewise.
2550 * ira.c (setup_prohibited_mode_move_regs): Likewise.
2552 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2554 * rtl.def (REG): Remove middle field.
2555 * rtl.h (rtx_def): Add orignal_regno to u2.
2556 (ORIGINAL_REGNO): Use it instead of field 1.
2557 (REG_ATTRS): Lower field index accordingly.
2558 * gengtype.c (adjust_field_rtx_def): Remove handling of
2559 ORIGINAL_REGNO. Move REG_ATTRS index down.
2560 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
2561 code that prints the REGNO.
2563 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2565 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
2568 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2570 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
2572 2014-05-13 Bin Cheng <bin.cheng@arm.com>
2574 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
2575 (alloc_iv): Lower base expressions containing ADDR_EXPR.
2577 2014-05-13 Ian Bolton <ian.bolton@arm.com>
2579 * config/aarch64/aarch64-protos.h
2580 (aarch64_hard_regno_caller_save_mode): New prototype.
2581 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
2583 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
2585 2014-05-13 Christian Bruel <christian.bruel@st.com>
2587 * target.def (mode_switching): New hook vector.
2588 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
2589 (mode_exit, modepriority_to_mode): Likewise.
2590 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
2591 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2592 * target.h: Include tm.h and hard-reg-set.h.
2593 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
2594 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
2595 * doc/tm.texi Regenerate.
2596 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
2597 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2598 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
2599 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
2600 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
2601 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2602 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
2603 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
2604 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
2605 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
2606 (ix86_emit_mode_set): Hookify.
2607 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
2609 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2610 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
2611 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
2612 (epiphany_mode_priority_to_mode): Remove declaration.
2613 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
2614 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
2615 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
2617 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
2618 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
2619 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
2621 2014-05-13 Jakub Jelinek <jakub@redhat.com>
2624 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
2625 is const0_rtx, return immediately. Don't test count == 0 when
2628 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2630 * Makefile.in: add shrink-wrap.o.
2631 * config/i386/i386.c: include "shrink-wrap.h"
2632 * function.c: Likewise.
2633 (requires_stack_frame_p, next_block_for_reg,
2634 move_insn_for_shrink_wrap, prepare_shrink_wrap,
2635 dup_block_and_redirect): Move to shrink-wrap.c
2636 (thread_prologue_and_epilogue_insns): Extract three code segments
2637 as functions in shrink-wrap.c
2638 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
2640 * shrink-wrap.c: New file.
2641 * shrink-wrap.h: New file.
2643 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
2645 * doc/extend.texi: Reflect current numbers of pragmas. Remove
2646 reference to Solaris.
2648 2014-05-12 Mike Stump <mikestump@comcast.net>
2651 * genattrtab.c (filename): Add.
2652 (convert_set_attr_alternative): Improve error message.
2653 (check_defs): Restore read_md_filename for error messages.
2654 (gen_insn): Save filename.
2656 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
2658 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
2659 -fno-local-ivars and -fivar-visibility.
2660 * c-family/c.opt: Make -Wshadow also implicitly enable
2663 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
2665 * doc/tm.texi: Remove reference to deleted macro.
2666 * doc/tm.texi.in: Likewise.
2668 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
2671 * config/avr/avr.c (avr_out_store_psi): Use correct constant
2674 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
2677 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
2678 * config/arm/aout.h (License): Same.
2679 * config/arm/bpabi.h (License): Same.
2680 * config/arm/elf.h (License): Same.
2681 * config/arm/linux-elf.h (License): Same.
2682 * config/arm/linux-gas.h (License): Same.
2683 * config/arm/netbsd-elf.h (License): Same.
2684 * config/arm/uclinux-eabi.h (License): Same.
2685 * config/arm/uclinux-elf.h (License): Same.
2686 * config/arm/vxworks.h (License): Same.
2688 2014-05-11 Jakub Jelinek <jakub@redhat.com>
2690 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
2691 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
2692 number of operands to 3.
2693 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
2694 * tree-nested.c (convert_nonlocal_omp_clauses,
2695 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
2696 * gimplify.c (gimplify_scan_omp_clauses): Handle
2697 OMP_CLAUSE_LINEAR_STMT.
2698 * omp-low.c (lower_rec_input_clauses): Fix typo.
2699 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
2700 cast between Fortran boolean_type_node and C _Bool if
2703 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
2705 PR tree-optimization/61136
2706 * wide-int.h (multiple_of_p): Define a version that doesn't return
2708 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
2709 integer_zerop/const_binop pair.
2710 (multiple_of_p): Likewise, converting both operands to widest_int
2713 2014-05-09 Teresa Johnson <tejohnson@google.com>
2715 * cgraphunit.c (analyze_functions): Use correct dump file.
2717 2014-05-09 Florian Weimer <fweimer@redhat.com>
2719 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
2721 (stack_protect_return_slot_p): New function.
2722 (expand_used_vars): Call stack_protect_decl_p and
2723 stack_protect_return_slot_p for -fstack-protector-strong.
2725 2014-05-09 David Wohlferd <LimeGreenSocks@yahoo.com>
2726 Andrew Haley <aph@redhat.com>
2727 Richard Sandiford <rdsandiford@googlemail.com>
2729 * doc/extend.texi: Rewrite inline asm page / re-org asm-related
2732 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
2735 * fold-const.c (fold_binary_loc): Changed width of mask.
2737 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
2739 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
2740 unsigned int initializers for regno_in, regno_out.
2742 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
2745 * config/avr/avr.md (cc): Add new attribute set_vzn.
2746 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
2747 Set cc insn attribute to set_vzn instead of set_zn for alternatives
2748 with INC, DEC or NEG.
2749 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
2750 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
2751 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
2753 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2756 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2758 * wide-int.cc (UTItype): Define.
2759 (UDWtype): Define for appropriate W_TYPE_SIZE.
2761 2014-05-09 Richard Biener <rguenther@suse.de>
2763 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
2764 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
2765 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
2766 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
2767 ssa_propagate): Adjust.
2769 2014-05-08 Jeff Law <law@redhat.com>
2771 PR tree-optimization/61009
2772 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
2773 tri-state rather than a boolean. When a block is too big to
2774 thread through, inform caller via negative return value.
2775 (thread_across_edge): If a block was too big for normal threading,
2776 then it's too big for a joiner too, so remove temporary equivalences
2777 and return immediately.
2779 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
2780 Matthias Klose <doko@ubuntu.com>
2783 * optc-gen.awk: Fix option handling for -Wunused-parameter.
2785 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
2788 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
2790 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
2793 * config/alpha/alpha.c: Include gimple-iterator.h.
2794 (alpha_gimple_fold_builtin): New function. Move
2795 ALPHA_BUILTIN_UMULH folding from ...
2796 (alpha_fold_builtin): ... here.
2797 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
2799 2014-05-08 Wei Mi <wmi@google.com>
2802 * config/i386/i386.c (ix86_compute_frame_layout): Update
2803 preferred_stack_boundary for call, expanded from tls descriptor.
2804 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
2805 to depend on SP register.
2806 (*tls_local_dynamic_base_32_gnu): Ditto.
2807 (*tls_local_dynamic_32_once): Ditto.
2808 (tls_global_dynamic_64_<mode>): Set
2809 ix86_tls_descriptor_calls_expanded_in_cfun.
2810 (tls_local_dynamic_base_64_<mode>): Ditto.
2811 (tls_global_dynamic_32): Set
2812 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
2813 to depend on SP register.
2814 (tls_local_dynamic_base_32): Ditto.
2816 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2818 * config/arm/arm_neon.h: Update comment.
2819 * config/arm/neon-docgen.ml: Delete.
2820 * config/arm/neon-gen.ml: Delete.
2821 * doc/arm-neon-intrinsics.texi: Update comment.
2823 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2825 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
2827 (vand, vorr, veor, vorn, vbic): Remove.
2828 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
2830 (neon_vsub_unspec): Likewise.
2831 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
2833 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2835 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
2836 (vadd_s16): Likewise.
2837 (vadd_s32): Likewise.
2838 (vadd_f32): Likewise.
2839 (vadd_u8): Likewise.
2840 (vadd_u16): Likewise.
2841 (vadd_u32): Likewise.
2842 (vadd_s64): Likewise.
2843 (vadd_u64): Likewise.
2844 (vaddq_s8): Likewise.
2845 (vaddq_s16): Likewise.
2846 (vaddq_s32): Likewise.
2847 (vaddq_s64): Likewise.
2848 (vaddq_f32): Likewise.
2849 (vaddq_u8): Likewise.
2850 (vaddq_u16): Likewise.
2851 (vaddq_u32): Likewise.
2852 (vaddq_u64): Likewise.
2853 (vmul_s8): Likewise.
2854 (vmul_s16): Likewise.
2855 (vmul_s32): Likewise.
2856 (vmul_f32): Likewise.
2857 (vmul_u8): Likewise.
2858 (vmul_u16): Likewise.
2859 (vmul_u32): Likewise.
2860 (vmul_p8): Likewise.
2861 (vmulq_s8): Likewise.
2862 (vmulq_s16): Likewise.
2863 (vmulq_s32): Likewise.
2864 (vmulq_f32): Likewise.
2865 (vmulq_u8): Likewise.
2866 (vmulq_u16): Likewise.
2867 (vmulq_u32): Likewise.
2868 (vsub_s8): Likewise.
2869 (vsub_s16): Likewise.
2870 (vsub_s32): Likewise.
2871 (vsub_f32): Likewise.
2872 (vsub_u8): Likewise.
2873 (vsub_u16): Likewise.
2874 (vsub_u32): Likewise.
2875 (vsub_s64): Likewise.
2876 (vsub_u64): Likewise.
2877 (vsubq_s8): Likewise.
2878 (vsubq_s16): Likewise.
2879 (vsubq_s32): Likewise.
2880 (vsubq_s64): Likewise.
2881 (vsubq_f32): Likewise.
2882 (vsubq_u8): Likewise.
2883 (vsubq_u16): Likewise.
2884 (vsubq_u32): Likewise.
2885 (vsubq_u64): Likewise.
2886 (vand_s8): Likewise.
2887 (vand_s16): Likewise.
2888 (vand_s32): Likewise.
2889 (vand_u8): Likewise.
2890 (vand_u16): Likewise.
2891 (vand_u32): Likewise.
2892 (vand_s64): Likewise.
2893 (vand_u64): Likewise.
2894 (vandq_s8): Likewise.
2895 (vandq_s16): Likewise.
2896 (vandq_s32): Likewise.
2897 (vandq_s64): Likewise.
2898 (vandq_u8): Likewise.
2899 (vandq_u16): Likewise.
2900 (vandq_u32): Likewise.
2901 (vandq_u64): Likewise.
2902 (vorr_s8): Likewise.
2903 (vorr_s16): Likewise.
2904 (vorr_s32): Likewise.
2905 (vorr_u8): Likewise.
2906 (vorr_u16): Likewise.
2907 (vorr_u32): Likewise.
2908 (vorr_s64): Likewise.
2909 (vorr_u64): Likewise.
2910 (vorrq_s8): Likewise.
2911 (vorrq_s16): Likewise.
2912 (vorrq_s32): Likewise.
2913 (vorrq_s64): Likewise.
2914 (vorrq_u8): Likewise.
2915 (vorrq_u16): Likewise.
2916 (vorrq_u32): Likewise.
2917 (vorrq_u64): Likewise.
2918 (veor_s8): Likewise.
2919 (veor_s16): Likewise.
2920 (veor_s32): Likewise.
2921 (veor_u8): Likewise.
2922 (veor_u16): Likewise.
2923 (veor_u32): Likewise.
2924 (veor_s64): Likewise.
2925 (veor_u64): Likewise.
2926 (veorq_s8): Likewise.
2927 (veorq_s16): Likewise.
2928 (veorq_s32): Likewise.
2929 (veorq_s64): Likewise.
2930 (veorq_u8): Likewise.
2931 (veorq_u16): Likewise.
2932 (veorq_u32): Likewise.
2933 (veorq_u64): Likewise.
2934 (vbic_s8): Likewise.
2935 (vbic_s16): Likewise.
2936 (vbic_s32): Likewise.
2937 (vbic_u8): Likewise.
2938 (vbic_u16): Likewise.
2939 (vbic_u32): Likewise.
2940 (vbic_s64): Likewise.
2941 (vbic_u64): Likewise.
2942 (vbicq_s8): Likewise.
2943 (vbicq_s16): Likewise.
2944 (vbicq_s32): Likewise.
2945 (vbicq_s64): Likewise.
2946 (vbicq_u8): Likewise.
2947 (vbicq_u16): Likewise.
2948 (vbicq_u32): Likewise.
2949 (vbicq_u64): Likewise.
2950 (vorn_s8): Likewise.
2951 (vorn_s16): Likewise.
2952 (vorn_s32): Likewise.
2953 (vorn_u8): Likewise.
2954 (vorn_u16): Likewise.
2955 (vorn_u32): Likewise.
2956 (vorn_s64): Likewise.
2957 (vorn_u64): Likewise.
2958 (vornq_s8): Likewise.
2959 (vornq_s16): Likewise.
2960 (vornq_s32): Likewise.
2961 (vornq_s64): Likewise.
2962 (vornq_u8): Likewise.
2963 (vornq_u16): Likewise.
2964 (vornq_u32): Likewise.
2965 (vornq_u64): Likewise.
2967 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2969 * wide-int.cc (UTItype): Define.
2970 (UDWtype): Define for appropriate W_TYPE_SIZE.
2972 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
2974 PR tree-optimization/59100
2975 * tree-ssa-phiopt.c: Include tree-inline.h.
2976 (neutral_element_p, absorbing_element_p): New functions.
2977 (value_replacement): Handle conditional binary operations with a
2978 neutral or absorbing element.
2980 2014-05-08 Richard Biener <rguenther@suse.de>
2982 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
2983 folding the expression.
2984 (valueize_expr): Remove.
2985 (visit_reference_op_load): Do not valueize the result of
2987 (simplify_binary_expression): Likewise.
2988 (simplify_unary_expression): Likewise.
2990 2014-05-08 Richard Biener <rguenther@suse.de>
2992 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
2993 looking at TYPE_ARG_TYPES.
2995 2014-05-08 Richard Biener <rguenther@suse.de>
2997 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
2998 pointer propagation special-case.
3000 2014-05-08 Bin Cheng <bin.cheng@arm.com>
3002 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
3003 core part of address expressions.
3005 2014-05-08 Alan Modra <amodra@gmail.com>
3008 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
3009 loads and stores when -mno-strict-align at any alignment.
3010 (expand_block_clear): Similarly. Also correct calculation of
3013 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
3016 * tree-complex.c (expand_complex_move): Keep line info when expanding
3018 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
3019 of complex expression. Use new argument to display correct location
3020 for values coming from phi statement.
3021 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
3022 (warn_uninitialized_phi): Pass location of phi argument to
3024 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
3025 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
3027 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
3029 * config/rs6000/predicates.md (indexed_address_mem): New.
3030 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
3031 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
3032 fpstore_ux, fpstore_u.
3033 (sign_extend, indexed, update): New.
3034 (cell_micro): Adjust.
3035 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
3036 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
3037 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
3038 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
3039 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
3040 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
3041 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
3042 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
3043 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
3044 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
3045 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
3046 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
3047 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
3048 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
3049 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
3051 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
3052 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
3053 *vsx_extract_<mode>_store): Adjust.
3054 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
3055 is_cracked_insn, insn_must_be_first_in_group,
3056 insn_must_be_last_in_group): Adjust.
3058 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
3060 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
3061 ppc440-fpstore): Adjust.
3062 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
3063 ppc476-fpstore): Adjust.
3064 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
3065 ppc601-fpstore): Adjust.
3066 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
3068 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
3070 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
3071 ppc7450-fpstore): Adjust.
3072 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
3073 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
3074 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
3076 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
3077 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
3078 cell-fpstore, cell-fpstore-update): Adjust.
3079 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
3080 ppce300c3_store, ppce300c3_fpstore): Adjust.
3081 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
3082 e500mc_fpstore): Adjust.
3083 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
3084 e500mc64_store, e500mc64_fpstore): Adjust.
3085 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
3086 e5500_fpstore): Adjust.
3087 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
3088 e6500_fpstore): Adjust.
3089 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
3091 * config/rs6000/power4.md (power4-load, power4-load-ext,
3092 power4-load-ext-update, power4-load-ext-update-indexed,
3093 power4-load-update-indexed, power4-load-update, power4-fpload,
3094 power4-fpload-update, power4-store, power4-store-update,
3095 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
3097 * config/rs6000/power5.md (power5-load, power5-load-ext,
3098 power5-load-ext-update, power5-load-ext-update-indexed,
3099 power5-load-update-indexed, power5-load-update, power5-fpload,
3100 power5-fpload-update, power5-store, power5-store-update,
3101 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
3103 * config/rs6000/power6.md (power6-load, power6-load-ext,
3104 power6-load-update, power6-load-update-indexed,
3105 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
3106 power6-fpload-update, power6-store, power6-store-update,
3107 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
3109 * config/rs6000/power7.md (power7-load, power7-load-ext,
3110 power7-load-update, power7-load-update-indexed,
3111 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
3112 power7-fpload-update, power7-store, power7-store-update,
3113 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
3115 * config/rs6000/power8.md (power8-load, power8-load-update,
3116 power8-load-ext, power8-load-ext-update, power8-fpload,
3117 power8-fpload-update, power8-store, power8-store-update-indexed,
3118 power8-fpstore, power8-fpstore-update): Adjust.
3119 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
3121 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
3122 titan_lsu_store, titan_lsu_fpstore): Adjust.
3123 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
3125 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
3128 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
3129 unrolled byte insns. Emit address increments after move insns.
3131 2014-05-07 David Malcolm <dmalcolm@redhat.com>
3133 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
3134 const_gimple, rather than a gimple.
3135 (gimple_call_builtin_p): Likewise, for the three variants.
3137 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
3138 (gimple_call_builtin_p): Likewise, for the three variants.
3140 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
3142 PR tree-optimization/61095
3143 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
3145 2014-05-07 Richard Biener <rguenther@suse.de>
3147 PR tree-optimization/61034
3148 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
3149 (maybe_skip_until): Use translate to take into account
3150 lattices when trying to do disambiguations.
3151 (get_continuation_for_phi_1): Likewise.
3152 (get_continuation_for_phi): Adjust for added translate arguments.
3153 (walk_non_aliased_vuses): Likewise.
3154 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
3155 (walk_non_aliased_vuses): Likewise.
3156 (call_may_clobber_ref_p_1): Declare.
3157 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
3158 calls. Stop early if we are only supposed to disambiguate.
3159 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
3161 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
3163 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
3164 Emit an error when the function has arguments.
3166 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
3168 * cfgloop.h (unswitch_loops): Remove.
3169 * doc/passes.texi: Remove references to loop-unswitch.c
3170 * timevar.def (TV_LOOP_UNSWITCH): Remove.
3172 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
3174 * tree-vect-data-refs.c (vect_grouped_load_supported): New
3175 check for loads group of length 3.
3176 (vect_permute_load_chain): New permutations for loads group of
3178 * tree-vect-stmts.c (vect_model_load_cost): Change cost
3179 of vec_perm_shuffle for the new permutations.
3181 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
3183 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
3184 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
3185 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
3186 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
3187 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
3188 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
3189 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
3190 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
3192 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
3194 * loop-unswitch.c: Delete.
3196 2014-05-07 Richard Biener <rguenther@suse.de>
3198 * config.gcc: Always set need_64bit_hwint to yes.
3200 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
3202 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
3203 of using optimize_size.
3205 2014-05-06 Mike Stump <mikestump@comcast.net>
3207 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
3209 2014-05-06 Joseph Myers <joseph@codesourcery.com>
3211 * config/i386/sse.md (*mov<mode>_internal)
3212 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
3213 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
3214 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
3215 (*<code><mode>3, *andnot<mode>3<mask_name>)
3216 (<mask_codefor><code><mode>3<mask_name>): Only consider
3217 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
3219 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
3222 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3224 * lra-constraints.c (valid_address_p): Move earlier in file.
3225 Add a constraint argument to the address_info version.
3226 (satisfies_memory_constraint_p): New function.
3227 (satisfies_address_constraint_p): Likewise.
3228 (process_alt_operands, curr_insn_transform): Use them.
3229 (process_address): Pass the constraint to valid_address_p when
3230 checking address operands.
3232 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
3234 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
3235 to their respective blocks. Fix inadvertent use of "node".
3237 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
3239 * emit-rtl.c (init_derived_machine_modes): New functionm, split
3241 (init_emit_once): ...here.
3242 * rtl.h (init_derived_machine_modes): Declare.
3243 * toplev.c (do_compile): Call it even if no_backend.
3245 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
3246 Mike Stump <mikestump@comcast.net>
3247 Richard Sandiford <rdsandiford@googlemail.com>
3248 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3250 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
3251 (rtx_equal_for_memref_p): Update comment.
3252 (adjust_offset_for_component_ref): Use wide-int interfaces.
3253 * builtins.c (get_object_alignment_2): Likewise.
3254 (c_readstr): Likewise.
3255 (target_char_cast): Add comment.
3256 (determine_block_size): Use wide-int interfaces.
3257 (expand_builtin_signbit): Likewise.
3258 (fold_builtin_int_roundingfn): Likewise.
3259 (fold_builtin_bitop): Likewise.
3260 (fold_builtin_bswap): Likewise.
3261 (fold_builtin_logarithm): Use signop.
3262 (fold_builtin_pow): Likewise.
3263 (fold_builtin_memory_op): Use wide-int interfaces.
3264 (fold_builtin_object_size): Likewise.
3265 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
3266 nb_iterations_estimate.
3267 (record_niter_bound): Use wide-int interfaces.
3268 (get_estimated_loop_iterations_int): Likewise.
3269 (get_estimated_loop_iterations): Likewise.
3270 (get_max_loop_iterations): Likewise.
3271 * cfgloop.h: Include wide-int.h.
3272 (struct nb_iter_bound): Change bound to widest_int.
3273 (struct loop): Change nb_iterations_upper_bound and
3274 nb_iterations_estimate to widest_int.
3275 (record_niter_bound): Switch to use widest_int.
3276 (get_estimated_loop_iterations): Likewise.
3277 (get_max_loop_iterations): Likewise.
3278 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
3279 update for wide-int.
3280 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
3281 * combine.c (try_combine): Likewise.
3282 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
3283 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
3285 (aarch64_float_const_representable_p): Likewise.
3286 * config/arc/arc.c: Include wide-int.h.
3287 (arc_can_use_doloop_p): Use wide-int interfaces.
3288 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
3289 (vfp3_const_double_index): Likewise.
3290 * config/avr/avr.c (avr_out_round): Likewise.
3291 (avr_fold_builtin): Likewise.
3292 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
3293 (bfin_can_use_doloop_p): Likewise.
3294 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
3295 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
3296 * config/i386/i386.c: Include wide-int.h.
3297 (ix86_data_alignment): Use wide-int interfaces.
3298 (ix86_local_alignment): Likewise.
3299 (ix86_emit_swsqrtsf): Update real_from_integer.
3300 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
3301 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
3302 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
3303 (zero_constant): Likewise.
3304 (input_operand): Likewise.
3305 (splat_input_operand): Likewise.
3306 (non_logical_cint_operand): Change const_double to const_wide_int.
3307 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
3308 (easy_altivec_constant): Remove comment.
3309 (paired_expand_vector_init): Use CONSTANT_P.
3310 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
3311 (rs6000_emit_move): Update checks.
3312 (rs6000_aggregate_candidate): Use wide-int interfaces.
3313 (rs6000_expand_ternop_builtin): Likewise.
3314 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
3315 (rs6000_assemble_integer): Likewise.
3316 (rs6000_hash_constant): Likewise.
3317 (output_toc): Likewise.
3318 (rs6000_rtx_costs): Likewise.
3319 (rs6000_emit_swrsqrt); Update call to real_from_integer.
3320 * config/rs6000/rs6000-c.c: Include wide-int.h.
3321 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
3322 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
3323 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
3324 Handle CONST_WIDE_INT.
3325 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
3326 Use tree_fits_uhwi_p.
3327 * config/sparc/sparc.c: Include wide-int.h.
3328 (sparc_fold_builtin): Use wide-int interfaces.
3329 * config/vax/vax.c: Include wide-int.h.
3330 (vax_float_literal): Use real_from_integer.
3331 * coretypes.h (struct hwivec_def): New.
3333 (const_hwivec): New.
3334 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
3335 (equiv_constant): Handle CONST_WIDE_INT.
3336 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
3337 (cselib_hash_rtx): Handle CONST_WIDE_INT.
3338 * dbxout.c (stabstr_U): Use wide-int interfaces.
3339 (dbxout_type): Update to use cst_fits_shwi_p.
3340 * defaults.h (LOG2_BITS_PER_UNIT): Define.
3341 (TARGET_SUPPORTS_WIDE_INT): Add default.
3342 * dfp.c: Include wide-int.h.
3343 (decimal_real_to_integer2): Use wide-int interfaces and rename to
3344 decimal_real_to_integer.
3345 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
3346 decimal_real_to_integer.
3347 * doc/generic.texi (Constant expressions): Update for wide_int.
3348 * doc/rtl.texi (const_double): Likewise.
3349 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
3350 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
3351 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
3352 (REAL_VALUE_FROM_INT): Remove.
3353 (TARGET_SUPPORTS_WIDE_INT): New.
3354 * doc/tm.texi: Regenerate.
3355 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
3356 * double-int.h: Include wide-int.h.
3357 (struct wi::int_traits): New.
3358 * dwarf2out.c (get_full_len): New.
3359 (dw_val_equal_p): Add case dw_val_class_wide_int.
3360 (size_of_loc_descr): Likewise.
3361 (output_loc_operands): Likewise.
3362 (insert_double): Remove.
3363 (insert_wide_int): New.
3365 (print_die): Add case dw_val_class_wide_int.
3366 (attr_checksum): Likewise.
3367 (attr_checksum_ordered): Likewise.
3368 (same_dw_val_p): Likewise.
3369 (size_of_die): Likewise.
3370 (value_format): Likewise.
3371 (output_die): Likewise.
3372 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
3374 (clz_loc_descriptor): Use wide-int interfaces.
3375 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
3376 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
3377 (round_up_to_align): Use wide-int interfaces.
3378 (field_byte_offset): Likewise.
3379 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
3380 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
3381 CONST_DOUBLE handling. Use wide-int interfaces.
3382 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
3383 (gen_enumeration_type_die): Use add_AT_wide.
3384 (hash_loc_operands): Add case dw_val_class_wide_int.
3385 (compare_loc_operands): Likewise.
3386 * dwarf2out.h: Include wide-int.h.
3387 (wide_int_ptr): New.
3388 (enum dw_val_class): Add dw_val_class_wide_int.
3389 (struct dw_val_struct): Add val_wide.
3390 * emit-rtl.c (const_wide_int_htab): New.
3391 (const_wide_int_htab_hash): New.
3392 (const_wide_int_htab_eq): New.
3393 (lookup_const_wide_int): New.
3394 (const_double_htab_hash): Use wide-int interfaces.
3395 (const_double_htab_eq): Likewise.
3396 (rtx_to_double_int): Conditionally compile for wide-int.
3397 (immed_double_int_const): Rename to immed_wide_int_const and
3398 update for wide-int.
3399 (immed_double_const): Conditionally compile for wide-int.
3400 (init_emit_once): Use wide-int interfaces.
3401 * explow.c (plus_constant): Likewise.
3402 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
3403 (lshift_value): Use wide-int interfaces.
3404 (expand_mult): Likewise.
3405 (choose_multiplier): Likewise.
3406 (expand_smod_pow2): Likewise.
3407 (make_tree): Likewise.
3408 * expr.c (convert_modes): Consolidate handling of constants.
3409 Use wide-int interfaces.
3410 (emit_group_load_1): Add note.
3411 (store_expr): Update comment.
3412 (get_inner_reference): Use wide-int interfaces.
3413 (expand_constructor): Update comment.
3414 (expand_expr_real_2): Use wide-int interfaces.
3415 (expand_expr_real_1): Likewise.
3416 (reduce_to_bit_field_precision): Likewise.
3417 (const_vector_from_tree): Likewise.
3418 * final.c: Include wide-int-print.h.
3419 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
3420 * fixed-value.c: Include wide-int.h.
3421 (fixed_from_string): Use wide-int interfaces.
3422 (fixed_to_decimal): Likewise.
3423 (fixed_convert_from_real): Likewise.
3424 (real_convert_from_fixed): Likewise.
3425 * fold-const.h (mem_ref_offset): Return an offset_int.
3426 (div_if_zero_remainder): Remove code parameter.
3427 * fold-const.c (div_if_zero_remainder): Remove code parameter.
3428 Use wide-int interfaces.
3429 (may_negate_without_overflow_p): Use wide-int interfaces.
3430 (negate_expr_p): Likewise.
3431 (fold_negate_expr): Likewise.
3432 (int_const_binop_1): Likewise.
3433 (const_binop): Likewise.
3434 (fold_convert_const_int_from_int): Likewise.
3435 (fold_convert_const_int_from_real): Likewise.
3436 (fold_convert_const_int_from_fixed): Likewise.
3437 (fold_convert_const_fixed_from_int): Likewise.
3438 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
3439 (sign_bit_p): Use wide-int interfaces.
3440 (make_range_step): Likewise.
3441 (build_range_check): Likewise. Pass an integer of the correct type
3442 instead of using integer_one_node.
3443 (range_predecessor): Pass an integer of the correct type instead
3444 of using integer_one_node.
3445 (range_successor): Likewise.
3446 (merge_ranges): Likewise.
3447 (unextend): Use wide-int interfaces.
3448 (extract_muldiv_1): Likewise.
3449 (fold_div_compare): Likewise.
3450 (fold_single_bit_test): Likewise.
3451 (fold_sign_changed_comparison): Likewise.
3452 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
3453 (fold_plusminus_mult_expr): Use wide-int interfaces.
3454 (native_encode_int): Likewise.
3455 (native_interpret_int): Likewise.
3456 (fold_unary_loc): Likewise.
3457 (pointer_may_wrap_p): Likewise.
3458 (size_low_cst): Likewise.
3459 (mask_with_tz): Likewise.
3460 (fold_binary_loc): Likewise.
3461 (fold_ternary_loc): Likewise.
3462 (multiple_of_p): Likewise.
3463 (tree_call_nonnegative_warnv_p): Update calls to
3464 tree_int_cst_min_precision and real_from_integer.
3465 (fold_negate_const): Use wide-int interfaces.
3466 (fold_abs_const): Likewise.
3467 (fold_relational_const): Use tree_int_cst_lt.
3468 (round_up_loc): Use wide-int interfaces.
3469 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
3470 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
3471 * gengtype.c: Remove include of double-int.h.
3472 (do_typedef): Use wide-int interfaces.
3473 (open_base_files): Add wide-int.h.
3474 (main): Add offset_int and widest_int typedefs.
3475 * gengtype-lex.l: Handle "^".
3476 (CXX_KEYWORD): Add "static".
3477 * gengtype-parse.c (require3): New.
3478 (require_template_declaration): Handle constant template arguments
3479 and nested templates.
3480 * gengtype-state.c: Don't include "double-int.h".
3481 * genpreds.c (write_one_predicate_function): Update comment.
3482 (write_tm_constrs_h): Add check for hval and lval use in
3484 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
3485 (add_to_sequence): Likewise.
3486 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
3487 and const_double_operand.
3488 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
3490 * gimple-fold.c (get_base_constructor): Likewise.
3491 (fold_array_ctor_reference): Likewise.
3492 (fold_nonarray_ctor_reference): Likewise.
3493 (fold_const_aggregate_ref_1): Likewise.
3494 (gimple_val_nonnegative_real_p): Likewise.
3495 (gimple_fold_indirect_ref): Likewise.
3496 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
3497 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
3498 (struct slsr_cand_d): Change index to be widest_int.
3499 (struct incr_info_d): Change incr to be widest_int.
3500 (alloc_cand_and_find_basis): Use wide-int interfaces.
3501 (slsr_process_phi): Likewise.
3502 (backtrace_base_for_ref): Likewise. Return a widest_int.
3503 (restructure_reference): Take a widest_int instead of a double_int.
3504 (slsr_process_ref): Use wide-int interfaces.
3505 (create_mul_ssa_cand): Likewise.
3506 (create_mul_imm_cand): Likewise.
3507 (create_add_ssa_cand): Likewise.
3508 (create_add_imm_cand): Take a widest_int instead of a double_int.
3509 (slsr_process_add): Use wide-int interfaces.
3510 (slsr_process_cast): Likewise.
3511 (slsr_process_copy): Likewise.
3512 (dump_candidate): Likewise.
3513 (dump_incr_vec): Likewise.
3514 (replace_ref): Likewise.
3515 (cand_increment): Likewise. Return a widest_int.
3516 (cand_abs_increment): Likewise.
3517 (replace_mult_candidate): Take a widest_int instead of a double_int.
3518 (replace_unconditional_candidate): Use wide-int interfaces.
3519 (incr_vec_index): Take a widest_int instead of a double_int.
3520 (create_add_on_incoming_edge): Likewise.
3521 (create_phi_basis): Use wide-int interfaces.
3522 (replace_conditional_candidate): Likewise.
3523 (record_increment): Take a widest_int instead of a double_int.
3524 (record_phi_increments): Use wide-int interfaces.
3525 (phi_incr_cost): Take a widest_int instead of a double_int.
3526 (lowest_cost_path): Likewise.
3527 (total_savings): Likewise.
3528 (analyze_increments): Use wide-int interfaces.
3529 (ncd_with_phi): Take a widest_int instead of a double_int.
3530 (ncd_of_cand_and_phis): Likewise.
3531 (nearest_common_dominator_for_cands): Likewise.
3532 (insert_initializers): Use wide-int interfaces.
3533 (all_phi_incrs_profitable): Likewise.
3534 (replace_one_candidate): Likewise.
3535 (replace_profitable_candidates): Likewise.
3536 * godump.c: Include wide-int-print.h.
3537 (go_output_typedef): Use wide-int interfaces.
3538 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
3539 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
3540 (build_loop_iteration_domains): Likewise.
3541 * hooks.h: Include wide-int.h rather than double-int.h.
3542 (hook_bool_dint_dint_uint_bool_true): Delete.
3543 (hook_bool_wint_wint_uint_bool_true): Declare.
3544 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
3545 (hook_bool_wint_wint_uint_bool_true): New.
3546 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
3548 (ubsan_expand_si_overflow_mul_check): Likewise.
3549 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
3550 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
3551 (get_ancestor_addr_info): Likewise.
3552 (ipa_modify_call_arguments): Likewise.
3553 * loop-doloop.c (doloop_modify): Likewise.
3554 (doloop_optimize): Likewise.
3555 * loop-iv.c (iv_number_of_iterations): Likewise.
3556 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
3557 (unroll_loop_constant_iterations): Likewise.
3558 (decide_unroll_runtime_iterations): Likewise.
3559 (unroll_loop_runtime_iterations): Likewise.
3560 (decide_peel_simple): Likewise.
3561 (decide_unroll_stupid): Likewise.
3562 * lto-streamer-in.c (streamer_read_wi): Add.
3563 (input_cfg): Use wide-int interfaces.
3564 (lto_input_tree_1): Likewise.
3565 * lto-streamer-out.c (streamer_write_wi): Add.
3566 (hash_tree): Use wide-int interfaces.
3567 (output_cfg): Likewise.
3568 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
3569 (GTFILES): Add wide-int.h and signop.h.
3570 (TAGS): Look for .cc files too.
3571 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
3572 * optabs.c (expand_subword_shift): Likewise.
3573 (expand_doubleword_shift): Likewise.
3574 (expand_absneg_bit): Likewise.
3575 (expand_copysign_absneg): Likewise.
3576 (expand_copysign_bit): Likewise.
3577 * postreload.c (reload_cse_simplify_set): Likewise.
3578 * predict.c (predict_iv_comparison): Likewise.
3579 * pretty-print.h: Include wide-int-print.h.
3581 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
3582 * print-tree.c: Include wide-int-print.h.
3583 (print_node_brief): Use wide-int interfaces.
3584 (print_node): Likewise.
3585 * read-rtl.c (validate_const_wide_int): New.
3586 (read_rtx_code): Add CONST_WIDE_INT case.
3587 * real.c: Include wide-int.h.
3588 (real_to_integer2): Delete.
3589 (real_to_integer): New function, returning a wide_int.
3590 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
3591 (ten_to_ptwo): Update call to real_from_integer.
3592 (real_digit): Likewise.
3593 * real.h: Include signop.h, wide-int.h and insn-modes.h.
3594 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
3595 (REAL_VALUE_TO_INT): Delete.
3596 (real_to_integer): Declare a wide-int form.
3597 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
3598 * recog.c (const_int_operand): Improve comment.
3599 (const_scalar_int_operand): New.
3600 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
3601 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
3602 (split_double): Likewise.
3603 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
3604 (rtx_size): Likewise.
3605 (rtx_alloc_stat_v): New.
3606 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
3607 (cwi_output_hex): New.
3608 (iterative_hash_rtx): Handle CONST_WIDE_INT.
3609 (cwi_check_failed_bounds): New.
3610 * rtl.def (CONST_WIDE_INT): New.
3611 * rtl.h: Include <utility> and wide-int.h.
3612 (struct hwivec_def): New.
3613 (CWI_GET_NUM_ELEM): New.
3614 (CWI_PUT_NUM_ELEM): New.
3615 (struct rtx_def): Add num_elem and hwiv.
3616 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
3617 (CASE_CONST_UNIQUE): Likewise.
3618 (CASE_CONST_ANY): Likewise.
3619 (CONST_SCALAR_INT_P): Likewise.
3620 (CONST_WIDE_INT_P): New.
3622 (HWIVEC_CHECK): New.
3623 (cwi_check_failed_bounds): New.
3625 (HWIVEC_CHECK): New.
3626 (CONST_WIDE_INT_VEC) New.
3627 (CONST_WIDE_INT_NUNITS) New.
3628 (CONST_WIDE_INT_ELT) New.
3629 (rtx_mode_t): New type.
3630 (wi::int_traits <rtx_mode_t>): New.
3632 (wi::min_value): New.
3633 (wi::max_value): New.
3635 (const_wide_int_alloc): New.
3636 (immed_wide_int_const): New.
3637 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
3638 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
3639 * signop.h: New file.
3640 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
3641 (simplify_const_unary_operation): Use wide-int interfaces.
3642 (simplify_binary_operation_1): Likewise.
3643 (simplify_const_binary_operation): Likewise.
3644 (simplify_const_relational_operation): Likewise.
3645 (simplify_immed_subreg): Likewise.
3646 * stmt.c (expand_case): Likewise.
3647 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
3648 signop rather than a bool.
3649 * stor-layout.c (layout_type): Use wide-int interfaces.
3650 (initialize_sizetypes): Update calls to
3651 set_min_and_max_values_for_integral_type.
3652 (set_min_and_max_values_for_integral_type): Take a signop rather
3653 than a bool. Use wide-int interfaces.
3654 (fixup_signed_type): Update accordingly. Remove
3655 HOST_BITS_PER_DOUBLE_INT limit.
3656 (fixup_unsigned_type): Likewise.
3657 * system.h (STATIC_CONSTANT_P): New.
3658 (STATIC_ASSERT): New.
3659 * target.def (can_use_doloop_p): Take widest_ints rather than
3661 * target.h: Include wide-int.h rather than double-int.h.
3662 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
3664 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
3665 rather than INT_CST_LT_UNSIGNED.
3666 (can_use_doloop_if_innermost): Take widest_ints rather than
3668 * tree-affine.c: Include wide-int-print.h.
3669 (double_int_ext_for_comb): Delete.
3670 (wide_int_ext_for_comb): New.
3671 (aff_combination_zero): Use wide-int interfaces.
3672 (aff_combination_const): Take a widest_int instead of a double_int.
3673 (aff_combination_elt): Use wide-int interfaces.
3674 (aff_combination_scale): Take a widest_int instead of a double_int.
3675 (aff_combination_add_elt): Likewise.
3676 (aff_combination_add_cst): Likewise.
3677 (aff_combination_add): Use wide-int interfaces.
3678 (aff_combination_convert): Likewise.
3679 (tree_to_aff_combination): Likewise.
3680 (add_elt_to_tree): Take a widest_int instead of a double_int.
3681 (aff_combination_to_tree): Use wide-int interfaces.
3682 (aff_combination_remove_elt): Likewise.
3683 (aff_combination_add_product): Take a widest_int instead of
3685 (aff_combination_mult): Use wide-int interfaces.
3686 (aff_combination_expand): Likewise.
3687 (double_int_constant_multiple_p): Delete.
3688 (wide_int_constant_multiple_p): New.
3689 (aff_combination_constant_multiple_p): Take a widest_int pointer
3690 instead of a double_int pointer.
3691 (print_aff): Use wide-int interfaces.
3692 (get_inner_reference_aff): Take a widest_int pointer
3693 instead of a double_int pointer.
3694 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
3695 * tree-affine.h: Include wide-int.h.
3696 (struct aff_comb_elt): Change type of coef to widest_int.
3697 (struct affine_tree_combination): Change type of offset to widest_int.
3698 (double_int_ext_for_comb): Delete.
3699 (wide_int_ext_for_comb): New.
3700 (aff_combination_const): Use widest_int instead of double_int.
3701 (aff_combination_scale): Likewise.
3702 (aff_combination_add_elt): Likewise.
3703 (aff_combination_constant_multiple_p): Likewise.
3704 (get_inner_reference_aff): Likewise.
3705 (aff_comb_cannot_overlap_p): Likewise.
3706 (aff_combination_zero_p): Use wide-int interfaces.
3707 * tree.c: Include tree.h.
3708 (init_ttree): Use make_int_cst.
3709 (tree_code_size): Removed code for INTEGER_CST case.
3710 (tree_size): Add INTEGER_CST case.
3711 (make_node_stat): Update comment.
3712 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
3713 (build_int_cst_type): Use wide-int interfaces.
3714 (double_int_to_tree): Likewise.
3715 (double_int_fits_to_tree_p): Delete.
3716 (force_fit_type_double): Delete.
3717 (force_fit_type): New.
3718 (int_cst_hash_hash): Use wide-int interfaces.
3719 (int_cst_hash_eq): Likewise.
3720 (build_int_cst_wide): Delete.
3721 (wide_int_to_tree): New.
3722 (cache_integer_cst): Use wide-int interfaces.
3723 (build_low_bits_mask): Likewise.
3724 (cst_and_fits_in_hwi): Likewise.
3725 (real_value_from_int_cst): Likewise.
3726 (make_int_cst_stat): New.
3727 (integer_zerop): Use wide_int interfaces.
3728 (integer_onep): Likewise.
3729 (integer_all_onesp): Likewise.
3730 (integer_pow2p): Likewise.
3731 (integer_nonzerop): Likewise.
3732 (tree_log2): Likewise.
3733 (tree_floor_log2): Likewise.
3734 (tree_ctz): Likewise.
3735 (int_size_in_bytes): Likewise.
3736 (mem_ref_offset): Return an offset_int rather than a double_int.
3737 (build_type_attribute_qual_variant): Use wide_int interfaces.
3738 (type_hash_eq): Likewise
3739 (tree_int_cst_equal): Likewise.
3740 (tree_int_cst_lt): Delete.
3741 (tree_int_cst_compare): Likewise.
3742 (tree_fits_shwi_p): Use wide_int interfaces.
3743 (tree_fits_uhwi_p): Likewise.
3744 (tree_int_cst_sign_bit): Likewise.
3745 (tree_int_cst_sgn): Likewise.
3746 (tree_int_cst_min_precision): Take a signop rather than a bool.
3747 (simple_cst_equal): Use wide_int interfaces.
3748 (compare_tree_int): Likewise.
3749 (iterative_hash_expr): Likewise.
3750 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
3752 (get_type_static_bounds): Use wide_int interfaces.
3753 (tree_int_cst_elt_check_failed): New.
3754 (build_common_tree_nodes): Reordered to set prec before filling in
3756 (int_cst_value): Check cst_and_fits_in_hwi.
3757 (widest_int_cst_value): Use wide_int interfaces.
3758 (upper_bound_in_type): Likewise.
3759 (lower_bound_in_type): Likewise.
3760 (num_ending_zeros): Likewise.
3761 (drop_tree_overflow): Likewise.
3762 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
3763 (gen_conditions_for_pow_cst_base): Likewise.
3764 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
3765 (group_case_labels_stmt): Use wide-int interfaces.
3766 (verify_gimple_assign_binary): Likewise.
3767 (print_loop): Likewise.
3768 * tree-chrec.c (tree_fold_binomial): Likewise.
3769 * tree-core.h (struct tree_base): Add int_length.
3770 (struct tree_int_cst): Change rep of value.
3771 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
3772 (dr_may_alias_p): Likewise.
3773 (max_stmt_executions_tree): Likewise.
3774 * tree.def (INTEGER_CST): Update comment.
3775 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
3776 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
3777 * tree-dump.c: Include wide-int.h and wide-int-print.h.
3778 (dequeue_and_dump): Use wide-int interfaces.
3779 * tree.h: Include wide-int.h.
3780 (NULL_TREE): Moved to earlier loc in file.
3781 (TREE_INT_CST_ELT_CHECK): New.
3782 (tree_int_cst_elt_check_failed): New.
3784 (TREE_INT_CST): Delete.
3785 (TREE_INT_CST_LOW): Use wide-int interfaces.
3786 (TREE_INT_CST_HIGH): Delete.
3787 (TREE_INT_CST_NUNITS): New.
3788 (TREE_INT_CST_EXT_NUNITS): Likewise.
3789 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
3790 (TREE_INT_CST_ELT): Likewise.
3791 (INT_CST_LT): Delete.
3792 (tree_int_cst_elt_check): New (two forms).
3793 (type_code_size): Update comment.
3794 (make_int_cst_stat, make_int_cst): New.
3795 (tree_to_double_int): Delete.
3796 (double_int_fits_to_tree_p): Delete.
3797 (force_fit_type_double): Delete.
3798 (build_int_cstu): Replace with out-of-line function.
3799 (build_int_cst_wide): Delete.
3800 (tree_int_cst_lt): Define inline.
3801 (tree_int_cst_le): New.
3802 (tree_int_cst_compare): Define inline.
3803 (tree_int_cst_min_precision): Take a signop rather than a bool.
3804 (wi::int_traits <const_tree>): New.
3805 (wi::int_traits <tree>): New.
3806 (wi::extended_tree): New.
3807 (wi::int_traits <wi::extended_tree>): New.
3808 (wi::to_widest): New.
3809 (wi::to_offset): New.
3810 (wi::fits_to_tree_p): New.
3811 (wi::min_value): New.
3812 (wi::max_value): New.
3813 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
3814 (copy_tree_body_r): Likewise.
3815 * tree-object-size.c (compute_object_offset): Likewise.
3816 (addr_object_size): Likewise.
3817 * tree-predcom.c: Include wide-int-print.h.
3818 (struct dref_d): Change type of offset to widest_int.
3819 (dump_dref): Call wide-int printer.
3820 (aff_combination_dr_offset): Use wide-int interfaces.
3821 (determine_offset): Take a widest_int pointer rather than a
3823 (split_data_refs_to_components): Use wide-int interfaces.
3824 (suitable_component_p): Likewise.
3825 (order_drefs): Likewise.
3826 (add_ref_to_chain): Likewise.
3827 (valid_initializer_p): Likewise.
3828 (determine_roots_comp): Likewise.
3829 * tree-pretty-print.c: Include wide-int-print.h.
3830 (dump_generic_node): Use wide-int interfaces.
3831 * tree-sra.c (sra_ipa_modify_expr): Likewise.
3832 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
3833 (move_fixed_address_to_symbol): Likewise.
3834 (move_hint_to_base): Likewise.
3835 (move_pointer_to_base): Likewise.
3836 (move_variant_to_index): Likewise.
3837 (most_expensive_mult_to_index): Likewise.
3838 (addr_to_parts): Likewise.
3839 (copy_ref_info): Likewise.
3840 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
3841 (indirect_refs_may_alias_p): Likewise.
3842 (stmt_kills_ref_p_1): Likewise.
3843 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
3844 * tree-ssa-ccp.c: Update comment at top of file. Include
3846 (struct prop_value_d): Change type of mask to widest_int.
3847 (extend_mask): New function.
3848 (dump_lattice_value): Use wide-int interfaces.
3849 (get_default_value): Likewise.
3850 (set_constant_value): Likewise.
3851 (set_value_varying): Likewise.
3852 (valid_lattice_transition): Likewise.
3853 (set_lattice_value): Likewise.
3854 (value_to_double_int): Delete.
3855 (value_to_wide_int): New.
3856 (get_value_from_alignment): Use wide-int interfaces.
3857 (get_value_for_expr): Likewise.
3858 (do_dbg_cnt): Likewise.
3859 (ccp_finalize): Likewise.
3860 (ccp_lattice_meet): Likewise.
3861 (bit_value_unop_1): Use widest_ints rather than double_ints.
3862 (bit_value_binop_1): Likewise.
3863 (bit_value_unop): Use wide-int interfaces.
3864 (bit_value_binop): Likewise.
3865 (bit_value_assume_aligned): Likewise.
3866 (evaluate_stmt): Likewise.
3867 (ccp_fold_stmt): Likewise.
3868 (visit_cond_stmt): Likewise.
3869 (ccp_visit_stmt): Likewise.
3870 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
3871 (constant_pointer_difference): Likewise.
3872 (associate_pointerplus): Likewise.
3873 (combine_conversions): Likewise.
3874 * tree-ssa-loop.h: Include wide-int.h.
3875 (struct tree_niter_desc): Change type of max to widest_int.
3876 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
3877 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
3878 (remove_redundant_iv_tests): Likewise.
3879 (canonicalize_loop_induction_variables): Likewise.
3880 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
3881 (constant_multiple_of): Take a widest_int pointer instead of
3882 a double_int pointer.
3883 (get_computation_aff): Use wide-int interfaces.
3884 (ptr_difference_cost): Likewise.
3885 (difference_cost): Likewise.
3886 (get_loop_invariant_expr_id): Likewise.
3887 (get_computation_cost_at): Likewise.
3888 (iv_elimination_compare_lt): Likewise.
3889 (may_eliminate_iv): Likewise.
3890 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
3891 instead of double_int.
3892 (max_loop_iterations): Likewise.
3893 (max_stmt_executions): Likewise.
3894 (estimated_stmt_executions): Likewise.
3895 * tree-ssa-loop-niter.c: Include wide-int-print.h.
3896 (split_to_var_and_offset): Use wide-int interfaces.
3897 (determine_value_range): Likewise.
3898 (bound_difference_of_offsetted_base): Likewise.
3899 (bounds_add): Take a widest_int instead of a double_int.
3900 (number_of_iterations_ne_max): Use wide-int interfaces.
3901 (number_of_iterations_ne): Likewise.
3902 (number_of_iterations_lt_to_ne): Likewise.
3903 (assert_loop_rolls_lt): Likewise.
3904 (number_of_iterations_lt): Likewise.
3905 (number_of_iterations_le): Likewise.
3906 (number_of_iterations_cond): Likewise.
3907 (number_of_iterations_exit): Likewise.
3908 (finite_loop_p): Likewise.
3909 (derive_constant_upper_bound_assign): Likewise.
3910 (derive_constant_upper_bound): Return a widest_int.
3911 (derive_constant_upper_bound_ops): Likewise.
3912 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
3913 (record_estimate): Take a widest_int rather than a double_int.
3914 (record_nonwrapping_iv): Use wide-int interfaces.
3915 (double_int_cmp): Delete.
3916 (wide_int_cmp): New.
3917 (bound_index): Take a widest_int rather than a double_int.
3918 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
3919 (maybe_lower_iteration_bound): Likewise.
3920 (estimate_numbers_of_iterations_loop): Likewise.
3921 (estimated_loop_iterations): Take a widest_int pointer than than
3922 a double_int pointer.
3923 (estimated_loop_iterations_int): Use wide-int interfaces.
3924 (max_loop_iterations): Take a widest_int pointer than than
3925 a double_int pointer.
3926 (max_loop_iterations_int): Use wide-int interfaces.
3927 (max_stmt_executions): Take a widest_int pointer than than
3928 a double_int pointer.
3929 (estimated_stmt_executions): Likewise.
3930 (n_of_executions_at_most): Use wide-int interfaces.
3931 (scev_probably_wraps_p): Likewise.
3932 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
3934 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
3936 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
3937 double_ints. Adjust for trailing_wide_ints <3> representation.
3938 (set_nonzero_bits): Likewise.
3939 (get_range_info): Return wide_ints rather than double_ints.
3940 Adjust for trailing_wide_ints <3> representation.
3941 (get_nonzero_bits): Likewise.
3942 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
3944 * tree-ssanames.h (struct range_info_def): Replace min, max and
3945 nonzero_bits with a trailing_wide_ints <3>.
3946 (set_range_info): Use wide_int_refs rather than double_ints.
3947 (set_nonzero_bits): Likewise.
3948 (get_range_info): Return wide_ints rather than double_ints.
3949 (get_nonzero_bits): Likewise.
3950 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
3951 * tree-ssa-pre.c (phi_translate_1): Likewise.
3952 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
3953 (acceptable_pow_call): Likewise.
3954 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
3956 (vn_reference_fold_indirect): Likewise.
3957 (vn_reference_maybe_forwprop_address): Likewise.
3958 (valueize_refs_1): Likewise.
3959 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
3960 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
3961 tree_int_cst_lt and tree_int_cst_le.
3962 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
3964 (streamer_alloc_tree): Likewise.
3965 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
3966 (streamer_write_tree_header): Likewise.
3967 (streamer_write_integer_cst): Likewise.
3968 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
3969 (build_constructors): Likewise.
3970 (array_value_type): Likewise.
3971 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
3972 (vect_check_gather): Likewise.
3973 * tree-vect-generic.c (build_replicated_const): Likewise.
3974 (expand_vector_divmod): Likewise.
3975 * tree-vect-loop.c (vect_transform_loop): Likewise.
3976 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
3977 (vect_do_peeling_for_alignment): Likewise.
3978 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
3979 * tree-vrp.c: Include wide-int.h.
3980 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
3981 (extract_range_from_assert): Use wide-int interfaces.
3982 (vrp_int_const_binop): Likewise.
3983 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
3984 double_int pointers.
3985 (ranges_from_anti_range): Use wide-int interfaces.
3986 (quad_int_cmp): Delete.
3987 (quad_int_pair_sort): Likewise.
3988 (extract_range_from_binary_expr_1): Use wide-int interfaces.
3989 (extract_range_from_unary_expr_1): Likewise.
3990 (adjust_range_with_scev): Likewise.
3991 (masked_increment): Take and return wide_ints rather than double_ints.
3992 (register_edge_assert_for_2): Use wide-int interfaces.
3993 (check_array_ref): Likewise.
3994 (search_for_addr_array): Likewise.
3995 (maybe_set_nonzero_bits): Likewise.
3996 (union_ranges): Pass an integer of the correct type instead of
3997 using integer_one_node.
3998 (intersect_ranges): Likewise.
3999 (simplify_truth_ops_using_ranges): Likewise.
4000 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
4001 (range_fits_type_p): Likewise.
4002 (simplify_cond_using_ranges): Likewise. Take a signop rather than
4004 (simplify_conversion_using_ranges): Use wide-int interfaces.
4005 (simplify_float_conversion_using_ranges): Likewise.
4006 (vrp_finalize): Likewise.
4007 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
4008 (gimple_stringops_transform): Likewise.
4009 * varasm.c (decode_addr_const): Likewise.
4010 (const_hash_1): Likewise.
4011 (const_rtx_hash_1): Likewise
4012 (output_constant): Likewise.
4013 (array_size_for_constructor): Likewise.
4014 (output_constructor_regular_field): Likewise.
4015 (output_constructor_bitfield): Likewise.
4016 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
4017 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
4019 * gencheck.c: Define BITS_PER_UNIT.
4022 * wide-int-print.cc: New.
4023 * wide-int-print.h: New.
4025 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4027 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
4029 2014-05-06 Richard Biener <rguenther@suse.de>
4031 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
4032 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
4033 (TODO_verify_all): Adjust.
4034 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
4035 TODO_verify_stmts and TODO_verify_rtl_sharing.
4036 * bb-reorder.c: Likewise.
4037 * cfgexpand.c: Likewise.
4038 * cprop.c: Likewise.
4040 * function.c: Likewise.
4041 * fwprop.c: Likewise.
4043 * gimple-ssa-isolate-paths.c: Likewise.
4044 * gimple-ssa-strength-reduction.c: Likewise.
4045 * ipa-split.c: Likewise.
4046 * loop-init.c: Likewise.
4047 * loop-unroll.c: Likewise.
4048 * lower-subreg.c: Likewise.
4049 * modulo-sched.c: Likewise.
4050 * postreload-gcse.c: Likewise.
4051 * predict.c: Likewise.
4052 * recog.c: Likewise.
4053 * sched-rgn.c: Likewise.
4054 * store-motion.c: Likewise.
4055 * tracer.c: Likewise.
4056 * trans-mem.c: Likewise.
4057 * tree-call-cdce.c: Likewise.
4058 * tree-cfg.c: Likewise.
4059 * tree-cfgcleanup.c: Likewise.
4060 * tree-complex.c: Likewise.
4061 * tree-eh.c: Likewise.
4062 * tree-emutls.c: Likewise.
4063 * tree-if-conv.c: Likewise.
4064 * tree-into-ssa.c: Likewise.
4065 * tree-loop-distribution.c: Likewise.
4066 * tree-object-size.c: Likewise.
4067 * tree-parloops.c: Likewise.
4068 * tree-pass.h: Likewise.
4069 * tree-sra.c: Likewise.
4070 * tree-ssa-ccp.c: Likewise.
4071 * tree-ssa-copy.c: Likewise.
4072 * tree-ssa-copyrename.c: Likewise.
4073 * tree-ssa-dce.c: Likewise.
4074 * tree-ssa-dom.c: Likewise.
4075 * tree-ssa-dse.c: Likewise.
4076 * tree-ssa-forwprop.c: Likewise.
4077 * tree-ssa-ifcombine.c: Likewise.
4078 * tree-ssa-loop-ch.c: Likewise.
4079 * tree-ssa-loop-ivcanon.c: Likewise.
4080 * tree-ssa-loop.c: Likewise.
4081 * tree-ssa-math-opts.c: Likewise.
4082 * tree-ssa-phiopt.c: Likewise.
4083 * tree-ssa-phiprop.c: Likewise.
4084 * tree-ssa-pre.c: Likewise.
4085 * tree-ssa-reassoc.c: Likewise.
4086 * tree-ssa-sink.c: Likewise.
4087 * tree-ssa-strlen.c: Likewise.
4088 * tree-ssa-tail-merge.c: Likewise.
4089 * tree-ssa-uncprop.c: Likewise.
4090 * tree-switch-conversion.c: Likewise.
4091 * tree-tailcall.c: Likewise.
4092 * tree-vect-generic.c: Likewise.
4093 * tree-vectorizer.c: Likewise.
4094 * tree-vrp.c: Likewise.
4096 * var-tracking.c: Likewise.
4097 * bt-load.c: Likewise.
4098 * cfgcleanup.c: Likewise.
4099 * combine-stack-adj.c: Likewise.
4100 * combine.c: Likewise.
4101 * compare-elim.c: Likewise.
4102 * config/epiphany/resolve-sw-modes.c: Likewise.
4103 * config/i386/i386.c: Likewise.
4104 * config/mips/mips.c: Likewise.
4105 * config/s390/s390.c: Likewise.
4106 * config/sh/sh_treg_combine.cc: Likewise.
4107 * config/sparc/sparc.c: Likewise.
4110 * final.c: Likewise.
4111 * ifcvt.c: Likewise.
4112 * mode-switching.c: Likewise.
4113 * passes.c: Likewise.
4114 * postreload.c: Likewise.
4116 * reg-stack.c: Likewise.
4117 * regcprop.c: Likewise.
4118 * regrename.c: Likewise.
4121 2014-05-06 Richard Biener <rguenther@suse.de>
4124 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
4125 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
4127 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
4130 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
4132 2014-05-05 Radovan Obradovic <robradovic@mips.com>
4133 Tom de Vries <tom@codesourcery.com>
4135 * target.def (call_fusage_contains_non_callee_clobbers): New
4137 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
4139 (@node Miscellaneous Register Hooks): New node.
4140 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
4141 * doc/tm.texi: Regenerate.
4143 2014-05-05 Marek Polacek <polacek@redhat.com>
4146 * opts.c (common_handle_option): Call error_at instead of warning_at.
4148 2014-05-05 Richard Biener <rguenther@suse.de>
4150 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
4151 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
4152 under the TODO_verify_il umbrella.
4154 2014-05-05 Richard Biener <rguenther@suse.de>
4156 * passes.c (execute_function_todo): Move TODO_verify_flow under
4157 the TODO_verify_ul umbrella.
4159 2014-05-05 Richard Biener <rguenther@suse.de>
4162 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
4163 X & CST away from a CST that is the mask of a mode.
4165 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4167 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
4168 int argument to enum machine_mode.
4169 (picochip_class_max_nregs): Ditto.
4170 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
4171 (picochip_class_max_nregs): Ditto.
4173 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4175 * target.def: Add new target hook.
4176 * doc/tm.texi: Regenerate.
4177 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
4178 * targhooks.c (default_keep_leaf_when_profiled): New function.
4180 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
4181 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
4183 2014-05-05 Bin Cheng <bin.cheng@arm.com>
4185 PR tree-optimization/60363
4186 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
4187 (copy_phi_args): New parameters. Call get_value_locus_in_path.
4188 (update_destination_phis): New parameter.
4189 (create_edge_and_update_destination_phis): Ditto.
4190 (ssa_fix_duplicate_block_edges): Pass new arguments.
4191 (thread_single_edge): Ditto.
4193 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
4195 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
4196 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
4197 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
4198 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
4199 Use RS6000_BTM_HARD_FLOAT.
4200 (BU_MISC_2): Likewise.
4201 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
4202 RS6000_BTM_HARD_FLOAT.
4203 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
4205 (rs6000_invalid_builtin): Add hard floating builtin support.
4206 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
4207 hard float builtins.
4208 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
4210 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4212 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
4213 Add missing function* argument.
4215 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
4217 * lra-constraints.c (valid_address_p): Move earlier in file.
4218 Add a constraint argument to the address_info version.
4219 (satisfies_memory_constraint_p): New function.
4220 (satisfies_address_constraint_p): Likewise.
4221 (process_alt_operands, curr_insn_transform): Use them.
4222 (process_address): Pass the constraint to valid_address_p when
4223 checking address operands.
4225 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
4227 * config/mips/mips.c (mips_isa_rev): New variable.
4228 (mips_set_architecture): Set it.
4229 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
4231 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
4232 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
4233 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
4234 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
4235 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
4236 conditions in terms of mips_isa_rev.
4237 (mips_isa_rev): Declare.
4239 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4241 * config/sh/sh-mem.cc: Use tabs instead of spaces.
4242 (prob_unlikely, prob_likely): Make variables const.
4244 2014-05-03 Denis Chertykov <chertykov@gmail.com>
4246 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
4248 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4250 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
4252 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4254 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
4255 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
4256 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
4258 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
4259 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
4261 Replace usage of ROUND_REG with sh_round_reg.
4262 Use CEIL instead of ROUND_ADVANCE.
4264 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4267 * config/sh/sh.c: Include stdlib headers before everything else.
4269 2014-05-02 Jakub Jelinek <jakub@redhat.com>
4271 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
4272 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
4273 (gimplify_adjust_omp_clauses): Simd region is never
4274 directly nested in combined parallel. Instead, for linear
4275 with copyin/copyout, if in combined for simd loop, make decl
4276 firstprivate/lastprivate on OMP_FOR.
4277 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
4278 expand_omp_for_static_chunk): When setting endvar, also set
4279 fd->loop.v to the same value.
4281 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
4283 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
4285 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
4287 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
4290 2014-05-02 Marek Polacek <polacek@redhat.com>
4292 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
4294 2014-05-02 Kito Cheng <kito@0xlab.org>
4296 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
4297 to a C expression marco.
4298 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
4299 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
4300 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
4301 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
4302 HONOR_REG_ALLOC_ORDER.
4303 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
4305 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4307 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
4309 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4311 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
4313 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
4315 * tree-if-conv.c (is_cond_scalar_reduction): New function.
4316 (convert_scalar_cond_reduction): Likewise.
4317 (predicate_scalar_phi): Add recognition and transformation
4318 of simple conditioanl reduction to be vectorizable.
4320 2014-05-01 Marek Polacek <polacek@redhat.com>
4323 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
4325 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
4327 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
4328 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
4329 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
4330 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
4331 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
4332 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
4333 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
4334 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
4336 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
4338 * config/arc/arc.opt (mlra): Move comment above option name
4339 to avoid mis-parsing as language options.
4341 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4343 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
4344 * config/sol2.h: ... here.
4345 * config/sol2-10.h: Remove.
4347 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
4348 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
4349 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
4350 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
4351 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
4352 * config/sol2.h: ... here.
4353 (SECTION_NAME_FORMAT): Don't redefine.
4354 (STARTFILE_ARCH32_SPEC): Rename to ...
4355 (STARTFILE_ARCH_SPEC): ... this.
4356 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
4357 * config/sparc/sol2.h: ... here.
4358 (SECTION_NAME_FORMAT): Don't undef.
4359 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
4360 (SUBTARGET_EXTRA_SPECS): Remove.
4361 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
4363 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
4364 (MD_STARTFILE_PREFIX): Remove.
4365 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
4366 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
4367 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
4368 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
4369 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
4370 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
4371 * config/i386/sol2.h: ... here.
4372 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
4373 * config/i386/sol2-bi.h: Remove.
4374 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
4375 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
4377 * config/i386/t-sol2-64: Rename to ...
4378 * config/i386/t-sol2: ... this.
4379 * config/sparc/t-sol2-64: Rename to ...
4380 * config/sparc/t-sol2: ... this.
4382 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
4383 sol2_tm_file_head, sol2_tm_file_tail.
4384 Include ${cpu_type}/sol2.h before sol2.h.
4386 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
4387 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
4388 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
4389 Reflect i386/t-sol2-64 renaming.
4390 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
4391 Reflect sparc/t-sol2-64 renaming.
4393 2014-04-30 Richard Biener <rguenther@suse.de>
4395 * passes.c (execute_function_todo): Move TODO_verify_stmts
4396 and TODO_verify_ssa under the TODO_verify_il umbrella.
4397 * tree-ssa.h (verify_ssa): Adjust prototype.
4398 * tree-ssa.c (verify_ssa): Add parameter to tell whether
4399 we should verify SSA operands.
4400 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
4401 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
4402 whether we should verify whether not throwing stmts have EH info.
4403 * graphite-scop-detection.c (create_sese_edges): Adjust.
4404 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
4405 * tree-eh.c (lower_try_finally_switch): Do not add the
4406 default case label twice.
4408 2014-04-30 Marek Polacek <polacek@redhat.com>
4410 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
4411 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
4412 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
4413 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
4415 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
4417 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
4418 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
4419 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
4420 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
4421 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
4422 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
4423 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
4424 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
4426 2014-04-29 David Malcolm <dmalcolm@redhat.com>
4428 * tree-cfg.c (dump_function_to_file): Dump the return type of
4429 functions, in a line to itself before the function body, mimicking
4430 the layout of a C function.
4432 2014-04-29 Jakub Jelinek <jakub@redhat.com>
4434 PR tree-optimization/60971
4435 * tree-tailcall.c (process_assignment): Reject conversions which
4438 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
4440 * calls.c (initialize_argument_information): Always treat
4441 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
4442 (expand_call): Likewise.
4443 (emit_library_call_calue_1): Likewise.
4444 * expr.c (PUSH_ARGS_REVERSED): Do not define.
4445 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
4448 2014-04-29 Nick Clifton <nickc@redhat.com>
4450 * config/msp430/msp430.md (umulsidi): Fix typo.
4451 (mulhisi3): Enable even inside interrupt handlers.
4452 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
4453 bigger return address pushed in large mode.
4455 2014-04-29 Nick Clifton <nickc@redhat.com>
4457 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
4458 (arc_init_reg_tables): Use a machine_mode enum to iterate over
4460 * config/m32r/m32r.c (init_reg_tables): Likewise.
4461 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
4462 enum to hold the modes.
4464 2014-04-29 Richard Biener <rguenther@suse.de>
4466 * dominance.c (free_dominance_info): Add overload with
4468 (dom_info_state): Likewise.
4469 (dom_info_available_p): Likewise.
4470 * basic-block.h (free_dominance_info, dom_info_state,
4471 dom_info_available_p): Declare overloads.
4472 * passes.c (execute_function_todo): Verify that verifiers
4473 don't change dominator info state. Drop dominator info
4474 for IPA pass invocations.
4475 * cgraph.c (release_function_body): Restore asserts that
4476 dominator information is released.
4478 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
4480 * doc/invoke.texi: Fix typo.
4481 * tree-vrp.c: Fix typos.
4482 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
4484 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4486 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
4488 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
4490 * config/aarch64/aarch64-builtins.c
4491 (aarch64_types_storestruct_lane_qualifiers): New.
4492 (TYPES_STORESTRUCT_LANE): Likewise.
4493 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
4494 (st3_lane): Likewise.
4495 (st4_lane): Likewise.
4496 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
4497 (vec_store_lanesci_lane<mode>): Likewise.
4498 (vec_store_lanesxi_lane<mode>): Likewise.
4499 (aarch64_st2_lane<VQ:mode>): Likewise.
4500 (aarch64_st3_lane<VQ:mode>): Likewise.
4501 (aarch64_st4_lane<VQ:mode>): Likewise.
4502 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
4503 * config/aarch64/arm_neon.h
4504 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
4505 use new macro arguments.
4506 (__ST3_LANE_FUNC): Likewise.
4507 (__ST4_LANE_FUNC): Likewise.
4508 * config/aarch64/iterators.md (V_TWO_ELEM): New.
4509 (V_THREE_ELEM): Likewise.
4510 (V_FOUR_ELEM): Likewise.
4512 2014-04-28 David Malcolm <dmalcolm@redhat.com>
4514 * doc/gimple.texi: Replace the description of the now-defunct
4515 union gimple_statement_d with a diagram showing the
4516 gimple_statement_base class hierarchy and its relationships to
4517 the GSS_ and GIMPLE_ enums.
4519 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
4521 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
4522 * config/aarch64/aarch64.c
4523 (aarch64_cannot_change_mode_class): Weaken conditions.
4524 (aarch64_modes_tieable_p): New.
4525 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
4527 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
4529 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4530 (loadsync_<mode>): Change mode.
4531 (load_quadpti, store_quadpti): New.
4532 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4533 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4535 2014-04-28 Martin Jambor <mjambor@suse.cz>
4537 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
4538 same alias type as the original statement.
4539 (subreplacement_assignment_data): New type.
4540 (handle_unscalarized_data_in_subtree): New type of parameter,
4541 generate new memory accesses with same alias type as the original
4543 (load_assign_lhs_subreplacements): Likewise.
4544 (sra_modify_constructor_assign): Generate new memory accesses with
4545 same alias type as the original statement.
4547 2014-04-28 Richard Biener <rguenther@suse.de>
4549 * tree-pass.h (TODO_verify_il): Define.
4550 (TODO_verify_all): Complete properly.
4551 * passes.c (execute_function_todo): Move existing loop-closed
4552 SSA verification under TODO_verify_il.
4553 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
4554 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
4555 Fix tree sharing issue.
4557 2014-04-28 Richard Biener <rguenther@suse.de>
4560 * builtins.def (DEF_C11_BUILTIN): Add.
4561 (BUILT_IN_ALIGNED_ALLOC): Likewise.
4562 * coretypes.h (enum function_class): Add function_c11_misc.
4563 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
4564 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
4565 (call_may_clobber_ref_p_1): Likewise.
4566 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
4567 (mark_all_reaching_defs_necessary_1): Likewise.
4568 (propagate_necessity): Likewise.
4569 (eliminate_unnecessary_stmts): Likewise.
4570 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
4572 2014-04-28 Richard Biener <rguenther@suse.de>
4574 * tree-vrp.c (vrp_var_may_overflow): Remove.
4575 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
4576 with overflow immediately bump to one before that value and
4577 let iteration figure out overflow status.
4579 2014-04-28 Richard Biener <rguenther@suse.de>
4581 * configure.ac: Do valgrind header checks unconditionally.
4582 Add --enable-valgrind-annotations.
4583 * system.h: Guard valgrind header inclusion with
4584 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
4585 * alloc-pool.c (pool_alloc, pool_free): Use
4586 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
4587 to guard possibly dead code.
4588 * config.in: Regenerated.
4589 * configure: Likewise.
4591 2014-04-28 Jeff Law <law@redhat.com>
4593 PR tree-optimization/60902
4594 * tree-ssa-threadedge.c
4595 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
4596 over real defs when invalidating outputs from statements that do not
4597 produce useful outputs for threading.
4599 2014-04-28 Richard Biener <rguenther@suse.de>
4601 PR tree-optimization/60979
4602 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
4603 SCOPs that end in a block with a successor with abnormal
4606 2014-04-28 Richard Biener <rguenther@suse.de>
4608 * tree-pass.h (execute_pass_list): Adjust prototype.
4609 * passes.c (pass_manager::execute_early_local_passes): Adjust.
4610 (do_per_function): Change callback signature, push all actual
4611 work to the callbals.
4612 (do_per_function_toporder): Likewise.
4613 (execute_function_dump): Adjust.
4614 (execute_function_todo): Likewise.
4615 (clear_last_verified): Likewise.
4616 (verify_curr_properties): Likewise.
4617 (update_properties_after_pass): Likewise.
4618 (execute_pass_list_1): Split out from ...
4619 (execute_pass_list): ... here. Adjust.
4620 (execute_ipa_pass_list): Likewise.
4621 * cgraphunit.c (cgraph_add_new_function): Adjust.
4622 (analyze_function): Likewise.
4623 (expand_function): Likewise.
4624 * cgraph.c (release_function_body): Free dominance info
4625 here instead of asserting it was magically freed elsewhere.
4627 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
4629 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
4630 * configure: Regenerate.
4631 * config/sparc/sparc.opt (muser-mode): New option.
4632 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
4634 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
4635 * doc/invoke.texi (SPARC options): Document -muser-mode.
4637 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
4639 * cselib.c (find_slot_memmode): Delete.
4640 (cselib_hasher): Change compare_type to a struct.
4641 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
4643 (preserve_constants_and_equivs): Adjust for new compare_type.
4644 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
4645 (wrap_constant): Delete.
4646 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
4648 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
4650 * doc/install.texi (Building with profile feedback): Remove
4653 2014-04-26 Tom de Vries <tom@codesourcery.com>
4655 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
4658 2014-04-25 Cary Coutant <ccoutant@google.com>
4661 * dwarf2out.c (should_move_die_to_comdat): A type definition
4662 can contain a subprogram definition, but don't move it to a
4664 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
4665 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
4667 (clone_tree_hash): Rename to...
4668 (clone_tree_partial): ...this; change callers. Copy
4669 DW_TAG_subprogram DIEs as declarations.
4670 (copy_decls_walk): Don't copy children of a declaration into a
4673 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
4676 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
4679 2014-04-25 Jiong Wang <jiong.wang@arm.com>
4681 * config/arm/predicates.md (call_insn_operand): Add long_call check.
4682 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
4684 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
4687 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4689 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
4691 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4693 PR tree-optimization/60930
4694 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
4695 creating a multiply candidate by folding two constant
4696 multiplicands when the result overflows.
4698 2014-04-25 Jakub Jelinek <jakub@redhat.com>
4700 PR tree-optimization/60960
4701 * tree-vect-generic.c (expand_vector_operation): Only call
4702 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
4704 2014-04-25 Tom de Vries <tom@codesourcery.com>
4706 * expr.c (clobber_reg_mode): New function.
4707 * expr.h (clobber_reg): New function.
4709 2014-04-25 Tom de Vries <tom@codesourcery.com>
4711 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
4714 2014-04-25 Radovan Obradovic <robradovic@mips.com>
4715 Tom de Vries <tom@codesourcery.com>
4717 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
4719 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
4720 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
4721 new argument to find_all_hard_reg_sets call.
4723 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4725 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
4726 Use HOST_WIDE_INT_C for mask literal.
4727 (aarch_rev16_shleft_mask_imm_p): Likewise.
4729 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
4732 * config/sparc/sparc.md (ashlsi3_extend): Delete.
4734 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
4736 PR preprocessor/56540
4737 * config/i386/i386-c.c (ix86_target_macros): Define
4738 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
4740 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4742 * configure.ac (tga_func): Remove.
4743 (LIB_TLS_SPEC): Remove.
4744 * configure: Regenerate.
4745 * config.in: Regenerate.
4746 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
4748 2014-04-25 Richard Biener <rguenther@suse.de>
4751 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
4752 call stmt use/clobber sets during stmt walk instead of
4753 walking the possibly incomplete set of caller edges.
4755 2014-04-25 Richard Biener <rguenther@suse.de>
4758 * passes.c (apply_ipa_transforms): Inline into only caller ...
4759 (execute_one_pass): ... here. Properly bring in function
4760 bodies for nodes we want to apply IPA transforms to.
4762 2014-04-24 Cong Hou <congh@google.com>
4764 PR tree-optimization/60896
4765 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
4766 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
4767 (vect_mark_pattern_stmts): Set the def type of all statements in
4768 PATTERN_DEF_SEQ as vect_internal_def.
4770 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
4772 * doc/extend.texi (PowerPC Built-in Functions): Document new
4773 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
4774 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
4776 * config/rs6000/predicates.md (const_0_to_3_operand): New
4777 predicate to match 0..3 integer constants.
4779 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
4780 to support adding miscellaneous builtin functions.
4781 (BU_DFP_MISC_2): Likewise.
4782 (BU_P7_MISC_1): Likewise.
4783 (BU_P7_MISC_2): Likewise.
4784 (BU_P8V_MISC_3): Likewise.
4785 (BU_MISC_1): Likewise.
4786 (BU_MISC_2): Likewise.
4787 (DIVWE): Add extended divide builtin functions.
4790 (DIVWEUO): Likewise.
4794 (DIVDEUO): Likewise.
4795 (DXEX): Add decimal floating-point builtin functions.
4798 (DDEDPDQ): Likewise.
4800 (DENBCDQ): Likewise.
4807 (CDTBCD): Add new BCD builtin functions.
4811 (BCDADD_LT): Likewise.
4812 (BCDADD_EQ): Likewise.
4813 (BCDADD_GT): Likewise.
4814 (BCDADD_OV): Likewise.
4816 (BCDSUB_LT): Likewise.
4817 (BCDSUB_EQ): Likewise.
4818 (BCDSUB_GT): Likewise.
4819 (BCDSUB_OV): Likewise.
4820 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
4821 (UNPACK_TD): Likewise.
4822 (PACK_TF): Likewise.
4823 (UNPACK_TF): Likewise.
4824 (UNPACK_TF_0): Likewise.
4825 (UNPACK_TF_1): Likewise.
4826 (PACK_V1TI): Likewise.
4827 (UNPACK_V1TI): Likewise.
4829 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
4830 support for decimal floating point builtin functions.
4831 (rs6000_expand_ternop_builtin): Add checks for the new builtin
4832 functions that take constant arguments.
4833 (rs6000_invalid_builtin): Add decimal floating point builtin support.
4834 (rs6000_init_builtins): Setup long double, _Decimal64, and
4835 _Decimal128 types for new builtin functions.
4836 (builtin_function_type): Set the unsigned flags appropriately for
4837 the new builtin functions.
4838 (rs6000_opt_masks): Add support for decimal floating point builtin
4841 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
4842 floating point builtin functions.
4843 (RS6000_BTM_COMMON): Likewise.
4844 (RS6000_BTI_long_double): Likewise.
4845 (RS6000_BTI_dfloat64): Likewise.
4846 (RS6000_BTI_dfloat128): Likewise.
4847 (long_double_type_internal_node): Likewise.
4848 (dfloat64_type_internal_node): Likewise.
4849 (dfloat128_type_internal_node): Likewise.
4851 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
4852 2.07 bcd arithmetic instructions.
4853 (UNSPEC_BCDSUB): Likewise.
4854 (UNSPEC_BCD_OVERFLOW): Likewise.
4855 (UNSPEC_BCD_ADD_SUB): Likewise.
4856 (bcd_add_sub): Likewise.
4857 (BCD_TEST): Likewise.
4858 (bcd<bcd_add_sub>): Likewise.
4859 (bcd<bcd_add_sub>_test): Likewise.
4860 (bcd<bcd_add_sub>_test2): Likewise.
4861 (bcd<bcd_add_sub>_<code>): Likewise.
4862 (peephole2 for combined bcd ops): Likewise.
4864 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
4865 decimal floating point builtin functions.
4866 (UNSPEC_DENBCD): Likewise.
4867 (UNSPEC_DXEX): Likewise.
4868 (UNSPEC_DIEX): Likewise.
4869 (UNSPEC_DSCLI): Likewise.
4870 (UNSPEC_DSCRI): Likewise.
4871 (D64_D128): Likewise.
4872 (dfp_suffix): Likewise.
4873 (dfp_ddedpd_<mode>): Likewise.
4874 (dfp_denbcd_<mode>): Likewise.
4875 (dfp_dxex_<mode>): Likewise.
4876 (dfp_diex_<mode>): Likewise.
4877 (dfp_dscli_<mode>): Likewise.
4878 (dfp_dscri_<mode>): Likewise.
4880 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
4882 (UNSPEC_CDTBCD): Likewise.
4883 (UNSPEC_CBCDTD): Likewise.
4884 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
4885 (UNSPEC_DIVEO): Likewise.
4886 (UNSPEC_DIVEU): Likewise.
4887 (UNSPEC_DIVEUO): Likewise.
4888 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
4889 pack/unpack 128-bit types.
4890 (UNSPEC_PACK_128BIT): Likewise.
4891 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
4892 (udiv<mode>3): Use idiv_ldiv mode attribute.
4893 (div<mode>3): Likewise.
4894 (addg6s): Add new BCD builtin functions.
4897 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
4898 (div_extend): Likewise.
4899 (div<div_extend>_<mode>"): Likewise.
4900 (FP128_64): Add support for new builtin functions to pack/unpack
4902 (unpack<mode>): Likewise.
4903 (unpacktf_0): Likewise.
4904 (unpacktf_1): Likewise.
4905 (unpack<mode>_dm): Likewise.
4906 (unpack<mode>_nodm): Likewise.
4907 (pack<mode>): Likewise.
4908 (unpackv1ti): Likewise.
4909 (packv1ti): Likewise.
4911 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
4913 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
4916 2014-04-24 Jakub Jelinek <jakub@redhat.com>
4918 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
4919 * gimplify.c (omp_is_private): Change last argument's type to int.
4920 Only diagnose lastprivate if the simd argument is 1, only diagnose
4921 linear if the simd argument is 2.
4922 (gimplify_omp_for): Adjust omp_is_private callers. When adding
4923 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
4924 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
4925 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
4926 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
4927 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
4928 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
4929 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
4930 * tree-nested.c (convert_nonlocal_omp_clauses,
4931 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
4933 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
4936 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
4939 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
4941 * flag-types.h (enum ivar_visibility): Add.
4943 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
4945 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
4946 function * argument.
4948 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
4950 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
4952 2014-04-24 Radovan Obradovic <robradovic@mips.com>
4953 Tom de Vries <tom@codesourcery.com>
4955 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
4956 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
4958 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
4959 * emit-rtl.c (try_split): Same.
4961 2014-04-24 Radovan Obradovic <robradovic@mips.com>
4962 Tom de Vries <tom@codesourcery.com>
4964 * common.opt (fuse-caller-save): New option.
4966 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
4968 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
4969 elements for big-endian.
4971 2014-04-24 Richard Biener <rguenther@suse.de>
4973 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
4974 during TER and instead use the sepops interface for expanding
4975 non-GIMPLE_SINGLE_RHS.
4977 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4979 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
4980 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
4982 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4984 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
4985 assembler 64-bit option.
4986 * configure: Regenerate.
4988 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4990 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
4991 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
4992 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
4993 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
4994 (TARGET_CRYPTO): Take TARGET_SIMD into account.
4996 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4998 * config/aarch64/aarch64-builtins.c
4999 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
5000 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
5001 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
5002 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
5004 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
5005 (Vrevsuff): New mode attribute.
5007 2014-04-24 Terry Guo <terry.guo@arm.com>
5009 * config/arm/arm.h (machine_function): Define variable
5010 after_arm_reorg here.
5011 * config/arm/arm.c (after_arm_reorg): Remove the definition.
5012 (arm_split_constant): Update the way to access variable
5015 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
5017 2014-04-23 Tom de Vries <tom@codesourcery.com>
5019 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
5021 2014-04-23 David Malcolm <dmalcolm@redhat.com>
5023 * is-a.h: Update comments to reflect the following changes to the
5024 "pointerness" of the API, making the template parameter match the
5025 return type, allowing use of is-a.h with typedefs of pointers.
5026 (is_a_helper::cast): Return a T rather then a pointer to a T, so
5027 that the return type matches the parameter to the is_a_helper.
5029 (dyn_cast): Likewise.
5031 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
5032 pointer from the is-a.h API.
5034 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
5035 (is_a_helper <cgraph_node *>::test): ...this, matching change to
5037 (is_a_helper <varpool_node>::test): Likewise, convert to...
5038 (is_a_helper <varpool_node *>::test): ...this.
5040 (varpool_first_variable): Update for removal of implicit pointer
5041 from the is-a.h API.
5042 (varpool_next_variable): Likewise.
5043 (varpool_first_static_initializer): Likewise.
5044 (varpool_next_static_initializer): Likewise.
5045 (varpool_first_defined_variable): Likewise.
5046 (varpool_next_defined_variable): Likewise.
5047 (cgraph_first_defined_function): Likewise.
5048 (cgraph_next_defined_function): Likewise.
5049 (cgraph_first_function): Likewise.
5050 (cgraph_next_function): Likewise.
5051 (cgraph_first_function_with_gimple_body): Likewise.
5052 (cgraph_next_function_with_gimple_body): Likewise.
5053 (cgraph_alias_target): Likewise.
5054 (varpool_alias_target): Likewise.
5055 (cgraph_function_or_thunk_node): Likewise.
5056 (varpool_variable_node): Likewise.
5057 (symtab_real_symbol_p): Likewise.
5058 * cgraphunit.c (referred_to_p): Likewise.
5059 (analyze_functions): Likewise.
5060 (handle_alias_pairs): Likewise.
5061 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
5062 * gimple-ssa.h (gimple_vuse_op): Likewise.
5063 (gimple_vdef_op): Likewise.
5064 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
5065 * gimple.c (gimple_build_asm_1): Likewise.
5066 (gimple_build_try): Likewise.
5067 (gimple_build_resx): Likewise.
5068 (gimple_build_eh_dispatch): Likewise.
5069 (gimple_build_omp_for): Likewise.
5070 (gimple_omp_for_set_clauses): Likewise.
5072 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
5073 (is_a_helper <gimple_statement_asm *>::test): ...this.
5074 (is_a_helper <gimple_statement_bind>::test): Convert to...
5075 (is_a_helper <gimple_statement_bind *>::test): ...this.
5076 (is_a_helper <gimple_statement_call>::test): Convert to...
5077 (is_a_helper <gimple_statement_call *>::test): ...this.
5078 (is_a_helper <gimple_statement_catch>::test): Convert to...
5079 (is_a_helper <gimple_statement_catch *>::test): ...this.
5080 (is_a_helper <gimple_statement_resx>::test): Convert to...
5081 (is_a_helper <gimple_statement_resx *>::test): ...this.
5082 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
5083 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
5084 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
5085 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
5086 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
5087 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
5088 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
5089 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
5090 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
5091 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
5092 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
5093 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
5094 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
5095 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
5096 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
5097 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
5098 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
5099 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
5100 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
5101 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
5102 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
5103 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
5104 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
5105 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
5106 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
5107 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
5108 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
5109 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
5110 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
5111 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
5112 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
5113 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
5114 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
5115 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
5116 (is_a_helper <gimple_statement_phi>::test): Convert to...
5117 (is_a_helper <gimple_statement_phi *>::test): ...this.
5118 (is_a_helper <gimple_statement_transaction>::test): Convert to...
5119 (is_a_helper <gimple_statement_transaction *>::test): ...this.
5120 (is_a_helper <gimple_statement_try>::test): Convert to...
5121 (is_a_helper <gimple_statement_try *>::test): ...this.
5122 (is_a_helper <gimple_statement_wce>::test): Convert to...
5123 (is_a_helper <gimple_statement_wce *>::test): ...this.
5124 (is_a_helper <const gimple_statement_asm>::test): Convert to...
5125 (is_a_helper <const gimple_statement_asm *>::test): ...this.
5126 (is_a_helper <const gimple_statement_bind>::test): Convert to...
5127 (is_a_helper <const gimple_statement_bind *>::test): ...this.
5128 (is_a_helper <const gimple_statement_call>::test): Convert to...
5129 (is_a_helper <const gimple_statement_call *>::test): ...this.
5130 (is_a_helper <const gimple_statement_catch>::test): Convert to...
5131 (is_a_helper <const gimple_statement_catch *>::test): ...this.
5132 (is_a_helper <const gimple_statement_resx>::test): Convert to...
5133 (is_a_helper <const gimple_statement_resx *>::test): ...this.
5134 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
5135 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
5136 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
5137 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
5138 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
5140 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
5142 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
5144 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
5146 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
5147 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
5148 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
5150 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
5151 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
5153 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
5154 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
5155 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
5156 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
5157 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
5158 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
5160 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
5161 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
5162 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
5163 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
5165 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
5166 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
5167 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
5168 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
5169 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
5170 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
5171 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
5172 (is_a_helper <const gimple_statement_phi>::test): Convert to...
5173 (is_a_helper <const gimple_statement_phi *>::test): ...this.
5174 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
5175 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
5176 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
5177 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
5178 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
5179 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
5180 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
5182 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
5184 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
5185 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
5187 (gimple_use_ops): Update for removal of implicit pointer from the
5189 (gimple_set_use_ops): Likewise.
5190 (gimple_vuse): Likewise.
5191 (gimple_vdef): Likewise.
5192 (gimple_vuse_ptr): Likewise.
5193 (gimple_vdef_ptr): Likewise.
5194 (gimple_set_vuse): Likewise.
5195 (gimple_set_vdef): Likewise.
5196 (gimple_omp_return_set_lhs): Likewise.
5197 (gimple_omp_return_lhs): Likewise.
5198 (gimple_omp_return_lhs_ptr): Likewise.
5199 (gimple_call_fntype): Likewise.
5200 (gimple_call_set_fntype): Likewise.
5201 (gimple_call_set_internal_fn): Likewise.
5202 (gimple_call_use_set): Likewise.
5203 (gimple_call_clobber_set): Likewise.
5204 (gimple_bind_vars): Likewise.
5205 (gimple_bind_set_vars): Likewise.
5206 (gimple_bind_body_ptr): Likewise.
5207 (gimple_bind_set_body): Likewise.
5208 (gimple_bind_add_stmt): Likewise.
5209 (gimple_bind_block): Likewise.
5210 (gimple_bind_set_block): Likewise.
5211 (gimple_asm_ninputs): Likewise.
5212 (gimple_asm_noutputs): Likewise.
5213 (gimple_asm_nclobbers): Likewise.
5214 (gimple_asm_nlabels): Likewise.
5215 (gimple_asm_input_op): Likewise.
5216 (gimple_asm_input_op_ptr): Likewise.
5217 (gimple_asm_output_op): Likewise.
5218 (gimple_asm_output_op_ptr): Likewise.
5219 (gimple_asm_set_output_op): Likewise.
5220 (gimple_asm_clobber_op): Likewise.
5221 (gimple_asm_set_clobber_op): Likewise.
5222 (gimple_asm_label_op): Likewise.
5223 (gimple_asm_set_label_op): Likewise.
5224 (gimple_asm_string): Likewise.
5225 (gimple_catch_types): Likewise.
5226 (gimple_catch_types_ptr): Likewise.
5227 (gimple_catch_handler_ptr): Likewise.
5228 (gimple_catch_set_types): Likewise.
5229 (gimple_catch_set_handler): Likewise.
5230 (gimple_eh_filter_types): Likewise.
5231 (gimple_eh_filter_types_ptr): Likewise.
5232 (gimple_eh_filter_failure_ptr): Likewise.
5233 (gimple_eh_filter_set_types): Likewise.
5234 (gimple_eh_filter_set_failure): Likewise.
5235 (gimple_eh_must_not_throw_fndecl): Likewise.
5236 (gimple_eh_must_not_throw_set_fndecl): Likewise.
5237 (gimple_eh_else_n_body_ptr): Likewise.
5238 (gimple_eh_else_e_body_ptr): Likewise.
5239 (gimple_eh_else_set_n_body): Likewise.
5240 (gimple_eh_else_set_e_body): Likewise.
5241 (gimple_try_eval_ptr): Likewise.
5242 (gimple_try_cleanup_ptr): Likewise.
5243 (gimple_try_set_eval): Likewise.
5244 (gimple_try_set_cleanup): Likewise.
5245 (gimple_wce_cleanup_ptr): Likewise.
5246 (gimple_wce_set_cleanup): Likewise.
5247 (gimple_phi_capacity): Likewise.
5248 (gimple_phi_num_args): Likewise.
5249 (gimple_phi_result): Likewise.
5250 (gimple_phi_result_ptr): Likewise.
5251 (gimple_phi_set_result): Likewise.
5252 (gimple_phi_arg): Likewise.
5253 (gimple_phi_set_arg): Likewise.
5254 (gimple_resx_region): Likewise.
5255 (gimple_resx_set_region): Likewise.
5256 (gimple_eh_dispatch_region): Likewise.
5257 (gimple_eh_dispatch_set_region): Likewise.
5258 (gimple_omp_critical_name): Likewise.
5259 (gimple_omp_critical_name_ptr): Likewise.
5260 (gimple_omp_critical_set_name): Likewise.
5261 (gimple_omp_for_clauses): Likewise.
5262 (gimple_omp_for_clauses_ptr): Likewise.
5263 (gimple_omp_for_set_clauses): Likewise.
5264 (gimple_omp_for_collapse): Likewise.
5265 (gimple_omp_for_index): Likewise.
5266 (gimple_omp_for_index_ptr): Likewise.
5267 (gimple_omp_for_set_index): Likewise.
5268 (gimple_omp_for_initial): Likewise.
5269 (gimple_omp_for_initial_ptr): Likewise.
5270 (gimple_omp_for_set_initial): Likewise.
5271 (gimple_omp_for_final): Likewise.
5272 (gimple_omp_for_final_ptr): Likewise.
5273 (gimple_omp_for_set_final): Likewise.
5274 (gimple_omp_for_incr): Likewise.
5275 (gimple_omp_for_incr_ptr): Likewise.
5276 (gimple_omp_for_set_incr): Likewise.
5277 (gimple_omp_for_pre_body_ptr): Likewise.
5278 (gimple_omp_for_set_pre_body): Likewise.
5279 (gimple_omp_parallel_clauses): Likewise.
5280 (gimple_omp_parallel_clauses_ptr): Likewise.
5281 (gimple_omp_parallel_set_clauses): Likewise.
5282 (gimple_omp_parallel_child_fn): Likewise.
5283 (gimple_omp_parallel_child_fn_ptr): Likewise.
5284 (gimple_omp_parallel_set_child_fn): Likewise.
5285 (gimple_omp_parallel_data_arg): Likewise.
5286 (gimple_omp_parallel_data_arg_ptr): Likewise.
5287 (gimple_omp_parallel_set_data_arg): Likewise.
5288 (gimple_omp_task_clauses): Likewise.
5289 (gimple_omp_task_clauses_ptr): Likewise.
5290 (gimple_omp_task_set_clauses): Likewise.
5291 (gimple_omp_task_child_fn): Likewise.
5292 (gimple_omp_task_child_fn_ptr): Likewise.
5293 (gimple_omp_task_set_child_fn): Likewise.
5294 (gimple_omp_task_data_arg): Likewise.
5295 (gimple_omp_task_data_arg_ptr): Likewise.
5296 (gimple_omp_task_set_data_arg): Likewise.
5297 (gimple_omp_taskreg_clauses): Likewise.
5298 (gimple_omp_taskreg_clauses_ptr): Likewise.
5299 (gimple_omp_taskreg_set_clauses): Likewise.
5300 (gimple_omp_taskreg_child_fn): Likewise.
5301 (gimple_omp_taskreg_child_fn_ptr): Likewise.
5302 (gimple_omp_taskreg_set_child_fn): Likewise.
5303 (gimple_omp_taskreg_data_arg): Likewise.
5304 (gimple_omp_taskreg_data_arg_ptr): Likewise.
5305 (gimple_omp_taskreg_set_data_arg): Likewise.
5306 (gimple_omp_task_copy_fn): Likewise.
5307 (gimple_omp_task_copy_fn_ptr): Likewise.
5308 (gimple_omp_task_set_copy_fn): Likewise.
5309 (gimple_omp_task_arg_size): Likewise.
5310 (gimple_omp_task_arg_size_ptr): Likewise.
5311 (gimple_omp_task_set_arg_size): Likewise.
5312 (gimple_omp_task_arg_align): Likewise.
5313 (gimple_omp_task_arg_align_ptr): Likewise.
5314 (gimple_omp_task_set_arg_align): Likewise.
5315 (gimple_omp_single_clauses): Likewise.
5316 (gimple_omp_single_clauses_ptr): Likewise.
5317 (gimple_omp_single_set_clauses): Likewise.
5318 (gimple_omp_target_clauses): Likewise.
5319 (gimple_omp_target_clauses_ptr): Likewise.
5320 (gimple_omp_target_set_clauses): Likewise.
5321 (gimple_omp_target_child_fn): Likewise.
5322 (gimple_omp_target_child_fn_ptr): Likewise.
5323 (gimple_omp_target_set_child_fn): Likewise.
5324 (gimple_omp_target_data_arg): Likewise.
5325 (gimple_omp_target_data_arg_ptr): Likewise.
5326 (gimple_omp_target_set_data_arg): Likewise.
5327 (gimple_omp_teams_clauses): Likewise.
5328 (gimple_omp_teams_clauses_ptr): Likewise.
5329 (gimple_omp_teams_set_clauses): Likewise.
5330 (gimple_omp_sections_clauses): Likewise.
5331 (gimple_omp_sections_clauses_ptr): Likewise.
5332 (gimple_omp_sections_set_clauses): Likewise.
5333 (gimple_omp_sections_control): Likewise.
5334 (gimple_omp_sections_control_ptr): Likewise.
5335 (gimple_omp_sections_set_control): Likewise.
5336 (gimple_omp_for_set_cond): Likewise.
5337 (gimple_omp_for_cond): Likewise.
5338 (gimple_omp_atomic_store_set_val): Likewise.
5339 (gimple_omp_atomic_store_val): Likewise.
5340 (gimple_omp_atomic_store_val_ptr): Likewise.
5341 (gimple_omp_atomic_load_set_lhs): Likewise.
5342 (gimple_omp_atomic_load_lhs): Likewise.
5343 (gimple_omp_atomic_load_lhs_ptr): Likewise.
5344 (gimple_omp_atomic_load_set_rhs): Likewise.
5345 (gimple_omp_atomic_load_rhs): Likewise.
5346 (gimple_omp_atomic_load_rhs_ptr): Likewise.
5347 (gimple_omp_continue_control_def): Likewise.
5348 (gimple_omp_continue_control_def_ptr): Likewise.
5349 (gimple_omp_continue_set_control_def): Likewise.
5350 (gimple_omp_continue_control_use): Likewise.
5351 (gimple_omp_continue_control_use_ptr): Likewise.
5352 (gimple_omp_continue_set_control_use): Likewise.
5353 (gimple_transaction_body_ptr): Likewise.
5354 (gimple_transaction_label): Likewise.
5355 (gimple_transaction_label_ptr): Likewise.
5356 (gimple_transaction_set_body): Likewise.
5357 (gimple_transaction_set_label): Likewise.
5359 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
5360 * ipa-inline-analysis.c (inline_write_summary): Likewise.
5361 * ipa-ref.c (ipa_record_reference): Likewise.
5362 * ipa-reference.c (analyze_function): Likewise.
5363 (ipa_reference_write_optimization_summary): Likewise.
5364 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
5365 (address_taken_from_non_vtable_p): Likewise.
5366 (comdat_can_be_unshared_p_1): Likewise.
5367 * lto-cgraph.c (lto_output_ref): Likewise.
5368 (add_references): Likewise.
5369 (compute_ltrans_boundary): Likewise.
5370 (output_symtab): Likewise.
5371 (input_ref): Likewise.
5372 (input_cgraph_1): Likewise.
5373 (output_cgraph_opt_summary): Likewise.
5374 * lto-streamer-out.c (lto_output): Likewise.
5375 (output_symbol_p): Likewise.
5376 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
5377 (lsei_start_function_in_partition): Likewise.
5378 (lsei_next_variable_in_partition): Likewise.
5379 (lsei_start_variable_in_partition): Likewise.
5380 * symtab.c (insert_to_assembler_name_hash): Likewise.
5381 (unlink_from_assembler_name_hash): Likewise.
5382 (symtab_unregister_node): Likewise.
5383 (symtab_remove_node): Likewise.
5384 (dump_symtab_node): Likewise.
5385 (verify_symtab_base): Likewise.
5386 (verify_symtab_node): Likewise.
5387 (symtab_make_decl_local): Likewise.
5388 (symtab_alias_ultimate_target): Likewise.
5389 (symtab_resolve_alias): Likewise.
5390 (symtab_get_symbol_partitioning_class): Likewise.
5391 * tree-phinodes.c (allocate_phi_node): Likewise.
5392 (reserve_phi_args_for_new_edge): Likewise.
5393 (remove_phi_args): Likewise.
5394 * varpool.c (varpool_node_for_asm): Likewise.
5395 (varpool_remove_unreferenced_decls): Likewise.
5397 2014-04-23 Jeff Law <law@redhat.com>
5399 PR tree-optimization/60902
5400 * tree-ssa-threadedge.c
5401 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
5402 invalidate outputs from statements that do not produce useful
5403 outputs for threading.
5405 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
5407 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
5408 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
5409 machine descriptions for Stack Smashing Protector.
5411 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
5413 * aarch64.md (<optab>_rol<mode>3): New pattern.
5414 (<optab>_rolsi3_uxtw): Likewise.
5415 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
5417 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
5419 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
5420 (arm_cortex_a12_tune): Likewise.
5422 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5424 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
5426 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5428 * config/arm/arm.md (arm_rev16si2): New pattern.
5429 (arm_rev16si2_alt): Likewise.
5430 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
5432 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5434 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
5435 (rev16<mode>2_alt): Likewise.
5436 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
5437 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
5438 (aarch_rev16_shleft_mask_imm_p): Likewise.
5439 (aarch_rev16_p_1): Likewise.
5440 (aarch_rev16_p): Likewise.
5441 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
5442 (aarch_rev16_shright_mask_imm_p): Likewise.
5443 (aarch_rev16_shleft_mask_imm_p): Likewise.
5445 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5447 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
5448 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
5450 (cortex_a53_extra_costs): Likewise.
5451 (cortex_a57_extra_costs): Likewise.
5452 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
5453 (cortexa7_extra_costs): Likewise.
5454 (cortexa8_extra_costs): Likewise.
5455 (cortexa12_extra_costs): Likewise.
5456 (cortexa15_extra_costs): Likewise.
5457 (v7m_extra_costs): Likewise.
5458 (arm_new_rtx_costs): Handle BSWAP.
5460 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5462 * config/arm/arm.c (cortexa8_extra_costs): New table.
5463 (arm_cortex_a8_tune): New tuning struct.
5464 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
5466 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5468 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
5470 2014-04-23 Richard Biener <rguenther@suse.de>
5472 * Makefile.in (OBJS): Remove loop-unswitch.o.
5473 * tree-pass.h (make_pass_rtl_unswitch): Remove.
5474 * passes.def (pass_rtl_unswitch): Likewise.
5475 * loop-init.c (gate_rtl_unswitch): Likewise.
5476 (rtl_unswitch): Likewise.
5477 (pass_data_rtl_unswitch): Likewise.
5478 (pass_rtl_unswitch): Likewise.
5479 (make_pass_rtl_unswitch): Likewise.
5480 * rtl.h (reversed_condition): Likewise.
5481 (compare_and_jump_seq): Likewise.
5482 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
5484 * loop-unroll.c (compare_and_jump_seq): Likewise.
5486 2014-04-23 Richard Biener <rguenther@suse.de>
5488 PR tree-optimization/60903
5489 * tree-ssa-loop-im.c (analyze_memory_references): Remove
5490 commented code block.
5491 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
5492 loop flags to newly created BBs and edges.
5494 2014-04-23 Nick Clifton <nickc@redhat.com>
5496 * config/msp430/msp430.c (msp430_handle_option): Move function
5498 (msp430_option_override): Simplify mcu and mcpu option handling.
5499 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
5500 support for -mhwmult command line option.
5501 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
5502 -mhwmult command line option.
5503 (msp430_hwmult_enabled): Delete.
5504 (msp43o_output_labelref): Add support for -mhwmult command line option.
5505 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
5506 (umulsidi3): Likewise.
5507 * config/msp430/msp430.opt (mmcu): Add Report attribute.
5508 (mcpu, mlarge, msmall): Likewise.
5509 (mhwmult): New option.
5510 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
5512 (msp430_is_f5_mcu): Remove prototype.
5513 (msp430_use_f5_series_hwmult): Add prototype.
5514 * config/msp430/msp430-opts.h: New file.
5515 * common/config/msp430: New directory.
5516 * common/config/msp430/msp430-common.c: New file.
5517 * config.gcc (msp430): Remove target_has_targetm_common.
5518 * doc/invoke.texi: Document -mhwmult command line option.
5520 2014-04-23 Nick Clifton <nickc@redhat.com>
5522 * config/i386/cygwin.h (ENDFILE_SPEC): Include
5523 default-manifest.o if it can be found in the search path.
5524 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
5526 2014-04-23 Terry Guo <terry.guo@arm.com>
5528 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
5530 2014-04-23 Richard Biener <rguenther@suse.de>
5533 * tree-inline.c (declare_return_variable): Use mark_addressable.
5535 2014-04-23 Richard Biener <rguenther@suse.de>
5538 * loop-init.c (loop_optimizer_init): Make sure to apply
5539 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
5541 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5544 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
5546 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
5547 if flag_sanitize_undefined_trap_on_error.
5548 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
5549 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
5550 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
5551 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
5552 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
5553 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
5554 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
5555 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
5556 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
5557 * ubsan.c (ubsan_instrument_unreachable): Return
5558 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
5559 (ubsan_expand_null_ifn): Emit __builtin_trap ()
5560 if flag_sanitize_undefined_trap_on_error and
5561 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
5562 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
5563 instrument_bool_enum_load): Emit __builtin_trap () if
5564 flag_sanitize_undefined_trap_on_error and
5565 __builtin_handle_*_abort () if !flag_sanitize_recover.
5566 * doc/invoke.texi (-fsanitize-recover,
5567 -fsanitize-undefined-trap-on-error): Document.
5569 2014-04-22 Christian Bruel <christian.bruel@st.com>
5571 * config/sh/sh.md (mov<mode>): Replace movQIHI.
5572 Force immediates to SImode.
5574 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
5576 * config/nios2/nios2.md (UNSPEC_ROUND): New.
5578 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
5579 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
5580 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
5581 (nios2_fpu_insn): Add entry for round.
5582 (N2FPU_NO_ERRNO_P): Define.
5583 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
5585 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
5587 2014-04-22 Richard Henderson <rth@redhat.com>
5589 * config/aarch64/aarch64 (addti3, subti3): New expanders.
5590 (add<GPI>3_compare0): Remove leading * from name.
5591 (add<GPI>3_carryin): Likewise.
5592 (sub<GPI>3_compare0): Likewise.
5593 (sub<GPI>3_carryin): Likewise.
5594 (<su_optab>mulditi3): New expander.
5595 (multi3): New expander.
5596 (madd<GPI>): Remove leading * from name.
5598 2014-04-22 Martin Jambor <mjambor@suse.cz>
5600 * cgraphclones.c (cgraph_function_versioning): Copy
5601 ipa_transforms_to_apply instead of asserting it is empty.
5603 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
5606 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
5607 on count_exp to get mode.
5609 2014-04-22 Andrew Pinski <apinski@cavium.com>
5611 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
5612 Handle TLS for ILP32.
5613 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
5614 (tlsie_small_<mode>): this and handle PTR.
5615 (tlsie_small_sidi): New pattern.
5616 (tlsle_small): Change to an expand to handle ILP32.
5617 (tlsle_small_<mode>): New pattern.
5618 (tlsdesc_small): Rename to ...
5619 (tlsdesc_small_<mode>): this and handle PTR.
5621 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5623 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
5625 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5627 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
5628 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
5629 (aarch64_types_signed_poly_qualifiers): Likewise.
5630 (aarch64_types_unsigned_signed_qualifiers): Likewise.
5631 (aarch64_types_poly_signed_qualifiers): Likewise.
5632 (TYPES_REINTERP_SS): Type macro added.
5633 (TYPES_REINTERP_SU): Likewise.
5634 (TYPES_REINTERP_SP): Likewise.
5635 (TYPES_REINTERP_US): Likewise.
5636 (TYPES_REINTERP_PS): Likewise.
5637 (aarch64_fold_builtin): New expression folding added.
5638 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
5639 Declarations removed.
5640 (REINTERP_SS): Declarations added.
5641 (REINTERP_US): Likewise.
5642 (REINTERP_PS): Likewise.
5643 (REINTERP_SU): Likewise.
5644 (REINTERP_SP): Likewise.
5645 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
5646 (vreinterpretq_p8_f64): Likewise.
5647 (vreinterpret_p16_f64): Likewise.
5648 (vreinterpretq_p16_f64): Likewise.
5649 (vreinterpret_f32_f64): Likewise.
5650 (vreinterpretq_f32_f64): Likewise.
5651 (vreinterpret_f64_f32): Likewise.
5652 (vreinterpret_f64_p8): Likewise.
5653 (vreinterpret_f64_p16): Likewise.
5654 (vreinterpret_f64_s8): Likewise.
5655 (vreinterpret_f64_s16): Likewise.
5656 (vreinterpret_f64_s32): Likewise.
5657 (vreinterpret_f64_s64): Likewise.
5658 (vreinterpret_f64_u8): Likewise.
5659 (vreinterpret_f64_u16): Likewise.
5660 (vreinterpret_f64_u32): Likewise.
5661 (vreinterpret_f64_u64): Likewise.
5662 (vreinterpretq_f64_f32): Likewise.
5663 (vreinterpretq_f64_p8): Likewise.
5664 (vreinterpretq_f64_p16): Likewise.
5665 (vreinterpretq_f64_s8): Likewise.
5666 (vreinterpretq_f64_s16): Likewise.
5667 (vreinterpretq_f64_s32): Likewise.
5668 (vreinterpretq_f64_s64): Likewise.
5669 (vreinterpretq_f64_u8): Likewise.
5670 (vreinterpretq_f64_u16): Likewise.
5671 (vreinterpretq_f64_u32): Likewise.
5672 (vreinterpretq_f64_u64): Likewise.
5673 (vreinterpret_s64_f64): Likewise.
5674 (vreinterpretq_s64_f64): Likewise.
5675 (vreinterpret_u64_f64): Likewise.
5676 (vreinterpretq_u64_f64): Likewise.
5677 (vreinterpret_s8_f64): Likewise.
5678 (vreinterpretq_s8_f64): Likewise.
5679 (vreinterpret_s16_f64): Likewise.
5680 (vreinterpretq_s16_f64): Likewise.
5681 (vreinterpret_s32_f64): Likewise.
5682 (vreinterpretq_s32_f64): Likewise.
5683 (vreinterpret_u8_f64): Likewise.
5684 (vreinterpretq_u8_f64): Likewise.
5685 (vreinterpret_u16_f64): Likewise.
5686 (vreinterpretq_u16_f64): Likewise.
5687 (vreinterpret_u32_f64): Likewise.
5688 (vreinterpretq_u32_f64): Likewise.
5690 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5692 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
5693 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
5694 (vreinterpret_p8_s8): Likewise.
5695 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
5696 (vreinterpret_p8_s16): Likewise.
5697 (vreinterpret_p8_s32): Likewise.
5698 (vreinterpret_p8_s64): Likewise.
5699 (vreinterpret_p8_f32): Likewise.
5700 (vreinterpret_p8_u8): Likewise.
5701 (vreinterpret_p8_u16): Likewise.
5702 (vreinterpret_p8_u32): Likewise.
5703 (vreinterpret_p8_u64): Likewise.
5704 (vreinterpret_p8_p16): Likewise.
5705 (vreinterpretq_p8_s8): Likewise.
5706 (vreinterpretq_p8_s16): Likewise.
5707 (vreinterpretq_p8_s32): Likewise.
5708 (vreinterpretq_p8_s64): Likewise.
5709 (vreinterpretq_p8_f32): Likewise.
5710 (vreinterpretq_p8_u8): Likewise.
5711 (vreinterpretq_p8_u16): Likewise.
5712 (vreinterpretq_p8_u32): Likewise.
5713 (vreinterpretq_p8_u64): Likewise.
5714 (vreinterpretq_p8_p16): Likewise.
5715 (vreinterpret_p16_s8): Likewise.
5716 (vreinterpret_p16_s16): Likewise.
5717 (vreinterpret_p16_s32): Likewise.
5718 (vreinterpret_p16_s64): Likewise.
5719 (vreinterpret_p16_f32): Likewise.
5720 (vreinterpret_p16_u8): Likewise.
5721 (vreinterpret_p16_u16): Likewise.
5722 (vreinterpret_p16_u32): Likewise.
5723 (vreinterpret_p16_u64): Likewise.
5724 (vreinterpret_p16_p8): Likewise.
5725 (vreinterpretq_p16_s8): Likewise.
5726 (vreinterpretq_p16_s16): Likewise.
5727 (vreinterpretq_p16_s32): Likewise.
5728 (vreinterpretq_p16_s64): Likewise.
5729 (vreinterpretq_p16_f32): Likewise.
5730 (vreinterpretq_p16_u8): Likewise.
5731 (vreinterpretq_p16_u16): Likewise.
5732 (vreinterpretq_p16_u32): Likewise.
5733 (vreinterpretq_p16_u64): Likewise.
5734 (vreinterpretq_p16_p8): Likewise.
5735 (vreinterpret_f32_s8): Likewise.
5736 (vreinterpret_f32_s16): Likewise.
5737 (vreinterpret_f32_s32): Likewise.
5738 (vreinterpret_f32_s64): Likewise.
5739 (vreinterpret_f32_u8): Likewise.
5740 (vreinterpret_f32_u16): Likewise.
5741 (vreinterpret_f32_u32): Likewise.
5742 (vreinterpret_f32_u64): Likewise.
5743 (vreinterpret_f32_p8): Likewise.
5744 (vreinterpret_f32_p16): Likewise.
5745 (vreinterpretq_f32_s8): Likewise.
5746 (vreinterpretq_f32_s16): Likewise.
5747 (vreinterpretq_f32_s32): Likewise.
5748 (vreinterpretq_f32_s64): Likewise.
5749 (vreinterpretq_f32_u8): Likewise.
5750 (vreinterpretq_f32_u16): Likewise.
5751 (vreinterpretq_f32_u32): Likewise.
5752 (vreinterpretq_f32_u64): Likewise.
5753 (vreinterpretq_f32_p8): Likewise.
5754 (vreinterpretq_f32_p16): Likewise.
5755 (vreinterpret_s64_s8): Likewise.
5756 (vreinterpret_s64_s16): Likewise.
5757 (vreinterpret_s64_s32): Likewise.
5758 (vreinterpret_s64_f32): Likewise.
5759 (vreinterpret_s64_u8): Likewise.
5760 (vreinterpret_s64_u16): Likewise.
5761 (vreinterpret_s64_u32): Likewise.
5762 (vreinterpret_s64_u64): Likewise.
5763 (vreinterpret_s64_p8): Likewise.
5764 (vreinterpret_s64_p16): Likewise.
5765 (vreinterpretq_s64_s8): Likewise.
5766 (vreinterpretq_s64_s16): Likewise.
5767 (vreinterpretq_s64_s32): Likewise.
5768 (vreinterpretq_s64_f32): Likewise.
5769 (vreinterpretq_s64_u8): Likewise.
5770 (vreinterpretq_s64_u16): Likewise.
5771 (vreinterpretq_s64_u32): Likewise.
5772 (vreinterpretq_s64_u64): Likewise.
5773 (vreinterpretq_s64_p8): Likewise.
5774 (vreinterpretq_s64_p16): Likewise.
5775 (vreinterpret_u64_s8): Likewise.
5776 (vreinterpret_u64_s16): Likewise.
5777 (vreinterpret_u64_s32): Likewise.
5778 (vreinterpret_u64_s64): Likewise.
5779 (vreinterpret_u64_f32): Likewise.
5780 (vreinterpret_u64_u8): Likewise.
5781 (vreinterpret_u64_u16): Likewise.
5782 (vreinterpret_u64_u32): Likewise.
5783 (vreinterpret_u64_p8): Likewise.
5784 (vreinterpret_u64_p16): Likewise.
5785 (vreinterpretq_u64_s8): Likewise.
5786 (vreinterpretq_u64_s16): Likewise.
5787 (vreinterpretq_u64_s32): Likewise.
5788 (vreinterpretq_u64_s64): Likewise.
5789 (vreinterpretq_u64_f32): Likewise.
5790 (vreinterpretq_u64_u8): Likewise.
5791 (vreinterpretq_u64_u16): Likewise.
5792 (vreinterpretq_u64_u32): Likewise.
5793 (vreinterpretq_u64_p8): Likewise.
5794 (vreinterpretq_u64_p16): Likewise.
5795 (vreinterpret_s8_s16): Likewise.
5796 (vreinterpret_s8_s32): Likewise.
5797 (vreinterpret_s8_s64): Likewise.
5798 (vreinterpret_s8_f32): Likewise.
5799 (vreinterpret_s8_u8): Likewise.
5800 (vreinterpret_s8_u16): Likewise.
5801 (vreinterpret_s8_u32): Likewise.
5802 (vreinterpret_s8_u64): Likewise.
5803 (vreinterpret_s8_p8): Likewise.
5804 (vreinterpret_s8_p16): Likewise.
5805 (vreinterpretq_s8_s16): Likewise.
5806 (vreinterpretq_s8_s32): Likewise.
5807 (vreinterpretq_s8_s64): Likewise.
5808 (vreinterpretq_s8_f32): Likewise.
5809 (vreinterpretq_s8_u8): Likewise.
5810 (vreinterpretq_s8_u16): Likewise.
5811 (vreinterpretq_s8_u32): Likewise.
5812 (vreinterpretq_s8_u64): Likewise.
5813 (vreinterpretq_s8_p8): Likewise.
5814 (vreinterpretq_s8_p16): Likewise.
5815 (vreinterpret_s16_s8): Likewise.
5816 (vreinterpret_s16_s32): Likewise.
5817 (vreinterpret_s16_s64): Likewise.
5818 (vreinterpret_s16_f32): Likewise.
5819 (vreinterpret_s16_u8): Likewise.
5820 (vreinterpret_s16_u16): Likewise.
5821 (vreinterpret_s16_u32): Likewise.
5822 (vreinterpret_s16_u64): Likewise.
5823 (vreinterpret_s16_p8): Likewise.
5824 (vreinterpret_s16_p16): Likewise.
5825 (vreinterpretq_s16_s8): Likewise.
5826 (vreinterpretq_s16_s32): Likewise.
5827 (vreinterpretq_s16_s64): Likewise.
5828 (vreinterpretq_s16_f32): Likewise.
5829 (vreinterpretq_s16_u8): Likewise.
5830 (vreinterpretq_s16_u16): Likewise.
5831 (vreinterpretq_s16_u32): Likewise.
5832 (vreinterpretq_s16_u64): Likewise.
5833 (vreinterpretq_s16_p8): Likewise.
5834 (vreinterpretq_s16_p16): Likewise.
5835 (vreinterpret_s32_s8): Likewise.
5836 (vreinterpret_s32_s16): Likewise.
5837 (vreinterpret_s32_s64): Likewise.
5838 (vreinterpret_s32_f32): Likewise.
5839 (vreinterpret_s32_u8): Likewise.
5840 (vreinterpret_s32_u16): Likewise.
5841 (vreinterpret_s32_u32): Likewise.
5842 (vreinterpret_s32_u64): Likewise.
5843 (vreinterpret_s32_p8): Likewise.
5844 (vreinterpret_s32_p16): Likewise.
5845 (vreinterpretq_s32_s8): Likewise.
5846 (vreinterpretq_s32_s16): Likewise.
5847 (vreinterpretq_s32_s64): Likewise.
5848 (vreinterpretq_s32_f32): Likewise.
5849 (vreinterpretq_s32_u8): Likewise.
5850 (vreinterpretq_s32_u16): Likewise.
5851 (vreinterpretq_s32_u32): Likewise.
5852 (vreinterpretq_s32_u64): Likewise.
5853 (vreinterpretq_s32_p8): Likewise.
5854 (vreinterpretq_s32_p16): Likewise.
5855 (vreinterpret_u8_s8): Likewise.
5856 (vreinterpret_u8_s16): Likewise.
5857 (vreinterpret_u8_s32): Likewise.
5858 (vreinterpret_u8_s64): Likewise.
5859 (vreinterpret_u8_f32): Likewise.
5860 (vreinterpret_u8_u16): Likewise.
5861 (vreinterpret_u8_u32): Likewise.
5862 (vreinterpret_u8_u64): Likewise.
5863 (vreinterpret_u8_p8): Likewise.
5864 (vreinterpret_u8_p16): Likewise.
5865 (vreinterpretq_u8_s8): Likewise.
5866 (vreinterpretq_u8_s16): Likewise.
5867 (vreinterpretq_u8_s32): Likewise.
5868 (vreinterpretq_u8_s64): Likewise.
5869 (vreinterpretq_u8_f32): Likewise.
5870 (vreinterpretq_u8_u16): Likewise.
5871 (vreinterpretq_u8_u32): Likewise.
5872 (vreinterpretq_u8_u64): Likewise.
5873 (vreinterpretq_u8_p8): Likewise.
5874 (vreinterpretq_u8_p16): Likewise.
5875 (vreinterpret_u16_s8): Likewise.
5876 (vreinterpret_u16_s16): Likewise.
5877 (vreinterpret_u16_s32): Likewise.
5878 (vreinterpret_u16_s64): Likewise.
5879 (vreinterpret_u16_f32): Likewise.
5880 (vreinterpret_u16_u8): Likewise.
5881 (vreinterpret_u16_u32): Likewise.
5882 (vreinterpret_u16_u64): Likewise.
5883 (vreinterpret_u16_p8): Likewise.
5884 (vreinterpret_u16_p16): Likewise.
5885 (vreinterpretq_u16_s8): Likewise.
5886 (vreinterpretq_u16_s16): Likewise.
5887 (vreinterpretq_u16_s32): Likewise.
5888 (vreinterpretq_u16_s64): Likewise.
5889 (vreinterpretq_u16_f32): Likewise.
5890 (vreinterpretq_u16_u8): Likewise.
5891 (vreinterpretq_u16_u32): Likewise.
5892 (vreinterpretq_u16_u64): Likewise.
5893 (vreinterpretq_u16_p8): Likewise.
5894 (vreinterpretq_u16_p16): Likewise.
5895 (vreinterpret_u32_s8): Likewise.
5896 (vreinterpret_u32_s16): Likewise.
5897 (vreinterpret_u32_s32): Likewise.
5898 (vreinterpret_u32_s64): Likewise.
5899 (vreinterpret_u32_f32): Likewise.
5900 (vreinterpret_u32_u8): Likewise.
5901 (vreinterpret_u32_u16): Likewise.
5902 (vreinterpret_u32_u64): Likewise.
5903 (vreinterpret_u32_p8): Likewise.
5904 (vreinterpret_u32_p16): Likewise.
5905 (vreinterpretq_u32_s8): Likewise.
5906 (vreinterpretq_u32_s16): Likewise.
5907 (vreinterpretq_u32_s32): Likewise.
5908 (vreinterpretq_u32_s64): Likewise.
5909 (vreinterpretq_u32_f32): Likewise.
5910 (vreinterpretq_u32_u8): Likewise.
5911 (vreinterpretq_u32_u16): Likewise.
5912 (vreinterpretq_u32_u64): Likewise.
5913 (vreinterpretq_u32_p8): Likewise.
5914 (vreinterpretq_u32_p16): Likewise.
5916 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5918 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
5920 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
5922 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
5923 (vqnegd_s64): Likewise.
5924 (vqabs_s64): Likewise.
5925 (vqabsd_s64): Likewise.
5927 2014-04-22 Richard Henderson <rth@redhat.com>
5929 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
5930 computation to the top of the loop.
5932 2014-04-22 Renlin <renlin.li@arm.com>
5933 Jiong Wang <jiong.wang@arm.com>
5935 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
5936 * config/aarch64/aarch64.c (aarch64_layout_frame)
5937 (aarch64_initial_elimination_offset): Likewise.
5939 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
5941 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
5944 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
5946 * machmode.h (bitwise_mode_for_mode): Declare.
5947 * stor-layout.h (bitwise_type_for_mode): Likewise.
5948 * stor-layout.c (bitwise_mode_for_mode): New function.
5949 (bitwise_type_for_mode): Likewise.
5950 * builtins.c (fold_builtin_memory_op): Use it instead of
5951 int_mode_for_mode and build_nonstandard_integer_type.
5953 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5955 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
5956 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
5957 (*-*-solaris2*): Simplify.
5958 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
5959 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
5960 *-*-solaris2.9* handling.
5962 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
5964 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
5965 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
5967 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
5968 * configure: Regenerate.
5970 * config/i386/sol2-9.h: Remove.
5972 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
5973 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
5974 Remove Solaris 9 references.
5976 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
5978 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
5979 (floatuns<GPI:mode><GPF:mode>2): Remove.
5980 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
5981 and floatuns conversions.
5982 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
5983 and floatuns conversions.
5984 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
5985 (w1,w2): New mode attributes for inequal width conversions.
5987 2014-04-22 Renlin Li <Renlin.Li@arm.com>
5989 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
5990 the output asm format.
5992 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
5994 * config/aarch64/aarch64-simd.md
5995 (aarch64_cm<optab>di): Always split.
5996 (*aarch64_cm<optab>di): New.
5997 (aarch64_cmtstdi): Always split.
5998 (*aarch64_cmtstdi): New.
6000 2014-04-22 Jakub Jelinek <jakub@redhat.com>
6002 PR tree-optimization/60823
6003 * omp-low.c (ipa_simd_modify_function_body): Go through
6004 all SSA_NAMEs and for those refering to vector arguments
6005 which are going to be replaced adjust SSA_NAME_VAR and,
6006 if it is a default definition, change it into a non-default
6007 definition assigned at the beginning of function from new_decl.
6008 (ipa_simd_modify_stmt_ops): Rewritten.
6009 * tree-dfa.c (set_ssa_default_def): When removing default def,
6010 check for NULL loc instead of NULL *loc.
6012 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6014 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
6015 restrictions on core registers for DImode values in Thumb2.
6017 2014-04-22 Ian Bolton <ian.bolton@arm.com>
6019 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
6020 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
6022 2014-04-22 Ian Bolton <ian.bolton@arm.com>
6024 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
6025 (*iordi_notzesidi_di): Likewise.
6026 (*iordi_notsesidi_di): Likewise.
6028 2014-04-22 Ian Bolton <ian.bolton@arm.com>
6030 * config/arm/arm-protos.h (tune_params): New struct members.
6031 * config/arm/arm.c: Initialise tune_params per processor.
6032 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
6033 for speed, based on new tune_params.
6035 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6037 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
6038 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
6039 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
6040 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
6041 * config/aarch64/arm_neon.h (vrnd_f64): Added.
6042 (vrnda_f64): Likewise.
6043 (vrndi_f64): Likewise.
6044 (vrndm_f64): Likewise.
6045 (vrndn_f64): Likewise.
6046 (vrndp_f64): Likewise.
6047 (vrndx_f64): Likewise.
6049 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6051 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
6052 GET_MODE_SIZE argument is enum machine_mode.
6054 2014-04-22 Jakub Jelinek <jakub@redhat.com>
6057 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
6058 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
6060 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
6063 * asan.c (asan_emit_stack_protection): Force the base to align to
6064 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
6065 appropriate bits if STRICT_ALIGNMENT.
6066 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
6068 (expand_used_vars): Leave a space in the stack frame for alignment
6069 if STRICT_ALIGNMENT.
6071 2014-04-21 David Malcolm <dmalcolm@redhat.com>
6073 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
6075 (gimple_store_p): Likewise.
6076 (gimple_assign_load_p): Likewise.
6077 (gimple_assign_cast_p): Likewise.
6078 (gimple_clobber_p): Likewise.
6080 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
6081 rather than a gimple.
6082 (gimple_assign_cast_p): Likewise.
6084 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
6087 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
6088 If mode is DDmode and TARGET_E500_DOUBLE allow move.
6090 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
6091 more debug information for E500 if -mdebug=reg.
6093 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
6096 * config/i386/i386.c (ix86_expand_builtin)
6097 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
6098 register for target RTX.
6099 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
6101 2014-04-18 Cong Hou <congh@google.com>
6103 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
6104 the widen-mult pattern by handling two operands with different sizes,
6105 and operands whose size is smaller than half of the result type.
6107 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
6109 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
6110 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
6111 (do_estimate_edge_time): Compute it.
6112 * ipa-inline.c (want_inline_small_function_p): Bypass
6113 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
6115 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
6117 * ipa-inline.c (spec_rem): New static variable.
6118 (dump_overall_stats): New function.
6119 (dump_inline_stats): New function.
6121 2014-04-18 Richard Henderson <rth@redhat.com>
6123 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
6124 to GET_MODE_SIZE, not a reg_class_t.
6126 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6128 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
6129 (vsx_xxmrglw_<mode>): Likewise.
6131 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
6134 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
6135 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
6136 (rs6000_init_hard_regno_mode_ok): Likewise.
6138 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
6140 * ipa-inline.c (inline_small_functions): Account only non-cold
6142 * doc/invoke.texi (inline-unit-growth): Update documentation.
6144 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
6146 * config/rs6000/rs6000.md (addti3, subti3): New.
6148 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
6151 * config/i386/i386.c (ix86_expand_clear): Remove outdated
6152 comment. Check optimize_insn_for_size_p instead of
6153 optimize_insn_for_speed_p.
6155 2014-04-17 Martin Jambor <mjambor@suse.cz>
6157 * gimple-iterator.c (gsi_start_edge): New function.
6158 * gimple-iterator.h (gsi_start_edge): Declare.
6159 * tree-sra.c (single_non_eh_succ): New function.
6160 (disqualify_ops_if_throwing_stmt): Renamed to
6161 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
6162 having one non-EH successor BB.
6163 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
6164 generate loads into replacements.
6165 (sra_modify_assign): Likewise and and also use the simple path for
6167 (sra_modify_function_body): Commit statements on edges.
6169 2014-04-17 Richard Biener <rguenther@suse.de>
6172 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
6173 comparison results and add clarifying comment.
6175 2014-04-17 Jakub Jelinek <jakub@redhat.com>
6177 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
6178 (blank_mode): Initialize it.
6179 (emit_mode_size_inline, emit_mode_nunits_inline,
6180 emit_mode_inner_inline): New functions.
6181 (emit_insn_modes_h): Call them and surround their output with
6182 #if GCC_VERSION >= 4001 ... #endif.
6183 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
6184 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
6185 mode_* arrays if the argument is __builtin_constant_p.
6186 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
6187 is enum machine_mode.
6189 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6191 * passes.c (opt_pass::execute): Adjust.
6192 (pass_manager::execute_pass_mode_switching): Likewise.
6193 (early_local_passes::execute): Likewise.
6194 (execute_one_pass): Pass cfun to the pass's execute method.
6195 * tree-pass.h (opt_pass::execute): Add function * argument.
6196 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
6197 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
6198 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
6199 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6200 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
6201 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
6202 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
6203 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
6204 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
6205 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
6206 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
6207 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
6208 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
6209 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
6210 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
6211 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
6212 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
6213 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
6214 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
6215 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6216 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6217 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6218 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6219 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6220 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6221 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6222 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
6223 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
6224 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
6227 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6229 * passes.c (opt_pass::gate): Take function * argument.
6230 (gate_all_early_local_passes): Merge into
6231 (early_local_passes::gate): this.
6232 (gate_all_early_optimizations): Merge into
6233 (all_early_optimizations::gate): this.
6234 (gate_all_optimizations): Mege into
6235 (all_optimizations::gate): this.
6236 (gate_all_optimizations_g): Merge into
6237 (all_optimizations_g::gate): this.
6238 (gate_rest_of_compilation): Mege into
6239 (rest_of_compilation::gate): this.
6240 (gate_postreload): Merge into
6241 (postreload::gate): this.
6242 (dump_one_pass): Pass cfun to the pass's gate method.
6243 (execute_ipa_summary_passes): Likewise.
6244 (execute_one_pass): Likewise.
6245 (ipa_write_summaries_2): Likewise.
6246 (ipa_write_optimization_summaries_1): Likewise.
6247 (ipa_read_summaries_1): Likewise.
6248 (ipa_read_optimization_summaries_1): Likewise.
6249 (execute_ipa_stmt_fixups): Likewise.
6250 * tree-pass.h (opt_pass::gate): Add function * argument.
6251 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
6252 combine-stack-adj.c, combine.c, compare-elim.c,
6253 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6254 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
6255 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
6256 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
6257 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
6258 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
6259 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
6260 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
6261 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
6262 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
6263 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
6264 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
6265 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
6266 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
6267 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6268 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6269 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6270 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6271 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6272 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6273 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6274 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6275 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
6276 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
6277 var-tracking.c, vtable-verify.c, web.c: Adjust.
6279 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6281 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
6282 * configure: Regenerate.
6284 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6286 * passes.c (dump_one_pass): don't check pass->has_gate.
6287 (execute_ipa_summary_passes): Likewise.
6288 (execute_one_pass): Likewise.
6289 (ipa_write_summaries_2): Likewise.
6290 (ipa_write_optimization_summaries_1): Likewise.
6291 (ipa_read_optimization_summaries_1): Likewise.
6292 (execute_ipa_stmt_fixups): Likewise.
6293 * tree-pass.h (pass_data::has_gate): Remove.
6294 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
6295 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
6296 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
6297 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6298 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
6299 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
6300 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
6301 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
6302 gimple-low.c, gimple-ssa-isolate-paths.c,
6303 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
6304 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
6305 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
6306 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
6307 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
6308 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
6309 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
6310 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
6311 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
6312 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
6313 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
6314 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6315 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6316 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6317 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6318 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6319 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6320 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6321 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6322 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
6323 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
6324 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
6327 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6329 * pass_manager.h (pass_manager::register_dump_files_1): Remove
6331 * passes.c (pass_manager::register_dump_files_1): Merge into
6332 (pass_manager::register_dump_files): this, and remove its handling of
6333 properties since the pass always has the properties anyway.
6334 (pass_manager::pass_manager): Adjust.
6336 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6338 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
6339 * passes.c (pass_manager::register_dump_files_1): Remove dead code
6340 dealing with properties.
6341 (pass_manager::register_dump_files): Adjust.
6343 2014-03-20 Mark Wielaard <mjw@redhat.com>
6345 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
6346 then represent the bound as normal constant value.
6348 2014-04-17 Jakub Jelinek <jakub@redhat.com>
6351 Forward port from 4.8 branch
6352 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
6354 * config/i386/bmiintrin.h (_blsi_u32): New.
6358 (_blsmsk_u32): Ditto.
6359 (_blsmsk_u64): Ditto.
6360 (_tzcnt_u32): Ditto.
6361 (_tzcnt_u64): Ditto.
6363 2014-04-17 Kito Cheng <kito@0xlab.org>
6365 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
6367 2014-04-17 Richard Biener <rguenther@suse.de>
6370 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
6371 boolean results for comparisons.
6373 2014-04-17 Richard Biener <rguenther@suse.de>
6375 PR tree-optimization/60836
6376 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
6377 initial PHI args to be gimple values.
6379 2014-04-17 Richard Biener <rguenther@suse.de>
6381 PR tree-optimization/60841
6382 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
6383 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
6384 of stmts to SLP build.
6385 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
6386 (vect_analyze_slp): Likewise.
6387 (vect_analyze_slp_instance): Likewise.
6388 (vect_build_slp_tree): Limit overall SLP tree growth.
6389 * tree-vectorizer.h (vect_analyze_data_refs,
6390 vect_analyze_slp): Adjust prototypes.
6392 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6394 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
6397 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6399 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
6400 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
6401 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
6402 for TARGET_SLOW_PSHUFB
6404 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6406 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
6407 * config/i386/i386.c (intel_cost): Ditto.
6409 2014-04-17 Joey Ye <joey.ye@arm.com>
6411 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
6413 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6415 * opts.c (common_handle_option): Disable -fipa-reference coorectly
6418 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6420 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
6421 (type_all_derivations_known_p): New predicate.
6422 (type_all_ctors_visible_p): New predicate.
6423 (type_possibly_instantiated_p): New predicate.
6424 (get_odr_type): Compute all_derivations_known.
6425 (dump_odr_type): Dump the flag.
6426 (maybe_record_type): Cleanup.
6427 (record_target_from_binfo): Add bases_to_consider array;
6428 record bases for types w/o instances and skip CXX destructor.
6429 (possible_polymorphic_call_targets_1): Add bases_to_consider
6430 and consider_construction parameters; check if type may have instance.
6431 (get_polymorphic_call_info): Set maybe_in_construction to true
6432 when we know nothing.
6433 (record_targets_from_bases): Skip CXX destructors; they are
6434 never called for types in construction.
6435 (possible_polymorphic_call_targets): Do not record target when
6436 type may not have instance.
6438 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6441 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
6442 external aliases alive, too.
6444 2014-04-16 Andrew Pinski <apinski@cavium.com>
6446 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
6449 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
6451 * final.c (compute_alignments): Do not apply loop alignment to a block
6452 falling through to the exit.
6454 2014-04-16 Catherine Moore <clm@codesourcery.com>
6456 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
6457 Adjust constraints for microMIPS store patterns.
6459 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
6461 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
6463 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
6465 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
6466 (append_use): Run at -O0.
6467 (append_vdef): Likewise.
6468 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
6469 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
6471 2014-04-16 Jakub Jelinek <jakub@redhat.com>
6473 PR tree-optimization/60844
6474 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
6475 (propagate_op_to_single_use, remove_visited_stmt_chain,
6476 linearize_expr, repropagate_negates, reassociate_bb): Use it
6477 instead of gsi_remove.
6479 2014-04-16 Martin Jambor <mjambor@suse.cz>
6481 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
6482 ipa_transforms_to_apply.
6483 (cgraph_function_versioning): Assert that old_node has empty
6484 ipa_transforms_to_apply.
6485 * trans-mem.c (ipa_tm_create_version): Likewise.
6486 * tree-inline.c (tree_function_versioning): Do not duplicate
6487 ipa_transforms_to_apply.
6489 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6492 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
6494 Pass necessary as flags on 64-bit Solaris/x86.
6495 Use lowercase relocs for x86_64-*-*.
6496 * configure: Regenerate.
6498 2014-04-15 Jan Hubicka <jh@suse.cz>
6500 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
6501 (maybe_record_node, likely_target_p): Use it.
6503 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6506 Revert following patch
6508 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6511 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
6512 software floating point or no floating point registers, do not
6513 allow any type in the FPRs. Eliminate a test for SPE SIMD types
6514 in GPRs that occurs after we tested for GPRs that would never be
6517 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
6518 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
6519 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
6520 specifically allow DDmode, since that does not use the SPE SIMD
6523 2014-03-21 Mark Wielaard <mjw@redhat.com>
6525 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
6526 as unsigned or int depending on type and value used.
6528 2014-04-15 Richard Biener <rguenther@suse.de>
6530 PR rtl-optimization/56965
6531 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
6532 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
6534 * alias.c (true_dependence_1): Do not call
6535 nonoverlapping_component_refs_p.
6536 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
6537 nonoverlapping_component_refs_p.
6538 (indirect_refs_may_alias_p): Likewise.
6540 2014-04-15 Teresa Johnson <tejohnson@google.com>
6542 * cfg.c (dump_bb_info): Fix flags check.
6543 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
6545 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6547 PR rtl-optimization/60663
6548 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
6551 2014-04-15 Richard Biener <rguenther@suse.de>
6553 * lto-streamer.h (LTO_major_version): Bump to 4.
6555 2014-04-15 Richard Biener <rguenther@suse.de>
6557 * common.opt (lto_partition_model): New enum.
6558 (flto-partition=): Merge separate options with a single with argument,
6559 add -flto-partition=one support.
6560 * flag-types.h (enum lto_partition_model): Declare.
6561 * opts.c (finish_options): Remove duplicate -flto-partition=
6563 * lto-wrapper.c (run_gcc): Adjust.
6565 2014-04-15 Richard Biener <rguenther@suse.de>
6567 * alias.c (ncr_compar): New function.
6568 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
6570 2014-04-15 Richard Biener <rguenther@suse.de>
6572 * alias.c (record_component_aliases): Do not walk BINFOs.
6574 2014-04-15 Richard Biener <rguenther@suse.de>
6576 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
6577 Add struct function argument and adjust.
6578 (find_func_aliases_for_call): Likewise.
6579 (find_func_aliases): Likewise.
6580 (find_func_clobbers): Likewise.
6581 (intra_create_variable_infos): Likewise.
6582 (compute_points_to_sets): Likewise.
6583 (ipa_pta_execute): Adjust. Do not push/pop cfun.
6585 2014-04-15 Richard Biener <rguenther@suse.de>
6587 * tree.c (iterative_hash_expr): Use enum tree_code_class
6588 to store TREE_CODE_CLASS.
6589 (tree_block): Likewise.
6590 (tree_set_block): Likewise.
6591 * tree.h (fold_build_pointer_plus_loc): Use
6592 convert_to_ptrofftype_loc.
6594 2014-04-15 Jakub Jelinek <jakub@redhat.com>
6597 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
6600 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
6602 * cfgloop.h (struct loop): Move force_vectorize down.
6603 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
6604 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
6605 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
6606 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
6607 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
6608 * tree-core.h (enum annot_expr_kind): Add new kind values.
6609 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
6610 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
6612 * tree.def (ANNOTATE_EXPR): Tweak comment.
6614 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6616 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
6619 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
6621 * tree.h (TYPE_IDENTIFIER): Declare.
6622 * tree.c (subrange_type_for_debug_p): Use it.
6623 * godump.c (go_format_type): Likewise.
6624 * dwarf2out.c (is_cxx_auto, modified_type_die,
6625 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
6626 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
6628 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6631 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
6633 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
6635 * config/i386/i386.c (examine_argument): Return bool. Return true if
6636 parameter should be passed in memory.
6637 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
6638 (construct_container): Update calls to examine_argument.
6639 (function_arg_advance_64): Ditto.
6640 (return_in_memory_32): Merge with ix86_return_in_memory.
6641 (return_in_memory_64): Ditto.
6642 (return_in_memory_ms_64): Ditto.
6644 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6646 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
6647 * coverage.c (coverage_compute_profile_id): Handle externally visible
6650 2014-04-14 Martin Jambor <mjambor@suse.cz>
6652 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
6653 DECL_DISREGARD_INLINE_LIMITS functions.
6655 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
6658 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
6660 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
6663 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
6664 optimize_insn_for_speed_p instead of
6665 optimize_function_for_speed_p.
6667 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
6669 * doc/invoke.texi (free): Document AArch64.
6671 2014-04-14 Richard Biener <rguenther@suse.de>
6673 PR tree-optimization/60042
6674 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
6675 (insert_into_preds_of_block): Do not prevent PHI insertion
6676 for REFERENCE exprs here ...
6677 (eliminate_dom_walker::before_dom_children): ... but prevent
6678 their use here under similar conditions when applied to the
6679 IL after PRE optimizations.
6681 2014-04-14 Richard Biener <rguenther@suse.de>
6683 * passes.def: Move early points-to after early SRA.
6685 2014-04-14 Richard Biener <rguenther@suse.de>
6687 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
6688 check for which sign-changes we allow when forwarding
6689 a converted value into a switch.
6691 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
6693 * stor-layout.c (place_field): Finalize non-constant offset for the
6696 2014-04-14 Richard Biener <rguenther@suse.de>
6698 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
6700 (expand_switch_using_bit_tests_p): Likewise.
6701 (process_switch): Compute and pass on speed_p based on the
6703 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
6704 optimize_bb_for_speed_p.
6706 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
6708 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
6709 * function.h (struct function): Rename has_force_vect_loops into
6710 has_force_vectorize_loops.
6711 * lto-streamer-in.c (input_cfg): Adjust for renaming.
6712 (input_struct_function_base): Likewise.
6713 * lto-streamer-out.c (output_cfg): Likewise.
6714 (output_struct_function_base): Likewise.
6715 * omp-low.c (expand_omp_simd): Likewise.
6716 * tree-cfg.c (move_sese_region_to_fn): Likewise.
6717 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
6718 (version_loop_for_if_conversion): Likewise.
6719 (tree_if_conversion): Likewise.
6720 (main_tree_if_conversion): Likewise.
6721 (gate_tree_if_conversion): Likewise.
6722 * tree-inline.c (copy_loops): Likewise.
6723 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
6724 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
6725 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
6726 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
6727 * tree-vectorizer.c (vectorize_loops): Likewise.
6728 * tree-vectorizer.h (unlimited_cost_model): Likewise.
6730 2014-04-14 Richard Biener <rguenther@suse.de>
6733 * lto-streamer-out.c (wrap_refs): New function.
6734 (lto_output): Wrap symbol references in global initializes in
6735 type-preserving MEM_REFs.
6737 2014-04-14 Christian Bruel <christian.bruel@st.com>
6739 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
6741 2014-04-14 Christian Bruel <christian.bruel@st.com>
6743 * config/sh/sh.md (setmemqi): New expand pattern.
6744 * config/sh/sh.h (CLEAR_RATIO): Define.
6745 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
6746 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
6748 2014-04-14 Richard Biener <rguenther@suse.de>
6751 * fold-const.c (negate_expr_p): Don't negate directional rounding
6753 (fold_negate_expr): Likewise.
6755 2014-04-14 Richard Biener <rguenther@suse.de>
6757 PR tree-optimization/59817
6758 PR tree-optimization/60453
6759 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
6760 recursion to catch all CHRECs in the scalar evolution and restrict
6761 the predicate for the remains appropriately.
6763 2014-04-12 Catherine Moore <clm@codesourcery.com>
6765 * config/mips/constraints.md: Add new register constraint "kb".
6766 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
6767 (*movhi_internal): Likewise.
6768 (*movqi_internal): Likewise.
6769 * config/mips/mips.h (M16_STORE_REGS): New register class.
6770 (REG_CLASS_NAMES): Add M16_STORE_REGS.
6771 (REG_CLASS_CONTENTS): Likewise.
6772 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
6774 2014-04-11 Tobias Burnus <burnus@net-b.de>
6777 * doc/invoke.texi (-Wformat-signedness): Document it.
6778 (Wformat=2): Mention that this enables -Wformat-signedness.
6780 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6782 * common/config/epiphany/epiphany-common.c
6783 (epiphany_option_optimization_table): Enable section anchors by
6784 default at -O1 or higher.
6785 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
6786 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
6787 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
6788 carries no extra cost.
6789 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
6790 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
6791 * config/epiphany/predicates.md (memclob_operand): New predicate.
6792 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
6793 Use memclob_operand predicate and X constraint for operand 3.
6795 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6797 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
6798 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
6801 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6803 PR rtl-optimization/60651
6804 * mode-switching.c (optimize_mode_switching): Make sure to emit
6805 sets of a lower numbered entity before sets of a higher numbered
6806 entity to a mode of the same or lower priority.
6807 When creating a seginfo for a basic block that starts with a code
6808 label, move the insertion point past the code label.
6809 (new_seginfo): Document and enforce requirement that
6810 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
6811 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
6812 * doc/tm.texi: Regenerate.
6814 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
6817 * config/arc/arc.c (arc_save_restore): Fix assert typo.
6819 2013-04-11 Jakub Jelinek <jakub@redhat.com>
6821 * BASE-VER: Set to 4.10.0.
6823 2014-04-11 Tobias Burnus <burnus@net-b.de>
6826 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
6827 * doc/gcc.texi (Service): Update description in the @menu
6828 * doc/invoke.texi (Option Summary): Remove misplaced and
6831 2014-04-11 Steve Ellcey <sellcey@mips.com>
6832 Jakub Jelinek <jakub@redhat.com>
6835 * expr.c (convert_move): Use emit_store_flag_force instead of
6836 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
6839 2014-04-11 Richard Biener <rguenther@suse.de>
6842 * varasm.c (assemble_alias): Avoid endless error reporting
6843 recursion by setting TREE_ASM_WRITTEN.
6845 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6847 * config/s390/s390.md: Add a splitter for NOT rtx.
6849 2014-04-11 Jakub Jelinek <jakub@redhat.com>
6851 PR rtl-optimization/60663
6852 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
6854 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
6855 Jakub Jelinek <jakub@redhat.com>
6858 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
6859 flag from decl_node to node.
6861 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6864 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
6865 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
6866 ameliorating the cases where it can be.
6868 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
6871 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
6873 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
6874 (loadsync_<mode>): Change mode.
6875 (load_quadpti, store_quadpti): New.
6876 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
6877 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
6878 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
6880 2014-04-09 Cong Hou <congh@google.com>
6883 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
6886 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6888 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
6889 instead of vnor to exploit possible fusion opportunity in the
6891 (altivec_expand_vec_perm_const_le): Likewise.
6893 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
6895 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
6896 (loadsync_<mode>): Change mode.
6897 (load_quadpti, store_quadpti): New.
6898 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
6899 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
6901 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
6904 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
6905 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
6906 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
6908 2014-04-08 Richard Biener <rguenther@suse.de>
6911 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
6912 a 64bit widest int print double-int similar to on HWI64 hosts.
6914 2014-04-08 Richard Biener <rguenther@suse.de>
6916 PR tree-optimization/60785
6917 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
6918 default defs properly.
6920 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
6922 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
6923 (Weffc++): Likewise.
6925 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
6927 * ipa-devirt.c (maybe_record_node): When node is not recorded,
6928 set completep to false rather than true.
6930 2014-04-07 Douglas B Rupp <rupp@adacore.com>
6933 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
6934 ARM_TARGET2_DWARF_FORMAT.
6936 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
6939 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
6940 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
6943 2014-04-07 Richard Biener <rguenther@suse.de>
6945 PR tree-optimization/60766
6946 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
6947 (may_eliminate_iv): Convert cand_value_at result to desired type.
6949 2014-04-07 Jason Merrill <jason@redhat.com>
6952 * common.opt (-fno-gnu-unique): Add.
6953 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
6955 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6957 * haifa-sched.c: Fix outdated function reference and minor
6958 grammar errors in introductory comment.
6960 2014-04-07 Richard Biener <rguenther@suse.de>
6963 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
6965 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
6967 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
6970 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
6971 size accounting for thunks.
6972 (pa_asm_output_mi_thunk): Use final_start_function() and
6973 final_end_function() to output function start and end directives.
6975 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
6977 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
6978 device specific ISA/ feature information. Remove short_sp and
6979 errata_skip ds. Add avr_device_specific_features enum to have device
6981 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
6982 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
6983 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
6984 updated device specific info.
6985 * config/avr/avr-mcus.def: Merge device specific details to
6986 dev_attribute field.
6987 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
6989 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
6990 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
6991 assembler if RMW isa supported by current device.
6992 * config/avr/genmultilib.awk: Update as device info structure changed.
6993 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
6995 2014-04-04 Cong Hou <congh@google.com>
6997 PR tree-optimization/60656
6998 * tree-vect-stmts.c (supportable_widening_operation):
6999 Fix a bug that elements in a vector with vect_used_by_reduction
7000 property are incorrectly reordered when the operation on it is not
7001 consistant with the one in reduction operation.
7003 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
7005 PR rtl-optimization/60155
7006 * gcse.c (record_set_data): New function.
7007 (single_set_gcse): New function.
7008 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
7009 (hoist_code): Likewise.
7010 (get_pressure_class_and_nregs): Likewise.
7012 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
7014 * explow.c (probe_stack_range): Emit a final optimization blockage.
7016 2014-04-04 Anthony Green <green@moxielogic.com>
7018 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
7021 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
7024 * lto-cgraph.c (input_overwrite_node): Check that partitioning
7025 flags are set only during streaming.
7026 * ipa.c (process_references, walk_polymorphic_call_targets,
7027 symtab_remove_unreachable_nodes): Drop bodies of always inline
7028 after early inlining.
7029 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
7031 2014-04-04 Jakub Jelinek <jakub@redhat.com>
7032 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7035 * dwarf2out.c (const_ok_for_output_1): Reject expressions
7038 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7041 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
7043 (cortex_a53_fdivd): Likewise.
7045 2014-04-04 Martin Jambor <mjambor@suse.cz>
7048 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
7050 * cgraph.c (clone_of_p): Also return true if thunks match.
7051 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
7052 cgraph_function_or_thunk_node and an obsolete comment.
7053 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
7055 (build_function_decl_skip_args): Likewise.
7056 (set_new_clone_decl_and_node_flags): New function.
7057 (duplicate_thunk_for_node): Likewise.
7058 (redirect_edge_duplicating_thunks): Likewise.
7059 (cgraph_clone_node): New parameter args_to_skip, pass it to
7060 redirect_edge_duplicating_thunks which is called instead of
7061 cgraph_redirect_edge_callee.
7062 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
7063 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
7065 2014-04-04 Jeff Law <law@redhat.com>
7068 * config/arm/predicates.md (const_int_I_operand): New predicate.
7069 (const_int_M_operand): Similarly.
7070 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
7072 (insv_t2, extv_reg, extzv_t2): Likewise.
7073 (load_multiple_with_writeback): Similarly for const_int_I_operand.
7074 (pop_multiple_with_writeback_and_return): Likewise.
7075 (vfp_pop_multiple_with_writeback): Likewise
7077 2014-04-04 Richard Biener <rguenther@suse.de>
7080 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
7081 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
7083 * gimplify.h (gimple_add_tmp_var_fn): Declare.
7084 * gimplify.c (gimple_add_tmp_var_fn): New function.
7085 * gimple-expr.h (create_tmp_reg_fn): Declare.
7086 * gimple-expr.c (create_tmp_reg_fn): New function.
7087 * gimple-low.c (record_vars_into): Don't change cfun.
7088 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
7089 code generation without cfun.
7091 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
7094 * Makefile.in (install-driver): Fix shell scripting.
7096 2014-04-03 Cong Hou <congh@google.com>
7098 PR tree-optimization/60505
7099 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
7100 threshold of number of iterations below which no vectorization
7102 * tree-vect-loop.c (new_loop_vec_info):
7103 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
7104 * tree-vect-loop.c (vect_analyze_loop_operations):
7105 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
7106 * tree-vect-loop.c (vect_transform_loop):
7107 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
7108 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
7109 of iterations of the loop and see if we should build the epilogue.
7111 2014-04-03 Richard Biener <rguenther@suse.de>
7113 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
7114 (streamer_tree_cache_create): Adjust.
7115 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
7116 to allow optional nodes array.
7117 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
7118 (streamer_tree_cache_append): Likewise.
7119 (streamer_tree_cache_create): Create nodes array optionally
7120 as specified by parameter.
7121 * lto-streamer-out.c (create_output_block): Avoid maintaining
7122 the node array in the writer cache.
7123 (DFS_write_tree): Remove assertion.
7124 (produce_asm_for_decls): Free the out decl state hash table early.
7125 * lto-streamer-in.c (lto_data_in_create): Adjust for
7126 streamer_tree_cache_create prototype change.
7128 2014-04-03 Richard Biener <rguenther@suse.de>
7130 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
7131 set TREE_CHAIN to NULL_TREE.
7133 2014-04-03 Richard Biener <rguenther@suse.de>
7135 PR tree-optimization/60740
7136 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
7137 over all GIMPLE_COND operands.
7139 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
7141 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
7142 (Weffc++): Remove Scott's numbering, merge lists and reference
7145 2014-04-03 Nick Clifton <nickc@redhat.com>
7147 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
7150 2014-04-03 Martin Jambor <mjambor@suse.cz>
7152 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
7153 mention gcc_unreachable before failing.
7154 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
7157 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
7160 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
7161 inconsistent code and instead mark the context inconsistent.
7162 (possible_polymorphic_call_targets): For inconsistent contexts
7163 return empty complete list.
7165 2014-04-02 Anthony Green <green@moxielogic.com>
7167 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
7168 (extendqisi2, extendhisi2): Define.
7169 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
7170 (WCHAR_TYPE): Change to unsigned int.
7172 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7174 PR tree-optimization/60733
7175 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
7176 insertion point for PHI candidates to be the end of the feeding
7177 block for the PHI argument.
7179 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
7181 PR rtl-optimization/60650
7182 * lra-constraints.c (process_alt_operands): Decrease reject for
7183 earlyclobber matching.
7185 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7187 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
7189 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7191 * config/spu/spu.c (pad_bb): Do not crash when the last
7192 insn is CODE_FOR_blockage.
7194 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7196 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
7197 lies outside the target mode.
7199 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
7202 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
7203 software floating point or no floating point registers, do not
7204 allow any type in the FPRs. Eliminate a test for SPE SIMD types
7205 in GPRs that occurs after we tested for GPRs that would never be
7208 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
7209 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
7210 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
7211 specifically allow DDmode, since that does not use the SPE SIMD
7214 2014-04-02 Richard Biener <rguenther@suse.de>
7217 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
7218 MODE_INTs. Properly use negv_optab.
7219 (expand_abs): Likewise.
7221 2014-04-02 Richard Biener <rguenther@suse.de>
7224 * Makefile.in (install-driver): Guard extra installs with special
7227 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
7229 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7232 2014-04-01 Richard Henderson <rth@redhat.com>
7235 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
7236 alternative enabled before register allocation.
7238 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
7240 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
7241 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
7243 (nios2_large_got_address): Remove unneeded 'sym' parameter.
7244 (nios2_got_address): Update nios2_large_got_address call site.
7245 (nios2_delegitimize_address): New function.
7246 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
7247 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
7248 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
7250 2014-04-01 Martin Husemann <martin@duskware.de>
7252 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
7255 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
7257 PR rtl-optimization/60604
7258 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
7259 check from register_operand.
7260 (register_operand): Redefine in terms of general_operand.
7261 (nonmemory_operand): Use register_operand for the non-constant cases.
7263 2014-04-01 Richard Biener <rguenther@suse.de>
7265 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
7267 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
7269 * doc/invoke.texi (mapp-regs): Clarify.
7271 2014-03-31 Ulrich Drepper <drepper@gmail.com>
7273 * config/i386/avx512fintrin.h (__v32hi): Define type.
7274 (__v64qi): Likewise.
7275 (_mm512_set1_epi8): Define.
7276 (_mm512_set1_epi16): Define.
7277 (_mm512_set4_epi32): Define.
7278 (_mm512_set4_epi64): Define.
7279 (_mm512_set4_pd): Define.
7280 (_mm512_set4_ps): Define.
7281 (_mm512_setr4_epi64): Define.
7282 (_mm512_setr4_epi32): Define.
7283 (_mm512_setr4_pd): Define.
7284 (_mm512_setr4_ps): Define.
7285 (_mm512_setzero_epi32): Define.
7287 2014-03-31 Martin Jambor <mjambor@suse.cz>
7290 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
7291 callsite_arguments_match_p. Updated all callers. Also check types of
7292 corresponding formal parameters and actual arguments.
7293 (not_all_callers_have_enough_arguments_p) Renamed to
7294 some_callers_have_mismatched_arguments_p.
7296 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
7298 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
7300 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
7303 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
7306 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
7308 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
7309 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
7311 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
7312 Use FMAMODE_NOVF512 mode iterator.
7313 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
7314 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
7315 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
7316 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
7318 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
7319 Use VF_128_256 mode iterator.
7320 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
7323 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7325 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
7326 static chain if needed.
7328 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
7331 * lra-constraints.c (index_part_to_reg): New.
7332 (process_address): Use it.
7334 2014-03-27 Jeff Law <law@redhat.com>
7335 Jakub Jelinek <jakub@redhat.com>
7338 * expr.c (do_tablejump): Use simplify_gen_binary rather than
7339 gen_rtx_{PLUS,MULT} to build up the address expression.
7341 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
7342 creating non-canonical RTL.
7344 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7347 * ipa-inline.c (want_inline_small_function_p): Short circuit large
7348 functions; reorganize to make cheap checks first.
7349 (inline_small_functions): Do not estimate growth when dumping;
7351 * ipa-inline.h (inline_summary): Add min_size.
7352 (growth_likely_positive): New function.
7353 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
7354 (set_cond_stmt_execution_predicate): Cleanup.
7355 (estimate_edge_size_and_time): Compute min_size.
7356 (estimate_calls_size_and_time): Likewise.
7357 (estimate_node_size_and_time): Likewise.
7358 (inline_update_overall_summary): Update min_size.
7359 (do_estimate_edge_time): Likewise.
7360 (do_estimate_edge_size): Update.
7361 (do_estimate_edge_hints): Update.
7362 (growth_likely_positive): New function.
7364 2014-03-28 Jakub Jelinek <jakub@redhat.com>
7367 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
7368 also if addr has VOIDmode.
7370 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7372 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
7373 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
7375 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
7376 instructions as well as AdvancedSIMD loads.
7378 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7380 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
7381 Use crypto_aese type.
7382 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
7383 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
7384 crypto_aese, crypto_aesmc. Move to types.md.
7385 * config/arm/types.md (crypto_aes): Split into crypto_aese,
7387 * config/arm/iterators.md (crypto_type): Likewise.
7389 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7391 * cgraph.c: Include expr.h and tree-dfa.h.
7392 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
7395 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
7398 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
7399 regs from checking multi-reg pseudos.
7401 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7403 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
7405 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7407 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
7408 if it would clobber the stack pointer, even temporarily.
7410 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
7412 * mode-switching.c: Make small adjustments to the top comment.
7414 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
7416 * config/rs6000/constraints.md (wD constraint): New constraint to
7417 match the constant integer to get the top DImode/DFmode out of a
7418 vector in a VSX register.
7420 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
7421 match the constant integer to get the top DImode/DFmode out of a
7422 vector in a VSX register.
7424 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
7427 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
7430 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
7431 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
7433 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
7434 Optimize vec_extract of 64-bit values, where the value being
7435 extracted is in the top word, where we can use scalar
7436 instructions. Add direct move and store support. Combine the big
7437 endian/little endian vector select load support into a single insn.
7438 (vsx_extract_<mode>_internal1): Likewise.
7439 (vsx_extract_<mode>_internal2): Likewise.
7440 (vsx_extract_<mode>_load): Likewise.
7441 (vsx_extract_<mode>_store): Likewise.
7442 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
7443 combined into vsx_extract_<mode>_load.
7444 (vsx_extract_<mode>_one_le): Likewise.
7446 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
7447 define the top 64-bit vector element.
7449 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
7452 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7453 Document vec_vbpermq builtin.
7456 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
7457 enable use of xxsldwi and xxpermdi builtin functions.
7458 (vec_xxpermdi): Likewise.
7460 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7461 Document use of vec_xxsldwi and vec_xxpermdi builtins.
7463 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
7465 PR rtl-optimization/60650
7466 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
7468 (find_spills_for): New.
7469 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
7470 Spill all pseudos on the second iteration.
7472 2014-03-27 Marek Polacek <polacek@redhat.com>
7475 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
7478 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7480 * config/s390/s390.c (s390_can_use_return_insn): Check for
7481 call-saved FPRs on 31 bit.
7483 2014-03-27 Jakub Jelinek <jakub@redhat.com>
7486 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
7487 if they need regimplification, just drop them instead of
7488 calling gimple_regimplify_operands on them.
7490 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
7493 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
7494 (aarch64_frame_pointer_required): Adjust logic.
7495 (aarch64_can_eliminate): Adjust logic.
7496 (aarch64_override_options_after_change): Adjust logic.
7498 2014-03-27 Dehao Chen <dehao@google.com>
7500 * ipa-inline.c (early_inliner): Update node's inline info.
7502 2014-03-26 Dehao Chen <dehao@google.com>
7504 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
7505 compiler inserted conditional jumps for NAN float check.
7507 2014-03-26 Jakub Jelinek <jakub@redhat.com>
7509 * ubsan.h (ubsan_create_data): Change second argument's type
7510 to const location_t *.
7511 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
7513 (ubsan_create_data): Change second argument to const location_t *PLOC.
7514 Create Loc field whenever PLOC is non-NULL.
7515 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
7516 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
7520 * real.c (real_to_integer2): Change type of low to UHWI.
7522 2014-03-26 Tobias Burnus <burnus@net-b.de>
7524 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
7525 (CILK_SELF_SPECS): New define.
7526 (driver_self_specs): Use it.
7528 2014-03-26 Richard Biener <rguenther@suse.de>
7530 * tree-pretty-print.c (percent_K_format): Implement special
7531 case for LTO and its stripped down BLOCK tree.
7533 2014-03-26 Jakub Jelinek <jakub@redhat.com>
7536 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
7538 * tree-vrp.c (simplify_internal_call_using_ranges): If only
7539 one range is range_int_cst_p, but not both, at least optimize
7540 addition/subtraction of 0 and multiplication by 0 or 1.
7541 * gimple-fold.c (gimple_fold_call): Fold
7542 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
7543 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
7544 INTEGER_CSTs, try to fold at least x * 0 and y - y.
7546 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
7548 PR rtl-optimization/60452
7549 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
7550 <case REG>: Return 1 for invalid offsets from the frame pointer.
7552 2014-03-26 Marek Polacek <polacek@redhat.com>
7555 * doc/extend.texi (C Extensions): Mention variable-length arrays in
7558 2014-03-26 Marek Polacek <polacek@redhat.com>
7561 * doc/extend.texi (Designated Inits): Describe what happens to omitted
7564 2014-03-26 Marek Polacek <polacek@redhat.com>
7567 * ira-color.c (update_conflict_hard_regno_costs): Perform the
7568 multiplication in unsigned type.
7570 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
7572 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
7574 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
7576 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
7578 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
7581 * cif-code.def (UNREACHABLE) New code.
7582 * ipa-inline.c (inline_small_functions): Skip edges to
7583 __builtlin_unreachable.
7584 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
7585 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
7586 predicate to __bulitin_unreachable.
7587 (set_cond_stmt_execution_predicate): Fix issue when
7588 invert_tree_comparison returns ERROR_MARK.
7589 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
7590 propagate to inline clones.
7591 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
7593 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
7594 * cgraphclones.c (cgraph_clone_node): If call destination is already
7595 ureachable, do not redirect it back.
7596 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
7599 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
7601 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
7602 Do not modify inline clones.
7604 2014-03-25 Jakub Jelinek <jakub@redhat.com>
7606 * config/i386/i386.md (general_sext_operand): New mode attr.
7607 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
7608 don't generate (sign_extend (const_int)).
7609 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
7610 operands[2]. Use We constraint instead of <i> and
7611 <general_sext_operand> predicate instead of <general_operand>.
7612 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
7613 * config/i386/constraints.md (We): New constraint.
7614 * config/i386/predicates.md (x86_64_sext_operand,
7615 sext_operand): New predicates.
7617 2014-03-25 Martin Jambor <mjambor@suse.cz>
7620 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
7621 inconsistent devirtualizations to __builtin_unreachable.
7623 2014-03-25 Marek Polacek <polacek@redhat.com>
7626 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
7628 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
7630 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
7631 order of elements for big-endian.
7633 2014-03-25 Richard Biener <rguenther@suse.de>
7636 * gimplify-me.c (gimple_regimplify_operands): Update the
7639 2014-03-25 Martin Jambor <mjambor@suse.cz>
7642 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
7643 (lto_output_varpool_node): Likewise.
7644 (input_overwrite_node): Likewise.
7645 (input_varpool_node): Likewise.
7647 2014-03-25 Richard Biener <rguenther@suse.de>
7649 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
7650 (run_gcc): Likewise.
7652 2014-03-25 Jakub Jelinek <jakub@redhat.com>
7654 * combine.c (simplify_compare_const): Add MODE argument.
7655 Handle mode_width 0 as very large mode_width.
7656 (try_combine, simplify_comparison): Adjust callers.
7658 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
7659 type to avoid signed integer overflow.
7660 * explow.c (plus_constant): Likewise.
7662 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
7664 * doc/generic.texi: Correct typos.
7666 2014-03-24 Tobias Burnus <burnus@net-b.de>
7668 * doc/invoke.texi (-flto): Expand section about
7669 using static libraries with LTO.
7671 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7673 PR rtl-optimization/60501
7674 * optabs.def (addptr3_optab): New optab.
7675 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
7676 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
7677 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
7679 * lra.c (emit_add3_insn): Use the addptr pattern if available.
7681 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
7683 2014-03-24 Ulrich Drepper <drepper@gmail.com>
7685 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
7688 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
7689 (_mm256_undefined_ps): Define.
7690 (_mm256_undefined_pd): Define.
7691 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
7692 (_mm_undefined_pd): Define.
7693 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
7694 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
7695 (_mm512_undefined_ps): Define.
7696 (_mm512_undefined_pd): Define.
7697 Use _mm*_undefined_*.
7698 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
7700 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
7702 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
7703 (lshr_simd): DI mode added.
7704 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
7705 (aarch64_ushr_simddi): Likewise.
7706 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
7707 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
7708 (vshrd_n_u64): Likewise.
7710 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7712 * Makefile.in (s-macro_list): Depend on cc1.
7714 2014-03-23 Teresa Johnson <tejohnson@google.com>
7716 * ipa-utils.c (ipa_print_order): Use specified dump file.
7718 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
7720 PR rtl-optimization/60601
7721 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
7723 * gcc.c (eval_spec_function): Initialize save_growing_value.
7725 2014-03-22 Jakub Jelinek <jakub@redhat.com>
7728 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
7729 code == MINUS_EXPR, never swap op0 with op1.
7731 * toplev.c (init_local_tick): Avoid signed integer multiplication
7733 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
7734 shift by first operand's bitsize.
7736 2014-03-21 Jakub Jelinek <jakub@redhat.com>
7739 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
7741 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
7742 TARGET_ISA_64BIT_P(x).
7744 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7746 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
7747 pattern for vector nor instead of subtract from splat(-1).
7748 (altivec_expand_vec_perm_const_le): Likewise.
7750 2014-03-21 Richard Henderson <rth@twiddle.net>
7753 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
7754 related insns after epilogue_completed.
7756 2014-03-21 Martin Jambor <mjambor@suse.cz>
7759 * cgraph.h (symtab_node): New flag body_removed.
7760 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
7761 when removing bodies.
7762 * symtab.c (dump_symtab_base): Dump body_removed flag.
7763 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
7764 had their bodies removed.
7766 2014-03-21 Martin Jambor <mjambor@suse.cz>
7769 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
7772 2014-03-21 Richard Biener <rguenther@suse.de>
7774 PR tree-optimization/60577
7775 * tree-core.h (struct tree_base): Document nothrow_flag use
7776 in VAR_DECL_NONALIASED.
7777 * tree.h (VAR_DECL_NONALIASED): New.
7778 (may_be_aliased): Adjust.
7779 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
7781 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7783 * expr.c (expand_expr_real_1): Remove outdated comment.
7785 2014-03-20 Jakub Jelinek <jakub@redhat.com>
7788 * ira.c (adjust_cleared_regs): Call copy_rtx on
7789 *reg_equiv[REGNO (loc)].src_p before passing it to
7790 simplify_replace_fn_rtx.
7793 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
7794 into CONST, put pic register as first operand of PLUS. Use
7795 gen_const_mem for both 32-bit and 64-bit PIC got loads.
7797 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7799 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
7801 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7803 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
7804 around for store forwarding issue in the FPU on the UT699.
7805 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
7806 loads and operations if -mfix-ut699 is specified.
7807 (divtf3_hq): Tweak attribute.
7808 (sqrttf2_hq): Likewise.
7810 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7812 * calls.c (store_one_arg): Remove incorrect const qualification on the
7813 type of the temporary.
7814 * cfgexpand.c (expand_return): Likewise.
7815 * expr.c (expand_constructor): Likewise.
7816 (expand_expr_real_1): Likewise.
7818 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7820 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
7823 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
7826 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
7828 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
7830 * config/arm/aarch-common-protos.h
7831 (alu_cost_table): Fix spelling of "extend".
7832 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
7834 2014-03-19 Richard Biener <rguenther@suse.de>
7837 * tree-core.h (tree_type_common): Re-order pointer members
7838 to reduce recursion depth during GC walks.
7840 2014-03-19 Marek Polacek <polacek@redhat.com>
7843 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
7844 before accessing it.
7846 2014-03-19 Richard Biener <rguenther@suse.de>
7849 * lto-streamer-in.c (input_function): In WPA stage do not drop
7852 2014-03-19 Jakub Jelinek <jakub@redhat.com>
7854 PR tree-optimization/60559
7855 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
7856 with build_zero_cst assignment.
7858 2014-03-18 Kai Tietz <ktietz@redhat.com>
7860 PR rtl-optimization/56356
7861 * sdbout.c (sdbout_parms): Verify that parms'
7862 incoming argument is valid.
7863 (sdbout_reg_parms): Likewise.
7865 2014-03-18 Richard Henderson <rth@redhat.com>
7868 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
7869 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
7870 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
7872 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
7874 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
7875 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
7876 Italicize plugin event names in description. Explain that
7877 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
7878 Remind that no GCC functions should be called after PLUGIN_FINISH.
7879 Explain what pragmas with expansion are.
7881 2014-03-18 Martin Liska <mliska@suse.cz>
7883 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
7884 gimple call statement is update.
7885 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
7886 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
7888 2014-03-18 Jakub Jelinek <jakub@redhat.com>
7891 * ubsan.c (ubsan_instrument_unreachable): Call
7892 initialize_sanitizer_builtins.
7893 (ubsan_pass): Likewise.
7896 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
7897 varpool_finalize_decl instead of rest_of_decl_compilation.
7899 2014-03-18 Richard Biener <rguenther@suse.de>
7901 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
7902 by using bitmap_and_compl instead of bitmap_and_compl_into.
7903 (df_rd_transfer_function): Likewise.
7905 2014-03-18 Richard Biener <rguenther@suse.de>
7907 * doc/lto.texi (fresolution): Fix typo.
7909 2014-03-18 Richard Biener <rguenther@suse.de>
7911 * doc/invoke.texi (flto): Update for changes in 4.9.
7913 2014-03-18 Richard Biener <rguenther@suse.de>
7915 * doc/loop.texi: Remove section on the removed lambda framework.
7916 Update loop docs with recent changes in preserving loop structure.
7918 2014-03-18 Richard Biener <rguenther@suse.de>
7920 * doc/lto.texi (-fresolution): Document.
7922 2014-03-18 Richard Biener <rguenther@suse.de>
7924 * doc/contrib.texi: Adjust my name.
7926 2014-03-18 Jakub Jelinek <jakub@redhat.com>
7929 * internal-fn.c: Include diagnostic-core.h.
7930 (expand_BUILTIN_EXPECT): New function.
7931 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
7932 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
7933 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
7934 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
7936 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
7937 Revert 3 argument __builtin_expect code.
7938 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
7939 * gimple-fold.c (gimple_fold_call): Likewise.
7940 * tree.h (fold_builtin_expect): New prototype.
7941 * builtins.c (build_builtin_expect_predicate): Add predictor
7942 argument, if non-NULL, create 3 argument __builtin_expect.
7943 (fold_builtin_expect): No longer static. Add ARG2 argument,
7944 pass it through to build_builtin_expect_predicate.
7945 (fold_builtin_2): Adjust caller.
7946 (fold_builtin_3): Handle BUILT_IN_EXPECT.
7947 * internal-fn.def (BUILTIN_EXPECT): New.
7949 2014-03-18 Tobias Burnus <burnus@net-b.de>
7952 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
7953 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
7954 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
7956 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
7959 * predict.c (combine_predictions_for_bb): Fix up formatting.
7960 (expr_expected_value_1, expr_expected_value): Add predictor argument,
7961 fill what it points to if non-NULL.
7962 (tree_predict_by_opcode): Adjust caller, use the predictor.
7963 * predict.def (PRED_COMPARE_AND_SWAP): Add.
7965 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
7967 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
7968 proper constant for the store mode.
7970 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
7972 * symtab.c (change_decl_assembler_name): Fix transparent alias
7975 2014-03-16 Renlin Li <Renlin.Li@arm.com>
7977 * config/aarch64/aarch64.c: Correct the comments about the
7978 aarch64 stack layout.
7980 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
7982 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
7983 check for GF_OMP_FOR_KIND_FOR.
7985 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
7987 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
7988 ymm and zmm register names.
7990 2014-03-17 Jakub Jelinek <jakub@redhat.com>
7993 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
7994 note creation for the 2010-08-31 changes.
7996 2014-03-17 Marek Polacek <polacek@redhat.com>
7999 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
8000 as -fno-tree-loop-vectorize.
8001 (expand_omp_simd): Likewise.
8003 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
8005 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
8006 (eligible_for_call_delay): New prototype.
8007 * config/sparc/sparc.c (tls_call_delay): Rename into...
8008 (eligible_for_call_delay): ...this. Return false if the instruction
8009 cannot be put in the delay slot of a branch.
8010 (eligible_for_restore_insn): Simplify.
8011 (eligible_for_return_delay): Return false if the instruction cannot be
8012 put in the delay slot of a branch and simplify.
8013 (eligible_for_sibcall_delay): Return false if the instruction cannot be
8014 put in the delay slot of a branch.
8015 * config/sparc/sparc.md (fix_ut699): New attribute.
8016 (tls_call_delay): Delete.
8017 (in_call_delay): Reimplement.
8018 (eligible_for_sibcall_delay): Rename into...
8019 (in_sibcall_delay): ...this.
8020 (eligible_for_return_delay): Rename into...
8021 (in_return_delay): ...this.
8022 (in_branch_delay): Reimplement.
8023 (in_uncond_branch_delay): Delete.
8024 (in_annul_branch_delay): Delete.
8026 2014-03-14 Richard Henderson <rth@redhat.com>
8029 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
8030 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
8031 (*floathi<X87MODEF>2_i387_with_temp): Remove.
8032 (floathi splitters): Remove.
8033 (float<SWI48x>xf2): New pattern.
8034 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
8035 code that tried to handle DImode for 32-bit, but which was excluded
8036 by the pattern's condition. Drop allocation of stack temporary.
8037 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
8038 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
8039 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
8040 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
8041 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
8042 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
8043 (*float<SWI48><MODEF>2_sse_interunit): Remove.
8044 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
8045 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
8046 (*float<SWI48x><X87MODEF>2_i387): Remove.
8047 (all float _with_temp splitters): Remove.
8048 (*float<SWI48x><MODEF>2_i387): New pattern.
8049 (*float<SWI48><MODEF>2_sse): New pattern.
8050 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
8051 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
8053 2014-03-14 Jakub Jelinek <jakub@redhat.com>
8054 Marek Polacek <polacek@redhat.com>
8057 * common.opt (dump_base_name_prefixed): New Variable.
8058 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
8059 if x_dump_base_name_prefixed is already set, set it at the end.
8061 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
8063 PR rtl-optimization/60508
8064 * lra-constraints.c (get_reload_reg): Add new parameter
8066 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
8067 Pass the new parameter values.
8069 2014-03-14 Richard Biener <rguenther@suse.de>
8071 * common.opt: Revert unintented changes from r205065.
8074 2014-03-14 Richard Biener <rguenther@suse.de>
8077 * cfghooks.c (split_block): Properly adjust all loops the
8078 block was a latch of.
8080 2014-03-14 Martin Jambor <mjambor@suse.cz>
8083 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
8086 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
8089 * config/avr/avr.c (avr_set_current_function): Pass function name
8090 through default_strip_name_encoding before sanity checking instead
8091 of skipping the first char of the assembler name.
8093 2014-03-13 Richard Henderson <rth@redhat.com>
8096 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
8097 (ix86_force_to_memory, ix86_free_from_memory): Remove.
8098 * config/i386/i386-protos.h: Likewise.
8099 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
8100 in the expander instead of a splitter.
8101 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
8102 any possibility of requiring a memory.
8103 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
8104 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
8105 (fp branch splitters): Update for ix86_split_fp_branch.
8106 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
8107 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
8108 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
8109 (*fop_<MODEF>_2_i387): Remove f/r alternative.
8110 (*fop_<MODEF>_3_i387): Likewise.
8111 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
8112 (splitters for the fop_* register patterns): Remove.
8113 (fscalexf4_i387): Rename from *fscalexf4_i387.
8114 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
8116 2014-03-13 Jakub Jelinek <jakub@redhat.com>
8118 PR tree-optimization/59779
8119 * tree-dfa.c (get_ref_base_and_extent): Use double_int
8120 type for bitsize and maxsize instead of HOST_WIDE_INT.
8122 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
8124 PR rtl-optimization/57320
8125 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
8126 the CFG after thread_prologue_and_epilogue_insns.
8128 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
8130 PR rtl-optimization/57189
8131 * lra-constraints.c (process_alt_operands): Disfavor spilling
8134 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
8136 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
8138 2014-03-13 Jakub Jelinek <jakub@redhat.com>
8140 PR tree-optimization/59025
8142 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
8143 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
8145 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
8148 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
8149 calls of avr_out_plus_1.
8151 2014-03-13 Bin Cheng <bin.cheng@arm.com>
8153 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
8154 BB's single pred and update the father loop's latch info later.
8156 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
8158 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
8162 (VEC_base): Likewise.
8163 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
8164 registers, we need to swap double words in little endian mode.
8166 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
8167 to be a container mode for 128-bit integer operations added in ISA
8168 2.07. Unlike TImode and PTImode, the preferred register set is
8169 the Altivec/VMX registers for the 128-bit operations.
8171 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
8173 (rs6000_split_128bit_ok_p): Likewise.
8175 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
8176 macros for creating ISA 2.07 normal and overloaded builtin
8177 functions with 3 arguments.
8178 (BU_P8V_OVERLOAD_3): Likewise.
8179 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
8180 for use as overloaded functions.
8181 (VPERM_1TI_UNS): Likewise.
8182 (VSEL_1TI): Likewise.
8183 (VSEL_1TI_UNS): Likewise.
8184 (ST_INTERNAL_1ti): Likewise.
8185 (LD_INTERNAL_1ti): Likewise.
8186 (XXSEL_1TI): Likewise.
8187 (XXSEL_1TI_UNS): Likewise.
8188 (VPERM_1TI): Likewise.
8189 (VPERM_1TI_UNS): Likewise.
8190 (XXPERMDI_1TI): Likewise.
8191 (SET_1TI): Likewise.
8192 (LXVD2X_V1TI): Likewise.
8193 (STXVD2X_V1TI): Likewise.
8194 (VEC_INIT_V1TI): Likewise.
8195 (VEC_SET_V1TI): Likewise.
8196 (VEC_EXT_V1TI): Likewise.
8197 (EQV_V1TI): Likewise.
8198 (NAND_V1TI): Likewise.
8199 (ORC_V1TI): Likewise.
8200 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
8201 added in ISA 2.07. Add both normal 'altivec' builtins, and the
8203 (VADDUQM): Likewise.
8204 (VSUBCUQ): Likewise.
8205 (VADDEUQM): Likewise.
8206 (VADDECUQ): Likewise.
8207 (VSUBEUQM): Likewise.
8208 (VSUBECUQ): Likewise.
8210 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
8211 __int128_t and __uint128_t types.
8212 (__uint128_type): Likewise.
8213 (altivec_categorize_keyword): Add support for vector __int128_t,
8214 vector __uint128_t, vector __int128, and vector unsigned __int128
8215 as a container type for TImode operations that need to be done in
8216 VSX/Altivec registers.
8217 (rs6000_macro_to_expand): Likewise.
8218 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
8219 to support 128-bit integer instructions vaddcuq, vadduqm,
8220 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
8221 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
8223 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
8224 for V1TImode, and set up preferences to use VSX/Altivec registers.
8225 Setup VSX reload handlers.
8226 (rs6000_debug_reg_global): Likewise.
8227 (rs6000_init_hard_regno_mode_ok): Likewise.
8228 (rs6000_preferred_simd_mode): Likewise.
8229 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
8230 (easy_altivec_constant): Likewise.
8231 (output_vec_const_move): Likewise.
8232 (rs6000_expand_vector_set): Convert V1TImode set and extract to
8234 (rs6000_expand_vector_extract): Likewise.
8235 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
8237 (rs6000_const_vec): Add support for V1TImode.
8238 (rs6000_emit_le_vsx_load): Swap double words when loading or
8239 storing TImode/V1TImode.
8240 (rs6000_emit_le_vsx_store): Likewise.
8241 (rs6000_emit_le_vsx_move): Likewise.
8242 (rs6000_emit_move): Add support for V1TImode.
8243 (altivec_expand_ld_builtin): Likewise.
8244 (altivec_expand_st_builtin): Likewise.
8245 (altivec_expand_vec_init_builtin): Likewise.
8246 (altivec_expand_builtin): Likewise.
8247 (rs6000_init_builtins): Add support for V1TImode type. Add
8248 support for ISA 2.07 128-bit integer builtins. Define type names
8249 for the VSX/Altivec vector types.
8250 (altivec_init_builtins): Add support for overloaded vector
8251 functions with V1TImode type.
8252 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
8253 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
8255 (rs6000_split_128bit_ok_p): Likewise.
8256 (rs6000_handle_altivec_attribute): Create V1TImode from vector
8257 __int128_t and vector __uint128_t.
8259 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
8260 and mode attributes.
8267 (VS_scalar): Likewise.
8268 (VS_double): Likewise.
8269 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
8271 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
8272 we support the ISA 2.07 128-bit integer arithmetic instructions.
8273 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
8274 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
8275 and TImode types for use with the builtin functions.
8276 (V1TI_type_node): Likewise.
8277 (unsigned_V1TI_type_node): Likewise.
8278 (intTI_type_internal_node): Likewise.
8279 (uintTI_type_internal_node): Likewise.
8281 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
8282 128-bit builtin functions.
8283 (UNSPEC_VADDEUQM): Likewise.
8284 (UNSPEC_VADDECUQ): Likewise.
8285 (UNSPEC_VSUBCUQ): Likewise.
8286 (UNSPEC_VSUBEUQM): Likewise.
8287 (UNSPEC_VSUBECUQ): Likewise.
8288 (VM): Add V1TImode to vector mode iterators.
8290 (VI_unit): Likewise.
8291 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
8292 (altivec_vaddcuq): Likewise.
8293 (altivec_vsubuqm): Likewise.
8294 (altivec_vsubcuq): Likewise.
8295 (altivec_vaddeuqm): Likewise.
8296 (altivec_vaddecuq): Likewise.
8297 (altivec_vsubeuqm): Likewise.
8298 (altivec_vsubecuq): Likewise.
8300 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
8302 (BOOL_128): Likewise.
8303 (BOOL_REGS_OUTPUT): Likewise.
8304 (BOOL_REGS_OP1): Likewise.
8305 (BOOL_REGS_OP2): Likewise.
8306 (BOOL_REGS_UNARY): Likewise.
8307 (BOOL_REGS_AND_CR0): Likewise.
8309 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
8310 128-bit integer builtin support.
8311 (vec_vadduqm): Likewise.
8312 (vec_vaddecuq): Likewise.
8313 (vec_vaddeuqm): Likewise.
8314 (vec_vsubecuq): Likewise.
8315 (vec_vsubeuqm): Likewise.
8316 (vec_vsubcuq): Likewise.
8317 (vec_vsubuqm): Likewise.
8319 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
8320 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
8321 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
8322 128-bit integer add/subtract to ISA 2.07.
8324 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
8326 * config/arc/arc.c (arc_predicate_delay_insns):
8327 Fix third argument passed to conditionalize_nonjump.
8329 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
8331 * config/aarch64/aarch64-builtins.c
8332 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
8333 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
8334 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
8335 instead of __builtin_lfloor.
8336 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
8338 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8340 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
8341 (tree_ssa_ifcombine_bb_1): New function.
8342 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
8343 is an empty forwarder block to then_bb or vice versa and then_bb
8344 and else_bb are effectively swapped.
8346 2014-03-12 Christian Bruel <christian.bruel@st.com>
8349 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
8350 REG_CFA_DEF_CFA note.
8351 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
8352 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
8354 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
8356 PR tree-optimization/60454
8357 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
8359 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8361 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
8362 Do not define target_cpu_default2 to generic.
8363 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
8364 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
8365 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
8367 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8368 Marc Glisse <marc.glisse@inria.fr>
8370 PR tree-optimization/60502
8371 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
8372 instead of build_low_bits_mask.
8374 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8377 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
8378 if there are multiple uses, but op doesn't live on E edge.
8379 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
8380 clobber stmts before __builtin_unreachable.
8382 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
8384 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
8385 hard_frame_pointer_rtx.
8386 * cse.c (cse_insn): Remove volatile check.
8387 * cselib.c (cselib_process_insn): Likewise.
8388 * dse.c (scan_insn): Likewise.
8390 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
8392 * config/arc/arc.c (conditionalize_nonjump): New function,
8394 (arc_ifcvt): ... this.
8395 (arc_predicate_delay_insns): Use it.
8397 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
8399 * config/arc/predicates.md (extend_operand): During/after reload,
8400 allow const_int_operand.
8401 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
8402 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
8403 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
8405 (umulsi3_highpart_i): Likewise.
8407 2014-03-11 Richard Biener <rguenther@suse.de>
8409 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
8410 Add asserts to guard possible wrong-code bugs.
8412 2014-03-11 Richard Biener <rguenther@suse.de>
8414 PR tree-optimization/60429
8415 PR tree-optimization/60485
8416 * tree-ssa-structalias.c (set_union_with_increment): Properly
8417 take into account all fields that overlap the shifted vars.
8418 (do_sd_constraint): Likewise.
8419 (do_ds_constraint): Likewise.
8420 (get_constraint_for_ptr_offset): Likewise.
8422 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
8424 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
8425 (nios2_compute_frame_layout):
8426 Add calculation of cfun->machine->fp_save_offset.
8427 (nios2_expand_prologue): Correct setting of frame pointer register
8429 (nios2_expand_epilogue): Update recovery of stack pointer from
8430 frame pointer accordingly.
8431 (nios2_initial_elimination_offset): Update calculation of offset
8432 for eliminating to HARD_FRAME_POINTER_REGNUM.
8434 2014-03-10 Jakub Jelinek <jakub@redhat.com>
8437 * ipa.c (symtab_remove_unreachable_nodes): Don't call
8438 cgraph_get_create_node on VAR_DECLs.
8440 2014-03-10 Richard Biener <rguenther@suse.de>
8443 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
8445 2014-03-08 Douglas B Rupp <rupp@gnat.com>
8447 * config/vms/vms.opt (vms_float_format): New variable.
8449 2014-03-08 Tobias Burnus <burnus@net-b.de>
8451 * doc/invoke.texi (-fcilkplus): Update implementation status.
8453 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
8454 Richard Biener <rguenther@suse.de>
8456 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
8457 consistently accross all TUs.
8458 (run_gcc): Enable -fshort-double automatically at link at link-time
8459 and disallow override.
8461 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
8464 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
8465 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
8466 if they can't be used.
8468 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8470 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
8471 for Solaris 11/x86 ld.
8472 * configure: Regenerate.
8474 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8476 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
8477 (LIB_TLS_SPEC): Save as ld_tls_libs.
8478 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
8479 (HAVE_AS_IX86_TLSLDM): New test.
8480 * configure, config.in: Regenerate.
8481 * config/i386/i386.c (legitimize_tls_address): Fall back to
8482 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
8483 cannot support TLS_MODEL_LOCAL_DYNAMIC.
8484 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
8485 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
8487 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
8489 * common.opt (fira-loop-pressure): Mark as optimization.
8491 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
8493 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
8494 an OpenMP mappable type.
8496 2014-03-06 Matthias Klose <doko@ubuntu.com>
8498 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
8499 MULTILIB_OSDIRNAMES is not defined.
8501 2014-03-06 Jakub Jelinek <jakub@redhat.com>
8502 Meador Inge <meadori@codesourcery.com>
8505 * config/arm/arm.c (arm_tls_symbol_p): Remove.
8506 (arm_legitimize_address): Call legitimize_tls_address for any
8507 arm_tls_referenced_p expression, handle constant addend. Call it
8508 before testing for !TARGET_ARM.
8509 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
8511 2014-03-06 Richard Biener <rguenther@suse.de>
8517 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
8519 * tree-streamer.c (record_common_node): Assert we don't record
8520 nodes with type double.
8521 (preload_common_node): Skip type double, complex double and double
8522 pointer since it is now frontend dependent due to fshort-double option.
8524 2014-03-06 Richard Biener <rguenther@suse.de>
8526 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
8527 or -fno-lto is specified and the linker has full plugin support.
8528 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
8529 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
8530 * lto-wrapper.c (merge_and_complain): Merge compile-time
8531 optimization levels.
8532 (run_gcc): And pass it through to the link options.
8534 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
8538 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8540 * cselib.c (remove_useless_values): Skip to avoid quadratic
8541 behavior if the condition moved from...
8542 (cselib_process_insn): ... here holds.
8544 2014-03-05 Jakub Jelinek <jakub@redhat.com>
8547 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
8548 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
8551 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
8552 (TM_H): Add x86-tune.def.
8554 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8556 * config/aarch64/aarch64.c (generic_tunings):
8557 Use cortexa57_extra_costs.
8559 2014-03-05 Jakub Jelinek <jakub@redhat.com>
8562 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
8563 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
8564 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
8567 2014-03-04 Heiher <r@hev.cc>
8569 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
8570 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
8572 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
8574 * config/i386/predicates.md (const2356_operand): Change to ...
8575 (const2367_operand): ... this.
8576 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
8578 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8579 (*avx512pf_scatterpf<mode>sf): Ditto.
8580 (avx512pf_scatterpf<mode>df): Ditto.
8581 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8582 (*avx512pf_scatterpf<mode>df): Ditto.
8583 * config/i386/i386.c (ix86_expand_builtin): Update
8584 incorrect hint operand error message.
8586 2014-03-04 Richard Biener <rguenther@suse.de>
8588 * lto-section-in.c (lto_get_section_data): Fix const cast.
8590 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
8592 * tree-streamer.c (record_common_node): Assert we don't record
8593 nodes with type double.
8594 (preload_common_node): Skip type double, complex double and double
8595 pointer since it is now frontend dependent due to fshort-double option.
8597 2014-03-04 Richard Biener <rguenther@suse.de>
8600 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
8601 (lto_input_toplevel_asms): Likewise.
8602 * lto-section-in.c (lto_get_section_data): Instead do it here
8605 2014-03-04 Richard Biener <rguenther@suse.de>
8607 PR tree-optimization/60382
8608 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
8609 dead PHIs a reduction.
8611 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
8613 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
8615 (_mm_prefetch): Move out of GCC target("sse") pragma.
8616 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
8617 GCC target("prfchw") pragma.
8618 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
8620 * config/i386/i386.c (ix86_option_override_internal): Enable
8621 -mprfchw with -mprefetchwt1.
8623 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
8625 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
8628 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
8630 * opts.h (CL_PCH_IGNORE): Define.
8631 * targhooks.c (option_affects_pch_p):
8632 Return false for options that have CL_PCH_IGNORE set.
8633 * opt-functions.awk: Process PchIgnore.
8634 * doc/options.texi: Document PchIgnore.
8636 * config/arc/arc.opt (misize): Add PchIgnore property.
8638 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8640 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
8641 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
8642 constraint on constants to permit them being loaded into
8643 GENERAL_REGS or BASE_REGS.
8645 2014-03-03 Nick Clifton <nickc@redhat.com>
8647 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
8648 anti-cacnonical alternatives.
8649 (negandhi3_real): New pattern.
8650 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
8652 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
8654 * config/avr/avr-mcus.def: Remove atxmega16x1.
8655 * config/avr/avr-tables.opt: Regenerate.
8656 * config/avr/t-multilib: Regenerate.
8657 * doc/avr-mmcu.texi: Regenerate.
8659 2014-03-03 Tobias Grosser <tobias@grosser.es>
8660 Mircea Namolaru <mircea.namolaru@inria.fr>
8662 PR tree-optimization/58028
8663 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
8666 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8668 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
8669 not handled by recognizers.
8671 2014-03-03 Jakub Jelinek <jakub@redhat.com>
8674 * function.c (expand_function_end): Don't emit
8675 clobber_return_register sequence if clobber_after is a BARRIER.
8676 * cfgexpand.c (construct_exit_block): Append instructions before
8677 return_label to prev_bb.
8679 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8681 * config/rs6000/constraints.md: Document reserved use of "wc".
8683 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
8686 * ipa.c (function_and_variable_visibility): When dissolving comdat
8687 group, also set all symbols to local.
8689 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
8694 2013-12-14 Jan Hubicka <jh@suse.cz>
8696 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
8698 2014-03-02 Jon Beniston <jon@beniston.com>
8704 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
8705 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
8706 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
8707 (simple_return, *simple_return): New patterns
8708 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
8709 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
8711 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
8713 * dwarf2out.c (gen_subprogram_die): Tidy.
8715 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
8718 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
8719 (*mov_t_msb_neg_negc): ... this new insn.
8721 2014-02-28 Jason Merrill <jason@redhat.com>
8724 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
8727 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
8730 * dwarf2out.c (decltype_auto_die): New static.
8731 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
8732 (gen_type_die_with_usage): Handle 'decltype(auto)'.
8733 (is_cxx_auto): Likewise.
8735 2014-02-28 Ian Bolton <ian.bolton@arm.com>
8737 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
8738 we are not using general regs only.
8740 2014-02-28 Richard Biener <rguenther@suse.de>
8743 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
8744 previous fix and only allow to remove trivial pre-headers
8745 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
8746 (remove_forwarder_block): Properly update the latch of a loop.
8748 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8751 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
8752 (cselib_preserved_hash_table): New.
8753 (preserve_constants_and_equivs): Move preserved vals to it.
8754 (cselib_find_slot): Look it up first.
8755 (cselib_init): Initialize it.
8756 (cselib_finish): Release it.
8757 (dump_cselib_table): Dump it.
8759 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8762 * cselib.c (remove_useless_values): Skip to avoid quadratic
8763 behavior if the condition moved from...
8764 (cselib_process_insn): ... here holds.
8766 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8769 * var-tracking.c (vt_initialize): Apply the same condition to
8770 preserve the CFA base value.
8772 2014-02-28 Joey Ye <joey.ye@arm.com>
8775 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
8776 if reload in progress or completed.
8778 2014-02-28 Tobias Burnus <burnus@net-b.de>
8781 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
8784 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
8786 * doc/tm.texi.in (Condition Code Status): Update documention for
8787 relative locations of cc0-setter and cc0-user.
8789 2014-02-27 Jeff Law <law@redhat.com>
8791 PR rtl-optimization/52714
8792 * combine.c (try_combine): When splitting an unrecognized PARALLEL
8793 into two independent simple sets, if I3 is a jump, ensure the
8794 pattern we place into I3 is a (set (pc) ...).
8796 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
8797 Jeff Law <law@redhat.com>
8799 PR rtl-optimization/49847
8800 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
8801 are in different blocks.
8802 * doc/tm.texi (Condition Code Status): Update documention for
8803 relative locations of cc0-setter and cc0-user.
8805 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
8808 * lra.c (lra_emit_add): Check SUBREG too.
8810 2014-02-27 Andreas Schwab <schwab@suse.de>
8812 * config/m68k/m68k.c (m68k_option_override): Disable
8813 -flive-range-shrinkage for classic m68k.
8814 (m68k_override_options_after_change): Likewise.
8816 2014-02-27 Marek Polacek <polacek@redhat.com>
8819 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
8820 -Wmaybe-uninitialized.
8822 2014-02-27 Alan Modra <amodra@gmail.com>
8825 * reload1.c (emit_input_reload_insns): When reload_override_in,
8826 set old to rl->in_reg when rl->in_reg is a subreg.
8828 2014-02-26 Richard Biener <rguenther@suse.de>
8831 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
8833 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
8835 * common/config/i386/predicates.md (const1256_operand): Remove.
8836 (const2356_operand): New.
8837 (const_1_to_2_operand): Remove.
8838 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
8839 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8840 (*avx512pf_gatherpf<mode>sf): Ditto.
8841 (avx512pf_gatherpf<mode>df): Ditto.
8842 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8843 (*avx512pf_gatherpf<mode>df): Ditto.
8844 (avx512pf_scatterpf<mode>sf): Ditto.
8845 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8846 (*avx512pf_scatterpf<mode>sf): Ditto.
8847 (avx512pf_scatterpf<mode>df): Ditto.
8848 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8849 (*avx512pf_scatterpf<mode>df): Ditto.
8850 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
8852 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
8854 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
8855 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
8856 (_mm512_mask_testn_epi64_mask): Move to ...
8857 * config/i386/avx512cdintrin.h: Here.
8858 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
8859 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
8860 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
8861 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
8862 TARGET_AVX512F from TARGET_AVX512CD.
8864 2014-02-26 Richard Biener <rguenther@suse.de>
8867 * ipa.c (walk_polymorphic_call_targets): Properly guard
8868 call to inline_update_overall_summary.
8870 2014-02-26 Bin Cheng <bin.cheng@arm.com>
8873 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
8874 and latches only if requested. Fix latch if it is removed.
8875 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
8876 LOOPS_HAVE_PREHEADERS.
8878 2014-02-25 Andrew Pinski <apinski@cavium.com>
8880 * builtins.c (expand_builtin_thread_pointer): Create a new target
8881 when the target is NULL.
8883 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
8885 PR rtl-optimization/60317
8886 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
8887 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
8888 * lra-assigns.c: Include params.h.
8889 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
8890 other reload pseudos considerations.
8892 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8894 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
8895 to use canonical form for nor<mode>3.
8897 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8900 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
8903 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
8905 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
8906 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
8907 (ix86_handle_option): Handle OPT_mprefetchwt1.
8908 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
8909 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8911 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
8912 OPTION_MASK_ISA_PREFETCHWT1.
8913 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
8914 (PTA_PREFETCHWT1): New.
8915 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
8916 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
8917 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
8918 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
8919 (*prefetch_avx512pf_<mode>_: Change into ...
8920 (*prefetch_prefetchwt1_<mode>: This.
8921 * config/i386/i386.opt (mprefetchwt1): New.
8922 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
8923 (_mm_prefetch): Handle intent to write.
8924 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
8926 2014-02-25 Richard Biener <rguenther@suse.de>
8929 * emit-rtl.c (mem_attrs_htab): Remove.
8930 (mem_attrs_htab_hash): Likewise.
8931 (mem_attrs_htab_eq): Likewise.
8932 (set_mem_attrs): Always allocate new mem-attrs when something changed.
8933 (init_emit_once): Do not allocate mem_attrs_htab.
8935 2014-02-25 Richard Biener <rguenther@suse.de>
8938 * lto-opts.c (lto_write_options): Output non-explicit conservative
8939 -fwrapv, -fno-trapv and -fno-strict-overflow.
8940 * lto-wrapper.c (merge_and_complain): Handle merging those options.
8941 (run_gcc): And pass them through.
8943 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
8945 * sel-sched.c (calculate_new_fences): New parameter ptime.
8946 Calculate it as a maximum over all fence cycles.
8947 (sel_sched_region_2): Adjust the call to calculate_new_fences.
8948 Print the final schedule timing when sched_verbose.
8950 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
8952 PR rtl-optimization/60292
8953 * sel-sched.c (fill_vec_av_set): Do not reset target availability
8954 bit fot the fence instruction.
8956 2014-02-24 Alangi Derick <alangiderick@gmail.com>
8958 * calls.h: Fix typo in comment.
8960 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
8962 * config/pa/pa.c (pa_output_move_double): Don't valididate when
8963 adjusting offsetable addresses.
8965 2014-02-24 Guozhi Wei <carrot@google.com>
8967 * sparseset.h (sparseset_pop): Fix the wrong index.
8969 2014-02-24 Walter Lee <walt@tilera.com>
8971 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
8972 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
8974 * common/config/tilegx/tilegx-common.c
8975 (TARGET_DEFAULT_TARGET_FLAGS): Define.
8976 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
8978 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
8979 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
8980 (tilegx_gimplify_va_arg_expr): Handle big endian.
8981 (tilegx_expand_unaligned_load): Ditto.
8982 (tilegx_expand_unaligned_store): Ditto.
8983 (TARGET_RETURN_IN_MSB): New.
8984 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
8985 (TARGET_ENDIAN_DEFAULT): New.
8986 (TARGET_BIG_ENDIAN): Handle big endian.
8987 (BYTES_BIG_ENDIAN): Ditto.
8988 (WORDS_BIG_ENDIAN): Ditto.
8989 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
8992 * config/tilegx/tilegx.md (extv): Handle big endian.
8994 (insn_st<n>): Ditto.
8995 (insn_st<n>_add<bitsuffix>): Ditto.
8996 (insn_stnt<n>): Ditto.
8997 (insn_stnt<n>_add<bitsuffix>):Ditto.
8998 (vec_interleave_highv8qi): Handle big endian.
8999 (vec_interleave_highv8qi_be): New.
9000 (vec_interleave_highv8qi_le): New.
9001 (insn_v1int_h): Handle big endian.
9002 (vec_interleave_lowv8qi): Handle big endian.
9003 (vec_interleave_lowv8qi_be): New.
9004 (vec_interleave_lowv8qi_le): New.
9005 (insn_v1int_l): Handle big endian.
9006 (vec_interleave_highv4hi): Handle big endian.
9007 (vec_interleave_highv4hi_be): New.
9008 (vec_interleave_highv4hi_le): New.
9009 (insn_v2int_h): Handle big endian.
9010 (vec_interleave_lowv4hi): Handle big endian.
9011 (vec_interleave_lowv4hi_be): New.
9012 (vec_interleave_lowv4hi_le): New.
9013 (insn_v2int_l): Handle big endian.
9014 (vec_interleave_highv2si): Handle big endian.
9015 (vec_interleave_highv2si_be): New.
9016 (vec_interleave_highv2si_le): New.
9017 (insn_v4int_h): Handle big endian.
9018 (vec_interleave_lowv2si): Handle big endian.
9019 (vec_interleave_lowv2si_be): New.
9020 (vec_interleave_lowv2si_le): New.
9021 (insn_v4int_l): Handle big endian.
9022 * config/tilegx/tilegx.opt (mbig-endian): New option.
9023 (mlittle-endian): New option.
9024 * doc/install.texi: Document tilegxbe-linux.
9025 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
9027 2014-02-24 Martin Jambor <mjambor@suse.cz>
9030 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
9031 there are no parameter descriptors.
9033 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
9035 PR rtl-optimization/60268
9036 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
9037 initialization to ...
9038 (sched_rgn_init): ... here.
9039 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
9041 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
9043 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
9046 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
9048 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
9051 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
9053 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
9054 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
9056 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
9058 * config/microblaze/predicates.md: Add cmp_op predicate.
9059 * config/microblaze/microblaze.md: Add branch_compare instruction
9060 which uses cmp_op predicate and emits cmp insn before branch.
9061 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
9062 to microblaze_expand_conditional_branch and consolidate logic.
9063 (microblaze_expand_conditional_branch): emit branch_compare
9064 insn instead of handling cmp op separate from branch insn.
9066 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9068 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
9071 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9073 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
9074 define_insn with define_expand and new define_insn
9075 *altivec_lve<VI_char>x_internal.
9076 (altivec_stve<VI_char>x): Replace define_insn with define_expand
9077 and new define_insn *altivec_stve<VI_char>x_internal.
9078 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
9080 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
9082 (altivec_expand_stvex_be): New function.
9084 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
9086 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
9087 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
9088 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
9089 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
9091 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
9094 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
9095 instead of emit_move_insn.
9097 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9099 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
9101 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
9102 gen_altivec_vsumsws.
9104 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9106 * config/rs6000/altivec.md (altivec_lvxl): Rename as
9107 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
9108 (altivec_lvxl_<mode>): New define_expand incorporating
9109 -maltivec=be semantics where needed.
9110 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
9111 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
9112 semantics where needed.
9113 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
9114 (altivec_stvx_<mode>): New define_expand incorporating
9115 -maltivec=be semantics where needed.
9116 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
9117 VM2 iterator instead of V4SI.
9118 (altivec_stvxl_<mode>): New define_expand incorporating
9119 -maltivec=be semantics where needed.
9120 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
9121 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
9122 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
9123 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
9124 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
9125 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
9126 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
9127 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
9128 ALTIVEC_BUILTIN_STVXL.
9129 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
9130 (altivec_expand_stvx_be): Likewise.
9131 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
9132 (altivec_expand_lvx_be): Likewise.
9133 (altivec_expand_stvx_be): Likewise.
9134 (altivec_expand_builtin): Add cases for
9135 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
9136 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
9137 (altivec_init_builtins): Add definitions for
9138 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
9139 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
9141 2014-02-21 Catherine Moore <clm@codesourcery.com>
9143 * doc/invoke.texi (mvirt, mno-virt): Document.
9144 * config/mips/mips.opt (mvirt): New option.
9145 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
9147 2014-02-21 Richard Biener <rguenther@suse.de>
9149 PR tree-optimization/60276
9150 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
9151 (STMT_VINFO_MIN_NEG_DIST): New macro.
9152 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
9153 STMT_VINFO_MIN_NEG_DIST.
9154 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
9155 made for negative dependence distances still hold.
9157 2014-02-21 Richard Biener <rguenther@suse.de>
9160 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
9161 DECL_INITIAL for globals not in the current function context.
9163 2014-02-21 Jakub Jelinek <jakub@redhat.com>
9165 PR tree-optimization/56490
9166 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
9167 * tree-ssa-uninit.c: Include params.h.
9168 (compute_control_dep_chain): Add num_calls argument, return false
9169 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
9170 num_calls to recursive call.
9171 (find_predicates): Change dep_chain into normal array,
9172 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
9173 variable and adjust compute_control_dep_chain caller.
9174 (find_def_preds): Likewise.
9176 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
9178 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
9179 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
9181 2014-02-21 Nick Clifton <nickc@redhat.com>
9183 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
9184 (pushhi1): Likewise.
9185 (popqi1): Add mode to pre_dec.
9188 2014-02-21 Jakub Jelinek <jakub@redhat.com>
9190 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
9191 mode for mask of V8SFmode permutation.
9193 2014-02-20 Richard Henderson <rth@redhat.com>
9196 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
9197 a new pseudo for OLDVAL.
9199 2014-02-20 Jakub Jelinek <jakub@redhat.com>
9202 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
9203 gen_reg_rtx if d->testing_p.
9204 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
9205 if d->testing_p and we will certainly return true.
9206 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
9209 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
9211 * emit-rtl.c (gen_reg_rtx): Assert that
9212 crtl->emit.regno_pointer_align_length is non-zero.
9214 2014-02-20 Richard Henderson <rth@redhat.com>
9217 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
9218 on failure the store back into EXPECT.
9220 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
9221 Sandra Loosemore <sandra@codesourcery.com>
9223 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
9224 * config/nios2/nios2.c (nios2_function_profiler): Add
9225 -fPIC (flag_pic == 2) support.
9226 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
9227 (nios2_large_offset_p): New function.
9228 (nios2_unspec_reloc_p): Move up position, update to use
9229 nios2_large_offset_p.
9230 (nios2_unspec_address): Remove function.
9231 (nios2_unspec_offset): New function.
9232 (nios2_large_got_address): New function.
9233 (nios2_got_address): Add large offset support.
9234 (nios2_legitimize_tls_address): Update usage of removed and new
9236 (nios2_symbol_binds_local_p): New function.
9237 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
9238 (nios2_legitimize_address): Update to use nios2_large_offset_p.
9239 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
9240 (nios2_print_operand): Merge H/L processing, add hiadj/lo
9241 processing for (const (unspec ...)).
9242 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
9244 2014-02-20 Richard Biener <rguenther@suse.de>
9246 * tree-cfg.c (replace_uses_by): Mark altered BBs before
9247 doing the substitution.
9248 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
9250 2014-02-20 Martin Jambor <mjambor@suse.cz>
9253 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
9254 info when checking whether lattices are bottom.
9256 2014-02-20 Richard Biener <rguenther@suse.de>
9259 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
9262 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
9265 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
9266 parameter specifying the scaling.
9267 (inline_call): Update.
9268 (want_inline_recursively): Guard division by zero.
9269 (recursive_inlining): Update.
9270 * ipa-inline.h (clone_inlined_nodes): Update.
9272 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
9275 * config/i386/i386.c (classify_argument): Pass structures of size
9276 64 bytes or less in register.
9278 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
9279 Kirill Yukhin <kirill.yukhin@intel.com>
9281 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
9282 (_mm_rcp28_round_ss): Ditto.
9283 (_mm_rsqrt28_round_sd): Ditto.
9284 (_mm_rsqrt28_round_ss): Ditto.
9285 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
9286 (_mm_rcp14_round_ss): Ditto.
9287 (_mm_rsqrt14_round_sd): Ditto.
9288 (_mm_rsqrt14_round_ss): Ditto.
9289 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
9290 the first input operand, get rid of match_dup.
9291 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
9293 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9295 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
9296 operand as the first input operand, set type attribute.
9297 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9299 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
9300 operand as the first input operand, set type attribute.
9302 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9304 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
9307 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
9310 * config/i386/i386.c (construct_container): Remove TFmode check
9311 for X86_64_INTEGER_CLASS.
9313 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
9316 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
9317 only when -Wpsabi is enabled.
9319 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
9322 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
9323 passing arrays in registers are the same as for structs, so remove the
9324 special case for them.
9326 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
9328 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
9329 destination type, extract only the valid bits if the source type is not
9330 integral and has a different mode.
9332 2014-02-19 Richard Biener <rguenther@suse.de>
9335 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
9338 2014-02-19 Richard Biener <rguenther@suse.de>
9341 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
9342 (ipa_modify_call_arguments): Emit an argument load explicitely and
9343 preserve virtual SSA form there and for the replacement call.
9344 Do not update SSA form nor free dominance info.
9346 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9348 * ipa.c (function_and_variable_visibility): Also clear WEAK
9349 flag when disolving COMDAT_GROUP.
9351 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9353 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
9354 * ipa-prop.c (ipa_set_jf_known_type): Return early when
9356 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
9357 do more sanity checks.
9358 (detect_type_change): Return true when giving up early.
9359 (compute_complex_assign_jump_func): Fix type parameter of
9360 ipa_set_ancestor_jf.
9361 (compute_complex_ancestor_jump_func): Likewise.
9362 (update_jump_functions_after_inlining): Fix updating of
9364 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
9366 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9368 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
9369 inline clones when edge disappears.
9371 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
9374 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
9375 Split 64-bit moves into 2 patterns. Do not allow the use of
9376 direct move for TDmode in little endian, since the decimal value
9377 has little endian bytes within a word, but the 64-bit pieces are
9378 ordered in a big endian fashion, and normal subreg's of TDmode are
9380 (mov<mode>_64bit_dm): Likewise.
9381 (movtd_64bit_nodm): Likewise.
9383 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
9385 PR tree-optimization/60174
9386 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
9387 statement of an SSA_NAME that occurs in an abnormal PHI node.
9389 2014-02-18 Jakub Jelinek <jakub@redhat.com>
9392 * final.c (SEEN_BB): Remove.
9393 (SEEN_NOTE, SEEN_EMITTED): Renumber.
9394 (final_scan_insn): Don't force_source_line on second
9395 NOTE_INSN_BASIC_BLOCK.
9397 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
9400 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
9401 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
9402 (type_natural_mode): Warn ABI change when %zmm register is not
9403 available for AVX512F vector value passing.
9405 2014-02-18 Kai Tietz <ktietz@redhat.com>
9408 * config/i386/i386.c (ix86_expand_prologue): Use value in
9409 rax register as displacement when restoring %r10 or %rax.
9410 Fix wrong offset when restoring both registers.
9412 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
9414 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
9415 assertion with conditional return.
9417 2014-02-18 Jakub Jelinek <jakub@redhat.com>
9418 Uros Bizjak <ubizjak@gmail.com>
9421 * config/i386/driver-i386.c (host_detect_local_cpu): If
9422 YMM state is not saved by the OS, also clear has_f16c. Move
9423 CPUID 0x80000001 handling before YMM state saving checking.
9425 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
9427 PR rtl-optimization/58960
9428 * haifa-sched.c (alloc_global_sched_pressure_data): New,
9429 factored out from ...
9430 (sched_init): ... here.
9431 (free_global_sched_pressure_data): New, factored out from ...
9432 (sched_finish): ... here.
9433 * sched-int.h (free_global_sched_pressure_data): Declare.
9434 * sched-rgn.c (nr_regions_initial): New static global.
9435 (haifa_find_rgns): Initialize it.
9436 (schedule_region): Disable sched-pressure for the newly
9439 2014-02-17 Richard Biener <rguenther@suse.de>
9441 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
9442 release SSA defs of pattern stmts.
9444 2014-02-17 Richard Biener <rguenther@suse.de>
9446 * tree-inline.c (expand_call_inline): Release the virtual
9447 operand defined by the call we are about to inline.
9449 2014-02-17 Richard Biener <rguenther@suse.de>
9451 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
9453 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
9454 Ilya Tocar <ilya.tocar@intel.com>
9456 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
9457 arguments order in builtin.
9458 (_mm512_permutexvar_epi64): Ditto.
9459 (_mm512_mask_permutexvar_epi64): Ditto
9460 (_mm512_maskz_permutexvar_epi32): Ditto
9461 (_mm512_permutexvar_epi32): Ditto
9462 (_mm512_mask_permutexvar_epi32): Ditto
9464 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9466 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
9467 (p8_vmrgow): Likewise.
9469 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9471 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
9474 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
9477 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
9478 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
9479 into 64-bit and 32-bit moves. On 64-bit moves, add support for
9480 using direct move instructions on ISA 2.07. Also adjust
9481 instruction length for 64-bit.
9482 (mov<mode>_64bit, TFmode/TDmode): Likewise.
9483 (mov<mode>_32bit, TFmode/TDmode): Likewise.
9485 2014-02-15 Alan Modra <amodra@gmail.com>
9489 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
9490 find_replacement on parts of insn rtl that might be reloaded.
9492 2014-02-15 Richard Biener <rguenther@suse.de>
9494 PR tree-optimization/60183
9495 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
9496 (tree_ssa_phiprop): Calculate and free post-dominators.
9498 2014-02-14 Jeff Law <law@redhat.com>
9500 PR rtl-optimization/60131
9501 * ree.c (get_extended_src_reg): New function.
9502 (combine_reaching_defs): Use it rather than assuming location of REG.
9503 (find_and_remove_re): Verify first operand of extension is
9504 a REG before adding the insns to the copy list.
9506 2014-02-14 Roland McGrath <mcgrathr@google.com>
9508 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
9509 * configure: Regenerated.
9510 * config.in: Regenerated.
9511 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
9512 instead of ASM_SHORT.
9514 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
9515 Richard Earnshaw <rearnsha@arm.com>
9517 PR rtl-optimization/59535
9518 * lra-constraints.c (process_alt_operands): Encourage alternative
9519 when unassigned pseudo class is superset of the alternative class.
9520 (inherit_reload_reg): Don't inherit when optimizing for code size.
9521 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
9522 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
9523 modes not less than 4 for Thumb1.
9525 2014-02-14 Kyle McMartin <kyle@redhat.com>
9528 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
9530 2014-02-14 Richard Biener <rguenther@suse.de>
9532 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
9533 (get_frame_arg): Drop the assert with langhook types_compatible_p.
9534 Do not strip INDIRECT_REFs.
9536 2014-02-14 Richard Biener <rguenther@suse.de>
9539 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
9540 DECL_FUNCTION_SPECIFIC_TARGET.
9541 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
9542 * tree-streamer-out.c (pack_ts_target_option): Remove.
9543 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
9544 (write_ts_function_decl_tree_pointers): Do not stream
9545 DECL_FUNCTION_SPECIFIC_TARGET.
9546 * tree-streamer-in.c (unpack_ts_target_option): Remove.
9547 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
9548 (lto_input_ts_function_decl_tree_pointers): Do not stream
9549 DECL_FUNCTION_SPECIFIC_TARGET.
9551 2014-02-14 Jakub Jelinek <jakub@redhat.com>
9553 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
9554 (get_initial_def_for_induction, vectorizable_induction): Ignore
9555 debug stmts when looking for exit_phi.
9556 (vectorizable_live_operation): Fix up condition.
9558 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9560 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
9561 nreverse() because it changes the content of original tree list.
9563 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9565 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
9566 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
9568 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9570 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
9571 GNU coding standards.
9573 2014-02-13 Jakub Jelinek <jakub@redhat.com>
9576 * dwarf2out.c (gen_subprogram_die): Don't call
9577 add_calling_convention_attribute if subr_die is old_die.
9579 2014-02-13 Sharad Singhai <singhai@google.com>
9581 * doc/optinfo.texi: Fix order of nodes.
9583 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
9585 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
9586 operands[2], not operands[3].
9588 2014-02-13 Richard Biener <rguenther@suse.de>
9591 * doc/install.texi (ISL): Update recommended version to 0.12.2,
9592 mention the possibility of an in-tree build.
9593 (CLooG): Update recommended version to 0.18.1, mention the
9594 possibility of an in-tree build and clarify that the ISL
9595 bundled with CLooG does not work.
9597 2014-02-13 Jakub Jelinek <jakub@redhat.com>
9600 * expr.c (compress_float_constant): If x is a hard register,
9601 extend into a pseudo and then move to x.
9603 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
9605 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
9606 caused by bad second argument to warning_at() with -mhotpatch and
9607 nested functions (e.g. with gfortran).
9609 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
9611 * opts.c (option_name): Remove "enabled by default" rider.
9613 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
9615 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
9617 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
9618 Uros Bizjak <ubizjak@gmail.com>
9621 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
9622 * configure: Regenerated.
9624 2014-02-12 Richard Biener <rguenther@suse.de>
9626 * vec.c (vec_prefix::calculate_allocation): Move as
9627 inline variant to vec.h.
9628 (vec_prefix::calculate_allocation_1): New out-of-line version.
9629 * vec.h (vec_prefix::calculate_allocation_1): Declare.
9630 (vec_prefix::m_has_auto_buf): Rename to ...
9631 (vec_prefix::m_using_auto_storage): ... this.
9632 (vec_prefix::calculate_allocation): Inline the easy cases
9633 and dispatch to calculate_allocation_1 which doesn't need the
9635 (va_heap::reserve): Use gcc_checking_assert.
9636 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
9637 m_using_auto_storage.
9638 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
9640 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
9641 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
9642 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
9644 2014-02-12 Richard Biener <rguenther@suse.de>
9646 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
9647 when we found a dependence.
9649 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
9651 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
9653 (maybe_fold_stmt): ... into this new function.
9654 * omp-low.c (lower_omp): Update comment.
9656 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
9659 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
9662 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
9664 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
9665 identifiers in comments.
9666 (cortexa53_extra_costs): Likewise.
9667 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
9668 (cortexa7_extra_costs): Likewise.
9669 (cortexa12_extra_costs): Likewise.
9670 (cortexa15_extra_costs): Likewise.
9671 (v7m_extra_costs): Likewise.
9673 2014-02-12 Richard Biener <rguenther@suse.de>
9676 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
9677 of posix_memalign being successful.
9678 (lower_stmt): Restrict lowering of posix_memalign to when
9679 -ftree-bit-ccp is enabled.
9681 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
9683 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
9685 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
9687 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
9689 PR rtl-optimization/60116
9690 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
9691 other_insn once the combination has been validated.
9693 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
9696 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
9698 * ipa-devirt.c: Include demangle.h
9699 (odr_violation_reported): New static variable.
9700 (add_type_duplicate): Update odr_violations.
9701 (maybe_record_node): Add completep parameter; update it.
9702 (record_target_from_binfo): Add COMPLETEP parameter;
9703 update it as needed.
9704 (possible_polymorphic_call_targets_1): Likewise.
9705 (struct polymorphic_call_target_d): Add nonconstruction_targets;
9706 rename FINAL to COMPLETE.
9707 (record_targets_from_bases): Sanity check we found the binfo;
9708 fix COMPLETEP updating.
9709 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
9710 parameter, fix computing of COMPLETEP.
9711 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
9712 at LTO time do demangling.
9713 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
9714 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
9716 (gimple_get_virt_method_for_binfo): Likewise.
9717 * gimple-fold.h (gimple_get_virt_method_for_binfo,
9718 gimple_get_virt_method_for_vtable): Update prototypes.
9720 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
9723 * genautomata.c (add_presence_absence): Fix typo with
9724 {final_}presence_list.
9726 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
9729 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
9730 for VSX/Altivec vectors that land in GPR registers.
9732 2014-02-11 Richard Henderson <rth@redhat.com>
9733 Jakub Jelinek <jakub@redhat.com>
9736 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
9737 around drhs if type conversion to lacc->type is not useless.
9739 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9741 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
9743 (cortex-a57.cortex-a53): Likewise.
9744 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
9746 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9748 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
9751 2014-02-11 Renlin Li <Renlin.Li@arm.com>
9753 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
9754 add_options_for_arm_vfp3.
9756 2014-02-11 Jeff Law <law@redhat.com>
9759 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
9760 object with an undesirable mode.
9762 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9765 * config/i386/sol2-9.h: New file.
9766 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
9767 *-*-solaris2.9*): Use it.
9769 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
9771 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
9772 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
9774 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
9776 * config/microblaze/microblaze.c: Extend mcpu version format
9778 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
9780 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
9782 2014-02-10 Richard Henderson <rth@redhat.com>
9785 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
9786 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
9787 ms-abi vs -mno-accumulate-outgoing-args.
9788 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
9789 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
9792 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
9795 * cfgexpand.c (expand_asm_operands): Attach source location to
9796 ASM_INPUT rtx objects.
9797 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
9799 2014-02-10 Nick Clifton <nickc@redhat.com>
9801 * config/mn10300/mn10300.c (popcount): New function.
9802 (mn10300_expand_prologue): Include saved registers in stack usage
9805 2014-02-10 Jeff Law <law@redhat.com>
9808 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
9809 when changing the SET_DEST of a prior insn to avoid an input reload.
9811 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
9813 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
9814 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
9815 -mcall-openbsd, or -mcall-linux.
9816 (CC1_ENDIAN_BIG_SPEC): Remove.
9817 (CC1_ENDIAN_LITTLE_SPEC): Remove.
9818 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
9819 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
9820 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
9821 and %cc1_endian_default.
9822 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
9824 2014-02-10 Richard Biener <rguenther@suse.de>
9826 PR tree-optimization/60115
9827 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
9828 MEM_REF handling. Properly verify that the accesses are not
9829 out of the objects bound.
9831 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9833 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
9836 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
9838 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
9839 proper constants and fix formatting.
9840 (possible_polymorphic_call_targets): Fix formatting.
9842 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
9843 Ilya Tocar <ilya.tocar@intel.com>
9845 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
9846 (_mm512_loadu_epi32): Renamed into...
9847 (_mm512_loadu_si512): This.
9848 (_mm512_storeu_epi32): Renamed into...
9849 (_mm512_storeu_si512): This.
9850 (_mm512_maskz_ceil_ps): Removed.
9851 (_mm512_maskz_ceil_pd): Ditto.
9852 (_mm512_maskz_floor_ps): Ditto.
9853 (_mm512_maskz_floor_pd): Ditto.
9854 (_mm512_floor_round_ps): Ditto.
9855 (_mm512_floor_round_pd): Ditto.
9856 (_mm512_ceil_round_ps): Ditto.
9857 (_mm512_ceil_round_pd): Ditto.
9858 (_mm512_mask_floor_round_ps): Ditto.
9859 (_mm512_mask_floor_round_pd): Ditto.
9860 (_mm512_mask_ceil_round_ps): Ditto.
9861 (_mm512_mask_ceil_round_pd): Ditto.
9862 (_mm512_maskz_floor_round_ps): Ditto.
9863 (_mm512_maskz_floor_round_pd): Ditto.
9864 (_mm512_maskz_ceil_round_ps): Ditto.
9865 (_mm512_maskz_ceil_round_pd): Ditto.
9866 (_mm512_expand_pd): Ditto.
9867 (_mm512_expand_ps): Ditto.
9868 * config/i386/i386.c (ix86_builtins): Remove
9869 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
9870 (bdesc_args): Ditto.
9871 * config/i386/predicates.md (const1256_operand): New.
9872 (const_1_to_2_operand): Ditto.
9873 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
9874 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9875 (*avx512pf_gatherpf<mode>sf): Ditto.
9876 (avx512pf_gatherpf<mode>df): Ditto.
9877 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9878 (*avx512pf_gatherpf<mode>df): Ditto.
9879 (avx512pf_scatterpf<mode>sf): Ditto.
9880 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9881 (*avx512pf_scatterpf<mode>sf): Ditto.
9882 (avx512pf_scatterpf<mode>df): Ditto.
9883 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9884 (*avx512pf_scatterpf<mode>df): Ditto.
9885 (avx512f_expand<mode>): Removed.
9886 (<shift_insn><mode>3<mask_name>): Change predicate type.
9888 2014-02-08 Jakub Jelinek <jakub@redhat.com>
9890 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
9891 not at the end of datarefs vector use ordered_remove to avoid
9892 reordering datarefs vector.
9895 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
9896 mark local addressable non-static vars as GOVD_PRIVATE
9897 instead of GOVD_LOCAL.
9898 * omp-low.c (lower_omp_for): Move gimple_bind_vars
9899 and BLOCK_VARS of gimple_bind_block to new_stmt rather
9903 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
9904 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
9905 assume_aligned or alloc_align attributes.
9906 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
9907 arguments. Handle also assume_aligned and alloc_align attributes.
9908 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
9909 calls to functions with assume_aligned or alloc_align attributes.
9910 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
9912 2014-02-08 Terry Guo <terry.guo@arm.com>
9914 * doc/invoke.texi: Document ARM -march=armv7e-m.
9916 2014-02-08 Jakub Jelinek <jakub@redhat.com>
9918 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
9919 flag on __cilkrts_rethrow builtin.
9922 * ipa-cp.c (determine_versionability): Fail at -O0
9923 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
9924 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
9927 2014-02-04 Jakub Jelinek <jakub@redhat.com>
9930 * tree-inline.c (copy_forbidden): Fail for
9931 __attribute__((optimize (0))) functions.
9933 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
9935 * varpool.c: Include pointer-set.h.
9936 (varpool_remove_unreferenced_decls): Variables in other partitions
9937 will not be output; be however careful to not lose information
9940 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
9942 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
9943 lookup in the vtable constructor.
9945 2014-02-07 Jeff Law <law@redhat.com>
9948 * config/m68k/m68k.md (ashldi_extsi): Turn into a
9949 define_insn_and_split.
9951 * ipa-inline.c (inline_small_functions): Fix typos.
9953 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9955 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
9956 (s390_can_use_return_insn): Declare.
9957 * config/s390/s390.h (EPILOGUE_USES): Define.
9958 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
9960 (s390_chunkify_start): Handle return JUMP_LABELs.
9961 (s390_early_mach): Emit a main_pool instruction on the entry edge.
9962 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
9963 (s390_can_use_return_insn): New functions.
9964 (s390_fix_long_loop_prediction): Handle conditional returns.
9965 (TARGET_SET_UP_BY_PROLOGUE): Define.
9966 * config/s390/s390.md (ANY_RETURN): New code iterator.
9967 (*creturn, *csimple_return, return, simple_return): New patterns.
9969 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9971 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
9972 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
9973 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
9974 REG_CFA_RESTORE list when deciding not to restore a register.
9976 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9978 * config/s390/s390.c: Include tree-pass.h and context.h.
9979 (s390_early_mach): New function, split out from...
9980 (s390_emit_prologue): ...here.
9981 (pass_data_s390_early_mach): New pass structure.
9982 (pass_s390_early_mach): New class.
9983 (s390_option_override): Create and register early_mach pass.
9984 Move to end of file.
9986 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9988 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
9989 to match for the exit block.
9991 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9993 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
9994 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
9995 Reject misaligned operands.
9997 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9999 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
10001 2014-02-07 Richard Biener <rguenther@suse.de>
10003 PR middle-end/60092
10004 * gimple-low.c (lower_builtin_posix_memalign): New function.
10005 (lower_stmt): Call it to lower posix_memalign in a way
10006 to make alignment info accessible.
10008 2014-02-07 Jakub Jelinek <jakub@redhat.com>
10011 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
10012 __builtin_setjmp_receiver.
10014 2014-02-07 Richard Biener <rguenther@suse.de>
10016 PR middle-end/60092
10017 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
10018 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
10019 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
10020 Handle BUILT_IN_POSIX_MEMALIGN.
10021 (find_func_clobbers): Likewise.
10022 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
10023 (call_may_clobber_ref_p_1): Likewise.
10025 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10028 * ipa-devirt.c (record_target_from_binfo): Remove overactive
10031 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10034 * lto-cgraph.c (lto_output_node): Use
10035 symtab_get_symbol_partitioning_class.
10036 (lto_output_varpool_node): likewise.
10037 (symtab_get_symbol_partitioning_class): Move here from
10038 lto/lto-partition.c
10039 * cgraph.h (symbol_partitioning_class): Likewise.
10040 (symtab_get_symbol_partitioning_class): Declare.
10042 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10044 * ggc.h (ggc_internal_cleared_alloc): New macro.
10045 * vec.h (vec_safe_copy): Handle memory stats.
10046 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
10047 * target-globals.c (save_target_globals): Likewise.
10049 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10052 * expr.c (emit_move_resolve_push): Export; be bit more selective
10053 on when to clear alias set.
10054 * expr.h (emit_move_resolve_push): Declare.
10055 * function.h (struct function): Add tail_call_marked.
10056 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
10057 * config/i386/i386-protos.h (ix86_expand_push): Remove.
10058 * config/i386/i386.md (TImode move expander): De not call
10060 (FP push expanders): Preserve memory attributes.
10061 * config/i386/sse.md (push<mode>1): Remove.
10062 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
10063 (ix86_expand_push): Remove.
10064 * config/i386/mmx.md (push<mode>1): Remove.
10066 2014-02-06 Jakub Jelinek <jakub@redhat.com>
10068 PR rtl-optimization/60030
10069 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
10070 lopart with paradoxical subreg before shifting it up by hprec.
10072 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10074 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
10075 Remove extra newline at end of file.
10076 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
10077 (arm_issue_rate): Handle cortexa57.
10078 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
10079 (cortex-a57.cortex-a53): Likewise.
10081 2014-02-06 Jakub Jelinek <jakub@redhat.com>
10084 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
10085 don't record in REG_FRAME_RELATED_EXPR registers not set in that
10087 (arm_expand_prologue): Adjust all callers.
10088 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
10089 info, registers also at the lowest numbered registers side. Use
10090 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
10094 * var-tracking.c (adjust_mems): Before adding a SET to
10095 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
10097 2014-02-06 Alan Modra <amodra@gmail.com>
10100 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
10101 change SDmode to DDmode when lra_in_progress.
10103 2014-02-06 Jakub Jelinek <jakub@redhat.com>
10105 PR middle-end/59150
10106 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
10107 free_data_ref on the dr first, and before goto again also set dr
10108 to the next dr. For simd_lane_access, free old datarefs[i] before
10109 overwriting it. For get_vectype_for_scalar_type failure, don't
10110 free_data_ref if simd_lane_access.
10112 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
10115 * tree.h (opts_for_fn): New inline function.
10116 (opt_for_fn): Define.
10117 * config/i386/i386.c (ix86_function_regparm): Use
10118 opt_for_fn (decl, optimize) instead of optimize.
10120 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
10122 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
10123 for SYMBOL_REF in large memory model.
10125 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10127 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
10128 and crypto support.
10129 (cortex-a57): Likewise.
10130 (cortex-a57.cortex-a53): Likewise.
10132 2014-02-06 Yury Gribov <y.gribov@samsung.com>
10133 Kugan Vivekanandarajah <kuganv@linaro.org>
10135 * config/arm/arm.c (arm_vector_alignment_reachable): Check
10137 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
10139 2014-02-06 Richard Biener <rguenther@suse.de>
10141 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
10142 set_loop_copy and initialize_original_copy_tables.
10144 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
10146 * config/aarch64/aarch64-simd.md
10147 (aarch64_ashr_simddi): Change QI to SI.
10149 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
10150 Jakub Jelinek <jakub@redhat.com>
10152 PR middle-end/60013
10153 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
10156 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10158 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
10159 CODE_FOR_altivec_vpku[hw]um to
10160 CODE_FOR_altivec_vpku[hw]um_direct.
10161 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
10162 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
10163 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
10164 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
10166 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10168 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
10169 generation for -maltivec=be.
10170 (altivec_vsumsws): Simplify redundant test.
10172 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10174 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
10175 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
10176 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
10177 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
10178 gen_altivec_vpkuwum.
10179 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
10181 (altivec_vpks<VI_char>ss): Likewise.
10182 (altivec_vpks<VI_char>us): Likewise.
10183 (altivec_vpku<VI_char>us): Likewise.
10184 (altivec_vpku<VI_char>um): Likewise.
10185 (altivec_vpku<VI_char>um_direct): New (copy of
10186 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
10188 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
10189 target is little endian and -maltivec=be is not specified.
10190 (*altivec_vupkhs<VU_char>_direct): New (copy of
10191 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
10192 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
10193 target is little endian and -maltivec=be is not specified.
10194 (*altivec_vupkls<VU_char>_direct): New (copy of
10195 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
10196 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
10197 little endian and -maltivec=be is not specified.
10198 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
10199 little endian and -maltivec=be is not specified.
10201 2014-02-05 Richard Henderson <rth@redhat.com>
10204 * combine-stack-adj.c: Revert r206943.
10205 * sched-int.h (struct deps_desc): Add last_args_size.
10206 * sched-deps.c (init_deps): Initialize it.
10207 (sched_analyze_insn): Add OUTPUT dependencies between insns that
10208 contain REG_ARGS_SIZE notes.
10210 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
10212 * lto-cgraph.c (asm_nodes_output): Make global.
10213 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
10214 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
10215 (driver_handle_option): Handle OPT_fwpa.
10217 2014-02-05 Jakub Jelinek <jakub@redhat.com>
10220 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
10221 a comment typo and formatting issue. If odr_hash hasn't been
10222 created, return vNULL and set *completep to false.
10224 PR middle-end/57499
10225 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
10226 bb with no successors.
10228 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
10231 * doc/invoke.texi (-march): Clarify documentation for ARM.
10232 (-mtune): Likewise.
10235 2014-02-05 Richard Biener <rguenther@suse.de>
10237 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
10238 when not vectorizing because of too many alias checks.
10239 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
10240 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
10242 2014-02-05 Nick Clifton <nickc@redhat.com>
10244 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
10245 accept extended registers in any mode when compiling for the MN10300.
10247 2014-02-05 Yury Gribov <y.gribov@samsung.com>
10249 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
10250 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
10251 sanitization attributes.
10252 (can_inline_edge_p): Likewise.
10253 (sanitize_attrs_match_for_inline_p): New function.
10255 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10257 * ipa-prop.c (detect_type_change): Shor circuit testing of
10258 type changes on THIS pointer.
10260 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
10263 * config/pa/pa.c (legitimize_tls_address): Return original address
10264 if not passed a SYMBOL_REF rtx.
10265 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
10267 (pa_emit_move_sequence): Simplify TLS source operands.
10268 (pa_legitimate_constant_p): Reject all TLS constants.
10269 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
10270 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
10272 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10274 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
10275 groups when we know they are controlled by LTO.
10276 * varasm.c (default_binds_local_p_1): If object is in other partition,
10277 it will be resolved locally.
10279 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
10281 * config/host-linux.c (linux_gt_pch_use_address): Don't
10282 use SSIZE_MAX because it is not always defined.
10284 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
10287 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
10288 threshold for pseudo splitting.
10289 (update_ebb_live_info): Process call argument hard registers and
10290 hard registers from insn definition too.
10291 (max_small_class_regs_num): New constant.
10292 (inherit_in_ebb): Update live hard regs through EBBs. Update
10293 reloads_num only for small register classes. Don't split for
10296 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
10299 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
10302 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10304 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
10305 visibility is safe.
10307 2014-02-04 Marek Polacek <polacek@redhat.com>
10309 * gdbinit.in (pel): Define.
10311 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
10313 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
10316 2014-02-04 Richard Biener <rguenther@suse.de>
10319 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
10320 in function context local.
10321 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
10322 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
10323 similar to LTO_imported_decl_ref.
10325 2014-02-04 Jakub Jelinek <jakub@redhat.com>
10327 PR tree-optimization/60002
10328 * cgraphclones.c (build_function_decl_skip_args): Clear
10329 DECL_LANG_SPECIFIC.
10331 PR tree-optimization/60023
10332 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
10333 false to gsi_replace.
10334 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
10335 has been in some EH region and vec_stmt could throw, add
10336 vec_stmt into the same EH region.
10337 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
10338 has no lhs, ignore it.
10339 * internal-fn.c (expand_MASK_LOAD): Likewise.
10342 * tree-inline.c (copy_forbidden): Fail for
10343 __attribute__((optimize (0))) functions.
10346 * omp-low.c (simd_clone_struct_copy): If from->inbranch
10347 is set, copy one less argument.
10348 (expand_simd_clones): Don't subtract clone_info->inbranch
10349 from simd_clone_struct_alloc argument.
10351 PR rtl-optimization/57915
10352 * recog.c (simplify_while_replacing): If all unary/binary/relational
10353 operation arguments are constant, attempt to simplify those.
10355 PR middle-end/59261
10356 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
10357 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
10359 2014-02-04 Richard Biener <rguenther@suse.de>
10361 PR tree-optimization/60012
10362 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
10363 TBAA disambiguation to all DDRs.
10365 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10368 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
10369 (LINK_SPEC): Use it for -shared, -shared-libgcc.
10371 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10374 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
10376 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10378 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
10379 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
10381 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10384 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
10385 to figure out targets of polymorphic calls with known decl.
10386 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10387 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
10388 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
10389 (get_polymorphic_call_info): ... here.
10390 (get_polymorphic_call_info_from_invariant): New function.
10392 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10394 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
10395 lookup via vtable pointer; check for type consistency
10396 and turn inconsitent facts into UNREACHABLE.
10397 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10398 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
10399 type inconsistent querries; return UNREACHABLE instead.
10401 2014-02-03 Richard Henderson <rth@twiddle.net>
10404 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
10405 already processed this node.
10406 (normalize_one_pred_1): Pass along mark_set.
10407 (normalize_one_pred): Create and destroy a pointer_set_t.
10408 (normalize_one_pred_chain): Likewise.
10410 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
10412 PR gcov-profile/58602
10413 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
10415 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10418 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
10419 -fno-devirtualize; try to devirtualize by the knowledge of
10420 virtual table pointer given by aggregate propagation.
10421 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10422 (ipa_print_node_jump_functions): Dump also offset that
10423 is relevant for polymorphic calls.
10424 (determine_known_aggregate_parts): Add arg_type parameter; use it
10425 instead of determining the type from pointer type.
10426 (ipa_compute_jump_functions_for_edge): Update call of
10427 determine_known_aggregate_parts.
10428 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
10429 (gimple_get_virt_method_for_binfo): ... here; simplify using
10430 vtable_pointer_value_to_vtable.
10431 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
10432 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
10433 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
10434 (vtable_pointer_value_to_vtable): Break out from ...; handle also
10436 (vtable_pointer_value_to_binfo): ... here.
10437 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
10439 2014-02-03 Teresa Johnson <tejohnson@google.com>
10441 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
10442 redef of outer loop index variable.
10444 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
10448 * doc/extend.texi (Function Attributes): Typo.
10450 2014-02-03 Cong Hou <congh@google.com>
10452 PR tree-optimization/60000
10453 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
10454 if the vectorized statement is a store. A store statement can only
10455 appear at the end of pattern statements.
10457 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
10459 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
10460 (ix86_option_override_internal): Default long double to 64-bit for
10461 32-bit Bionic and to 128-bit for 64-bit Bionic.
10463 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
10464 TARGET_LONG_DOUBLE_128 is true.
10465 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
10467 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
10468 (mlong-double-64): Negate -mlong-double-128.
10469 (mlong-double-128): New option.
10471 * config/i386/i386-c.c (ix86_target_macros): Define
10472 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
10474 * doc/invoke.texi: Document -mlong-double-128.
10476 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
10478 PR rtl-optimization/60024
10479 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
10481 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
10483 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
10485 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
10487 PR rtl-optimization/57662
10488 * sel-sched.c (code_motion_path_driver): Do not mark already not
10489 existing blocks in the visiting bitmap.
10491 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
10493 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
10494 on the insn being emitted.
10496 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
10497 Will Deacon <will.deacon@arm.com>
10499 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
10501 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10503 * config/arm/arm-tables.opt: Regenerate.
10505 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10507 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
10508 for vector types other than V16QImode.
10509 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
10510 define_expand, and call altivec_expand_vec_perm_le when producing
10511 code with little endian element order.
10512 (*altivec_vperm_<mode>_internal): New insn having previous
10513 behavior of altivec_vperm_<mode>.
10514 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
10515 altivec_expand_vec_perm_le when producing code with little endian
10517 (*altivec_vperm_<mode>_uns_internal): New insn having previous
10518 behavior of altivec_vperm_<mode>_uns.
10520 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10522 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
10523 (altivec_vsumsws): Add handling for -maltivec=be with a little
10525 (altivec_vsumsws_direct): New.
10526 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
10527 gen_altivec_vsumsws.
10529 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
10531 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
10532 vtable_pointer_value_to_binfo): New functions.
10533 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
10534 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
10536 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
10538 * config/nios2/nios2.md (load_got_register): Initialize GOT
10539 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
10540 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
10542 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
10544 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
10545 preserverd by passthrough, do not propagate the type.
10547 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10549 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
10550 (mips_atomic_assign_expand_fenv): New function.
10551 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
10553 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10555 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
10556 (__builtin_mips_set_fcsr): Likewise.
10557 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
10558 MIPS_USI_FTYPE_VOID.
10559 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
10560 (mips16_expand_set_fcsr): Likewise.
10561 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
10562 (mips16_set_fcsr_stub): Likewise.
10563 (mips16_get_fcsr_one_only_stub): New class.
10564 (mips16_set_fcsr_one_only_stub): Likewise.
10565 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
10566 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
10567 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
10568 (hard_float): New availability predicate.
10569 (mips_builtins): Add get_fcsr and set_fcsr.
10570 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
10571 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
10572 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
10573 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
10574 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
10577 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10579 * config/mips/mips.c (mips_one_only_stub): New class.
10580 (mips_need_mips16_rdhwr_p): Replace with...
10581 (mips16_rdhwr_stub): ...this new variable.
10582 (mips16_stub_call_address): New function.
10583 (mips16_rdhwr_one_only_stub): New class.
10584 (mips_expand_thread_pointer): Use mips16_stub_call_address.
10585 (mips_output_mips16_rdhwr): Delete.
10586 (mips_finish_stub): New function.
10587 (mips_code_end): Use it to handle rdhwr stubs.
10589 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
10592 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
10593 when calculating size of integer atomic types.
10595 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
10597 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
10599 2014-02-01 Jakub Jelinek <jakub@redhat.com>
10601 PR tree-optimization/60003
10602 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
10603 * profile.c (branch_prob): Use gimple_call_builtin_p
10604 to check for BUILT_IN_SETJMP_RECEIVER.
10605 * tree-inline.c (copy_bb): Call notice_special_calls.
10607 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
10610 * lra-constraints.c (process_alt_operands): Update reload_sum only
10613 2014-01-31 Richard Henderson <rth@redhat.com>
10615 PR middle-end/60004
10616 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
10617 until after else_eh is processed.
10619 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
10621 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
10622 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
10623 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
10624 in smmintrin.h, remove them.
10625 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
10626 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
10627 * config/i386/i386.md (ROUND_SAE): Fix value.
10628 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
10629 (const48_operand): New.
10630 * config/i386/subst.md (round), (round_expand): Use
10631 const_4_or_8_to_11_operand.
10632 (round_saeonly), (round_saeonly_expand): Use const48_operand.
10634 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
10636 * config/i386/constraints.md (Yk): Swap meaning with k.
10637 * config/i386/i386.md (movhi_internal): Change Yk to k.
10638 (movqi_internal): Ditto.
10639 (*k<logic><mode>): Ditto.
10642 (kandn<mode>): Ditto.
10643 (*<code>hi_1): Ditto.
10644 (*<code>qi_1): Ditto.
10645 (kxnor<mode>): Ditto.
10646 (kortestzhi): Ditto.
10647 (kortestchi): Ditto.
10649 (*one_cmplhi2_1): Ditto.
10650 (*one_cmplqi2_1): Ditto.
10651 * config/i386/sse.md (): Change k to Yk.
10652 (avx512f_load<mode>_mask): Ditto.
10653 (avx512f_blendm<mode>): Ditto.
10654 (avx512f_store<mode>_mask): Ditto.
10655 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
10656 (avx512f_storedqu<mode>_mask): Ditto.
10657 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
10659 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
10660 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
10661 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
10662 (avx512f_maskcmp<mode>3): Ditto.
10663 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
10664 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
10665 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
10666 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
10667 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
10668 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
10669 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
10670 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
10671 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
10672 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
10673 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
10674 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
10675 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
10676 (vec_extract_lo_<mode>_maskm): Ditto.
10677 (vec_extract_hi_<mode>_maskm): Ditto.
10678 (avx512f_vternlog<mode>_mask): Ditto.
10679 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
10680 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
10681 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
10682 (avx512f_<code>v8div16qi2_mask): Ditto.
10683 (avx512f_<code>v8div16qi2_mask_store): Ditto.
10684 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
10685 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
10686 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
10687 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
10688 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
10689 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10690 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
10691 (*avx512pf_scatterpf<mode>df_mask): Ditto.
10692 (avx512cd_maskb_vec_dupv8di): Ditto.
10693 (avx512cd_maskw_vec_dupv16si): Ditto.
10694 (avx512f_vpermi2var<mode>3_maskz): Ditto.
10695 (avx512f_vpermi2var<mode>3_mask): Ditto.
10696 (avx512f_vpermi2var<mode>3_mask): Ditto.
10697 (avx512f_vpermt2var<mode>3_maskz): Ditto.
10698 (*avx512f_gathersi<mode>): Ditto.
10699 (*avx512f_gathersi<mode>_2): Ditto.
10700 (*avx512f_gatherdi<mode>): Ditto.
10701 (*avx512f_gatherdi<mode>_2): Ditto.
10702 (*avx512f_scattersi<mode>): Ditto.
10703 (*avx512f_scatterdi<mode>): Ditto.
10704 (avx512f_compress<mode>_mask): Ditto.
10705 (avx512f_compressstore<mode>_mask): Ditto.
10706 (avx512f_expand<mode>_mask): Ditto.
10707 * config/i386/subst.md (mask): Change k to Yk.
10708 (mask_scalar_merge): Ditto.
10711 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
10713 * doc/extend.texi (Vector Extensions): Document ?: in C++.
10715 2014-01-31 Richard Biener <rguenther@suse.de>
10717 PR middle-end/59990
10718 * builtins.c (fold_builtin_memory_op): Make sure to not
10719 use a floating-point mode or a boolean or enumeral type for
10720 the copy operation.
10722 2014-01-30 DJ Delorie <dj@redhat.com>
10724 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
10725 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
10726 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
10727 whenever main() has an epilogue.
10729 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10731 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
10732 unused variable "field".
10733 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
10734 (vsx_mergeh_<mode>): Likewise.
10735 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
10736 (altivec_vmrghh): Likewise.
10737 (altivec_vmrghw): Likewise.
10738 (altivec_vmrglb): Likewise.
10739 (altivec_vmrglh): Likewise.
10740 (altivec_vmrglw): Likewise.
10741 (altivec_vspltb): Add missing uses.
10742 (altivec_vsplth): Likewise.
10743 (altivec_vspltw): Likewise.
10744 (altivec_vspltsf): Likewise.
10746 2014-01-30 Jakub Jelinek <jakub@redhat.com>
10749 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
10750 frame related instructions.
10752 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
10754 PR rtl-optimization/59959
10755 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
10756 any reload of register whose subreg is invalid.
10758 2014-01-30 Jakub Jelinek <jakub@redhat.com>
10760 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
10761 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
10762 Add missing return type - void.
10764 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10766 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
10767 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
10768 remove element index adjustment for endian (now handled in vsx.md
10770 (altivec_expand_vec_perm_const): Use
10771 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
10772 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
10773 (vsx_xxspltw_<mode>): Adjust element index for little endian.
10774 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
10775 define_expand and a new define_insn *altivec_vspltb_internal;
10776 adjust for -maltivec=be on a little endian target.
10777 (altivec_vspltb_direct): New.
10778 (altivec_vsplth): Divide into a define_expand and a new
10779 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
10780 little endian target.
10781 (altivec_vsplth_direct): New.
10782 (altivec_vspltw): Divide into a define_expand and a new
10783 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
10784 little endian target.
10785 (altivec_vspltw_direct): New.
10786 (altivec_vspltsf): Divide into a define_expand and a new
10787 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
10788 a little endian target.
10790 2014-01-30 Richard Biener <rguenther@suse.de>
10792 PR tree-optimization/59993
10793 * tree-ssa-forwprop.c (associate_pointerplus): Check we
10794 can propagate form the earlier stmt and avoid the transform
10795 when the intermediate result is needed.
10797 2014-01-30 Alangi Derick <alangiderick@gmail.com>
10799 * README.Portability: Fix typo.
10801 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
10803 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
10804 comparison_operator with ordered_comparison_operator.
10806 2014-01-30 Nick Clifton <nickc@redhat.com>
10808 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
10809 Rename to mn10300_store_multiple_regs.
10810 * config/mn10300/mn10300.c: Likewise.
10811 * config/mn10300/mn10300.md (store_movm): Fix typo: call
10812 store_multiple_regs.
10813 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
10814 Call mn10300_store_multiple_regs.
10816 2014-01-30 Nick Clifton <nickc@redhat.com>
10817 DJ Delorie <dj@redhat.com>
10819 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
10820 %fp 2 to keep registers after it properly word-aligned.
10821 (rl78_alloc_physical_registers_umul): Handle the case where both
10822 input operands are the same.
10824 2014-01-30 Richard Biener <rguenther@suse.de>
10826 PR tree-optimization/59903
10827 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
10830 2014-01-30 Jason Merrill <jason@redhat.com>
10833 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
10836 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
10838 2014-01-30 Richard Biener <rguenther@suse.de>
10840 PR tree-optimization/59951
10841 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
10843 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
10846 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
10847 SFmode to DFmode case.
10849 2014-01-29 DJ Delorie <dj@redhat.com>
10851 * config/msp430/msp430.opt (-minrt): New.
10852 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
10854 (ENDFILE_SPEC): Likewise.
10856 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
10858 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
10859 (estimate_function_body_sizes): Use it.
10861 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
10864 * dwarf2out.c (is_cxx_auto): New.
10865 (is_base_type): Use it.
10866 (gen_type_die_with_usage): Likewise.
10868 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10870 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
10871 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
10872 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
10873 -maltivec=be with LE targets.
10874 (vsx_mergeh_<mode>): Likewise.
10875 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
10876 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
10877 (altivec_vmrghb): Replace with define_expand and new
10878 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
10879 (altivec_vmrghb_direct): New define_insn.
10880 (altivec_vmrghh): Replace with define_expand and new
10881 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
10882 (altivec_vmrghh_direct): New define_insn.
10883 (altivec_vmrghw): Replace with define_expand and new
10884 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
10885 (altivec_vmrghw_direct): New define_insn.
10886 (*altivec_vmrghsf): Adjust for endianness.
10887 (altivec_vmrglb): Replace with define_expand and new
10888 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
10889 (altivec_vmrglb_direct): New define_insn.
10890 (altivec_vmrglh): Replace with define_expand and new
10891 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
10892 (altivec_vmrglh_direct): New define_insn.
10893 (altivec_vmrglw): Replace with define_expand and new
10894 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
10895 (altivec_vmrglw_direct): New define_insn.
10896 (*altivec_vmrglsf): Adjust for endianness.
10897 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
10898 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
10899 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
10900 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
10901 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
10902 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
10903 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
10904 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
10906 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
10908 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
10909 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
10912 2014-01-29 Richard Biener <rguenther@suse.de>
10914 PR tree-optimization/58742
10915 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
10916 associate_pointerplus_align.
10917 (associate_pointerplus_diff): New function.
10918 (associate_pointerplus): Likewise. Call associate_pointerplus_align
10919 and associate_pointerplus_diff.
10921 2014-01-29 Richard Biener <rguenther@suse.de>
10923 * lto-streamer.h (LTO_major_version): Bump to 3.
10924 (LTO_minor_version): Reset to 0.
10926 2014-01-29 Renlin Li <Renlin.Li@arm.com>
10928 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
10929 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
10930 (arm_file_start): Generate correct asm header for armv7ve.
10931 * config/arm/bpabi.h: Add multilib support for armv7ve.
10932 * config/arm/driver-arm.c: Change the architectures of cortex-a7
10933 and cortex-a15 to armv7ve.
10934 * config/arm/t-aprofile: Add multilib support for armv7ve.
10935 * doc/invoke.texi: Document -march=armv7ve.
10937 2014-01-29 Richard Biener <rguenther@suse.de>
10939 PR tree-optimization/58742
10940 * tree-ssa-forwprop.c (associate_plusminus): Return true
10941 if we changed sth, defer EH cleanup to ...
10942 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
10943 (simplify_mult): New function.
10945 2014-01-29 Jakub Jelinek <jakub@redhat.com>
10947 PR middle-end/59917
10948 PR tree-optimization/59920
10949 * tree.c (build_common_builtin_nodes): Remove
10950 __builtin_setjmp_dispatcher initialization.
10951 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
10952 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
10953 instead of gsi_after_labels + manually skipping debug stmts.
10954 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
10955 ignore bbs with IFN_ABNORMAL_DISPATCHER.
10956 * tree-inline.c (copy_edges_for_bb): Remove
10957 can_make_abnormal_goto argument, instead add abnormal_goto_dest
10958 argument. Ignore computed_goto_p stmts. Don't call
10959 make_abnormal_goto_edges. If a call might need abnormal edges
10960 for non-local gotos, see if it already has an edge to
10961 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
10962 with true argument, don't do anything then, otherwise add
10963 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
10964 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
10966 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
10967 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
10968 (lower_stmt): Don't set data->calls_builtin_setjmp.
10969 (lower_builtin_setjmp): Adjust comment.
10970 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
10971 * tree-cfg.c (found_computed_goto): Remove.
10972 (factor_computed_gotos): Remove.
10973 (make_goto_expr_edges): Return bool, true for computed gotos.
10974 Don't call make_abnormal_goto_edges.
10975 (build_gimple_cfg): Don't set found_computed_goto, don't call
10976 factor_computed_gotos.
10977 (computed_goto_p): No longer static.
10978 (make_blocks): Don't set found_computed_goto.
10979 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
10980 (make_edges): If make_goto_expr_edges returns true, push bb
10981 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
10982 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
10983 vector. Record mapping between bbs and OpenMP regions if there
10984 are any, adjust make_gimple_omp_edges caller. Call
10985 handle_abnormal_edges.
10986 (make_abnormal_goto_edges): Remove.
10987 * tree-cfg.h (make_abnormal_goto_edges): Remove.
10988 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
10989 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
10990 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
10991 * internal-fn.def (ABNORMAL_DISPATCHER): New.
10992 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
10993 filling *region also set *region_idx to (*region)->entry->index.
10996 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
10997 For REGs set ORIGINAL_REGNO.
10999 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
11001 * doc/md.texi: Mention that a target shouldn't implement
11002 vec_widen_(s|u)mul_even/odd pair if it is less efficient
11005 2014-01-29 Jakub Jelinek <jakub@redhat.com>
11007 PR tree-optimization/59594
11008 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
11009 a copy of the datarefs vector rather than the vector itself.
11011 2014-01-28 Jason Merrill <jason@redhat.com>
11014 * dwarf2out.c (auto_die): New static.
11015 (gen_type_die_with_usage): Handle C++1y 'auto'.
11016 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
11019 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
11022 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
11023 (SPEC_X32): Likewise.
11024 (SPEC_64): Likewise.
11025 * config/i386/i386.c (ix86_option_override_internal): Turn off
11026 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
11028 (x86_file_start): Output .code16gcc for TARGET_16BIT.
11029 * config/i386/i386.h (TARGET_16BIT): New macro.
11030 (TARGET_16BIT_P): Likewise.
11031 * config/i386/i386.opt: Add m16.
11032 * doc/invoke.texi: Document -m16.
11034 2014-01-28 Jakub Jelinek <jakub@redhat.com>
11036 PR preprocessor/59935
11037 * input.c (location_get_source_line): Bail out on when line number
11038 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
11040 2014-01-28 Richard Biener <rguenther@suse.de>
11042 PR tree-optimization/58742
11043 * tree-ssa-forwprop.c (associate_plusminus): Handle
11044 pointer subtraction of the form (T)(P + A) - (T)P.
11046 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11048 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
11051 2014-01-28 Richard Biener <rguenther@suse.de>
11054 2014-01-28 Richard Biener <rguenther@suse.de>
11056 PR rtl-optimization/45364
11057 PR rtl-optimization/59890
11058 * var-tracking.c (local_get_addr_clear_given_value): Handle
11059 already cleared slot.
11060 (val_reset): Handle not allocated local_get_addr_cache.
11061 (vt_find_locations): Use post-order on the inverted CFG.
11063 2014-01-28 Richard Biener <rguenther@suse.de>
11065 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
11067 2014-01-28 Richard Biener <rguenther@suse.de>
11069 PR rtl-optimization/45364
11070 PR rtl-optimization/59890
11071 * var-tracking.c (local_get_addr_clear_given_value): Handle
11072 already cleared slot.
11073 (val_reset): Handle not allocated local_get_addr_cache.
11074 (vt_find_locations): Use post-order on the inverted CFG.
11076 2014-01-28 Alan Modra <amodra@gmail.com>
11078 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
11079 * configure.ac <recursive call for build != host>: Define
11080 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
11081 and LD_FOR_BUILD too.
11082 * configure: Regenerate.
11084 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
11086 * config/i386/i386.c (get_builtin_code_for_version): Separate
11087 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
11088 Broadwell from Haswell.
11090 2014-01-27 Steve Ellcey <sellcey@mips.com>
11092 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
11093 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
11094 * config/mips/mips.c (mips_option_override): Change setting
11096 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
11097 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
11098 Change from Mask to Var.
11100 2014-01-27 Jeff Law <law@redhat.com>
11102 * ipa-inline.c (inline_small_functions): Fix typo.
11104 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
11106 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
11107 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
11108 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
11109 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
11110 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
11111 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
11112 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
11113 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
11114 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
11115 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
11116 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
11117 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
11118 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
11119 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
11120 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
11121 (_mm512_storeu_epi64): Ditto.
11122 (_mm512_cmpge_epi32_mask): Ditto.
11123 (_mm512_cmpge_epu32_mask): Ditto.
11124 (_mm512_cmpge_epi64_mask): Ditto.
11125 (_mm512_cmpge_epu64_mask): Ditto.
11126 (_mm512_cmple_epi32_mask): Ditto.
11127 (_mm512_cmple_epu32_mask): Ditto.
11128 (_mm512_cmple_epi64_mask): Ditto.
11129 (_mm512_cmple_epu64_mask): Ditto.
11130 (_mm512_cmplt_epi32_mask): Ditto.
11131 (_mm512_cmplt_epu32_mask): Ditto.
11132 (_mm512_cmplt_epi64_mask): Ditto.
11133 (_mm512_cmplt_epu64_mask): Ditto.
11134 (_mm512_cmpneq_epi32_mask): Ditto.
11135 (_mm512_cmpneq_epu32_mask): Ditto.
11136 (_mm512_cmpneq_epi64_mask): Ditto.
11137 (_mm512_cmpneq_epu64_mask): Ditto.
11138 (_mm512_expand_pd): Ditto.
11139 (_mm512_expand_ps): Ditto.
11140 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
11141 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
11142 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
11143 * config/i386/i386.c (ix86_builtins): Add
11144 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
11145 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
11146 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
11147 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
11148 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
11149 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
11150 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
11151 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
11152 IX86_BUILTIN_PMOVUSQW512_MEM.
11153 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
11154 __builtin_ia32_pmovsqd512mem_mask,
11155 __builtin_ia32_pmovqd512mem_mask,
11156 __builtin_ia32_pmovusqw512mem_mask,
11157 __builtin_ia32_pmovsqw512mem_mask,
11158 __builtin_ia32_pmovqw512mem_mask,
11159 __builtin_ia32_pmovusdw512mem_mask,
11160 __builtin_ia32_pmovsdw512mem_mask,
11161 __builtin_ia32_pmovdw512mem_mask,
11162 __builtin_ia32_pmovqb512mem_mask,
11163 __builtin_ia32_pmovusqb512mem_mask,
11164 __builtin_ia32_pmovsqb512mem_mask,
11165 __builtin_ia32_pmovusdb512mem_mask,
11166 __builtin_ia32_pmovsdb512mem_mask,
11167 __builtin_ia32_pmovdb512mem_mask.
11168 (bdesc_args): Add __builtin_ia32_expanddf512,
11169 __builtin_ia32_expandsf512.
11170 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
11171 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
11172 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
11173 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
11174 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
11175 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
11176 (avx512f_<code>v8div16qi2_mask_store): This.
11177 (avx512f_expand<mode>): New.
11179 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
11181 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
11183 (_mm512_mask_prefetch_i64gather_pd): Ditto.
11184 (_mm512_prefetch_i32scatter_pd): Ditto.
11185 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
11186 (_mm512_prefetch_i64scatter_pd): Ditto.
11187 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
11188 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
11189 (_mm512_mask_prefetch_i64gather_ps): Ditto.
11190 (_mm512_prefetch_i32scatter_ps): Ditto.
11191 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
11192 (_mm512_prefetch_i64scatter_ps): Ditto.
11193 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
11194 * config/i386/i386-builtin-types.def: Define
11195 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
11196 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
11197 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
11198 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
11199 IX86_BUILTIN_SCATTERPFQPD.
11200 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
11201 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
11202 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
11203 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
11204 __builtin_ia32_scatterpfqps.
11205 (ix86_expand_builtin): Expand new built-ins.
11206 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
11207 fix memory access data type.
11208 (*avx512pf_gatherpf<mode>_mask): Ditto.
11209 (*avx512pf_gatherpf<mode>): Ditto.
11210 (avx512pf_scatterpf<mode>): Ditto.
11211 (*avx512pf_scatterpf<mode>_mask): Ditto.
11212 (*avx512pf_scatterpf<mode>): Ditto.
11213 (GATHER_SCATTER_SF_MEM_MODE): New.
11214 (avx512pf_gatherpf<mode>df): Ditto.
11215 (*avx512pf_gatherpf<mode>df_mask): Ditto.
11216 (*avx512pf_scatterpf<mode>df): Ditto.
11218 2014-01-27 Jakub Jelinek <jakub@redhat.com>
11221 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
11222 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
11225 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
11227 * common/config/arm/arm-common.c
11228 (arm_rewrite_mcpu): Handle multiple names.
11230 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
11232 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
11234 * gimple-builder.h (create_gimple_tmp): Delete.
11236 2014-01-27 Christian Bruel <christian.bruel@st.com>
11238 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
11241 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
11243 * config/pa/pa.md (call): Generate indirect long calls to non-local
11244 functions when outputing 32-bit code.
11245 (call_value): Likewise except for special call to buggy powf function.
11247 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
11248 portable runtime and PIC indirect calls.
11249 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
11250 and PIC call sequences. Use ldo instead of blr to set return register
11251 in PIC call sequence.
11253 2014-01-25 Walter Lee <walt@tilera.com>
11255 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
11256 avoid clobbering a live register.
11258 2014-01-25 Walter Lee <walt@tilera.com>
11260 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
11261 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
11262 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
11263 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
11265 2014-01-25 Walter Lee <walt@tilera.com>
11267 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
11268 arguments on even registers.
11269 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
11271 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
11272 (BIGGEST_ALIGNMENT): Ditto.
11273 (BIGGEST_FIELD_ALIGNMENT): Ditto.
11275 2014-01-25 Walter Lee <walt@tilera.com>
11277 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
11278 insns before bundling.
11279 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
11281 2014-01-25 Walter Lee <walt@tilera.com>
11283 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
11284 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
11285 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
11287 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
11289 * config/mips/constraints.md (kl): Delete.
11290 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
11291 define expands, using...
11292 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
11293 instructions for MIPS16.
11294 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
11295 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
11297 2014-01-25 Walter Lee <walt@tilera.com>
11299 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
11303 2014-01-25 Walter Lee <walt@tilera.com>
11305 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
11306 (TARGET_EXPAND_TO_RTL_HOOK): Define.
11308 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
11310 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
11313 2014-01-25 Jakub Jelinek <jakub@redhat.com>
11315 * print-rtl.c (in_call_function_usage): New var.
11316 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
11317 EXPR_LIST mode as mode and not as reg note name.
11319 PR middle-end/59561
11320 * cfgloopmanip.c (copy_loop_info): If
11321 loop->warned_aggressive_loop_optimizations, make sure
11322 the flag is set in target loop too.
11324 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
11326 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
11328 * builtins.def: Likewise.
11329 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
11330 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
11331 * ira.c (ira_setup_eliminable_regset): Likewise.
11332 * omp-low.c (gate_expand_omp): Likewise.
11333 (execute_lower_omp): Likewise.
11334 (diagnose_sb_0): Likewise.
11335 (gate_diagnose_omp_blocks): Likewise.
11336 (simd_clone_clauses_extract): Likewise.
11339 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11341 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
11342 correction for little endian...
11343 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
11346 2014-01-24 Jeff Law <law@redhat.com>
11348 PR tree-optimization/59919
11349 * tree-vrp.c (find_assert_locations_1): Do not register asserts
11350 for non-returning calls.
11352 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
11354 * common/config/aarch64/aarch64-common.c
11355 (aarch64_rewrite_mcpu): Handle multiple names.
11356 * config/aarch64/aarch64.h
11357 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
11359 2014-01-24 Dodji Seketeli <dodji@redhat.com>
11361 * input.c (add_file_to_cache_tab): Handle the case where fopen
11364 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
11367 * config/i386/i386.md (pushsf splitter): Get stack adjustment
11368 from push operand if code of push isn't PRE_DEC.
11370 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11373 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
11374 -mquad-memory-atomic. Update -mquad-memory documentation to say
11375 it is only used for non-atomic loads/stores.
11377 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
11378 -mquad-memory or -mquad-memory-atomic switches.
11380 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
11381 -mquad-memory-atomic to ISA 2.07 support.
11383 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
11384 to separate support of normal quad word memory operations (ldq, stq)
11385 from the atomic quad word memory operations.
11387 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
11388 support to separate non-atomic quad word operations from atomic
11389 quad word operations. Disable non-atomic quad word operations in
11390 little endian mode so that we don't have to swap words after the
11391 load and before the store.
11392 (quad_load_store_p): Add comment about atomic quad word support.
11393 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
11394 options printed with -mdebug=reg.
11396 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
11397 -mquad-memory-atomic as the test for whether we have quad word
11398 atomic instructions.
11399 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
11400 or -mp8-vector are used, allow byte/half-word atomic operations.
11402 * config/rs6000/sync.md (load_lockedti): Insure that the address
11403 is a proper indexed or indirect address for the lqarx instruction.
11404 On little endian systems, swap the hi/lo registers after the lqarx
11406 (load_lockedpti): Use indexed_or_indirect_operand predicate to
11407 insure the address is valid for the lqarx instruction.
11408 (store_conditionalti): Insure that the address is a proper indexed
11409 or indirect address for the stqcrx. instruction. On little endian
11410 systems, swap the hi/lo registers before doing the stqcrx.
11412 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
11413 insure the address is valid for the stqcrx. instruction.
11415 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
11416 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
11417 type of quad memory support is available.
11419 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
11421 PR regression/59915
11422 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
11423 there is a danger of looping.
11425 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
11427 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
11428 force flag_ira_loop_pressure if set via command line.
11430 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
11432 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
11433 (ashr_simd): New builtin handling DI mode.
11434 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
11435 (aarch64_sshr_simddi): New match pattern.
11436 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
11437 (vshrd_n_s64): Likewise.
11438 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
11440 2014-01-23 Nick Clifton <nickc@redhat.com>
11442 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
11443 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
11444 favour of mcu specific scripts.
11445 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
11448 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
11449 Alex Velenko <Alex.Velenko@arm.com>
11451 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
11452 (vaddv_s16): Likewise.
11453 (vaddv_s32): Likewise.
11454 (vaddv_u8): Likewise.
11455 (vaddv_u16): Likewise.
11456 (vaddv_u32): Likewise.
11457 (vaddvq_s8): Likewise.
11458 (vaddvq_s16): Likewise.
11459 (vaddvq_s32): Likewise.
11460 (vaddvq_s64): Likewise.
11461 (vaddvq_u8): Likewise.
11462 (vaddvq_u16): Likewise.
11463 (vaddvq_u32): Likewise.
11464 (vaddvq_u64): Likewise.
11465 (vaddv_f32): Likewise.
11466 (vaddvq_f32): Likewise.
11467 (vaddvq_f64): Likewise.
11468 (vmaxv_f32): Likewise.
11469 (vmaxv_s8): Likewise.
11470 (vmaxv_s16): Likewise.
11471 (vmaxv_s32): Likewise.
11472 (vmaxv_u8): Likewise.
11473 (vmaxv_u16): Likewise.
11474 (vmaxv_u32): Likewise.
11475 (vmaxvq_f32): Likewise.
11476 (vmaxvq_f64): Likewise.
11477 (vmaxvq_s8): Likewise.
11478 (vmaxvq_s16): Likewise.
11479 (vmaxvq_s32): Likewise.
11480 (vmaxvq_u8): Likewise.
11481 (vmaxvq_u16): Likewise.
11482 (vmaxvq_u32): Likewise.
11483 (vmaxnmv_f32): Likewise.
11484 (vmaxnmvq_f32): Likewise.
11485 (vmaxnmvq_f64): Likewise.
11486 (vminv_f32): Likewise.
11487 (vminv_s8): Likewise.
11488 (vminv_s16): Likewise.
11489 (vminv_s32): Likewise.
11490 (vminv_u8): Likewise.
11491 (vminv_u16): Likewise.
11492 (vminv_u32): Likewise.
11493 (vminvq_f32): Likewise.
11494 (vminvq_f64): Likewise.
11495 (vminvq_s8): Likewise.
11496 (vminvq_s16): Likewise.
11497 (vminvq_s32): Likewise.
11498 (vminvq_u8): Likewise.
11499 (vminvq_u16): Likewise.
11500 (vminvq_u32): Likewise.
11501 (vminnmv_f32): Likewise.
11502 (vminnmvq_f32): Likewise.
11503 (vminnmvq_f64): Likewise.
11505 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
11507 * config/aarch64/aarch64-simd.md
11508 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
11509 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
11510 (*aarch64_mul3_elt<mode>): Likewise.
11511 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
11512 (*aarch64_mul3_elt_to_64v2df): Likewise.
11513 (*aarch64_mla_elt<mode>): Likewise.
11514 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
11515 (*aarch64_mls_elt<mode>): Likewise.
11516 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
11517 (*aarch64_fma4_elt<mode>): Likewise.
11518 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
11519 (*aarch64_fma4_elt_to_64v2df): Likewise.
11520 (*aarch64_fnma4_elt<mode>): Likewise.
11521 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
11522 (*aarch64_fnma4_elt_to_64v2df): Likewise.
11523 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
11524 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
11525 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
11526 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
11527 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
11528 (aarch64_sqdmull_lane<mode>_internal): Likewise.
11529 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
11531 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
11533 * config/aarch64/aarch64-simd.md
11534 (aarch64_be_checked_get_lane<mode>): New define_expand.
11535 * config/aarch64/aarch64-simd-builtins.def
11536 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
11537 New builtin definition.
11538 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
11539 Use new safe be builtin.
11541 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
11543 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
11545 (aarch64_be_st1<mode>): Likewise.
11546 (aarch_ld1<VALL:mode>): Define_expand modified.
11547 (aarch_st1<VALL:mode>): Likewise.
11548 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
11549 (UNSPEC_ST1): Likewise.
11551 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
11553 * config/microblaze/microblaze.md: Add trap insn and attribute
11555 2014-01-23 Dodji Seketeli <dodji@redhat.com>
11557 PR preprocessor/58580
11558 * input.h (location_get_source_line): Take an additional line_size
11560 (void diagnostics_file_cache_fini): Declare new function.
11561 * input.c (struct fcache): New type.
11562 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
11563 New static constants.
11564 (diagnostic_file_cache_init, total_lines_num)
11565 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
11566 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
11567 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
11568 (get_next_line, read_next_line, goto_next_line, read_line_num):
11569 New static function definitions.
11570 (diagnostic_file_cache_fini): New function.
11571 (location_get_source_line): Take an additional output line_len
11572 parameter. Re-write using lookup_or_add_file_to_cache_tab and
11574 * diagnostic.c (diagnostic_finish): Call
11575 diagnostic_file_cache_fini.
11576 (adjust_line): Take an additional input parameter for the length
11577 of the line, rather than calculating it with strlen.
11578 (diagnostic_show_locus): Adjust the use of
11579 location_get_source_line and adjust_line with respect to their new
11580 signature. While displaying a line now, do not stop at the first
11581 null byte. Rather, display the zero byte as a space and keep
11582 going until we reach the size of the line.
11583 * Makefile.in: Add vec.o to OBJS-libcommon
11585 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
11586 Ilya Tocar <ilya.tocar@intel.com>
11588 * config/i386/avx512fintrin.h (_mm512_kmov): New.
11589 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
11590 (__builtin_ia32_kmov16): Ditto.
11591 * config/i386/i386.md (UNSPEC_KMOV): New.
11594 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
11596 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
11597 (_mm512_storeu_si512): Ditto.
11599 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
11602 * rtl.h (get_referenced_operands): Declare.
11603 * recog.c (get_referenced_operands): New function.
11604 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
11605 operands have been referenced when recording LO_SUM references.
11607 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
11609 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
11611 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
11613 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
11614 Enable for generic and recent AMD targets.
11616 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
11618 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
11619 ARG_SIZE note when adjustment was eliminated.
11621 2014-01-22 Jeff Law <law@redhat.com>
11623 PR tree-optimization/59597
11624 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
11625 in file. Accept new argument REGISTERING and use it to modify
11626 dump output appropriately.
11627 (register_jump_thread): Corresponding changes.
11628 (mark_threaded_blocks): Reinstate code to cancel unprofitable
11629 thread paths involving joiner blocks. Add code to dump cancelled
11630 jump threading paths.
11632 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
11634 PR rtl-optimization/59477
11635 * lra-constraints.c (inherit_in_ebb): Process call for living hard
11636 regs. Update reloads_num and potential_reload_hard_regs for all insns.
11638 2014-01-22 Tom Tromey <tromey@redhat.com>
11640 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
11642 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
11644 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
11646 PR rtl-optimization/59896
11647 * lra-constraints.c (process_alt_operands): Check unused note for
11648 matched operands of insn with no output reloads.
11650 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
11652 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
11653 (mips_move_from_gpr_cost): Likewise.
11655 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
11657 PR rtl-optimization/59858
11658 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
11659 ira_class_hard_regs_num.
11660 (process_alt_operands): Increase reject for dying matched operand.
11662 2014-01-21 Jakub Jelinek <jakub@redhat.com>
11665 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
11666 smaller than size, perform several stores or loads and stores
11667 at dst + count - size to store or copy all of size bytes, rather
11668 than just last modesize bytes.
11670 2014-01-20 DJ Delorie <dj@redhat.com>
11672 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
11673 that CLOBBERs are REGs before propogating their values.
11675 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
11677 PR middle-end/59789
11678 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
11679 (cgraph_inline_failed_type): New function.
11680 * cgraph.h (DEFCIFCODE): Add type.
11681 (cgraph_inline_failed_type_t): New enum.
11682 (cgraph_inline_failed_type): New prototype.
11683 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
11684 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
11685 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
11686 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
11687 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
11688 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
11689 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
11690 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
11691 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
11692 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
11693 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
11694 OPTIMIZATION_MISMATCH.
11695 * tree-inline.c (expand_call_inline): Emit errors during
11696 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
11698 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
11701 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
11702 mode attribute in insn output.
11704 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
11706 * output.h (output_constant): Delete.
11707 * varasm.c (output_constant): Make private.
11709 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
11711 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
11713 2014-01-20 Jakub Jelinek <jakub@redhat.com>
11715 PR middle-end/59860
11716 * tree.h (fold_builtin_strcat): New prototype.
11717 * builtins.c (fold_builtin_strcat): No longer static. Add len
11718 argument, if non-NULL, don't call c_strlen. Optimize
11719 directly into __builtin_memcpy instead of __builtin_strcpy.
11720 (fold_builtin_2): Adjust fold_builtin_strcat caller.
11721 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
11723 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
11725 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
11726 for SImode_address_operand operands, having only a REG argument.
11728 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
11730 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
11731 loader name using mbig-endian.
11732 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
11734 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
11736 * doc/invoke.texi (-march): Clarify documentation for AArch64.
11737 (-mtune): Likewise.
11740 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
11742 * config/aarch64/aarch64-protos.h
11743 (aarch64_cannot_change_mode_class_ptr): Declare.
11744 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
11745 aarch64_cannot_change_mode_class_ptr): New.
11746 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
11747 backend hook aarch64_cannot_change_mode_class.
11749 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
11751 * common/config/aarch64/aarch64-common.c
11752 (aarch64_handle_option): Don't handle any option order logic here.
11753 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
11754 selected_cpu, warn on architecture version mismatch.
11755 (aarch64_override_options): Fix parsing order for option strings.
11757 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11758 Iain Sandoe <iain@codesourcery.com>
11761 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
11762 warning. Amend comment to reflect current functionality.
11764 2014-01-20 Richard Biener <rguenther@suse.de>
11766 PR middle-end/59860
11767 * builtins.c (fold_builtin_strcat): Remove case better handled
11768 by tree-ssa-strlen.c.
11770 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
11772 * config/aarch64/aarch64.opt
11773 (mcpu, march, mtune): Make case-insensitive.
11775 2014-01-20 Jakub Jelinek <jakub@redhat.com>
11778 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
11779 if operands[1] is a REG or ZERO_EXTEND of a REG.
11781 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
11783 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
11785 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
11787 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
11788 long non-pic millicode calls.
11790 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11792 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
11794 2014-01-19 Kito Cheng <kito@0xlab.org>
11796 * builtins.c (expand_movstr): Check movstr expand done or fail.
11798 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11799 H.J. Lu <hongjiu.lu@intel.com>
11802 * config/i386/i386.md (*lea<mode>): Zero-extend return register
11803 to DImode for zero-extended addresses.
11805 2014-01-19 Jakub Jelinek <jakub@redhat.com>
11807 PR rtl-optimization/57763
11808 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
11809 on the new indirect jump_insn and increment LABEL_NUSES (label).
11811 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
11815 * config.gcc (x86_archs): New variable.
11816 (x86_64_archs): Likewise.
11817 (x86_cpus): Likewise.
11818 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
11819 --with-arch/--with-cpu= options.
11820 Support --with-arch=/--with-cpu={nehalem,westmere,
11821 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
11823 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11825 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
11826 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
11828 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11830 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
11832 2014-01-18 Jakub Jelinek <jakub@redhat.com>
11835 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
11836 clear cpp_get_options (parse_in)->warn_unused_macros for
11837 ix86_target_macros_internal with cpp_define.
11839 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
11841 * jump.c (delete_related_insns): Keep (use (insn))s.
11842 * reorg.c (redundant_insn): Check for barriers too.
11844 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11846 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
11848 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
11850 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
11851 call to $$dyncall when TARGET_LONG_CALLS is true.
11853 2014-01-17 Jeff Law <law@redhat.com>
11855 * ree.c (combine_set_extension): Temporarily disable test for
11856 changing number of hard registers.
11858 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
11860 PR middle-end/58125
11861 * ipa-inline-analysis.c (inline_free_summary):
11862 Do not free summary of aliases.
11864 2014-01-17 Jakub Jelinek <jakub@redhat.com>
11866 PR middle-end/59706
11867 * gimplify.c (gimplify_expr): Use create_tmp_var
11868 instead of create_tmp_var_raw. If cond doesn't have
11869 integral type, don't add the IFN_ANNOTATE builtin at all.
11871 2014-01-17 Martin Jambor <mjambor@suse.cz>
11874 * ipa-cp.c (prev_edge_clone): New variable.
11875 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
11876 Also resize prev_edge_clone vector.
11877 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
11878 (ipcp_edge_removal_hook): New function.
11879 (ipcp_driver): Register ipcp_edge_removal_hook.
11881 2014-01-17 Andrew Pinski <apinski@cavium.com>
11882 Steve Ellcey <sellcey@mips.com>
11885 * config/mips/mips.c (mips_print_operand): Check operand mode instead
11888 2014-01-17 Jeff Law <law@redhat.com>
11890 PR middle-end/57904
11891 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
11892 so that pass_ccp runs first.
11894 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11896 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
11897 (ix86_adjust_cost): Use !TARGET_XXX.
11898 (do_reorder_for_imul): Likewise.
11899 (swap_top_of_ready_list): Likewise.
11900 (ix86_sched_reorder): Likewise.
11902 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11904 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
11905 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
11906 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
11907 (intel_memset): New. Duplicate slm_memset.
11908 (intel_cost): New. Duplicate slm_cost.
11909 (m_INTEL): New macro.
11910 (processor_target_table): Add "intel".
11911 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
11912 with PROCESSOR_INTEL for "intel".
11913 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
11914 PROCESSOR_SILVERMONT.
11915 (ix86_issue_rate): Likewise.
11916 (ix86_adjust_cost): Likewise.
11917 (ia32_multipass_dfa_lookahead): Likewise.
11918 (swap_top_of_ready_list): Likewise.
11919 (ix86_sched_reorder): Likewise.
11920 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
11921 instead of TARGET_OPT_AGU.
11922 * config/i386/i386.h (TARGET_INTEL): New.
11923 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
11924 (processor_type): Add PROCESSOR_INTEL.
11925 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
11926 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
11928 2014-01-17 Marek Polacek <polacek@redhat.com>
11931 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
11934 2014-01-17 Richard Biener <rguenther@suse.de>
11936 PR tree-optimization/46590
11937 * opts.c (default_options_table): Add entries for
11938 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
11939 all enabled at -O1 but not for -Og.
11940 * common.opt (fbranch-count-reg): Remove Init(1).
11941 (fmove-loop-invariants): Likewise.
11942 (ftree-pta): Likewise.
11944 2014-01-17 Jakub Jelinek <jakub@redhat.com>
11946 * config/i386/i386.c (ix86_data_alignment): For compatibility with
11947 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
11948 decls to at least the GCC 4.8 used alignments.
11951 * tree-nested.c (convert_nonlocal_reference_stmt,
11952 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
11953 of GIMPLE_BIND stmts, adjust associated decls.
11955 2014-01-17 Richard Biener <rguenther@suse.de>
11957 PR tree-optimization/46590
11958 * vec.h (vec<>::bseach): New member function implementing
11959 binary search according to C89 bsearch.
11960 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
11961 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
11962 bitmap pointer again. Make accesses_in_loop a flat array.
11963 (mem_ref_obstack): New global.
11964 (outermost_indep_loop): Adjust for mem_ref->stored changes.
11965 (mark_ref_stored): Likewise.
11966 (ref_indep_loop_p_2): Likewise.
11967 (set_ref_stored_in_loop): New helper function.
11968 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
11969 (memref_free): Adjust.
11970 (record_mem_ref_loc): Simplify.
11971 (gather_mem_refs_stmt): Adjust.
11972 (sort_locs_in_loop_postorder_cmp): New function.
11973 (analyze_memory_references): Sort accesses_in_loop after
11974 loop postorder number.
11975 (find_ref_loc_in_loop_cmp): New function.
11976 (for_all_locs_in_loop): Find relevant cluster of locs in
11977 accesses_in_loop and iterate without recursion.
11978 (execute_sm): Avoid uninit warning.
11979 (struct ref_always_accessed): Simplify.
11980 (ref_always_accessed::operator ()): Likewise.
11981 (ref_always_accessed_p): Likewise.
11982 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
11983 loop postorder numbers here.
11984 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
11987 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
11990 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
11991 on decls for which assemble_alias has been called.
11993 2014-01-17 Nick Clifton <nickc@redhat.com>
11995 * config/msp430/msp430.opt: (mcpu): New option.
11996 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
11997 (msp430_option_override): Parse target_cpu. If the MCU name
11998 matches a generic string, clear target_mcu.
11999 (msp430_attr): Allow numeric interrupt values up to 63.
12000 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
12001 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
12003 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
12005 * config/msp430/msp430.md (popm): Use %J rather than %I.
12006 (addsi3): Use msp430_nonimmediate_operand for operand 2.
12007 (addhi_cy_i): Use immediate_operand for operand 2.
12008 * doc/invoke.texi: Document -mcpu option.
12010 2014-01-17 Richard Biener <rguenther@suse.de>
12012 PR rtl-optimization/38518
12013 * df.h (df_analyze_loop): Declare.
12014 * df-core.c: Include cfgloop.h.
12015 (df_analyze_1): Split out main part of df_analyze.
12016 (df_analyze): Adjust.
12017 (loop_inverted_post_order_compute): New function.
12018 (loop_post_order_compute): Likewise.
12019 (df_analyze_loop): New function avoiding whole-function
12020 postorder computes.
12021 * loop-invariant.c (find_defs): Use df_analyze_loop.
12022 (find_invariants): Adjust.
12023 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
12025 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
12027 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
12028 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
12030 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
12032 * ipa-ref.c (ipa_remove_stmt_references): Fix references
12033 traversal when removing references.
12035 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
12038 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
12040 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
12042 PR middle-end/56791
12043 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
12044 pushing a reload for an autoinc when we had previously reloaded an
12045 inner part of the address.
12047 2014-01-16 Jakub Jelinek <jakub@redhat.com>
12049 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
12051 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
12052 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
12053 when not giving up or versioning for alias only because of
12055 (vect_analyze_data_ref_dependences): Set to true.
12056 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
12058 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
12059 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
12062 PR middle-end/58344
12063 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
12066 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
12067 operand 0 predicate for gathers, use a new pseudo as subtarget.
12069 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
12071 PR middle-end/59609
12072 * lra-constraints.c (process_alt_operands): Add printing debug info.
12073 Check absence of input/output reloads for matched operands too.
12075 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
12077 PR rtl-optimization/59835
12078 * ira.c (ira_init_register_move_cost): Increase cost for
12081 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
12083 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
12085 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
12088 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
12089 non-register objects. Use gen_(high/low)part more consistently.
12092 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
12095 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
12096 endian support, remove tests for WORDS_BIG_ENDIAN.
12097 (p8_mfvsrd_3_<mode>): Likewise.
12098 (reload_gpr_from_vsx<mode>): Likewise.
12099 (reload_gpr_from_vsxsf): Likewise.
12100 (p8_mfvsrd_4_disf): Likewise.
12102 2014-01-16 Richard Biener <rguenther@suse.de>
12104 PR rtl-optimization/46590
12105 * lcm.c (compute_antinout_edge): Use postorder iteration.
12106 (compute_laterin): Use inverted postorder iteration.
12108 2014-01-16 Nick Clifton <nickc@redhat.com>
12110 PR middle-end/28865
12111 * varasm.c (output_constant): Return the number of bytes actually
12113 (output_constructor_array_range): Update the field size with the
12114 number of bytes emitted by output_constant.
12115 (output_constructor_regular_field): Likewise. Also do not
12116 complain if the total number of bytes emitted is now greater
12117 than the expected fieldpos.
12118 * output.h (output_constant): Update prototype and descriptive comment.
12120 2014-01-16 Marek Polacek <polacek@redhat.com>
12122 PR middle-end/59827
12123 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
12124 it is error_mark_node.
12126 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
12128 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
12129 VALID_AVX256_REG_OR_OI_MODE.
12131 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
12133 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
12134 current procedure should be profiled.
12136 2014-01-15 Andrew Pinski <apinski@cavium.com>
12138 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
12139 of moving from/to the STACK_REG register class.
12141 2014-01-15 Richard Henderson <rth@redhat.com>
12144 * reginfo.c (global_regs_decl): Globalize.
12145 * rtl.h (global_regs_decl): Declare.
12146 * ira.c (do_reload): Diagnose frame_pointer_needed and it
12147 reserved via global_regs.
12149 2014-01-15 Teresa Johnson <tejohnson@google.com>
12151 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
12153 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12155 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
12156 and vmulosh rather than call gen_vec_widen_smult_*.
12157 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
12158 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
12159 (vec_widen_smult_even_v16qi): Likewise.
12160 (vec_widen_umult_even_v8hi): Likewise.
12161 (vec_widen_smult_even_v8hi): Likewise.
12162 (vec_widen_umult_odd_v16qi): Likewise.
12163 (vec_widen_smult_odd_v16qi): Likewise.
12164 (vec_widen_umult_odd_v8hi): Likewise.
12165 (vec_widen_smult_odd_v8hi): Likewise.
12166 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
12167 vmuloub rather than call gen_vec_widen_umult_*.
12168 (vec_widen_umult_lo_v16qi): Likewise.
12169 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
12170 vmulosb rather than call gen_vec_widen_smult_*.
12171 (vec_widen_smult_lo_v16qi): Likewise.
12172 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
12173 rather than call gen_vec_widen_umult_*.
12174 (vec_widen_umult_lo_v8hi): Likewise.
12175 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
12176 rather than call gen_vec_widen_smult_*.
12177 (vec_widen_smult_lo_v8hi): Likewise.
12179 2014-01-15 Jeff Law <law@redhat.com>
12181 PR tree-optimization/59747
12182 * ree.c (find_and_remove_re): Properly handle case where a second
12183 eliminated extension requires widening a copy created for elimination
12184 of a prior extension.
12185 (combine_set_extension): Ensure that the number of hard regs needed
12186 for a destination register does not change when we widen it.
12188 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
12190 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
12191 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
12192 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
12193 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
12194 (avr-*-rtems*): Likewise.
12195 (bfin*-rtems*): Likewise.
12196 (moxie-*-rtems*): Likewise.
12197 (h8300-*-rtems*): Likewise.
12198 (i[34567]86-*-rtems*): Likewise.
12199 (lm32-*-rtems*): Likewise.
12200 (m32r-*-rtems*): Likewise.
12201 (m68k-*-rtems*): Likewise.
12202 (microblaze*-*-rtems*): Likewise.
12203 (mips*-*-rtems*): Likewise.
12204 (powerpc-*-rtems*): Likewise.
12205 (sh-*-rtems*): Likewise.
12206 (sparc-*-rtems*): Likewise.
12207 (sparc64-*-rtems*): Likewise.
12208 (v850-*-rtems*): Likewise.
12209 (m32c-*-rtems*): Likewise.
12211 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
12213 PR rtl-optimization/59511
12214 * ira.c (ira_init_register_move_cost): Use memory costs for some
12215 cases of register move cost calculations.
12216 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
12217 instead of BB frequency.
12218 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
12219 * lra-assigns.c (find_hard_regno_for): Ditto.
12221 2014-01-15 Richard Biener <rguenther@suse.de>
12223 PR tree-optimization/59822
12224 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
12225 (vectorizable_load): Use it to hoist defs of uses of invariant
12226 loads out of the loop.
12228 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
12229 Kugan Vivekanandarajah <kuganv@linaro.org>
12232 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
12235 2014-01-15 Richard Biener <rguenther@suse.de>
12237 PR rtl-optimization/59802
12238 * lcm.c (compute_available): Use inverted postorder to seed
12239 the initial worklist.
12241 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12244 * config/s390/s390.c (s390_preferred_reload_class): Don't return
12245 ADDR_REGS for invalid symrefs in non-PIC code.
12247 2014-01-15 Jakub Jelinek <jakub@redhat.com>
12250 * builtins.c (determine_block_size): Initialize *probable_max_size
12251 even if len_rtx is CONST_INT.
12253 2014-01-14 Andrew Pinski <apinski@cavium.com>
12255 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
12256 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
12257 (cortexa53_tunings): Likewise.
12258 (aarch64_sched_issue_rate): New function.
12259 (TARGET_SCHED_ISSUE_RATE): Define.
12261 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
12263 * ira-costs.c (find_costs_and_classes): Add missed
12264 ira_init_register_move_cost_if_necessary.
12266 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
12269 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
12271 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
12274 * config/i386/i386.c (type_natural_mode): Add a bool parameter
12275 to indicate if type is used for function return value. Warn ABI
12276 change if the vector mode isn't available for function return value.
12277 (ix86_function_arg_advance): Pass false to type_natural_mode.
12278 (ix86_function_arg): Likewise.
12279 (ix86_gimplify_va_arg): Likewise.
12280 (function_arg_32): Don't warn ABI change.
12281 (ix86_function_value): Pass true to type_natural_mode.
12282 (ix86_return_in_memory): Likewise.
12283 (ix86_struct_value_rtx): Removed.
12284 (TARGET_STRUCT_VALUE_RTX): Likewise.
12286 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
12288 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
12289 converting a conditional jump into a conditional return.
12291 2014-01-14 Richard Biener <rguenther@suse.de>
12293 PR tree-optimization/58921
12294 PR tree-optimization/59006
12295 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
12296 hoisting invariant stmts.
12297 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
12298 invariant loads on the preheader edge if possible.
12300 2014-01-14 Joey Ye <joey.ye@arm.com>
12302 * doc/plugin.texi (Building GCC plugins): Update to C++.
12304 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
12306 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
12307 (_mm_rcp28_round_ss): Ditto.
12308 (_mm_rsqrt28_round_sd): Ditto.
12309 (_mm_rsqrt28_round_ss): Ditto.
12310 (_mm_rcp28_sd): Ditto.
12311 (_mm_rcp28_ss): Ditto.
12312 (_mm_rsqrt28_sd): Ditto.
12313 (_mm_rsqrt28_ss): Ditto.
12314 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
12315 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
12316 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
12317 (IX86_BUILTIN_RCP28SD): Ditto.
12318 (IX86_BUILTIN_RCP28SS): Ditto.
12319 (IX86_BUILTIN_RSQRT28SD): Ditto.
12320 (IX86_BUILTIN_RSQRT28SS): Ditto.
12321 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
12322 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
12323 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
12324 (ix86_expand_special_args_builtin): Expand new FTYPE.
12325 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
12326 (srcp14<mode>): Make insn unary.
12327 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
12328 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
12329 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
12330 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
12331 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
12332 Fix rounding: make it SAE only.
12333 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
12335 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
12337 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
12338 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
12339 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
12340 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
12341 (round_saeonly_mask_scalar_operand4): Ditto.
12342 (round_saeonly_mask_scalar_op3): Ditto.
12343 (round_saeonly_mask_scalar_op4): Ditto.
12345 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12347 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
12348 Implement -maltivec=be for vec_insert and vec_extract.
12350 2014-01-10 DJ Delorie <dj@redhat.com>
12352 * config/msp430/msp430.md (call_internal): Don't allow memory
12353 references with SP as the base register.
12354 (call_value_internal): Likewise.
12355 * config/msp430/constraints.md (Yc): New. For memory references
12356 that don't use SP as a base register.
12358 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
12359 "an integer without a # prefix"
12360 * config/msp430/msp430.md (epilogue_helper): Use it.
12362 2014-01-13 Jakub Jelinek <jakub@redhat.com>
12365 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
12366 AVX512F gather builtins.
12367 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
12368 on gather decls with INTEGER_TYPE masktype.
12369 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
12370 directly into the builtin rather than hoisting it before loop.
12372 PR tree-optimization/59387
12373 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
12374 (scev_const_prop): If folded_casts and type has undefined overflow,
12375 use force_gimple_operand instead of force_gimple_operand_gsi and
12376 for each added stmt if it is assign with
12377 arith_code_with_undefined_signed_overflow, call
12378 rewrite_to_defined_overflow.
12379 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
12380 gimple-fold.h instead.
12381 (arith_code_with_undefined_signed_overflow,
12382 rewrite_to_defined_overflow): Moved to ...
12383 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
12384 rewrite_to_defined_overflow): ... here. No longer static.
12385 Include gimplify-me.h.
12386 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
12387 rewrite_to_defined_overflow): New prototypes.
12389 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12391 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
12393 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
12395 * builtins.c (get_object_alignment_2): Minor tweak.
12396 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
12398 2014-01-13 Christian Bruel <christian.bruel@st.com>
12400 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
12401 optimized non constant lengths.
12403 2014-01-13 Jakub Jelinek <jakub@redhat.com>
12406 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
12407 load as __atomic_load_N if possible.
12409 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
12411 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
12413 (rs6000_expand_builtin): Adjust call.
12415 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
12418 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
12419 * config/rs6000/rs6000.c: Include target-globals.h.
12420 (rs6000_set_current_function): Instead of doing target_reinit
12421 unconditionally, use save_target_globals_default_opts and
12422 restore_target_globals.
12424 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
12426 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
12427 (rs6000_expand_builtin): Handle mffs and mtfsf.
12428 (rs6000_init_builtins): Define mffs and mtfsf.
12429 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
12430 (rs6000_mffs): New pattern.
12431 (rs6000_mtfsf): New pattern.
12433 2014-01-11 Bin Cheng <bin.cheng@arm.com>
12435 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
12436 Start narrowing with START. Apply candidate-use pair
12437 and check overall cost in narrowing.
12438 (iv_ca_prune): Pass new argument.
12440 2014-01-10 Jeff Law <law@redhat.com>
12442 PR middle-end/59743
12443 * ree.c (combine_reaching_defs): Ensure the defining statement
12444 occurs before the extension when optimizing extensions with
12445 different source and destination hard registers.
12447 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
12450 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
12451 vtables into the type inheritance graph.
12453 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12455 PR rtl-optimization/59754
12456 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
12457 modes in the REGNO != REGNO case.
12459 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12461 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
12463 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12465 PR tree-optimization/59745
12466 * tree-predcom.c (tree_predictive_commoning_loop): Call
12467 free_affine_expand_cache if giving up because components is NULL.
12469 * target-globals.c (save_target_globals): Allocate < 4KB structs using
12470 GC in payload of target_globals struct instead of allocating them on
12471 the heap and the larger structs separately using GC.
12472 * target-globals.h (struct target_globals): Make regs, hard_regs,
12473 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
12474 of GTY((skip)) and change type to void *.
12475 (reset_target_globals): Cast loads from those fields to corresponding
12478 2014-01-10 Steve Ellcey <sellcey@mips.com>
12481 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
12482 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
12483 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
12485 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
12488 * aarch64-modes.def (CC_Zmode): New flags mode.
12489 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
12490 represents an equality.
12491 (aarch64_get_condition_code): Handle CC_Zmode.
12492 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
12494 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12496 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
12497 extraction in good case.
12499 2014-01-10 Richard Biener <rguenther@suse.de>
12501 PR tree-optimization/59374
12502 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
12503 checking after SLP discovery. Mark stmts not participating
12504 in any SLP instance properly.
12506 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12508 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
12509 when handling a SET rtx.
12511 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12513 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
12514 (cortex-a57): Likewise.
12515 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
12517 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12519 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
12520 non-iwmmxt builtins.
12522 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
12526 * ipa-devirt.c record_target_from_binfo): Take as argument
12527 stack of binfos and lookup matching one for virtual inheritance.
12528 (possible_polymorphic_call_targets_1): Update.
12530 2014-01-10 Huacai Chen <chenhc@lemote.com>
12532 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
12533 kernel strings for Loongson-2E/2F/3A.
12535 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12537 PR middle-end/59670
12538 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
12539 is_gimple_call before calling gimple_call_internal_p.
12541 2014-01-09 Steve Ellcey <sellcey@mips.com>
12543 * Makefile.in (TREE_FLOW_H): Remove.
12544 (TREE_SSA_H): Add file names from tree-flow.h.
12545 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
12546 * tree.h: Remove tree-flow.h reference.
12547 * hash-table.h: Remove tree-flow.h reference.
12548 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
12549 reference with tree-ssa-loop.h.
12551 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12553 * doc/invoke.texi: Add -maltivec={be,le} options, and document
12554 default element-order behavior for -maltivec.
12555 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
12556 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
12557 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
12558 when targeting big endian, at least for now.
12559 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
12561 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12563 PR middle-end/47735
12564 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
12565 var satisfies use_register_for_decl, just take into account type
12566 alignment, rather than decl alignment.
12568 PR tree-optimization/59622
12569 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
12570 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
12571 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
12572 Don't devirtualize for inplace at all. For targets.length () == 1,
12573 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
12575 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
12577 * config/i386/i386.md (cpu): Remove the unused btver1.
12579 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
12581 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
12583 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12586 * tree-core.h (struct target_globals): New forward declaration.
12587 (struct tree_target_option): Add globals field.
12588 * tree.h (TREE_TARGET_GLOBALS): Define.
12589 (prepare_target_option_nodes_for_pch): New prototype.
12590 * target-globals.h (struct target_globals): Define even if
12591 !SWITCHABLE_TARGET.
12592 * tree.c (prepare_target_option_node_for_pch,
12593 prepare_target_option_nodes_for_pch): New functions.
12594 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
12595 * config/i386/i386.c: Include target-globals.h.
12596 (ix86_set_current_function): Instead of doing target_reinit
12597 unconditionally, use save_target_globals_default_opts and
12598 restore_target_globals.
12600 2014-01-09 Richard Biener <rguenther@suse.de>
12602 PR tree-optimization/59715
12603 * tree-cfg.h (split_critical_edges): Declare.
12604 * tree-cfg.c (split_critical_edges): Export.
12605 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
12607 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
12609 * cfgexpand.c (expand_stack_vars): Optionally disable
12610 asan stack protection.
12611 (expand_used_vars): Likewise.
12612 (partition_stack_vars): Likewise.
12613 * asan.c (asan_emit_stack_protection): Optionally disable
12614 after return stack usage.
12615 (instrument_derefs): Optionally disable memory access instrumentation.
12616 (instrument_builtin_call): Likewise.
12617 (instrument_strlen_call): Likewise.
12618 (asan_protect_global): Optionally disable global variables protection.
12619 * doc/invoke.texi: Added doc for new options.
12620 * params.def: Added new options.
12621 * params.h: Likewise.
12623 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12625 PR rtl-optimization/59724
12626 * ifcvt.c (cond_exec_process_if_block): Don't call
12627 flow_find_head_matching_sequence with 0 longest_match.
12628 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
12629 non-active insns if !stop_after.
12630 (try_head_merge_bb): Revert 2014-01-07 changes.
12632 2014-01-08 Jeff Law <law@redhat.com>
12634 * ree.c (get_sub_rtx): New function, extracted from...
12635 (merge_def_and_ext): Here.
12636 (combine_reaching_defs): Use get_sub_rtx.
12638 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
12640 * cgraph.h (varpool_variable_node): Do not choke on null node.
12642 2014-01-08 Catherine Moore <clm@codesourcery.com>
12644 * config/mips/mips.md (simple_return): Attempt to use JRC
12646 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
12648 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
12650 PR rtl-optimization/59137
12651 * reorg.c (steal_delay_list_from_target): Call update_block for
12653 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
12655 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12657 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
12658 two duplicate entries.
12660 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
12663 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
12665 * config/mips/mips.c (mips_truncated_op_cost): New function.
12666 (mips_rtx_costs): Adjust test for BADDU.
12667 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
12669 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
12671 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
12672 (*baddu_si): ...this new pattern.
12674 2014-01-08 Jakub Jelinek <jakub@redhat.com>
12677 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
12679 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
12681 PR middle-end/57748
12682 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
12684 (expand_expr, expand_normal): Adjust.
12685 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
12686 inner_reference_p. Use inner_reference_p to expand inner references.
12687 (store_expr): Adjust.
12688 * cfgexpand.c (expand_call_stmt): Adjust.
12690 2014-01-08 Rong Xu <xur@google.com>
12692 * gcov-io.c (gcov_var): Move from gcov-io.h.
12693 (gcov_position): Ditto.
12694 (gcov_is_error): Ditto.
12695 (gcov_rewrite): Ditto.
12696 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
12697 only part to libgcc/libgcov.h.
12699 2014-01-08 Marek Polacek <polacek@redhat.com>
12701 PR middle-end/59669
12702 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
12704 2014-01-08 Marek Polacek <polacek@redhat.com>
12707 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
12709 2014-01-08 Jakub Jelinek <jakub@redhat.com>
12711 PR rtl-optimization/59649
12712 * stor-layout.c (get_mode_bounds): For BImode return
12713 0 and STORE_FLAG_VALUE.
12715 2014-01-08 Richard Biener <rguenther@suse.de>
12717 PR middle-end/59630
12718 * gimple.h (is_gimple_builtin_call): Remove.
12719 (gimple_builtin_call_types_compatible_p): New.
12720 (gimple_call_builtin_p): New overload.
12721 * gimple.c (is_gimple_builtin_call): Remove.
12722 (validate_call): Rename to ...
12723 (gimple_builtin_call_types_compatible_p): ... this and export. Also
12724 check return types.
12725 (validate_type): New static function.
12726 (gimple_call_builtin_p): New overload and adjust.
12727 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
12728 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
12729 (gimple_fold_stmt_to_constant_1): Likewise.
12730 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
12732 2014-01-08 Richard Biener <rguenther@suse.de>
12734 PR middle-end/59471
12735 * gimplify.c (gimplify_expr): Gimplify register-register type
12736 VIEW_CONVERT_EXPRs to separate stmts.
12738 2014-01-07 Jeff Law <law@redhat.com>
12740 PR middle-end/53623
12741 * ree.c (combine_set_extension): Handle case where source
12742 and destination registers in an extension insn are different.
12743 (combine_reaching_defs): Allow source and destination registers
12744 in extension to be different under limited circumstances.
12745 (add_removable_extension): Remove restriction that the
12746 source and destination registers in the extension are the same.
12747 (find_and_remove_re): Emit a copy from the extension's
12748 destination to its source after the defining insn if
12749 the source and destination registers are different.
12751 PR middle-end/59285
12752 * ifcvt.c (merge_if_block): If we are merging a block with more than
12753 one successor with a block with no successors, remove any BARRIER
12754 after the second block.
12756 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
12758 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
12760 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
12763 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
12764 for 14-bit register offsets when INT14_OK_STRICT is false.
12766 2014-01-07 Roland Stigge <stigge@antcom.de>
12767 Michael Meissner <meissner@linux.vnet.ibm.com>
12770 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
12771 Only check TFmode for SPE constants. Don't check TImode or TDmode.
12773 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
12775 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
12778 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
12780 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
12781 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
12782 rtx is const0_rtx or not.
12784 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
12787 * target-globals.c (save_target_globals): Remove this_fn_optab
12789 * toplev.c: Include optabs.h.
12790 (target_reinit): Temporarily restore the global options if another
12791 set of options are in force.
12793 2014-01-07 Jakub Jelinek <jakub@redhat.com>
12795 PR rtl-optimization/58668
12796 * cfgcleanup.c (flow_find_cross_jump): Don't count
12797 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
12798 to determine what is counted.
12799 (flow_find_head_matching_sequence): Use active_insn_p to determine
12801 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
12803 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
12804 determine what is counted.
12806 PR tree-optimization/59643
12807 * tree-predcom.c (split_data_refs_to_components): If one dr is
12808 read and one write, determine_offset fails and the write isn't
12809 in the bad component, just put the read into the bad component.
12811 2014-01-07 Mike Stump <mikestump@comcast.net>
12812 Jakub Jelinek <jakub@redhat.com>
12815 * tree-core.h (struct tree_optimization_option): Change optabs
12816 type from unsigned char * to void *.
12817 * optabs.c (init_tree_optimization_optabs): Adjust
12818 TREE_OPTIMIZATION_OPTABS initialization.
12820 2014-01-06 Jakub Jelinek <jakub@redhat.com>
12823 * config/i386/i386.h (struct machine_function): Add
12824 no_drap_save_restore field.
12825 * config/i386/i386.c (ix86_save_reg): Use
12826 !cfun->machine->no_drap_save_restore instead of
12827 crtl->stack_realign_needed.
12828 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
12829 this function clears frame_pointer_needed. Set
12830 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
12831 and DRAP reg is needed.
12833 2014-01-06 Marek Polacek <polacek@redhat.com>
12836 * doc/implement-c.texi: Mention that other integer types are
12837 permitted as bit-field types in strictly conforming mode.
12839 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
12841 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
12842 is newly allocated.
12844 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
12846 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
12848 2014-01-06 Martin Jambor <mjambor@suse.cz>
12851 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
12853 * ipa-prop.c (ipa_print_node_params): Fix indentation.
12855 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
12859 * var-tracking.c (add_stores): Preserve the value of the source even if
12860 we don't record the store.
12862 2014-01-06 Terry Guo <terry.guo@arm.com>
12864 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
12866 2014-01-05 Iain Sandoe <iain@codesourcery.com>
12869 * config/darwin.c (darwin_function_section): Adjust return values to
12870 correspond to optimisation changes made in r206070.
12872 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
12874 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
12875 from prefetch_block tune setting.
12876 (nocona_cost): Correct size of prefetch block to 64.
12878 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
12880 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
12881 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
12882 used to save the static chain register in the computation of the offset
12883 from which the FP registers need to be restored.
12885 2014-01-04 Jakub Jelinek <jakub@redhat.com>
12887 PR tree-optimization/59519
12888 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
12889 ICE if get_current_def (current_new_name) is already non-NULL, as long
12890 as it is a phi result of some other phi in *new_exit_bb that has
12893 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
12894 or vmovdqu* for misaligned_operand.
12895 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
12896 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
12897 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
12898 aligned_mem for AVX512F masked aligned load and store builtins and for
12899 non-temporal moves.
12901 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
12903 PR tree-optimization/59651
12904 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
12905 Address range for negative step should be added by TYPE_SIZE_UNIT.
12907 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
12909 * config/m68k/m68k.c (handle_move_double): Handle pushes with
12910 overlapping registers also for registers other than the stack pointer.
12912 2014-01-03 Marek Polacek <polacek@redhat.com>
12915 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
12918 2014-01-03 Jakub Jelinek <jakub@redhat.com>
12921 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
12924 * config/i386/i386.md (MODE_SIZE): New mode attribute.
12925 (push splitter): Use <P:MODE_SIZE> instead of
12926 GET_MODE_SIZE (<P:MODE>mode).
12927 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
12928 (mov -1, reg peephole2): Likewise.
12929 * config/i386/sse.md (*mov<mode>_internal,
12930 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
12931 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
12932 *<code><mode>3, *andnot<mode>3<mask_name>,
12933 <mask_codefor><code><mode>3<mask_name>): Likewise.
12934 * config/i386/subst.md (mask_mode512bit_condition,
12935 sd_mask_mode512bit_condition): Likewise.
12937 2014-01-02 Xinliang David Li <davidxl@google.com>
12939 PR tree-optimization/59303
12940 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
12941 (dump_predicates): Better output format.
12942 (pred_equal_p): New function.
12943 (is_neq_relop_p): Ditto.
12944 (is_neq_zero_form_p): Ditto.
12945 (pred_expr_equal_p): Ditto.
12946 (pred_neg_p): Ditto.
12947 (simplify_pred): Ditto.
12948 (simplify_preds_2): Ditto.
12949 (simplify_preds_3): Ditto.
12950 (simplify_preds_4): Ditto.
12951 (simplify_preds): Ditto.
12952 (push_pred): Ditto.
12953 (push_to_worklist): Ditto.
12954 (get_pred_info_from_cmp): Ditto.
12955 (is_degenerated_phi): Ditto.
12956 (normalize_one_pred_1): Ditto.
12957 (normalize_one_pred): Ditto.
12958 (normalize_one_pred_chain): Ditto.
12959 (normalize_preds): Ditto.
12960 (normalize_cond_1): Remove function.
12961 (normalize_cond): Ditto.
12962 (is_gcond_subset_of): Ditto.
12963 (is_subset_of_any): Ditto.
12964 (is_or_set_subset_of): Ditto.
12965 (is_and_set_subset_of): Ditto.
12966 (is_norm_cond_subset_of): Ditto.
12967 (pred_chain_length_cmp): Ditto.
12968 (convert_control_dep_chain_into_preds): Type change.
12969 (find_predicates): Ditto.
12970 (find_def_preds): Ditto.
12971 (destroy_predicates_vecs): Ditto.
12972 (find_matching_predicates_in_rest_chains): Ditto.
12973 (use_pred_not_overlap_with_undef_path_pred): Ditto.
12974 (is_pred_expr_subset): Ditto.
12975 (is_pred_chain_subset_of): Ditto.
12976 (is_included_in): Ditto.
12977 (is_superset_of): Ditto.
12979 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
12981 Update copyright years.
12983 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
12985 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
12986 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
12987 config/arc/arc.md, config/arc/arc.opt,
12988 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
12989 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
12990 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
12991 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
12992 config/linux-protos.h, config/linux.c, config/winnt-c.c,
12993 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
12994 vtable-verify.c, vtable-verify.h: Use the standard form for the
12997 2014-01-02 Tobias Burnus <burnus@net-b.de>
12999 * gcc.c (process_command): Update copyright notice dates.
13000 * gcov-dump.c: Ditto.
13002 * doc/cpp.texi: Bump @copying's copyright year.
13003 * doc/cppinternals.texi: Ditto.
13004 * doc/gcc.texi: Ditto.
13005 * doc/gccint.texi: Ditto.
13006 * doc/gcov.texi: Ditto.
13007 * doc/install.texi: Ditto.
13008 * doc/invoke.texi: Ditto.
13010 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
13012 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
13014 2014-01-01 Jakub Jelinek <jakub@redhat.com>
13016 * config/i386/sse.md (*mov<mode>_internal): Guard
13017 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
13019 PR rtl-optimization/59647
13020 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
13021 new_rtx into UNSIGNED_FLOAT rtxes.
13023 Copyright (C) 2014 Free Software Foundation, Inc.
13025 Copying and distribution of this file, with or without modification,
13026 are permitted in any medium without royalty provided the copyright
13027 notice and this notice are preserved.