1 2014-06-07 Eric Botcazou <ebotcazou@adacore.com>
3 * tree-ssa-tail-merge.c (same_succ_hash): Hash the static chain of a
4 call statement, if any.
5 (gimple_equal_p) <GIMPLE_CALL>: Compare the static chain of the call
6 statements, if any. Tidy up.
8 2014-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
11 * config/rs6000/vsx.md (VSX_LE): Split VSX_D into 2 separate
12 iterators, VSX_D that handles 64-bit types, and VSX_LE that
13 handles swapping the two 64-bit double words on little endian
14 systems. Include V1TImode and optionally TImode in VSX_LE so that
15 these types are properly swapped. Change all of the insns and
16 splits that do the 64-bit swaps to use VSX_LE.
17 (vsx_le_perm_load_<mode>): Likewise.
18 (vsx_le_perm_store_<mode>): Likewise.
19 (splitters for little endian memory operations): Likewise.
20 (vsx_xxpermdi2_le_<mode>): Likewise.
21 (vsx_lxvd2x2_le_<mode>): Likewise.
22 (vsx_stxvd2x2_le_<mode>): Likewise.
24 2014-06-06 Uros Bizjak <ubizjak@gmail.com>
27 * config/i386/i386.md (*floatunssi<mode>2_i387_with_xmm): New
28 define_insn_and_split pattern, merged from *floatunssi<mode>2_1
29 and corresponding splitters. Zero extend general register
30 or memory input operand to XMM temporary. Enable for
31 TARGET_SSE2 and TARGET_INTER_UNIT_MOVES_TO_VEC only.
32 (floatunssi<mode>2): Update expander predicate.
34 2014-06-06 Vladimir Makarov <vmakarov@redhat.com>
36 PR rtl-optimization/61325
37 * lra-constraints.c (process_address_1): Check scale equal to one
38 to prevent transformation: base + scale * index => base + new_reg.
40 2014-06-06 Richard Biener <rguenther@suse.de>
42 PR tree-optimization/59299
43 * tree-ssa-sink.c (all_immediate_uses_same_place): Work on
45 (nearest_common_dominator_of_uses): Likewise.
46 (statement_sink_location): Adjust. Support sinking loads.
48 2014-06-06 Martin Jambor <mjambor@suse.cz>
50 * ipa-prop.c (get_place_in_agg_contents_list): New function.
51 (build_agg_jump_func_from_list): Likewise.
52 (determine_known_aggregate_parts): Renamed to
53 determine_locally_known_aggregate_parts. Moved some functionality
54 to the two functions above, removed bound checks.
56 2014-06-06 James Greenhalgh <james.greenhalgh@arm.com>
58 * config/aarch64/aarch64-protos.h (aarch64_expand_movmem): New.
59 * config/aarch64/aarch64.c (aarch64_move_pointer): New.
60 (aarch64_progress_pointer): Likewise.
61 (aarch64_copy_one_part_and_move_pointers): Likewise.
62 (aarch64_expand_movmen): Likewise.
63 * config/aarch64/aarch64.h (MOVE_RATIO): Set low.
64 * config/aarch64/aarch64.md (movmem<mode>): New.
66 2014-06-06 Bingfeng Mei <bmei@broadcom.com>
68 * targhooks.c (default_add_stmt_cost): Call target specific
69 hook instead of default one.
71 2014-06-06 Thomas Preud'homme <thomas.preudhomme@arm.com>
73 * ChangeLog (2014-05-23): Fix ChangeLog entry to refer to target
74 endianness instead of host endianness.
75 * tree-ssa-math-opts.c (find_bswap_or_nop_1): Likewise in dumps and
78 2014-06-06 Eric Botcazou <ebotcazou@adacore.com>
81 * function.c (instantiate_decls): Process the saved static chain.
82 (expand_function_start): If not optimizing, save the static chain
84 * tree-nested.c (convert_all_function_calls): Always create the static
85 chain for nested functions if not optimizing.
87 2014-06-06 Eric Botcazou <ebotcazou@adacore.com>
89 * tree-cfg.c (make_edges) <GIMPLE_RETURN>: Put a location on the edge.
91 2014-06-06 Richard Biener <rguenther@suse.de>
93 * cfgexpand.c (expand_gimple_cond): Remove check for current_loops.
94 (construct_init_block): Likewise.
95 (construct_exit_block): Likewise.
96 (pass_expand::execute): Likewise.
97 * graphite.c (graphite_transforms): Replace check for current_loops
98 with a check for > 1 loops.
99 (pass_graphite_transforms::execute): Adjust.
100 * ipa-split.c (split_function): Remove check for current_loops.
101 * omp-low.c (expand_parallel_call): Likewise.
102 (expand_omp_for_init_counts): Likewise.
103 (extract_omp_for_update_vars): Likewise.
104 (expand_omp_for_generic): Likewise.
105 (expand_omp_sections): Likewise.
106 (expand_omp_target): Likewise.
107 * tracer.c (tail_duplicate): Likewise.
108 (pass_tracer::execute): Likewise.
109 * trans-mem.c (expand_transaction): Likewise.
110 * tree-complex.c (expand_complex_div_wide): Likewise.
111 * tree-eh.c (lower_resx): Likewise.
112 (cleanup_empty_eh_merge_phis): Likewise.
113 * tree-predcom.c (run_tree_predictive_commoning): Replace check for
114 current_loops with a check for > 1 loops.
115 (pass_predcom::execute): Adjust.
116 * tree-scalar-evolution.c (scev_reset): Remove check for current_loops.
117 * tree-ssa-copy.c (copy_prop_visit_phi_node): Likewise.
118 * tree-ssa-dom.c (pass_phi_only_cprop::execute): Likewise.
119 * tree-ssa-tail-merge.c (tail_merge_optimize): Likewise.
120 * tree-ssa-threadupdate.c (thread_through_all_blocks): Likewise.
121 * tree-switch-conversion.c (process_switch): Likewise.
122 * tree-tailcall.c (tree_optimize_tail_calls_1): Likewise.
123 * tree-vrp.c (vrp_visit_phi_node): Likewise.
124 (execute_vrp): Likewise.
125 * ubsan.c (ubsan_expand_null_ifn): Likewise.
127 2014-06-06 Eric Botcazou <ebotcazou@adacore.com>
129 * rtl.h (insn_location): Declare.
130 * cfgcleanup.c (try_forward_edges): Compare the locus of locations
131 with UNKNOWN_LOCATION.
132 * emit-rtl.c (insn_location): New function.
133 * final.c (notice_source_line): Check that the instruction has a
134 location before retrieving it and use insn_location.
135 * modulo-sched.c (loop_single_full_bb_p): Likewise.
136 * print-rtl.c (print_rtx): Likewise.
138 2014-06-06 Richard Biener <rguenther@suse.de>
140 * passes.def: Move 2nd VRP pass before phi-only-cprop.
142 2014-06-06 Christian Bruel <christian.bruel@st.com>
144 PR tree-optimization/43934
145 * tree-ssa-loop-im.c (determine_max_movement): Add PHI def constant
148 2014-06-06 Richard Sandiford <rdsandiford@googlemail.com>
150 * ira-lives.c (single_reg_class): Add missing break. Explicitly
151 return NO_REGS for extra address and memory constraints. Handle
152 operands that match (or are equivalent to something that matches)
153 extra constant constraints. Ignore other non-register operands.
155 2014-06-06 Alan Modra <amodra@gmail.com>
158 * doc/tm.texi.in (INCOMING_REG_PARM_STACK_SPACE): Document.
159 * doc/tm.texi: Regenerate.
160 * function.c (INCOMING_REG_PARM_STACK_SPACE): Provide default.
161 Use throughout in place of REG_PARM_STACK_SPACE.
162 * config/rs6000/rs6000.c (rs6000_reg_parm_stack_space): Add
163 "incoming" param. Pass to rs6000_function_parms_need_stack.
164 (rs6000_function_parms_need_stack): Add "incoming" param, ignore
165 prototype_p when incoming. Use function decl when incoming
166 to handle K&R style functions.
167 * config/rs6000/rs6000.h (REG_PARM_STACK_SPACE): Adjust.
168 (INCOMING_REG_PARM_STACK_SPACE): Define.
170 2014-06-05 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
173 * cfgexpand.c (expand_debug_expr): Use address space of nested
174 TREE_TYPE for ADDR_EXPR and MEM_REF.
176 2014-06-05 Jeff Law <law@redhat.com>
178 PR tree-optimization/61289
179 * tree-ssa-threadedge.c (invalidate_equivalences): Remove SRC_MAP and
180 DST_MAP parameters. Invalidate by walking all the SSA_NAME_VALUES
181 looking for those which match LHS. All callers changed.
182 (record_temporary_equivalences_from_phis): Remove SRC_MAP and DST_MAP
183 parameters and code which manipulated them. All callers changed.
184 (record_temporary_equivalences_from_stmts_at_dest): Remove SRC_MAP
185 and DST_MAP parameters. Simplify invalidation code by just calling
186 invalidate_equivalences. All callers changed.
187 (thread_across_edge): Simplify now that we don't need to maintain
188 the map of equivalences to invalidate.
190 2014-06-05 Kai Tietz <ktietz@redhat.com>
191 Richard Henderson <rth@redhat.com>
194 * config/i386/predicates.md (memory_nox32_operand): Add memory_operand
195 checking for !TARGET_X32.
196 * config/i386/i386.md (UNSPEC_PEEPSIB): New unspec constant.
197 (sibcall_intern): New define_insn, plus required peepholes.
198 (sibcall_pop_intern): Likewise.
199 (sibcall_value_intern): Likewise.
200 (sibcall_value_pop_intern): Likewise.
202 2014-06-05 Ilya Enkovich <ilya.enkovich@intel.com>
204 * tree-inline.c (tree_function_versioning): Check DF info existence
207 2014-06-05 Marcus Shawcroft <marcus.shawcroft@arm.com>
209 * config/aarch64/aarch64.h (aarch64_frame): Add hard_fp_offset and
211 * config/aarch64/aarch64.c (aarch64_layout_frame): Initialize
212 aarch64_frame hard_fp_offset and frame_size.
213 (aarch64_expand_prologue): Use aarch64_frame hard_fp_offset and
214 frame_size; remove original_frame_size.
215 (aarch64_expand_epilogue, aarch64_final_eh_return_addr): Likewise.
216 (aarch64_initial_elimination_offset): Remove frame_size and
217 offset. Use aarch64_frame frame_size.
219 2014-06-05 Marcus Shawcroft <marcus.shawcroft@arm.com>
220 Jiong Wang <jiong.wang@arm.com>
221 Renlin <renlin.li@arm.com>
223 * config/aarch64/aarch64.c (aarch64_layout_frame): Correct
224 initialization of R30 offset. Update offset. Iterate core
225 regisers upto X30. Remove X29, X30 specific code.
227 2014-06-05 Marcus Shawcroft <marcus.shawcroft@arm.com>
228 Jiong Wang <jiong.wang@arm.com>
230 * config/aarch64/aarch64.c (SLOT_NOT_REQUIRED, SLOT_REQUIRED): Define.
231 (aarch64_layout_frame): Use SLOT_NOT_REQUIRED and SLOT_REQUIRED.
232 (aarch64_register_saved_on_entry): Adjust test.
234 2014-06-05 Marcus Shawcroft <marcus.shawcroft@arm.com>
236 * config/aarch64/aarch64.h (machine_function): Move
237 saved_varargs_size from here...
238 (aarch64_frame): ... to here.
240 * config/aarch64/aarch64.c (aarch64_expand_prologue)
241 (aarch64_expand_epilogue, aarch64_final_eh_return_addr)
242 (aarch64_initial_elimination_offset)
243 (aarch64_setup_incoming_varargs): Adjust location of
246 2014-06-05 Marcus Shawcroft <marcus.shawcroft@arm.com>
248 * config/aarch64/aarch64.c (aarch64_expand_prologue): Update stack
251 2014-06-05 Jaydeep Patil <Jaydeep.Patil@imgtec.com>
252 Prachi Godbole <Prachi.Godbole@imgtec.com>
254 * config/mips/mips-cpus.def: Add definition for p5600. Updated
255 mips32r5 entry to use PROCESSOR_P5600.
256 * config/mips/mips-tables.opt: Regenerate.
257 * config/mips/mips-protos.h (mips_fmadd_bypass): Add prototype.
258 * config/mips/mips.c (mips_fmadd_bypass): New function.
259 (mips_rtx_cost_data): Add costs for p5600.
260 (mips_issue_rate): Add support for p5600.
261 (mips_multipass_dfa_lookahead): Likewise.
262 * config/mips/mips.h (TUNE_P5600): New define.
263 (TUNE_MACC_CHAINS): Add TUNE_P5600.
264 (MIPS_ISA_LEVEL_SPEC): Map -march=p5600 to -mips32r5.
265 * config/mips/mips.md: Include p5600.md.
266 (processor): Add p5600.
267 * config/mips/p5600.md: New file.
269 2014-06-05 Evgeny Stupachenko <evstupac@gmail.com>
271 * config/i386/sse.md (*ssse3_palignr<mode>_perm): New.
272 * config/i386/predicates.md (palignr_operand): New.
273 Indicates if permutation is suitable for palignr instruction.
275 2014-06-05 Yuri Rumyantsev <ysrumyan@gmail.com>
277 PR tree-optimization/61319
278 * tree-if-conv.c (is_cond_scalar_reduction): Add missed check that
279 stmt belongs to loop.
281 2014-06-05 Richard Biener <rguenther@suse.de>
283 * gimplify.c (create_tmp_from_val): Remove is_formal parameter
284 and set DECL_GIMPLE_REG_P unconditionally if appropriate.
285 (lookup_tmp_var): Adjust.
286 (prepare_gimple_addressable): Unset DECL_GIMPLE_REG_P here.
288 2014-06-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
290 * config/arm/arm.md (enabled): Disable opt_enabled attribute.
292 2014-06-05 Marek Polacek <polacek@redhat.com>
295 * doc/invoke.texi: Document -Wlogical-not-parentheses.
297 2014-06-04 Tom de Vries <tom@codesourcery.com>
299 * config/s390/s390.md ("addptrdi3", "addptrsi3"): Use INTVAL only on
302 2014-06-04 Marc Glisse <marc.glisse@inria.fr>
304 PR tree-optimization/61385
305 * tree-ssa-phiopt.c (value_replacement): Punt if there are PHI nodes.
307 2014-06-04 Bernd Schmidt <bernds@codesourcery.com>
309 * lto-wrapper.c (fatal, fatal_perror): Remove functions. All callers
310 changed to use fatal_error.
311 (main): Ensure lto_wrapper_cleanup is run atexit.
313 2014-06-04 Richard Sandiford <rdsandiford@googlemail.com>
315 * lra-constraints.c (valid_address_p): Move earlier in file.
316 (address_eliminator): New structure.
317 (satisfies_memory_constraint_p): New function.
318 (satisfies_address_constraint_p): Likewise.
319 (process_alt_operands, process_address, curr_insn_transform): Use them.
321 2014-06-04 Richard Sandiford <rdsandiford@googlemail.com>
323 * lra-int.h (lra_static_insn_data): Make operand_alternative a
325 (target_lra_int, default_target_lra_int, this_target_lra_int)
326 (op_alt_data): Delete.
327 * lra.h (lra_init): Delete.
328 * lra.c (default_target_lra_int, this_target_lra_int): Delete.
329 (init_insn_code_data_once): Remove op_alt_data handling.
330 (finish_insn_code_data_once): Likewise.
331 (init_op_alt_data): Delete.
332 (get_static_insn_data): Initialize operand_alternative to null.
333 (free_insn_recog_data): Cast operand_alternative before freeing it.
334 (setup_operand_alternative): Take the operand_alternative as
335 parameter and assume it isn't already cached in the static
337 (lra_set_insn_recog_data): Update accordingly.
339 * ira.c (ira_init): Don't call lra_init.
340 * target-globals.h (this_target_lra_int): Declare.
341 (target_globals): Remove lra_int.
342 (restore_target_globals): Update accordingly.
343 * target-globals.c: Don't include lra-int.h.
344 (default_target_globals, save_target_globals): Remove lra_int.
346 2014-06-04 Richard Sandiford <rdsandiford@googlemail.com>
348 * recog.h (operand_alternative): Convert reg_class, reject,
349 matched and matches into bitfields.
350 (preprocess_constraints): New overload.
351 (preprocess_insn_constraints): New function.
352 (preprocess_constraints): Take the insn as parameter.
353 (recog_op_alt): Change into a pointer.
354 (target_recog): Add x_op_alt.
355 * recog.c (asm_op_alt): New variable.
356 (recog_op_alt): Change into a pointer.
357 (preprocess_constraints): New overload, replacing the old function
358 definition with one that doesn't use global state.
359 (preprocess_insn_constraints): New function.
360 (preprocess_constraints): Use them. Take the insn as parameter.
361 Use asm_op_alt for asms.
362 (recog_init): Free existing x_op_alt entries.
363 * ira-lives.c (check_and_make_def_conflict): Make operand_alternative
365 (make_early_clobber_and_input_conflicts): Likewise.
366 (process_bb_node_lives): Pass the insn to process_constraints.
367 * reg-stack.c (check_asm_stack_operands): Likewise.
368 (subst_asm_stack_regs): Likewise.
369 * regcprop.c (copyprop_hardreg_forward_1): Likewise.
370 * regrename.c (build_def_use): Likewise.
371 * sched-deps.c (sched_analyze_insn): Likewise.
372 * sel-sched.c (get_reg_class, implicit_clobber_conflict_p): Likewise.
373 * config/arm/arm.c (xscale_sched_adjust_cost): Likewise.
374 (note_invalid_constants): Likewise.
375 * config/i386/i386.c (ix86_legitimate_combined_insn): Likewise.
376 (ix86_legitimate_combined_insn): Make operand_alternative pointer
379 2014-06-04 Richard Sandiford <rdsandiford@googlemail.com>
381 * recog.c (preprocess_constraints): Don't skip disabled alternatives.
382 * ira-lives.c (check_and_make_def_conflict): Check for disabled
384 (make_early_clobber_and_input_conflicts): Likewise.
385 * config/i386/i386.c (ix86_legitimate_combined_insn): Likewise.
387 2014-06-04 Richard Sandiford <rdsandiford@googlemail.com>
389 * recog.h (alternative_class): New function.
390 (which_op_alt): Return a const recog_op_alt.
391 * reg-stack.c (check_asm_stack_operands): Update type accordingly.
392 (subst_asm_stack_regs): Likewise.
393 * config/arm/arm.c (note_invalid_constants): Likewise.
394 * regcprop.c (copyprop_hardreg_forward_1): Likewise. Don't modify
395 the operand_alternative; use alternative class instead.
396 * sel-sched.c (get_reg_class): Likewise.
397 * regrename.c (build_def_use): Likewise.
398 (hide_operands, restore_operands, record_out_operands): Update type
401 2014-06-04 Richard Sandiford <rdsandiford@googlemail.com>
403 * recog.h (recog_op_alt): Convert to a flat array.
404 (which_op_alt): New function.
405 * recog.c (recog_op_alt): Convert to a flat array.
406 (preprocess_constraints): Update accordingly, grouping all
407 operands of the same alternative together, rather than the
409 * ira-lives.c (check_and_make_def_conflict): Likewise.
410 (make_early_clobber_and_input_conflicts): Likewise.
411 * config/i386/i386.c (ix86_legitimate_combined_insn): Likewise.
412 * reg-stack.c (check_asm_stack_operands): Use which_op_alt.
413 (subst_asm_stack_regs): Likewise.
414 * regcprop.c (copyprop_hardreg_forward_1): Likewise.
415 * regrename.c (hide_operands, record_out_operands): Likewise.
416 (build_def_use): Likewise.
417 * sel-sched.c (get_reg_class): Likewise.
418 * config/arm/arm.c (note_invalid_constants): Likewise.
420 2014-06-04 Jason Merrill <jason@redhat.com>
424 * gimplify.c (gimplify_arg): Non-static.
425 * gimplify.h: Declare it.
427 2014-06-04 Richard Biener <rguenther@suse.de>
429 * tree.h (may_be_aliased): Trust TREE_ADDRESSABLE from
430 TREE_PUBLIC and DECL_EXTERNAL decls.
432 2014-06-04 Matthew Fortune <matthew.fortune@imgtec.com>
434 * regcprop.c (copyprop_hardreg_forward_1): Account for
435 HARD_REGNO_CALL_PART_CLOBBERED.
437 2014-06-04 Richard Biener <rguenther@suse.de>
439 * configure.ac: Check whether the underlying type of int64_t
440 is long or long long.
441 * configure: Regenerate.
442 * config.in: Likewise.
443 * hwint.h (HOST_WIDE_INT): Match the underlying type of int64_t.
444 (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
446 2014-06-04 Richard Biener <rguenther@suse.de>
448 PR tree-optimization/60098
449 * tree-ssa-dse.c (dse_possible_dead_store_p): Walk until
451 (dse_optimize_stmt): Simplify, now that we found a kill
454 2014-06-04 Richard Biener <rguenther@suse.de>
456 * tree-ssa-alias.c (stmt_may_clobber_ref_p): Improve handling
457 of accesses with non-invariant address.
459 2014-06-04 Martin Liska <mliska@suse.cz>
461 * cgraph.h (cgraph_make_wrapper): New function introduced.
462 * cgraphunit.c (cgraph_make_wrapper): The function implementation.
463 * ipa-inline.h (inline_analyze_function): The function is global.
464 * ipa-inline-analysis.c (inline_analyze_function): Likewise.
466 2014-06-04 Martin Liska <mliska@suse.cz>
468 * tree.h (private_lookup_attribute_starting): New function.
469 (lookup_attribute_starting): Likewise.
470 * tree.c (private_lookup_attribute_starting): Likewise.
472 2014-06-04 Martin Liska <mliska@suse.cz>
474 * cgraph.h (expand_thunk): New argument added.
475 (address_taken_from_non_vtable_p): New global function.
476 * ipa-visibility.c (address_taken_from_non_vtable_p): Likewise.
477 * cgraphclones.c (duplicate_thunk_for_node): Argument added to call.
478 * cgraphunit.c (analyze_function): Likewise.
479 (assemble_thunks_and_aliases): Argument added to call.
480 (expand_thunk): New argument forces to produce GIMPLE thunk.
482 2014-06-04 Martin Liska <mliska@suse.cz>
484 * coverage.h (coverage_compute_cfg_checksum): Argument added.
485 * coverage.c (coverage_compute_cfg_checksum): Likewise.
486 * profile.c (branch_prob): Likewise.
488 2014-06-04 Martin Jambor <mjambor@suse.cz>
491 * ipa-pure-const.c (propagate_pure_const): Add unreachable default
492 handler for switch on an ipa_ref_use enum.
493 * ipa-reference.c (analyze_function): Likewise.
495 2014-06-04 Kai Tietz <ktietz@redhat.com>
497 * recog.c (peep2_attempt): Copy SIBLING_CALL_P flag
498 from old call-instruction.
500 2014-06-04 Bin Cheng <bin.cheng@arm.com>
502 * config/aarch64/aarch64.c (aarch64_classify_address)
503 (aarch64_legitimize_reload_address): Support full addressing modes
505 * config/aarch64/aarch64.md (mov<mode>, movmisalign<mode>)
506 (*aarch64_simd_mov<mode>, *aarch64_simd_mov<mode>): Relax predicates.
508 2014-06-03 Andrew Pinski <apinski@cavium.com>
510 * config/aarch64/aarch64.c (aarch64_if_then_else_costs): Allow non comparisons
513 2014-06-03 Andrew Pinski <apinski@cavium.com>
515 * config/aarch64/aarch64.c (aarch64_if_then_else_costs): New function.
516 (aarch64_rtx_costs): Use aarch64_if_then_else_costs.
518 2014-06-03 Kai Tietz <ktietz@redhat.com>
520 * config/i386/i386.c (ix86_function_value_regno_p): Disallow DX_REG
523 2014-06-03 Dehao Chen <dehao@google.com>
525 * tree-cfg.c (gimple_merge_blocks): Only reset count when BBs are in
528 2014-06-03 Marek Polacek <polacek@redhat.com>
531 * doc/invoke.texi: Document -Wswitch-bool.
532 * function.c (stack_protect_epilogue): Cast controlling expression of
534 * gengtype.c (walk_type): Generate switch expression with its
535 controlling expression cast to int.
537 2014-06-03 Vishnu K S <Vishnu.k_s@atmel.com>
539 * config/avr/avr-mcus.def: Add new avr25 devices attiny441, attiny828
541 * config/avr/avr-tables.opt: Regenerate.
542 * config/avr/t-multilib: Regenerate.
543 * doc/avr-mmcu.texi: Regenerate.
545 2014-06-03 Vishnu K S <vishnu.k_s@atmel.com>
546 Pitchumani Sivanupandi <pitchumani.s@atmel.com>
548 * config/avr/avr-mcus.def (ata6616c): Add new avr25 device.
549 (ata6617c, ata664251): Add new avr35 devices.
550 (ata6612c): Add new avr4 device.
551 (ata6613c, ata6614q): Add new avr5 devices.
552 * config/avr/avr-tables.opt: Regenerate.
553 * config/avr/t-multilib: Regenerate.
554 * doc/avr-mmcu.texi: Regenerate.
556 2014-06-03 Alan Lawrence <alan.lawrence@arm.com>
558 * gcc/config/aarch64/aarch64-builtins.c
559 (aarch64_types_binop_ssu_qualifiers): New static data.
560 (TYPES_BINOP_SSU): Define.
561 * gcc/config/aarch64/aarch64-simd-builtins.def (suqadd, ushl, urshl,
562 urshr_n, ushll_n): Use appropriate unsigned qualifiers.
563 * gcc/config/aarch64/arm_neon.h (vrshl_u8, vrshl_u16, vrshl_u32,
564 vrshl_u64, vrshlq_u8, vrshlq_u16, vrshlq_u32, vrshlq_u64, vrshld_u64,
565 vrshr_n_u8, vrshr_n_u16, vrshr_n_u32, vrshr_n_u64, vrshrq_n_u8,
566 vrshrq_n_u16, vrshrq_n_u32, vrshrq_n_u64, vrshrd_n_u64, vshll_n_u8,
567 vshll_n_u16, vshll_n_u32, vuqadd_s8, vuqadd_s16, vuqadd_s32,
568 vuqadd_s64, vuqaddq_s8, vuqaddq_s16, vuqaddq_s32, vuqaddq_s64,
569 vuqaddb_s8, vuqaddh_s16, vuqadds_s32, vuqaddd_s64): Add signedness
570 suffix to builtin function name, remove cast.
571 (vshl_s8, vshl_s16, vshl_s32, vshl_s64, vshl_u8, vshl_u16, vshl_u32,
572 vshl_u64, vshlq_s8, vshlq_s16, vshlq_s32, vshlq_s64, vshlq_u8,
573 vshlq_u16, vshlq_u32, vshlq_u64, vshld_s64, vshld_u64): Remove cast.
575 2014-06-03 Alan Lawrence <alan.lawrence@arm.com>
577 * gcc/config/aarch64/aarch64-builtins.c
578 (aarch64_types_binop_uus_qualifiers,
579 aarch64_types_shift_to_unsigned_qualifiers,
580 aarch64_types_unsigned_shiftacc_qualifiers): Define.
581 * gcc/config/aarch64/aarch64-simd-builtins.def (uqshl, uqrshl, uqadd,
582 uqsub, usqadd, usra_n, ursra_n, uqshrn_n, uqrshrn_n, usri_n, usli_n,
583 sqshlu_n, uqshl_n): Update qualifiers.
584 * gcc/config/aarch64/arm_neon.h (vqadd_u8, vqadd_u16, vqadd_u32,
585 vqadd_u64, vqaddq_u8, vqaddq_u16, vqaddq_u32, vqaddq_u64, vqsub_u8,
586 vqsub_u16, vqsub_u32, vqsub_u64, vqsubq_u8, vqsubq_u16, vqsubq_u32,
587 vqsubq_u64, vqaddb_u8, vqaddh_u16, vqadds_u32, vqaddd_u64, vqrshl_u8,
588 vqrshl_u16, vqrshl_u32, vqrshl_u64, vqrshlq_u8, vqrshlq_u16,
589 vqrshlq_u32, vqrshlq_u64, vqrshlb_u8, vqrshlh_u16, vqrshls_u32,
590 vqrshld_u64, vqrshrn_n_u16, vqrshrn_n_u32, vqrshrn_n_u64,
591 vqrshrnh_n_u16, vqrshrns_n_u32, vqrshrnd_n_u64, vqshl_u8, vqshl_u16,
592 vqshl_u32, vqshl_u64, vqshlq_u8, vqshlq_u16, vqshlq_u32, vqshlq_u64,
593 vqshlb_u8, vqshlh_u16, vqshls_u32, vqshld_u64, vqshl_n_u8, vqshl_n_u16,
594 vqshl_n_u32, vqshl_n_u64, vqshlq_n_u8, vqshlq_n_u16, vqshlq_n_u32,
595 vqshlq_n_u64, vqshlb_n_u8, vqshlh_n_u16, vqshls_n_u32, vqshld_n_u64,
596 vqshlu_n_s8, vqshlu_n_s16, vqshlu_n_s32, vqshlu_n_s64, vqshluq_n_s8,
597 vqshluq_n_s16, vqshluq_n_s32, vqshluq_n_s64, vqshlub_n_s8,
598 vqshluh_n_s16, vqshlus_n_s32, vqshlud_n_s64, vqshrn_n_u16,
599 vqshrn_n_u32, vqshrn_n_u64, vqshrnh_n_u16, vqshrns_n_u32,
600 vqshrnd_n_u64, vqsubb_u8, vqsubh_u16, vqsubs_u32, vqsubd_u64,
601 vrsra_n_u8, vrsra_n_u16, vrsra_n_u32, vrsra_n_u64, vrsraq_n_u8,
602 vrsraq_n_u16, vrsraq_n_u32, vrsraq_n_u64, vrsrad_n_u64, vsli_n_u8,
603 vsli_n_u16, vsli_n_u32,vsli_n_u64, vsliq_n_u8, vsliq_n_u16,
604 vsliq_n_u32, vsliq_n_u64, vslid_n_u64, vsqadd_u8, vsqadd_u16,
605 vsqadd_u32, vsqadd_u64, vsqaddq_u8, vsqaddq_u16, vsqaddq_u32,
606 vsqaddq_u64, vsqaddb_u8, vsqaddh_u16, vsqadds_u32, vsqaddd_u64,
607 vsra_n_u8, vsra_n_u16, vsra_n_u32, vsra_n_u64, vsraq_n_u8,
608 vsraq_n_u16, vsraq_n_u32, vsraq_n_u64, vsrad_n_u64, vsri_n_u8,
609 vsri_n_u16, vsri_n_u32, vsri_n_u64, vsriq_n_u8, vsriq_n_u16,
610 vsriq_n_u32, vsriq_n_u64, vsrid_n_u64): Remove casts.
612 2014-06-03 Teresa Johnson <tejohnson@google.com>
614 * tree-sra.c (modify_function): Record caller nodes after rebuild.
616 2014-06-02 Jason Merrill <jason@redhat.com>
619 * varpool.c (ctor_for_folding): Handle uninitialized vtables.
621 2014-06-03 Alan Lawrence <alan.lawrence@arm.com>
623 * config/aarch64/aarch64.c (aarch64_evpc_ext): allow and handle
626 2014-06-03 Alan Lawrence <alan.lawrence@arm.com>
628 * config/aarch64/aarch64-simd.md (aarch64_rev<REVERSE:rev-op><mode>):
630 * config/aarch64/aarch64.c (aarch64_evpc_rev): New function.
631 (aarch64_expand_vec_perm_const_1): Add call to aarch64_evpc_rev.
632 * config/aarch64/iterators.md (REVERSE): New iterator.
633 (UNSPEC_REV64, UNSPEC_REV32, UNSPEC_REV16): New enum elements.
634 (rev_op): New int_attribute.
635 * config/aarch64/arm_neon.h (vrev16_p8, vrev16_s8, vrev16_u8,
636 vrev16q_p8, vrev16q_s8, vrev16q_u8, vrev32_p8, vrev32_p16, vrev32_s8,
637 vrev32_s16, vrev32_u8, vrev32_u16, vrev32q_p8, vrev32q_p16, vrev32q_s8,
638 vrev32q_s16, vrev32q_u8, vrev32q_u16, vrev64_f32, vrev64_p8,
639 vrev64_p16, vrev64_s8, vrev64_s16, vrev64_s32, vrev64_u8, vrev64_u16,
640 vrev64_u32, vrev64q_f32, vrev64q_p8, vrev64q_p16, vrev64q_s8,
641 vrev64q_s16, vrev64q_s32, vrev64q_u8, vrev64q_u16, vrev64q_u32):
642 Replace temporary __asm__ with __builtin_shuffle.
644 2014-06-03 Andrew Bennett <andrew.bennett@imgtec.com>
646 * config/mips/mips-cpus.def: Add mips32r3, mips32r5, mips64r3 and
648 * config/mips/mips-tables.opt: Regenerate.
649 * config/mips/mips.c (mips_compute_frame_info): Changed if statement
650 to use mips_isa_rev rather than ISA_MIPS32R2.
651 * config/mips/mips.h (ISA_MIPS32R3): New define.
652 (ISA_MIPS32R5): New define.
653 (ISA_MIPS64R3): New define.
654 (ISA_MIPS64R5): New define.
655 (TARGET_CPU_CPP_BUILTINS): Added support for ISA_MIPS32R3,
656 ISA_MIPS32R5, ISA_MIPS64R3 and ISA_MIPS64R5.
657 (MIPS_ISA_LEVEL_SPEC): Added support for mips32r3, mips32r5, mips64r3
659 (MIPS_ISA_SYNCI_SPEC): Likewise.
660 (ISA_HAS_64BIT_REGS): Added ISA_MIPS64R3 and ISA_MIPS64R5.
661 (LINK_SPEC): Added mips32r3 and mips32r5.
662 * config/mips/t-isa3264 (MULTILIB_MATCHES): Map mips32r3 and mips32r5
663 to mips32r2; and mips64r3 and mips64r5 to mips64r2.
664 * config/mips/t-mti-elf (MULTILIB_MATCHES): Likewise.
665 * config/mips/t-mti-linux (MULTILIB_MATCHES): Likewise.
666 * config/mips/t-sde (MULTILIB_MATCHES): Likewise.
667 * config/mips/t-sdemtk (MULTILIB_MATCHES): New define.
668 * doc/invoke.texi: Document mips32r3, mips32r5, mips64r3 and mips64r5.
670 2014-06-03 Andrew Bennett <andrew.bennett@imgtec.com>
672 * doc/invoke.texi: Document -mxpa and -mno-xpa MIPS command line
674 * config/mips/mips.opt (mxpa): New option.
675 * config/mips/mips.h (ASM_SPEC): Pass mxpa and mno-xpa to the
678 2014-06-03 Martin Jambor <mjambor@suse.cz>
681 * ipa-cp.c (cgraph_edge_brings_value_p): Handle edges leading to
684 2014-06-03 Thomas Preud'homme <thomas.preudhomme@arm.com>
686 PR tree-optimization/61328
687 * tree-ssa-math-opts.c (init_symbolic_number): Extract symbolic number
688 initialization from find_bswap_or_nop_1.
689 (find_bswap_or_nop_1): Test return value of find_bswap_or_nop_1 stored
690 in source_expr2 before using the size value the function sets. Also
691 make use of init_symbolic_number () in both the old place and
692 find_bswap_or_nop_load () to avoid reading uninitialized memory when
693 doing recursion in the GIMPLE_BINARY_RHS case.
695 2014-06-03 Richard Biener <rguenther@suse.de>
697 PR tree-optimization/61383
698 * tree-ssa-ifcombine.c (bb_no_side_effects_p): Make sure
701 2014-06-03 Richard Sandiford <rdsandiford@googlemail.com>
703 * defaults.h (USE_MD_CONSTRAINTS, EXTRA_MEMORY_CONSTRAINT)
704 (EXTRA_ADDRESS_CONSTRAINT, DEFAULT_CONSTRAINT_LEN, CONSTRAINT_LEN)
705 (CONST_OK_FOR_CONSTRAINT_P, CONST_DOUBLE_OK_FOR_LETTER_P)
706 (REG_CLASS_FROM_CONSTRAINT, EXTRA_CONSTRAINT_STR): Delete definitions
708 (REG_CLASS_FROM_LETTER, CONST_OK_FOR_LETTER_P)
709 (CONST_DOUBLE_OK_FOR_LETTER_P, EXTRA_CONSTRAINT): Move poising to...
710 * system.h: ...here and make it unconditional.
711 * target.def (conditional_register_usage): Mention
712 define_register_constraint instead of old-style constraint macros.
713 * doc/tm.texi.in: Remove documentation for old-style constraint macros.
714 * doc/tm.texi: Regenerate.
715 * genoutput.c: Remove USE_MD_CONSTRAINTS conditions and all code
716 protected by !USE_MD_CONSTRAINTS.
717 * config/frv/frv.md: Remove quote from old version of documentation.
718 * config/frv/frv.c (frv_conditional_register_usage): Likewise.
719 * config/m32r/m32r.c (easy_di_const, easy_df_const): Avoid mentioning
720 CONST_DOUBLE_OK_FOR_LETTER.
721 * config/sh/constraints.md: Likewise EXTRA_CONSTRAINT.
723 2014-06-02 Andrew Pinski <apinski@cavium.com>
725 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER):
726 /lib/ld-linux32-aarch64.so.1 is used for ILP32.
727 (LINUX_TARGET_LINK_SPEC): Update linker script for ILP32.
728 file whose name depends on -mabi= and -mbig-endian.
729 * config/aarch64/t-aarch64-linux (MULTILIB_OSDIRNAMES):
730 Handle LP64 better and handle ilp32 too.
731 (MULTILIB_OPTIONS): Delete.
732 (MULTILIB_DIRNAMES): Delete.
734 2014-06-02 Andrew MacLeod <amacleod@redhat.com>
736 * expr.h: Remove prototypes of functions defined in builtins.c.
737 * tree.h: (build_call_expr_*, build_string_literal): Add prototypes.
738 Remove prototypes of functions defined in builtins.c.
739 * builtins.h: Update prototype list to include all exported functions.
740 * builtins.c: (default_libc_has_function, gnu_libc_has_function,
741 no_c99_libc_has_function): Move to targhooks.c
742 (build_string_literal, build_call_expr_loc_array,
743 build_call_expr_loc_vec, build_call_expr_loc, build_call_expr): Move
745 (expand_builtin_object_size, fold_builtin_object_size): Make static.
746 * targhooks.c (default_libc_has_function, gnu_libc_has_function,
747 no_c99_libc_has_function): Relocate from builtins.c.
748 * tree.c: Include builtins.h.
749 (build_call_expr_loc_array, build_call_expr_loc_vec,
750 build_call_expr_loc, build_call_expr, build_string_literal): Relocate
752 * fold-const.h (fold_fma): Move prototype to builtins.h.
753 * realmpfr.h (do_mpc_arg2): Move prototype to builtins.h.
754 * asan.c: Include builtins.h.
755 * cfgexpand.c: Likewise.
756 * convert.c: Likewise.
757 * emit-rtl.c: Likewise.
758 * except.c: Likewise.
760 * fold-const.c: Likewise.
761 * gimple-fold.c: Likewise.
762 * gimple-ssa-strength-reduction.c: Likewise.
763 * gimplify.c: Likewise.
764 * ipa-inline.c: Likewise.
765 * ipa-prop.c: Likewise.
766 * lto-streamer-out.c: Likewise.
768 * tree-inline.c: Likewise.
769 * tree-object-size.c: Likewise.
770 * tree-sra.c: Likewise.
771 * tree-ssa-ccp.c: Likewise.
772 * tree-ssa-forwprop.c: Likewise.
773 * tree-ssa-loop-ivcanon.c: Likewise.
774 * tree-ssa-loop-ivopts.c: Likewise.
775 * tree-ssa-math-opts.c: Likewise.
776 * tree-ssa-reassoc.c: Likewise.
777 * tree-ssa-threadedge.c: Likewise.
778 * tree-streamer-in.c: Likewise.
779 * tree-vect-data-refs.c: Likewise.
780 * tree-vect-patterns.c: Likewise.
781 * tree-vect-stmts.c: Likewise.
782 * config/aarch64/aarch64.c: Likewise.
783 * config/alpha/alpha.c: Likewise.
784 * config/arc/arc.c: Likewise.
785 * config/arm/arm.c: Likewise.
786 * config/avr/avr.c: Likewise.
787 * config/bfin/bfin.c: Likewise.
788 * config/c6x/c6x.c: Likewise.
789 * config/cr16/cr16.c: Likewise.
790 * config/cris/cris.c: Likewise.
791 * config/epiphany/epiphany.c: Likewise.
792 * config/fr30/fr30.c: Likewise.
793 * config/frv/frv.c: Likewise.
794 * config/h8300/h8300.c: Likewise.
795 * config/i386/i386.c: Likewise.
796 * config/i386/winnt.c: Likewise.
797 * config/ia64/ia64.c: Likewise.
798 * config/iq2000/iq2000.c: Likewise.
799 * config/lm32/lm32.c: Likewise.
800 * config/m32c/m32c.c: Likewise.
801 * config/m32r/m32r.c: Likewise.
802 * config/m68k/m68k.c: Likewise.
803 * config/mcore/mcore.c: Likewise.
804 * config/mep/mep.c: Likewise.
805 * config/microblaze/microblaze.c: Likewise.
806 * config/mips/mips.c: Likewise.
807 * config/mmix/mmix.c: Likewise.
808 * config/mn10300/mn10300.c: Likewise.
809 * config/moxie/moxie.c: Likewise.
810 * config/msp430/msp430.c: Likewise.
811 * config/nds32/nds32.c: Likewise.
812 * config/pa/pa.c: Likewise.
813 * config/pdp11/pdp11.c: Likewise.
814 * config/picochip/picochip.c: Likewise.
815 * config/rl78/rl78.c: Likewise.
816 * config/rs6000/rs6000.c: Likewise.
817 * config/rx/rx.c: Likewise.
818 * config/s390/s390.c: Likewise.
819 * config/score/score.c: Likewise.
820 * config/sh/sh.c: Likewise.
821 * config/sparc/sparc.c: Likewise.
822 * config/spu/spu.c: Likewise.
823 * config/stormy16/stormy16.c: Likewise.
824 * config/tilegx/tilegx.c: Likewise.
825 * config/tilepro/tilepro.c: Likewise.
826 * config/v850/v850.c: Likewise.
827 * config/vax/vax.c: Likewise.
828 * config/xtensa/xtensa.c: Likewise.
830 2014-06-02 Jeff Law <law@redhat.com>
832 PR rtl-optimization/61094
833 * ree.c (combine_reaching_defs): Do not reextend an insn if it
834 was marked as do_no_reextend. If a copy is needed to eliminate
835 an extension, then mark it as do_not_reextend.
837 2014-06-02 Marcus Shawcroft <marcus.shawcroft@arm.com>
839 * config/aarch64/aarch64.md (set_fpcr): Drop ISB after FPCR write.
841 2014-06-02 Richard Henderson <rth@redhat.com>
844 * config/alpha/alpha.c (print_operand_address): Allow symbolic
845 addresses inside asms. Use output_operand_lossage instead of
848 2014-06-02 Uros Bizjak <ubizjak@gmail.com>
851 * config/i386/i386.c (ix86_expand_vec_perm) [case V32QImode]: Use
852 GEN_INT (-128) instead of GEN_INT (128) to set MSB of QImode constant.
854 2014-06-02 Tom de Vries <tom@codesourcery.com>
856 * config/aarch64/aarch64.c (aarch64_float_const_representable_p): Handle
857 case that x has VOIDmode.
859 2014-06-02 Bernd Schmidt <bernds@codesourcery.com>
861 * varasm.c (copy_constant): Delete function.
862 (build_constant_desc): Don't call it.
864 2014-06-02 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
867 * config/arm/arm.h (TARGET_SUPPORTS_WIDE_INT): Define.
868 * config/arm/arm.md (mov64 splitter): Replace const_double_operand
869 with immediate_operand.
871 2014-06-02 Andreas Schwab <schwab@suse.de>
874 (ia64_first_cycle_multipass_dfa_lookahead_guard): Check
875 pending_data_specs first.
877 2014-06-02 Richard Biener <rguenther@suse.de>
879 PR tree-optimization/61378
880 * tree-ssa-sccvn.c (vn_reference_lookup_3): Initialize
883 2014-06-01 Uros Bizjak <ubizjak@gmail.com>
885 * config/i386/constraints.md (Bw): Rename from 'w'.
886 (Bz): Rename from 'z'.
887 * config/i386/i386.md: Change 'w' to 'Bw' and 'z' to 'Bz' globally.
889 2014-06-01 Kai Tietz <ktietz@redhat.com>
892 * config/i386/constrains.md (define_constrain): New 'Bs' constraint.
893 * config/i386/i386.md (sibcall_insn_operand): Use Bs
894 instead of m constraint.
896 2014-05-31 Andreas Schwab <schwab@linux-m68k.org>
898 * config/m68k/m68k.md (beq0_di, bne0_di): Make the "o" constraint
899 a separate alternative where the scratch operand 2 is marked as
902 2014-05-31 Kugan Vivekanandarajah <kuganv@linaro.org>
904 * config/arm/arm.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New define.
905 (arm_builtins) : Add ARM_BUILTIN_GET_FPSCR and ARM_BUILTIN_SET_FPSCR.
906 (bdesc_2arg) : Add description for builtins __builtins_arm_set_fpscr
907 and __builtins_arm_get_fpscr.
908 (arm_init_builtins) : Initialize builtins __builtins_arm_set_fpscr and
909 __builtins_arm_get_fpscr.
910 (arm_expand_builtin) : Expand builtins __builtins_arm_set_fpscr and
911 __builtins_arm_ldfpscr.
912 (arm_atomic_assign_expand_fenv): New function.
913 * config/arm/vfp.md (set_fpscr): New pattern.
914 (get_fpscr) : Likewise.
915 * config/arm/unspecs.md (unspecv): Add VUNSPEC_GET_FPSCR and
917 * doc/extend.texi (AARCH64 Built-in Functions) : Document
918 __builtins_arm_set_fpscr, __builtins_arm_get_fpscr.
920 2014-05-30 Jakub Jelinek <jakub@redhat.com>
922 * asan.c (report_error_func): Add SLOW_P argument, use
923 BUILT_IN_ASAN_*_N if set.
924 (build_check_stmt): Likewise.
925 (instrument_derefs): If T has insufficient alignment,
926 force same handling as for odd sizes.
928 * sanitizer.def (BUILT_IN_ASAN_REPORT_LOAD_N,
929 BUILT_IN_ASAN_REPORT_STORE_N): New.
930 * asan.c (struct asan_mem_ref): Change access_size type to
932 (asan_mem_ref_init, asan_mem_ref_new, get_mem_refs_of_builtin_call,
933 update_mem_ref_hash_table): Likewise.
934 (asan_mem_ref_hasher::hash): Hash in a HWI.
935 (report_error_func): Change size_in_bytes argument to HWI.
936 Use *_N builtins if size_in_bytes is larger than 16 or not power of
938 (build_shadow_mem_access): New function.
939 (build_check_stmt): Use it. Change size_in_bytes argument to HWI.
940 Handle size_in_bytes not power of two or larger than 16.
941 (instrument_derefs): Don't give up if size_in_bytes is not
942 power of two or is larger than 16.
944 2014-05-30 Kai Tietz <ktietz@redhat.com>
947 * config/i386/i386.c (x86_output_mi_thunk): Add memory case
948 for sibling-tail-calls.
949 * config/i386/i386.md (sibcall_insn_operand): Add memory-constrain
951 * config/i386/predicates.md (sibcall_memory_operand): New predicate.
952 (sibcall_insn_operand): Add check for sibcall_memory_operand.
954 2014-05-30 Pitchumani Sivanupandi <pitchumani.s@atmel.com>
956 * config/avr/avr-mcus.def: Change ATA6289 ISA to AVR4
957 * config/avr/avr-tables.opt: Regenerate.
958 * config/avr/t-multilib: Regenerate.
959 * doc/avr-mmcu.texi: Regenerate.
961 2014-05-30 Ian Lance Taylor <iant@google.com>
963 * config/i386/xmmintrin.h (_mm_pause): Move out of scope of pragma
966 2014-05-30 Tom de Vries <tom@codesourcery.com>
968 * config/i386/i386.c (TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS):
971 2014-05-30 Tom de Vries <tom@codesourcery.com>
973 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
974 * lra.c (initialize_lra_reg_info_element): Add init of
975 actual_call_used_reg_set field.
976 (lra): Call lra_create_live_ranges before lra_inheritance for
978 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
980 * lra-constraints.c (need_for_call_save_p): Use actual_call_used_reg_set
981 instead of call_used_reg_set for -fuse-caller-save.
982 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
984 2014-05-30 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
986 * config/arm/thumb2.md (*thumb2_movhi_insn): Set type of movw
988 * config/arm/vfp.md (*thumb2_movsi_vfp): Likewise.
990 2014-05-30 Richard Sandiford <rdsandiford@googlemail.com>
992 * ira.c (ira_get_dup_out_num): Check for output operands at
993 the start of the loop. Handle cases where an included alternative
994 follows an excluded one.
996 2014-05-29 Mike Stump <mikestump@comcast.net>
999 * collect2.c (maybe_run_lto_and_relink): Be sure to always run
1000 post ld passes when lto is used.
1002 2014-05-29 Vladimir Makarov <vmakarov@redhat.com>
1004 PR rtl-optimization/61325
1005 * lra-constraints.c (process_address): Rename to process_address_1.
1006 (process_address): New function.
1008 2014-05-29 Alan Lawrence <alan.lawrence@arm.com>
1010 * config/aarch64/aarch64-builtins.c (aarch64_types_binopv_qualifiers,
1011 TYPES_BINOPV): New static data.
1012 * config/aarch64/aarch64-simd-builtins.def (im_lane_bound):
1014 * config/aarch64/aarch64-simd.md (aarch64_ext,
1015 aarch64_im_lane_boundsi): New patterns.
1016 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_const_1): Match
1018 (aarch64_evpc_ext): New function.
1020 * config/aarch64/iterators.md (UNSPEC_EXT): New enum element.
1022 * config/aarch64/arm_neon.h (vext_f32, vext_f64, vext_p8, vext_p16,
1023 vext_s8, vext_s16, vext_s32, vext_s64, vext_u8, vext_u16, vext_u32,
1024 vext_u64, vextq_f32, vextq_f64, vextq_p8, vextq_p16, vextq_s8,
1025 vextq_s16, vextq_s32, vextq_s64, vextq_u8, vextq_u16, vextq_u32,
1026 vextq_u64): Replace __asm with __builtin_shuffle and im_lane_boundsi.
1028 2014-05-29 Tom de Vries <tom@codesourcery.com>
1030 * rtl.h (BLOCK_SYMBOL_CHECK): Use SYMBOL_REF_FLAGS.
1032 2014-05-29 Richard Earnshaw <rearnsha@arm.com>
1033 Richard Sandiford <rdsandiford@googlemail.com>
1035 * arm/iterators.md (shiftable_ops): New code iterator.
1036 (t2_binop0, arith_shift_insn): New code attributes.
1037 * arm/predicates.md (shift_nomul_operator): New predicate.
1038 * arm/arm.md (insn_enabled): Delete.
1039 (enabled): Remove insn_enabled test.
1040 (*arith_shiftsi): Delete. Replace with ...
1041 (*<arith_shift_insn>_multsi): ... new pattern.
1042 (*<arith_shift_insn>_shiftsi): ... new pattern.
1043 * config/arm/arm.c (arm_print_operand): Handle operand format 'b'.
1045 2014-05-29 Radovan Obradovic <robradovic@mips.com>
1046 Tom de Vries <tom@codesourcery.com>
1048 * config/mips/mips.h (POST_CALL_TMP_REG): Define.
1049 * config/mips/mips.c (mips_emit_call_insn): Add POST_CALL_TMP_REG
1051 (mips_split_call): Use POST_CALL_TMP_REG.
1052 (TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): Redefine to true.
1054 2014-05-29 Tom de Vries <tom@codesourcery.com>
1056 * final.c (collect_fn_hard_reg_usage): Guard variable declaration
1057 with #ifdef STACK_REGS.
1059 2014-05-28 Jan Hubicka <hubicka@ucw.cz>
1061 * varasm.c (get_variable_section): Walk aliases.
1062 (place_block_symbol): Walk aliases.
1064 2014-05-28 Tom de Vries <tom@codesourcery.com>
1067 2014-05-28 Tom de Vries <tom@codesourcery.com>
1069 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
1070 * lra.c (initialize_lra_reg_info_element): Add init of
1071 actual_call_used_reg_set field.
1072 (lra): Call lra_create_live_ranges before lra_inheritance for
1074 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
1076 * lra-constraints.c (need_for_call_save_p): Use
1077 actual_call_used_reg_set instead of call_used_reg_set for
1079 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
1081 2014-05-28 Richard Sandiford <rdsandiford@googlemail.com>
1083 * doc/md.texi: Document that the % constraint character must
1084 be at the beginning of the string.
1085 * genoutput.c (validate_insn_alternatives): Check that '=',
1086 '+' and '%' only appear at the beginning of a constraint.
1087 * ira.c (commutative_constraint_p): Delete.
1088 (ira_get_dup_out_num): Expect the '%' commutativity marker to be
1089 at the start of the string.
1090 * config/alpha/alpha.md (*movmemdi_1, *clrmemdi_1): Remove
1092 * config/arm/neon.md (bicdi3_neon): Likewise.
1093 * config/iq2000/iq2000.md (addsi3_internal, subsi3_internal, sgt_si)
1094 (slt_si, sltu_si): Likewise.
1095 * config/vax/vax.md (sbcdi3): Likewise.
1096 * config/h8300/h8300.md (*cmpstz): Remove duplicate '+'.
1097 * config/arc/arc.md (mulsi_600, mulsidi_600, umulsidi_600)
1098 (mul64): Move '%' to beginning of constraint.
1099 * config/arm/arm.md (*xordi3_insn): Likewise.
1100 * config/nds32/nds32.md (add<mode>3, mulsi3, andsi3, iorsi3)
1103 2014-05-28 Richard Sandiford <rdsandiford@googlemail.com>
1105 * doc/md.texi: Document the restrictions on the "enabled" attribute.
1107 2014-05-28 Jason Merrill <jason@redhat.com>
1110 * cgraph.h (symtab_node::get_comdat_group_id): New.
1111 * cgraphunit.c (analyze_functions): Call it.
1112 * symtab.c (dump_symtab_node): Likewise.
1113 * tree.c (decl_comdat_group_id): New.
1114 * tree.h: Declare it.
1115 * lto-streamer-out.c (write_symbol): Use it.
1116 * trans-mem.c (ipa_tm_create_version_alias): Likewise.
1118 2014-05-28 Francois-Xavier Coudert <fxcoudert@gcc.gnu.org>
1120 PR bootstrap/PR61146
1121 * wide-int.cc: Do not include longlong.h when compiling with clang.
1123 2014-05-28 Richard Biener <rguenther@suse.de>
1125 * tree-ssa-propagate.c (add_control_edge): Print less vertical space.
1126 * tree-vrp.c (extract_range_from_ssa_name): Also copy VR_UNDEFINED.
1127 (vrp_visit_assignment_or_call): Print less vertical space.
1128 (vrp_visit_stmt): Likewise.
1129 (vrp_visit_phi_node): Likewise. For a PHI argument with
1130 VR_VARYING range consider recording it as copy.
1132 2014-05-28 Richard Biener <rguenther@suse.de>
1135 2014-05-28 Richard Biener <rguenther@suse.de>
1137 * hwint.h (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
1139 2014-05-28 Bernd Edlinger <bernd.edlinger@hotmail.de>
1141 * expr.c (expand_assignment): Fold the bitpos in the to_rtx if
1142 sufficiently aligned and an offset is used at the same time.
1143 (expand_expr_real_1): Likewise.
1145 2014-05-28 Richard Biener <rguenther@suse.de>
1148 * fold-const.c (fold_comparison): When folding
1149 X +- C1 CMP Y +- C2 to X CMP Y +- C2 +- C1 also ensure
1150 the sign of the remaining constant operand stays the same.
1152 2014-05-28 Kaushik Phatak <kaushik.phatak@kpit.com>
1154 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
1155 __RL78_64BIT_DOUBLES__ or __RL78_32BIT_DOUBLES__.
1156 (ASM_SPEC): Pass -m64bit-doubles or -m32bit-doubles on
1158 (DOUBLE_TYPE_SIZE): Use 64 bit if TARGET_64BIT_DOUBLES is true.
1159 * gcc/config/rl78/rl78.opt (m64bit-doubles): New option.
1160 (m32bit-doubles) Likewise.
1161 * gcc/config/rl78/t-rl78: Add 64-bit-double multilib.
1162 * doc/invoke.texi: Document -m32bit-doubles and -m64bit-doubles
1165 2014-05-28 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1167 * configure.ac ($gcc_cv_ld_clearcap): New test.
1168 * configure: Regenerate.
1169 * config.in: Regenerate.
1170 * config/sol2.opt (mclear-hwcap): New option.
1171 * config/sol2.h (LINK_CLEARCAP_SPEC): Define.
1172 * config/sol2-clearcap.map: Moved here from
1173 testsuite/gcc.target/i386/clearcap.map.
1174 * config/sol2-clearcapv2.map: Move here from
1175 gcc.target/i386/clearcapv2.map.
1176 * config/t-sol2 (install): Depend on install-clearcap-map.
1177 (install-clearcap-map): New target.
1178 * doc/invoke.texi (Option Summary, Solaris 2 Options): Document
1181 2014-05-28 Richard Biener <rguenther@suse.de>
1183 * hwint.h (*_HALF_WIDE_INT*): Move to ...
1184 * wide-int.cc (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT):
1185 ... here and remove the rest.
1186 * hwint.h (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
1188 2014-05-28 Richard Biener <rguenther@suse.de>
1190 PR tree-optimization/61335
1191 * tree-vrp.c (vrp_visit_phi_node): If the compare of old and
1192 new range fails, drop to varying.
1194 2014-05-28 Tom de Vries <tom@codesourcery.com>
1196 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
1197 * lra.c (initialize_lra_reg_info_element): Add init of
1198 actual_call_used_reg_set field.
1199 (lra): Call lra_create_live_ranges before lra_inheritance for
1201 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
1203 * lra-constraints.c (need_for_call_save_p): Use
1204 actual_call_used_reg_set instead of call_used_reg_set for
1206 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
1208 2014-05-28 Radovan Obradovic <robradovic@mips.com>
1209 Tom de Vries <tom@codesourcery.com>
1211 * doc/invoke.texi (@item Optimization Options): Add -fuse-caller-save
1213 (@item -fuse-caller-save): New item.
1215 2014-05-28 Radovan Obradovic <robradovic@mips.com>
1216 Tom de Vries <tom@codesourcery.com>
1218 * opts.c (default_options_table): Add OPT_LEVELS_2_PLUS entry with
1219 OPT_fuse_caller_save.
1221 2014-05-28 Radovan Obradovic <robradovic@mips.com>
1222 Tom de Vries <tom@codesourcery.com>
1224 * df-scan.c (df_get_call_refs): Use get_call_reg_set_usage.
1225 * caller-save.c (setup_save_areas, save_call_clobbered_regs): Use
1226 get_call_reg_set_usage.
1227 * resource.c (mark_set_resources, mark_target_live_regs): Use
1228 get_call_reg_set_usage.
1229 * ira-int.h (struct ira_allocno): Add crossed_calls_clobbered_regs
1231 (ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS): Define.
1232 * ira-lives.c (process_bb_node_lives): Use get_call_reg_set_usage.
1233 Calculate ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
1234 * ira-build.c (ira_create_allocno): Init
1235 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
1236 (create_cap_allocno, propagate_allocno_info)
1237 (propagate_some_info_from_allocno)
1238 (copy_info_to_removed_store_destinations): Handle
1239 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
1240 * ira-costs.c (ira_tune_allocno_costs): Use
1241 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS to adjust costs.
1243 2014-05-28 Radovan Obradovic <robradovic@mips.com>
1244 Tom de Vries <tom@codesourcery.com>
1246 * cgraph.h (struct cgraph_rtl_info): Add function_used_regs
1247 and function_used_regs_valid fields.
1248 * final.c: Move include of hard-reg-set.h to before rtl.h to declare
1249 find_all_hard_reg_sets.
1250 (collect_fn_hard_reg_usage, get_call_fndecl, get_call_cgraph_rtl_info)
1251 (get_call_reg_set_usage): New function.
1252 (rest_of_handle_final): Use collect_fn_hard_reg_usage.
1253 * regs.h (get_call_reg_set_usage): Declare.
1255 2014-05-28 Georg-Johann Lay <avr@gjlay.de>
1258 * config/dbx.h (License): Add Runtime Library Exception.
1259 * config/newlib-stdint.h (License): Same.
1260 * config/rtems.h (License): Same
1261 * config/initfini-array.h (License): Same
1262 * config/v850/v850.h (License): Same.
1263 * config/v850/v850-opts.h (License): Same
1264 * config/v850/rtems.h (License): Same.
1266 2014-05-28 Georg-Johann Lay <avr@gjlay.de>
1269 * doc/extend.texi (Local Labels): Note that label differences are
1270 not supported for AVR.
1272 2014-05-27 Andrew Pinski <apinski@cavium.com>
1274 * config/aarch64/aarch64.md (stack_protect_set_<mode>):
1275 Use <w> for the register in assembly template.
1276 (stack_protect_test): Use the mode of operands[0] for the result.
1277 (stack_protect_test_<mode>): Use <w> for the register
1278 in assembly template.
1280 2014-05-27 DJ Delorie <dj@redhat.com>
1282 * config/rx/rx.c (add_vector_labels): New.
1283 (rx_output_function_prologue): Call it.
1284 (rx_handle_func_attribute): Don't require empty arguments.
1285 (rx_handle_vector_attribute): New.
1286 (rx_attribute_table): Add "vector" attribute.
1287 * doc/extend.texi (interrupt, vector): Document new/changed
1288 RX-specific attributes.
1290 * config/rx/rx.c (rx_adjust_insn_length): Skip for non-insns.
1292 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
1294 * double-int.c (div_and_round_double) <ROUND_DIV_EXPR>: Use the proper
1295 predicate to detect a negative quotient.
1297 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
1299 * fold-const.c (fold_comparison): Clean up and extend X +- C1 CMP C2
1300 to X CMP C2 -+ C1 transformation to EQ_EXPR/NE_EXPR.
1301 Add X - Y CMP 0 to X CMP Y transformation.
1302 (fold_binary_loc) <EQ_EXPR/NE_EXPR>: Remove same transformations.
1304 2014-05-27 Segher Boessenkool <segher@kernel.crashing.org>
1306 * stmt.c (dump_case_nodes): Don't convert values to HOST_WIDE_INT
1309 2014-05-27 Steve Ellcey <sellcey@mips.com>
1311 * config/mips/mips.c: Add include of cgraph.h.
1313 2014-05-27 Richard Biener <rguenther@suse.de>
1315 * system.h (__STDC_FORMAT_MACROS): Define as very first thing.
1317 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
1320 * config/arm/arm.h (License): Add note to COPYING.RUNTIME.
1321 * config/arm/arm-cores.def (License): Same.
1322 * config/arm/arm-opts.h (License): Same.
1323 * config/arm/aout.h (License): Same.
1324 * config/arm/bpabi.h (License): Same.
1325 * config/arm/elf.h (License): Same.
1326 * config/arm/linux-elf.h (License): Same.
1327 * config/arm/linux-gas.h (License): Same.
1328 * config/arm/netbsd-elf.h (License): Same.
1329 * config/arm/uclinux-eabi.h (License): Same.
1330 * config/arm/uclinux-elf.h (License): Same.
1331 * config/arm/vxworks.h (License): Same.
1333 2014-05-27 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1335 * config/arm/neon.md (neon_bswap<mode>): New pattern.
1336 * config/arm/arm.c (neon_itype): Add NEON_BSWAP.
1337 (arm_init_neon_builtins): Handle NEON_BSWAP.
1338 Define required type nodes.
1339 (arm_expand_neon_builtin): Handle NEON_BSWAP.
1340 (arm_builtin_vectorized_function): Handle BUILTIN_BSWAP builtins.
1341 * config/arm/arm_neon_builtins.def (bswap): Define builtins.
1342 * config/arm/iterators.md (VDQHSD): New mode iterator.
1344 2014-05-27 Richard Biener <rguenther@suse.de>
1346 * tree-vrp.c (vrp_evaluate_conditional_warnv_with_ops_using_ranges):
1347 Try using literal operands when comparing value-ranges failed.
1349 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
1351 * ira.c (commutative_operand): Adjust for change to recog_data.
1352 [Missing from previous commit.]
1354 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
1356 * system.h (TEST_BIT): New macro.
1357 * recog.h (alternative_mask): New type.
1358 (ALL_ALTERNATIVES, ALTERNATIVE_BIT): New macros.
1359 (recog_data_d): Replace alternative_enabled_p array with
1360 enabled_alternatives.
1361 (target_recog): New structure.
1362 (default_target_recog, this_target_recog): Declare.
1363 (get_enabled_alternatives, recog_init): Likewise.
1364 * recog.c (default_target_recog, this_target_recog): New variables.
1365 (get_enabled_alternatives): New function.
1366 (extract_insn): Use it.
1367 (recog_init): New function.
1368 (preprocess_constraints, constrain_operands): Adjust for change to
1370 * postreload.c (reload_cse_simplify_operands): Likewise.
1371 * reload.c (find_reloads): Likewise.
1372 * ira-costs.c (record_reg_classes): Likewise.
1373 * ira-lives.c (single_reg_class): Likewise. Fix bug in which
1374 all alternatives after a disabled one would be skipped.
1375 (ira_implicitly_set_insn_hard_regs): Likewise.
1376 * ira.c (ira_setup_alts): Adjust for change to recog_data.
1377 * lra-int.h (lra_insn_recog_data): Replace alternative_enabled_p
1378 with enabled_alternatives.
1379 * lra.c (free_insn_recog_data): Update accordingly.
1380 (lra_update_insn_recog_data): Likewise.
1381 (lra_set_insn_recog_data): Likewise. Use get_enabled_alternatives.
1382 * lra-constraints.c (process_alt_operands): Likewise. Handle
1383 only_alternative as part of the enabled mask.
1384 * target-globals.h (this_target_recog): Declare.
1385 (target_globals): Add a recog field.
1386 (restore_target_globals): Restore this_target_recog.
1387 * target-globals.c: Include recog.h.
1388 (default_target_globals): Initialize recog field.
1389 (save_target_globals): Likewise.
1390 * reginfo.c (reinit_regs): Call recog_init.
1391 * toplev.c (backend_init_target): Likewise.
1393 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
1395 * gencodes.c (main): Make LAST_INSN_CODE higher than any insn code,
1396 rather than any named insn's code.
1398 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
1401 * config/arm/arm-opts.h (License): Add GCC Runtime Library Exception.
1402 * config/arm/arm-cores.def (License): Same.
1404 2014-05-26 Jan Hubicka <hubicka@ucw.cz>
1406 * tree.h (decl_comdat_group): Declare.
1407 * cgraph.h (symtab_in_same_comdat_p): Move offline to ...
1408 * tree.c (decl_comdat_group): Here.
1410 2014-05-26 Richard Sandiford <r.sandiford@uk.ibm.com>
1412 PR rtl-optimization/61222
1413 * combine.c (simplify_shift_const_1): When moving a PLUS outside
1414 the shift, truncate the PLUS operand to the result mode.
1416 2014-05-26 Uros Bizjak <ubizjak@gmail.com>
1419 * config/i386/i386.c (ix86_rtx_costs)
1420 <case CONST_INT, case CONST, case LABEL_REF, case SYMBOL_REF>:
1423 2014-05-26 Martin Jambor <mjambor@suse.cz>
1425 * ira.c (split_live_ranges_for_shrink_wrap): Remove bailout on
1428 2014-05-26 Richard Biener <rguenther@suse.de>
1430 * wide-int.h (wi::int_traits <long>, wi::int_traits <unsigned long>,
1431 wi::int_traits <long long>, wi::int_traits <unsigned long long>):
1432 Provide specializations.
1433 (wi::int_traits <HOST_WIDE_INT>,
1434 wi::int_traits <unsigned HOST_WIDE_INT>): Remove specializations.
1436 2014-05-26 Alan Modra <amodra@gmail.com>
1439 * config/rs6000/rs6000.c (rs6000_emit_set_const): Remove unneeded
1440 params and return a bool. Remove dead code. Update comment.
1441 Assert we have a const_int source. Remove bogus code from
1442 32-bit HWI days. Move !TARGET_POWERPC64 handling, and correct
1443 handling of constants > 2G and reg_equal note, from..
1444 (rs6000_emit_set_long_const): ..here. Remove unneeded param and
1445 return value. Update comment. If we can, use a new pseudo
1446 for intermediate calculations.
1447 * config/rs6000/rs6000-protos.h (rs6000_emit_set_const): Update
1449 * config/rs6000/rs6000.md (movsi_internal1_single+1): Update
1450 call to rs6000_emit_set_const in splitter.
1451 (movdi_internal64+2, +3): Likewise.
1453 2014-05-26 Richard Biener <rguenther@suse.de>
1455 * system.h: Define __STDC_FORMAT_MACROS before
1456 including inttypes.h.
1457 * hwint.h (HOST_WIDEST_INT, HOST_BITS_PER_WIDEST_INT,
1458 HOST_WIDEST_INT_PRINT, HOST_WIDEST_INT_PRINT_DEC,
1459 HOST_WIDEST_INT_PRINT_DEC_C, HOST_WIDEST_INT_PRINT_UNSIGNED,
1460 HOST_WIDEST_INT_PRINT_HEX, HOST_WIDEST_INT_PRINT_DOUBLE_HEX,
1461 HOST_WIDEST_INT_C): Remove.
1462 (PRId64, PRIi64, PRIo64, PRIu64, PRIx64, PRIX64): Define
1463 if C99 inttypes.h is not available.
1464 * coretypes.h (gcov_type, gcov_type_unsigned): Use [u]int64_t.
1465 * gcov-io.h (gcov_type, gcov_type_unsigned): Likewise.
1466 * gcov-io.c (gcov_histo_index): Drop non-64bit hwi case.
1467 * cfgloop.h (struct niter_desc): Use uint64_t for niter field.
1468 * bitmap.c (struct bitmap_descriptor_d): Use uint64_t for counters.
1469 (struct output_info): Likewise.
1470 (print_statistics): Adjust.
1471 (dump_bitmap_statistics): Likewise.
1472 * bt-load.c (migrate_btr_defs): Print with PRId64.
1473 * cfg.c (dump_edge_info, dump_bb_info): Likewise.
1474 (MAX_SAFE_MULTIPLIER): Adjust.
1475 * cfghooks.c (dump_bb_for_graph): Print with PRId64.
1476 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee,
1477 dump_cgraph_node): Likewise.
1478 * final.c (dump_basic_block_info): Likewise.
1479 * gcov-dump.c (tag_counters, tag_summary, dump_working_sets): Likewise.
1480 * gcov.c (format_gcov): Likewise.
1481 * ipa-cp.c (good_cloning_opportunity_p): Likewise. Use int64_t
1483 (get_clone_agg_value): Use HOST_WIDE_INT for offset.
1484 * ipa-inline.c (compute_max_insns): Use int64_t for calcuation.
1485 (inline_small_functions, dump_overall_stats, dump_inline_stats):
1486 Use PRId64 for dumping.
1487 * ipa-profile.c (dump_histogram, ipa_profile): Likewise.
1488 * ira-color.c (struct allocno_hard_regs): Use int64_t for cost.
1489 (add_allocno_hard_regs): Adjust.
1490 * loop-doloop.c (doloop_modify): Print using PRId64.
1491 * loop-iv.c (inverse): Compute in uint64_t.
1492 (determine_max_iter, iv_number_of_iterations): Likewise.
1493 * loop-unroll.c (decide_peel_completely, decide_peel_simple):
1495 * lto-streamer-out.c (write_symbol): Use uint64_t.
1496 * mcf.c (CAP_INFINITY): Use int64_t maximum.
1497 (dump_fixup_edge, create_fixup_graph, cancel_negative_cycle,
1498 find_max_flow, adjust_cfg_counts): Use int64_t and dump with PRId64.
1499 * modulo-sched.c (const_iteration_count): Use int64_t.
1500 (sms_schedule): Dump using PRId64.
1501 * predict.c (dump_prediction): Likewise.
1502 * pretty-print.h (pp_widest_integer): Remove.
1503 * profile.c (get_working_sets, is_edge_inconsistent,
1504 is_inconsistent, read_profile_edge_counts): Dump using PRId64.
1505 * tree-pretty-print.c (pp_double_int): Remove case handling
1506 HOST_BITS_PER_DOUBLE_INT == HOST_BITS_PER_WIDEST_INT.
1507 * tree-ssa-math-opts.c (struct symbolic_number): Use uint64_t
1509 (pass_optimize_bswap::execute): Remove restriction on hosts.
1510 * tree-streamer-in.c (streamer_alloc_tree): Use HOST_WIDE_INT.
1511 * tree-streamer-out.c (streamer_write_tree_header): Likewise.
1512 * tree.c (widest_int_cst_value): Remove.
1513 * tree.h (widest_int_cst_value): Likewise.
1514 * value-prof.c (dump_histogram_value): Print using PRId64.
1515 * gengtype.c (main): Also inject int64_t.
1516 * ggc-page.c (struct max_alignment): Use int64_t.
1517 * alloc-pool.c (struct allocation_object_def): Likewise.
1518 * ira-conflicts.c (build_conflict_bit_table): Use uint64_t
1520 * doc/tm.texi.in: Remove reference to HOST_WIDEST_INT.
1521 * doc/tm.texi: Regenerated.
1522 * gengtype-lex.l (IWORD): Handle [u]int64_t.
1523 * config/sh/sh.c (expand_cbranchdi4): Use gcov_type.
1524 * config/mmix/mmix-protos.h (mmix_intval, mmix_shiftable_wyde_value,
1525 mmix_output_register_setting): Use [u]int64_t in prototypes.
1526 * config/mmix/mmix.c (mmix_print_operand, mmix_output_register_setting,
1527 mmix_shiftable_wyde_value, mmix_output_shiftvalue_op_from_str,
1528 mmix_output_octa, mmix_output_shifted_value): Adjust.
1529 (mmix_intval): Adjust. Remove unreachable case.
1530 * config/mmix/mmix.md (*nonlocal_goto_receiver_expanded): Use int64_t.
1532 2014-05-26 Richard Biener <rguenther@suse.de>
1534 * configure.ac: Drop __int64 type check. Insist that we
1535 found uint64_t and int64_t.
1536 * hwint.h (HOST_BITS_PER___INT64): Remove.
1537 (HOST_BITS_PER_WIDE_INT): Define to 64 and remove __int64 case.
1538 (HOST_WIDE_INT_PRINT_*): Remove 32bit case.
1539 (HOST_WIDEST_INT*): Define to HOST_WIDE_INT*.
1540 (HOST_WIDEST_FAST_INT): Remove __int64 case.
1541 * vmsdbg.h (struct _DST_SRC_COMMAND): Use int64_t
1542 for dst_q_src_df_rms_cdt.
1543 * configure: Regenerate.
1544 * config.in: Likewise.
1546 2014-05-26 Michael Tautschnig <mt@debian.org>
1549 * doc/extend.texi (X86 Built-in Functions): Fix parameter lists of
1550 __builtin_ia32_vfrczs[sd] and __builtin_ia32_mpsadbw256.
1552 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1554 PR rtl-optimization/61278
1555 * shrink-wrap.c (move_insn_for_shrink_wrap): Check df_live.
1557 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1559 PR rtl-optimization/61220
1560 Part of PR rtl-optimization/61225
1561 * shrink-wrap.c (move_insn_for_shrink_wrap): Skip SP and FP adjustment
1562 insn; skip split_edge for a block with only one successor.
1564 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1566 * symtab.c (symtab_nonoverwritable_alias): Copy READONLY flag
1569 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1571 * ipa-visibility.c (can_replace_by_local_alias_in_vtable): New function.
1572 (update_vtable_references): New function.
1573 (function_and_variable_visibility): Rewrite also vtable initializers.
1574 * varpool.c (cgraph_variable_initializer_availability): Remove assert.
1576 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1578 * ggc.h (ggc_grow): New function.
1579 * ggc-none.c (ggc_grow): New function.
1580 * ggc-page.c (ggc_grow): Likewise.
1582 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1584 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
1585 address_taken_from_non_vtable_p, comdat_can_be_unshared_p_1,
1586 comdat_can_be_unshared_p, cgraph_externally_visible_p,
1587 varpool_externally_visible_p, can_replace_by_local_alias,
1588 update_visibility_by_resolution_info, function_and_variable_visibility,
1589 pass_data_ipa_function_and_variable_visibility,
1590 make_pass_ipa_function_and_variable_visibility,
1591 whole_program_function_and_variable_visibility,
1592 pass_data_ipa_whole_program_visibility,
1593 make_pass_ipa_whole_program_visibility): Move to ipa-visibility.c
1594 * cgraph.h (cgraph_local_node_p): Declare.
1595 * ipa-visibility.c: New file.
1596 * Makefile.in (OBJS): Add ipa-visiblity.o
1598 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1600 * gimple-fold.c (can_refer_decl_in_current_unit_p): Be sure
1601 that var decl is available.
1603 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1605 * tree-core.h (tree_decl_with_vis): Replace comdat_group by
1606 symtab_node pointer.
1607 * tree.c (copy_node_stat): Be sure to not copy symtab_node pointer.
1608 (find_decls_types_r): Do not walk COMDAT_GROUP.
1609 * tree.h (DECL_COMDAT_GROUP): Revamp to use decl_comdat_group.
1610 * varasm.c (make_decl_one_only): Use set_comdat_group;
1611 create node if needed.
1612 * ipa-inline-transform.c (save_inline_function_body): Update
1613 way we decl->symtab mapping.
1614 * symtab.c (symtab_hash, hash_node, eq_node
1615 symtab_insert_node_to_hashtable): Remove.
1616 (symtab_register_node): Update.
1617 (symtab_unregister_node): Update.
1618 (symtab_get_node): Reimplement as inline function.
1619 (symtab_add_to_same_comdat_group): Update.
1620 (symtab_dissolve_same_comdat_group_list): Update.
1621 (dump_symtab_base): Update.
1622 (verify_symtab_base): Update.
1623 (symtab_make_decl_local): Update.
1624 (fixup_same_cpp_alias_visibility): Update.
1625 (symtab_nonoverwritable_alias): Update.
1626 * cgraphclones.c (set_new_clone_decl_and_node_flags): Update.
1627 * ipa.c (update_visibility_by_resolution_info): UPdate.
1628 * bb-reorder.c: Include cgraph.h
1629 * lto-streamer-out.c (DFS_write_tree_body, hash_tree): Do not deal
1631 * ipa-comdats.c (set_comdat_group, ipa_comdats): Update.
1632 * cgraph.c (cgraph_get_create_node): Update.
1633 * cgraph.h (struct symtab_node): Add get_comdat_group, set_comdat_group
1635 (symtab_get_node): Make inline.
1636 (symtab_insert_node_to_hashtable): Remove.
1637 (symtab_can_be_discarded): Update.
1638 (decl_comdat_group): New function.
1639 * tree-streamer-in.c (lto_input_ts_decl_with_vis_tree_pointers):
1641 * lto-cgraph.c (lto_output_node, lto_output_varpool_node): Stream out
1643 (read_comdat_group): New function.
1644 (input_node, input_varpool_node): Use it.
1645 * trans-mem.c (ipa_tm_create_version_alias): Update code creating
1647 * mips.c (mips_start_unique_function): Likewise.
1648 (ix86_code_end): Likewise.
1649 (rs6000_code_end): Likweise.
1650 * tree-streamer-out.c (DECL_COMDAT_GROUP): Do not stream comdat group.
1652 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1654 * gengtype-state.c (fatal_reading_state): Bring offline.
1655 * optabs.c (widening_optab_handler): Bring offline.
1656 * optabs.h (widening_optab_handler): Likewise.
1657 * final.c (get_attr_length_1): Likewise.
1659 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1661 * sched-int.h (sd_iterator_cond): Manually tail recurse.
1663 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1665 * config/rs6000/440.md (ppc440-integer): Include shift without dot.
1666 (ppc440-compare): Include shift with dot.
1667 * config/rs6000/e300c2c3.md (ppce300c3_iu): Include shift without dot.
1668 * config/rs6000/e5500.md (e5500_sfx2): Include constant shift
1670 * config/rs6000/e6500.md (e6500_sfx): Exclude constant shift
1672 (e6500_sfx2): Include it.
1673 * config/rs6000/rs6000.md ( *zero_extend<mode>di2_internal1,
1674 *zero_extend<mode>di2_internal2, *zero_extend<mode>di2_internal3,
1675 *zero_extendsidi2_lfiwzx, andsi3_mc, andsi3_nomc,
1676 andsi3_internal0_nomc, extzvsi_internal, extzvdi_internal,
1677 *extzvdi_internal1, *extzvdi_internal2, rotlsi3, *rotlsi3_64,
1678 *rotlsi3_internal4, *rotlsi3_internal7le, *rotlsi3_internal7be,
1679 *rotlsi3_internal10le, *rotlsi3_internal10be, rlwinm,
1680 *lshiftrt_internal1le, *lshiftrt_internal1be,
1681 *lshiftrt_internal4le, *lshiftrt_internal4be, rotldi3,
1682 *rotldi3_internal4, *rotldi3_internal7le, *rotldi3_internal7be,
1683 *rotldi3_internal10le, *rotldi3_internal10be,
1684 *rotldi3_internal13le, *rotldi3_internal13be, *ashldi3_internal4,
1685 ashldi3_internal5, *ashldi3_internal6, *ashldi3_internal7,
1686 ashldi3_internal8, *ashldi3_internal9, anddi3_mc, anddi3_nomc,
1687 *anddi3_internal2_mc, *anddi3_internal3_mc, and 4 anonymous
1688 define_insns): Use type "shift" in the appropriate alternatives.
1690 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1692 * config/rs6000/rs6000.md (type): Add "logical". Delete
1694 (dot): Adjust comment.
1695 (andsi3_mc, *andsi3_internal2_mc, *andsi3_internal3_mc,
1696 *andsi3_internal4, *andsi3_internal5_mc, *boolsi3_internal2,
1697 *boolsi3_internal3, *boolccsi3_internal2, *boolccsi3_internal3,
1698 anddi3_mc, *anddi3_internal2_mc, *anddi3_internal3_mc,
1699 *booldi3_internal2, *booldi3_internal3, *boolcdi3_internal2,
1700 *boolcdi3_internal3, *boolccdi3_internal2, *boolccdi3_internal3,
1701 *mov<mode>_internal2, and 10 anonymous define_insns): Use "logical".
1702 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
1704 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
1705 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
1706 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
1707 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
1708 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
1709 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
1710 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
1711 * config/rs6000/8540.md (ppc8540_su): Adjust.
1712 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
1713 cell-cmp-microcoded): Adjust.
1714 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
1715 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1716 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
1717 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
1718 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
1719 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
1720 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
1721 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
1722 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
1724 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
1725 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare):
1726 Adjust. Adjust comment.
1727 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
1728 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
1730 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1732 * config/rs6000/rs6000.md (type): Add "add".
1733 (*add<mode>3_internal1, addsi3_high, *add<mode>3_internal2,
1734 *add<mode>3_internal3, *neg<mode>2_internal, and 5 anonymous
1735 define_insns): Use it.
1736 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
1738 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
1739 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
1740 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
1741 * config/rs6000/601.md (ppc601-integer): Adjust.
1742 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
1743 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
1744 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
1745 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
1746 * config/rs6000/8540.md (ppc8540_su): Adjust.
1747 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
1748 cell-cmp-microcoded): Adjust.
1749 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
1750 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1751 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
1752 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
1753 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
1754 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
1755 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
1756 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
1757 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
1759 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
1760 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare): Adjust.
1761 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
1762 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
1764 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1766 * config/rs6000/rs6000.md (type): Delete "var_shift_rotate",
1767 "delayed_compare", "var_delayed_compare".
1768 (var_shift): New attribute.
1769 (cell_micro): Adjust.
1770 (*andsi3_internal2_mc, *andsi3_internal3_mc, *andsi3_internal4,
1771 *andsi3_internal5_mc, *extzvsi_internal1, *extzvsi_internal2,
1772 rotlsi3, *rotlsi3_64, *rotlsi3_internal2, *rotlsi3_internal3,
1773 *rotlsi3_internal4, *rotlsi3_internal5, *rotlsi3_internal6,
1774 *rotlsi3_internal8le, *rotlsi3_internal8be, *rotlsi3_internal9le,
1775 *rotlsi3_internal9be, *rotlsi3_internal10le, *rotlsi3_internal10be,
1776 *rotlsi3_internal11le, *rotlsi3_internal11be, *rotlsi3_internal12le,
1777 *rotlsi3_internal12be, ashlsi3, *ashlsi3_64, lshrsi3, *lshrsi3_64,
1778 *lshiftrt_internal2le, *lshiftrt_internal2be, *lshiftrt_internal3le,
1779 *lshiftrt_internal3be, *lshiftrt_internal5le, *lshiftrt_internal5be,
1780 *lshiftrt_internal5le, *lshiftrt_internal5be, ashrsi3, *ashrsi3_64,
1781 rotldi3, *rotldi3_internal2, *rotldi3_internal3, *rotldi3_internal4,
1782 *rotldi3_internal5, *rotldi3_internal6, *rotldi3_internal7le,
1783 *rotldi3_internal7be, *rotldi3_internal8le, *rotldi3_internal8be,
1784 *rotldi3_internal9le, *rotldi3_internal9be, *rotldi3_internal10le,
1785 *rotldi3_internal10be, *rotldi3_internal11le, *rotldi3_internal11be,
1786 *rotldi3_internal12le, *rotldi3_internal12be, *rotldi3_internal13le,
1787 *rotldi3_internal13be, *rotldi3_internal14le, *rotldi3_internal14be,
1788 *rotldi3_internal15le, *rotldi3_internal15be, *ashldi3_internal1,
1789 *ashldi3_internal2, *ashldi3_internal3, *lshrdi3_internal1,
1790 *lshrdi3_internal2, *lshrdi3_internal3, *ashrdi3_internal1,
1791 *ashrdi3_internal2, *ashrdi3_internal3, *anddi3_internal2_mc,
1792 *anddi3_internal3_mc, as well as 11 anonymous define_insns): Adjust.
1793 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1794 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
1796 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
1797 * config/rs6000/440.md (ppc440-integer): Adjust.
1798 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
1799 * config/rs6000/601.md (ppc601-integer, ppc601-compare): Adjust.
1800 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
1801 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
1802 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
1803 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
1804 * config/rs6000/8540.md (ppc8540_su): Adjust.
1805 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
1806 cell-cmp-microcoded): Adjust.
1807 * config/rs6000/e300c2c3.md (ppce300c3_cmp): Adjust.
1808 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1809 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2,
1810 e500mc64_delayed): Adjust.
1811 * config/rs6000/e5500.md (e5500_sfx, e5500_delayed): Adjust.
1812 * config/rs6000/e6500.md (e6500_sfx, e6500_delayed): Adjust.
1813 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
1814 * config/rs6000/power4.md (power4-integer, power4-compare): Adjust.
1815 * config/rs6000/power5.md (power5-integer, power5-compare): Adjust.
1816 * config/rs6000/power6.md (power6-shift, power6-var-rotate,
1817 power6-delayed-compare, power6-var-delayed-compare): Adjust.
1818 * config/rs6000/power7.md (power7-integer, power7-compare): Adjust.
1819 * config/rs6000/power8.md (power8-1cyc, power8-compare): Adjust.
1821 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
1822 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
1824 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1826 * config/rs6000/rs6000.md (type): Delete "idiv", "ldiv". Add "div".
1827 (bits): New mode_attr.
1828 (idiv_ldiv): Delete mode_attr.
1829 (udiv<mode>3, *div<mode>3, div<div_extend>_<mode>): Adjust.
1830 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1831 rs6000_adjust_priority, is_nonpipeline_insn,
1832 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
1834 * config/rs6000/40x.md (ppc403-idiv): Adjust.
1835 * config/rs6000/440.md (ppc440-idiv): Adjust.
1836 * config/rs6000/476.md (ppc476-idiv): Adjust.
1837 * config/rs6000/601.md (ppc601-idiv): Adjust.
1838 * config/rs6000/603.md (ppc603-idiv): Adjust.
1839 * config/rs6000/6xx.md (ppc604-idiv, ppc620-idiv, ppc630-idiv,
1840 ppc620-ldiv): Adjust.
1841 * config/rs6000/7450.md (ppc7450-idiv): Adjust.
1842 * config/rs6000/7xx.md (ppc750-idiv): Adjust.
1843 * config/rs6000/8540.md (ppc8540_divide): Adjust.
1844 * config/rs6000/a2.md (ppca2-idiv, ppca2-ldiv): Adjust.
1845 * config/rs6000/cell.md (cell-idiv, cell-ldiv): Adjust.
1846 * config/rs6000/e300c2c3.md (ppce300c3_divide): Adjust.
1847 * config/rs6000/e500mc.md (e500mc_divide): Adjust.
1848 * config/rs6000/e500mc64.md (e500mc64_divide): Adjust.
1849 * config/rs6000/e5500.md (e5500_divide, e5500_divide_d): Adjust.
1850 * config/rs6000/e6500.md (e6500_divide, e6500_divide_d): Adjust.
1851 * config/rs6000/mpc.md (mpccore-idiv): Adjust.
1852 * config/rs6000/power4.md (power4-idiv, power4-ldiv): Adjust.
1853 * config/rs6000/power5.md (power5-idiv, power5-ldiv): Adjust.
1854 * config/rs6000/power6.md (power6-idiv, power6-ldiv): Adjust.
1855 * config/rs6000/power7.md (power7-idiv, power7-ldiv): Adjust.
1856 * config/rs6000/power8.md (power8-idiv, power8-ldiv): Adjust.
1857 * config/rs6000/rs64.md (rs64a-idiv, rs64a-ldiv): Adjust.
1858 * config/rs6000/titan.md (titan_fxu_div): Adjust.
1860 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1862 * config/rs6000/rs6000.md (type): Delete "insert_word",
1863 "insert_dword". Add "insert".
1864 (size): Update comment.
1865 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1866 insn_must_be_first_in_group): Adjust.
1867 (insvsi_internal, *insvsi_internal1, *insvsi_internal2,
1868 *insvsi_internal3, *insvsi_internal4, *insvsi_internal5,
1869 *insvsi_internal6, insvdi_internal): Adjust.
1871 * config/rs6000/40x.md (ppc403-integer): Adjust.
1872 * config/rs6000/440.md (ppc440-integer): Adjust.
1873 * config/rs6000/476.md (ppc476-simple-integer): Adjust.
1874 * config/rs6000/601.md (ppc601-integer): Adjust.
1875 * config/rs6000/603.md (ppc603-integer): Adjust.
1876 * config/rs6000/6xx.md (ppc604-integer): Adjust.
1877 * config/rs6000/7450.md (ppc7450-integer): Adjust.
1878 * config/rs6000/7xx.md (ppc750-integer): Adjust.
1879 * config/rs6000/8540.md (ppc8540_su): Adjust.
1880 * config/rs6000/cell.md (cell-integer, cell-insert): Adjust.
1881 * config/rs6000/e300c2c3.md (ppce300c3_iu): Adjust.
1882 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1883 * config/rs6000/e500mc64.md (e500mc64_su): Adjust.
1884 * config/rs6000/e5500.md (e5500_sfx): Adjust.
1885 * config/rs6000/e6500.md (e6500_sfx): Adjust.
1886 * config/rs6000/mpc.md (mpccore-integer): Adjust.
1887 * config/rs6000/power4.md (power4-integer, power4-insert): Adjust.
1888 * config/rs6000/power5.md (power5-integer, power5-insert): Adjust.
1889 * config/rs6000/power6.md (power6-insert, power6-insert-dword): Adjust.
1890 * config/rs6000/power7.md (power7-integer): Adjust.
1891 * config/rs6000/power8.md (power8-1cyc): Adjust.
1892 * config/rs6000/rs64.md (rs64a-integer): Adjust.
1893 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
1895 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1897 * config/rs6000/rs6000.md (type): Add "mul". Delete "imul",
1898 "imul2", "imul3", "lmul", "imul_compare", "lmul_compare".
1899 (size): New attribute.
1900 (dot): New attribute.
1901 (cell_micro): Adjust.
1902 (mulsi3, *mulsi3_internal1, *mulsi3_internal2, mulsidi3,
1903 umulsidi3, smulsi3_highpart, umulsi3_highpart, muldi3,
1904 *muldi3_internal1, *muldi3_internal2, smuldi3_highpart,
1905 umuldi3_highpart): Adjust.
1906 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1907 rs6000_adjust_priority, is_nonpipeline_insn,
1908 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
1910 * config/rs6000/40x.md (ppc403-imul, ppc405-imul, ppc405-imul2,
1911 ppc405-imul3): Adjust.
1912 * config/rs6000/440.md (ppc440-imul, ppc440-imul2): Adjust.
1913 * config/rs6000/476.md (ppc476-imul): Adjust.
1914 * config/rs6000/601.md (ppc601-imul): Adjust.
1915 * config/rs6000/603.md (ppc603-imul, ppc603-imul2): Adjust.
1916 * config/rs6000/6xx.md (ppc604-imul, ppc604e-imul, ppc620-imul,
1917 ppc620-imul2, ppc620-imul3, ppc620-lmul): Adjust.
1918 * config/rs6000/7450.md (ppc7450-imul, ppc7450-imul2): Adjust.
1919 * config/rs6000/7xx.md (ppc750-imul, ppc750-imul2, ppc750-imul3):
1921 * config/rs6000/8540.md (ppc8540_multiply): Adjust.
1922 * config/rs6000/a2.md (ppca2-imul, ppca2-lmul): Adjust.
1923 * config/rs6000/cell.md (cell-lmul, cell-lmul-cmp, cell-imul23,
1925 * config/rs6000/e300c2c3.md (ppce300c3_multiply): Adjust.
1926 * config/rs6000/e500mc.md (e500mc_multiply): Adjust.
1927 * config/rs6000/e500mc64.md (e500mc64_multiply): Adjust.
1928 * config/rs6000/e5500.md (e5500_multiply, e5500_multiply_i): Adjust.
1929 * config/rs6000/e6500.md (e6500_multiply, e6500_multiply_i): Adjust.
1930 * config/rs6000/mpc.md (mpccore-imul): Adjust.
1931 * config/rs6000/power4.md (power4-lmul-cmp, power4-imul-cmp,
1932 power4-lmul, power4-imul, power4-imul3): Adjust.
1933 * config/rs6000/power5.md (power5-lmul-cmp, power5-imul-cmp,
1934 power5-lmul, power5-imul, power5-imul3): Adjust.
1935 * config/rs6000/power6.md (power6-lmul-cmp, power6-imul-cmp,
1936 power6-lmul, power6-imul, power6-imul3): Adjust.
1937 * config/rs6000/power7.md (power7-mul, power7-mul-compare): Adjust.
1938 * config/rs6000/power8.md (power8-mul, power8-mul-compare): Adjust.
1940 * config/rs6000/rs64.md (rs64a-imul, rs64a-imul2, rs64a-imul3,
1941 rs64a-lmul): Adjust.
1942 * config/rs6000/titan.md (titan_imul): Adjust.
1944 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1946 * config/rs6000/rs6000.md (type): Add new value "halfmul".
1947 (*macchwc, *macchw, *macchwuc, *macchwu, *machhwc, *machhw,
1948 *machhwuc, *machhwu, *maclhwc, *maclhw, *maclhwuc, *maclhwu,
1949 *nmacchwc, *nmacchw, *nmachhwc, *nmachhw, *nmaclhwc, *nmaclhw,
1950 *mulchwc, *mulchw, *mulchwuc, *mulchwu, *mulhhwc, *mulhhw,
1951 *mulhhwuc, *mulhhwu, *mullhwc, *mullhw, *mullhwuc, *mullhwu): Use it.
1952 * config/rs6000/40x.md (ppc405-imul3): Add type halfmul.
1953 * config/rs6000/440.md (ppc440-imul2): Add type halfmul.
1954 * config/rs6000/476.md (ppc476-imul): Add type halfmul.
1955 * config/rs6000/titan.md: Delete nonsensical comment.
1956 (titan_imul): Add type imul3.
1957 (titan_mulhw): Remove type imul3; add type halfmul.
1959 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1961 * config/rs6000/rs6000.md (type): Reorder, reformat.
1963 2014-05-23 Martin Jambor <mjambor@suse.cz>
1965 PR tree-optimization/53787
1966 * params.def (PARAM_IPA_MAX_AA_STEPS): New param.
1967 * ipa-prop.h (ipa_node_params): Rename uses_analysis_done to
1968 analysis_done, update all uses.
1969 * ipa-prop.c: Include domwalk.h
1970 (param_analysis_info): Removed.
1971 (param_aa_status): New type.
1972 (ipa_bb_info): Likewise.
1973 (func_body_info): Likewise.
1974 (ipa_get_bb_info): New function.
1975 (aa_overwalked): Likewise.
1976 (find_dominating_aa_status): Likewise.
1977 (parm_bb_aa_status_for_bb): Likewise.
1978 (parm_preserved_before_stmt_p): Changed to use new param AA info.
1979 (load_from_unmodified_param): Accept func_body_info as a parameter
1980 instead of parms_ainfo.
1981 (parm_ref_data_preserved_p): Changed to use new param AA info.
1982 (parm_ref_data_pass_through_p): Likewise.
1983 (ipa_load_from_parm_agg_1): Likewise. Update callers.
1984 (compute_complex_assign_jump_func): Changed to use new param AA info.
1985 (compute_complex_ancestor_jump_func): Likewise.
1986 (ipa_compute_jump_functions_for_edge): Likewise.
1987 (ipa_compute_jump_functions): Removed.
1988 (ipa_compute_jump_functions_for_bb): New function.
1989 (ipa_analyze_indirect_call_uses): Likewise, moved variable
1991 (ipa_analyze_virtual_call_uses): Accept func_body_info instead of node
1992 and info, moved variable declarations down.
1993 (ipa_analyze_call_uses): Accept and pass on func_body_info instead of
1995 (ipa_analyze_stmt_uses): Likewise.
1996 (ipa_analyze_params_uses): Removed.
1997 (ipa_analyze_params_uses_in_bb): New function.
1998 (ipa_analyze_controlled_uses): Likewise.
1999 (free_ipa_bb_info): Likewise.
2000 (analysis_dom_walker): New class.
2001 (ipa_analyze_node): Handle node-specific forbidden analysis,
2002 initialize and free func_body_info, use dominator walker.
2003 (ipcp_modif_dom_walker): New class.
2004 (ipcp_transform_function): Create and free func_body_info, use
2005 ipcp_modif_dom_walker, moved a lot of functionality there.
2007 2014-05-23 Marek Polacek <polacek@redhat.com>
2008 Jakub Jelinek <jakub@redhat.com>
2010 * builtins.def: Change SANITIZE_FLOAT_DIVIDE to SANITIZE_NONDEFAULT.
2011 * gcc.c (sanitize_spec_function): Likewise.
2012 * convert.c (convert_to_integer): Include "ubsan.h". Add
2013 floating-point to integer instrumentation.
2014 * doc/invoke.texi: Document -fsanitize=float-cast-overflow.
2015 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_CAST and
2016 SANITIZE_NONDEFAULT.
2017 * opts.c (common_handle_option): Handle -fsanitize=float-cast-overflow.
2018 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW,
2019 BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW_ABORT): Add.
2020 * ubsan.c: Include "realmpfr.h" and "dfp.h".
2021 (get_ubsan_type_info_for_type): Handle REAL_TYPEs.
2022 (ubsan_type_descriptor): Set tkind to 0xffff for types other than
2023 float/double/long double.
2024 (ubsan_instrument_float_cast): New function.
2025 * ubsan.h (ubsan_instrument_float_cast): Declare.
2027 2014-05-23 Jiong Wang <jiong.wang@arm.com>
2029 * config/aarch64/predicates.md (aarch64_call_insn_operand): New
2031 * config/aarch64/constraints.md ("Ucs", "Usf"): New constraints.
2032 * config/aarch64/aarch64.md (*sibcall_insn, *sibcall_value_insn):
2033 Adjust for tailcalling through registers.
2034 * config/aarch64/aarch64.h (enum reg_class): New caller save
2036 (REG_CLASS_NAMES): Likewise.
2037 (REG_CLASS_CONTENTS): Likewise.
2038 * config/aarch64/aarch64.c (aarch64_function_ok_for_sibcall):
2039 Allow tailcalling without decls.
2041 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
2043 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
2044 Rewrite check for ORT_PARALLEL and ORT_COMBINED_PARALLEL.
2046 * omp-low.c (expand_omp_for_static_chunk): Rename variable si to
2047 gsi, and variables v_* to v*.
2049 2014-05-23 Eric Botcazou <ebotcazou@adacore.com>
2051 * varasm.c (output_constructor_bitfield): Fix thinkos in latest change.
2053 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
2055 * gimple.h (enum gf_mask): Add and use GF_OMP_FOR_SIMD.
2056 * omp-low.c: Update accordingly.
2058 * gimple.h (enum gf_mask): Rewrite "<< 0" shift expressions used
2059 for GF_OMP_FOR_KIND_MASK, GF_OMP_FOR_KIND_FOR,
2060 GF_OMP_FOR_KIND_DISTRIBUTE, GF_OMP_FOR_KIND_SIMD,
2061 GF_OMP_FOR_KIND_CILKSIMD, GF_OMP_TARGET_KIND_MASK,
2062 GF_OMP_TARGET_KIND_REGION, GF_OMP_TARGET_KIND_DATA,
2063 GF_OMP_TARGET_KIND_UPDATE.
2065 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
2066 Explicitly enumerate the expected region types.
2068 2014-05-23 Paul Eggert <eggert@cs.ucla.edu>
2071 * doc/extend.texi (Function Attributes): Fix __attribute__ ((malloc))
2072 documentation; the old documentation didn't clearly state the
2073 constraints on the contents of the pointed-to storage.
2075 2014-05-23 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
2077 Fix bootstrap error on ia64
2078 * config/ia64/ia64.c (ia64_first_cycle_multipass_dfa_lookahead_guard):
2079 Return default value.
2081 2014-05-23 Thomas Preud'homme <thomas.preudhomme@arm.com>
2083 PR tree-optimization/54733
2084 * tree-ssa-math-opts.c (nop_stats): New "bswap_stats" structure.
2086 (find_bswap_or_nop_load): New.
2087 (find_bswap_1): Renamed to ...
2088 (find_bswap_or_nop_1): This. Also add support for memory source.
2089 (find_bswap): Renamed to ...
2090 (find_bswap_or_nop): This. Also add support for memory source and
2091 detection of bitwise operations equivalent to load in target
2093 (execute_optimize_bswap): Likewise. Also move its leading comment back
2094 in place and split statement transformation into ...
2095 (bswap_replace): This.
2097 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
2099 PR rtl-optimization/61215
2100 * lra-elelimination.c (lra_eliminate_regs_1): Don't use
2101 simplify_gen_subreg until final substitution.
2103 2014-05-23 Alan Modra <amodra@gmail.com>
2106 * config/rs6000/rs6000.c (mem_operand_gpr): Handle SImode.
2107 * config/rs6000/rs6000.md (extendsidi2_lfiwax, extendsidi2_nocell):
2108 Use "Y" constraint rather than "m".
2110 2014-05-23 Kugan Vivekanandarajah <kuganv@linaro.org>
2112 * config/aarch64/aarch64.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New
2114 * config/aarch64/aarch64-protos.h (aarch64_atomic_assign_expand_fenv):
2115 New function declaration.
2116 * config/aarch64/aarch64-builtins.c (aarch64_builtins) : Add
2117 AARCH64_BUILTIN_GET_FPCR, AARCH64_BUILTIN_SET_FPCR.
2118 AARCH64_BUILTIN_GET_FPSR and AARCH64_BUILTIN_SET_FPSR.
2119 (aarch64_init_builtins) : Initialize builtins
2120 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
2121 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
2122 (aarch64_expand_builtin) : Expand builtins __builtins_aarch64_set_fpcr
2123 __builtins_aarch64_get_fpcr, __builtins_aarch64_get_fpsr,
2124 and __builtins_aarch64_set_fpsr.
2125 (aarch64_atomic_assign_expand_fenv): New function.
2126 * config/aarch64/aarch64.md (set_fpcr): New pattern.
2127 (get_fpcr) : Likewise.
2128 (set_fpsr) : Likewise.
2129 (get_fpsr) : Likewise.
2130 (unspecv): Add UNSPECV_GET_FPCR and UNSPECV_SET_FPCR, UNSPECV_GET_FPSR
2131 and UNSPECV_SET_FPSR.
2132 * doc/extend.texi (AARCH64 Built-in Functions) : Document
2133 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
2134 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
2136 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
2138 PR rtl-optimization/60969
2139 * ira-costs.c (record_reg_classes): Process NO_REGS for matching
2140 constraints. Set up mem cost for NO_REGS case.
2142 2014-05-22 Thomas Schwinge <thomas@codesourcery.com>
2144 * builtin-types.def: Simplify examples for DEF_FUNCTION_TYPE_*.
2146 2012-05-22 Bernd Schmidt <bernds@codesourcery.com>
2148 * config/darwin.c: Include "lto-section-names.h".
2149 (LTO_SEGMENT_NAME): Don't define.
2150 * config/i386/winnt.c: Include "lto-section-names.h".
2151 * lto-streamer.c: Include "lto-section-names.h".
2152 * lto-streamer.h (LTO_SECTION_NAME_PREFIX): Don't define.
2153 * lto-wrapper.c: Include "lto-section-names.h".
2154 (LTO_SECTION_NAME_PREFIX): Don't define.
2155 * lto-section-names.h: New file.
2156 * cgraphunit.c: Include "lto-section-names.h".
2158 2014-05-22 Peter Bergner <bergner@vnet.ibm.com>
2160 * config/rs6000/htm.md (ttest): Use correct shift value to get CR0.
2162 2014-05-22 Richard Earnshaw <rearnsha@arm.com>
2165 * arm.md (arm_cmpdi_unsigned): Fix length calculation for Thumb2.
2167 2014-05-22 Nick Clifton <nickc@redhat.com>
2169 * config/msp430/msp430.h (ASM_SPEC): Add spaces after inserted options.
2171 2014-05-22 Eric Botcazou <ebotcazou@adacore.com>
2173 * tree-ssa-forwprop.c (associate_plusminus): Extend (T)(P + A) - (T)P
2174 -> (T)A transformation to integer types.
2176 2014-05-22 Teresa Johnson <tejohnson@google.com>
2178 * gcov-io.c (gcov_position): Use gcov_nonruntime_assert.
2179 (gcov_is_error): Remove gcc_assert from IN_LIBGCOV code.
2180 (gcov_rewrite): Use gcov_nonruntime_assert.
2182 (gcov_write_words): Ditto.
2183 (gcov_write_length): Ditto.
2184 (gcov_read_words): Use gcov_nonruntime_assert, and remove
2185 gcc_assert from IN_LIBGCOV code.
2186 (gcov_read_summary): Use gcov_error to flag profile corruption.
2187 (gcov_sync): Use gcov_nonruntime_assert.
2188 (gcov_seek): Remove gcc_assert from IN_LIBGCOV code.
2189 (gcov_histo_index): Use gcov_nonruntime_assert.
2190 (static void gcov_histogram_merge): Ditto.
2191 (compute_working_sets): Ditto.
2192 * gcov-io.h (gcov_nonruntime_assert): Define.
2193 (gcov_error): Define for !IN_LIBGCOV
2195 2014-05-22 Richard Biener <rguenther@suse.de>
2197 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
2198 BUILT_IN_REALLOC like BUILT_IN_STRDUP.
2199 (call_may_clobber_ref_p_1): Handle BUILT_IN_REALLOC as allocation
2200 and deallocation site.
2201 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2202 Handle BUILT_IN_REALLOC similar to BUILT_IN_STRDUP with also
2203 passing through the incoming points-to set.
2204 (handle_lhs_call): Use flags argument instead of recomputing it.
2205 (find_func_aliases_for_call): Call handle_lhs_call with proper
2208 2014-05-22 Jakub Jelinek <jakub@redhat.com>
2210 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Make sure
2211 all padding bits in REAL_VALUE_TYPE are cleared.
2213 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
2215 Cleanup and improve multipass_dfa_lookahead_guard
2216 * config/i386/i386.c (core2i7_first_cycle_multipass_filter_ready_try,)
2217 (core2i7_first_cycle_multipass_begin,)
2218 (core2i7_first_cycle_multipass_issue,)
2219 (core2i7_first_cycle_multipass_backtrack): Update signature.
2220 * config/ia64/ia64.c
2221 (ia64_first_cycle_multipass_dfa_lookahead_guard_spec): Remove.
2222 (ia64_first_cycle_multipass_dfa_lookahead_guard): Update signature.
2223 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove
2225 (ia64_first_cycle_multipass_dfa_lookahead_guard): Merge logic from
2226 ia64_first_cycle_multipass_dfa_lookahead_guard_spec. Update return
2228 * config/rs6000/rs6000.c (rs6000_use_sched_lookahead_guard): Update
2230 * doc/tm.texi: Regenerate.
2232 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove.
2233 * haifa-sched.c (ready_try): Make signed to allow negative values.
2234 (rebug_ready_list_1): Update.
2235 (choose_ready): Simplify.
2236 (sched_extend_ready_list): Update.
2238 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
2240 Remove IA64 speculation tweaking flags
2241 * config/ia64/ia64.c (ia64_set_sched_flags): Delete handling of
2242 speculation tuning flags.
2243 (msched-prefer-non-data-spec-insns,)
2244 (msched-prefer-non-control-spec-insns): Obsolete options.
2245 * haifa-sched.c (choose_ready): Remove handling of
2246 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
2247 * sched-int.h (enum SPEC_SCHED_FLAGS): Remove PREFER_NON_CONTROL_SPEC
2248 and PREFER_NON_DATA_SPEC.
2249 * sel-sched.c (process_spec_exprs): Remove handling of
2250 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
2252 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
2254 Improve scheduling debug output
2255 * haifa-sched.c (debug_ready_list): Remove unnecessary prototype.
2256 (advance_one_cycle): Update.
2257 (schedule_insn, queue_to_ready): Add debug printouts.
2258 (debug_ready_list_1): New static function.
2259 (debug_ready_list): Update.
2260 (max_issue): Add debug printouts.
2261 (dump_insn_stream): New static function.
2262 (schedule_block): Use it. Also better indent printouts.
2264 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
2266 Fix sched_insn debug counter
2267 * haifa-sched.c (schedule_insn): Update.
2268 (struct haifa_saved_data): Add nonscheduled_insns_begin.
2269 (save_backtrack_point, restore_backtrack_point): Update.
2270 (first_nonscheduled_insn): New static function.
2271 (queue_to_ready, choose_ready): Use it.
2272 (schedule_block): Init nonscheduled_insns_begin.
2273 (sched_emit_insn): Update.
2276 2014-05-22 Kugan Vivekanandarajah <kuganv@linaro.org>
2278 * config/aarch64/aarch64.c (aarch64_regno_regclass) : Change CORE_REGS
2280 (aarch64_secondary_reload) : LikeWise.
2281 (aarch64_class_max_nregs) : Remove CORE_REGS.
2282 * config/aarch64/aarch64.h (enum reg_class) : Remove CORE_REGS.
2283 (REG_CLASS_NAMES) : Likewise.
2284 (REG_CLASS_CONTENTS) : LikeWise.
2285 (INDEX_REG_CLASS) : Change CORE_REGS to GENERAL_REGS.
2287 2014-05-21 Guozhi Wei <carrot@google.com>
2290 * config/aarch64/arm_neon.h (vqdmulh_n_s16): Change the last operand's
2292 (vqdmulhq_n_s16): Likewise.
2294 2014-05-21 Segher Boessenkool <segher@kernel.crashing.org>
2296 * config/rs6000/predicates.md (update_indexed_address_mem): Delete.
2298 2014-05-21 Marek Polacek <polacek@redhat.com>
2301 * ubsan.c (is_ubsan_builtin_p): Turn assert into a condition.
2303 2014-05-21 Martin Jambor <mjambor@suse.cz>
2305 * doc/invoke.texi (Optimize Options): Document parameters
2306 ipa-cp-eval-threshold, ipa-max-agg-items, ipa-cp-loop-hint-bonus and
2307 ipa-cp-array-index-hint-bonus.
2309 2014-05-21 Mark Wielaard <mjw@redhat.com>
2312 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_type if DWARF
2313 version >= 3 or not strict DWARF.
2314 * langhooks.h (struct lang_hooks_for_types): Add
2315 enum_underlying_base_type.
2316 * langhooks.c (lhd_enum_underlying_base_type): New function.
2317 * gcc/langhooks.h (struct lang_hooks_for_types): Add
2318 enum_underlying_base_type.
2319 * langhooks-def.h (lhd_enum_underlying_base_type): New declaration.
2320 (LANG_HOOKS_ENUM_UNDERLYING_BASE_TYPE): New define.
2321 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add new lang hook.
2323 2014-05-21 Richard Biener <rguenther@suse.de>
2325 * doc/invoke.texi (-flto-partition=): Document one and none algorithms.
2327 2014-05-21 John Marino <gnugcc@marino.st>
2329 * config.gcc (*-*-dragonfly*): New target.
2330 * configure.ac: Detect dl_iterate_phdr (*freebsd*, *dragonfly*).
2331 * configure: Regenerate.
2332 * config/dragonfly-stdint.h: New.
2333 * config/dragonfly.h: New.
2334 * config/dragonfly.opt: New.
2335 * config/i386/dragonfly.h: New.
2336 * ginclude/stddef.h: Detect _PTRDIFF_T_DECLARED for DragonFly.
2338 2014-05-21 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2340 * tree.def (VOID_CST): New.
2341 * tree-core.h (TI_VOID): New.
2342 * tree.h (void_node): New.
2343 * tree.c (tree_node_structure_for_code, tree_code_size)
2344 (iterative_hash_expr): Handle VOID_CST.
2345 (build_common_tree_nodes): Initialize void_node.
2347 2014-05-21 Bernd Schmidt <bernds@codesourcery.com>
2349 * reload1.c (remove_init_insns, will_delete_init_insn_p): New static
2351 (reload, calculate_needs_all_insns, reload_as_needed): Use them.
2353 * config/bfin/bfin.c (split_load_immediate): Use gen_int_mode in a few
2356 * cfgrtl.c (cfg_layout_initialize): Weaken assert to only trigger if
2357 flag_reorder_blocks_and_partition.
2358 * hw-doloop.c (reorg_loops): Avoid reordering if that flag is set.
2360 2014-05-21 Oleg Endo <olegendo@gcc.gnu.org>
2363 * config/sh/sh.md (*addc_r_1): Rename to addc_t_r. Remove empty
2365 (*addc_r_t): Add new insn_and_split.
2367 2014-05-21 Jakub Jelinek <jakub@redhat.com>
2370 * omp-low.c (handle_simd_reference): New function.
2371 (lower_rec_input_clauses): Use it. Defer adding reference
2372 initialization even for reduction without placeholder if in simd,
2373 handle it properly later on.
2375 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
2377 PR tree-optimization/60899
2378 * gimple-fold.c (can_refer_decl_in_current_unit_p): Cleanup;
2379 assume all static symbols will have definition wile parsing and
2380 check the do have definition later in compilation; check that
2381 variable referring symbol will be output before concluding that
2382 reference is safe; be conservative for referring local statics;
2383 be more precise about when comdat is output in other partition.
2385 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
2388 * ipa-inline-transform.c (inline_call): Use add CALLEE_REMOVED
2390 * ipa-inline.c (inline_to_all_callers): If callee was removed; return.
2391 (ipa_inline): Loop inline_to_all_callers until no more aliases
2394 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
2396 * ipa.c (ipa_discover_readonly_nonaddressable_var): Fix dumping;
2397 set writeonly flag only for vars actually written to.
2399 2014-05-20 Dehao Chen <dehao@google.com>
2401 * ipa-inline-transform.c (clone_inlined_nodes): Use min of edge count
2402 and callee count to get clone count.
2403 * tree-inline.c (expand_call_inline): Use callee count instead of bb
2406 2014-05-20 Richard Sandiford <rdsandiford@googlemail.com>
2408 PR rtl-optimization/61243
2409 * emit-rtl.c (emit_copy_of_insn_after): Copy CROSSING_JUMP_P.
2411 2014-05-20 Xinliang David Li <davidxl@google.com>
2413 * cgraphunit.c (walk_polymorphic_call_targets): Add
2414 dbgcnt and fopt-info support.
2415 * ipa-prop.c (ipa_make_edge_direct_to_target): Ditto.
2416 * ipa-devirt.c (ipa_devirt): Ditto.
2417 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Ditto.
2418 * ipa.c (walk_polymorphic_call_targets): Ditto.
2419 * gimple-fold.c (fold_gimple_assign): Ditto.
2420 (gimple_fold_call): Ditto.
2421 * dbgcnt.def: New counter.
2423 2014-05-20 DJ Delorie <dj@redhat.com>
2425 * config/msp430/msp430.md (split): Don't allow subregs when
2426 splitting SImode adds.
2427 (andneghi): Fix subtraction logic.
2428 * config/msp430/predicates.md (msp430_nonsubreg_or_imm_operand): New.
2430 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
2432 * tree.h (DECL_ONE_ONLY): Return true only for externally visible
2434 * except.c (switch_to_exception_section, resolve_unique_section,
2435 get_named_text_section, default_function_rodata_section,
2436 align_variable, get_block_for_decl, default_section_type_flags):
2437 Use DECL_COMDAT_GROUP instead of DECL_ONE_ONLY.
2438 * symtab.c (symtab_add_to_same_comdat_group,
2439 symtab_make_decl_local, fixup_same_cpp_alias_visibility,
2440 symtab_nonoverwritable_alias, symtab_get_symbol_partitioning_class):
2442 * cgraphclones.c (cgraph_create_virtual_clone): Likewise.
2443 * bb-reorder.c (pass_partition_blocks::gate): Likewise.
2444 * config/c6x/c6x.c (c6x_elf_unique_section): Likewise.
2445 (c6x_function_in_section_p): Likewise.
2446 * config/darwin.c (machopic_select_section): Likewise.
2447 * config/arm/arm.c (arm_function_in_section_p): Likewise.
2448 * config/mips/mips.c (mips_function_rodata_section): Likewise.
2449 * config/mep/mep.c (mep_select_section): LIkewise.
2450 * config/i386/i386.c (x86_64_elf_unique_section): Likewise.
2452 2014-05-20 Eric Botcazou <ebotcazou@adacore.com>
2454 * tree-ssa-dom.c (hashable_expr_equal_p) <EXPR_CALL>: Also compare the
2455 EH region of calls to pure functions that can throw an exception.
2456 * tree-ssa-sccvn.c (vn_reference_eq): Remove duplicated test.
2457 (copy_reference_ops_from_call): Also copy the EH region of the call if
2458 it can throw an exception.
2460 2014-05-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2462 * simplify-rtx.c (simplify_binary_operation_1): Optimize case of
2463 nested VEC_SELECTs that are inverses of each other.
2465 2014-05-20 Richard Biener <rguenther@suse.de>
2467 * tree-ssa-sccvn.c (process_scc): Dump SCC here, when iterating,
2468 (extract_and_process_scc_for_name): not here.
2469 (cond_dom_walker::before_dom_children): Only process
2470 stmts that end the BB in interesting ways.
2471 (run_scc_vn): Mark param uses as visited.
2473 2014-05-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2475 * config/arm/arm.md (arith_shiftsi): Do not predicate for
2478 2014-05-20 Nick Clifton <nickc@redhat.com>
2480 * config/msp430/msp430.c (TARGET_GIMPLIFY_VA_ARG_EXPR): Define.
2481 (msp430_gimplify_va_arg_expr): New function.
2482 (msp430_print_operand): Handle (CONST (ZERO_EXTRACT)).
2484 * config/msp430/msp430.md (zero_extendpsisi2): Use + constraint on
2485 operand 0 in order to prevent confusion about the number of
2488 2014-05-20 Richard Biener <rguenther@suse.de>
2490 PR tree-optimization/61221
2491 * tree-ssa-pre.c (el_to_update): Remove.
2492 (eliminate_dom_walker::before_dom_children): Handle released
2493 VDEFs by value-numbering them to the associated VUSE. Update
2494 stmt immediately for substituted call address.
2495 (eliminate): Remove delayed stmt updating code.
2496 * tree-ssa-sccvn.c (vuse_ssa_val): New function valueizing
2497 possibly late re-numbered vuses.
2498 (vn_reference_lookup_2): Adjust.
2499 (vn_reference_lookup_pieces): Likewise.
2500 (vn_reference_lookup): Likewise.
2502 2014-05-20 Richard Biener <rguenther@suse.de>
2504 * config.gcc: Remove need_64bit_hwint.
2505 * configure.ac: Do not define NEED_64BIT_HOST_WIDE_INT.
2506 * hwint.h: Do not check NEED_64BIT_HOST_WIDE_INT but assume
2508 * config.in: Regenerate.
2509 * configure: Likewise.
2511 2014-05-19 David Wohlferd <dw@LimeGreenSocks.com>
2513 * doc/extend.texi: Create Label Attributes section,
2514 move all label attributes into it and reference it.
2516 2014-05-19 Richard Earnshaw <rearnsha@arm.com>
2518 * arm.c (thumb1_reorg): When scanning backwards skip anything
2519 that's not a proper insn.
2521 2014-05-19 Richard Biener <rguenther@suse.de>
2523 PR tree-optimization/61221
2524 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
2525 Do nothing for unreachable blocks.
2526 * tree-ssa-sccvn.c (cond_dom_walker::before_dom_children):
2527 Improve unreachability detection.
2529 2014-05-19 Richard Biener <rguenther@suse.de>
2531 PR tree-optimization/61209
2532 * tree-ssa-sccvn.c (visit_phi): Avoid setting expr to VN_TOP.
2534 2014-05-19 Nick Clifton <nickc@redhat.com>
2536 * except.c (init_eh): Fix computation of builtin setjmp buffer
2537 size to allow for targets where POINTER_SIZE > BITS_PER_WORD.
2539 2014-05-19 Richard Biener <rguenther@suse.de>
2541 PR tree-optimization/61184
2542 * tree-vrp.c (is_negative_overflow_infinity): Use
2543 TREE_OVERFLOW_P and do that check first.
2544 (is_positive_overflow_infinity): Likewise.
2545 (is_overflow_infinity): Likewise.
2546 (vrp_operand_equal_p): Properly treat operands with
2547 differing overflow as not equal.
2549 2014-05-19 Bernd Schmidt <bernds@codesourcery.com>
2551 * simplify-rtx.c (simplify_unary_operation_1): Use CONST_INT_P in
2552 shift simplification where it was intended.
2554 2014-05-19 Christian Bruel <christian.bruel@st.com>
2557 * config/sh/sh.md (movsf_ie): Unset fp_mode for fmov.
2559 2014-05-19 Richard Sandiford <r.sandiford@uk.ibm.com>
2562 * config/sparc/sparc.c (sparc_fold_builtin): Use widest_int rather
2565 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
2567 * reg-notes.def (CROSSING_JUMP): Likewise.
2568 * rtl.h (rtx_def): Update comment for jump flag.
2569 (CROSSING_JUMP_P): Define.
2570 * cfgcleanup.c (try_forward_edges, try_optimize_cfg): Use it instead
2571 of a REG_CROSSING_JUMP note.
2572 * cfghooks.c (tidy_fallthru_edges): Likewise.
2573 * cfgrtl.c (fixup_partition_crossing, rtl_verify_edges): Likewise.
2574 * emit-rtl.c (try_split): Likewise.
2575 * haifa-sched.c (sched_create_recovery_edges): Likewise.
2576 * ifcvt.c (find_if_case_1, find_if_case_2): Likewise.
2577 * jump.c (redirect_jump_2): Likewise.
2578 * reorg.c (follow_jumps, fill_slots_from_thread): Likewise.
2579 (relax_delay_slots): Likewise.
2580 * config/arc/arc.md (jump_i, cbranchsi4_scratch, *bbit): Likewise.
2581 (bbit_di): Likewise.
2582 * config/arc/arc.c (arc_reorg, arc_can_follow_jump): Likewise.
2583 * config/sh/sh.md (jump_compact): Likewise.
2584 * bb-reorder.c (rotate_loop): Likewise.
2585 (pass_duplicate_computed_gotos::execute): Likewise.
2586 (add_reg_crossing_jump_notes): Rename to...
2587 (update_crossing_jump_flags): ...this.
2588 (pass_partition_blocks::execute): Update accordingly.
2590 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
2592 * tree.h: Remove extraneous template <>.
2594 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
2596 * ipa.c (symtab_remove_unreachable_nodes): Remove
2597 symbol from comdat group if its body was eliminated.
2598 (comdat_can_be_unshared_p_1): Static symbols can always be privatized.
2599 * symtab.c (symtab_remove_from_same_comdat_group): Break out from ...
2600 (symtab_unregister_node): ... this one.
2601 (verify_symtab_base): More strict checking of comdats.
2602 * cgraph.h (symtab_remove_from_same_comdat_group): Declare.
2604 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
2606 * tree-pass.h (make_pass_ipa_comdats): New pass.
2607 * timevar.def (TV_IPA_COMDATS): New timevar.
2608 * passes.def (pass_ipa_comdats): Add.
2609 * Makefile.in (OBJS): Add ipa-comdats.o
2610 * ipa-comdats.c: New file.
2612 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
2614 * ipa.c (update_visibility_by_resolution_info): New function.
2615 (function_and_variable_visibility): Use it.
2617 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
2619 * cgraph.h (symtab_first_defined_symbol, symtab_next_defined_symbol):
2621 (FOR_EACH_DEFINED_SYMBOL): New macro.
2622 (varpool_first_static_initializer, varpool_next_static_initializer,
2623 varpool_first_defined_variable, varpool_next_defined_variable):
2625 (symtab_in_same_comdat_p): Correctly deal with inline functions.
2627 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2629 * ggc-page.c (ggc_handle_finalizers): Add comment.
2631 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2633 * ggc-common.c (ggc_internal_cleared_alloc): Adjust.
2634 * ggc-none.c (ggc_internal_alloc): Assert if a finalizer is passed.
2635 (ggc_internal_cleared_alloc): Likewise.
2636 * ggc-page.c (finalizer): New class.
2637 (vec_finalizer): Likewise.
2638 (globals::finalizers): New member.
2639 (globals::vec_finalizers): Likewise.
2640 (ggc_internal_alloc): Record the finalizer if any for the block being
2642 (ggc_handle_finalizers): New function.
2643 (ggc_collect): Call ggc_handle_finalizers.
2644 * ggc.h (ggc_internal_alloc): Add arguments to allow installing a
2646 (ggc_internal_cleared_alloc): Likewise.
2647 (finalize): New function.
2648 (need_finalization_p): Likewise.
2649 (ggc_alloc): Install the type's destructor as the finalizer if it
2651 (ggc_cleared_alloc): Likewise.
2652 (ggc_vec_alloc): Likewise.
2653 (ggc_cleared_vec_alloc): Likewise.
2655 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2657 * ggc.h (ggc_alloc_cleared_simd_clone_stat): Remove function.
2659 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2661 * alias.c (record_alias_subset): Adjust.
2662 * bitmap.c (bitmap_element_allocate): Likewise.
2663 (bitmap_gc_alloc_stat): Likewise.
2664 * cfg.c (init_flow): Likewise.
2665 (alloc_block): Likewise.
2666 (unchecked_make_edge): Likewise.
2667 * cfgloop.c (alloc_loop): Likewise.
2668 (flow_loops_find): Likewise.
2669 (rescan_loop_exit): Likewise.
2670 * cfgrtl.c (init_rtl_bb_info): Likewise.
2671 * cgraph.c (insert_new_cgraph_node_version): Likewise.
2672 (cgraph_allocate_node): Likewise.
2673 (cgraph_create_edge_1): Likewise.
2674 (cgraph_allocate_init_indirect_info): Likewise.
2675 * cgraphclones.c (cgraph_clone_edge): Likewise.
2676 * cgraphunit.c (add_asm_node): Likewise.
2677 (init_lowered_empty_function): Likewise.
2678 * config/aarch64/aarch64.c (aarch64_init_machine_status): Likewise.
2679 * config/alpha/alpha.c (alpha_init_machine_status): Likewise.
2680 (alpha_use_linkage): Likewise.
2681 * config/arc/arc.c (arc_init_machine_status): Likewise.
2682 * config/arm/arm.c (arm_init_machine_status): Likewise.
2683 * config/avr/avr.c (avr_init_machine_status): Likewise.
2684 * config/bfin/bfin.c (bfin_init_machine_status): Likewise.
2685 * config/c6x/c6x.c (c6x_init_machine_status): Likewise.
2686 * config/cris/cris.c (cris_init_machine_status): Likewise.
2687 * config/darwin.c (machopic_indirection_name): Likewise.
2688 (darwin_build_constant_cfstring): Likewise.
2689 (darwin_enter_string_into_cfstring_table): Likewise.
2690 * config/epiphany/epiphany.c (epiphany_init_machine_status): Likewise.
2691 * config/frv/frv.c (frv_init_machine_status): Likewise.
2692 * config/i386/i386.c (get_dllimport_decl): Likewise.
2693 (ix86_init_machine_status): Likewise.
2694 (assign_386_stack_local): Likewise.
2695 * config/i386/winnt.c (i386_pe_record_external_function): Likewise.
2696 (i386_pe_maybe_record_exported_symbol): Likewise.
2697 (i386_pe_record_stub): Likewise.
2698 * config/ia64/ia64.c (ia64_init_machine_status): Likewise.
2699 * config/iq2000/iq2000.c (iq2000_init_machine_status): Likewise.
2700 * config/m32c/m32c.c (m32c_init_machine_status): Likewise.
2701 (m32c_note_pragma_address): Likewise.
2702 * config/mep/mep.c (mep_init_machine_status): Likewise.
2703 (mep_note_pragma_flag): Likewise.
2704 * config/mips/mips.c (mflip_mips16_use_mips16_p): Likewise.
2705 (mips16_local_alias): Likewise.
2706 (mips_init_machine_status): Likewise.
2707 * config/mmix/mmix.c (mmix_init_machine_status): Likewise.
2708 * config/moxie/moxie.c (moxie_init_machine_status): Likewise.
2709 * config/msp430/msp430.c (msp430_init_machine_status): Likewise.
2710 * config/nds32/nds32.c (nds32_init_machine_status): Likewise.
2711 * config/nios2/nios2.c (nios2_init_machine_status): Likewise.
2712 * config/pa/pa.c (pa_init_machine_status): Likewise.
2713 (pa_get_deferred_plabel): Likewise.
2714 * config/rl78/rl78.c (rl78_init_machine_status): Likewise.
2715 * config/rs6000/rs6000.c (builtin_function_type): Likewise.
2716 (rs6000_init_machine_status): Likewise.
2717 (output_toc): Likewise.
2718 * config/s390/s390.c (s390_init_machine_status): Likewise.
2719 * config/score/score.c (score_output_external): Likewise.
2720 * config/sparc/sparc.c (sparc_init_machine_status): Likewise.
2721 * config/spu/spu.c (spu_init_machine_status): Likewise.
2722 * config/tilegx/tilegx.c (tilegx_init_machine_status): Likewise.
2723 * config/tilepro/tilepro.c (tilepro_init_machine_status): Likewise.
2724 * config/xtensa/xtensa.c (xtensa_init_machine_status): Likewise.
2725 * coverage.c (coverage_end_function): Likewise.
2726 * dbxout.c (dbxout_init): Likewise.
2727 * doc/gty.texi: Don't mention variable_size attribute.
2728 * dwarf2cfi.c (new_cfi): Adjust.
2729 (new_cfi_row): Likewise.
2730 (copy_cfi_row): Likewise.
2731 (create_cie_data): Likewise.
2732 * dwarf2out.c (dwarf2out_alloc_current_fde): Likewise.
2733 (new_loc_descr): Likewise.
2734 (find_AT_string_in_table): Likewise.
2735 (add_addr_table_entry): Likewise.
2736 (new_die): Likewise.
2737 (add_var_loc_to_decl): Likewise.
2738 (clone_die): Likewise.
2739 (clone_as_declaration): Likewise.
2740 (break_out_comdat_types): Likewise.
2741 (new_loc_list): Likewise.
2742 (add_loc_descr_to_each): Likewise.
2743 (add_location_or_const_value_attribute): Likewise.
2744 (add_linkage_name): Likewise.
2745 (lookup_filename): Likewise.
2746 (dwarf2out_var_location): Likewise.
2747 (new_line_info_table): Likewise.
2748 (dwarf2out_init): Likewise.
2749 (mem_loc_descriptor): Likewise.
2750 (loc_descriptor): Likewise.
2751 (add_const_value_attribute): Likewise.
2752 (tree_add_const_value_attribute): Likewise.
2753 (comp_dir_string): Likewise.
2754 (dwarf2out_vms_debug_main_pointer): Likewise.
2755 (string_cst_pool_decl): Likewise.
2756 * emit-rtl.c (set_mem_attrs): Likewise.
2757 (get_reg_attrs): Likewise.
2758 (start_sequence): Likewise.
2759 (init_emit): Likewise.
2760 (init_emit_regs): Likewise.
2761 * except.c (init_eh_for_function): Likewise.
2762 (gen_eh_region): Likewise.
2763 (gen_eh_region_catch): Likewise.
2764 (gen_eh_landing_pad): Likewise.
2765 (add_call_site): Likewise.
2766 * function.c (add_frame_space): Likewise.
2767 (insert_temp_slot_address): Likewise.
2768 (assign_stack_temp_for_type): Likewise.
2769 (get_hard_reg_initial_val): Likewise.
2770 (allocate_struct_function): Likewise.
2771 (prepare_function_start): Likewise.
2772 (types_used_by_var_decl_insert): Likewise.
2773 * gengtype.c (variable_size_p): Remove function.
2774 (enum alloc_quantity): Remove enum.
2775 (write_typed_alloc_def): Remove function.
2776 (write_typed_struct_alloc_def): Likewise.
2777 (write_typed_typedef_alloc_def): Likewise.
2778 (write_typed_alloc_defns): Likewise.
2780 * ggc-common.c (ggc_cleared_alloc_htab_ignore_args): Adjust.
2781 (ggc_cleared_alloc_ptr_array_two_args): Likewise.
2782 * ggc.h (ggc_alloc): new function.
2783 (ggc_cleared_alloc): Likewise.
2784 (ggc_vec_alloc): Template on type of vector element, and remove
2785 element size argument.
2786 (ggc_cleared_vec_alloc): Likewise.
2787 * gimple.c (gimple_build_omp_for): Adjust.
2788 (gimple_copy): Likewise.
2789 * ipa-cp.c (get_replacement_map): Likewise.
2790 (find_aggregate_values_for_callers_subset): Likewise.
2791 (known_aggs_to_agg_replacement_list): Likewise.
2792 * ipa-devirt.c (get_odr_type): Likewise.
2793 * ipa-prop.c (ipa_node_duplication_hook): Likewise.
2794 (read_agg_replacement_chain): Likewise.
2795 * loop-iv.c (get_simple_loop_desc): Likewise.
2796 * lto-cgraph.c (input_node_opt_summary): Likewise.
2797 * lto-section-in.c (lto_new_in_decl_state): Likewise.
2798 * lto-streamer-in.c (lto_input_eh_catch_list): Likewise.
2799 (input_eh_region): Likewise.
2800 (input_eh_lp): Likewise.
2801 (input_cfg): Likewise.
2802 * optabs.c (set_optab_libfunc): Likewise.
2803 (init_tree_optimization_optabs): Likewise.
2804 (set_conv_libfunc): Likewise.
2805 * passes.c (do_per_function_toporder): Likewise.
2806 * rtl.h: Don't use variable_size gty attribute.
2807 * sese.c (if_region_set_false_region): Adjust.
2808 * stringpool.c (gt_pch_save_stringpool): Likewise.
2809 * target-globals.c (save_target_globals): Likewise.
2810 * toplev.c (general_init): Likewise.
2811 * trans-mem.c (record_tm_replacement): Likewise.
2812 (split_bb_make_tm_edge): Likewise.
2813 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2814 * tree-data-ref.h (lambda_vector_new): Likewise.
2815 * tree-eh.c (add_stmt_to_eh_lp_fn): Likewise.
2816 * tree-iterator.c (tsi_link_before): Likewise.
2817 (tsi_link_after): Likewise.
2818 * tree-scalar-evolution.c (new_scev_info_str): Likewise.
2819 * tree-ssa-loop-niter.c (record_estimate): Likewise.
2820 * tree-ssa-operands.c (ssa_operand_alloc): Likewise.
2821 * tree-ssa-operands.h: Don't use variable_size gty attribute.
2822 * tree-ssa.c (init_tree_ssa): Adjust.
2823 * tree-ssanames.c (set_range_info): Likewise.
2824 (get_ptr_info): Likewise.
2825 (duplicate_ssa_name_ptr_info): Likewise.
2826 (duplicate_ssa_name_range_info): Likewise.
2827 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Likewise.
2828 (unpack_ts_fixed_cst_value_fields): Likewise.
2829 * tree.c (build_fixed): Likewise.
2830 (build_real): Likewise.
2831 (build_string): Likewise.
2832 (decl_priority_info): Likewise.
2833 (decl_debug_expr_insert): Likewise.
2834 (decl_value_expr_insert): Likewise.
2835 (decl_debug_args_insert): Likewise.
2836 (type_hash_add): Likewise.
2837 (build_omp_clause): Likewise.
2838 * ubsan.c (decl_for_type_insert): Likewise.
2839 * varasm.c (get_unnamed_section): Likewise.
2840 (get_noswitch_section): Likewise.
2841 (get_section): Likewise.
2842 (get_block_for_section): Likewise.
2843 (create_block_symbol): Likewise.
2844 (build_constant_desc): Likewise.
2845 (create_constant_pool): Likewise.
2846 (force_const_mem): Likewise.
2847 (record_tm_clone_pair): Likewise.
2848 * varpool.c (varpool_create_empty_node): Likewise.
2850 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2852 * dwarf2out.c (tree_add_const_value_attribute): Call
2853 ggc_internal_cleared_alloc instead of ggc_alloc_cleared_atomic.
2854 * gengtype.c (write_typed_alloc_def): Call ggc_internal_<x>alloc
2855 instead of ggc_internal_<x>alloc_stat.
2856 * ggc-common.c (ggc_internal_cleared_alloc): Drop _stat suffix.
2857 (ggc_realloc): Likewise.
2858 * ggc-none.c (ggc_internal_alloc): Likewise.
2859 (ggc_internal_cleared_alloc): Likewise.
2860 * ggc-page.c: Likewise.
2861 * ggc.h (ggc_internal_alloc_stat): Likewise.
2862 (ggc_internal_alloc): Remove macro.
2863 (ggc_internal_cleared_alloc_stat): Drop _stat suffix.
2864 (ggc_internal_cleared_alloc): Remove macro.
2865 (GGC_RESIZEVEC): Adjust.
2866 (ggc_resizevar): Remove macro.
2867 (ggc_internal_vec_alloc_stat): Drop _stat suffix.
2868 (ggc_internal_cleared_vec_alloc_stat): Likewise.
2869 (ggc_internal_vec_cleared_alloc): Remove macro.
2870 (ggc_alloc_atomic_stat): Drop _stat suffix.
2871 (ggc_alloc_atomic): Remove macro.
2872 (ggc_alloc_cleared_atomic): Remove macro.
2873 (ggc_alloc_string_stat): Drop _stat suffix.
2874 (ggc_alloc_string): Remove macro.
2875 (ggc_alloc_rtx_def_stat): Adjust.
2876 (ggc_alloc_tree_node_stat): Likewise.
2877 (ggc_alloc_cleared_tree_node_stat): Likewise.
2878 (ggc_alloc_cleared_gimple_statement_stat): Likewise.
2879 (ggc_alloc_cleared_simd_clone_stat): Likewise.
2880 * gimple.c (gimple_build_omp_for): Likewise.
2881 (gimple_copy): Likewise.
2882 * stringpool.c (ggc_alloc_string_stat): Drop _stat suffix.
2883 * toplev.c (realloc_for_line_map): Adjust.
2884 * tree-data-ref.h (lambda_vector_new): Likewise.
2885 * tree-phinodes.c (allocate_phi_node): Likewise.
2886 * tree.c (grow_tree_vec_stat): Likewise.
2887 * vec.h (va_gc::reserve): Adjust.
2889 2014-05-17 Ajit Agarwal <ajitkum@xilinx.com>
2891 * config/microblaze/microblaze.c (break_handler): New Declaration.
2892 (microblaze_break_function_p,microblaze_is_break_handler): New.
2893 (compute_frame_size): Use microblaze_break_function_p.
2894 Add the test of break_handler.
2895 (microblaze_function_prologue) : Add the test of variable
2896 break_handler. Check the fnname by BREAK_HANDLER_NAME.
2897 (microblaze_function_epilogue) : Add the test of break_handler.
2898 (microblaze_globalize_label) : Add the test of break_handler.
2899 Check the name by BREAK_HANDLER_NAME.
2901 * config/microblaze/microblaze.h (BREAK_HANDLER_NAME): New macro
2903 * config/microblaze/microblaze.md (*<optab>,<optab>_internal): Add
2904 microblaze_is_break_handler test.
2905 (call_internal1,call_value_intern): Use microblaze_break_function_p.
2906 Use SYMBOL_REF_DECL.
2908 * config/microblaze/microblaze-protos.h
2909 (microblaze_break_function_p,microblaze_is_break_handler):
2912 * doc/extend.texi (MicroBlaze break_handler Functions): Document
2913 new MicroBlaze break_handler functions.
2915 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
2917 * doc/extend.texi (Size of an asm): Move node text according
2918 to its @menu entry position.
2920 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
2922 PR tree-optimization/61140
2923 PR tree-optimization/61150
2924 PR tree-optimization/61197
2925 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
2927 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
2929 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
2931 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
2933 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
2934 __SIZEOF_INT128__ is defined.
2936 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
2938 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
2939 (rs6000_delegitimize_address): Use it.
2941 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
2943 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
2944 inplace argument. Store the new address in the original MEM when true.
2945 * emit-rtl.c (change_address_1): Likewise.
2946 (adjust_address_1, adjust_automodify_address_1, offset_address):
2948 * rtl.h (plus_constant): Add an inplace argument.
2949 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
2950 when true. Avoid generating (plus X (const_int 0)).
2951 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
2952 in-place. Pass true to plus_constant.
2953 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
2955 2014-05-16 Dehao Chen <dehao@google.com>
2957 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
2959 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
2962 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
2964 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
2966 2014-05-16 Dehao Chen <dehao@google.com>
2968 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
2969 optimize_function_for_size_p.
2970 * regs.h (REG_FREQ_FROM_BB): Likewise.
2972 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
2975 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
2976 negt_reg_operand cases.
2977 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
2979 * config/sh/predicates.md (cbranch_treg_value): Simplify.
2981 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
2983 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
2986 2014-05-16 David Malcolm <dmalcolm@redhat.com>
2989 2014-04-29 David Malcolm <dmalcolm@redhat.com>
2991 * tree-cfg.c (dump_function_to_file): Dump the return type of
2992 functions, in a line to itself before the function body, mimicking
2993 the layout of a C function.
2995 2014-05-16 Dehao Chen <dehao@google.com>
2997 * cfghooks.c (make_forwarder_block): Use direct computation to
2998 get fall-through edge's count and frequency.
3000 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
3002 * config/arc/arc.c (arc_init): Fix typo in error message.
3003 * config/i386/i386.c (ix86_expand_builtin): Likewise.
3004 (split_stack_prologue_scratch_regno): Likewise.
3005 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
3006 word from error message.
3008 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
3010 * ira-costs.c: Fix typo in comment.
3012 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
3014 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
3016 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
3018 * varpool.c (dump_varpool_node): Dump write-only flag.
3019 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
3021 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
3022 write-only variables.
3023 * ipa.c (process_references): New function.
3024 (set_readonly_bit): New function.
3025 (set_writeonly_bit): New function.
3026 (clear_addressable_bit): New function.
3027 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
3028 fix handling of aliases.
3029 * cgraph.h (struct varpool_node): Add writeonly flag.
3031 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
3033 PR rtl-optimization/60969
3034 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
3035 Calculate costs for this case.
3037 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
3039 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
3040 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
3042 2014-05-16 Richard Biener <rguenther@suse.de>
3044 PR tree-optimization/61194
3045 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
3046 bool patterns ending in a COND_EXPR.
3048 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3050 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
3052 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3054 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
3055 where we were unable to cost an RTX.
3057 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3059 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
3062 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3063 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
3065 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
3067 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3068 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
3070 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
3071 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
3073 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3074 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
3076 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
3079 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3080 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
3082 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
3085 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3086 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
3088 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
3089 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
3091 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3092 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
3094 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
3097 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3098 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
3100 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
3101 ZERO_EXTEND and SIGN_EXTEND better.
3103 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3104 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
3106 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
3109 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3110 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
3112 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
3113 costs when costing loads and stores to memory.
3115 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3116 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
3118 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
3121 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3123 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
3125 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3126 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
3128 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
3130 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
3132 (aarch64_rtx_mult_cost): New.
3133 (aarch64_rtx_costs): Use it, refactor as appropriate.
3135 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3136 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
3138 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
3139 emit instructions, return number of instructions which would
3141 (aarch64_add_constant): Update call to aarch64_build_constant.
3142 (aarch64_output_mi_thunk): Likewise.
3143 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
3146 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3148 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
3149 (TARGET_RTX_COSTS): Call it.
3151 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3153 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
3154 (cortexa57_vector_cost): Likewise.
3155 (cortexa57_tunings): Use them.
3157 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3159 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
3160 (cpu_addrcost_table): Use it.
3161 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
3162 (aarch64_address_cost): Rewrite using aarch64_classify_address,
3165 2014-05-16 Richard Biener <rguenther@suse.de>
3167 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
3168 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
3169 (visit_phi): Ignore edges marked as not executable.
3170 (class cond_dom_walker): New.
3171 (cond_dom_walker::before_dom_children): Value-number
3172 control statements and mark successor edges as not
3173 executable if possible.
3174 (run_scc_vn): First walk all control statements in
3175 dominator order, marking edges as not executable.
3176 * tree-inline.c (copy_edges_for_bb): Be not confused
3177 about random edge flags.
3179 2014-05-16 Richard Biener <rguenther@suse.de>
3181 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
3183 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
3186 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
3187 (__TM_simple_begin): Use it.
3188 (__TM_begin): Likewise.
3190 2014-05-15 Martin Jambor <mjambor@suse.cz>
3193 * ipa-prop.c (update_indirect_edges_after_inlining): Check
3194 type_preserved flag when the indirect edge is polymorphic.
3196 2014-05-15 Martin Jambor <mjambor@suse.cz>
3198 PR tree-optimization/61090
3199 * tree-sra.c (sra_modify_expr): Pass the current gsi to
3200 build_ref_for_model.
3202 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3204 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
3205 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
3207 2014-05-15 Jakub Jelinek <jakub@redhat.com>
3209 PR tree-optimization/61158
3210 * fold-const.c (fold_binary_loc): If X is zero-extended and
3211 shiftc >= prec, make sure zerobits is all ones instead of
3212 invoking undefined behavior.
3214 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3216 * regcprop.h: New file.
3217 * regcprop.c (skip_debug_insn_p): New decl.
3218 (replace_oldest_value_reg): Check skip_debug_insn_p.
3219 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
3220 * shrink-wrap.c: Include regcprop.h.
3221 (prepare_shrink_wrap): Call
3222 copyprop_hardreg_forward_bb_without_debug_insn.
3224 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3226 * shrink-wrap.h: Update comment.
3227 * shrink-wrap.c: Update comment.
3228 (next_block_for_reg): Rename to live_edge_for_reg.
3229 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
3230 (move_insn_for_shrink_wrap): Split live_edge.
3231 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
3233 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
3235 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
3237 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
3238 * config/sparc/sparc.md (fptype_ut699): New attribute.
3239 (in_branch_delay): Return false if -mfix-ut699 is specified and
3240 fptype_ut699 is set to single.
3241 (truncdfsf2): Add fptype_ut699 attribute.
3242 (fix_truncdfsi2): Likewise.
3243 (floatsisf2): Change fptype attribute.
3244 (fix_truncsfsi2): Likewise.
3245 (negtf2_notv9): Delete.
3246 (negtf2_v9): Likewise.
3247 (negtf2_hq): New instruction.
3248 (negtf2): New instruction and splitter.
3249 (negdf2_notv9): Rewrite.
3250 (abstf2_notv9): Delete.
3251 (abstf2_hq_v9): Likewise.
3252 (abstf2_v9): Likewise.
3253 (abstf2_hq): New instruction.
3254 (abstf2): New instruction and splitter.
3255 (absdf2_notv9): Rewrite.
3257 2014-05-14 Cary Coutant <ccoutant@google.com>
3260 * opts.c (common_handle_option): Don't special-case "-g".
3261 (set_debug_level): Default to at least level 2 with "-g".
3263 2014-05-14 DJ Delorie <dj@redhat.com>
3265 * config/msp430/msp430.c (msp430_builtin): Add
3266 MSP430_BUILTIN_DELAY_CYCLES.
3267 (msp430_init_builtins): Register void __delay_cycles(long long).
3268 (msp430_builtin_decl): Add it.
3269 (cg_magic_constant): New.
3270 (msp430_expand_delay_cycles): New.
3271 (msp430_expand_builtin): Call it.
3272 (msp430_print_operand_raw): Change integer printing from "int" to
3274 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
3275 (delay_cycles_start): New.
3276 (delay_cycles_end): New.
3277 (delay_cycles_32): New.
3278 (delay_cycles_32x): New.
3279 (delay_cycles_16): New.
3280 (delay_cycles_16x): New.
3281 (delay_cycles_2): New.
3282 (delay_cycles_1): New.
3283 * doc/extend.texi: Document __delay_cycles().
3285 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
3287 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
3288 length attribute computation.
3290 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
3293 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
3295 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
3298 * config/sparc/sparc.md: Fix types of low and high in DI constant
3299 splitter. Use gen_int_mode in some other splitters.
3301 2014-05-14 Martin Jambor <mjambor@suse.cz>
3304 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
3306 2014-05-14 James Norris <jnorris@codesourcery.com>
3308 * omp-low.c (expand_parallel_call): Remove shadow variable.
3309 (expand_omp_taskreg): Likewise.
3311 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
3313 * common/config/i386/i386-common.c
3314 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
3315 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
3316 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
3317 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
3318 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
3319 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
3320 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
3321 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
3322 xsavecintrin.h, xsavesintrin.h.
3323 (x86_64-*-*): Ditto.
3324 * config/i386/clflushoptintrin.h: New.
3325 * config/i386/xsavecintrin.h: Ditto.
3326 * config/i386/xsavesintrin.h: Ditto.
3327 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
3328 (bit_XSAVES): Ditto.
3329 (bit_XSAVES): Ditto.
3330 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
3331 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
3333 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
3334 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
3335 OPTION_MASK_ISA_XSAVES.
3336 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
3338 (PTA_CLFLUSHOPT) Define.
3339 (PTA_XSAVEC): Ditto.
3340 (PTA_XSAVES): Ditto.
3341 (ix86_option_override_internal): Handle new options.
3342 (ix86_valid_target_attribute_inner_p): Ditto.
3343 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
3344 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
3345 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
3346 (bdesc_special_args): Add __builtin_ia32_xsaves,
3347 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
3348 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
3349 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
3350 (ix86_expand_builtin): Handle new builtins.
3351 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
3352 (TARGET_CLFLUSHOPT_P): Ditto.
3353 (TARGET_XSAVEC): Ditto.
3354 (TARGET_XSAVEC_P): Ditto.
3355 (TARGET_XSAVES): Ditto.
3356 (TARGET_XSAVES_P): Ditto.
3357 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
3358 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
3359 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
3361 (ANY_XRSTOR64): Ditto.
3363 (xrstor): Change into <xrstor>.
3364 (xrstor_rex64): Change into <xrstor>_rex64.
3365 (xrstor64): Change into <xrstor>64
3367 * config/i386/i386.opt (mclflushopt): New.
3370 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
3372 * doc/invoke.texi: Document new options.
3374 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
3376 PR rtl-optimization/60866
3377 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
3378 Default it to -1. Pass it down to init_simplejump_data.
3379 (init_simplejump_data): New parameter old_seqno. Pass it down
3380 to get_seqno_for_a_jump.
3381 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
3382 initializing new jump seqno as a last resort. Add comment.
3383 (sel_redirect_edge_and_branch): Save old seqno of the conditional
3384 jump and pass it down to sel_init_new_insn.
3385 (sel_redirect_edge_and_branch_force): Likewise.
3387 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
3389 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
3390 shifted values to avoid build warning.
3392 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
3394 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
3395 * cfgrtl.c (rtl_merge_blocks): Fix comment.
3396 (cfg_layout_merge_blocks): Likewise.
3397 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
3399 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
3401 PR rtl-optimization/60901
3402 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
3403 bb predecessor belongs to the same scheduling region. Adjust comment.
3405 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
3407 * doc/sourcebuild.texi: (dfp_hw): Document.
3408 (p8vector_hw): Likewise.
3409 (powerpc_eabi_ok): Likewise.
3410 (powerpc_elfv2): Likewise.
3411 (powerpc_htm_ok): Likewise.
3412 (ppc_recip_hw): Likewise.
3415 2014-05-13 Cary Coutant <ccoutant@google.com>
3417 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
3419 2014-05-13 David Malcolm <dmalcolm@redhat.com>
3421 * gengtype-parse.c (require3): Eliminate in favor of...
3423 (require_template_declaration): Update to support optional single *
3426 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
3427 (create_user_defined_type): Handle a single level of explicit
3428 pointerness within template arguments.
3429 (struct write_types_data): Add field "kind".
3430 (filter_type_name): Handle "*" character.
3431 (write_user_func_for_structure_ptr): Require a write_types_data
3432 rather than just a prefix string, so that we can look up the kind
3433 of the wtd and use it as an index into wrote_user_func_for_ptr,
3434 ensuring that such functions are written at most once. Support
3435 subclasses by invoking the marking function of the ultimate base class.
3436 (write_user_func_for_structure_body): Require a write_types_data
3437 rather than just a prefix string, so that we can pass this to
3438 write_user_func_for_structure_ptr.
3439 (write_func_for_structure): Likewise.
3440 (ggc_wtd): Add initializer of new "kind" field.
3441 (pch_wtd): Likewise.
3443 * gengtype.h (enum write_types_kinds): New.
3444 (struct type): Add field wrote_user_func_for_ptr to the "s"
3447 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
3449 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
3450 instead of const_binop.
3451 (fold_binary_loc): Likewise.
3453 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
3455 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
3456 calculation to match get_ref_base_and_extent.
3458 2014-05-13 Catherine Moore <clm@codesourcery.com>
3459 Sandra Loosemore <sandra@codesourcery.com>
3461 * configure.ac: Fix assembly for explicit JALR relocation check.
3462 * configure: Regenerate.
3464 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3466 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
3467 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
3468 Remove associated type declarations and initialisations.
3469 (arm_expand_neon_builtin): Likewise.
3470 (neon_emit_pair_result_insn): Delete.
3471 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
3472 * config/arm/neon.md (neon_vtrn<mode>): Delete.
3473 (neon_vzip<mode>): Likewise.
3474 (neon_vuzp<mode>): Likewise.
3476 2014-05-13 Richard Biener <rguenther@suse.de>
3479 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
3480 it needs revisiting whether the call still may be tail-called.
3482 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3484 * rtl.def (SYMBOL_REF): Remove middle "0" field.
3485 * rtl.h (block_symbol): Reduce number of fields to 2.
3486 (rtx_def): Add u2.symbol_ref_flags.
3487 (SYMBOL_REF_FLAGS): Use it.
3488 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
3489 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
3490 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
3491 Lower index of SYMBOL_REF_DATA.
3492 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
3493 Print SYMBOL_REF_FLAGS at the same time.
3494 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
3496 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3498 * rtl.def (VAR_LOCATION): Remove "i" field.
3499 * rtl.h (rtx_def): Add u2.var_location_status.
3500 (PAT_VAR_LOCATION_STATUS): Use it.
3501 (gen_rtx_VAR_LOCATION): Declare.
3502 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
3503 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
3504 * var-tracking.c (emit_note_insn_var_location): Remove casts.
3506 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3508 * rtl.def (scratch): Fix outdated comment and remove "0" field.
3509 * gengtype.c (adjust_field_rtx_def): Update accordingly.
3511 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3513 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
3514 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
3515 * rtl.h (rtx_def): Add insn_uid to u2 field.
3516 (RTX_FLAG_CHECK8): Delete in favor of...
3517 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
3518 (INSN_DELETED_P): Update accordingly.
3519 (INSN_UID): Use u2.insn_uid.
3520 (INSN_CHAIN_CODE_P): Define.
3521 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
3522 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
3523 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
3524 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
3525 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
3526 indices accordingly.
3527 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
3528 Update indices for insn-chain rtxes.
3529 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
3530 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
3531 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
3532 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
3533 * combine.c (try_combine): Likewise.
3534 * ira.c (setup_prohibited_mode_move_regs): Likewise.
3536 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3538 * rtl.def (REG): Remove middle field.
3539 * rtl.h (rtx_def): Add orignal_regno to u2.
3540 (ORIGINAL_REGNO): Use it instead of field 1.
3541 (REG_ATTRS): Lower field index accordingly.
3542 * gengtype.c (adjust_field_rtx_def): Remove handling of
3543 ORIGINAL_REGNO. Move REG_ATTRS index down.
3544 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
3545 code that prints the REGNO.
3547 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3549 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
3552 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3554 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
3556 2014-05-13 Bin Cheng <bin.cheng@arm.com>
3558 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
3559 (alloc_iv): Lower base expressions containing ADDR_EXPR.
3561 2014-05-13 Ian Bolton <ian.bolton@arm.com>
3563 * config/aarch64/aarch64-protos.h
3564 (aarch64_hard_regno_caller_save_mode): New prototype.
3565 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
3567 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
3569 2014-05-13 Christian Bruel <christian.bruel@st.com>
3571 * target.def (mode_switching): New hook vector.
3572 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
3573 (mode_exit, modepriority_to_mode): Likewise.
3574 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
3575 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
3576 * target.h: Include tm.h and hard-reg-set.h.
3577 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
3578 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
3579 * doc/tm.texi Regenerate.
3580 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
3581 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
3582 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
3583 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
3584 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
3585 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
3586 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
3587 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
3588 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
3589 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
3590 (ix86_emit_mode_set): Hookify.
3591 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
3593 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
3594 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
3595 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
3596 (epiphany_mode_priority_to_mode): Remove declaration.
3597 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
3598 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
3599 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
3601 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
3602 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
3603 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
3605 2014-05-13 Jakub Jelinek <jakub@redhat.com>
3608 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
3609 is const0_rtx, return immediately. Don't test count == 0 when
3612 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3614 * Makefile.in: add shrink-wrap.o.
3615 * config/i386/i386.c: include "shrink-wrap.h"
3616 * function.c: Likewise.
3617 (requires_stack_frame_p, next_block_for_reg,
3618 move_insn_for_shrink_wrap, prepare_shrink_wrap,
3619 dup_block_and_redirect): Move to shrink-wrap.c
3620 (thread_prologue_and_epilogue_insns): Extract three code segments
3621 as functions in shrink-wrap.c
3622 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
3624 * shrink-wrap.c: New file.
3625 * shrink-wrap.h: New file.
3627 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
3629 * doc/extend.texi: Reflect current numbers of pragmas. Remove
3630 reference to Solaris.
3632 2014-05-12 Mike Stump <mikestump@comcast.net>
3635 * genattrtab.c (filename): Add.
3636 (convert_set_attr_alternative): Improve error message.
3637 (check_defs): Restore read_md_filename for error messages.
3638 (gen_insn): Save filename.
3640 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
3642 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
3643 -fno-local-ivars and -fivar-visibility.
3644 * c-family/c.opt: Make -Wshadow also implicitly enable
3647 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
3649 * doc/tm.texi: Remove reference to deleted macro.
3650 * doc/tm.texi.in: Likewise.
3652 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
3655 * config/avr/avr.c (avr_out_store_psi): Use correct constant
3658 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
3661 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
3662 * config/arm/aout.h (License): Same.
3663 * config/arm/bpabi.h (License): Same.
3664 * config/arm/elf.h (License): Same.
3665 * config/arm/linux-elf.h (License): Same.
3666 * config/arm/linux-gas.h (License): Same.
3667 * config/arm/netbsd-elf.h (License): Same.
3668 * config/arm/uclinux-eabi.h (License): Same.
3669 * config/arm/uclinux-elf.h (License): Same.
3670 * config/arm/vxworks.h (License): Same.
3672 2014-05-11 Jakub Jelinek <jakub@redhat.com>
3674 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
3675 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
3676 number of operands to 3.
3677 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
3678 * tree-nested.c (convert_nonlocal_omp_clauses,
3679 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
3680 * gimplify.c (gimplify_scan_omp_clauses): Handle
3681 OMP_CLAUSE_LINEAR_STMT.
3682 * omp-low.c (lower_rec_input_clauses): Fix typo.
3683 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
3684 cast between Fortran boolean_type_node and C _Bool if
3687 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
3689 PR tree-optimization/61136
3690 * wide-int.h (multiple_of_p): Define a version that doesn't return
3692 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
3693 integer_zerop/const_binop pair.
3694 (multiple_of_p): Likewise, converting both operands to widest_int
3697 2014-05-09 Teresa Johnson <tejohnson@google.com>
3699 * cgraphunit.c (analyze_functions): Use correct dump file.
3701 2014-05-09 Florian Weimer <fweimer@redhat.com>
3703 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
3705 (stack_protect_return_slot_p): New function.
3706 (expand_used_vars): Call stack_protect_decl_p and
3707 stack_protect_return_slot_p for -fstack-protector-strong.
3709 2014-05-09 David Wohlferd <LimeGreenSocks@yahoo.com>
3710 Andrew Haley <aph@redhat.com>
3711 Richard Sandiford <rdsandiford@googlemail.com>
3713 * doc/extend.texi: Rewrite inline asm page / re-org asm-related
3716 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
3719 * fold-const.c (fold_binary_loc): Changed width of mask.
3721 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
3723 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
3724 unsigned int initializers for regno_in, regno_out.
3726 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
3729 * config/avr/avr.md (cc): Add new attribute set_vzn.
3730 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
3731 Set cc insn attribute to set_vzn instead of set_zn for alternatives
3732 with INC, DEC or NEG.
3733 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
3734 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
3735 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
3737 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3740 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3742 * wide-int.cc (UTItype): Define.
3743 (UDWtype): Define for appropriate W_TYPE_SIZE.
3745 2014-05-09 Richard Biener <rguenther@suse.de>
3747 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
3748 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
3749 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
3750 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
3751 ssa_propagate): Adjust.
3753 2014-05-08 Jeff Law <law@redhat.com>
3755 PR tree-optimization/61009
3756 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
3757 tri-state rather than a boolean. When a block is too big to
3758 thread through, inform caller via negative return value.
3759 (thread_across_edge): If a block was too big for normal threading,
3760 then it's too big for a joiner too, so remove temporary equivalences
3761 and return immediately.
3763 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
3764 Matthias Klose <doko@ubuntu.com>
3767 * optc-gen.awk: Fix option handling for -Wunused-parameter.
3769 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
3772 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
3774 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
3777 * config/alpha/alpha.c: Include gimple-iterator.h.
3778 (alpha_gimple_fold_builtin): New function. Move
3779 ALPHA_BUILTIN_UMULH folding from ...
3780 (alpha_fold_builtin): ... here.
3781 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
3783 2014-05-08 Wei Mi <wmi@google.com>
3786 * config/i386/i386.c (ix86_compute_frame_layout): Update
3787 preferred_stack_boundary for call, expanded from tls descriptor.
3788 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
3789 to depend on SP register.
3790 (*tls_local_dynamic_base_32_gnu): Ditto.
3791 (*tls_local_dynamic_32_once): Ditto.
3792 (tls_global_dynamic_64_<mode>): Set
3793 ix86_tls_descriptor_calls_expanded_in_cfun.
3794 (tls_local_dynamic_base_64_<mode>): Ditto.
3795 (tls_global_dynamic_32): Set
3796 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
3797 to depend on SP register.
3798 (tls_local_dynamic_base_32): Ditto.
3800 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3802 * config/arm/arm_neon.h: Update comment.
3803 * config/arm/neon-docgen.ml: Delete.
3804 * config/arm/neon-gen.ml: Delete.
3805 * doc/arm-neon-intrinsics.texi: Update comment.
3807 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3809 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
3811 (vand, vorr, veor, vorn, vbic): Remove.
3812 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
3814 (neon_vsub_unspec): Likewise.
3815 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
3817 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3819 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
3820 (vadd_s16): Likewise.
3821 (vadd_s32): Likewise.
3822 (vadd_f32): Likewise.
3823 (vadd_u8): Likewise.
3824 (vadd_u16): Likewise.
3825 (vadd_u32): Likewise.
3826 (vadd_s64): Likewise.
3827 (vadd_u64): Likewise.
3828 (vaddq_s8): Likewise.
3829 (vaddq_s16): Likewise.
3830 (vaddq_s32): Likewise.
3831 (vaddq_s64): Likewise.
3832 (vaddq_f32): Likewise.
3833 (vaddq_u8): Likewise.
3834 (vaddq_u16): Likewise.
3835 (vaddq_u32): Likewise.
3836 (vaddq_u64): Likewise.
3837 (vmul_s8): Likewise.
3838 (vmul_s16): Likewise.
3839 (vmul_s32): Likewise.
3840 (vmul_f32): Likewise.
3841 (vmul_u8): Likewise.
3842 (vmul_u16): Likewise.
3843 (vmul_u32): Likewise.
3844 (vmul_p8): Likewise.
3845 (vmulq_s8): Likewise.
3846 (vmulq_s16): Likewise.
3847 (vmulq_s32): Likewise.
3848 (vmulq_f32): Likewise.
3849 (vmulq_u8): Likewise.
3850 (vmulq_u16): Likewise.
3851 (vmulq_u32): Likewise.
3852 (vsub_s8): Likewise.
3853 (vsub_s16): Likewise.
3854 (vsub_s32): Likewise.
3855 (vsub_f32): Likewise.
3856 (vsub_u8): Likewise.
3857 (vsub_u16): Likewise.
3858 (vsub_u32): Likewise.
3859 (vsub_s64): Likewise.
3860 (vsub_u64): Likewise.
3861 (vsubq_s8): Likewise.
3862 (vsubq_s16): Likewise.
3863 (vsubq_s32): Likewise.
3864 (vsubq_s64): Likewise.
3865 (vsubq_f32): Likewise.
3866 (vsubq_u8): Likewise.
3867 (vsubq_u16): Likewise.
3868 (vsubq_u32): Likewise.
3869 (vsubq_u64): Likewise.
3870 (vand_s8): Likewise.
3871 (vand_s16): Likewise.
3872 (vand_s32): Likewise.
3873 (vand_u8): Likewise.
3874 (vand_u16): Likewise.
3875 (vand_u32): Likewise.
3876 (vand_s64): Likewise.
3877 (vand_u64): Likewise.
3878 (vandq_s8): Likewise.
3879 (vandq_s16): Likewise.
3880 (vandq_s32): Likewise.
3881 (vandq_s64): Likewise.
3882 (vandq_u8): Likewise.
3883 (vandq_u16): Likewise.
3884 (vandq_u32): Likewise.
3885 (vandq_u64): Likewise.
3886 (vorr_s8): Likewise.
3887 (vorr_s16): Likewise.
3888 (vorr_s32): Likewise.
3889 (vorr_u8): Likewise.
3890 (vorr_u16): Likewise.
3891 (vorr_u32): Likewise.
3892 (vorr_s64): Likewise.
3893 (vorr_u64): Likewise.
3894 (vorrq_s8): Likewise.
3895 (vorrq_s16): Likewise.
3896 (vorrq_s32): Likewise.
3897 (vorrq_s64): Likewise.
3898 (vorrq_u8): Likewise.
3899 (vorrq_u16): Likewise.
3900 (vorrq_u32): Likewise.
3901 (vorrq_u64): Likewise.
3902 (veor_s8): Likewise.
3903 (veor_s16): Likewise.
3904 (veor_s32): Likewise.
3905 (veor_u8): Likewise.
3906 (veor_u16): Likewise.
3907 (veor_u32): Likewise.
3908 (veor_s64): Likewise.
3909 (veor_u64): Likewise.
3910 (veorq_s8): Likewise.
3911 (veorq_s16): Likewise.
3912 (veorq_s32): Likewise.
3913 (veorq_s64): Likewise.
3914 (veorq_u8): Likewise.
3915 (veorq_u16): Likewise.
3916 (veorq_u32): Likewise.
3917 (veorq_u64): Likewise.
3918 (vbic_s8): Likewise.
3919 (vbic_s16): Likewise.
3920 (vbic_s32): Likewise.
3921 (vbic_u8): Likewise.
3922 (vbic_u16): Likewise.
3923 (vbic_u32): Likewise.
3924 (vbic_s64): Likewise.
3925 (vbic_u64): Likewise.
3926 (vbicq_s8): Likewise.
3927 (vbicq_s16): Likewise.
3928 (vbicq_s32): Likewise.
3929 (vbicq_s64): Likewise.
3930 (vbicq_u8): Likewise.
3931 (vbicq_u16): Likewise.
3932 (vbicq_u32): Likewise.
3933 (vbicq_u64): Likewise.
3934 (vorn_s8): Likewise.
3935 (vorn_s16): Likewise.
3936 (vorn_s32): Likewise.
3937 (vorn_u8): Likewise.
3938 (vorn_u16): Likewise.
3939 (vorn_u32): Likewise.
3940 (vorn_s64): Likewise.
3941 (vorn_u64): Likewise.
3942 (vornq_s8): Likewise.
3943 (vornq_s16): Likewise.
3944 (vornq_s32): Likewise.
3945 (vornq_s64): Likewise.
3946 (vornq_u8): Likewise.
3947 (vornq_u16): Likewise.
3948 (vornq_u32): Likewise.
3949 (vornq_u64): Likewise.
3951 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3953 * wide-int.cc (UTItype): Define.
3954 (UDWtype): Define for appropriate W_TYPE_SIZE.
3956 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
3958 PR tree-optimization/59100
3959 * tree-ssa-phiopt.c: Include tree-inline.h.
3960 (neutral_element_p, absorbing_element_p): New functions.
3961 (value_replacement): Handle conditional binary operations with a
3962 neutral or absorbing element.
3964 2014-05-08 Richard Biener <rguenther@suse.de>
3966 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
3967 folding the expression.
3968 (valueize_expr): Remove.
3969 (visit_reference_op_load): Do not valueize the result of
3971 (simplify_binary_expression): Likewise.
3972 (simplify_unary_expression): Likewise.
3974 2014-05-08 Richard Biener <rguenther@suse.de>
3976 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
3977 looking at TYPE_ARG_TYPES.
3979 2014-05-08 Richard Biener <rguenther@suse.de>
3981 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
3982 pointer propagation special-case.
3984 2014-05-08 Bin Cheng <bin.cheng@arm.com>
3986 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
3987 core part of address expressions.
3989 2014-05-08 Alan Modra <amodra@gmail.com>
3992 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
3993 loads and stores when -mno-strict-align at any alignment.
3994 (expand_block_clear): Similarly. Also correct calculation of
3997 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
4000 * tree-complex.c (expand_complex_move): Keep line info when expanding
4002 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
4003 of complex expression. Use new argument to display correct location
4004 for values coming from phi statement.
4005 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
4006 (warn_uninitialized_phi): Pass location of phi argument to
4008 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
4009 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
4011 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
4013 * config/rs6000/predicates.md (indexed_address_mem): New.
4014 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
4015 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
4016 fpstore_ux, fpstore_u.
4017 (sign_extend, indexed, update): New.
4018 (cell_micro): Adjust.
4019 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
4020 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
4021 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
4022 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
4023 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
4024 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
4025 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
4026 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
4027 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
4028 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
4029 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
4030 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
4031 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
4032 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
4033 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
4035 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
4036 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
4037 *vsx_extract_<mode>_store): Adjust.
4038 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
4039 is_cracked_insn, insn_must_be_first_in_group,
4040 insn_must_be_last_in_group): Adjust.
4042 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
4044 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
4045 ppc440-fpstore): Adjust.
4046 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
4047 ppc476-fpstore): Adjust.
4048 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
4049 ppc601-fpstore): Adjust.
4050 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
4052 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
4054 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
4055 ppc7450-fpstore): Adjust.
4056 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
4057 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
4058 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
4060 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
4061 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
4062 cell-fpstore, cell-fpstore-update): Adjust.
4063 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
4064 ppce300c3_store, ppce300c3_fpstore): Adjust.
4065 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
4066 e500mc_fpstore): Adjust.
4067 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
4068 e500mc64_store, e500mc64_fpstore): Adjust.
4069 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
4070 e5500_fpstore): Adjust.
4071 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
4072 e6500_fpstore): Adjust.
4073 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
4075 * config/rs6000/power4.md (power4-load, power4-load-ext,
4076 power4-load-ext-update, power4-load-ext-update-indexed,
4077 power4-load-update-indexed, power4-load-update, power4-fpload,
4078 power4-fpload-update, power4-store, power4-store-update,
4079 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
4081 * config/rs6000/power5.md (power5-load, power5-load-ext,
4082 power5-load-ext-update, power5-load-ext-update-indexed,
4083 power5-load-update-indexed, power5-load-update, power5-fpload,
4084 power5-fpload-update, power5-store, power5-store-update,
4085 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
4087 * config/rs6000/power6.md (power6-load, power6-load-ext,
4088 power6-load-update, power6-load-update-indexed,
4089 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
4090 power6-fpload-update, power6-store, power6-store-update,
4091 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
4093 * config/rs6000/power7.md (power7-load, power7-load-ext,
4094 power7-load-update, power7-load-update-indexed,
4095 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
4096 power7-fpload-update, power7-store, power7-store-update,
4097 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
4099 * config/rs6000/power8.md (power8-load, power8-load-update,
4100 power8-load-ext, power8-load-ext-update, power8-fpload,
4101 power8-fpload-update, power8-store, power8-store-update-indexed,
4102 power8-fpstore, power8-fpstore-update): Adjust.
4103 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
4105 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
4106 titan_lsu_store, titan_lsu_fpstore): Adjust.
4107 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
4109 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
4112 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
4113 unrolled byte insns. Emit address increments after move insns.
4115 2014-05-07 David Malcolm <dmalcolm@redhat.com>
4117 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
4118 const_gimple, rather than a gimple.
4119 (gimple_call_builtin_p): Likewise, for the three variants.
4121 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
4122 (gimple_call_builtin_p): Likewise, for the three variants.
4124 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
4126 PR tree-optimization/61095
4127 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
4129 2014-05-07 Richard Biener <rguenther@suse.de>
4131 PR tree-optimization/61034
4132 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
4133 (maybe_skip_until): Use translate to take into account
4134 lattices when trying to do disambiguations.
4135 (get_continuation_for_phi_1): Likewise.
4136 (get_continuation_for_phi): Adjust for added translate arguments.
4137 (walk_non_aliased_vuses): Likewise.
4138 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
4139 (walk_non_aliased_vuses): Likewise.
4140 (call_may_clobber_ref_p_1): Declare.
4141 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
4142 calls. Stop early if we are only supposed to disambiguate.
4143 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
4145 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
4147 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
4148 Emit an error when the function has arguments.
4150 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
4152 * cfgloop.h (unswitch_loops): Remove.
4153 * doc/passes.texi: Remove references to loop-unswitch.c
4154 * timevar.def (TV_LOOP_UNSWITCH): Remove.
4156 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
4158 * tree-vect-data-refs.c (vect_grouped_load_supported): New
4159 check for loads group of length 3.
4160 (vect_permute_load_chain): New permutations for loads group of
4162 * tree-vect-stmts.c (vect_model_load_cost): Change cost
4163 of vec_perm_shuffle for the new permutations.
4165 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
4167 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
4168 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
4169 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
4170 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
4171 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
4172 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
4173 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
4174 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
4176 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
4178 * loop-unswitch.c: Delete.
4180 2014-05-07 Richard Biener <rguenther@suse.de>
4182 * config.gcc: Always set need_64bit_hwint to yes.
4184 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
4186 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
4187 of using optimize_size.
4189 2014-05-06 Mike Stump <mikestump@comcast.net>
4191 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
4193 2014-05-06 Joseph Myers <joseph@codesourcery.com>
4195 * config/i386/sse.md (*mov<mode>_internal)
4196 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
4197 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
4198 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
4199 (*<code><mode>3, *andnot<mode>3<mask_name>)
4200 (<mask_codefor><code><mode>3<mask_name>): Only consider
4201 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
4203 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
4206 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
4208 * lra-constraints.c (valid_address_p): Move earlier in file.
4209 Add a constraint argument to the address_info version.
4210 (satisfies_memory_constraint_p): New function.
4211 (satisfies_address_constraint_p): Likewise.
4212 (process_alt_operands, curr_insn_transform): Use them.
4213 (process_address): Pass the constraint to valid_address_p when
4214 checking address operands.
4216 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
4218 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
4219 to their respective blocks. Fix inadvertent use of "node".
4221 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
4223 * emit-rtl.c (init_derived_machine_modes): New functionm, split
4225 (init_emit_once): ...here.
4226 * rtl.h (init_derived_machine_modes): Declare.
4227 * toplev.c (do_compile): Call it even if no_backend.
4229 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
4230 Mike Stump <mikestump@comcast.net>
4231 Richard Sandiford <rdsandiford@googlemail.com>
4232 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4234 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
4235 (rtx_equal_for_memref_p): Update comment.
4236 (adjust_offset_for_component_ref): Use wide-int interfaces.
4237 * builtins.c (get_object_alignment_2): Likewise.
4238 (c_readstr): Likewise.
4239 (target_char_cast): Add comment.
4240 (determine_block_size): Use wide-int interfaces.
4241 (expand_builtin_signbit): Likewise.
4242 (fold_builtin_int_roundingfn): Likewise.
4243 (fold_builtin_bitop): Likewise.
4244 (fold_builtin_bswap): Likewise.
4245 (fold_builtin_logarithm): Use signop.
4246 (fold_builtin_pow): Likewise.
4247 (fold_builtin_memory_op): Use wide-int interfaces.
4248 (fold_builtin_object_size): Likewise.
4249 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
4250 nb_iterations_estimate.
4251 (record_niter_bound): Use wide-int interfaces.
4252 (get_estimated_loop_iterations_int): Likewise.
4253 (get_estimated_loop_iterations): Likewise.
4254 (get_max_loop_iterations): Likewise.
4255 * cfgloop.h: Include wide-int.h.
4256 (struct nb_iter_bound): Change bound to widest_int.
4257 (struct loop): Change nb_iterations_upper_bound and
4258 nb_iterations_estimate to widest_int.
4259 (record_niter_bound): Switch to use widest_int.
4260 (get_estimated_loop_iterations): Likewise.
4261 (get_max_loop_iterations): Likewise.
4262 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
4263 update for wide-int.
4264 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
4265 * combine.c (try_combine): Likewise.
4266 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
4267 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
4269 (aarch64_float_const_representable_p): Likewise.
4270 * config/arc/arc.c: Include wide-int.h.
4271 (arc_can_use_doloop_p): Use wide-int interfaces.
4272 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
4273 (vfp3_const_double_index): Likewise.
4274 * config/avr/avr.c (avr_out_round): Likewise.
4275 (avr_fold_builtin): Likewise.
4276 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
4277 (bfin_can_use_doloop_p): Likewise.
4278 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
4279 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
4280 * config/i386/i386.c: Include wide-int.h.
4281 (ix86_data_alignment): Use wide-int interfaces.
4282 (ix86_local_alignment): Likewise.
4283 (ix86_emit_swsqrtsf): Update real_from_integer.
4284 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
4285 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
4286 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
4287 (zero_constant): Likewise.
4288 (input_operand): Likewise.
4289 (splat_input_operand): Likewise.
4290 (non_logical_cint_operand): Change const_double to const_wide_int.
4291 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
4292 (easy_altivec_constant): Remove comment.
4293 (paired_expand_vector_init): Use CONSTANT_P.
4294 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
4295 (rs6000_emit_move): Update checks.
4296 (rs6000_aggregate_candidate): Use wide-int interfaces.
4297 (rs6000_expand_ternop_builtin): Likewise.
4298 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
4299 (rs6000_assemble_integer): Likewise.
4300 (rs6000_hash_constant): Likewise.
4301 (output_toc): Likewise.
4302 (rs6000_rtx_costs): Likewise.
4303 (rs6000_emit_swrsqrt); Update call to real_from_integer.
4304 * config/rs6000/rs6000-c.c: Include wide-int.h.
4305 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
4306 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
4307 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
4308 Handle CONST_WIDE_INT.
4309 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
4310 Use tree_fits_uhwi_p.
4311 * config/sparc/sparc.c: Include wide-int.h.
4312 (sparc_fold_builtin): Use wide-int interfaces.
4313 * config/vax/vax.c: Include wide-int.h.
4314 (vax_float_literal): Use real_from_integer.
4315 * coretypes.h (struct hwivec_def): New.
4317 (const_hwivec): New.
4318 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
4319 (equiv_constant): Handle CONST_WIDE_INT.
4320 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
4321 (cselib_hash_rtx): Handle CONST_WIDE_INT.
4322 * dbxout.c (stabstr_U): Use wide-int interfaces.
4323 (dbxout_type): Update to use cst_fits_shwi_p.
4324 * defaults.h (LOG2_BITS_PER_UNIT): Define.
4325 (TARGET_SUPPORTS_WIDE_INT): Add default.
4326 * dfp.c: Include wide-int.h.
4327 (decimal_real_to_integer2): Use wide-int interfaces and rename to
4328 decimal_real_to_integer.
4329 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
4330 decimal_real_to_integer.
4331 * doc/generic.texi (Constant expressions): Update for wide_int.
4332 * doc/rtl.texi (const_double): Likewise.
4333 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
4334 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
4335 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
4336 (REAL_VALUE_FROM_INT): Remove.
4337 (TARGET_SUPPORTS_WIDE_INT): New.
4338 * doc/tm.texi: Regenerate.
4339 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
4340 * double-int.h: Include wide-int.h.
4341 (struct wi::int_traits): New.
4342 * dwarf2out.c (get_full_len): New.
4343 (dw_val_equal_p): Add case dw_val_class_wide_int.
4344 (size_of_loc_descr): Likewise.
4345 (output_loc_operands): Likewise.
4346 (insert_double): Remove.
4347 (insert_wide_int): New.
4349 (print_die): Add case dw_val_class_wide_int.
4350 (attr_checksum): Likewise.
4351 (attr_checksum_ordered): Likewise.
4352 (same_dw_val_p): Likewise.
4353 (size_of_die): Likewise.
4354 (value_format): Likewise.
4355 (output_die): Likewise.
4356 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
4358 (clz_loc_descriptor): Use wide-int interfaces.
4359 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
4360 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
4361 (round_up_to_align): Use wide-int interfaces.
4362 (field_byte_offset): Likewise.
4363 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
4364 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
4365 CONST_DOUBLE handling. Use wide-int interfaces.
4366 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
4367 (gen_enumeration_type_die): Use add_AT_wide.
4368 (hash_loc_operands): Add case dw_val_class_wide_int.
4369 (compare_loc_operands): Likewise.
4370 * dwarf2out.h: Include wide-int.h.
4371 (wide_int_ptr): New.
4372 (enum dw_val_class): Add dw_val_class_wide_int.
4373 (struct dw_val_struct): Add val_wide.
4374 * emit-rtl.c (const_wide_int_htab): New.
4375 (const_wide_int_htab_hash): New.
4376 (const_wide_int_htab_eq): New.
4377 (lookup_const_wide_int): New.
4378 (const_double_htab_hash): Use wide-int interfaces.
4379 (const_double_htab_eq): Likewise.
4380 (rtx_to_double_int): Conditionally compile for wide-int.
4381 (immed_double_int_const): Rename to immed_wide_int_const and
4382 update for wide-int.
4383 (immed_double_const): Conditionally compile for wide-int.
4384 (init_emit_once): Use wide-int interfaces.
4385 * explow.c (plus_constant): Likewise.
4386 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
4387 (lshift_value): Use wide-int interfaces.
4388 (expand_mult): Likewise.
4389 (choose_multiplier): Likewise.
4390 (expand_smod_pow2): Likewise.
4391 (make_tree): Likewise.
4392 * expr.c (convert_modes): Consolidate handling of constants.
4393 Use wide-int interfaces.
4394 (emit_group_load_1): Add note.
4395 (store_expr): Update comment.
4396 (get_inner_reference): Use wide-int interfaces.
4397 (expand_constructor): Update comment.
4398 (expand_expr_real_2): Use wide-int interfaces.
4399 (expand_expr_real_1): Likewise.
4400 (reduce_to_bit_field_precision): Likewise.
4401 (const_vector_from_tree): Likewise.
4402 * final.c: Include wide-int-print.h.
4403 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
4404 * fixed-value.c: Include wide-int.h.
4405 (fixed_from_string): Use wide-int interfaces.
4406 (fixed_to_decimal): Likewise.
4407 (fixed_convert_from_real): Likewise.
4408 (real_convert_from_fixed): Likewise.
4409 * fold-const.h (mem_ref_offset): Return an offset_int.
4410 (div_if_zero_remainder): Remove code parameter.
4411 * fold-const.c (div_if_zero_remainder): Remove code parameter.
4412 Use wide-int interfaces.
4413 (may_negate_without_overflow_p): Use wide-int interfaces.
4414 (negate_expr_p): Likewise.
4415 (fold_negate_expr): Likewise.
4416 (int_const_binop_1): Likewise.
4417 (const_binop): Likewise.
4418 (fold_convert_const_int_from_int): Likewise.
4419 (fold_convert_const_int_from_real): Likewise.
4420 (fold_convert_const_int_from_fixed): Likewise.
4421 (fold_convert_const_fixed_from_int): Likewise.
4422 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
4423 (sign_bit_p): Use wide-int interfaces.
4424 (make_range_step): Likewise.
4425 (build_range_check): Likewise. Pass an integer of the correct type
4426 instead of using integer_one_node.
4427 (range_predecessor): Pass an integer of the correct type instead
4428 of using integer_one_node.
4429 (range_successor): Likewise.
4430 (merge_ranges): Likewise.
4431 (unextend): Use wide-int interfaces.
4432 (extract_muldiv_1): Likewise.
4433 (fold_div_compare): Likewise.
4434 (fold_single_bit_test): Likewise.
4435 (fold_sign_changed_comparison): Likewise.
4436 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
4437 (fold_plusminus_mult_expr): Use wide-int interfaces.
4438 (native_encode_int): Likewise.
4439 (native_interpret_int): Likewise.
4440 (fold_unary_loc): Likewise.
4441 (pointer_may_wrap_p): Likewise.
4442 (size_low_cst): Likewise.
4443 (mask_with_tz): Likewise.
4444 (fold_binary_loc): Likewise.
4445 (fold_ternary_loc): Likewise.
4446 (multiple_of_p): Likewise.
4447 (tree_call_nonnegative_warnv_p): Update calls to
4448 tree_int_cst_min_precision and real_from_integer.
4449 (fold_negate_const): Use wide-int interfaces.
4450 (fold_abs_const): Likewise.
4451 (fold_relational_const): Use tree_int_cst_lt.
4452 (round_up_loc): Use wide-int interfaces.
4453 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
4454 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
4455 * gengtype.c: Remove include of double-int.h.
4456 (do_typedef): Use wide-int interfaces.
4457 (open_base_files): Add wide-int.h.
4458 (main): Add offset_int and widest_int typedefs.
4459 * gengtype-lex.l: Handle "^".
4460 (CXX_KEYWORD): Add "static".
4461 * gengtype-parse.c (require3): New.
4462 (require_template_declaration): Handle constant template arguments
4463 and nested templates.
4464 * gengtype-state.c: Don't include "double-int.h".
4465 * genpreds.c (write_one_predicate_function): Update comment.
4466 (write_tm_constrs_h): Add check for hval and lval use in
4468 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
4469 (add_to_sequence): Likewise.
4470 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
4471 and const_double_operand.
4472 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
4474 * gimple-fold.c (get_base_constructor): Likewise.
4475 (fold_array_ctor_reference): Likewise.
4476 (fold_nonarray_ctor_reference): Likewise.
4477 (fold_const_aggregate_ref_1): Likewise.
4478 (gimple_val_nonnegative_real_p): Likewise.
4479 (gimple_fold_indirect_ref): Likewise.
4480 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
4481 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
4482 (struct slsr_cand_d): Change index to be widest_int.
4483 (struct incr_info_d): Change incr to be widest_int.
4484 (alloc_cand_and_find_basis): Use wide-int interfaces.
4485 (slsr_process_phi): Likewise.
4486 (backtrace_base_for_ref): Likewise. Return a widest_int.
4487 (restructure_reference): Take a widest_int instead of a double_int.
4488 (slsr_process_ref): Use wide-int interfaces.
4489 (create_mul_ssa_cand): Likewise.
4490 (create_mul_imm_cand): Likewise.
4491 (create_add_ssa_cand): Likewise.
4492 (create_add_imm_cand): Take a widest_int instead of a double_int.
4493 (slsr_process_add): Use wide-int interfaces.
4494 (slsr_process_cast): Likewise.
4495 (slsr_process_copy): Likewise.
4496 (dump_candidate): Likewise.
4497 (dump_incr_vec): Likewise.
4498 (replace_ref): Likewise.
4499 (cand_increment): Likewise. Return a widest_int.
4500 (cand_abs_increment): Likewise.
4501 (replace_mult_candidate): Take a widest_int instead of a double_int.
4502 (replace_unconditional_candidate): Use wide-int interfaces.
4503 (incr_vec_index): Take a widest_int instead of a double_int.
4504 (create_add_on_incoming_edge): Likewise.
4505 (create_phi_basis): Use wide-int interfaces.
4506 (replace_conditional_candidate): Likewise.
4507 (record_increment): Take a widest_int instead of a double_int.
4508 (record_phi_increments): Use wide-int interfaces.
4509 (phi_incr_cost): Take a widest_int instead of a double_int.
4510 (lowest_cost_path): Likewise.
4511 (total_savings): Likewise.
4512 (analyze_increments): Use wide-int interfaces.
4513 (ncd_with_phi): Take a widest_int instead of a double_int.
4514 (ncd_of_cand_and_phis): Likewise.
4515 (nearest_common_dominator_for_cands): Likewise.
4516 (insert_initializers): Use wide-int interfaces.
4517 (all_phi_incrs_profitable): Likewise.
4518 (replace_one_candidate): Likewise.
4519 (replace_profitable_candidates): Likewise.
4520 * godump.c: Include wide-int-print.h.
4521 (go_output_typedef): Use wide-int interfaces.
4522 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
4523 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
4524 (build_loop_iteration_domains): Likewise.
4525 * hooks.h: Include wide-int.h rather than double-int.h.
4526 (hook_bool_dint_dint_uint_bool_true): Delete.
4527 (hook_bool_wint_wint_uint_bool_true): Declare.
4528 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
4529 (hook_bool_wint_wint_uint_bool_true): New.
4530 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
4532 (ubsan_expand_si_overflow_mul_check): Likewise.
4533 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
4534 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
4535 (get_ancestor_addr_info): Likewise.
4536 (ipa_modify_call_arguments): Likewise.
4537 * loop-doloop.c (doloop_modify): Likewise.
4538 (doloop_optimize): Likewise.
4539 * loop-iv.c (iv_number_of_iterations): Likewise.
4540 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
4541 (unroll_loop_constant_iterations): Likewise.
4542 (decide_unroll_runtime_iterations): Likewise.
4543 (unroll_loop_runtime_iterations): Likewise.
4544 (decide_peel_simple): Likewise.
4545 (decide_unroll_stupid): Likewise.
4546 * lto-streamer-in.c (streamer_read_wi): Add.
4547 (input_cfg): Use wide-int interfaces.
4548 (lto_input_tree_1): Likewise.
4549 * lto-streamer-out.c (streamer_write_wi): Add.
4550 (hash_tree): Use wide-int interfaces.
4551 (output_cfg): Likewise.
4552 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
4553 (GTFILES): Add wide-int.h and signop.h.
4554 (TAGS): Look for .cc files too.
4555 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
4556 * optabs.c (expand_subword_shift): Likewise.
4557 (expand_doubleword_shift): Likewise.
4558 (expand_absneg_bit): Likewise.
4559 (expand_copysign_absneg): Likewise.
4560 (expand_copysign_bit): Likewise.
4561 * postreload.c (reload_cse_simplify_set): Likewise.
4562 * predict.c (predict_iv_comparison): Likewise.
4563 * pretty-print.h: Include wide-int-print.h.
4565 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
4566 * print-tree.c: Include wide-int-print.h.
4567 (print_node_brief): Use wide-int interfaces.
4568 (print_node): Likewise.
4569 * read-rtl.c (validate_const_wide_int): New.
4570 (read_rtx_code): Add CONST_WIDE_INT case.
4571 * real.c: Include wide-int.h.
4572 (real_to_integer2): Delete.
4573 (real_to_integer): New function, returning a wide_int.
4574 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
4575 (ten_to_ptwo): Update call to real_from_integer.
4576 (real_digit): Likewise.
4577 * real.h: Include signop.h, wide-int.h and insn-modes.h.
4578 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
4579 (REAL_VALUE_TO_INT): Delete.
4580 (real_to_integer): Declare a wide-int form.
4581 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
4582 * recog.c (const_int_operand): Improve comment.
4583 (const_scalar_int_operand): New.
4584 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
4585 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
4586 (split_double): Likewise.
4587 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
4588 (rtx_size): Likewise.
4589 (rtx_alloc_stat_v): New.
4590 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
4591 (cwi_output_hex): New.
4592 (iterative_hash_rtx): Handle CONST_WIDE_INT.
4593 (cwi_check_failed_bounds): New.
4594 * rtl.def (CONST_WIDE_INT): New.
4595 * rtl.h: Include <utility> and wide-int.h.
4596 (struct hwivec_def): New.
4597 (CWI_GET_NUM_ELEM): New.
4598 (CWI_PUT_NUM_ELEM): New.
4599 (struct rtx_def): Add num_elem and hwiv.
4600 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
4601 (CASE_CONST_UNIQUE): Likewise.
4602 (CASE_CONST_ANY): Likewise.
4603 (CONST_SCALAR_INT_P): Likewise.
4604 (CONST_WIDE_INT_P): New.
4606 (HWIVEC_CHECK): New.
4607 (cwi_check_failed_bounds): New.
4609 (HWIVEC_CHECK): New.
4610 (CONST_WIDE_INT_VEC) New.
4611 (CONST_WIDE_INT_NUNITS) New.
4612 (CONST_WIDE_INT_ELT) New.
4613 (rtx_mode_t): New type.
4614 (wi::int_traits <rtx_mode_t>): New.
4616 (wi::min_value): New.
4617 (wi::max_value): New.
4619 (const_wide_int_alloc): New.
4620 (immed_wide_int_const): New.
4621 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
4622 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
4623 * signop.h: New file.
4624 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
4625 (simplify_const_unary_operation): Use wide-int interfaces.
4626 (simplify_binary_operation_1): Likewise.
4627 (simplify_const_binary_operation): Likewise.
4628 (simplify_const_relational_operation): Likewise.
4629 (simplify_immed_subreg): Likewise.
4630 * stmt.c (expand_case): Likewise.
4631 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
4632 signop rather than a bool.
4633 * stor-layout.c (layout_type): Use wide-int interfaces.
4634 (initialize_sizetypes): Update calls to
4635 set_min_and_max_values_for_integral_type.
4636 (set_min_and_max_values_for_integral_type): Take a signop rather
4637 than a bool. Use wide-int interfaces.
4638 (fixup_signed_type): Update accordingly. Remove
4639 HOST_BITS_PER_DOUBLE_INT limit.
4640 (fixup_unsigned_type): Likewise.
4641 * system.h (STATIC_CONSTANT_P): New.
4642 (STATIC_ASSERT): New.
4643 * target.def (can_use_doloop_p): Take widest_ints rather than
4645 * target.h: Include wide-int.h rather than double-int.h.
4646 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
4648 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
4649 rather than INT_CST_LT_UNSIGNED.
4650 (can_use_doloop_if_innermost): Take widest_ints rather than
4652 * tree-affine.c: Include wide-int-print.h.
4653 (double_int_ext_for_comb): Delete.
4654 (wide_int_ext_for_comb): New.
4655 (aff_combination_zero): Use wide-int interfaces.
4656 (aff_combination_const): Take a widest_int instead of a double_int.
4657 (aff_combination_elt): Use wide-int interfaces.
4658 (aff_combination_scale): Take a widest_int instead of a double_int.
4659 (aff_combination_add_elt): Likewise.
4660 (aff_combination_add_cst): Likewise.
4661 (aff_combination_add): Use wide-int interfaces.
4662 (aff_combination_convert): Likewise.
4663 (tree_to_aff_combination): Likewise.
4664 (add_elt_to_tree): Take a widest_int instead of a double_int.
4665 (aff_combination_to_tree): Use wide-int interfaces.
4666 (aff_combination_remove_elt): Likewise.
4667 (aff_combination_add_product): Take a widest_int instead of
4669 (aff_combination_mult): Use wide-int interfaces.
4670 (aff_combination_expand): Likewise.
4671 (double_int_constant_multiple_p): Delete.
4672 (wide_int_constant_multiple_p): New.
4673 (aff_combination_constant_multiple_p): Take a widest_int pointer
4674 instead of a double_int pointer.
4675 (print_aff): Use wide-int interfaces.
4676 (get_inner_reference_aff): Take a widest_int pointer
4677 instead of a double_int pointer.
4678 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
4679 * tree-affine.h: Include wide-int.h.
4680 (struct aff_comb_elt): Change type of coef to widest_int.
4681 (struct affine_tree_combination): Change type of offset to widest_int.
4682 (double_int_ext_for_comb): Delete.
4683 (wide_int_ext_for_comb): New.
4684 (aff_combination_const): Use widest_int instead of double_int.
4685 (aff_combination_scale): Likewise.
4686 (aff_combination_add_elt): Likewise.
4687 (aff_combination_constant_multiple_p): Likewise.
4688 (get_inner_reference_aff): Likewise.
4689 (aff_comb_cannot_overlap_p): Likewise.
4690 (aff_combination_zero_p): Use wide-int interfaces.
4691 * tree.c: Include tree.h.
4692 (init_ttree): Use make_int_cst.
4693 (tree_code_size): Removed code for INTEGER_CST case.
4694 (tree_size): Add INTEGER_CST case.
4695 (make_node_stat): Update comment.
4696 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
4697 (build_int_cst_type): Use wide-int interfaces.
4698 (double_int_to_tree): Likewise.
4699 (double_int_fits_to_tree_p): Delete.
4700 (force_fit_type_double): Delete.
4701 (force_fit_type): New.
4702 (int_cst_hash_hash): Use wide-int interfaces.
4703 (int_cst_hash_eq): Likewise.
4704 (build_int_cst_wide): Delete.
4705 (wide_int_to_tree): New.
4706 (cache_integer_cst): Use wide-int interfaces.
4707 (build_low_bits_mask): Likewise.
4708 (cst_and_fits_in_hwi): Likewise.
4709 (real_value_from_int_cst): Likewise.
4710 (make_int_cst_stat): New.
4711 (integer_zerop): Use wide_int interfaces.
4712 (integer_onep): Likewise.
4713 (integer_all_onesp): Likewise.
4714 (integer_pow2p): Likewise.
4715 (integer_nonzerop): Likewise.
4716 (tree_log2): Likewise.
4717 (tree_floor_log2): Likewise.
4718 (tree_ctz): Likewise.
4719 (int_size_in_bytes): Likewise.
4720 (mem_ref_offset): Return an offset_int rather than a double_int.
4721 (build_type_attribute_qual_variant): Use wide_int interfaces.
4722 (type_hash_eq): Likewise
4723 (tree_int_cst_equal): Likewise.
4724 (tree_int_cst_lt): Delete.
4725 (tree_int_cst_compare): Likewise.
4726 (tree_fits_shwi_p): Use wide_int interfaces.
4727 (tree_fits_uhwi_p): Likewise.
4728 (tree_int_cst_sign_bit): Likewise.
4729 (tree_int_cst_sgn): Likewise.
4730 (tree_int_cst_min_precision): Take a signop rather than a bool.
4731 (simple_cst_equal): Use wide_int interfaces.
4732 (compare_tree_int): Likewise.
4733 (iterative_hash_expr): Likewise.
4734 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
4736 (get_type_static_bounds): Use wide_int interfaces.
4737 (tree_int_cst_elt_check_failed): New.
4738 (build_common_tree_nodes): Reordered to set prec before filling in
4740 (int_cst_value): Check cst_and_fits_in_hwi.
4741 (widest_int_cst_value): Use wide_int interfaces.
4742 (upper_bound_in_type): Likewise.
4743 (lower_bound_in_type): Likewise.
4744 (num_ending_zeros): Likewise.
4745 (drop_tree_overflow): Likewise.
4746 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
4747 (gen_conditions_for_pow_cst_base): Likewise.
4748 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
4749 (group_case_labels_stmt): Use wide-int interfaces.
4750 (verify_gimple_assign_binary): Likewise.
4751 (print_loop): Likewise.
4752 * tree-chrec.c (tree_fold_binomial): Likewise.
4753 * tree-core.h (struct tree_base): Add int_length.
4754 (struct tree_int_cst): Change rep of value.
4755 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
4756 (dr_may_alias_p): Likewise.
4757 (max_stmt_executions_tree): Likewise.
4758 * tree.def (INTEGER_CST): Update comment.
4759 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
4760 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
4761 * tree-dump.c: Include wide-int.h and wide-int-print.h.
4762 (dequeue_and_dump): Use wide-int interfaces.
4763 * tree.h: Include wide-int.h.
4764 (NULL_TREE): Moved to earlier loc in file.
4765 (TREE_INT_CST_ELT_CHECK): New.
4766 (tree_int_cst_elt_check_failed): New.
4768 (TREE_INT_CST): Delete.
4769 (TREE_INT_CST_LOW): Use wide-int interfaces.
4770 (TREE_INT_CST_HIGH): Delete.
4771 (TREE_INT_CST_NUNITS): New.
4772 (TREE_INT_CST_EXT_NUNITS): Likewise.
4773 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
4774 (TREE_INT_CST_ELT): Likewise.
4775 (INT_CST_LT): Delete.
4776 (tree_int_cst_elt_check): New (two forms).
4777 (type_code_size): Update comment.
4778 (make_int_cst_stat, make_int_cst): New.
4779 (tree_to_double_int): Delete.
4780 (double_int_fits_to_tree_p): Delete.
4781 (force_fit_type_double): Delete.
4782 (build_int_cstu): Replace with out-of-line function.
4783 (build_int_cst_wide): Delete.
4784 (tree_int_cst_lt): Define inline.
4785 (tree_int_cst_le): New.
4786 (tree_int_cst_compare): Define inline.
4787 (tree_int_cst_min_precision): Take a signop rather than a bool.
4788 (wi::int_traits <const_tree>): New.
4789 (wi::int_traits <tree>): New.
4790 (wi::extended_tree): New.
4791 (wi::int_traits <wi::extended_tree>): New.
4792 (wi::to_widest): New.
4793 (wi::to_offset): New.
4794 (wi::fits_to_tree_p): New.
4795 (wi::min_value): New.
4796 (wi::max_value): New.
4797 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
4798 (copy_tree_body_r): Likewise.
4799 * tree-object-size.c (compute_object_offset): Likewise.
4800 (addr_object_size): Likewise.
4801 * tree-predcom.c: Include wide-int-print.h.
4802 (struct dref_d): Change type of offset to widest_int.
4803 (dump_dref): Call wide-int printer.
4804 (aff_combination_dr_offset): Use wide-int interfaces.
4805 (determine_offset): Take a widest_int pointer rather than a
4807 (split_data_refs_to_components): Use wide-int interfaces.
4808 (suitable_component_p): Likewise.
4809 (order_drefs): Likewise.
4810 (add_ref_to_chain): Likewise.
4811 (valid_initializer_p): Likewise.
4812 (determine_roots_comp): Likewise.
4813 * tree-pretty-print.c: Include wide-int-print.h.
4814 (dump_generic_node): Use wide-int interfaces.
4815 * tree-sra.c (sra_ipa_modify_expr): Likewise.
4816 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
4817 (move_fixed_address_to_symbol): Likewise.
4818 (move_hint_to_base): Likewise.
4819 (move_pointer_to_base): Likewise.
4820 (move_variant_to_index): Likewise.
4821 (most_expensive_mult_to_index): Likewise.
4822 (addr_to_parts): Likewise.
4823 (copy_ref_info): Likewise.
4824 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
4825 (indirect_refs_may_alias_p): Likewise.
4826 (stmt_kills_ref_p_1): Likewise.
4827 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
4828 * tree-ssa-ccp.c: Update comment at top of file. Include
4830 (struct prop_value_d): Change type of mask to widest_int.
4831 (extend_mask): New function.
4832 (dump_lattice_value): Use wide-int interfaces.
4833 (get_default_value): Likewise.
4834 (set_constant_value): Likewise.
4835 (set_value_varying): Likewise.
4836 (valid_lattice_transition): Likewise.
4837 (set_lattice_value): Likewise.
4838 (value_to_double_int): Delete.
4839 (value_to_wide_int): New.
4840 (get_value_from_alignment): Use wide-int interfaces.
4841 (get_value_for_expr): Likewise.
4842 (do_dbg_cnt): Likewise.
4843 (ccp_finalize): Likewise.
4844 (ccp_lattice_meet): Likewise.
4845 (bit_value_unop_1): Use widest_ints rather than double_ints.
4846 (bit_value_binop_1): Likewise.
4847 (bit_value_unop): Use wide-int interfaces.
4848 (bit_value_binop): Likewise.
4849 (bit_value_assume_aligned): Likewise.
4850 (evaluate_stmt): Likewise.
4851 (ccp_fold_stmt): Likewise.
4852 (visit_cond_stmt): Likewise.
4853 (ccp_visit_stmt): Likewise.
4854 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
4855 (constant_pointer_difference): Likewise.
4856 (associate_pointerplus): Likewise.
4857 (combine_conversions): Likewise.
4858 * tree-ssa-loop.h: Include wide-int.h.
4859 (struct tree_niter_desc): Change type of max to widest_int.
4860 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
4861 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
4862 (remove_redundant_iv_tests): Likewise.
4863 (canonicalize_loop_induction_variables): Likewise.
4864 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
4865 (constant_multiple_of): Take a widest_int pointer instead of
4866 a double_int pointer.
4867 (get_computation_aff): Use wide-int interfaces.
4868 (ptr_difference_cost): Likewise.
4869 (difference_cost): Likewise.
4870 (get_loop_invariant_expr_id): Likewise.
4871 (get_computation_cost_at): Likewise.
4872 (iv_elimination_compare_lt): Likewise.
4873 (may_eliminate_iv): Likewise.
4874 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
4875 instead of double_int.
4876 (max_loop_iterations): Likewise.
4877 (max_stmt_executions): Likewise.
4878 (estimated_stmt_executions): Likewise.
4879 * tree-ssa-loop-niter.c: Include wide-int-print.h.
4880 (split_to_var_and_offset): Use wide-int interfaces.
4881 (determine_value_range): Likewise.
4882 (bound_difference_of_offsetted_base): Likewise.
4883 (bounds_add): Take a widest_int instead of a double_int.
4884 (number_of_iterations_ne_max): Use wide-int interfaces.
4885 (number_of_iterations_ne): Likewise.
4886 (number_of_iterations_lt_to_ne): Likewise.
4887 (assert_loop_rolls_lt): Likewise.
4888 (number_of_iterations_lt): Likewise.
4889 (number_of_iterations_le): Likewise.
4890 (number_of_iterations_cond): Likewise.
4891 (number_of_iterations_exit): Likewise.
4892 (finite_loop_p): Likewise.
4893 (derive_constant_upper_bound_assign): Likewise.
4894 (derive_constant_upper_bound): Return a widest_int.
4895 (derive_constant_upper_bound_ops): Likewise.
4896 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
4897 (record_estimate): Take a widest_int rather than a double_int.
4898 (record_nonwrapping_iv): Use wide-int interfaces.
4899 (double_int_cmp): Delete.
4900 (wide_int_cmp): New.
4901 (bound_index): Take a widest_int rather than a double_int.
4902 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
4903 (maybe_lower_iteration_bound): Likewise.
4904 (estimate_numbers_of_iterations_loop): Likewise.
4905 (estimated_loop_iterations): Take a widest_int pointer than than
4906 a double_int pointer.
4907 (estimated_loop_iterations_int): Use wide-int interfaces.
4908 (max_loop_iterations): Take a widest_int pointer than than
4909 a double_int pointer.
4910 (max_loop_iterations_int): Use wide-int interfaces.
4911 (max_stmt_executions): Take a widest_int pointer than than
4912 a double_int pointer.
4913 (estimated_stmt_executions): Likewise.
4914 (n_of_executions_at_most): Use wide-int interfaces.
4915 (scev_probably_wraps_p): Likewise.
4916 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
4918 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
4920 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
4921 double_ints. Adjust for trailing_wide_ints <3> representation.
4922 (set_nonzero_bits): Likewise.
4923 (get_range_info): Return wide_ints rather than double_ints.
4924 Adjust for trailing_wide_ints <3> representation.
4925 (get_nonzero_bits): Likewise.
4926 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
4928 * tree-ssanames.h (struct range_info_def): Replace min, max and
4929 nonzero_bits with a trailing_wide_ints <3>.
4930 (set_range_info): Use wide_int_refs rather than double_ints.
4931 (set_nonzero_bits): Likewise.
4932 (get_range_info): Return wide_ints rather than double_ints.
4933 (get_nonzero_bits): Likewise.
4934 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
4935 * tree-ssa-pre.c (phi_translate_1): Likewise.
4936 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
4937 (acceptable_pow_call): Likewise.
4938 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
4940 (vn_reference_fold_indirect): Likewise.
4941 (vn_reference_maybe_forwprop_address): Likewise.
4942 (valueize_refs_1): Likewise.
4943 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
4944 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
4945 tree_int_cst_lt and tree_int_cst_le.
4946 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
4948 (streamer_alloc_tree): Likewise.
4949 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
4950 (streamer_write_tree_header): Likewise.
4951 (streamer_write_integer_cst): Likewise.
4952 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
4953 (build_constructors): Likewise.
4954 (array_value_type): Likewise.
4955 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
4956 (vect_check_gather): Likewise.
4957 * tree-vect-generic.c (build_replicated_const): Likewise.
4958 (expand_vector_divmod): Likewise.
4959 * tree-vect-loop.c (vect_transform_loop): Likewise.
4960 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
4961 (vect_do_peeling_for_alignment): Likewise.
4962 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
4963 * tree-vrp.c: Include wide-int.h.
4964 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
4965 (extract_range_from_assert): Use wide-int interfaces.
4966 (vrp_int_const_binop): Likewise.
4967 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
4968 double_int pointers.
4969 (ranges_from_anti_range): Use wide-int interfaces.
4970 (quad_int_cmp): Delete.
4971 (quad_int_pair_sort): Likewise.
4972 (extract_range_from_binary_expr_1): Use wide-int interfaces.
4973 (extract_range_from_unary_expr_1): Likewise.
4974 (adjust_range_with_scev): Likewise.
4975 (masked_increment): Take and return wide_ints rather than double_ints.
4976 (register_edge_assert_for_2): Use wide-int interfaces.
4977 (check_array_ref): Likewise.
4978 (search_for_addr_array): Likewise.
4979 (maybe_set_nonzero_bits): Likewise.
4980 (union_ranges): Pass an integer of the correct type instead of
4981 using integer_one_node.
4982 (intersect_ranges): Likewise.
4983 (simplify_truth_ops_using_ranges): Likewise.
4984 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
4985 (range_fits_type_p): Likewise.
4986 (simplify_cond_using_ranges): Likewise. Take a signop rather than
4988 (simplify_conversion_using_ranges): Use wide-int interfaces.
4989 (simplify_float_conversion_using_ranges): Likewise.
4990 (vrp_finalize): Likewise.
4991 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
4992 (gimple_stringops_transform): Likewise.
4993 * varasm.c (decode_addr_const): Likewise.
4994 (const_hash_1): Likewise.
4995 (const_rtx_hash_1): Likewise
4996 (output_constant): Likewise.
4997 (array_size_for_constructor): Likewise.
4998 (output_constructor_regular_field): Likewise.
4999 (output_constructor_bitfield): Likewise.
5000 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
5001 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
5003 * gencheck.c: Define BITS_PER_UNIT.
5006 * wide-int-print.cc: New.
5007 * wide-int-print.h: New.
5009 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5011 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
5013 2014-05-06 Richard Biener <rguenther@suse.de>
5015 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
5016 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
5017 (TODO_verify_all): Adjust.
5018 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
5019 TODO_verify_stmts and TODO_verify_rtl_sharing.
5020 * bb-reorder.c: Likewise.
5021 * cfgexpand.c: Likewise.
5022 * cprop.c: Likewise.
5024 * function.c: Likewise.
5025 * fwprop.c: Likewise.
5027 * gimple-ssa-isolate-paths.c: Likewise.
5028 * gimple-ssa-strength-reduction.c: Likewise.
5029 * ipa-split.c: Likewise.
5030 * loop-init.c: Likewise.
5031 * loop-unroll.c: Likewise.
5032 * lower-subreg.c: Likewise.
5033 * modulo-sched.c: Likewise.
5034 * postreload-gcse.c: Likewise.
5035 * predict.c: Likewise.
5036 * recog.c: Likewise.
5037 * sched-rgn.c: Likewise.
5038 * store-motion.c: Likewise.
5039 * tracer.c: Likewise.
5040 * trans-mem.c: Likewise.
5041 * tree-call-cdce.c: Likewise.
5042 * tree-cfg.c: Likewise.
5043 * tree-cfgcleanup.c: Likewise.
5044 * tree-complex.c: Likewise.
5045 * tree-eh.c: Likewise.
5046 * tree-emutls.c: Likewise.
5047 * tree-if-conv.c: Likewise.
5048 * tree-into-ssa.c: Likewise.
5049 * tree-loop-distribution.c: Likewise.
5050 * tree-object-size.c: Likewise.
5051 * tree-parloops.c: Likewise.
5052 * tree-pass.h: Likewise.
5053 * tree-sra.c: Likewise.
5054 * tree-ssa-ccp.c: Likewise.
5055 * tree-ssa-copy.c: Likewise.
5056 * tree-ssa-copyrename.c: Likewise.
5057 * tree-ssa-dce.c: Likewise.
5058 * tree-ssa-dom.c: Likewise.
5059 * tree-ssa-dse.c: Likewise.
5060 * tree-ssa-forwprop.c: Likewise.
5061 * tree-ssa-ifcombine.c: Likewise.
5062 * tree-ssa-loop-ch.c: Likewise.
5063 * tree-ssa-loop-ivcanon.c: Likewise.
5064 * tree-ssa-loop.c: Likewise.
5065 * tree-ssa-math-opts.c: Likewise.
5066 * tree-ssa-phiopt.c: Likewise.
5067 * tree-ssa-phiprop.c: Likewise.
5068 * tree-ssa-pre.c: Likewise.
5069 * tree-ssa-reassoc.c: Likewise.
5070 * tree-ssa-sink.c: Likewise.
5071 * tree-ssa-strlen.c: Likewise.
5072 * tree-ssa-tail-merge.c: Likewise.
5073 * tree-ssa-uncprop.c: Likewise.
5074 * tree-switch-conversion.c: Likewise.
5075 * tree-tailcall.c: Likewise.
5076 * tree-vect-generic.c: Likewise.
5077 * tree-vectorizer.c: Likewise.
5078 * tree-vrp.c: Likewise.
5080 * var-tracking.c: Likewise.
5081 * bt-load.c: Likewise.
5082 * cfgcleanup.c: Likewise.
5083 * combine-stack-adj.c: Likewise.
5084 * combine.c: Likewise.
5085 * compare-elim.c: Likewise.
5086 * config/epiphany/resolve-sw-modes.c: Likewise.
5087 * config/i386/i386.c: Likewise.
5088 * config/mips/mips.c: Likewise.
5089 * config/s390/s390.c: Likewise.
5090 * config/sh/sh_treg_combine.cc: Likewise.
5091 * config/sparc/sparc.c: Likewise.
5094 * final.c: Likewise.
5095 * ifcvt.c: Likewise.
5096 * mode-switching.c: Likewise.
5097 * passes.c: Likewise.
5098 * postreload.c: Likewise.
5100 * reg-stack.c: Likewise.
5101 * regcprop.c: Likewise.
5102 * regrename.c: Likewise.
5105 2014-05-06 Richard Biener <rguenther@suse.de>
5108 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
5109 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
5111 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
5114 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
5116 2014-05-05 Radovan Obradovic <robradovic@mips.com>
5117 Tom de Vries <tom@codesourcery.com>
5119 * target.def (call_fusage_contains_non_callee_clobbers): New
5121 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
5123 (@node Miscellaneous Register Hooks): New node.
5124 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
5125 * doc/tm.texi: Regenerate.
5127 2014-05-05 Marek Polacek <polacek@redhat.com>
5130 * opts.c (common_handle_option): Call error_at instead of warning_at.
5132 2014-05-05 Richard Biener <rguenther@suse.de>
5134 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
5135 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
5136 under the TODO_verify_il umbrella.
5138 2014-05-05 Richard Biener <rguenther@suse.de>
5140 * passes.c (execute_function_todo): Move TODO_verify_flow under
5141 the TODO_verify_ul umbrella.
5143 2014-05-05 Richard Biener <rguenther@suse.de>
5146 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
5147 X & CST away from a CST that is the mask of a mode.
5149 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5151 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
5152 int argument to enum machine_mode.
5153 (picochip_class_max_nregs): Ditto.
5154 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
5155 (picochip_class_max_nregs): Ditto.
5157 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5159 * target.def: Add new target hook.
5160 * doc/tm.texi: Regenerate.
5161 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
5162 * targhooks.c (default_keep_leaf_when_profiled): New function.
5164 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
5165 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
5167 2014-05-05 Bin Cheng <bin.cheng@arm.com>
5169 PR tree-optimization/60363
5170 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
5171 (copy_phi_args): New parameters. Call get_value_locus_in_path.
5172 (update_destination_phis): New parameter.
5173 (create_edge_and_update_destination_phis): Ditto.
5174 (ssa_fix_duplicate_block_edges): Pass new arguments.
5175 (thread_single_edge): Ditto.
5177 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
5179 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
5180 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
5181 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
5182 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
5183 Use RS6000_BTM_HARD_FLOAT.
5184 (BU_MISC_2): Likewise.
5185 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
5186 RS6000_BTM_HARD_FLOAT.
5187 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
5189 (rs6000_invalid_builtin): Add hard floating builtin support.
5190 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
5191 hard float builtins.
5192 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
5194 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
5196 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
5197 Add missing function* argument.
5199 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
5201 * lra-constraints.c (valid_address_p): Move earlier in file.
5202 Add a constraint argument to the address_info version.
5203 (satisfies_memory_constraint_p): New function.
5204 (satisfies_address_constraint_p): Likewise.
5205 (process_alt_operands, curr_insn_transform): Use them.
5206 (process_address): Pass the constraint to valid_address_p when
5207 checking address operands.
5209 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
5211 * config/mips/mips.c (mips_isa_rev): New variable.
5212 (mips_set_architecture): Set it.
5213 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
5215 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
5216 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
5217 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
5218 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
5219 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
5220 conditions in terms of mips_isa_rev.
5221 (mips_isa_rev): Declare.
5223 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
5225 * config/sh/sh-mem.cc: Use tabs instead of spaces.
5226 (prob_unlikely, prob_likely): Make variables const.
5228 2014-05-03 Denis Chertykov <chertykov@gmail.com>
5230 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
5232 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
5234 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
5236 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
5238 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
5239 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
5240 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
5242 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
5243 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
5245 Replace usage of ROUND_REG with sh_round_reg.
5246 Use CEIL instead of ROUND_ADVANCE.
5248 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
5251 * config/sh/sh.c: Include stdlib headers before everything else.
5253 2014-05-02 Jakub Jelinek <jakub@redhat.com>
5255 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
5256 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
5257 (gimplify_adjust_omp_clauses): Simd region is never
5258 directly nested in combined parallel. Instead, for linear
5259 with copyin/copyout, if in combined for simd loop, make decl
5260 firstprivate/lastprivate on OMP_FOR.
5261 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
5262 expand_omp_for_static_chunk): When setting endvar, also set
5263 fd->loop.v to the same value.
5265 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5267 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
5269 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
5271 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
5274 2014-05-02 Marek Polacek <polacek@redhat.com>
5276 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
5278 2014-05-02 Kito Cheng <kito@0xlab.org>
5280 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
5281 to a C expression marco.
5282 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
5283 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
5284 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
5285 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
5286 HONOR_REG_ALLOC_ORDER.
5287 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
5289 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5291 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
5293 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5295 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
5297 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
5299 * tree-if-conv.c (is_cond_scalar_reduction): New function.
5300 (convert_scalar_cond_reduction): Likewise.
5301 (predicate_scalar_phi): Add recognition and transformation
5302 of simple conditioanl reduction to be vectorizable.
5304 2014-05-01 Marek Polacek <polacek@redhat.com>
5307 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
5309 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
5311 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
5312 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
5313 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
5314 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
5315 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
5316 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
5317 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
5318 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
5320 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
5322 * config/arc/arc.opt (mlra): Move comment above option name
5323 to avoid mis-parsing as language options.
5325 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5327 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
5328 * config/sol2.h: ... here.
5329 * config/sol2-10.h: Remove.
5331 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
5332 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
5333 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
5334 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
5335 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
5336 * config/sol2.h: ... here.
5337 (SECTION_NAME_FORMAT): Don't redefine.
5338 (STARTFILE_ARCH32_SPEC): Rename to ...
5339 (STARTFILE_ARCH_SPEC): ... this.
5340 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
5341 * config/sparc/sol2.h: ... here.
5342 (SECTION_NAME_FORMAT): Don't undef.
5343 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
5344 (SUBTARGET_EXTRA_SPECS): Remove.
5345 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
5347 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
5348 (MD_STARTFILE_PREFIX): Remove.
5349 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
5350 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
5351 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
5352 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
5353 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
5354 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
5355 * config/i386/sol2.h: ... here.
5356 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
5357 * config/i386/sol2-bi.h: Remove.
5358 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
5359 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
5361 * config/i386/t-sol2-64: Rename to ...
5362 * config/i386/t-sol2: ... this.
5363 * config/sparc/t-sol2-64: Rename to ...
5364 * config/sparc/t-sol2: ... this.
5366 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
5367 sol2_tm_file_head, sol2_tm_file_tail.
5368 Include ${cpu_type}/sol2.h before sol2.h.
5370 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
5371 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
5372 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
5373 Reflect i386/t-sol2-64 renaming.
5374 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
5375 Reflect sparc/t-sol2-64 renaming.
5377 2014-04-30 Richard Biener <rguenther@suse.de>
5379 * passes.c (execute_function_todo): Move TODO_verify_stmts
5380 and TODO_verify_ssa under the TODO_verify_il umbrella.
5381 * tree-ssa.h (verify_ssa): Adjust prototype.
5382 * tree-ssa.c (verify_ssa): Add parameter to tell whether
5383 we should verify SSA operands.
5384 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
5385 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
5386 whether we should verify whether not throwing stmts have EH info.
5387 * graphite-scop-detection.c (create_sese_edges): Adjust.
5388 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
5389 * tree-eh.c (lower_try_finally_switch): Do not add the
5390 default case label twice.
5392 2014-04-30 Marek Polacek <polacek@redhat.com>
5394 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
5395 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
5396 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
5397 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
5399 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
5401 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
5402 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
5403 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
5404 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
5405 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
5406 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
5407 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
5408 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
5410 2014-04-29 David Malcolm <dmalcolm@redhat.com>
5412 * tree-cfg.c (dump_function_to_file): Dump the return type of
5413 functions, in a line to itself before the function body, mimicking
5414 the layout of a C function.
5416 2014-04-29 Jakub Jelinek <jakub@redhat.com>
5418 PR tree-optimization/60971
5419 * tree-tailcall.c (process_assignment): Reject conversions which
5422 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
5424 * calls.c (initialize_argument_information): Always treat
5425 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
5426 (expand_call): Likewise.
5427 (emit_library_call_calue_1): Likewise.
5428 * expr.c (PUSH_ARGS_REVERSED): Do not define.
5429 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
5432 2014-04-29 Nick Clifton <nickc@redhat.com>
5434 * config/msp430/msp430.md (umulsidi): Fix typo.
5435 (mulhisi3): Enable even inside interrupt handlers.
5436 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
5437 bigger return address pushed in large mode.
5439 2014-04-29 Nick Clifton <nickc@redhat.com>
5441 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
5442 (arc_init_reg_tables): Use a machine_mode enum to iterate over
5444 * config/m32r/m32r.c (init_reg_tables): Likewise.
5445 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
5446 enum to hold the modes.
5448 2014-04-29 Richard Biener <rguenther@suse.de>
5450 * dominance.c (free_dominance_info): Add overload with
5452 (dom_info_state): Likewise.
5453 (dom_info_available_p): Likewise.
5454 * basic-block.h (free_dominance_info, dom_info_state,
5455 dom_info_available_p): Declare overloads.
5456 * passes.c (execute_function_todo): Verify that verifiers
5457 don't change dominator info state. Drop dominator info
5458 for IPA pass invocations.
5459 * cgraph.c (release_function_body): Restore asserts that
5460 dominator information is released.
5462 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
5464 * doc/invoke.texi: Fix typo.
5465 * tree-vrp.c: Fix typos.
5466 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
5468 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5470 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
5472 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
5474 * config/aarch64/aarch64-builtins.c
5475 (aarch64_types_storestruct_lane_qualifiers): New.
5476 (TYPES_STORESTRUCT_LANE): Likewise.
5477 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
5478 (st3_lane): Likewise.
5479 (st4_lane): Likewise.
5480 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
5481 (vec_store_lanesci_lane<mode>): Likewise.
5482 (vec_store_lanesxi_lane<mode>): Likewise.
5483 (aarch64_st2_lane<VQ:mode>): Likewise.
5484 (aarch64_st3_lane<VQ:mode>): Likewise.
5485 (aarch64_st4_lane<VQ:mode>): Likewise.
5486 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
5487 * config/aarch64/arm_neon.h
5488 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
5489 use new macro arguments.
5490 (__ST3_LANE_FUNC): Likewise.
5491 (__ST4_LANE_FUNC): Likewise.
5492 * config/aarch64/iterators.md (V_TWO_ELEM): New.
5493 (V_THREE_ELEM): Likewise.
5494 (V_FOUR_ELEM): Likewise.
5496 2014-04-28 David Malcolm <dmalcolm@redhat.com>
5498 * doc/gimple.texi: Replace the description of the now-defunct
5499 union gimple_statement_d with a diagram showing the
5500 gimple_statement_base class hierarchy and its relationships to
5501 the GSS_ and GIMPLE_ enums.
5503 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
5505 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
5506 * config/aarch64/aarch64.c
5507 (aarch64_cannot_change_mode_class): Weaken conditions.
5508 (aarch64_modes_tieable_p): New.
5509 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
5511 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
5513 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
5514 (loadsync_<mode>): Change mode.
5515 (load_quadpti, store_quadpti): New.
5516 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
5517 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
5519 2014-04-28 Martin Jambor <mjambor@suse.cz>
5521 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
5522 same alias type as the original statement.
5523 (subreplacement_assignment_data): New type.
5524 (handle_unscalarized_data_in_subtree): New type of parameter,
5525 generate new memory accesses with same alias type as the original
5527 (load_assign_lhs_subreplacements): Likewise.
5528 (sra_modify_constructor_assign): Generate new memory accesses with
5529 same alias type as the original statement.
5531 2014-04-28 Richard Biener <rguenther@suse.de>
5533 * tree-pass.h (TODO_verify_il): Define.
5534 (TODO_verify_all): Complete properly.
5535 * passes.c (execute_function_todo): Move existing loop-closed
5536 SSA verification under TODO_verify_il.
5537 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
5538 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
5539 Fix tree sharing issue.
5541 2014-04-28 Richard Biener <rguenther@suse.de>
5544 * builtins.def (DEF_C11_BUILTIN): Add.
5545 (BUILT_IN_ALIGNED_ALLOC): Likewise.
5546 * coretypes.h (enum function_class): Add function_c11_misc.
5547 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
5548 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
5549 (call_may_clobber_ref_p_1): Likewise.
5550 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
5551 (mark_all_reaching_defs_necessary_1): Likewise.
5552 (propagate_necessity): Likewise.
5553 (eliminate_unnecessary_stmts): Likewise.
5554 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
5556 2014-04-28 Richard Biener <rguenther@suse.de>
5558 * tree-vrp.c (vrp_var_may_overflow): Remove.
5559 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
5560 with overflow immediately bump to one before that value and
5561 let iteration figure out overflow status.
5563 2014-04-28 Richard Biener <rguenther@suse.de>
5565 * configure.ac: Do valgrind header checks unconditionally.
5566 Add --enable-valgrind-annotations.
5567 * system.h: Guard valgrind header inclusion with
5568 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
5569 * alloc-pool.c (pool_alloc, pool_free): Use
5570 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
5571 to guard possibly dead code.
5572 * config.in: Regenerated.
5573 * configure: Likewise.
5575 2014-04-28 Jeff Law <law@redhat.com>
5577 PR tree-optimization/60902
5578 * tree-ssa-threadedge.c
5579 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
5580 over real defs when invalidating outputs from statements that do not
5581 produce useful outputs for threading.
5583 2014-04-28 Richard Biener <rguenther@suse.de>
5585 PR tree-optimization/60979
5586 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
5587 SCOPs that end in a block with a successor with abnormal
5590 2014-04-28 Richard Biener <rguenther@suse.de>
5592 * tree-pass.h (execute_pass_list): Adjust prototype.
5593 * passes.c (pass_manager::execute_early_local_passes): Adjust.
5594 (do_per_function): Change callback signature, push all actual
5595 work to the callbals.
5596 (do_per_function_toporder): Likewise.
5597 (execute_function_dump): Adjust.
5598 (execute_function_todo): Likewise.
5599 (clear_last_verified): Likewise.
5600 (verify_curr_properties): Likewise.
5601 (update_properties_after_pass): Likewise.
5602 (execute_pass_list_1): Split out from ...
5603 (execute_pass_list): ... here. Adjust.
5604 (execute_ipa_pass_list): Likewise.
5605 * cgraphunit.c (cgraph_add_new_function): Adjust.
5606 (analyze_function): Likewise.
5607 (expand_function): Likewise.
5608 * cgraph.c (release_function_body): Free dominance info
5609 here instead of asserting it was magically freed elsewhere.
5611 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
5613 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
5614 * configure: Regenerate.
5615 * config/sparc/sparc.opt (muser-mode): New option.
5616 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
5618 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
5619 * doc/invoke.texi (SPARC options): Document -muser-mode.
5621 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
5623 * cselib.c (find_slot_memmode): Delete.
5624 (cselib_hasher): Change compare_type to a struct.
5625 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
5627 (preserve_constants_and_equivs): Adjust for new compare_type.
5628 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
5629 (wrap_constant): Delete.
5630 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
5632 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
5634 * doc/install.texi (Building with profile feedback): Remove
5637 2014-04-26 Tom de Vries <tom@codesourcery.com>
5639 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
5642 2014-04-25 Cary Coutant <ccoutant@google.com>
5645 * dwarf2out.c (should_move_die_to_comdat): A type definition
5646 can contain a subprogram definition, but don't move it to a
5648 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
5649 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
5651 (clone_tree_hash): Rename to...
5652 (clone_tree_partial): ...this; change callers. Copy
5653 DW_TAG_subprogram DIEs as declarations.
5654 (copy_decls_walk): Don't copy children of a declaration into a
5657 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
5660 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
5663 2014-04-25 Jiong Wang <jiong.wang@arm.com>
5665 * config/arm/predicates.md (call_insn_operand): Add long_call check.
5666 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
5668 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
5671 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5673 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
5675 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5677 PR tree-optimization/60930
5678 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
5679 creating a multiply candidate by folding two constant
5680 multiplicands when the result overflows.
5682 2014-04-25 Jakub Jelinek <jakub@redhat.com>
5684 PR tree-optimization/60960
5685 * tree-vect-generic.c (expand_vector_operation): Only call
5686 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
5688 2014-04-25 Tom de Vries <tom@codesourcery.com>
5690 * expr.c (clobber_reg_mode): New function.
5691 * expr.h (clobber_reg): New function.
5693 2014-04-25 Tom de Vries <tom@codesourcery.com>
5695 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
5698 2014-04-25 Radovan Obradovic <robradovic@mips.com>
5699 Tom de Vries <tom@codesourcery.com>
5701 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
5703 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
5704 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
5705 new argument to find_all_hard_reg_sets call.
5707 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5709 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
5710 Use HOST_WIDE_INT_C for mask literal.
5711 (aarch_rev16_shleft_mask_imm_p): Likewise.
5713 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
5716 * config/sparc/sparc.md (ashlsi3_extend): Delete.
5718 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
5720 PR preprocessor/56540
5721 * config/i386/i386-c.c (ix86_target_macros): Define
5722 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
5724 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5726 * configure.ac (tga_func): Remove.
5727 (LIB_TLS_SPEC): Remove.
5728 * configure: Regenerate.
5729 * config.in: Regenerate.
5730 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
5732 2014-04-25 Richard Biener <rguenther@suse.de>
5735 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
5736 call stmt use/clobber sets during stmt walk instead of
5737 walking the possibly incomplete set of caller edges.
5739 2014-04-25 Richard Biener <rguenther@suse.de>
5742 * passes.c (apply_ipa_transforms): Inline into only caller ...
5743 (execute_one_pass): ... here. Properly bring in function
5744 bodies for nodes we want to apply IPA transforms to.
5746 2014-04-24 Cong Hou <congh@google.com>
5748 PR tree-optimization/60896
5749 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
5750 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
5751 (vect_mark_pattern_stmts): Set the def type of all statements in
5752 PATTERN_DEF_SEQ as vect_internal_def.
5754 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
5756 * doc/extend.texi (PowerPC Built-in Functions): Document new
5757 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
5758 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
5760 * config/rs6000/predicates.md (const_0_to_3_operand): New
5761 predicate to match 0..3 integer constants.
5763 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
5764 to support adding miscellaneous builtin functions.
5765 (BU_DFP_MISC_2): Likewise.
5766 (BU_P7_MISC_1): Likewise.
5767 (BU_P7_MISC_2): Likewise.
5768 (BU_P8V_MISC_3): Likewise.
5769 (BU_MISC_1): Likewise.
5770 (BU_MISC_2): Likewise.
5771 (DIVWE): Add extended divide builtin functions.
5774 (DIVWEUO): Likewise.
5778 (DIVDEUO): Likewise.
5779 (DXEX): Add decimal floating-point builtin functions.
5782 (DDEDPDQ): Likewise.
5784 (DENBCDQ): Likewise.
5791 (CDTBCD): Add new BCD builtin functions.
5795 (BCDADD_LT): Likewise.
5796 (BCDADD_EQ): Likewise.
5797 (BCDADD_GT): Likewise.
5798 (BCDADD_OV): Likewise.
5800 (BCDSUB_LT): Likewise.
5801 (BCDSUB_EQ): Likewise.
5802 (BCDSUB_GT): Likewise.
5803 (BCDSUB_OV): Likewise.
5804 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
5805 (UNPACK_TD): Likewise.
5806 (PACK_TF): Likewise.
5807 (UNPACK_TF): Likewise.
5808 (UNPACK_TF_0): Likewise.
5809 (UNPACK_TF_1): Likewise.
5810 (PACK_V1TI): Likewise.
5811 (UNPACK_V1TI): Likewise.
5813 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
5814 support for decimal floating point builtin functions.
5815 (rs6000_expand_ternop_builtin): Add checks for the new builtin
5816 functions that take constant arguments.
5817 (rs6000_invalid_builtin): Add decimal floating point builtin support.
5818 (rs6000_init_builtins): Setup long double, _Decimal64, and
5819 _Decimal128 types for new builtin functions.
5820 (builtin_function_type): Set the unsigned flags appropriately for
5821 the new builtin functions.
5822 (rs6000_opt_masks): Add support for decimal floating point builtin
5825 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
5826 floating point builtin functions.
5827 (RS6000_BTM_COMMON): Likewise.
5828 (RS6000_BTI_long_double): Likewise.
5829 (RS6000_BTI_dfloat64): Likewise.
5830 (RS6000_BTI_dfloat128): Likewise.
5831 (long_double_type_internal_node): Likewise.
5832 (dfloat64_type_internal_node): Likewise.
5833 (dfloat128_type_internal_node): Likewise.
5835 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
5836 2.07 bcd arithmetic instructions.
5837 (UNSPEC_BCDSUB): Likewise.
5838 (UNSPEC_BCD_OVERFLOW): Likewise.
5839 (UNSPEC_BCD_ADD_SUB): Likewise.
5840 (bcd_add_sub): Likewise.
5841 (BCD_TEST): Likewise.
5842 (bcd<bcd_add_sub>): Likewise.
5843 (bcd<bcd_add_sub>_test): Likewise.
5844 (bcd<bcd_add_sub>_test2): Likewise.
5845 (bcd<bcd_add_sub>_<code>): Likewise.
5846 (peephole2 for combined bcd ops): Likewise.
5848 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
5849 decimal floating point builtin functions.
5850 (UNSPEC_DENBCD): Likewise.
5851 (UNSPEC_DXEX): Likewise.
5852 (UNSPEC_DIEX): Likewise.
5853 (UNSPEC_DSCLI): Likewise.
5854 (UNSPEC_DSCRI): Likewise.
5855 (D64_D128): Likewise.
5856 (dfp_suffix): Likewise.
5857 (dfp_ddedpd_<mode>): Likewise.
5858 (dfp_denbcd_<mode>): Likewise.
5859 (dfp_dxex_<mode>): Likewise.
5860 (dfp_diex_<mode>): Likewise.
5861 (dfp_dscli_<mode>): Likewise.
5862 (dfp_dscri_<mode>): Likewise.
5864 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
5866 (UNSPEC_CDTBCD): Likewise.
5867 (UNSPEC_CBCDTD): Likewise.
5868 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
5869 (UNSPEC_DIVEO): Likewise.
5870 (UNSPEC_DIVEU): Likewise.
5871 (UNSPEC_DIVEUO): Likewise.
5872 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
5873 pack/unpack 128-bit types.
5874 (UNSPEC_PACK_128BIT): Likewise.
5875 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
5876 (udiv<mode>3): Use idiv_ldiv mode attribute.
5877 (div<mode>3): Likewise.
5878 (addg6s): Add new BCD builtin functions.
5881 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
5882 (div_extend): Likewise.
5883 (div<div_extend>_<mode>"): Likewise.
5884 (FP128_64): Add support for new builtin functions to pack/unpack
5886 (unpack<mode>): Likewise.
5887 (unpacktf_0): Likewise.
5888 (unpacktf_1): Likewise.
5889 (unpack<mode>_dm): Likewise.
5890 (unpack<mode>_nodm): Likewise.
5891 (pack<mode>): Likewise.
5892 (unpackv1ti): Likewise.
5893 (packv1ti): Likewise.
5895 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
5897 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
5900 2014-04-24 Jakub Jelinek <jakub@redhat.com>
5902 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
5903 * gimplify.c (omp_is_private): Change last argument's type to int.
5904 Only diagnose lastprivate if the simd argument is 1, only diagnose
5905 linear if the simd argument is 2.
5906 (gimplify_omp_for): Adjust omp_is_private callers. When adding
5907 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
5908 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
5909 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
5910 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
5911 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
5912 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
5913 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
5914 * tree-nested.c (convert_nonlocal_omp_clauses,
5915 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
5917 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
5920 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
5923 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
5925 * flag-types.h (enum ivar_visibility): Add.
5927 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
5929 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
5930 function * argument.
5932 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
5934 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
5936 2014-04-24 Radovan Obradovic <robradovic@mips.com>
5937 Tom de Vries <tom@codesourcery.com>
5939 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
5940 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
5942 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
5943 * emit-rtl.c (try_split): Same.
5945 2014-04-24 Radovan Obradovic <robradovic@mips.com>
5946 Tom de Vries <tom@codesourcery.com>
5948 * common.opt (fuse-caller-save): New option.
5950 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
5952 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
5953 elements for big-endian.
5955 2014-04-24 Richard Biener <rguenther@suse.de>
5957 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
5958 during TER and instead use the sepops interface for expanding
5959 non-GIMPLE_SINGLE_RHS.
5961 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5963 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
5964 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
5966 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5968 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
5969 assembler 64-bit option.
5970 * configure: Regenerate.
5972 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5974 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
5975 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
5976 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
5977 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
5978 (TARGET_CRYPTO): Take TARGET_SIMD into account.
5980 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5982 * config/aarch64/aarch64-builtins.c
5983 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
5984 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
5985 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
5986 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
5988 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
5989 (Vrevsuff): New mode attribute.
5991 2014-04-24 Terry Guo <terry.guo@arm.com>
5993 * config/arm/arm.h (machine_function): Define variable
5994 after_arm_reorg here.
5995 * config/arm/arm.c (after_arm_reorg): Remove the definition.
5996 (arm_split_constant): Update the way to access variable
5999 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
6001 2014-04-23 Tom de Vries <tom@codesourcery.com>
6003 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
6005 2014-04-23 David Malcolm <dmalcolm@redhat.com>
6007 * is-a.h: Update comments to reflect the following changes to the
6008 "pointerness" of the API, making the template parameter match the
6009 return type, allowing use of is-a.h with typedefs of pointers.
6010 (is_a_helper::cast): Return a T rather then a pointer to a T, so
6011 that the return type matches the parameter to the is_a_helper.
6013 (dyn_cast): Likewise.
6015 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
6016 pointer from the is-a.h API.
6018 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
6019 (is_a_helper <cgraph_node *>::test): ...this, matching change to
6021 (is_a_helper <varpool_node>::test): Likewise, convert to...
6022 (is_a_helper <varpool_node *>::test): ...this.
6024 (varpool_first_variable): Update for removal of implicit pointer
6025 from the is-a.h API.
6026 (varpool_next_variable): Likewise.
6027 (varpool_first_static_initializer): Likewise.
6028 (varpool_next_static_initializer): Likewise.
6029 (varpool_first_defined_variable): Likewise.
6030 (varpool_next_defined_variable): Likewise.
6031 (cgraph_first_defined_function): Likewise.
6032 (cgraph_next_defined_function): Likewise.
6033 (cgraph_first_function): Likewise.
6034 (cgraph_next_function): Likewise.
6035 (cgraph_first_function_with_gimple_body): Likewise.
6036 (cgraph_next_function_with_gimple_body): Likewise.
6037 (cgraph_alias_target): Likewise.
6038 (varpool_alias_target): Likewise.
6039 (cgraph_function_or_thunk_node): Likewise.
6040 (varpool_variable_node): Likewise.
6041 (symtab_real_symbol_p): Likewise.
6042 * cgraphunit.c (referred_to_p): Likewise.
6043 (analyze_functions): Likewise.
6044 (handle_alias_pairs): Likewise.
6045 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
6046 * gimple-ssa.h (gimple_vuse_op): Likewise.
6047 (gimple_vdef_op): Likewise.
6048 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
6049 * gimple.c (gimple_build_asm_1): Likewise.
6050 (gimple_build_try): Likewise.
6051 (gimple_build_resx): Likewise.
6052 (gimple_build_eh_dispatch): Likewise.
6053 (gimple_build_omp_for): Likewise.
6054 (gimple_omp_for_set_clauses): Likewise.
6056 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
6057 (is_a_helper <gimple_statement_asm *>::test): ...this.
6058 (is_a_helper <gimple_statement_bind>::test): Convert to...
6059 (is_a_helper <gimple_statement_bind *>::test): ...this.
6060 (is_a_helper <gimple_statement_call>::test): Convert to...
6061 (is_a_helper <gimple_statement_call *>::test): ...this.
6062 (is_a_helper <gimple_statement_catch>::test): Convert to...
6063 (is_a_helper <gimple_statement_catch *>::test): ...this.
6064 (is_a_helper <gimple_statement_resx>::test): Convert to...
6065 (is_a_helper <gimple_statement_resx *>::test): ...this.
6066 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
6067 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
6068 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
6069 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
6070 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
6071 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
6072 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
6073 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
6074 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
6075 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
6076 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
6077 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
6078 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
6079 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
6080 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
6081 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
6082 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
6083 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
6084 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
6085 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
6086 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
6087 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
6088 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
6089 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
6090 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
6091 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
6092 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
6093 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
6094 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
6095 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
6096 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
6097 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
6098 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
6099 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
6100 (is_a_helper <gimple_statement_phi>::test): Convert to...
6101 (is_a_helper <gimple_statement_phi *>::test): ...this.
6102 (is_a_helper <gimple_statement_transaction>::test): Convert to...
6103 (is_a_helper <gimple_statement_transaction *>::test): ...this.
6104 (is_a_helper <gimple_statement_try>::test): Convert to...
6105 (is_a_helper <gimple_statement_try *>::test): ...this.
6106 (is_a_helper <gimple_statement_wce>::test): Convert to...
6107 (is_a_helper <gimple_statement_wce *>::test): ...this.
6108 (is_a_helper <const gimple_statement_asm>::test): Convert to...
6109 (is_a_helper <const gimple_statement_asm *>::test): ...this.
6110 (is_a_helper <const gimple_statement_bind>::test): Convert to...
6111 (is_a_helper <const gimple_statement_bind *>::test): ...this.
6112 (is_a_helper <const gimple_statement_call>::test): Convert to...
6113 (is_a_helper <const gimple_statement_call *>::test): ...this.
6114 (is_a_helper <const gimple_statement_catch>::test): Convert to...
6115 (is_a_helper <const gimple_statement_catch *>::test): ...this.
6116 (is_a_helper <const gimple_statement_resx>::test): Convert to...
6117 (is_a_helper <const gimple_statement_resx *>::test): ...this.
6118 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
6119 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
6120 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
6121 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
6122 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
6124 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
6126 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
6128 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
6130 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
6131 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
6132 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
6134 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
6135 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
6137 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
6138 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
6139 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
6140 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
6141 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
6142 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
6144 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
6145 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
6146 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
6147 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
6149 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
6150 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
6151 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
6152 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
6153 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
6154 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
6155 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
6156 (is_a_helper <const gimple_statement_phi>::test): Convert to...
6157 (is_a_helper <const gimple_statement_phi *>::test): ...this.
6158 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
6159 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
6160 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
6161 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
6162 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
6163 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
6164 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
6166 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
6168 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
6169 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
6171 (gimple_use_ops): Update for removal of implicit pointer from the
6173 (gimple_set_use_ops): Likewise.
6174 (gimple_vuse): Likewise.
6175 (gimple_vdef): Likewise.
6176 (gimple_vuse_ptr): Likewise.
6177 (gimple_vdef_ptr): Likewise.
6178 (gimple_set_vuse): Likewise.
6179 (gimple_set_vdef): Likewise.
6180 (gimple_omp_return_set_lhs): Likewise.
6181 (gimple_omp_return_lhs): Likewise.
6182 (gimple_omp_return_lhs_ptr): Likewise.
6183 (gimple_call_fntype): Likewise.
6184 (gimple_call_set_fntype): Likewise.
6185 (gimple_call_set_internal_fn): Likewise.
6186 (gimple_call_use_set): Likewise.
6187 (gimple_call_clobber_set): Likewise.
6188 (gimple_bind_vars): Likewise.
6189 (gimple_bind_set_vars): Likewise.
6190 (gimple_bind_body_ptr): Likewise.
6191 (gimple_bind_set_body): Likewise.
6192 (gimple_bind_add_stmt): Likewise.
6193 (gimple_bind_block): Likewise.
6194 (gimple_bind_set_block): Likewise.
6195 (gimple_asm_ninputs): Likewise.
6196 (gimple_asm_noutputs): Likewise.
6197 (gimple_asm_nclobbers): Likewise.
6198 (gimple_asm_nlabels): Likewise.
6199 (gimple_asm_input_op): Likewise.
6200 (gimple_asm_input_op_ptr): Likewise.
6201 (gimple_asm_output_op): Likewise.
6202 (gimple_asm_output_op_ptr): Likewise.
6203 (gimple_asm_set_output_op): Likewise.
6204 (gimple_asm_clobber_op): Likewise.
6205 (gimple_asm_set_clobber_op): Likewise.
6206 (gimple_asm_label_op): Likewise.
6207 (gimple_asm_set_label_op): Likewise.
6208 (gimple_asm_string): Likewise.
6209 (gimple_catch_types): Likewise.
6210 (gimple_catch_types_ptr): Likewise.
6211 (gimple_catch_handler_ptr): Likewise.
6212 (gimple_catch_set_types): Likewise.
6213 (gimple_catch_set_handler): Likewise.
6214 (gimple_eh_filter_types): Likewise.
6215 (gimple_eh_filter_types_ptr): Likewise.
6216 (gimple_eh_filter_failure_ptr): Likewise.
6217 (gimple_eh_filter_set_types): Likewise.
6218 (gimple_eh_filter_set_failure): Likewise.
6219 (gimple_eh_must_not_throw_fndecl): Likewise.
6220 (gimple_eh_must_not_throw_set_fndecl): Likewise.
6221 (gimple_eh_else_n_body_ptr): Likewise.
6222 (gimple_eh_else_e_body_ptr): Likewise.
6223 (gimple_eh_else_set_n_body): Likewise.
6224 (gimple_eh_else_set_e_body): Likewise.
6225 (gimple_try_eval_ptr): Likewise.
6226 (gimple_try_cleanup_ptr): Likewise.
6227 (gimple_try_set_eval): Likewise.
6228 (gimple_try_set_cleanup): Likewise.
6229 (gimple_wce_cleanup_ptr): Likewise.
6230 (gimple_wce_set_cleanup): Likewise.
6231 (gimple_phi_capacity): Likewise.
6232 (gimple_phi_num_args): Likewise.
6233 (gimple_phi_result): Likewise.
6234 (gimple_phi_result_ptr): Likewise.
6235 (gimple_phi_set_result): Likewise.
6236 (gimple_phi_arg): Likewise.
6237 (gimple_phi_set_arg): Likewise.
6238 (gimple_resx_region): Likewise.
6239 (gimple_resx_set_region): Likewise.
6240 (gimple_eh_dispatch_region): Likewise.
6241 (gimple_eh_dispatch_set_region): Likewise.
6242 (gimple_omp_critical_name): Likewise.
6243 (gimple_omp_critical_name_ptr): Likewise.
6244 (gimple_omp_critical_set_name): Likewise.
6245 (gimple_omp_for_clauses): Likewise.
6246 (gimple_omp_for_clauses_ptr): Likewise.
6247 (gimple_omp_for_set_clauses): Likewise.
6248 (gimple_omp_for_collapse): Likewise.
6249 (gimple_omp_for_index): Likewise.
6250 (gimple_omp_for_index_ptr): Likewise.
6251 (gimple_omp_for_set_index): Likewise.
6252 (gimple_omp_for_initial): Likewise.
6253 (gimple_omp_for_initial_ptr): Likewise.
6254 (gimple_omp_for_set_initial): Likewise.
6255 (gimple_omp_for_final): Likewise.
6256 (gimple_omp_for_final_ptr): Likewise.
6257 (gimple_omp_for_set_final): Likewise.
6258 (gimple_omp_for_incr): Likewise.
6259 (gimple_omp_for_incr_ptr): Likewise.
6260 (gimple_omp_for_set_incr): Likewise.
6261 (gimple_omp_for_pre_body_ptr): Likewise.
6262 (gimple_omp_for_set_pre_body): Likewise.
6263 (gimple_omp_parallel_clauses): Likewise.
6264 (gimple_omp_parallel_clauses_ptr): Likewise.
6265 (gimple_omp_parallel_set_clauses): Likewise.
6266 (gimple_omp_parallel_child_fn): Likewise.
6267 (gimple_omp_parallel_child_fn_ptr): Likewise.
6268 (gimple_omp_parallel_set_child_fn): Likewise.
6269 (gimple_omp_parallel_data_arg): Likewise.
6270 (gimple_omp_parallel_data_arg_ptr): Likewise.
6271 (gimple_omp_parallel_set_data_arg): Likewise.
6272 (gimple_omp_task_clauses): Likewise.
6273 (gimple_omp_task_clauses_ptr): Likewise.
6274 (gimple_omp_task_set_clauses): Likewise.
6275 (gimple_omp_task_child_fn): Likewise.
6276 (gimple_omp_task_child_fn_ptr): Likewise.
6277 (gimple_omp_task_set_child_fn): Likewise.
6278 (gimple_omp_task_data_arg): Likewise.
6279 (gimple_omp_task_data_arg_ptr): Likewise.
6280 (gimple_omp_task_set_data_arg): Likewise.
6281 (gimple_omp_taskreg_clauses): Likewise.
6282 (gimple_omp_taskreg_clauses_ptr): Likewise.
6283 (gimple_omp_taskreg_set_clauses): Likewise.
6284 (gimple_omp_taskreg_child_fn): Likewise.
6285 (gimple_omp_taskreg_child_fn_ptr): Likewise.
6286 (gimple_omp_taskreg_set_child_fn): Likewise.
6287 (gimple_omp_taskreg_data_arg): Likewise.
6288 (gimple_omp_taskreg_data_arg_ptr): Likewise.
6289 (gimple_omp_taskreg_set_data_arg): Likewise.
6290 (gimple_omp_task_copy_fn): Likewise.
6291 (gimple_omp_task_copy_fn_ptr): Likewise.
6292 (gimple_omp_task_set_copy_fn): Likewise.
6293 (gimple_omp_task_arg_size): Likewise.
6294 (gimple_omp_task_arg_size_ptr): Likewise.
6295 (gimple_omp_task_set_arg_size): Likewise.
6296 (gimple_omp_task_arg_align): Likewise.
6297 (gimple_omp_task_arg_align_ptr): Likewise.
6298 (gimple_omp_task_set_arg_align): Likewise.
6299 (gimple_omp_single_clauses): Likewise.
6300 (gimple_omp_single_clauses_ptr): Likewise.
6301 (gimple_omp_single_set_clauses): Likewise.
6302 (gimple_omp_target_clauses): Likewise.
6303 (gimple_omp_target_clauses_ptr): Likewise.
6304 (gimple_omp_target_set_clauses): Likewise.
6305 (gimple_omp_target_child_fn): Likewise.
6306 (gimple_omp_target_child_fn_ptr): Likewise.
6307 (gimple_omp_target_set_child_fn): Likewise.
6308 (gimple_omp_target_data_arg): Likewise.
6309 (gimple_omp_target_data_arg_ptr): Likewise.
6310 (gimple_omp_target_set_data_arg): Likewise.
6311 (gimple_omp_teams_clauses): Likewise.
6312 (gimple_omp_teams_clauses_ptr): Likewise.
6313 (gimple_omp_teams_set_clauses): Likewise.
6314 (gimple_omp_sections_clauses): Likewise.
6315 (gimple_omp_sections_clauses_ptr): Likewise.
6316 (gimple_omp_sections_set_clauses): Likewise.
6317 (gimple_omp_sections_control): Likewise.
6318 (gimple_omp_sections_control_ptr): Likewise.
6319 (gimple_omp_sections_set_control): Likewise.
6320 (gimple_omp_for_set_cond): Likewise.
6321 (gimple_omp_for_cond): Likewise.
6322 (gimple_omp_atomic_store_set_val): Likewise.
6323 (gimple_omp_atomic_store_val): Likewise.
6324 (gimple_omp_atomic_store_val_ptr): Likewise.
6325 (gimple_omp_atomic_load_set_lhs): Likewise.
6326 (gimple_omp_atomic_load_lhs): Likewise.
6327 (gimple_omp_atomic_load_lhs_ptr): Likewise.
6328 (gimple_omp_atomic_load_set_rhs): Likewise.
6329 (gimple_omp_atomic_load_rhs): Likewise.
6330 (gimple_omp_atomic_load_rhs_ptr): Likewise.
6331 (gimple_omp_continue_control_def): Likewise.
6332 (gimple_omp_continue_control_def_ptr): Likewise.
6333 (gimple_omp_continue_set_control_def): Likewise.
6334 (gimple_omp_continue_control_use): Likewise.
6335 (gimple_omp_continue_control_use_ptr): Likewise.
6336 (gimple_omp_continue_set_control_use): Likewise.
6337 (gimple_transaction_body_ptr): Likewise.
6338 (gimple_transaction_label): Likewise.
6339 (gimple_transaction_label_ptr): Likewise.
6340 (gimple_transaction_set_body): Likewise.
6341 (gimple_transaction_set_label): Likewise.
6343 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
6344 * ipa-inline-analysis.c (inline_write_summary): Likewise.
6345 * ipa-ref.c (ipa_record_reference): Likewise.
6346 * ipa-reference.c (analyze_function): Likewise.
6347 (ipa_reference_write_optimization_summary): Likewise.
6348 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
6349 (address_taken_from_non_vtable_p): Likewise.
6350 (comdat_can_be_unshared_p_1): Likewise.
6351 * lto-cgraph.c (lto_output_ref): Likewise.
6352 (add_references): Likewise.
6353 (compute_ltrans_boundary): Likewise.
6354 (output_symtab): Likewise.
6355 (input_ref): Likewise.
6356 (input_cgraph_1): Likewise.
6357 (output_cgraph_opt_summary): Likewise.
6358 * lto-streamer-out.c (lto_output): Likewise.
6359 (output_symbol_p): Likewise.
6360 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
6361 (lsei_start_function_in_partition): Likewise.
6362 (lsei_next_variable_in_partition): Likewise.
6363 (lsei_start_variable_in_partition): Likewise.
6364 * symtab.c (insert_to_assembler_name_hash): Likewise.
6365 (unlink_from_assembler_name_hash): Likewise.
6366 (symtab_unregister_node): Likewise.
6367 (symtab_remove_node): Likewise.
6368 (dump_symtab_node): Likewise.
6369 (verify_symtab_base): Likewise.
6370 (verify_symtab_node): Likewise.
6371 (symtab_make_decl_local): Likewise.
6372 (symtab_alias_ultimate_target): Likewise.
6373 (symtab_resolve_alias): Likewise.
6374 (symtab_get_symbol_partitioning_class): Likewise.
6375 * tree-phinodes.c (allocate_phi_node): Likewise.
6376 (reserve_phi_args_for_new_edge): Likewise.
6377 (remove_phi_args): Likewise.
6378 * varpool.c (varpool_node_for_asm): Likewise.
6379 (varpool_remove_unreferenced_decls): Likewise.
6381 2014-04-23 Jeff Law <law@redhat.com>
6383 PR tree-optimization/60902
6384 * tree-ssa-threadedge.c
6385 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
6386 invalidate outputs from statements that do not produce useful
6387 outputs for threading.
6389 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
6391 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
6392 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
6393 machine descriptions for Stack Smashing Protector.
6395 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
6397 * aarch64.md (<optab>_rol<mode>3): New pattern.
6398 (<optab>_rolsi3_uxtw): Likewise.
6399 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
6401 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
6403 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
6404 (arm_cortex_a12_tune): Likewise.
6406 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6408 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
6410 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6412 * config/arm/arm.md (arm_rev16si2): New pattern.
6413 (arm_rev16si2_alt): Likewise.
6414 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
6416 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6418 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
6419 (rev16<mode>2_alt): Likewise.
6420 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
6421 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
6422 (aarch_rev16_shleft_mask_imm_p): Likewise.
6423 (aarch_rev16_p_1): Likewise.
6424 (aarch_rev16_p): Likewise.
6425 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
6426 (aarch_rev16_shright_mask_imm_p): Likewise.
6427 (aarch_rev16_shleft_mask_imm_p): Likewise.
6429 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6431 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
6432 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
6434 (cortex_a53_extra_costs): Likewise.
6435 (cortex_a57_extra_costs): Likewise.
6436 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
6437 (cortexa7_extra_costs): Likewise.
6438 (cortexa8_extra_costs): Likewise.
6439 (cortexa12_extra_costs): Likewise.
6440 (cortexa15_extra_costs): Likewise.
6441 (v7m_extra_costs): Likewise.
6442 (arm_new_rtx_costs): Handle BSWAP.
6444 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6446 * config/arm/arm.c (cortexa8_extra_costs): New table.
6447 (arm_cortex_a8_tune): New tuning struct.
6448 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
6450 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6452 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
6454 2014-04-23 Richard Biener <rguenther@suse.de>
6456 * Makefile.in (OBJS): Remove loop-unswitch.o.
6457 * tree-pass.h (make_pass_rtl_unswitch): Remove.
6458 * passes.def (pass_rtl_unswitch): Likewise.
6459 * loop-init.c (gate_rtl_unswitch): Likewise.
6460 (rtl_unswitch): Likewise.
6461 (pass_data_rtl_unswitch): Likewise.
6462 (pass_rtl_unswitch): Likewise.
6463 (make_pass_rtl_unswitch): Likewise.
6464 * rtl.h (reversed_condition): Likewise.
6465 (compare_and_jump_seq): Likewise.
6466 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
6468 * loop-unroll.c (compare_and_jump_seq): Likewise.
6470 2014-04-23 Richard Biener <rguenther@suse.de>
6472 PR tree-optimization/60903
6473 * tree-ssa-loop-im.c (analyze_memory_references): Remove
6474 commented code block.
6475 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
6476 loop flags to newly created BBs and edges.
6478 2014-04-23 Nick Clifton <nickc@redhat.com>
6480 * config/msp430/msp430.c (msp430_handle_option): Move function
6482 (msp430_option_override): Simplify mcu and mcpu option handling.
6483 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
6484 support for -mhwmult command line option.
6485 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
6486 -mhwmult command line option.
6487 (msp430_hwmult_enabled): Delete.
6488 (msp43o_output_labelref): Add support for -mhwmult command line option.
6489 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
6490 (umulsidi3): Likewise.
6491 * config/msp430/msp430.opt (mmcu): Add Report attribute.
6492 (mcpu, mlarge, msmall): Likewise.
6493 (mhwmult): New option.
6494 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
6496 (msp430_is_f5_mcu): Remove prototype.
6497 (msp430_use_f5_series_hwmult): Add prototype.
6498 * config/msp430/msp430-opts.h: New file.
6499 * common/config/msp430: New directory.
6500 * common/config/msp430/msp430-common.c: New file.
6501 * config.gcc (msp430): Remove target_has_targetm_common.
6502 * doc/invoke.texi: Document -mhwmult command line option.
6504 2014-04-23 Nick Clifton <nickc@redhat.com>
6506 * config/i386/cygwin.h (ENDFILE_SPEC): Include
6507 default-manifest.o if it can be found in the search path.
6508 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
6510 2014-04-23 Terry Guo <terry.guo@arm.com>
6512 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
6514 2014-04-23 Richard Biener <rguenther@suse.de>
6517 * tree-inline.c (declare_return_variable): Use mark_addressable.
6519 2014-04-23 Richard Biener <rguenther@suse.de>
6522 * loop-init.c (loop_optimizer_init): Make sure to apply
6523 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
6525 2014-04-22 Jakub Jelinek <jakub@redhat.com>
6528 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
6530 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
6531 if flag_sanitize_undefined_trap_on_error.
6532 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
6533 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
6534 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
6535 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
6536 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
6537 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
6538 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
6539 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
6540 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
6541 * ubsan.c (ubsan_instrument_unreachable): Return
6542 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
6543 (ubsan_expand_null_ifn): Emit __builtin_trap ()
6544 if flag_sanitize_undefined_trap_on_error and
6545 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
6546 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
6547 instrument_bool_enum_load): Emit __builtin_trap () if
6548 flag_sanitize_undefined_trap_on_error and
6549 __builtin_handle_*_abort () if !flag_sanitize_recover.
6550 * doc/invoke.texi (-fsanitize-recover,
6551 -fsanitize-undefined-trap-on-error): Document.
6553 2014-04-22 Christian Bruel <christian.bruel@st.com>
6555 * config/sh/sh.md (mov<mode>): Replace movQIHI.
6556 Force immediates to SImode.
6558 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
6560 * config/nios2/nios2.md (UNSPEC_ROUND): New.
6562 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
6563 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
6564 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
6565 (nios2_fpu_insn): Add entry for round.
6566 (N2FPU_NO_ERRNO_P): Define.
6567 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
6569 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
6571 2014-04-22 Richard Henderson <rth@redhat.com>
6573 * config/aarch64/aarch64 (addti3, subti3): New expanders.
6574 (add<GPI>3_compare0): Remove leading * from name.
6575 (add<GPI>3_carryin): Likewise.
6576 (sub<GPI>3_compare0): Likewise.
6577 (sub<GPI>3_carryin): Likewise.
6578 (<su_optab>mulditi3): New expander.
6579 (multi3): New expander.
6580 (madd<GPI>): Remove leading * from name.
6582 2014-04-22 Martin Jambor <mjambor@suse.cz>
6584 * cgraphclones.c (cgraph_function_versioning): Copy
6585 ipa_transforms_to_apply instead of asserting it is empty.
6587 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
6590 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
6591 on count_exp to get mode.
6593 2014-04-22 Andrew Pinski <apinski@cavium.com>
6595 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
6596 Handle TLS for ILP32.
6597 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
6598 (tlsie_small_<mode>): this and handle PTR.
6599 (tlsie_small_sidi): New pattern.
6600 (tlsle_small): Change to an expand to handle ILP32.
6601 (tlsle_small_<mode>): New pattern.
6602 (tlsdesc_small): Rename to ...
6603 (tlsdesc_small_<mode>): this and handle PTR.
6605 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6607 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
6609 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6611 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
6612 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
6613 (aarch64_types_signed_poly_qualifiers): Likewise.
6614 (aarch64_types_unsigned_signed_qualifiers): Likewise.
6615 (aarch64_types_poly_signed_qualifiers): Likewise.
6616 (TYPES_REINTERP_SS): Type macro added.
6617 (TYPES_REINTERP_SU): Likewise.
6618 (TYPES_REINTERP_SP): Likewise.
6619 (TYPES_REINTERP_US): Likewise.
6620 (TYPES_REINTERP_PS): Likewise.
6621 (aarch64_fold_builtin): New expression folding added.
6622 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
6623 Declarations removed.
6624 (REINTERP_SS): Declarations added.
6625 (REINTERP_US): Likewise.
6626 (REINTERP_PS): Likewise.
6627 (REINTERP_SU): Likewise.
6628 (REINTERP_SP): Likewise.
6629 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
6630 (vreinterpretq_p8_f64): Likewise.
6631 (vreinterpret_p16_f64): Likewise.
6632 (vreinterpretq_p16_f64): Likewise.
6633 (vreinterpret_f32_f64): Likewise.
6634 (vreinterpretq_f32_f64): Likewise.
6635 (vreinterpret_f64_f32): Likewise.
6636 (vreinterpret_f64_p8): Likewise.
6637 (vreinterpret_f64_p16): Likewise.
6638 (vreinterpret_f64_s8): Likewise.
6639 (vreinterpret_f64_s16): Likewise.
6640 (vreinterpret_f64_s32): Likewise.
6641 (vreinterpret_f64_s64): Likewise.
6642 (vreinterpret_f64_u8): Likewise.
6643 (vreinterpret_f64_u16): Likewise.
6644 (vreinterpret_f64_u32): Likewise.
6645 (vreinterpret_f64_u64): Likewise.
6646 (vreinterpretq_f64_f32): Likewise.
6647 (vreinterpretq_f64_p8): Likewise.
6648 (vreinterpretq_f64_p16): Likewise.
6649 (vreinterpretq_f64_s8): Likewise.
6650 (vreinterpretq_f64_s16): Likewise.
6651 (vreinterpretq_f64_s32): Likewise.
6652 (vreinterpretq_f64_s64): Likewise.
6653 (vreinterpretq_f64_u8): Likewise.
6654 (vreinterpretq_f64_u16): Likewise.
6655 (vreinterpretq_f64_u32): Likewise.
6656 (vreinterpretq_f64_u64): Likewise.
6657 (vreinterpret_s64_f64): Likewise.
6658 (vreinterpretq_s64_f64): Likewise.
6659 (vreinterpret_u64_f64): Likewise.
6660 (vreinterpretq_u64_f64): Likewise.
6661 (vreinterpret_s8_f64): Likewise.
6662 (vreinterpretq_s8_f64): Likewise.
6663 (vreinterpret_s16_f64): Likewise.
6664 (vreinterpretq_s16_f64): Likewise.
6665 (vreinterpret_s32_f64): Likewise.
6666 (vreinterpretq_s32_f64): Likewise.
6667 (vreinterpret_u8_f64): Likewise.
6668 (vreinterpretq_u8_f64): Likewise.
6669 (vreinterpret_u16_f64): Likewise.
6670 (vreinterpretq_u16_f64): Likewise.
6671 (vreinterpret_u32_f64): Likewise.
6672 (vreinterpretq_u32_f64): Likewise.
6674 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6676 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
6677 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
6678 (vreinterpret_p8_s8): Likewise.
6679 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
6680 (vreinterpret_p8_s16): Likewise.
6681 (vreinterpret_p8_s32): Likewise.
6682 (vreinterpret_p8_s64): Likewise.
6683 (vreinterpret_p8_f32): Likewise.
6684 (vreinterpret_p8_u8): Likewise.
6685 (vreinterpret_p8_u16): Likewise.
6686 (vreinterpret_p8_u32): Likewise.
6687 (vreinterpret_p8_u64): Likewise.
6688 (vreinterpret_p8_p16): Likewise.
6689 (vreinterpretq_p8_s8): Likewise.
6690 (vreinterpretq_p8_s16): Likewise.
6691 (vreinterpretq_p8_s32): Likewise.
6692 (vreinterpretq_p8_s64): Likewise.
6693 (vreinterpretq_p8_f32): Likewise.
6694 (vreinterpretq_p8_u8): Likewise.
6695 (vreinterpretq_p8_u16): Likewise.
6696 (vreinterpretq_p8_u32): Likewise.
6697 (vreinterpretq_p8_u64): Likewise.
6698 (vreinterpretq_p8_p16): Likewise.
6699 (vreinterpret_p16_s8): Likewise.
6700 (vreinterpret_p16_s16): Likewise.
6701 (vreinterpret_p16_s32): Likewise.
6702 (vreinterpret_p16_s64): Likewise.
6703 (vreinterpret_p16_f32): Likewise.
6704 (vreinterpret_p16_u8): Likewise.
6705 (vreinterpret_p16_u16): Likewise.
6706 (vreinterpret_p16_u32): Likewise.
6707 (vreinterpret_p16_u64): Likewise.
6708 (vreinterpret_p16_p8): Likewise.
6709 (vreinterpretq_p16_s8): Likewise.
6710 (vreinterpretq_p16_s16): Likewise.
6711 (vreinterpretq_p16_s32): Likewise.
6712 (vreinterpretq_p16_s64): Likewise.
6713 (vreinterpretq_p16_f32): Likewise.
6714 (vreinterpretq_p16_u8): Likewise.
6715 (vreinterpretq_p16_u16): Likewise.
6716 (vreinterpretq_p16_u32): Likewise.
6717 (vreinterpretq_p16_u64): Likewise.
6718 (vreinterpretq_p16_p8): Likewise.
6719 (vreinterpret_f32_s8): Likewise.
6720 (vreinterpret_f32_s16): Likewise.
6721 (vreinterpret_f32_s32): Likewise.
6722 (vreinterpret_f32_s64): Likewise.
6723 (vreinterpret_f32_u8): Likewise.
6724 (vreinterpret_f32_u16): Likewise.
6725 (vreinterpret_f32_u32): Likewise.
6726 (vreinterpret_f32_u64): Likewise.
6727 (vreinterpret_f32_p8): Likewise.
6728 (vreinterpret_f32_p16): Likewise.
6729 (vreinterpretq_f32_s8): Likewise.
6730 (vreinterpretq_f32_s16): Likewise.
6731 (vreinterpretq_f32_s32): Likewise.
6732 (vreinterpretq_f32_s64): Likewise.
6733 (vreinterpretq_f32_u8): Likewise.
6734 (vreinterpretq_f32_u16): Likewise.
6735 (vreinterpretq_f32_u32): Likewise.
6736 (vreinterpretq_f32_u64): Likewise.
6737 (vreinterpretq_f32_p8): Likewise.
6738 (vreinterpretq_f32_p16): Likewise.
6739 (vreinterpret_s64_s8): Likewise.
6740 (vreinterpret_s64_s16): Likewise.
6741 (vreinterpret_s64_s32): Likewise.
6742 (vreinterpret_s64_f32): Likewise.
6743 (vreinterpret_s64_u8): Likewise.
6744 (vreinterpret_s64_u16): Likewise.
6745 (vreinterpret_s64_u32): Likewise.
6746 (vreinterpret_s64_u64): Likewise.
6747 (vreinterpret_s64_p8): Likewise.
6748 (vreinterpret_s64_p16): Likewise.
6749 (vreinterpretq_s64_s8): Likewise.
6750 (vreinterpretq_s64_s16): Likewise.
6751 (vreinterpretq_s64_s32): Likewise.
6752 (vreinterpretq_s64_f32): Likewise.
6753 (vreinterpretq_s64_u8): Likewise.
6754 (vreinterpretq_s64_u16): Likewise.
6755 (vreinterpretq_s64_u32): Likewise.
6756 (vreinterpretq_s64_u64): Likewise.
6757 (vreinterpretq_s64_p8): Likewise.
6758 (vreinterpretq_s64_p16): Likewise.
6759 (vreinterpret_u64_s8): Likewise.
6760 (vreinterpret_u64_s16): Likewise.
6761 (vreinterpret_u64_s32): Likewise.
6762 (vreinterpret_u64_s64): Likewise.
6763 (vreinterpret_u64_f32): Likewise.
6764 (vreinterpret_u64_u8): Likewise.
6765 (vreinterpret_u64_u16): Likewise.
6766 (vreinterpret_u64_u32): Likewise.
6767 (vreinterpret_u64_p8): Likewise.
6768 (vreinterpret_u64_p16): Likewise.
6769 (vreinterpretq_u64_s8): Likewise.
6770 (vreinterpretq_u64_s16): Likewise.
6771 (vreinterpretq_u64_s32): Likewise.
6772 (vreinterpretq_u64_s64): Likewise.
6773 (vreinterpretq_u64_f32): Likewise.
6774 (vreinterpretq_u64_u8): Likewise.
6775 (vreinterpretq_u64_u16): Likewise.
6776 (vreinterpretq_u64_u32): Likewise.
6777 (vreinterpretq_u64_p8): Likewise.
6778 (vreinterpretq_u64_p16): Likewise.
6779 (vreinterpret_s8_s16): Likewise.
6780 (vreinterpret_s8_s32): Likewise.
6781 (vreinterpret_s8_s64): Likewise.
6782 (vreinterpret_s8_f32): Likewise.
6783 (vreinterpret_s8_u8): Likewise.
6784 (vreinterpret_s8_u16): Likewise.
6785 (vreinterpret_s8_u32): Likewise.
6786 (vreinterpret_s8_u64): Likewise.
6787 (vreinterpret_s8_p8): Likewise.
6788 (vreinterpret_s8_p16): Likewise.
6789 (vreinterpretq_s8_s16): Likewise.
6790 (vreinterpretq_s8_s32): Likewise.
6791 (vreinterpretq_s8_s64): Likewise.
6792 (vreinterpretq_s8_f32): Likewise.
6793 (vreinterpretq_s8_u8): Likewise.
6794 (vreinterpretq_s8_u16): Likewise.
6795 (vreinterpretq_s8_u32): Likewise.
6796 (vreinterpretq_s8_u64): Likewise.
6797 (vreinterpretq_s8_p8): Likewise.
6798 (vreinterpretq_s8_p16): Likewise.
6799 (vreinterpret_s16_s8): Likewise.
6800 (vreinterpret_s16_s32): Likewise.
6801 (vreinterpret_s16_s64): Likewise.
6802 (vreinterpret_s16_f32): Likewise.
6803 (vreinterpret_s16_u8): Likewise.
6804 (vreinterpret_s16_u16): Likewise.
6805 (vreinterpret_s16_u32): Likewise.
6806 (vreinterpret_s16_u64): Likewise.
6807 (vreinterpret_s16_p8): Likewise.
6808 (vreinterpret_s16_p16): Likewise.
6809 (vreinterpretq_s16_s8): Likewise.
6810 (vreinterpretq_s16_s32): Likewise.
6811 (vreinterpretq_s16_s64): Likewise.
6812 (vreinterpretq_s16_f32): Likewise.
6813 (vreinterpretq_s16_u8): Likewise.
6814 (vreinterpretq_s16_u16): Likewise.
6815 (vreinterpretq_s16_u32): Likewise.
6816 (vreinterpretq_s16_u64): Likewise.
6817 (vreinterpretq_s16_p8): Likewise.
6818 (vreinterpretq_s16_p16): Likewise.
6819 (vreinterpret_s32_s8): Likewise.
6820 (vreinterpret_s32_s16): Likewise.
6821 (vreinterpret_s32_s64): Likewise.
6822 (vreinterpret_s32_f32): Likewise.
6823 (vreinterpret_s32_u8): Likewise.
6824 (vreinterpret_s32_u16): Likewise.
6825 (vreinterpret_s32_u32): Likewise.
6826 (vreinterpret_s32_u64): Likewise.
6827 (vreinterpret_s32_p8): Likewise.
6828 (vreinterpret_s32_p16): Likewise.
6829 (vreinterpretq_s32_s8): Likewise.
6830 (vreinterpretq_s32_s16): Likewise.
6831 (vreinterpretq_s32_s64): Likewise.
6832 (vreinterpretq_s32_f32): Likewise.
6833 (vreinterpretq_s32_u8): Likewise.
6834 (vreinterpretq_s32_u16): Likewise.
6835 (vreinterpretq_s32_u32): Likewise.
6836 (vreinterpretq_s32_u64): Likewise.
6837 (vreinterpretq_s32_p8): Likewise.
6838 (vreinterpretq_s32_p16): Likewise.
6839 (vreinterpret_u8_s8): Likewise.
6840 (vreinterpret_u8_s16): Likewise.
6841 (vreinterpret_u8_s32): Likewise.
6842 (vreinterpret_u8_s64): Likewise.
6843 (vreinterpret_u8_f32): Likewise.
6844 (vreinterpret_u8_u16): Likewise.
6845 (vreinterpret_u8_u32): Likewise.
6846 (vreinterpret_u8_u64): Likewise.
6847 (vreinterpret_u8_p8): Likewise.
6848 (vreinterpret_u8_p16): Likewise.
6849 (vreinterpretq_u8_s8): Likewise.
6850 (vreinterpretq_u8_s16): Likewise.
6851 (vreinterpretq_u8_s32): Likewise.
6852 (vreinterpretq_u8_s64): Likewise.
6853 (vreinterpretq_u8_f32): Likewise.
6854 (vreinterpretq_u8_u16): Likewise.
6855 (vreinterpretq_u8_u32): Likewise.
6856 (vreinterpretq_u8_u64): Likewise.
6857 (vreinterpretq_u8_p8): Likewise.
6858 (vreinterpretq_u8_p16): Likewise.
6859 (vreinterpret_u16_s8): Likewise.
6860 (vreinterpret_u16_s16): Likewise.
6861 (vreinterpret_u16_s32): Likewise.
6862 (vreinterpret_u16_s64): Likewise.
6863 (vreinterpret_u16_f32): Likewise.
6864 (vreinterpret_u16_u8): Likewise.
6865 (vreinterpret_u16_u32): Likewise.
6866 (vreinterpret_u16_u64): Likewise.
6867 (vreinterpret_u16_p8): Likewise.
6868 (vreinterpret_u16_p16): Likewise.
6869 (vreinterpretq_u16_s8): Likewise.
6870 (vreinterpretq_u16_s16): Likewise.
6871 (vreinterpretq_u16_s32): Likewise.
6872 (vreinterpretq_u16_s64): Likewise.
6873 (vreinterpretq_u16_f32): Likewise.
6874 (vreinterpretq_u16_u8): Likewise.
6875 (vreinterpretq_u16_u32): Likewise.
6876 (vreinterpretq_u16_u64): Likewise.
6877 (vreinterpretq_u16_p8): Likewise.
6878 (vreinterpretq_u16_p16): Likewise.
6879 (vreinterpret_u32_s8): Likewise.
6880 (vreinterpret_u32_s16): Likewise.
6881 (vreinterpret_u32_s32): Likewise.
6882 (vreinterpret_u32_s64): Likewise.
6883 (vreinterpret_u32_f32): Likewise.
6884 (vreinterpret_u32_u8): Likewise.
6885 (vreinterpret_u32_u16): Likewise.
6886 (vreinterpret_u32_u64): Likewise.
6887 (vreinterpret_u32_p8): Likewise.
6888 (vreinterpret_u32_p16): Likewise.
6889 (vreinterpretq_u32_s8): Likewise.
6890 (vreinterpretq_u32_s16): Likewise.
6891 (vreinterpretq_u32_s32): Likewise.
6892 (vreinterpretq_u32_s64): Likewise.
6893 (vreinterpretq_u32_f32): Likewise.
6894 (vreinterpretq_u32_u8): Likewise.
6895 (vreinterpretq_u32_u16): Likewise.
6896 (vreinterpretq_u32_u64): Likewise.
6897 (vreinterpretq_u32_p8): Likewise.
6898 (vreinterpretq_u32_p16): Likewise.
6900 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6902 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
6904 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
6906 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
6907 (vqnegd_s64): Likewise.
6908 (vqabs_s64): Likewise.
6909 (vqabsd_s64): Likewise.
6911 2014-04-22 Richard Henderson <rth@redhat.com>
6913 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
6914 computation to the top of the loop.
6916 2014-04-22 Renlin <renlin.li@arm.com>
6917 Jiong Wang <jiong.wang@arm.com>
6919 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
6920 * config/aarch64/aarch64.c (aarch64_layout_frame)
6921 (aarch64_initial_elimination_offset): Likewise.
6923 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
6925 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
6928 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
6930 * machmode.h (bitwise_mode_for_mode): Declare.
6931 * stor-layout.h (bitwise_type_for_mode): Likewise.
6932 * stor-layout.c (bitwise_mode_for_mode): New function.
6933 (bitwise_type_for_mode): Likewise.
6934 * builtins.c (fold_builtin_memory_op): Use it instead of
6935 int_mode_for_mode and build_nonstandard_integer_type.
6937 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6939 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
6940 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
6941 (*-*-solaris2*): Simplify.
6942 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
6943 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
6944 *-*-solaris2.9* handling.
6946 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
6948 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
6949 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
6951 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
6952 * configure: Regenerate.
6954 * config/i386/sol2-9.h: Remove.
6956 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
6957 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
6958 Remove Solaris 9 references.
6960 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
6962 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
6963 (floatuns<GPI:mode><GPF:mode>2): Remove.
6964 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
6965 and floatuns conversions.
6966 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
6967 and floatuns conversions.
6968 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
6969 (w1,w2): New mode attributes for inequal width conversions.
6971 2014-04-22 Renlin Li <Renlin.Li@arm.com>
6973 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
6974 the output asm format.
6976 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
6978 * config/aarch64/aarch64-simd.md
6979 (aarch64_cm<optab>di): Always split.
6980 (*aarch64_cm<optab>di): New.
6981 (aarch64_cmtstdi): Always split.
6982 (*aarch64_cmtstdi): New.
6984 2014-04-22 Jakub Jelinek <jakub@redhat.com>
6986 PR tree-optimization/60823
6987 * omp-low.c (ipa_simd_modify_function_body): Go through
6988 all SSA_NAMEs and for those refering to vector arguments
6989 which are going to be replaced adjust SSA_NAME_VAR and,
6990 if it is a default definition, change it into a non-default
6991 definition assigned at the beginning of function from new_decl.
6992 (ipa_simd_modify_stmt_ops): Rewritten.
6993 * tree-dfa.c (set_ssa_default_def): When removing default def,
6994 check for NULL loc instead of NULL *loc.
6996 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6998 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
6999 restrictions on core registers for DImode values in Thumb2.
7001 2014-04-22 Ian Bolton <ian.bolton@arm.com>
7003 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
7004 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
7006 2014-04-22 Ian Bolton <ian.bolton@arm.com>
7008 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
7009 (*iordi_notzesidi_di): Likewise.
7010 (*iordi_notsesidi_di): Likewise.
7012 2014-04-22 Ian Bolton <ian.bolton@arm.com>
7014 * config/arm/arm-protos.h (tune_params): New struct members.
7015 * config/arm/arm.c: Initialise tune_params per processor.
7016 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
7017 for speed, based on new tune_params.
7019 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
7021 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
7022 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
7023 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
7024 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
7025 * config/aarch64/arm_neon.h (vrnd_f64): Added.
7026 (vrnda_f64): Likewise.
7027 (vrndi_f64): Likewise.
7028 (vrndm_f64): Likewise.
7029 (vrndn_f64): Likewise.
7030 (vrndp_f64): Likewise.
7031 (vrndx_f64): Likewise.
7033 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7035 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
7036 GET_MODE_SIZE argument is enum machine_mode.
7038 2014-04-22 Jakub Jelinek <jakub@redhat.com>
7041 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
7042 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
7044 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
7047 * asan.c (asan_emit_stack_protection): Force the base to align to
7048 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
7049 appropriate bits if STRICT_ALIGNMENT.
7050 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
7052 (expand_used_vars): Leave a space in the stack frame for alignment
7053 if STRICT_ALIGNMENT.
7055 2014-04-21 David Malcolm <dmalcolm@redhat.com>
7057 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
7059 (gimple_store_p): Likewise.
7060 (gimple_assign_load_p): Likewise.
7061 (gimple_assign_cast_p): Likewise.
7062 (gimple_clobber_p): Likewise.
7064 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
7065 rather than a gimple.
7066 (gimple_assign_cast_p): Likewise.
7068 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
7071 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
7072 If mode is DDmode and TARGET_E500_DOUBLE allow move.
7074 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
7075 more debug information for E500 if -mdebug=reg.
7077 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
7080 * config/i386/i386.c (ix86_expand_builtin)
7081 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
7082 register for target RTX.
7083 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
7085 2014-04-18 Cong Hou <congh@google.com>
7087 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
7088 the widen-mult pattern by handling two operands with different sizes,
7089 and operands whose size is smaller than half of the result type.
7091 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
7093 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
7094 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
7095 (do_estimate_edge_time): Compute it.
7096 * ipa-inline.c (want_inline_small_function_p): Bypass
7097 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
7099 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
7101 * ipa-inline.c (spec_rem): New static variable.
7102 (dump_overall_stats): New function.
7103 (dump_inline_stats): New function.
7105 2014-04-18 Richard Henderson <rth@redhat.com>
7107 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
7108 to GET_MODE_SIZE, not a reg_class_t.
7110 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7112 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
7113 (vsx_xxmrglw_<mode>): Likewise.
7115 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
7118 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
7119 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
7120 (rs6000_init_hard_regno_mode_ok): Likewise.
7122 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
7124 * ipa-inline.c (inline_small_functions): Account only non-cold
7126 * doc/invoke.texi (inline-unit-growth): Update documentation.
7128 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
7130 * config/rs6000/rs6000.md (addti3, subti3): New.
7132 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
7135 * config/i386/i386.c (ix86_expand_clear): Remove outdated
7136 comment. Check optimize_insn_for_size_p instead of
7137 optimize_insn_for_speed_p.
7139 2014-04-17 Martin Jambor <mjambor@suse.cz>
7141 * gimple-iterator.c (gsi_start_edge): New function.
7142 * gimple-iterator.h (gsi_start_edge): Declare.
7143 * tree-sra.c (single_non_eh_succ): New function.
7144 (disqualify_ops_if_throwing_stmt): Renamed to
7145 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
7146 having one non-EH successor BB.
7147 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
7148 generate loads into replacements.
7149 (sra_modify_assign): Likewise and and also use the simple path for
7151 (sra_modify_function_body): Commit statements on edges.
7153 2014-04-17 Richard Biener <rguenther@suse.de>
7156 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
7157 comparison results and add clarifying comment.
7159 2014-04-17 Jakub Jelinek <jakub@redhat.com>
7161 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
7162 (blank_mode): Initialize it.
7163 (emit_mode_size_inline, emit_mode_nunits_inline,
7164 emit_mode_inner_inline): New functions.
7165 (emit_insn_modes_h): Call them and surround their output with
7166 #if GCC_VERSION >= 4001 ... #endif.
7167 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
7168 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
7169 mode_* arrays if the argument is __builtin_constant_p.
7170 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
7171 is enum machine_mode.
7173 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
7175 * passes.c (opt_pass::execute): Adjust.
7176 (pass_manager::execute_pass_mode_switching): Likewise.
7177 (early_local_passes::execute): Likewise.
7178 (execute_one_pass): Pass cfun to the pass's execute method.
7179 * tree-pass.h (opt_pass::execute): Add function * argument.
7180 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
7181 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
7182 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
7183 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
7184 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
7185 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
7186 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
7187 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
7188 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
7189 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
7190 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
7191 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
7192 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
7193 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
7194 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
7195 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
7196 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
7197 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
7198 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
7199 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
7200 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
7201 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
7202 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
7203 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
7204 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
7205 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
7206 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
7207 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
7208 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
7211 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
7213 * passes.c (opt_pass::gate): Take function * argument.
7214 (gate_all_early_local_passes): Merge into
7215 (early_local_passes::gate): this.
7216 (gate_all_early_optimizations): Merge into
7217 (all_early_optimizations::gate): this.
7218 (gate_all_optimizations): Mege into
7219 (all_optimizations::gate): this.
7220 (gate_all_optimizations_g): Merge into
7221 (all_optimizations_g::gate): this.
7222 (gate_rest_of_compilation): Mege into
7223 (rest_of_compilation::gate): this.
7224 (gate_postreload): Merge into
7225 (postreload::gate): this.
7226 (dump_one_pass): Pass cfun to the pass's gate method.
7227 (execute_ipa_summary_passes): Likewise.
7228 (execute_one_pass): Likewise.
7229 (ipa_write_summaries_2): Likewise.
7230 (ipa_write_optimization_summaries_1): Likewise.
7231 (ipa_read_summaries_1): Likewise.
7232 (ipa_read_optimization_summaries_1): Likewise.
7233 (execute_ipa_stmt_fixups): Likewise.
7234 * tree-pass.h (opt_pass::gate): Add function * argument.
7235 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
7236 combine-stack-adj.c, combine.c, compare-elim.c,
7237 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
7238 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
7239 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
7240 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
7241 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
7242 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
7243 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
7244 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
7245 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
7246 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
7247 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
7248 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
7249 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
7250 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
7251 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
7252 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
7253 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
7254 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
7255 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
7256 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
7257 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
7258 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
7259 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
7260 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
7261 var-tracking.c, vtable-verify.c, web.c: Adjust.
7263 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
7265 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
7266 * configure: Regenerate.
7268 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
7270 * passes.c (dump_one_pass): don't check pass->has_gate.
7271 (execute_ipa_summary_passes): Likewise.
7272 (execute_one_pass): Likewise.
7273 (ipa_write_summaries_2): Likewise.
7274 (ipa_write_optimization_summaries_1): Likewise.
7275 (ipa_read_optimization_summaries_1): Likewise.
7276 (execute_ipa_stmt_fixups): Likewise.
7277 * tree-pass.h (pass_data::has_gate): Remove.
7278 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
7279 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
7280 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
7281 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
7282 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
7283 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
7284 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
7285 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
7286 gimple-low.c, gimple-ssa-isolate-paths.c,
7287 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
7288 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
7289 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
7290 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
7291 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
7292 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
7293 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
7294 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
7295 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
7296 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
7297 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
7298 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
7299 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
7300 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
7301 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
7302 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
7303 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
7304 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
7305 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
7306 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
7307 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
7308 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
7311 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
7313 * pass_manager.h (pass_manager::register_dump_files_1): Remove
7315 * passes.c (pass_manager::register_dump_files_1): Merge into
7316 (pass_manager::register_dump_files): this, and remove its handling of
7317 properties since the pass always has the properties anyway.
7318 (pass_manager::pass_manager): Adjust.
7320 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
7322 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
7323 * passes.c (pass_manager::register_dump_files_1): Remove dead code
7324 dealing with properties.
7325 (pass_manager::register_dump_files): Adjust.
7327 2014-03-20 Mark Wielaard <mjw@redhat.com>
7329 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
7330 then represent the bound as normal constant value.
7332 2014-04-17 Jakub Jelinek <jakub@redhat.com>
7335 Forward port from 4.8 branch
7336 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
7338 * config/i386/bmiintrin.h (_blsi_u32): New.
7342 (_blsmsk_u32): Ditto.
7343 (_blsmsk_u64): Ditto.
7344 (_tzcnt_u32): Ditto.
7345 (_tzcnt_u64): Ditto.
7347 2014-04-17 Kito Cheng <kito@0xlab.org>
7349 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
7351 2014-04-17 Richard Biener <rguenther@suse.de>
7354 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
7355 boolean results for comparisons.
7357 2014-04-17 Richard Biener <rguenther@suse.de>
7359 PR tree-optimization/60836
7360 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
7361 initial PHI args to be gimple values.
7363 2014-04-17 Richard Biener <rguenther@suse.de>
7365 PR tree-optimization/60841
7366 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
7367 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
7368 of stmts to SLP build.
7369 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
7370 (vect_analyze_slp): Likewise.
7371 (vect_analyze_slp_instance): Likewise.
7372 (vect_build_slp_tree): Limit overall SLP tree growth.
7373 * tree-vectorizer.h (vect_analyze_data_refs,
7374 vect_analyze_slp): Adjust prototypes.
7376 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
7378 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
7381 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
7383 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
7384 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
7385 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
7386 for TARGET_SLOW_PSHUFB
7388 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
7390 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
7391 * config/i386/i386.c (intel_cost): Ditto.
7393 2014-04-17 Joey Ye <joey.ye@arm.com>
7395 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
7397 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
7399 * opts.c (common_handle_option): Disable -fipa-reference coorectly
7402 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
7404 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
7405 (type_all_derivations_known_p): New predicate.
7406 (type_all_ctors_visible_p): New predicate.
7407 (type_possibly_instantiated_p): New predicate.
7408 (get_odr_type): Compute all_derivations_known.
7409 (dump_odr_type): Dump the flag.
7410 (maybe_record_type): Cleanup.
7411 (record_target_from_binfo): Add bases_to_consider array;
7412 record bases for types w/o instances and skip CXX destructor.
7413 (possible_polymorphic_call_targets_1): Add bases_to_consider
7414 and consider_construction parameters; check if type may have instance.
7415 (get_polymorphic_call_info): Set maybe_in_construction to true
7416 when we know nothing.
7417 (record_targets_from_bases): Skip CXX destructors; they are
7418 never called for types in construction.
7419 (possible_polymorphic_call_targets): Do not record target when
7420 type may not have instance.
7422 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
7425 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
7426 external aliases alive, too.
7428 2014-04-16 Andrew Pinski <apinski@cavium.com>
7430 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
7433 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
7435 * final.c (compute_alignments): Do not apply loop alignment to a block
7436 falling through to the exit.
7438 2014-04-16 Catherine Moore <clm@codesourcery.com>
7440 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
7441 Adjust constraints for microMIPS store patterns.
7443 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
7445 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
7447 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
7449 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
7450 (append_use): Run at -O0.
7451 (append_vdef): Likewise.
7452 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
7453 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
7455 2014-04-16 Jakub Jelinek <jakub@redhat.com>
7457 PR tree-optimization/60844
7458 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
7459 (propagate_op_to_single_use, remove_visited_stmt_chain,
7460 linearize_expr, repropagate_negates, reassociate_bb): Use it
7461 instead of gsi_remove.
7463 2014-04-16 Martin Jambor <mjambor@suse.cz>
7465 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
7466 ipa_transforms_to_apply.
7467 (cgraph_function_versioning): Assert that old_node has empty
7468 ipa_transforms_to_apply.
7469 * trans-mem.c (ipa_tm_create_version): Likewise.
7470 * tree-inline.c (tree_function_versioning): Do not duplicate
7471 ipa_transforms_to_apply.
7473 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7476 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
7478 Pass necessary as flags on 64-bit Solaris/x86.
7479 Use lowercase relocs for x86_64-*-*.
7480 * configure: Regenerate.
7482 2014-04-15 Jan Hubicka <jh@suse.cz>
7484 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
7485 (maybe_record_node, likely_target_p): Use it.
7487 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7490 Revert following patch
7492 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
7495 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
7496 software floating point or no floating point registers, do not
7497 allow any type in the FPRs. Eliminate a test for SPE SIMD types
7498 in GPRs that occurs after we tested for GPRs that would never be
7501 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
7502 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
7503 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
7504 specifically allow DDmode, since that does not use the SPE SIMD
7507 2014-03-21 Mark Wielaard <mjw@redhat.com>
7509 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
7510 as unsigned or int depending on type and value used.
7512 2014-04-15 Richard Biener <rguenther@suse.de>
7514 PR rtl-optimization/56965
7515 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
7516 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
7518 * alias.c (true_dependence_1): Do not call
7519 nonoverlapping_component_refs_p.
7520 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
7521 nonoverlapping_component_refs_p.
7522 (indirect_refs_may_alias_p): Likewise.
7524 2014-04-15 Teresa Johnson <tejohnson@google.com>
7526 * cfg.c (dump_bb_info): Fix flags check.
7527 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
7529 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7531 PR rtl-optimization/60663
7532 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
7535 2014-04-15 Richard Biener <rguenther@suse.de>
7537 * lto-streamer.h (LTO_major_version): Bump to 4.
7539 2014-04-15 Richard Biener <rguenther@suse.de>
7541 * common.opt (lto_partition_model): New enum.
7542 (flto-partition=): Merge separate options with a single with argument,
7543 add -flto-partition=one support.
7544 * flag-types.h (enum lto_partition_model): Declare.
7545 * opts.c (finish_options): Remove duplicate -flto-partition=
7547 * lto-wrapper.c (run_gcc): Adjust.
7549 2014-04-15 Richard Biener <rguenther@suse.de>
7551 * alias.c (ncr_compar): New function.
7552 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
7554 2014-04-15 Richard Biener <rguenther@suse.de>
7556 * alias.c (record_component_aliases): Do not walk BINFOs.
7558 2014-04-15 Richard Biener <rguenther@suse.de>
7560 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7561 Add struct function argument and adjust.
7562 (find_func_aliases_for_call): Likewise.
7563 (find_func_aliases): Likewise.
7564 (find_func_clobbers): Likewise.
7565 (intra_create_variable_infos): Likewise.
7566 (compute_points_to_sets): Likewise.
7567 (ipa_pta_execute): Adjust. Do not push/pop cfun.
7569 2014-04-15 Richard Biener <rguenther@suse.de>
7571 * tree.c (iterative_hash_expr): Use enum tree_code_class
7572 to store TREE_CODE_CLASS.
7573 (tree_block): Likewise.
7574 (tree_set_block): Likewise.
7575 * tree.h (fold_build_pointer_plus_loc): Use
7576 convert_to_ptrofftype_loc.
7578 2014-04-15 Jakub Jelinek <jakub@redhat.com>
7581 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
7584 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
7586 * cfgloop.h (struct loop): Move force_vectorize down.
7587 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
7588 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
7589 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
7590 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
7591 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
7592 * tree-core.h (enum annot_expr_kind): Add new kind values.
7593 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
7594 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
7596 * tree.def (ANNOTATE_EXPR): Tweak comment.
7598 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
7600 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
7603 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
7605 * tree.h (TYPE_IDENTIFIER): Declare.
7606 * tree.c (subrange_type_for_debug_p): Use it.
7607 * godump.c (go_format_type): Likewise.
7608 * dwarf2out.c (is_cxx_auto, modified_type_die,
7609 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
7610 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
7612 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
7615 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
7617 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
7619 * config/i386/i386.c (examine_argument): Return bool. Return true if
7620 parameter should be passed in memory.
7621 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
7622 (construct_container): Update calls to examine_argument.
7623 (function_arg_advance_64): Ditto.
7624 (return_in_memory_32): Merge with ix86_return_in_memory.
7625 (return_in_memory_64): Ditto.
7626 (return_in_memory_ms_64): Ditto.
7628 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
7630 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
7631 * coverage.c (coverage_compute_profile_id): Handle externally visible
7634 2014-04-14 Martin Jambor <mjambor@suse.cz>
7636 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
7637 DECL_DISREGARD_INLINE_LIMITS functions.
7639 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
7642 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
7644 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
7647 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
7648 optimize_insn_for_speed_p instead of
7649 optimize_function_for_speed_p.
7651 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
7653 * doc/invoke.texi (free): Document AArch64.
7655 2014-04-14 Richard Biener <rguenther@suse.de>
7657 PR tree-optimization/60042
7658 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
7659 (insert_into_preds_of_block): Do not prevent PHI insertion
7660 for REFERENCE exprs here ...
7661 (eliminate_dom_walker::before_dom_children): ... but prevent
7662 their use here under similar conditions when applied to the
7663 IL after PRE optimizations.
7665 2014-04-14 Richard Biener <rguenther@suse.de>
7667 * passes.def: Move early points-to after early SRA.
7669 2014-04-14 Richard Biener <rguenther@suse.de>
7671 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
7672 check for which sign-changes we allow when forwarding
7673 a converted value into a switch.
7675 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
7677 * stor-layout.c (place_field): Finalize non-constant offset for the
7680 2014-04-14 Richard Biener <rguenther@suse.de>
7682 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
7684 (expand_switch_using_bit_tests_p): Likewise.
7685 (process_switch): Compute and pass on speed_p based on the
7687 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
7688 optimize_bb_for_speed_p.
7690 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
7692 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
7693 * function.h (struct function): Rename has_force_vect_loops into
7694 has_force_vectorize_loops.
7695 * lto-streamer-in.c (input_cfg): Adjust for renaming.
7696 (input_struct_function_base): Likewise.
7697 * lto-streamer-out.c (output_cfg): Likewise.
7698 (output_struct_function_base): Likewise.
7699 * omp-low.c (expand_omp_simd): Likewise.
7700 * tree-cfg.c (move_sese_region_to_fn): Likewise.
7701 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
7702 (version_loop_for_if_conversion): Likewise.
7703 (tree_if_conversion): Likewise.
7704 (main_tree_if_conversion): Likewise.
7705 (gate_tree_if_conversion): Likewise.
7706 * tree-inline.c (copy_loops): Likewise.
7707 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
7708 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
7709 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
7710 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
7711 * tree-vectorizer.c (vectorize_loops): Likewise.
7712 * tree-vectorizer.h (unlimited_cost_model): Likewise.
7714 2014-04-14 Richard Biener <rguenther@suse.de>
7717 * lto-streamer-out.c (wrap_refs): New function.
7718 (lto_output): Wrap symbol references in global initializes in
7719 type-preserving MEM_REFs.
7721 2014-04-14 Christian Bruel <christian.bruel@st.com>
7723 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
7725 2014-04-14 Christian Bruel <christian.bruel@st.com>
7727 * config/sh/sh.md (setmemqi): New expand pattern.
7728 * config/sh/sh.h (CLEAR_RATIO): Define.
7729 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
7730 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
7732 2014-04-14 Richard Biener <rguenther@suse.de>
7735 * fold-const.c (negate_expr_p): Don't negate directional rounding
7737 (fold_negate_expr): Likewise.
7739 2014-04-14 Richard Biener <rguenther@suse.de>
7741 PR tree-optimization/59817
7742 PR tree-optimization/60453
7743 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
7744 recursion to catch all CHRECs in the scalar evolution and restrict
7745 the predicate for the remains appropriately.
7747 2014-04-12 Catherine Moore <clm@codesourcery.com>
7749 * config/mips/constraints.md: Add new register constraint "kb".
7750 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
7751 (*movhi_internal): Likewise.
7752 (*movqi_internal): Likewise.
7753 * config/mips/mips.h (M16_STORE_REGS): New register class.
7754 (REG_CLASS_NAMES): Add M16_STORE_REGS.
7755 (REG_CLASS_CONTENTS): Likewise.
7756 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
7758 2014-04-11 Tobias Burnus <burnus@net-b.de>
7761 * doc/invoke.texi (-Wformat-signedness): Document it.
7762 (Wformat=2): Mention that this enables -Wformat-signedness.
7764 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
7766 * common/config/epiphany/epiphany-common.c
7767 (epiphany_option_optimization_table): Enable section anchors by
7768 default at -O1 or higher.
7769 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
7770 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
7771 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
7772 carries no extra cost.
7773 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
7774 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
7775 * config/epiphany/predicates.md (memclob_operand): New predicate.
7776 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
7777 Use memclob_operand predicate and X constraint for operand 3.
7779 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
7781 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
7782 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
7785 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
7787 PR rtl-optimization/60651
7788 * mode-switching.c (optimize_mode_switching): Make sure to emit
7789 sets of a lower numbered entity before sets of a higher numbered
7790 entity to a mode of the same or lower priority.
7791 When creating a seginfo for a basic block that starts with a code
7792 label, move the insertion point past the code label.
7793 (new_seginfo): Document and enforce requirement that
7794 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
7795 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
7796 * doc/tm.texi: Regenerate.
7798 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
7801 * config/arc/arc.c (arc_save_restore): Fix assert typo.
7803 2013-04-11 Jakub Jelinek <jakub@redhat.com>
7805 * BASE-VER: Set to 4.10.0.
7807 2014-04-11 Tobias Burnus <burnus@net-b.de>
7810 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
7811 * doc/gcc.texi (Service): Update description in the @menu
7812 * doc/invoke.texi (Option Summary): Remove misplaced and
7815 2014-04-11 Steve Ellcey <sellcey@mips.com>
7816 Jakub Jelinek <jakub@redhat.com>
7819 * expr.c (convert_move): Use emit_store_flag_force instead of
7820 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
7823 2014-04-11 Richard Biener <rguenther@suse.de>
7826 * varasm.c (assemble_alias): Avoid endless error reporting
7827 recursion by setting TREE_ASM_WRITTEN.
7829 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7831 * config/s390/s390.md: Add a splitter for NOT rtx.
7833 2014-04-11 Jakub Jelinek <jakub@redhat.com>
7835 PR rtl-optimization/60663
7836 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
7838 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
7839 Jakub Jelinek <jakub@redhat.com>
7842 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
7843 flag from decl_node to node.
7845 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7848 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
7849 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
7850 ameliorating the cases where it can be.
7852 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
7855 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
7857 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
7858 (loadsync_<mode>): Change mode.
7859 (load_quadpti, store_quadpti): New.
7860 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
7861 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
7862 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
7864 2014-04-09 Cong Hou <congh@google.com>
7867 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
7870 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7872 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
7873 instead of vnor to exploit possible fusion opportunity in the
7875 (altivec_expand_vec_perm_const_le): Likewise.
7877 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
7879 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
7880 (loadsync_<mode>): Change mode.
7881 (load_quadpti, store_quadpti): New.
7882 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
7883 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
7885 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
7888 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
7889 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
7890 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
7892 2014-04-08 Richard Biener <rguenther@suse.de>
7895 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
7896 a 64bit widest int print double-int similar to on HWI64 hosts.
7898 2014-04-08 Richard Biener <rguenther@suse.de>
7900 PR tree-optimization/60785
7901 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
7902 default defs properly.
7904 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
7906 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
7907 (Weffc++): Likewise.
7909 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
7911 * ipa-devirt.c (maybe_record_node): When node is not recorded,
7912 set completep to false rather than true.
7914 2014-04-07 Douglas B Rupp <rupp@adacore.com>
7917 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
7918 ARM_TARGET2_DWARF_FORMAT.
7920 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
7923 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
7924 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
7927 2014-04-07 Richard Biener <rguenther@suse.de>
7929 PR tree-optimization/60766
7930 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
7931 (may_eliminate_iv): Convert cand_value_at result to desired type.
7933 2014-04-07 Jason Merrill <jason@redhat.com>
7936 * common.opt (-fno-gnu-unique): Add.
7937 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
7939 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7941 * haifa-sched.c: Fix outdated function reference and minor
7942 grammar errors in introductory comment.
7944 2014-04-07 Richard Biener <rguenther@suse.de>
7947 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
7949 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
7951 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
7954 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
7955 size accounting for thunks.
7956 (pa_asm_output_mi_thunk): Use final_start_function() and
7957 final_end_function() to output function start and end directives.
7959 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
7961 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
7962 device specific ISA/ feature information. Remove short_sp and
7963 errata_skip ds. Add avr_device_specific_features enum to have device
7965 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
7966 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
7967 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
7968 updated device specific info.
7969 * config/avr/avr-mcus.def: Merge device specific details to
7970 dev_attribute field.
7971 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
7973 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
7974 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
7975 assembler if RMW isa supported by current device.
7976 * config/avr/genmultilib.awk: Update as device info structure changed.
7977 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
7979 2014-04-04 Cong Hou <congh@google.com>
7981 PR tree-optimization/60656
7982 * tree-vect-stmts.c (supportable_widening_operation):
7983 Fix a bug that elements in a vector with vect_used_by_reduction
7984 property are incorrectly reordered when the operation on it is not
7985 consistant with the one in reduction operation.
7987 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
7989 PR rtl-optimization/60155
7990 * gcse.c (record_set_data): New function.
7991 (single_set_gcse): New function.
7992 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
7993 (hoist_code): Likewise.
7994 (get_pressure_class_and_nregs): Likewise.
7996 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
7998 * explow.c (probe_stack_range): Emit a final optimization blockage.
8000 2014-04-04 Anthony Green <green@moxielogic.com>
8002 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
8005 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
8008 * lto-cgraph.c (input_overwrite_node): Check that partitioning
8009 flags are set only during streaming.
8010 * ipa.c (process_references, walk_polymorphic_call_targets,
8011 symtab_remove_unreachable_nodes): Drop bodies of always inline
8012 after early inlining.
8013 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
8015 2014-04-04 Jakub Jelinek <jakub@redhat.com>
8016 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8019 * dwarf2out.c (const_ok_for_output_1): Reject expressions
8022 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8025 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
8027 (cortex_a53_fdivd): Likewise.
8029 2014-04-04 Martin Jambor <mjambor@suse.cz>
8032 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
8034 * cgraph.c (clone_of_p): Also return true if thunks match.
8035 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
8036 cgraph_function_or_thunk_node and an obsolete comment.
8037 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
8039 (build_function_decl_skip_args): Likewise.
8040 (set_new_clone_decl_and_node_flags): New function.
8041 (duplicate_thunk_for_node): Likewise.
8042 (redirect_edge_duplicating_thunks): Likewise.
8043 (cgraph_clone_node): New parameter args_to_skip, pass it to
8044 redirect_edge_duplicating_thunks which is called instead of
8045 cgraph_redirect_edge_callee.
8046 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
8047 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
8049 2014-04-04 Jeff Law <law@redhat.com>
8052 * config/arm/predicates.md (const_int_I_operand): New predicate.
8053 (const_int_M_operand): Similarly.
8054 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
8056 (insv_t2, extv_reg, extzv_t2): Likewise.
8057 (load_multiple_with_writeback): Similarly for const_int_I_operand.
8058 (pop_multiple_with_writeback_and_return): Likewise.
8059 (vfp_pop_multiple_with_writeback): Likewise
8061 2014-04-04 Richard Biener <rguenther@suse.de>
8064 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
8065 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
8067 * gimplify.h (gimple_add_tmp_var_fn): Declare.
8068 * gimplify.c (gimple_add_tmp_var_fn): New function.
8069 * gimple-expr.h (create_tmp_reg_fn): Declare.
8070 * gimple-expr.c (create_tmp_reg_fn): New function.
8071 * gimple-low.c (record_vars_into): Don't change cfun.
8072 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
8073 code generation without cfun.
8075 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
8078 * Makefile.in (install-driver): Fix shell scripting.
8080 2014-04-03 Cong Hou <congh@google.com>
8082 PR tree-optimization/60505
8083 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
8084 threshold of number of iterations below which no vectorization
8086 * tree-vect-loop.c (new_loop_vec_info):
8087 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
8088 * tree-vect-loop.c (vect_analyze_loop_operations):
8089 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
8090 * tree-vect-loop.c (vect_transform_loop):
8091 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
8092 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
8093 of iterations of the loop and see if we should build the epilogue.
8095 2014-04-03 Richard Biener <rguenther@suse.de>
8097 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
8098 (streamer_tree_cache_create): Adjust.
8099 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
8100 to allow optional nodes array.
8101 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
8102 (streamer_tree_cache_append): Likewise.
8103 (streamer_tree_cache_create): Create nodes array optionally
8104 as specified by parameter.
8105 * lto-streamer-out.c (create_output_block): Avoid maintaining
8106 the node array in the writer cache.
8107 (DFS_write_tree): Remove assertion.
8108 (produce_asm_for_decls): Free the out decl state hash table early.
8109 * lto-streamer-in.c (lto_data_in_create): Adjust for
8110 streamer_tree_cache_create prototype change.
8112 2014-04-03 Richard Biener <rguenther@suse.de>
8114 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
8115 set TREE_CHAIN to NULL_TREE.
8117 2014-04-03 Richard Biener <rguenther@suse.de>
8119 PR tree-optimization/60740
8120 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
8121 over all GIMPLE_COND operands.
8123 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
8125 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
8126 (Weffc++): Remove Scott's numbering, merge lists and reference
8129 2014-04-03 Nick Clifton <nickc@redhat.com>
8131 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
8134 2014-04-03 Martin Jambor <mjambor@suse.cz>
8136 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
8137 mention gcc_unreachable before failing.
8138 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
8141 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
8144 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
8145 inconsistent code and instead mark the context inconsistent.
8146 (possible_polymorphic_call_targets): For inconsistent contexts
8147 return empty complete list.
8149 2014-04-02 Anthony Green <green@moxielogic.com>
8151 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
8152 (extendqisi2, extendhisi2): Define.
8153 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
8154 (WCHAR_TYPE): Change to unsigned int.
8156 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8158 PR tree-optimization/60733
8159 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
8160 insertion point for PHI candidates to be the end of the feeding
8161 block for the PHI argument.
8163 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
8165 PR rtl-optimization/60650
8166 * lra-constraints.c (process_alt_operands): Decrease reject for
8167 earlyclobber matching.
8169 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8171 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
8173 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
8175 * config/spu/spu.c (pad_bb): Do not crash when the last
8176 insn is CODE_FOR_blockage.
8178 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
8180 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
8181 lies outside the target mode.
8183 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
8186 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
8187 software floating point or no floating point registers, do not
8188 allow any type in the FPRs. Eliminate a test for SPE SIMD types
8189 in GPRs that occurs after we tested for GPRs that would never be
8192 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
8193 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
8194 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
8195 specifically allow DDmode, since that does not use the SPE SIMD
8198 2014-04-02 Richard Biener <rguenther@suse.de>
8201 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
8202 MODE_INTs. Properly use negv_optab.
8203 (expand_abs): Likewise.
8205 2014-04-02 Richard Biener <rguenther@suse.de>
8208 * Makefile.in (install-driver): Guard extra installs with special
8211 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
8213 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
8216 2014-04-01 Richard Henderson <rth@redhat.com>
8219 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
8220 alternative enabled before register allocation.
8222 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
8224 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
8225 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
8227 (nios2_large_got_address): Remove unneeded 'sym' parameter.
8228 (nios2_got_address): Update nios2_large_got_address call site.
8229 (nios2_delegitimize_address): New function.
8230 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
8231 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
8232 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
8234 2014-04-01 Martin Husemann <martin@duskware.de>
8236 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
8239 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
8241 PR rtl-optimization/60604
8242 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
8243 check from register_operand.
8244 (register_operand): Redefine in terms of general_operand.
8245 (nonmemory_operand): Use register_operand for the non-constant cases.
8247 2014-04-01 Richard Biener <rguenther@suse.de>
8249 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
8251 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
8253 * doc/invoke.texi (mapp-regs): Clarify.
8255 2014-03-31 Ulrich Drepper <drepper@gmail.com>
8257 * config/i386/avx512fintrin.h (__v32hi): Define type.
8258 (__v64qi): Likewise.
8259 (_mm512_set1_epi8): Define.
8260 (_mm512_set1_epi16): Define.
8261 (_mm512_set4_epi32): Define.
8262 (_mm512_set4_epi64): Define.
8263 (_mm512_set4_pd): Define.
8264 (_mm512_set4_ps): Define.
8265 (_mm512_setr4_epi64): Define.
8266 (_mm512_setr4_epi32): Define.
8267 (_mm512_setr4_pd): Define.
8268 (_mm512_setr4_ps): Define.
8269 (_mm512_setzero_epi32): Define.
8271 2014-03-31 Martin Jambor <mjambor@suse.cz>
8274 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
8275 callsite_arguments_match_p. Updated all callers. Also check types of
8276 corresponding formal parameters and actual arguments.
8277 (not_all_callers_have_enough_arguments_p) Renamed to
8278 some_callers_have_mismatched_arguments_p.
8280 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
8282 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
8284 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
8287 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
8290 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
8292 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
8293 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
8295 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
8296 Use FMAMODE_NOVF512 mode iterator.
8297 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
8298 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
8299 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
8300 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
8302 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
8303 Use VF_128_256 mode iterator.
8304 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
8307 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
8309 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
8310 static chain if needed.
8312 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
8315 * lra-constraints.c (index_part_to_reg): New.
8316 (process_address): Use it.
8318 2014-03-27 Jeff Law <law@redhat.com>
8319 Jakub Jelinek <jakub@redhat.com>
8322 * expr.c (do_tablejump): Use simplify_gen_binary rather than
8323 gen_rtx_{PLUS,MULT} to build up the address expression.
8325 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
8326 creating non-canonical RTL.
8328 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
8331 * ipa-inline.c (want_inline_small_function_p): Short circuit large
8332 functions; reorganize to make cheap checks first.
8333 (inline_small_functions): Do not estimate growth when dumping;
8335 * ipa-inline.h (inline_summary): Add min_size.
8336 (growth_likely_positive): New function.
8337 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
8338 (set_cond_stmt_execution_predicate): Cleanup.
8339 (estimate_edge_size_and_time): Compute min_size.
8340 (estimate_calls_size_and_time): Likewise.
8341 (estimate_node_size_and_time): Likewise.
8342 (inline_update_overall_summary): Update min_size.
8343 (do_estimate_edge_time): Likewise.
8344 (do_estimate_edge_size): Update.
8345 (do_estimate_edge_hints): Update.
8346 (growth_likely_positive): New function.
8348 2014-03-28 Jakub Jelinek <jakub@redhat.com>
8351 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
8352 also if addr has VOIDmode.
8354 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8356 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
8357 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
8359 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
8360 instructions as well as AdvancedSIMD loads.
8362 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8364 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
8365 Use crypto_aese type.
8366 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
8367 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
8368 crypto_aese, crypto_aesmc. Move to types.md.
8369 * config/arm/types.md (crypto_aes): Split into crypto_aese,
8371 * config/arm/iterators.md (crypto_type): Likewise.
8373 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
8375 * cgraph.c: Include expr.h and tree-dfa.h.
8376 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
8379 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
8382 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
8383 regs from checking multi-reg pseudos.
8385 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8387 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
8389 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
8391 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
8392 if it would clobber the stack pointer, even temporarily.
8394 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
8396 * mode-switching.c: Make small adjustments to the top comment.
8398 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
8400 * config/rs6000/constraints.md (wD constraint): New constraint to
8401 match the constant integer to get the top DImode/DFmode out of a
8402 vector in a VSX register.
8404 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
8405 match the constant integer to get the top DImode/DFmode out of a
8406 vector in a VSX register.
8408 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
8411 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
8414 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
8415 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
8417 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
8418 Optimize vec_extract of 64-bit values, where the value being
8419 extracted is in the top word, where we can use scalar
8420 instructions. Add direct move and store support. Combine the big
8421 endian/little endian vector select load support into a single insn.
8422 (vsx_extract_<mode>_internal1): Likewise.
8423 (vsx_extract_<mode>_internal2): Likewise.
8424 (vsx_extract_<mode>_load): Likewise.
8425 (vsx_extract_<mode>_store): Likewise.
8426 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
8427 combined into vsx_extract_<mode>_load.
8428 (vsx_extract_<mode>_one_le): Likewise.
8430 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
8431 define the top 64-bit vector element.
8433 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
8436 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
8437 Document vec_vbpermq builtin.
8440 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
8441 enable use of xxsldwi and xxpermdi builtin functions.
8442 (vec_xxpermdi): Likewise.
8444 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
8445 Document use of vec_xxsldwi and vec_xxpermdi builtins.
8447 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
8449 PR rtl-optimization/60650
8450 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
8452 (find_spills_for): New.
8453 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
8454 Spill all pseudos on the second iteration.
8456 2014-03-27 Marek Polacek <polacek@redhat.com>
8459 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
8462 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8464 * config/s390/s390.c (s390_can_use_return_insn): Check for
8465 call-saved FPRs on 31 bit.
8467 2014-03-27 Jakub Jelinek <jakub@redhat.com>
8470 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
8471 if they need regimplification, just drop them instead of
8472 calling gimple_regimplify_operands on them.
8474 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
8477 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
8478 (aarch64_frame_pointer_required): Adjust logic.
8479 (aarch64_can_eliminate): Adjust logic.
8480 (aarch64_override_options_after_change): Adjust logic.
8482 2014-03-27 Dehao Chen <dehao@google.com>
8484 * ipa-inline.c (early_inliner): Update node's inline info.
8486 2014-03-26 Dehao Chen <dehao@google.com>
8488 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
8489 compiler inserted conditional jumps for NAN float check.
8491 2014-03-26 Jakub Jelinek <jakub@redhat.com>
8493 * ubsan.h (ubsan_create_data): Change second argument's type
8494 to const location_t *.
8495 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
8497 (ubsan_create_data): Change second argument to const location_t *PLOC.
8498 Create Loc field whenever PLOC is non-NULL.
8499 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
8500 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
8504 * real.c (real_to_integer2): Change type of low to UHWI.
8506 2014-03-26 Tobias Burnus <burnus@net-b.de>
8508 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
8509 (CILK_SELF_SPECS): New define.
8510 (driver_self_specs): Use it.
8512 2014-03-26 Richard Biener <rguenther@suse.de>
8514 * tree-pretty-print.c (percent_K_format): Implement special
8515 case for LTO and its stripped down BLOCK tree.
8517 2014-03-26 Jakub Jelinek <jakub@redhat.com>
8520 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
8522 * tree-vrp.c (simplify_internal_call_using_ranges): If only
8523 one range is range_int_cst_p, but not both, at least optimize
8524 addition/subtraction of 0 and multiplication by 0 or 1.
8525 * gimple-fold.c (gimple_fold_call): Fold
8526 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
8527 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
8528 INTEGER_CSTs, try to fold at least x * 0 and y - y.
8530 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
8532 PR rtl-optimization/60452
8533 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
8534 <case REG>: Return 1 for invalid offsets from the frame pointer.
8536 2014-03-26 Marek Polacek <polacek@redhat.com>
8539 * doc/extend.texi (C Extensions): Mention variable-length arrays in
8542 2014-03-26 Marek Polacek <polacek@redhat.com>
8545 * doc/extend.texi (Designated Inits): Describe what happens to omitted
8548 2014-03-26 Marek Polacek <polacek@redhat.com>
8551 * ira-color.c (update_conflict_hard_regno_costs): Perform the
8552 multiplication in unsigned type.
8554 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
8556 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
8558 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
8560 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
8562 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
8565 * cif-code.def (UNREACHABLE) New code.
8566 * ipa-inline.c (inline_small_functions): Skip edges to
8567 __builtlin_unreachable.
8568 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
8569 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
8570 predicate to __bulitin_unreachable.
8571 (set_cond_stmt_execution_predicate): Fix issue when
8572 invert_tree_comparison returns ERROR_MARK.
8573 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
8574 propagate to inline clones.
8575 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
8577 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
8578 * cgraphclones.c (cgraph_clone_node): If call destination is already
8579 ureachable, do not redirect it back.
8580 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
8583 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
8585 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
8586 Do not modify inline clones.
8588 2014-03-25 Jakub Jelinek <jakub@redhat.com>
8590 * config/i386/i386.md (general_sext_operand): New mode attr.
8591 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
8592 don't generate (sign_extend (const_int)).
8593 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
8594 operands[2]. Use We constraint instead of <i> and
8595 <general_sext_operand> predicate instead of <general_operand>.
8596 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
8597 * config/i386/constraints.md (We): New constraint.
8598 * config/i386/predicates.md (x86_64_sext_operand,
8599 sext_operand): New predicates.
8601 2014-03-25 Martin Jambor <mjambor@suse.cz>
8604 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
8605 inconsistent devirtualizations to __builtin_unreachable.
8607 2014-03-25 Marek Polacek <polacek@redhat.com>
8610 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
8612 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
8614 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
8615 order of elements for big-endian.
8617 2014-03-25 Richard Biener <rguenther@suse.de>
8620 * gimplify-me.c (gimple_regimplify_operands): Update the
8623 2014-03-25 Martin Jambor <mjambor@suse.cz>
8626 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
8627 (lto_output_varpool_node): Likewise.
8628 (input_overwrite_node): Likewise.
8629 (input_varpool_node): Likewise.
8631 2014-03-25 Richard Biener <rguenther@suse.de>
8633 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
8634 (run_gcc): Likewise.
8636 2014-03-25 Jakub Jelinek <jakub@redhat.com>
8638 * combine.c (simplify_compare_const): Add MODE argument.
8639 Handle mode_width 0 as very large mode_width.
8640 (try_combine, simplify_comparison): Adjust callers.
8642 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
8643 type to avoid signed integer overflow.
8644 * explow.c (plus_constant): Likewise.
8646 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
8648 * doc/generic.texi: Correct typos.
8650 2014-03-24 Tobias Burnus <burnus@net-b.de>
8652 * doc/invoke.texi (-flto): Expand section about
8653 using static libraries with LTO.
8655 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8657 PR rtl-optimization/60501
8658 * optabs.def (addptr3_optab): New optab.
8659 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
8660 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
8661 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
8663 * lra.c (emit_add3_insn): Use the addptr pattern if available.
8665 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
8667 2014-03-24 Ulrich Drepper <drepper@gmail.com>
8669 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
8672 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
8673 (_mm256_undefined_ps): Define.
8674 (_mm256_undefined_pd): Define.
8675 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
8676 (_mm_undefined_pd): Define.
8677 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
8678 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
8679 (_mm512_undefined_ps): Define.
8680 (_mm512_undefined_pd): Define.
8681 Use _mm*_undefined_*.
8682 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
8684 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
8686 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
8687 (lshr_simd): DI mode added.
8688 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
8689 (aarch64_ushr_simddi): Likewise.
8690 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
8691 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
8692 (vshrd_n_u64): Likewise.
8694 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8696 * Makefile.in (s-macro_list): Depend on cc1.
8698 2014-03-23 Teresa Johnson <tejohnson@google.com>
8700 * ipa-utils.c (ipa_print_order): Use specified dump file.
8702 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
8704 PR rtl-optimization/60601
8705 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
8707 * gcc.c (eval_spec_function): Initialize save_growing_value.
8709 2014-03-22 Jakub Jelinek <jakub@redhat.com>
8712 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
8713 code == MINUS_EXPR, never swap op0 with op1.
8715 * toplev.c (init_local_tick): Avoid signed integer multiplication
8717 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
8718 shift by first operand's bitsize.
8720 2014-03-21 Jakub Jelinek <jakub@redhat.com>
8723 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
8725 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
8726 TARGET_ISA_64BIT_P(x).
8728 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8730 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
8731 pattern for vector nor instead of subtract from splat(-1).
8732 (altivec_expand_vec_perm_const_le): Likewise.
8734 2014-03-21 Richard Henderson <rth@twiddle.net>
8737 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
8738 related insns after epilogue_completed.
8740 2014-03-21 Martin Jambor <mjambor@suse.cz>
8743 * cgraph.h (symtab_node): New flag body_removed.
8744 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
8745 when removing bodies.
8746 * symtab.c (dump_symtab_base): Dump body_removed flag.
8747 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
8748 had their bodies removed.
8750 2014-03-21 Martin Jambor <mjambor@suse.cz>
8753 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
8756 2014-03-21 Richard Biener <rguenther@suse.de>
8758 PR tree-optimization/60577
8759 * tree-core.h (struct tree_base): Document nothrow_flag use
8760 in VAR_DECL_NONALIASED.
8761 * tree.h (VAR_DECL_NONALIASED): New.
8762 (may_be_aliased): Adjust.
8763 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
8765 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
8767 * expr.c (expand_expr_real_1): Remove outdated comment.
8769 2014-03-20 Jakub Jelinek <jakub@redhat.com>
8772 * ira.c (adjust_cleared_regs): Call copy_rtx on
8773 *reg_equiv[REGNO (loc)].src_p before passing it to
8774 simplify_replace_fn_rtx.
8777 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
8778 into CONST, put pic register as first operand of PLUS. Use
8779 gen_const_mem for both 32-bit and 64-bit PIC got loads.
8781 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8783 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
8785 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
8787 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
8788 around for store forwarding issue in the FPU on the UT699.
8789 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
8790 loads and operations if -mfix-ut699 is specified.
8791 (divtf3_hq): Tweak attribute.
8792 (sqrttf2_hq): Likewise.
8794 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
8796 * calls.c (store_one_arg): Remove incorrect const qualification on the
8797 type of the temporary.
8798 * cfgexpand.c (expand_return): Likewise.
8799 * expr.c (expand_constructor): Likewise.
8800 (expand_expr_real_1): Likewise.
8802 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8804 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
8807 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
8810 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
8812 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
8814 * config/arm/aarch-common-protos.h
8815 (alu_cost_table): Fix spelling of "extend".
8816 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
8818 2014-03-19 Richard Biener <rguenther@suse.de>
8821 * tree-core.h (tree_type_common): Re-order pointer members
8822 to reduce recursion depth during GC walks.
8824 2014-03-19 Marek Polacek <polacek@redhat.com>
8827 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
8828 before accessing it.
8830 2014-03-19 Richard Biener <rguenther@suse.de>
8833 * lto-streamer-in.c (input_function): In WPA stage do not drop
8836 2014-03-19 Jakub Jelinek <jakub@redhat.com>
8838 PR tree-optimization/60559
8839 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
8840 with build_zero_cst assignment.
8842 2014-03-18 Kai Tietz <ktietz@redhat.com>
8844 PR rtl-optimization/56356
8845 * sdbout.c (sdbout_parms): Verify that parms'
8846 incoming argument is valid.
8847 (sdbout_reg_parms): Likewise.
8849 2014-03-18 Richard Henderson <rth@redhat.com>
8852 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
8853 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
8854 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
8856 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
8858 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
8859 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
8860 Italicize plugin event names in description. Explain that
8861 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
8862 Remind that no GCC functions should be called after PLUGIN_FINISH.
8863 Explain what pragmas with expansion are.
8865 2014-03-18 Martin Liska <mliska@suse.cz>
8867 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
8868 gimple call statement is update.
8869 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
8870 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
8872 2014-03-18 Jakub Jelinek <jakub@redhat.com>
8875 * ubsan.c (ubsan_instrument_unreachable): Call
8876 initialize_sanitizer_builtins.
8877 (ubsan_pass): Likewise.
8880 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
8881 varpool_finalize_decl instead of rest_of_decl_compilation.
8883 2014-03-18 Richard Biener <rguenther@suse.de>
8885 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
8886 by using bitmap_and_compl instead of bitmap_and_compl_into.
8887 (df_rd_transfer_function): Likewise.
8889 2014-03-18 Richard Biener <rguenther@suse.de>
8891 * doc/lto.texi (fresolution): Fix typo.
8893 2014-03-18 Richard Biener <rguenther@suse.de>
8895 * doc/invoke.texi (flto): Update for changes in 4.9.
8897 2014-03-18 Richard Biener <rguenther@suse.de>
8899 * doc/loop.texi: Remove section on the removed lambda framework.
8900 Update loop docs with recent changes in preserving loop structure.
8902 2014-03-18 Richard Biener <rguenther@suse.de>
8904 * doc/lto.texi (-fresolution): Document.
8906 2014-03-18 Richard Biener <rguenther@suse.de>
8908 * doc/contrib.texi: Adjust my name.
8910 2014-03-18 Jakub Jelinek <jakub@redhat.com>
8913 * internal-fn.c: Include diagnostic-core.h.
8914 (expand_BUILTIN_EXPECT): New function.
8915 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
8916 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
8917 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
8918 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
8920 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
8921 Revert 3 argument __builtin_expect code.
8922 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
8923 * gimple-fold.c (gimple_fold_call): Likewise.
8924 * tree.h (fold_builtin_expect): New prototype.
8925 * builtins.c (build_builtin_expect_predicate): Add predictor
8926 argument, if non-NULL, create 3 argument __builtin_expect.
8927 (fold_builtin_expect): No longer static. Add ARG2 argument,
8928 pass it through to build_builtin_expect_predicate.
8929 (fold_builtin_2): Adjust caller.
8930 (fold_builtin_3): Handle BUILT_IN_EXPECT.
8931 * internal-fn.def (BUILTIN_EXPECT): New.
8933 2014-03-18 Tobias Burnus <burnus@net-b.de>
8936 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
8937 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
8938 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
8940 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
8943 * predict.c (combine_predictions_for_bb): Fix up formatting.
8944 (expr_expected_value_1, expr_expected_value): Add predictor argument,
8945 fill what it points to if non-NULL.
8946 (tree_predict_by_opcode): Adjust caller, use the predictor.
8947 * predict.def (PRED_COMPARE_AND_SWAP): Add.
8949 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
8951 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
8952 proper constant for the store mode.
8954 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
8956 * symtab.c (change_decl_assembler_name): Fix transparent alias
8959 2014-03-16 Renlin Li <Renlin.Li@arm.com>
8961 * config/aarch64/aarch64.c: Correct the comments about the
8962 aarch64 stack layout.
8964 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
8966 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
8967 check for GF_OMP_FOR_KIND_FOR.
8969 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
8971 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
8972 ymm and zmm register names.
8974 2014-03-17 Jakub Jelinek <jakub@redhat.com>
8977 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
8978 note creation for the 2010-08-31 changes.
8980 2014-03-17 Marek Polacek <polacek@redhat.com>
8983 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
8984 as -fno-tree-loop-vectorize.
8985 (expand_omp_simd): Likewise.
8987 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
8989 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
8990 (eligible_for_call_delay): New prototype.
8991 * config/sparc/sparc.c (tls_call_delay): Rename into...
8992 (eligible_for_call_delay): ...this. Return false if the instruction
8993 cannot be put in the delay slot of a branch.
8994 (eligible_for_restore_insn): Simplify.
8995 (eligible_for_return_delay): Return false if the instruction cannot be
8996 put in the delay slot of a branch and simplify.
8997 (eligible_for_sibcall_delay): Return false if the instruction cannot be
8998 put in the delay slot of a branch.
8999 * config/sparc/sparc.md (fix_ut699): New attribute.
9000 (tls_call_delay): Delete.
9001 (in_call_delay): Reimplement.
9002 (eligible_for_sibcall_delay): Rename into...
9003 (in_sibcall_delay): ...this.
9004 (eligible_for_return_delay): Rename into...
9005 (in_return_delay): ...this.
9006 (in_branch_delay): Reimplement.
9007 (in_uncond_branch_delay): Delete.
9008 (in_annul_branch_delay): Delete.
9010 2014-03-14 Richard Henderson <rth@redhat.com>
9013 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
9014 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
9015 (*floathi<X87MODEF>2_i387_with_temp): Remove.
9016 (floathi splitters): Remove.
9017 (float<SWI48x>xf2): New pattern.
9018 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
9019 code that tried to handle DImode for 32-bit, but which was excluded
9020 by the pattern's condition. Drop allocation of stack temporary.
9021 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
9022 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
9023 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
9024 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
9025 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
9026 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
9027 (*float<SWI48><MODEF>2_sse_interunit): Remove.
9028 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
9029 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
9030 (*float<SWI48x><X87MODEF>2_i387): Remove.
9031 (all float _with_temp splitters): Remove.
9032 (*float<SWI48x><MODEF>2_i387): New pattern.
9033 (*float<SWI48><MODEF>2_sse): New pattern.
9034 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
9035 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
9037 2014-03-14 Jakub Jelinek <jakub@redhat.com>
9038 Marek Polacek <polacek@redhat.com>
9041 * common.opt (dump_base_name_prefixed): New Variable.
9042 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
9043 if x_dump_base_name_prefixed is already set, set it at the end.
9045 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
9047 PR rtl-optimization/60508
9048 * lra-constraints.c (get_reload_reg): Add new parameter
9050 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
9051 Pass the new parameter values.
9053 2014-03-14 Richard Biener <rguenther@suse.de>
9055 * common.opt: Revert unintented changes from r205065.
9058 2014-03-14 Richard Biener <rguenther@suse.de>
9061 * cfghooks.c (split_block): Properly adjust all loops the
9062 block was a latch of.
9064 2014-03-14 Martin Jambor <mjambor@suse.cz>
9067 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
9070 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
9073 * config/avr/avr.c (avr_set_current_function): Pass function name
9074 through default_strip_name_encoding before sanity checking instead
9075 of skipping the first char of the assembler name.
9077 2014-03-13 Richard Henderson <rth@redhat.com>
9080 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
9081 (ix86_force_to_memory, ix86_free_from_memory): Remove.
9082 * config/i386/i386-protos.h: Likewise.
9083 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
9084 in the expander instead of a splitter.
9085 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
9086 any possibility of requiring a memory.
9087 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
9088 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
9089 (fp branch splitters): Update for ix86_split_fp_branch.
9090 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
9091 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
9092 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
9093 (*fop_<MODEF>_2_i387): Remove f/r alternative.
9094 (*fop_<MODEF>_3_i387): Likewise.
9095 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
9096 (splitters for the fop_* register patterns): Remove.
9097 (fscalexf4_i387): Rename from *fscalexf4_i387.
9098 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
9100 2014-03-13 Jakub Jelinek <jakub@redhat.com>
9102 PR tree-optimization/59779
9103 * tree-dfa.c (get_ref_base_and_extent): Use double_int
9104 type for bitsize and maxsize instead of HOST_WIDE_INT.
9106 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
9108 PR rtl-optimization/57320
9109 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
9110 the CFG after thread_prologue_and_epilogue_insns.
9112 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
9114 PR rtl-optimization/57189
9115 * lra-constraints.c (process_alt_operands): Disfavor spilling
9118 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
9120 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
9122 2014-03-13 Jakub Jelinek <jakub@redhat.com>
9124 PR tree-optimization/59025
9126 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
9127 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
9129 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
9132 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
9133 calls of avr_out_plus_1.
9135 2014-03-13 Bin Cheng <bin.cheng@arm.com>
9137 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
9138 BB's single pred and update the father loop's latch info later.
9140 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
9142 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
9146 (VEC_base): Likewise.
9147 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
9148 registers, we need to swap double words in little endian mode.
9150 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
9151 to be a container mode for 128-bit integer operations added in ISA
9152 2.07. Unlike TImode and PTImode, the preferred register set is
9153 the Altivec/VMX registers for the 128-bit operations.
9155 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
9157 (rs6000_split_128bit_ok_p): Likewise.
9159 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
9160 macros for creating ISA 2.07 normal and overloaded builtin
9161 functions with 3 arguments.
9162 (BU_P8V_OVERLOAD_3): Likewise.
9163 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
9164 for use as overloaded functions.
9165 (VPERM_1TI_UNS): Likewise.
9166 (VSEL_1TI): Likewise.
9167 (VSEL_1TI_UNS): Likewise.
9168 (ST_INTERNAL_1ti): Likewise.
9169 (LD_INTERNAL_1ti): Likewise.
9170 (XXSEL_1TI): Likewise.
9171 (XXSEL_1TI_UNS): Likewise.
9172 (VPERM_1TI): Likewise.
9173 (VPERM_1TI_UNS): Likewise.
9174 (XXPERMDI_1TI): Likewise.
9175 (SET_1TI): Likewise.
9176 (LXVD2X_V1TI): Likewise.
9177 (STXVD2X_V1TI): Likewise.
9178 (VEC_INIT_V1TI): Likewise.
9179 (VEC_SET_V1TI): Likewise.
9180 (VEC_EXT_V1TI): Likewise.
9181 (EQV_V1TI): Likewise.
9182 (NAND_V1TI): Likewise.
9183 (ORC_V1TI): Likewise.
9184 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
9185 added in ISA 2.07. Add both normal 'altivec' builtins, and the
9187 (VADDUQM): Likewise.
9188 (VSUBCUQ): Likewise.
9189 (VADDEUQM): Likewise.
9190 (VADDECUQ): Likewise.
9191 (VSUBEUQM): Likewise.
9192 (VSUBECUQ): Likewise.
9194 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
9195 __int128_t and __uint128_t types.
9196 (__uint128_type): Likewise.
9197 (altivec_categorize_keyword): Add support for vector __int128_t,
9198 vector __uint128_t, vector __int128, and vector unsigned __int128
9199 as a container type for TImode operations that need to be done in
9200 VSX/Altivec registers.
9201 (rs6000_macro_to_expand): Likewise.
9202 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
9203 to support 128-bit integer instructions vaddcuq, vadduqm,
9204 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
9205 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
9207 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
9208 for V1TImode, and set up preferences to use VSX/Altivec registers.
9209 Setup VSX reload handlers.
9210 (rs6000_debug_reg_global): Likewise.
9211 (rs6000_init_hard_regno_mode_ok): Likewise.
9212 (rs6000_preferred_simd_mode): Likewise.
9213 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
9214 (easy_altivec_constant): Likewise.
9215 (output_vec_const_move): Likewise.
9216 (rs6000_expand_vector_set): Convert V1TImode set and extract to
9218 (rs6000_expand_vector_extract): Likewise.
9219 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
9221 (rs6000_const_vec): Add support for V1TImode.
9222 (rs6000_emit_le_vsx_load): Swap double words when loading or
9223 storing TImode/V1TImode.
9224 (rs6000_emit_le_vsx_store): Likewise.
9225 (rs6000_emit_le_vsx_move): Likewise.
9226 (rs6000_emit_move): Add support for V1TImode.
9227 (altivec_expand_ld_builtin): Likewise.
9228 (altivec_expand_st_builtin): Likewise.
9229 (altivec_expand_vec_init_builtin): Likewise.
9230 (altivec_expand_builtin): Likewise.
9231 (rs6000_init_builtins): Add support for V1TImode type. Add
9232 support for ISA 2.07 128-bit integer builtins. Define type names
9233 for the VSX/Altivec vector types.
9234 (altivec_init_builtins): Add support for overloaded vector
9235 functions with V1TImode type.
9236 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
9237 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
9239 (rs6000_split_128bit_ok_p): Likewise.
9240 (rs6000_handle_altivec_attribute): Create V1TImode from vector
9241 __int128_t and vector __uint128_t.
9243 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
9244 and mode attributes.
9251 (VS_scalar): Likewise.
9252 (VS_double): Likewise.
9253 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
9255 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
9256 we support the ISA 2.07 128-bit integer arithmetic instructions.
9257 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
9258 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
9259 and TImode types for use with the builtin functions.
9260 (V1TI_type_node): Likewise.
9261 (unsigned_V1TI_type_node): Likewise.
9262 (intTI_type_internal_node): Likewise.
9263 (uintTI_type_internal_node): Likewise.
9265 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
9266 128-bit builtin functions.
9267 (UNSPEC_VADDEUQM): Likewise.
9268 (UNSPEC_VADDECUQ): Likewise.
9269 (UNSPEC_VSUBCUQ): Likewise.
9270 (UNSPEC_VSUBEUQM): Likewise.
9271 (UNSPEC_VSUBECUQ): Likewise.
9272 (VM): Add V1TImode to vector mode iterators.
9274 (VI_unit): Likewise.
9275 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
9276 (altivec_vaddcuq): Likewise.
9277 (altivec_vsubuqm): Likewise.
9278 (altivec_vsubcuq): Likewise.
9279 (altivec_vaddeuqm): Likewise.
9280 (altivec_vaddecuq): Likewise.
9281 (altivec_vsubeuqm): Likewise.
9282 (altivec_vsubecuq): Likewise.
9284 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
9286 (BOOL_128): Likewise.
9287 (BOOL_REGS_OUTPUT): Likewise.
9288 (BOOL_REGS_OP1): Likewise.
9289 (BOOL_REGS_OP2): Likewise.
9290 (BOOL_REGS_UNARY): Likewise.
9291 (BOOL_REGS_AND_CR0): Likewise.
9293 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
9294 128-bit integer builtin support.
9295 (vec_vadduqm): Likewise.
9296 (vec_vaddecuq): Likewise.
9297 (vec_vaddeuqm): Likewise.
9298 (vec_vsubecuq): Likewise.
9299 (vec_vsubeuqm): Likewise.
9300 (vec_vsubcuq): Likewise.
9301 (vec_vsubuqm): Likewise.
9303 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
9304 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
9305 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
9306 128-bit integer add/subtract to ISA 2.07.
9308 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
9310 * config/arc/arc.c (arc_predicate_delay_insns):
9311 Fix third argument passed to conditionalize_nonjump.
9313 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
9315 * config/aarch64/aarch64-builtins.c
9316 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
9317 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
9318 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
9319 instead of __builtin_lfloor.
9320 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
9322 2014-03-12 Jakub Jelinek <jakub@redhat.com>
9324 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
9325 (tree_ssa_ifcombine_bb_1): New function.
9326 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
9327 is an empty forwarder block to then_bb or vice versa and then_bb
9328 and else_bb are effectively swapped.
9330 2014-03-12 Christian Bruel <christian.bruel@st.com>
9333 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
9334 REG_CFA_DEF_CFA note.
9335 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
9336 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
9338 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
9340 PR tree-optimization/60454
9341 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
9343 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9345 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
9346 Do not define target_cpu_default2 to generic.
9347 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
9348 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
9349 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
9351 2014-03-12 Jakub Jelinek <jakub@redhat.com>
9352 Marc Glisse <marc.glisse@inria.fr>
9354 PR tree-optimization/60502
9355 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
9356 instead of build_low_bits_mask.
9358 2014-03-12 Jakub Jelinek <jakub@redhat.com>
9361 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
9362 if there are multiple uses, but op doesn't live on E edge.
9363 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
9364 clobber stmts before __builtin_unreachable.
9366 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
9368 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
9369 hard_frame_pointer_rtx.
9370 * cse.c (cse_insn): Remove volatile check.
9371 * cselib.c (cselib_process_insn): Likewise.
9372 * dse.c (scan_insn): Likewise.
9374 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
9376 * config/arc/arc.c (conditionalize_nonjump): New function,
9378 (arc_ifcvt): ... this.
9379 (arc_predicate_delay_insns): Use it.
9381 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
9383 * config/arc/predicates.md (extend_operand): During/after reload,
9384 allow const_int_operand.
9385 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
9386 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
9387 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
9389 (umulsi3_highpart_i): Likewise.
9391 2014-03-11 Richard Biener <rguenther@suse.de>
9393 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
9394 Add asserts to guard possible wrong-code bugs.
9396 2014-03-11 Richard Biener <rguenther@suse.de>
9398 PR tree-optimization/60429
9399 PR tree-optimization/60485
9400 * tree-ssa-structalias.c (set_union_with_increment): Properly
9401 take into account all fields that overlap the shifted vars.
9402 (do_sd_constraint): Likewise.
9403 (do_ds_constraint): Likewise.
9404 (get_constraint_for_ptr_offset): Likewise.
9406 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
9408 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
9409 (nios2_compute_frame_layout):
9410 Add calculation of cfun->machine->fp_save_offset.
9411 (nios2_expand_prologue): Correct setting of frame pointer register
9413 (nios2_expand_epilogue): Update recovery of stack pointer from
9414 frame pointer accordingly.
9415 (nios2_initial_elimination_offset): Update calculation of offset
9416 for eliminating to HARD_FRAME_POINTER_REGNUM.
9418 2014-03-10 Jakub Jelinek <jakub@redhat.com>
9421 * ipa.c (symtab_remove_unreachable_nodes): Don't call
9422 cgraph_get_create_node on VAR_DECLs.
9424 2014-03-10 Richard Biener <rguenther@suse.de>
9427 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
9429 2014-03-08 Douglas B Rupp <rupp@gnat.com>
9431 * config/vms/vms.opt (vms_float_format): New variable.
9433 2014-03-08 Tobias Burnus <burnus@net-b.de>
9435 * doc/invoke.texi (-fcilkplus): Update implementation status.
9437 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
9438 Richard Biener <rguenther@suse.de>
9440 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
9441 consistently accross all TUs.
9442 (run_gcc): Enable -fshort-double automatically at link at link-time
9443 and disallow override.
9445 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
9448 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
9449 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
9450 if they can't be used.
9452 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9454 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
9455 for Solaris 11/x86 ld.
9456 * configure: Regenerate.
9458 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9460 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
9461 (LIB_TLS_SPEC): Save as ld_tls_libs.
9462 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
9463 (HAVE_AS_IX86_TLSLDM): New test.
9464 * configure, config.in: Regenerate.
9465 * config/i386/i386.c (legitimize_tls_address): Fall back to
9466 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
9467 cannot support TLS_MODEL_LOCAL_DYNAMIC.
9468 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
9469 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
9471 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
9473 * common.opt (fira-loop-pressure): Mark as optimization.
9475 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
9477 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
9478 an OpenMP mappable type.
9480 2014-03-06 Matthias Klose <doko@ubuntu.com>
9482 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
9483 MULTILIB_OSDIRNAMES is not defined.
9485 2014-03-06 Jakub Jelinek <jakub@redhat.com>
9486 Meador Inge <meadori@codesourcery.com>
9489 * config/arm/arm.c (arm_tls_symbol_p): Remove.
9490 (arm_legitimize_address): Call legitimize_tls_address for any
9491 arm_tls_referenced_p expression, handle constant addend. Call it
9492 before testing for !TARGET_ARM.
9493 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
9495 2014-03-06 Richard Biener <rguenther@suse.de>
9501 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
9503 * tree-streamer.c (record_common_node): Assert we don't record
9504 nodes with type double.
9505 (preload_common_node): Skip type double, complex double and double
9506 pointer since it is now frontend dependent due to fshort-double option.
9508 2014-03-06 Richard Biener <rguenther@suse.de>
9510 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
9511 or -fno-lto is specified and the linker has full plugin support.
9512 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
9513 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
9514 * lto-wrapper.c (merge_and_complain): Merge compile-time
9515 optimization levels.
9516 (run_gcc): And pass it through to the link options.
9518 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
9522 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9524 * cselib.c (remove_useless_values): Skip to avoid quadratic
9525 behavior if the condition moved from...
9526 (cselib_process_insn): ... here holds.
9528 2014-03-05 Jakub Jelinek <jakub@redhat.com>
9531 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
9532 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
9535 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
9536 (TM_H): Add x86-tune.def.
9538 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9540 * config/aarch64/aarch64.c (generic_tunings):
9541 Use cortexa57_extra_costs.
9543 2014-03-05 Jakub Jelinek <jakub@redhat.com>
9546 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
9547 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
9548 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
9551 2014-03-04 Heiher <r@hev.cc>
9553 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
9554 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
9556 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
9558 * config/i386/predicates.md (const2356_operand): Change to ...
9559 (const2367_operand): ... this.
9560 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
9562 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9563 (*avx512pf_scatterpf<mode>sf): Ditto.
9564 (avx512pf_scatterpf<mode>df): Ditto.
9565 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9566 (*avx512pf_scatterpf<mode>df): Ditto.
9567 * config/i386/i386.c (ix86_expand_builtin): Update
9568 incorrect hint operand error message.
9570 2014-03-04 Richard Biener <rguenther@suse.de>
9572 * lto-section-in.c (lto_get_section_data): Fix const cast.
9574 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
9576 * tree-streamer.c (record_common_node): Assert we don't record
9577 nodes with type double.
9578 (preload_common_node): Skip type double, complex double and double
9579 pointer since it is now frontend dependent due to fshort-double option.
9581 2014-03-04 Richard Biener <rguenther@suse.de>
9584 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
9585 (lto_input_toplevel_asms): Likewise.
9586 * lto-section-in.c (lto_get_section_data): Instead do it here
9589 2014-03-04 Richard Biener <rguenther@suse.de>
9591 PR tree-optimization/60382
9592 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
9593 dead PHIs a reduction.
9595 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
9597 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
9599 (_mm_prefetch): Move out of GCC target("sse") pragma.
9600 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
9601 GCC target("prfchw") pragma.
9602 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
9604 * config/i386/i386.c (ix86_option_override_internal): Enable
9605 -mprfchw with -mprefetchwt1.
9607 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
9609 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
9612 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
9614 * opts.h (CL_PCH_IGNORE): Define.
9615 * targhooks.c (option_affects_pch_p):
9616 Return false for options that have CL_PCH_IGNORE set.
9617 * opt-functions.awk: Process PchIgnore.
9618 * doc/options.texi: Document PchIgnore.
9620 * config/arc/arc.opt (misize): Add PchIgnore property.
9622 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9624 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
9625 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
9626 constraint on constants to permit them being loaded into
9627 GENERAL_REGS or BASE_REGS.
9629 2014-03-03 Nick Clifton <nickc@redhat.com>
9631 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
9632 anti-cacnonical alternatives.
9633 (negandhi3_real): New pattern.
9634 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
9636 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
9638 * config/avr/avr-mcus.def: Remove atxmega16x1.
9639 * config/avr/avr-tables.opt: Regenerate.
9640 * config/avr/t-multilib: Regenerate.
9641 * doc/avr-mmcu.texi: Regenerate.
9643 2014-03-03 Tobias Grosser <tobias@grosser.es>
9644 Mircea Namolaru <mircea.namolaru@inria.fr>
9646 PR tree-optimization/58028
9647 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
9650 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9652 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
9653 not handled by recognizers.
9655 2014-03-03 Jakub Jelinek <jakub@redhat.com>
9658 * function.c (expand_function_end): Don't emit
9659 clobber_return_register sequence if clobber_after is a BARRIER.
9660 * cfgexpand.c (construct_exit_block): Append instructions before
9661 return_label to prev_bb.
9663 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9665 * config/rs6000/constraints.md: Document reserved use of "wc".
9667 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
9670 * ipa.c (function_and_variable_visibility): When dissolving comdat
9671 group, also set all symbols to local.
9673 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
9678 2013-12-14 Jan Hubicka <jh@suse.cz>
9680 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
9682 2014-03-02 Jon Beniston <jon@beniston.com>
9688 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
9689 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
9690 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
9691 (simple_return, *simple_return): New patterns
9692 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
9693 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
9695 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
9697 * dwarf2out.c (gen_subprogram_die): Tidy.
9699 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
9702 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
9703 (*mov_t_msb_neg_negc): ... this new insn.
9705 2014-02-28 Jason Merrill <jason@redhat.com>
9708 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
9711 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
9714 * dwarf2out.c (decltype_auto_die): New static.
9715 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
9716 (gen_type_die_with_usage): Handle 'decltype(auto)'.
9717 (is_cxx_auto): Likewise.
9719 2014-02-28 Ian Bolton <ian.bolton@arm.com>
9721 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
9722 we are not using general regs only.
9724 2014-02-28 Richard Biener <rguenther@suse.de>
9727 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
9728 previous fix and only allow to remove trivial pre-headers
9729 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
9730 (remove_forwarder_block): Properly update the latch of a loop.
9732 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9735 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
9736 (cselib_preserved_hash_table): New.
9737 (preserve_constants_and_equivs): Move preserved vals to it.
9738 (cselib_find_slot): Look it up first.
9739 (cselib_init): Initialize it.
9740 (cselib_finish): Release it.
9741 (dump_cselib_table): Dump it.
9743 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9746 * cselib.c (remove_useless_values): Skip to avoid quadratic
9747 behavior if the condition moved from...
9748 (cselib_process_insn): ... here holds.
9750 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9753 * var-tracking.c (vt_initialize): Apply the same condition to
9754 preserve the CFA base value.
9756 2014-02-28 Joey Ye <joey.ye@arm.com>
9759 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
9760 if reload in progress or completed.
9762 2014-02-28 Tobias Burnus <burnus@net-b.de>
9765 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
9768 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
9770 * doc/tm.texi.in (Condition Code Status): Update documention for
9771 relative locations of cc0-setter and cc0-user.
9773 2014-02-27 Jeff Law <law@redhat.com>
9775 PR rtl-optimization/52714
9776 * combine.c (try_combine): When splitting an unrecognized PARALLEL
9777 into two independent simple sets, if I3 is a jump, ensure the
9778 pattern we place into I3 is a (set (pc) ...).
9780 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
9781 Jeff Law <law@redhat.com>
9783 PR rtl-optimization/49847
9784 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
9785 are in different blocks.
9786 * doc/tm.texi (Condition Code Status): Update documention for
9787 relative locations of cc0-setter and cc0-user.
9789 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
9792 * lra.c (lra_emit_add): Check SUBREG too.
9794 2014-02-27 Andreas Schwab <schwab@suse.de>
9796 * config/m68k/m68k.c (m68k_option_override): Disable
9797 -flive-range-shrinkage for classic m68k.
9798 (m68k_override_options_after_change): Likewise.
9800 2014-02-27 Marek Polacek <polacek@redhat.com>
9803 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
9804 -Wmaybe-uninitialized.
9806 2014-02-27 Alan Modra <amodra@gmail.com>
9809 * reload1.c (emit_input_reload_insns): When reload_override_in,
9810 set old to rl->in_reg when rl->in_reg is a subreg.
9812 2014-02-26 Richard Biener <rguenther@suse.de>
9815 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
9817 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
9819 * common/config/i386/predicates.md (const1256_operand): Remove.
9820 (const2356_operand): New.
9821 (const_1_to_2_operand): Remove.
9822 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
9823 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9824 (*avx512pf_gatherpf<mode>sf): Ditto.
9825 (avx512pf_gatherpf<mode>df): Ditto.
9826 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9827 (*avx512pf_gatherpf<mode>df): Ditto.
9828 (avx512pf_scatterpf<mode>sf): Ditto.
9829 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9830 (*avx512pf_scatterpf<mode>sf): Ditto.
9831 (avx512pf_scatterpf<mode>df): Ditto.
9832 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9833 (*avx512pf_scatterpf<mode>df): Ditto.
9834 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
9836 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
9838 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
9839 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
9840 (_mm512_mask_testn_epi64_mask): Move to ...
9841 * config/i386/avx512cdintrin.h: Here.
9842 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
9843 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
9844 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
9845 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
9846 TARGET_AVX512F from TARGET_AVX512CD.
9848 2014-02-26 Richard Biener <rguenther@suse.de>
9851 * ipa.c (walk_polymorphic_call_targets): Properly guard
9852 call to inline_update_overall_summary.
9854 2014-02-26 Bin Cheng <bin.cheng@arm.com>
9857 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
9858 and latches only if requested. Fix latch if it is removed.
9859 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
9860 LOOPS_HAVE_PREHEADERS.
9862 2014-02-25 Andrew Pinski <apinski@cavium.com>
9864 * builtins.c (expand_builtin_thread_pointer): Create a new target
9865 when the target is NULL.
9867 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
9869 PR rtl-optimization/60317
9870 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
9871 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
9872 * lra-assigns.c: Include params.h.
9873 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
9874 other reload pseudos considerations.
9876 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9878 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
9879 to use canonical form for nor<mode>3.
9881 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9884 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
9887 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
9889 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
9890 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
9891 (ix86_handle_option): Handle OPT_mprefetchwt1.
9892 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
9893 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
9895 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9896 OPTION_MASK_ISA_PREFETCHWT1.
9897 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
9898 (PTA_PREFETCHWT1): New.
9899 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
9900 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
9901 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
9902 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
9903 (*prefetch_avx512pf_<mode>_: Change into ...
9904 (*prefetch_prefetchwt1_<mode>: This.
9905 * config/i386/i386.opt (mprefetchwt1): New.
9906 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
9907 (_mm_prefetch): Handle intent to write.
9908 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
9910 2014-02-25 Richard Biener <rguenther@suse.de>
9913 * emit-rtl.c (mem_attrs_htab): Remove.
9914 (mem_attrs_htab_hash): Likewise.
9915 (mem_attrs_htab_eq): Likewise.
9916 (set_mem_attrs): Always allocate new mem-attrs when something changed.
9917 (init_emit_once): Do not allocate mem_attrs_htab.
9919 2014-02-25 Richard Biener <rguenther@suse.de>
9922 * lto-opts.c (lto_write_options): Output non-explicit conservative
9923 -fwrapv, -fno-trapv and -fno-strict-overflow.
9924 * lto-wrapper.c (merge_and_complain): Handle merging those options.
9925 (run_gcc): And pass them through.
9927 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
9929 * sel-sched.c (calculate_new_fences): New parameter ptime.
9930 Calculate it as a maximum over all fence cycles.
9931 (sel_sched_region_2): Adjust the call to calculate_new_fences.
9932 Print the final schedule timing when sched_verbose.
9934 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
9936 PR rtl-optimization/60292
9937 * sel-sched.c (fill_vec_av_set): Do not reset target availability
9938 bit fot the fence instruction.
9940 2014-02-24 Alangi Derick <alangiderick@gmail.com>
9942 * calls.h: Fix typo in comment.
9944 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
9946 * config/pa/pa.c (pa_output_move_double): Don't valididate when
9947 adjusting offsetable addresses.
9949 2014-02-24 Guozhi Wei <carrot@google.com>
9951 * sparseset.h (sparseset_pop): Fix the wrong index.
9953 2014-02-24 Walter Lee <walt@tilera.com>
9955 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
9956 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
9958 * common/config/tilegx/tilegx-common.c
9959 (TARGET_DEFAULT_TARGET_FLAGS): Define.
9960 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
9962 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
9963 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
9964 (tilegx_gimplify_va_arg_expr): Handle big endian.
9965 (tilegx_expand_unaligned_load): Ditto.
9966 (tilegx_expand_unaligned_store): Ditto.
9967 (TARGET_RETURN_IN_MSB): New.
9968 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
9969 (TARGET_ENDIAN_DEFAULT): New.
9970 (TARGET_BIG_ENDIAN): Handle big endian.
9971 (BYTES_BIG_ENDIAN): Ditto.
9972 (WORDS_BIG_ENDIAN): Ditto.
9973 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
9976 * config/tilegx/tilegx.md (extv): Handle big endian.
9978 (insn_st<n>): Ditto.
9979 (insn_st<n>_add<bitsuffix>): Ditto.
9980 (insn_stnt<n>): Ditto.
9981 (insn_stnt<n>_add<bitsuffix>):Ditto.
9982 (vec_interleave_highv8qi): Handle big endian.
9983 (vec_interleave_highv8qi_be): New.
9984 (vec_interleave_highv8qi_le): New.
9985 (insn_v1int_h): Handle big endian.
9986 (vec_interleave_lowv8qi): Handle big endian.
9987 (vec_interleave_lowv8qi_be): New.
9988 (vec_interleave_lowv8qi_le): New.
9989 (insn_v1int_l): Handle big endian.
9990 (vec_interleave_highv4hi): Handle big endian.
9991 (vec_interleave_highv4hi_be): New.
9992 (vec_interleave_highv4hi_le): New.
9993 (insn_v2int_h): Handle big endian.
9994 (vec_interleave_lowv4hi): Handle big endian.
9995 (vec_interleave_lowv4hi_be): New.
9996 (vec_interleave_lowv4hi_le): New.
9997 (insn_v2int_l): Handle big endian.
9998 (vec_interleave_highv2si): Handle big endian.
9999 (vec_interleave_highv2si_be): New.
10000 (vec_interleave_highv2si_le): New.
10001 (insn_v4int_h): Handle big endian.
10002 (vec_interleave_lowv2si): Handle big endian.
10003 (vec_interleave_lowv2si_be): New.
10004 (vec_interleave_lowv2si_le): New.
10005 (insn_v4int_l): Handle big endian.
10006 * config/tilegx/tilegx.opt (mbig-endian): New option.
10007 (mlittle-endian): New option.
10008 * doc/install.texi: Document tilegxbe-linux.
10009 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
10011 2014-02-24 Martin Jambor <mjambor@suse.cz>
10014 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
10015 there are no parameter descriptors.
10017 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
10019 PR rtl-optimization/60268
10020 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
10021 initialization to ...
10022 (sched_rgn_init): ... here.
10023 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
10025 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
10027 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
10030 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
10032 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
10035 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
10037 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
10038 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
10040 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
10042 * config/microblaze/predicates.md: Add cmp_op predicate.
10043 * config/microblaze/microblaze.md: Add branch_compare instruction
10044 which uses cmp_op predicate and emits cmp insn before branch.
10045 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
10046 to microblaze_expand_conditional_branch and consolidate logic.
10047 (microblaze_expand_conditional_branch): emit branch_compare
10048 insn instead of handling cmp op separate from branch insn.
10050 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10052 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
10055 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10057 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
10058 define_insn with define_expand and new define_insn
10059 *altivec_lve<VI_char>x_internal.
10060 (altivec_stve<VI_char>x): Replace define_insn with define_expand
10061 and new define_insn *altivec_stve<VI_char>x_internal.
10062 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
10064 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
10066 (altivec_expand_stvex_be): New function.
10068 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
10070 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
10071 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
10072 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
10073 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
10075 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
10078 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
10079 instead of emit_move_insn.
10081 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10083 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
10084 vspltw with vsldoi.
10085 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
10086 gen_altivec_vsumsws.
10088 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10090 * config/rs6000/altivec.md (altivec_lvxl): Rename as
10091 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
10092 (altivec_lvxl_<mode>): New define_expand incorporating
10093 -maltivec=be semantics where needed.
10094 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
10095 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
10096 semantics where needed.
10097 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
10098 (altivec_stvx_<mode>): New define_expand incorporating
10099 -maltivec=be semantics where needed.
10100 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
10101 VM2 iterator instead of V4SI.
10102 (altivec_stvxl_<mode>): New define_expand incorporating
10103 -maltivec=be semantics where needed.
10104 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
10105 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
10106 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
10107 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
10108 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
10109 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
10110 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
10111 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
10112 ALTIVEC_BUILTIN_STVXL.
10113 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
10114 (altivec_expand_stvx_be): Likewise.
10115 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
10116 (altivec_expand_lvx_be): Likewise.
10117 (altivec_expand_stvx_be): Likewise.
10118 (altivec_expand_builtin): Add cases for
10119 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
10120 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
10121 (altivec_init_builtins): Add definitions for
10122 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
10123 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
10125 2014-02-21 Catherine Moore <clm@codesourcery.com>
10127 * doc/invoke.texi (mvirt, mno-virt): Document.
10128 * config/mips/mips.opt (mvirt): New option.
10129 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
10131 2014-02-21 Richard Biener <rguenther@suse.de>
10133 PR tree-optimization/60276
10134 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
10135 (STMT_VINFO_MIN_NEG_DIST): New macro.
10136 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
10137 STMT_VINFO_MIN_NEG_DIST.
10138 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
10139 made for negative dependence distances still hold.
10141 2014-02-21 Richard Biener <rguenther@suse.de>
10143 PR middle-end/60291
10144 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
10145 DECL_INITIAL for globals not in the current function context.
10147 2014-02-21 Jakub Jelinek <jakub@redhat.com>
10149 PR tree-optimization/56490
10150 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
10151 * tree-ssa-uninit.c: Include params.h.
10152 (compute_control_dep_chain): Add num_calls argument, return false
10153 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
10154 num_calls to recursive call.
10155 (find_predicates): Change dep_chain into normal array,
10156 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
10157 variable and adjust compute_control_dep_chain caller.
10158 (find_def_preds): Likewise.
10160 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
10162 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
10163 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
10165 2014-02-21 Nick Clifton <nickc@redhat.com>
10167 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
10168 (pushhi1): Likewise.
10169 (popqi1): Add mode to pre_dec.
10170 (pophi1): Likewise.
10172 2014-02-21 Jakub Jelinek <jakub@redhat.com>
10174 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
10175 mode for mask of V8SFmode permutation.
10177 2014-02-20 Richard Henderson <rth@redhat.com>
10180 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
10181 a new pseudo for OLDVAL.
10183 2014-02-20 Jakub Jelinek <jakub@redhat.com>
10186 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
10187 gen_reg_rtx if d->testing_p.
10188 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
10189 if d->testing_p and we will certainly return true.
10190 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
10193 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
10195 * emit-rtl.c (gen_reg_rtx): Assert that
10196 crtl->emit.regno_pointer_align_length is non-zero.
10198 2014-02-20 Richard Henderson <rth@redhat.com>
10201 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
10202 on failure the store back into EXPECT.
10204 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
10205 Sandra Loosemore <sandra@codesourcery.com>
10207 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
10208 * config/nios2/nios2.c (nios2_function_profiler): Add
10209 -fPIC (flag_pic == 2) support.
10210 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
10211 (nios2_large_offset_p): New function.
10212 (nios2_unspec_reloc_p): Move up position, update to use
10213 nios2_large_offset_p.
10214 (nios2_unspec_address): Remove function.
10215 (nios2_unspec_offset): New function.
10216 (nios2_large_got_address): New function.
10217 (nios2_got_address): Add large offset support.
10218 (nios2_legitimize_tls_address): Update usage of removed and new
10220 (nios2_symbol_binds_local_p): New function.
10221 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
10222 (nios2_legitimize_address): Update to use nios2_large_offset_p.
10223 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
10224 (nios2_print_operand): Merge H/L processing, add hiadj/lo
10225 processing for (const (unspec ...)).
10226 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
10228 2014-02-20 Richard Biener <rguenther@suse.de>
10230 * tree-cfg.c (replace_uses_by): Mark altered BBs before
10231 doing the substitution.
10232 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
10234 2014-02-20 Martin Jambor <mjambor@suse.cz>
10237 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
10238 info when checking whether lattices are bottom.
10240 2014-02-20 Richard Biener <rguenther@suse.de>
10242 PR middle-end/60221
10243 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
10246 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
10249 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
10250 parameter specifying the scaling.
10251 (inline_call): Update.
10252 (want_inline_recursively): Guard division by zero.
10253 (recursive_inlining): Update.
10254 * ipa-inline.h (clone_inlined_nodes): Update.
10256 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
10259 * config/i386/i386.c (classify_argument): Pass structures of size
10260 64 bytes or less in register.
10262 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
10263 Kirill Yukhin <kirill.yukhin@intel.com>
10265 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
10266 (_mm_rcp28_round_ss): Ditto.
10267 (_mm_rsqrt28_round_sd): Ditto.
10268 (_mm_rsqrt28_round_ss): Ditto.
10269 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
10270 (_mm_rcp14_round_ss): Ditto.
10271 (_mm_rsqrt14_round_sd): Ditto.
10272 (_mm_rsqrt14_round_ss): Ditto.
10273 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
10274 the first input operand, get rid of match_dup.
10275 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
10277 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
10279 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
10280 operand as the first input operand, set type attribute.
10281 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
10282 Set type attribute.
10283 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
10284 operand as the first input operand, set type attribute.
10286 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10288 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
10291 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
10294 * config/i386/i386.c (construct_container): Remove TFmode check
10295 for X86_64_INTEGER_CLASS.
10297 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
10300 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
10301 only when -Wpsabi is enabled.
10303 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
10306 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
10307 passing arrays in registers are the same as for structs, so remove the
10308 special case for them.
10310 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
10312 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
10313 destination type, extract only the valid bits if the source type is not
10314 integral and has a different mode.
10316 2014-02-19 Richard Biener <rguenther@suse.de>
10319 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
10322 2014-02-19 Richard Biener <rguenther@suse.de>
10325 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
10326 (ipa_modify_call_arguments): Emit an argument load explicitely and
10327 preserve virtual SSA form there and for the replacement call.
10328 Do not update SSA form nor free dominance info.
10330 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
10332 * ipa.c (function_and_variable_visibility): Also clear WEAK
10333 flag when disolving COMDAT_GROUP.
10335 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
10337 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
10338 * ipa-prop.c (ipa_set_jf_known_type): Return early when
10339 not devirtualizing.
10340 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
10341 do more sanity checks.
10342 (detect_type_change): Return true when giving up early.
10343 (compute_complex_assign_jump_func): Fix type parameter of
10344 ipa_set_ancestor_jf.
10345 (compute_complex_ancestor_jump_func): Likewise.
10346 (update_jump_functions_after_inlining): Fix updating of
10348 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
10350 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
10352 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
10353 inline clones when edge disappears.
10355 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
10358 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
10359 Split 64-bit moves into 2 patterns. Do not allow the use of
10360 direct move for TDmode in little endian, since the decimal value
10361 has little endian bytes within a word, but the 64-bit pieces are
10362 ordered in a big endian fashion, and normal subreg's of TDmode are
10364 (mov<mode>_64bit_dm): Likewise.
10365 (movtd_64bit_nodm): Likewise.
10367 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
10369 PR tree-optimization/60174
10370 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
10371 statement of an SSA_NAME that occurs in an abnormal PHI node.
10373 2014-02-18 Jakub Jelinek <jakub@redhat.com>
10376 * final.c (SEEN_BB): Remove.
10377 (SEEN_NOTE, SEEN_EMITTED): Renumber.
10378 (final_scan_insn): Don't force_source_line on second
10379 NOTE_INSN_BASIC_BLOCK.
10381 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
10384 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
10385 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
10386 (type_natural_mode): Warn ABI change when %zmm register is not
10387 available for AVX512F vector value passing.
10389 2014-02-18 Kai Tietz <ktietz@redhat.com>
10392 * config/i386/i386.c (ix86_expand_prologue): Use value in
10393 rax register as displacement when restoring %r10 or %rax.
10394 Fix wrong offset when restoring both registers.
10396 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
10398 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
10399 assertion with conditional return.
10401 2014-02-18 Jakub Jelinek <jakub@redhat.com>
10402 Uros Bizjak <ubizjak@gmail.com>
10405 * config/i386/driver-i386.c (host_detect_local_cpu): If
10406 YMM state is not saved by the OS, also clear has_f16c. Move
10407 CPUID 0x80000001 handling before YMM state saving checking.
10409 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
10411 PR rtl-optimization/58960
10412 * haifa-sched.c (alloc_global_sched_pressure_data): New,
10413 factored out from ...
10414 (sched_init): ... here.
10415 (free_global_sched_pressure_data): New, factored out from ...
10416 (sched_finish): ... here.
10417 * sched-int.h (free_global_sched_pressure_data): Declare.
10418 * sched-rgn.c (nr_regions_initial): New static global.
10419 (haifa_find_rgns): Initialize it.
10420 (schedule_region): Disable sched-pressure for the newly
10423 2014-02-17 Richard Biener <rguenther@suse.de>
10425 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
10426 release SSA defs of pattern stmts.
10428 2014-02-17 Richard Biener <rguenther@suse.de>
10430 * tree-inline.c (expand_call_inline): Release the virtual
10431 operand defined by the call we are about to inline.
10433 2014-02-17 Richard Biener <rguenther@suse.de>
10435 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
10437 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
10438 Ilya Tocar <ilya.tocar@intel.com>
10440 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
10441 arguments order in builtin.
10442 (_mm512_permutexvar_epi64): Ditto.
10443 (_mm512_mask_permutexvar_epi64): Ditto
10444 (_mm512_maskz_permutexvar_epi32): Ditto
10445 (_mm512_permutexvar_epi32): Ditto
10446 (_mm512_mask_permutexvar_epi32): Ditto
10448 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10450 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
10451 (p8_vmrgow): Likewise.
10453 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10455 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
10458 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
10461 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
10462 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
10463 into 64-bit and 32-bit moves. On 64-bit moves, add support for
10464 using direct move instructions on ISA 2.07. Also adjust
10465 instruction length for 64-bit.
10466 (mov<mode>_64bit, TFmode/TDmode): Likewise.
10467 (mov<mode>_32bit, TFmode/TDmode): Likewise.
10469 2014-02-15 Alan Modra <amodra@gmail.com>
10473 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
10474 find_replacement on parts of insn rtl that might be reloaded.
10476 2014-02-15 Richard Biener <rguenther@suse.de>
10478 PR tree-optimization/60183
10479 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
10480 (tree_ssa_phiprop): Calculate and free post-dominators.
10482 2014-02-14 Jeff Law <law@redhat.com>
10484 PR rtl-optimization/60131
10485 * ree.c (get_extended_src_reg): New function.
10486 (combine_reaching_defs): Use it rather than assuming location of REG.
10487 (find_and_remove_re): Verify first operand of extension is
10488 a REG before adding the insns to the copy list.
10490 2014-02-14 Roland McGrath <mcgrathr@google.com>
10492 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
10493 * configure: Regenerated.
10494 * config.in: Regenerated.
10495 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
10496 instead of ASM_SHORT.
10498 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
10499 Richard Earnshaw <rearnsha@arm.com>
10501 PR rtl-optimization/59535
10502 * lra-constraints.c (process_alt_operands): Encourage alternative
10503 when unassigned pseudo class is superset of the alternative class.
10504 (inherit_reload_reg): Don't inherit when optimizing for code size.
10505 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
10506 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
10507 modes not less than 4 for Thumb1.
10509 2014-02-14 Kyle McMartin <kyle@redhat.com>
10512 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
10514 2014-02-14 Richard Biener <rguenther@suse.de>
10516 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
10517 (get_frame_arg): Drop the assert with langhook types_compatible_p.
10518 Do not strip INDIRECT_REFs.
10520 2014-02-14 Richard Biener <rguenther@suse.de>
10523 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
10524 DECL_FUNCTION_SPECIFIC_TARGET.
10525 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
10526 * tree-streamer-out.c (pack_ts_target_option): Remove.
10527 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
10528 (write_ts_function_decl_tree_pointers): Do not stream
10529 DECL_FUNCTION_SPECIFIC_TARGET.
10530 * tree-streamer-in.c (unpack_ts_target_option): Remove.
10531 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
10532 (lto_input_ts_function_decl_tree_pointers): Do not stream
10533 DECL_FUNCTION_SPECIFIC_TARGET.
10535 2014-02-14 Jakub Jelinek <jakub@redhat.com>
10537 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
10538 (get_initial_def_for_induction, vectorizable_induction): Ignore
10539 debug stmts when looking for exit_phi.
10540 (vectorizable_live_operation): Fix up condition.
10542 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
10544 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
10545 nreverse() because it changes the content of original tree list.
10547 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
10549 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
10550 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
10552 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
10554 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
10555 GNU coding standards.
10557 2014-02-13 Jakub Jelinek <jakub@redhat.com>
10560 * dwarf2out.c (gen_subprogram_die): Don't call
10561 add_calling_convention_attribute if subr_die is old_die.
10563 2014-02-13 Sharad Singhai <singhai@google.com>
10565 * doc/optinfo.texi: Fix order of nodes.
10567 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
10569 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
10570 operands[2], not operands[3].
10572 2014-02-13 Richard Biener <rguenther@suse.de>
10575 * doc/install.texi (ISL): Update recommended version to 0.12.2,
10576 mention the possibility of an in-tree build.
10577 (CLooG): Update recommended version to 0.18.1, mention the
10578 possibility of an in-tree build and clarify that the ISL
10579 bundled with CLooG does not work.
10581 2014-02-13 Jakub Jelinek <jakub@redhat.com>
10584 * expr.c (compress_float_constant): If x is a hard register,
10585 extend into a pseudo and then move to x.
10587 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
10589 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
10590 caused by bad second argument to warning_at() with -mhotpatch and
10591 nested functions (e.g. with gfortran).
10593 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
10595 * opts.c (option_name): Remove "enabled by default" rider.
10597 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
10599 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
10601 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
10602 Uros Bizjak <ubizjak@gmail.com>
10605 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
10606 * configure: Regenerated.
10608 2014-02-12 Richard Biener <rguenther@suse.de>
10610 * vec.c (vec_prefix::calculate_allocation): Move as
10611 inline variant to vec.h.
10612 (vec_prefix::calculate_allocation_1): New out-of-line version.
10613 * vec.h (vec_prefix::calculate_allocation_1): Declare.
10614 (vec_prefix::m_has_auto_buf): Rename to ...
10615 (vec_prefix::m_using_auto_storage): ... this.
10616 (vec_prefix::calculate_allocation): Inline the easy cases
10617 and dispatch to calculate_allocation_1 which doesn't need the
10619 (va_heap::reserve): Use gcc_checking_assert.
10620 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
10621 m_using_auto_storage.
10622 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
10624 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
10625 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
10626 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
10628 2014-02-12 Richard Biener <rguenther@suse.de>
10630 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
10631 when we found a dependence.
10633 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
10635 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
10637 (maybe_fold_stmt): ... into this new function.
10638 * omp-low.c (lower_omp): Update comment.
10640 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
10643 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
10646 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
10648 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
10649 identifiers in comments.
10650 (cortexa53_extra_costs): Likewise.
10651 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
10652 (cortexa7_extra_costs): Likewise.
10653 (cortexa12_extra_costs): Likewise.
10654 (cortexa15_extra_costs): Likewise.
10655 (v7m_extra_costs): Likewise.
10657 2014-02-12 Richard Biener <rguenther@suse.de>
10659 PR middle-end/60092
10660 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
10661 of posix_memalign being successful.
10662 (lower_stmt): Restrict lowering of posix_memalign to when
10663 -ftree-bit-ccp is enabled.
10665 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
10667 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
10669 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
10671 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
10673 PR rtl-optimization/60116
10674 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
10675 other_insn once the combination has been validated.
10677 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
10680 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
10682 * ipa-devirt.c: Include demangle.h
10683 (odr_violation_reported): New static variable.
10684 (add_type_duplicate): Update odr_violations.
10685 (maybe_record_node): Add completep parameter; update it.
10686 (record_target_from_binfo): Add COMPLETEP parameter;
10687 update it as needed.
10688 (possible_polymorphic_call_targets_1): Likewise.
10689 (struct polymorphic_call_target_d): Add nonconstruction_targets;
10690 rename FINAL to COMPLETE.
10691 (record_targets_from_bases): Sanity check we found the binfo;
10692 fix COMPLETEP updating.
10693 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
10694 parameter, fix computing of COMPLETEP.
10695 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
10696 at LTO time do demangling.
10697 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
10698 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
10700 (gimple_get_virt_method_for_binfo): Likewise.
10701 * gimple-fold.h (gimple_get_virt_method_for_binfo,
10702 gimple_get_virt_method_for_vtable): Update prototypes.
10704 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
10707 * genautomata.c (add_presence_absence): Fix typo with
10708 {final_}presence_list.
10710 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
10713 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
10714 for VSX/Altivec vectors that land in GPR registers.
10716 2014-02-11 Richard Henderson <rth@redhat.com>
10717 Jakub Jelinek <jakub@redhat.com>
10720 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
10721 around drhs if type conversion to lacc->type is not useless.
10723 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10725 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
10727 (cortex-a57.cortex-a53): Likewise.
10728 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
10730 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10732 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
10735 2014-02-11 Renlin Li <Renlin.Li@arm.com>
10737 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
10738 add_options_for_arm_vfp3.
10740 2014-02-11 Jeff Law <law@redhat.com>
10742 PR middle-end/54041
10743 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
10744 object with an undesirable mode.
10746 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10749 * config/i386/sol2-9.h: New file.
10750 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
10751 *-*-solaris2.9*): Use it.
10753 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
10755 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
10756 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
10758 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
10760 * config/microblaze/microblaze.c: Extend mcpu version format
10762 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
10764 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
10766 2014-02-10 Richard Henderson <rth@redhat.com>
10769 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
10770 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
10771 ms-abi vs -mno-accumulate-outgoing-args.
10772 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
10773 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
10776 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
10778 PR middle-end/60080
10779 * cfgexpand.c (expand_asm_operands): Attach source location to
10780 ASM_INPUT rtx objects.
10781 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
10783 2014-02-10 Nick Clifton <nickc@redhat.com>
10785 * config/mn10300/mn10300.c (popcount): New function.
10786 (mn10300_expand_prologue): Include saved registers in stack usage
10789 2014-02-10 Jeff Law <law@redhat.com>
10791 PR middle-end/52306
10792 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
10793 when changing the SET_DEST of a prior insn to avoid an input reload.
10795 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
10797 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
10798 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
10799 -mcall-openbsd, or -mcall-linux.
10800 (CC1_ENDIAN_BIG_SPEC): Remove.
10801 (CC1_ENDIAN_LITTLE_SPEC): Remove.
10802 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
10803 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
10804 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
10805 and %cc1_endian_default.
10806 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
10808 2014-02-10 Richard Biener <rguenther@suse.de>
10810 PR tree-optimization/60115
10811 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
10812 MEM_REF handling. Properly verify that the accesses are not
10813 out of the objects bound.
10815 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10817 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
10820 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
10822 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
10823 proper constants and fix formatting.
10824 (possible_polymorphic_call_targets): Fix formatting.
10826 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
10827 Ilya Tocar <ilya.tocar@intel.com>
10829 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
10830 (_mm512_loadu_epi32): Renamed into...
10831 (_mm512_loadu_si512): This.
10832 (_mm512_storeu_epi32): Renamed into...
10833 (_mm512_storeu_si512): This.
10834 (_mm512_maskz_ceil_ps): Removed.
10835 (_mm512_maskz_ceil_pd): Ditto.
10836 (_mm512_maskz_floor_ps): Ditto.
10837 (_mm512_maskz_floor_pd): Ditto.
10838 (_mm512_floor_round_ps): Ditto.
10839 (_mm512_floor_round_pd): Ditto.
10840 (_mm512_ceil_round_ps): Ditto.
10841 (_mm512_ceil_round_pd): Ditto.
10842 (_mm512_mask_floor_round_ps): Ditto.
10843 (_mm512_mask_floor_round_pd): Ditto.
10844 (_mm512_mask_ceil_round_ps): Ditto.
10845 (_mm512_mask_ceil_round_pd): Ditto.
10846 (_mm512_maskz_floor_round_ps): Ditto.
10847 (_mm512_maskz_floor_round_pd): Ditto.
10848 (_mm512_maskz_ceil_round_ps): Ditto.
10849 (_mm512_maskz_ceil_round_pd): Ditto.
10850 (_mm512_expand_pd): Ditto.
10851 (_mm512_expand_ps): Ditto.
10852 * config/i386/i386.c (ix86_builtins): Remove
10853 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
10854 (bdesc_args): Ditto.
10855 * config/i386/predicates.md (const1256_operand): New.
10856 (const_1_to_2_operand): Ditto.
10857 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
10858 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
10859 (*avx512pf_gatherpf<mode>sf): Ditto.
10860 (avx512pf_gatherpf<mode>df): Ditto.
10861 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10862 (*avx512pf_gatherpf<mode>df): Ditto.
10863 (avx512pf_scatterpf<mode>sf): Ditto.
10864 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
10865 (*avx512pf_scatterpf<mode>sf): Ditto.
10866 (avx512pf_scatterpf<mode>df): Ditto.
10867 (*avx512pf_scatterpf<mode>df_mask): Ditto.
10868 (*avx512pf_scatterpf<mode>df): Ditto.
10869 (avx512f_expand<mode>): Removed.
10870 (<shift_insn><mode>3<mask_name>): Change predicate type.
10872 2014-02-08 Jakub Jelinek <jakub@redhat.com>
10874 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
10875 not at the end of datarefs vector use ordered_remove to avoid
10876 reordering datarefs vector.
10879 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
10880 mark local addressable non-static vars as GOVD_PRIVATE
10881 instead of GOVD_LOCAL.
10882 * omp-low.c (lower_omp_for): Move gimple_bind_vars
10883 and BLOCK_VARS of gimple_bind_block to new_stmt rather
10886 PR middle-end/60092
10887 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
10888 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
10889 assume_aligned or alloc_align attributes.
10890 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
10891 arguments. Handle also assume_aligned and alloc_align attributes.
10892 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
10893 calls to functions with assume_aligned or alloc_align attributes.
10894 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
10896 2014-02-08 Terry Guo <terry.guo@arm.com>
10898 * doc/invoke.texi: Document ARM -march=armv7e-m.
10900 2014-02-08 Jakub Jelinek <jakub@redhat.com>
10902 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
10903 flag on __cilkrts_rethrow builtin.
10906 * ipa-cp.c (determine_versionability): Fail at -O0
10907 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
10908 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
10911 2014-02-04 Jakub Jelinek <jakub@redhat.com>
10914 * tree-inline.c (copy_forbidden): Fail for
10915 __attribute__((optimize (0))) functions.
10917 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
10919 * varpool.c: Include pointer-set.h.
10920 (varpool_remove_unreferenced_decls): Variables in other partitions
10921 will not be output; be however careful to not lose information
10922 about partitioning.
10924 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
10926 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
10927 lookup in the vtable constructor.
10929 2014-02-07 Jeff Law <law@redhat.com>
10932 * config/m68k/m68k.md (ashldi_extsi): Turn into a
10933 define_insn_and_split.
10935 * ipa-inline.c (inline_small_functions): Fix typos.
10937 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10939 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
10940 (s390_can_use_return_insn): Declare.
10941 * config/s390/s390.h (EPILOGUE_USES): Define.
10942 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
10944 (s390_chunkify_start): Handle return JUMP_LABELs.
10945 (s390_early_mach): Emit a main_pool instruction on the entry edge.
10946 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
10947 (s390_can_use_return_insn): New functions.
10948 (s390_fix_long_loop_prediction): Handle conditional returns.
10949 (TARGET_SET_UP_BY_PROLOGUE): Define.
10950 * config/s390/s390.md (ANY_RETURN): New code iterator.
10951 (*creturn, *csimple_return, return, simple_return): New patterns.
10953 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10955 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
10956 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
10957 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
10958 REG_CFA_RESTORE list when deciding not to restore a register.
10960 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10962 * config/s390/s390.c: Include tree-pass.h and context.h.
10963 (s390_early_mach): New function, split out from...
10964 (s390_emit_prologue): ...here.
10965 (pass_data_s390_early_mach): New pass structure.
10966 (pass_s390_early_mach): New class.
10967 (s390_option_override): Create and register early_mach pass.
10968 Move to end of file.
10970 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10972 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
10973 to match for the exit block.
10975 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10977 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
10978 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
10979 Reject misaligned operands.
10981 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10983 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
10985 2014-02-07 Richard Biener <rguenther@suse.de>
10987 PR middle-end/60092
10988 * gimple-low.c (lower_builtin_posix_memalign): New function.
10989 (lower_stmt): Call it to lower posix_memalign in a way
10990 to make alignment info accessible.
10992 2014-02-07 Jakub Jelinek <jakub@redhat.com>
10995 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
10996 __builtin_setjmp_receiver.
10998 2014-02-07 Richard Biener <rguenther@suse.de>
11000 PR middle-end/60092
11001 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
11002 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
11003 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
11004 Handle BUILT_IN_POSIX_MEMALIGN.
11005 (find_func_clobbers): Likewise.
11006 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
11007 (call_may_clobber_ref_p_1): Likewise.
11009 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
11012 * ipa-devirt.c (record_target_from_binfo): Remove overactive
11015 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
11018 * lto-cgraph.c (lto_output_node): Use
11019 symtab_get_symbol_partitioning_class.
11020 (lto_output_varpool_node): likewise.
11021 (symtab_get_symbol_partitioning_class): Move here from
11022 lto/lto-partition.c
11023 * cgraph.h (symbol_partitioning_class): Likewise.
11024 (symtab_get_symbol_partitioning_class): Declare.
11026 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
11028 * ggc.h (ggc_internal_cleared_alloc): New macro.
11029 * vec.h (vec_safe_copy): Handle memory stats.
11030 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
11031 * target-globals.c (save_target_globals): Likewise.
11033 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
11036 * expr.c (emit_move_resolve_push): Export; be bit more selective
11037 on when to clear alias set.
11038 * expr.h (emit_move_resolve_push): Declare.
11039 * function.h (struct function): Add tail_call_marked.
11040 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
11041 * config/i386/i386-protos.h (ix86_expand_push): Remove.
11042 * config/i386/i386.md (TImode move expander): De not call
11044 (FP push expanders): Preserve memory attributes.
11045 * config/i386/sse.md (push<mode>1): Remove.
11046 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
11047 (ix86_expand_push): Remove.
11048 * config/i386/mmx.md (push<mode>1): Remove.
11050 2014-02-06 Jakub Jelinek <jakub@redhat.com>
11052 PR rtl-optimization/60030
11053 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
11054 lopart with paradoxical subreg before shifting it up by hprec.
11056 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11058 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
11059 Remove extra newline at end of file.
11060 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
11061 (arm_issue_rate): Handle cortexa57.
11062 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
11063 (cortex-a57.cortex-a53): Likewise.
11065 2014-02-06 Jakub Jelinek <jakub@redhat.com>
11068 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
11069 don't record in REG_FRAME_RELATED_EXPR registers not set in that
11071 (arm_expand_prologue): Adjust all callers.
11072 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
11073 info, registers also at the lowest numbered registers side. Use
11074 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
11078 * var-tracking.c (adjust_mems): Before adding a SET to
11079 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
11081 2014-02-06 Alan Modra <amodra@gmail.com>
11084 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
11085 change SDmode to DDmode when lra_in_progress.
11087 2014-02-06 Jakub Jelinek <jakub@redhat.com>
11089 PR middle-end/59150
11090 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
11091 free_data_ref on the dr first, and before goto again also set dr
11092 to the next dr. For simd_lane_access, free old datarefs[i] before
11093 overwriting it. For get_vectype_for_scalar_type failure, don't
11094 free_data_ref if simd_lane_access.
11096 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
11099 * tree.h (opts_for_fn): New inline function.
11100 (opt_for_fn): Define.
11101 * config/i386/i386.c (ix86_function_regparm): Use
11102 opt_for_fn (decl, optimize) instead of optimize.
11104 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
11106 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
11107 for SYMBOL_REF in large memory model.
11109 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11111 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
11112 and crypto support.
11113 (cortex-a57): Likewise.
11114 (cortex-a57.cortex-a53): Likewise.
11116 2014-02-06 Yury Gribov <y.gribov@samsung.com>
11117 Kugan Vivekanandarajah <kuganv@linaro.org>
11119 * config/arm/arm.c (arm_vector_alignment_reachable): Check
11121 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
11123 2014-02-06 Richard Biener <rguenther@suse.de>
11125 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
11126 set_loop_copy and initialize_original_copy_tables.
11128 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
11130 * config/aarch64/aarch64-simd.md
11131 (aarch64_ashr_simddi): Change QI to SI.
11133 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
11134 Jakub Jelinek <jakub@redhat.com>
11136 PR middle-end/60013
11137 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
11140 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11142 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
11143 CODE_FOR_altivec_vpku[hw]um to
11144 CODE_FOR_altivec_vpku[hw]um_direct.
11145 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
11146 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
11147 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
11148 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
11150 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11152 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
11153 generation for -maltivec=be.
11154 (altivec_vsumsws): Simplify redundant test.
11156 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11158 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
11159 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
11160 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
11161 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
11162 gen_altivec_vpkuwum.
11163 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
11165 (altivec_vpks<VI_char>ss): Likewise.
11166 (altivec_vpks<VI_char>us): Likewise.
11167 (altivec_vpku<VI_char>us): Likewise.
11168 (altivec_vpku<VI_char>um): Likewise.
11169 (altivec_vpku<VI_char>um_direct): New (copy of
11170 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
11172 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
11173 target is little endian and -maltivec=be is not specified.
11174 (*altivec_vupkhs<VU_char>_direct): New (copy of
11175 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
11176 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
11177 target is little endian and -maltivec=be is not specified.
11178 (*altivec_vupkls<VU_char>_direct): New (copy of
11179 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
11180 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
11181 little endian and -maltivec=be is not specified.
11182 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
11183 little endian and -maltivec=be is not specified.
11185 2014-02-05 Richard Henderson <rth@redhat.com>
11188 * combine-stack-adj.c: Revert r206943.
11189 * sched-int.h (struct deps_desc): Add last_args_size.
11190 * sched-deps.c (init_deps): Initialize it.
11191 (sched_analyze_insn): Add OUTPUT dependencies between insns that
11192 contain REG_ARGS_SIZE notes.
11194 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
11196 * lto-cgraph.c (asm_nodes_output): Make global.
11197 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
11198 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
11199 (driver_handle_option): Handle OPT_fwpa.
11201 2014-02-05 Jakub Jelinek <jakub@redhat.com>
11204 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
11205 a comment typo and formatting issue. If odr_hash hasn't been
11206 created, return vNULL and set *completep to false.
11208 PR middle-end/57499
11209 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
11210 bb with no successors.
11212 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
11215 * doc/invoke.texi (-march): Clarify documentation for ARM.
11216 (-mtune): Likewise.
11219 2014-02-05 Richard Biener <rguenther@suse.de>
11221 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
11222 when not vectorizing because of too many alias checks.
11223 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
11224 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
11226 2014-02-05 Nick Clifton <nickc@redhat.com>
11228 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
11229 accept extended registers in any mode when compiling for the MN10300.
11231 2014-02-05 Yury Gribov <y.gribov@samsung.com>
11233 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
11234 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
11235 sanitization attributes.
11236 (can_inline_edge_p): Likewise.
11237 (sanitize_attrs_match_for_inline_p): New function.
11239 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
11241 * ipa-prop.c (detect_type_change): Shor circuit testing of
11242 type changes on THIS pointer.
11244 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
11247 * config/pa/pa.c (legitimize_tls_address): Return original address
11248 if not passed a SYMBOL_REF rtx.
11249 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
11251 (pa_emit_move_sequence): Simplify TLS source operands.
11252 (pa_legitimate_constant_p): Reject all TLS constants.
11253 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
11254 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
11256 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
11258 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
11259 groups when we know they are controlled by LTO.
11260 * varasm.c (default_binds_local_p_1): If object is in other partition,
11261 it will be resolved locally.
11263 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
11265 * config/host-linux.c (linux_gt_pch_use_address): Don't
11266 use SSIZE_MAX because it is not always defined.
11268 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
11271 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
11272 threshold for pseudo splitting.
11273 (update_ebb_live_info): Process call argument hard registers and
11274 hard registers from insn definition too.
11275 (max_small_class_regs_num): New constant.
11276 (inherit_in_ebb): Update live hard regs through EBBs. Update
11277 reloads_num only for small register classes. Don't split for
11280 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
11283 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
11286 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
11288 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
11289 visibility is safe.
11291 2014-02-04 Marek Polacek <polacek@redhat.com>
11293 * gdbinit.in (pel): Define.
11295 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
11297 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
11300 2014-02-04 Richard Biener <rguenther@suse.de>
11303 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
11304 in function context local.
11305 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
11306 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
11307 similar to LTO_imported_decl_ref.
11309 2014-02-04 Jakub Jelinek <jakub@redhat.com>
11311 PR tree-optimization/60002
11312 * cgraphclones.c (build_function_decl_skip_args): Clear
11313 DECL_LANG_SPECIFIC.
11315 PR tree-optimization/60023
11316 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
11317 false to gsi_replace.
11318 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
11319 has been in some EH region and vec_stmt could throw, add
11320 vec_stmt into the same EH region.
11321 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
11322 has no lhs, ignore it.
11323 * internal-fn.c (expand_MASK_LOAD): Likewise.
11326 * tree-inline.c (copy_forbidden): Fail for
11327 __attribute__((optimize (0))) functions.
11330 * omp-low.c (simd_clone_struct_copy): If from->inbranch
11331 is set, copy one less argument.
11332 (expand_simd_clones): Don't subtract clone_info->inbranch
11333 from simd_clone_struct_alloc argument.
11335 PR rtl-optimization/57915
11336 * recog.c (simplify_while_replacing): If all unary/binary/relational
11337 operation arguments are constant, attempt to simplify those.
11339 PR middle-end/59261
11340 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
11341 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
11343 2014-02-04 Richard Biener <rguenther@suse.de>
11345 PR tree-optimization/60012
11346 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
11347 TBAA disambiguation to all DDRs.
11349 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11352 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
11353 (LINK_SPEC): Use it for -shared, -shared-libgcc.
11355 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
11358 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
11360 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
11362 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
11363 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
11365 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
11368 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
11369 to figure out targets of polymorphic calls with known decl.
11370 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
11371 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
11372 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
11373 (get_polymorphic_call_info): ... here.
11374 (get_polymorphic_call_info_from_invariant): New function.
11376 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
11378 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
11379 lookup via vtable pointer; check for type consistency
11380 and turn inconsitent facts into UNREACHABLE.
11381 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
11382 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
11383 type inconsistent querries; return UNREACHABLE instead.
11385 2014-02-03 Richard Henderson <rth@twiddle.net>
11388 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
11389 already processed this node.
11390 (normalize_one_pred_1): Pass along mark_set.
11391 (normalize_one_pred): Create and destroy a pointer_set_t.
11392 (normalize_one_pred_chain): Likewise.
11394 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
11396 PR gcov-profile/58602
11397 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
11399 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
11402 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
11403 -fno-devirtualize; try to devirtualize by the knowledge of
11404 virtual table pointer given by aggregate propagation.
11405 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
11406 (ipa_print_node_jump_functions): Dump also offset that
11407 is relevant for polymorphic calls.
11408 (determine_known_aggregate_parts): Add arg_type parameter; use it
11409 instead of determining the type from pointer type.
11410 (ipa_compute_jump_functions_for_edge): Update call of
11411 determine_known_aggregate_parts.
11412 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
11413 (gimple_get_virt_method_for_binfo): ... here; simplify using
11414 vtable_pointer_value_to_vtable.
11415 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
11416 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
11417 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
11418 (vtable_pointer_value_to_vtable): Break out from ...; handle also
11420 (vtable_pointer_value_to_binfo): ... here.
11421 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
11423 2014-02-03 Teresa Johnson <tejohnson@google.com>
11425 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
11426 redef of outer loop index variable.
11428 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
11432 * doc/extend.texi (Function Attributes): Typo.
11434 2014-02-03 Cong Hou <congh@google.com>
11436 PR tree-optimization/60000
11437 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
11438 if the vectorized statement is a store. A store statement can only
11439 appear at the end of pattern statements.
11441 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
11443 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
11444 (ix86_option_override_internal): Default long double to 64-bit for
11445 32-bit Bionic and to 128-bit for 64-bit Bionic.
11447 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
11448 TARGET_LONG_DOUBLE_128 is true.
11449 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
11451 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
11452 (mlong-double-64): Negate -mlong-double-128.
11453 (mlong-double-128): New option.
11455 * config/i386/i386-c.c (ix86_target_macros): Define
11456 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
11458 * doc/invoke.texi: Document -mlong-double-128.
11460 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
11462 PR rtl-optimization/60024
11463 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
11465 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
11467 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
11469 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
11471 PR rtl-optimization/57662
11472 * sel-sched.c (code_motion_path_driver): Do not mark already not
11473 existing blocks in the visiting bitmap.
11475 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
11477 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
11478 on the insn being emitted.
11480 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
11481 Will Deacon <will.deacon@arm.com>
11483 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
11485 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11487 * config/arm/arm-tables.opt: Regenerate.
11489 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11491 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
11492 for vector types other than V16QImode.
11493 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
11494 define_expand, and call altivec_expand_vec_perm_le when producing
11495 code with little endian element order.
11496 (*altivec_vperm_<mode>_internal): New insn having previous
11497 behavior of altivec_vperm_<mode>.
11498 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
11499 altivec_expand_vec_perm_le when producing code with little endian
11501 (*altivec_vperm_<mode>_uns_internal): New insn having previous
11502 behavior of altivec_vperm_<mode>_uns.
11504 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11506 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
11507 (altivec_vsumsws): Add handling for -maltivec=be with a little
11509 (altivec_vsumsws_direct): New.
11510 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
11511 gen_altivec_vsumsws.
11513 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
11515 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
11516 vtable_pointer_value_to_binfo): New functions.
11517 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
11518 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
11520 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
11522 * config/nios2/nios2.md (load_got_register): Initialize GOT
11523 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
11524 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
11526 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
11528 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
11529 preserverd by passthrough, do not propagate the type.
11531 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
11533 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
11534 (mips_atomic_assign_expand_fenv): New function.
11535 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
11537 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
11539 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
11540 (__builtin_mips_set_fcsr): Likewise.
11541 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
11542 MIPS_USI_FTYPE_VOID.
11543 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
11544 (mips16_expand_set_fcsr): Likewise.
11545 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
11546 (mips16_set_fcsr_stub): Likewise.
11547 (mips16_get_fcsr_one_only_stub): New class.
11548 (mips16_set_fcsr_one_only_stub): Likewise.
11549 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
11550 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
11551 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
11552 (hard_float): New availability predicate.
11553 (mips_builtins): Add get_fcsr and set_fcsr.
11554 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
11555 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
11556 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
11557 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
11558 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
11561 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
11563 * config/mips/mips.c (mips_one_only_stub): New class.
11564 (mips_need_mips16_rdhwr_p): Replace with...
11565 (mips16_rdhwr_stub): ...this new variable.
11566 (mips16_stub_call_address): New function.
11567 (mips16_rdhwr_one_only_stub): New class.
11568 (mips_expand_thread_pointer): Use mips16_stub_call_address.
11569 (mips_output_mips16_rdhwr): Delete.
11570 (mips_finish_stub): New function.
11571 (mips_code_end): Use it to handle rdhwr stubs.
11573 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
11576 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
11577 when calculating size of integer atomic types.
11579 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
11581 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
11583 2014-02-01 Jakub Jelinek <jakub@redhat.com>
11585 PR tree-optimization/60003
11586 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
11587 * profile.c (branch_prob): Use gimple_call_builtin_p
11588 to check for BUILT_IN_SETJMP_RECEIVER.
11589 * tree-inline.c (copy_bb): Call notice_special_calls.
11591 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
11594 * lra-constraints.c (process_alt_operands): Update reload_sum only
11597 2014-01-31 Richard Henderson <rth@redhat.com>
11599 PR middle-end/60004
11600 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
11601 until after else_eh is processed.
11603 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
11605 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
11606 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
11607 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
11608 in smmintrin.h, remove them.
11609 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
11610 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
11611 * config/i386/i386.md (ROUND_SAE): Fix value.
11612 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
11613 (const48_operand): New.
11614 * config/i386/subst.md (round), (round_expand): Use
11615 const_4_or_8_to_11_operand.
11616 (round_saeonly), (round_saeonly_expand): Use const48_operand.
11618 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
11620 * config/i386/constraints.md (Yk): Swap meaning with k.
11621 * config/i386/i386.md (movhi_internal): Change Yk to k.
11622 (movqi_internal): Ditto.
11623 (*k<logic><mode>): Ditto.
11626 (kandn<mode>): Ditto.
11627 (*<code>hi_1): Ditto.
11628 (*<code>qi_1): Ditto.
11629 (kxnor<mode>): Ditto.
11630 (kortestzhi): Ditto.
11631 (kortestchi): Ditto.
11633 (*one_cmplhi2_1): Ditto.
11634 (*one_cmplqi2_1): Ditto.
11635 * config/i386/sse.md (): Change k to Yk.
11636 (avx512f_load<mode>_mask): Ditto.
11637 (avx512f_blendm<mode>): Ditto.
11638 (avx512f_store<mode>_mask): Ditto.
11639 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
11640 (avx512f_storedqu<mode>_mask): Ditto.
11641 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
11643 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
11644 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
11645 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
11646 (avx512f_maskcmp<mode>3): Ditto.
11647 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
11648 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
11649 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
11650 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
11651 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
11652 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
11653 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
11654 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
11655 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
11656 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
11657 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
11658 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
11659 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
11660 (vec_extract_lo_<mode>_maskm): Ditto.
11661 (vec_extract_hi_<mode>_maskm): Ditto.
11662 (avx512f_vternlog<mode>_mask): Ditto.
11663 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
11664 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
11665 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
11666 (avx512f_<code>v8div16qi2_mask): Ditto.
11667 (avx512f_<code>v8div16qi2_mask_store): Ditto.
11668 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
11669 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
11670 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
11671 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
11672 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
11673 (*avx512pf_gatherpf<mode>df_mask): Ditto.
11674 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
11675 (*avx512pf_scatterpf<mode>df_mask): Ditto.
11676 (avx512cd_maskb_vec_dupv8di): Ditto.
11677 (avx512cd_maskw_vec_dupv16si): Ditto.
11678 (avx512f_vpermi2var<mode>3_maskz): Ditto.
11679 (avx512f_vpermi2var<mode>3_mask): Ditto.
11680 (avx512f_vpermi2var<mode>3_mask): Ditto.
11681 (avx512f_vpermt2var<mode>3_maskz): Ditto.
11682 (*avx512f_gathersi<mode>): Ditto.
11683 (*avx512f_gathersi<mode>_2): Ditto.
11684 (*avx512f_gatherdi<mode>): Ditto.
11685 (*avx512f_gatherdi<mode>_2): Ditto.
11686 (*avx512f_scattersi<mode>): Ditto.
11687 (*avx512f_scatterdi<mode>): Ditto.
11688 (avx512f_compress<mode>_mask): Ditto.
11689 (avx512f_compressstore<mode>_mask): Ditto.
11690 (avx512f_expand<mode>_mask): Ditto.
11691 * config/i386/subst.md (mask): Change k to Yk.
11692 (mask_scalar_merge): Ditto.
11695 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
11697 * doc/extend.texi (Vector Extensions): Document ?: in C++.
11699 2014-01-31 Richard Biener <rguenther@suse.de>
11701 PR middle-end/59990
11702 * builtins.c (fold_builtin_memory_op): Make sure to not
11703 use a floating-point mode or a boolean or enumeral type for
11704 the copy operation.
11706 2014-01-30 DJ Delorie <dj@redhat.com>
11708 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
11709 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
11710 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
11711 whenever main() has an epilogue.
11713 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11715 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
11716 unused variable "field".
11717 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
11718 (vsx_mergeh_<mode>): Likewise.
11719 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
11720 (altivec_vmrghh): Likewise.
11721 (altivec_vmrghw): Likewise.
11722 (altivec_vmrglb): Likewise.
11723 (altivec_vmrglh): Likewise.
11724 (altivec_vmrglw): Likewise.
11725 (altivec_vspltb): Add missing uses.
11726 (altivec_vsplth): Likewise.
11727 (altivec_vspltw): Likewise.
11728 (altivec_vspltsf): Likewise.
11730 2014-01-30 Jakub Jelinek <jakub@redhat.com>
11733 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
11734 frame related instructions.
11736 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
11738 PR rtl-optimization/59959
11739 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
11740 any reload of register whose subreg is invalid.
11742 2014-01-30 Jakub Jelinek <jakub@redhat.com>
11744 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
11745 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
11746 Add missing return type - void.
11748 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11750 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
11751 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
11752 remove element index adjustment for endian (now handled in vsx.md
11754 (altivec_expand_vec_perm_const): Use
11755 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
11756 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
11757 (vsx_xxspltw_<mode>): Adjust element index for little endian.
11758 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
11759 define_expand and a new define_insn *altivec_vspltb_internal;
11760 adjust for -maltivec=be on a little endian target.
11761 (altivec_vspltb_direct): New.
11762 (altivec_vsplth): Divide into a define_expand and a new
11763 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
11764 little endian target.
11765 (altivec_vsplth_direct): New.
11766 (altivec_vspltw): Divide into a define_expand and a new
11767 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
11768 little endian target.
11769 (altivec_vspltw_direct): New.
11770 (altivec_vspltsf): Divide into a define_expand and a new
11771 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
11772 a little endian target.
11774 2014-01-30 Richard Biener <rguenther@suse.de>
11776 PR tree-optimization/59993
11777 * tree-ssa-forwprop.c (associate_pointerplus): Check we
11778 can propagate form the earlier stmt and avoid the transform
11779 when the intermediate result is needed.
11781 2014-01-30 Alangi Derick <alangiderick@gmail.com>
11783 * README.Portability: Fix typo.
11785 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
11787 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
11788 comparison_operator with ordered_comparison_operator.
11790 2014-01-30 Nick Clifton <nickc@redhat.com>
11792 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
11793 Rename to mn10300_store_multiple_regs.
11794 * config/mn10300/mn10300.c: Likewise.
11795 * config/mn10300/mn10300.md (store_movm): Fix typo: call
11796 store_multiple_regs.
11797 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
11798 Call mn10300_store_multiple_regs.
11800 2014-01-30 Nick Clifton <nickc@redhat.com>
11801 DJ Delorie <dj@redhat.com>
11803 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
11804 %fp 2 to keep registers after it properly word-aligned.
11805 (rl78_alloc_physical_registers_umul): Handle the case where both
11806 input operands are the same.
11808 2014-01-30 Richard Biener <rguenther@suse.de>
11810 PR tree-optimization/59903
11811 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
11814 2014-01-30 Jason Merrill <jason@redhat.com>
11817 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
11820 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
11822 2014-01-30 Richard Biener <rguenther@suse.de>
11824 PR tree-optimization/59951
11825 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
11827 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
11830 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
11831 SFmode to DFmode case.
11833 2014-01-29 DJ Delorie <dj@redhat.com>
11835 * config/msp430/msp430.opt (-minrt): New.
11836 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
11838 (ENDFILE_SPEC): Likewise.
11840 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
11842 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
11843 (estimate_function_body_sizes): Use it.
11845 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
11848 * dwarf2out.c (is_cxx_auto): New.
11849 (is_base_type): Use it.
11850 (gen_type_die_with_usage): Likewise.
11852 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11854 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
11855 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
11856 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
11857 -maltivec=be with LE targets.
11858 (vsx_mergeh_<mode>): Likewise.
11859 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
11860 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
11861 (altivec_vmrghb): Replace with define_expand and new
11862 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
11863 (altivec_vmrghb_direct): New define_insn.
11864 (altivec_vmrghh): Replace with define_expand and new
11865 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
11866 (altivec_vmrghh_direct): New define_insn.
11867 (altivec_vmrghw): Replace with define_expand and new
11868 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
11869 (altivec_vmrghw_direct): New define_insn.
11870 (*altivec_vmrghsf): Adjust for endianness.
11871 (altivec_vmrglb): Replace with define_expand and new
11872 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
11873 (altivec_vmrglb_direct): New define_insn.
11874 (altivec_vmrglh): Replace with define_expand and new
11875 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
11876 (altivec_vmrglh_direct): New define_insn.
11877 (altivec_vmrglw): Replace with define_expand and new
11878 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
11879 (altivec_vmrglw_direct): New define_insn.
11880 (*altivec_vmrglsf): Adjust for endianness.
11881 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
11882 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
11883 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
11884 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
11885 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
11886 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
11887 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
11888 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
11890 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
11892 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
11893 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
11896 2014-01-29 Richard Biener <rguenther@suse.de>
11898 PR tree-optimization/58742
11899 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
11900 associate_pointerplus_align.
11901 (associate_pointerplus_diff): New function.
11902 (associate_pointerplus): Likewise. Call associate_pointerplus_align
11903 and associate_pointerplus_diff.
11905 2014-01-29 Richard Biener <rguenther@suse.de>
11907 * lto-streamer.h (LTO_major_version): Bump to 3.
11908 (LTO_minor_version): Reset to 0.
11910 2014-01-29 Renlin Li <Renlin.Li@arm.com>
11912 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
11913 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
11914 (arm_file_start): Generate correct asm header for armv7ve.
11915 * config/arm/bpabi.h: Add multilib support for armv7ve.
11916 * config/arm/driver-arm.c: Change the architectures of cortex-a7
11917 and cortex-a15 to armv7ve.
11918 * config/arm/t-aprofile: Add multilib support for armv7ve.
11919 * doc/invoke.texi: Document -march=armv7ve.
11921 2014-01-29 Richard Biener <rguenther@suse.de>
11923 PR tree-optimization/58742
11924 * tree-ssa-forwprop.c (associate_plusminus): Return true
11925 if we changed sth, defer EH cleanup to ...
11926 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
11927 (simplify_mult): New function.
11929 2014-01-29 Jakub Jelinek <jakub@redhat.com>
11931 PR middle-end/59917
11932 PR tree-optimization/59920
11933 * tree.c (build_common_builtin_nodes): Remove
11934 __builtin_setjmp_dispatcher initialization.
11935 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
11936 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
11937 instead of gsi_after_labels + manually skipping debug stmts.
11938 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
11939 ignore bbs with IFN_ABNORMAL_DISPATCHER.
11940 * tree-inline.c (copy_edges_for_bb): Remove
11941 can_make_abnormal_goto argument, instead add abnormal_goto_dest
11942 argument. Ignore computed_goto_p stmts. Don't call
11943 make_abnormal_goto_edges. If a call might need abnormal edges
11944 for non-local gotos, see if it already has an edge to
11945 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
11946 with true argument, don't do anything then, otherwise add
11947 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
11948 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
11950 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
11951 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
11952 (lower_stmt): Don't set data->calls_builtin_setjmp.
11953 (lower_builtin_setjmp): Adjust comment.
11954 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
11955 * tree-cfg.c (found_computed_goto): Remove.
11956 (factor_computed_gotos): Remove.
11957 (make_goto_expr_edges): Return bool, true for computed gotos.
11958 Don't call make_abnormal_goto_edges.
11959 (build_gimple_cfg): Don't set found_computed_goto, don't call
11960 factor_computed_gotos.
11961 (computed_goto_p): No longer static.
11962 (make_blocks): Don't set found_computed_goto.
11963 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
11964 (make_edges): If make_goto_expr_edges returns true, push bb
11965 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
11966 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
11967 vector. Record mapping between bbs and OpenMP regions if there
11968 are any, adjust make_gimple_omp_edges caller. Call
11969 handle_abnormal_edges.
11970 (make_abnormal_goto_edges): Remove.
11971 * tree-cfg.h (make_abnormal_goto_edges): Remove.
11972 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
11973 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
11974 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
11975 * internal-fn.def (ABNORMAL_DISPATCHER): New.
11976 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
11977 filling *region also set *region_idx to (*region)->entry->index.
11980 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
11981 For REGs set ORIGINAL_REGNO.
11983 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
11985 * doc/md.texi: Mention that a target shouldn't implement
11986 vec_widen_(s|u)mul_even/odd pair if it is less efficient
11989 2014-01-29 Jakub Jelinek <jakub@redhat.com>
11991 PR tree-optimization/59594
11992 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
11993 a copy of the datarefs vector rather than the vector itself.
11995 2014-01-28 Jason Merrill <jason@redhat.com>
11998 * dwarf2out.c (auto_die): New static.
11999 (gen_type_die_with_usage): Handle C++1y 'auto'.
12000 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
12003 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
12006 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
12007 (SPEC_X32): Likewise.
12008 (SPEC_64): Likewise.
12009 * config/i386/i386.c (ix86_option_override_internal): Turn off
12010 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
12012 (x86_file_start): Output .code16gcc for TARGET_16BIT.
12013 * config/i386/i386.h (TARGET_16BIT): New macro.
12014 (TARGET_16BIT_P): Likewise.
12015 * config/i386/i386.opt: Add m16.
12016 * doc/invoke.texi: Document -m16.
12018 2014-01-28 Jakub Jelinek <jakub@redhat.com>
12020 PR preprocessor/59935
12021 * input.c (location_get_source_line): Bail out on when line number
12022 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
12024 2014-01-28 Richard Biener <rguenther@suse.de>
12026 PR tree-optimization/58742
12027 * tree-ssa-forwprop.c (associate_plusminus): Handle
12028 pointer subtraction of the form (T)(P + A) - (T)P.
12030 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12032 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
12035 2014-01-28 Richard Biener <rguenther@suse.de>
12038 2014-01-28 Richard Biener <rguenther@suse.de>
12040 PR rtl-optimization/45364
12041 PR rtl-optimization/59890
12042 * var-tracking.c (local_get_addr_clear_given_value): Handle
12043 already cleared slot.
12044 (val_reset): Handle not allocated local_get_addr_cache.
12045 (vt_find_locations): Use post-order on the inverted CFG.
12047 2014-01-28 Richard Biener <rguenther@suse.de>
12049 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
12051 2014-01-28 Richard Biener <rguenther@suse.de>
12053 PR rtl-optimization/45364
12054 PR rtl-optimization/59890
12055 * var-tracking.c (local_get_addr_clear_given_value): Handle
12056 already cleared slot.
12057 (val_reset): Handle not allocated local_get_addr_cache.
12058 (vt_find_locations): Use post-order on the inverted CFG.
12060 2014-01-28 Alan Modra <amodra@gmail.com>
12062 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
12063 * configure.ac <recursive call for build != host>: Define
12064 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
12065 and LD_FOR_BUILD too.
12066 * configure: Regenerate.
12068 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
12070 * config/i386/i386.c (get_builtin_code_for_version): Separate
12071 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
12072 Broadwell from Haswell.
12074 2014-01-27 Steve Ellcey <sellcey@mips.com>
12076 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
12077 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
12078 * config/mips/mips.c (mips_option_override): Change setting
12080 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
12081 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
12082 Change from Mask to Var.
12084 2014-01-27 Jeff Law <law@redhat.com>
12086 * ipa-inline.c (inline_small_functions): Fix typo.
12088 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
12090 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
12091 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
12092 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
12093 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
12094 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
12095 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
12096 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
12097 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
12098 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
12099 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
12100 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
12101 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
12102 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
12103 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
12104 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
12105 (_mm512_storeu_epi64): Ditto.
12106 (_mm512_cmpge_epi32_mask): Ditto.
12107 (_mm512_cmpge_epu32_mask): Ditto.
12108 (_mm512_cmpge_epi64_mask): Ditto.
12109 (_mm512_cmpge_epu64_mask): Ditto.
12110 (_mm512_cmple_epi32_mask): Ditto.
12111 (_mm512_cmple_epu32_mask): Ditto.
12112 (_mm512_cmple_epi64_mask): Ditto.
12113 (_mm512_cmple_epu64_mask): Ditto.
12114 (_mm512_cmplt_epi32_mask): Ditto.
12115 (_mm512_cmplt_epu32_mask): Ditto.
12116 (_mm512_cmplt_epi64_mask): Ditto.
12117 (_mm512_cmplt_epu64_mask): Ditto.
12118 (_mm512_cmpneq_epi32_mask): Ditto.
12119 (_mm512_cmpneq_epu32_mask): Ditto.
12120 (_mm512_cmpneq_epi64_mask): Ditto.
12121 (_mm512_cmpneq_epu64_mask): Ditto.
12122 (_mm512_expand_pd): Ditto.
12123 (_mm512_expand_ps): Ditto.
12124 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
12125 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
12126 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
12127 * config/i386/i386.c (ix86_builtins): Add
12128 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
12129 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
12130 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
12131 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
12132 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
12133 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
12134 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
12135 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
12136 IX86_BUILTIN_PMOVUSQW512_MEM.
12137 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
12138 __builtin_ia32_pmovsqd512mem_mask,
12139 __builtin_ia32_pmovqd512mem_mask,
12140 __builtin_ia32_pmovusqw512mem_mask,
12141 __builtin_ia32_pmovsqw512mem_mask,
12142 __builtin_ia32_pmovqw512mem_mask,
12143 __builtin_ia32_pmovusdw512mem_mask,
12144 __builtin_ia32_pmovsdw512mem_mask,
12145 __builtin_ia32_pmovdw512mem_mask,
12146 __builtin_ia32_pmovqb512mem_mask,
12147 __builtin_ia32_pmovusqb512mem_mask,
12148 __builtin_ia32_pmovsqb512mem_mask,
12149 __builtin_ia32_pmovusdb512mem_mask,
12150 __builtin_ia32_pmovsdb512mem_mask,
12151 __builtin_ia32_pmovdb512mem_mask.
12152 (bdesc_args): Add __builtin_ia32_expanddf512,
12153 __builtin_ia32_expandsf512.
12154 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
12155 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
12156 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
12157 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
12158 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
12159 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
12160 (avx512f_<code>v8div16qi2_mask_store): This.
12161 (avx512f_expand<mode>): New.
12163 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
12165 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
12167 (_mm512_mask_prefetch_i64gather_pd): Ditto.
12168 (_mm512_prefetch_i32scatter_pd): Ditto.
12169 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
12170 (_mm512_prefetch_i64scatter_pd): Ditto.
12171 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
12172 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
12173 (_mm512_mask_prefetch_i64gather_ps): Ditto.
12174 (_mm512_prefetch_i32scatter_ps): Ditto.
12175 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
12176 (_mm512_prefetch_i64scatter_ps): Ditto.
12177 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
12178 * config/i386/i386-builtin-types.def: Define
12179 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
12180 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
12181 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
12182 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
12183 IX86_BUILTIN_SCATTERPFQPD.
12184 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
12185 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
12186 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
12187 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
12188 __builtin_ia32_scatterpfqps.
12189 (ix86_expand_builtin): Expand new built-ins.
12190 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
12191 fix memory access data type.
12192 (*avx512pf_gatherpf<mode>_mask): Ditto.
12193 (*avx512pf_gatherpf<mode>): Ditto.
12194 (avx512pf_scatterpf<mode>): Ditto.
12195 (*avx512pf_scatterpf<mode>_mask): Ditto.
12196 (*avx512pf_scatterpf<mode>): Ditto.
12197 (GATHER_SCATTER_SF_MEM_MODE): New.
12198 (avx512pf_gatherpf<mode>df): Ditto.
12199 (*avx512pf_gatherpf<mode>df_mask): Ditto.
12200 (*avx512pf_scatterpf<mode>df): Ditto.
12202 2014-01-27 Jakub Jelinek <jakub@redhat.com>
12205 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
12206 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
12209 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
12211 * common/config/arm/arm-common.c
12212 (arm_rewrite_mcpu): Handle multiple names.
12214 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
12216 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
12218 * gimple-builder.h (create_gimple_tmp): Delete.
12220 2014-01-27 Christian Bruel <christian.bruel@st.com>
12222 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
12225 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
12227 * config/pa/pa.md (call): Generate indirect long calls to non-local
12228 functions when outputing 32-bit code.
12229 (call_value): Likewise except for special call to buggy powf function.
12231 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
12232 portable runtime and PIC indirect calls.
12233 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
12234 and PIC call sequences. Use ldo instead of blr to set return register
12235 in PIC call sequence.
12237 2014-01-25 Walter Lee <walt@tilera.com>
12239 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
12240 avoid clobbering a live register.
12242 2014-01-25 Walter Lee <walt@tilera.com>
12244 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
12245 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
12246 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
12247 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
12249 2014-01-25 Walter Lee <walt@tilera.com>
12251 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
12252 arguments on even registers.
12253 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
12255 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
12256 (BIGGEST_ALIGNMENT): Ditto.
12257 (BIGGEST_FIELD_ALIGNMENT): Ditto.
12259 2014-01-25 Walter Lee <walt@tilera.com>
12261 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
12262 insns before bundling.
12263 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
12265 2014-01-25 Walter Lee <walt@tilera.com>
12267 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
12268 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
12269 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
12271 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
12273 * config/mips/constraints.md (kl): Delete.
12274 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
12275 define expands, using...
12276 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
12277 instructions for MIPS16.
12278 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
12279 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
12281 2014-01-25 Walter Lee <walt@tilera.com>
12283 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
12287 2014-01-25 Walter Lee <walt@tilera.com>
12289 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
12290 (TARGET_EXPAND_TO_RTL_HOOK): Define.
12292 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
12294 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
12297 2014-01-25 Jakub Jelinek <jakub@redhat.com>
12299 * print-rtl.c (in_call_function_usage): New var.
12300 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
12301 EXPR_LIST mode as mode and not as reg note name.
12303 PR middle-end/59561
12304 * cfgloopmanip.c (copy_loop_info): If
12305 loop->warned_aggressive_loop_optimizations, make sure
12306 the flag is set in target loop too.
12308 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
12310 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
12312 * builtins.def: Likewise.
12313 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
12314 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
12315 * ira.c (ira_setup_eliminable_regset): Likewise.
12316 * omp-low.c (gate_expand_omp): Likewise.
12317 (execute_lower_omp): Likewise.
12318 (diagnose_sb_0): Likewise.
12319 (gate_diagnose_omp_blocks): Likewise.
12320 (simd_clone_clauses_extract): Likewise.
12323 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12325 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
12326 correction for little endian...
12327 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
12330 2014-01-24 Jeff Law <law@redhat.com>
12332 PR tree-optimization/59919
12333 * tree-vrp.c (find_assert_locations_1): Do not register asserts
12334 for non-returning calls.
12336 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
12338 * common/config/aarch64/aarch64-common.c
12339 (aarch64_rewrite_mcpu): Handle multiple names.
12340 * config/aarch64/aarch64.h
12341 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
12343 2014-01-24 Dodji Seketeli <dodji@redhat.com>
12345 * input.c (add_file_to_cache_tab): Handle the case where fopen
12348 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
12351 * config/i386/i386.md (pushsf splitter): Get stack adjustment
12352 from push operand if code of push isn't PRE_DEC.
12354 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
12357 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
12358 -mquad-memory-atomic. Update -mquad-memory documentation to say
12359 it is only used for non-atomic loads/stores.
12361 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
12362 -mquad-memory or -mquad-memory-atomic switches.
12364 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
12365 -mquad-memory-atomic to ISA 2.07 support.
12367 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
12368 to separate support of normal quad word memory operations (ldq, stq)
12369 from the atomic quad word memory operations.
12371 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
12372 support to separate non-atomic quad word operations from atomic
12373 quad word operations. Disable non-atomic quad word operations in
12374 little endian mode so that we don't have to swap words after the
12375 load and before the store.
12376 (quad_load_store_p): Add comment about atomic quad word support.
12377 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
12378 options printed with -mdebug=reg.
12380 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
12381 -mquad-memory-atomic as the test for whether we have quad word
12382 atomic instructions.
12383 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
12384 or -mp8-vector are used, allow byte/half-word atomic operations.
12386 * config/rs6000/sync.md (load_lockedti): Insure that the address
12387 is a proper indexed or indirect address for the lqarx instruction.
12388 On little endian systems, swap the hi/lo registers after the lqarx
12390 (load_lockedpti): Use indexed_or_indirect_operand predicate to
12391 insure the address is valid for the lqarx instruction.
12392 (store_conditionalti): Insure that the address is a proper indexed
12393 or indirect address for the stqcrx. instruction. On little endian
12394 systems, swap the hi/lo registers before doing the stqcrx.
12396 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
12397 insure the address is valid for the stqcrx. instruction.
12399 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
12400 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
12401 type of quad memory support is available.
12403 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
12405 PR regression/59915
12406 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
12407 there is a danger of looping.
12409 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
12411 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
12412 force flag_ira_loop_pressure if set via command line.
12414 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
12416 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
12417 (ashr_simd): New builtin handling DI mode.
12418 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
12419 (aarch64_sshr_simddi): New match pattern.
12420 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
12421 (vshrd_n_s64): Likewise.
12422 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
12424 2014-01-23 Nick Clifton <nickc@redhat.com>
12426 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
12427 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
12428 favour of mcu specific scripts.
12429 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
12432 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
12433 Alex Velenko <Alex.Velenko@arm.com>
12435 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
12436 (vaddv_s16): Likewise.
12437 (vaddv_s32): Likewise.
12438 (vaddv_u8): Likewise.
12439 (vaddv_u16): Likewise.
12440 (vaddv_u32): Likewise.
12441 (vaddvq_s8): Likewise.
12442 (vaddvq_s16): Likewise.
12443 (vaddvq_s32): Likewise.
12444 (vaddvq_s64): Likewise.
12445 (vaddvq_u8): Likewise.
12446 (vaddvq_u16): Likewise.
12447 (vaddvq_u32): Likewise.
12448 (vaddvq_u64): Likewise.
12449 (vaddv_f32): Likewise.
12450 (vaddvq_f32): Likewise.
12451 (vaddvq_f64): Likewise.
12452 (vmaxv_f32): Likewise.
12453 (vmaxv_s8): Likewise.
12454 (vmaxv_s16): Likewise.
12455 (vmaxv_s32): Likewise.
12456 (vmaxv_u8): Likewise.
12457 (vmaxv_u16): Likewise.
12458 (vmaxv_u32): Likewise.
12459 (vmaxvq_f32): Likewise.
12460 (vmaxvq_f64): Likewise.
12461 (vmaxvq_s8): Likewise.
12462 (vmaxvq_s16): Likewise.
12463 (vmaxvq_s32): Likewise.
12464 (vmaxvq_u8): Likewise.
12465 (vmaxvq_u16): Likewise.
12466 (vmaxvq_u32): Likewise.
12467 (vmaxnmv_f32): Likewise.
12468 (vmaxnmvq_f32): Likewise.
12469 (vmaxnmvq_f64): Likewise.
12470 (vminv_f32): Likewise.
12471 (vminv_s8): Likewise.
12472 (vminv_s16): Likewise.
12473 (vminv_s32): Likewise.
12474 (vminv_u8): Likewise.
12475 (vminv_u16): Likewise.
12476 (vminv_u32): Likewise.
12477 (vminvq_f32): Likewise.
12478 (vminvq_f64): Likewise.
12479 (vminvq_s8): Likewise.
12480 (vminvq_s16): Likewise.
12481 (vminvq_s32): Likewise.
12482 (vminvq_u8): Likewise.
12483 (vminvq_u16): Likewise.
12484 (vminvq_u32): Likewise.
12485 (vminnmv_f32): Likewise.
12486 (vminnmvq_f32): Likewise.
12487 (vminnmvq_f64): Likewise.
12489 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
12491 * config/aarch64/aarch64-simd.md
12492 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
12493 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
12494 (*aarch64_mul3_elt<mode>): Likewise.
12495 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
12496 (*aarch64_mul3_elt_to_64v2df): Likewise.
12497 (*aarch64_mla_elt<mode>): Likewise.
12498 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
12499 (*aarch64_mls_elt<mode>): Likewise.
12500 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
12501 (*aarch64_fma4_elt<mode>): Likewise.
12502 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
12503 (*aarch64_fma4_elt_to_64v2df): Likewise.
12504 (*aarch64_fnma4_elt<mode>): Likewise.
12505 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
12506 (*aarch64_fnma4_elt_to_64v2df): Likewise.
12507 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
12508 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
12509 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
12510 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
12511 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
12512 (aarch64_sqdmull_lane<mode>_internal): Likewise.
12513 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
12515 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
12517 * config/aarch64/aarch64-simd.md
12518 (aarch64_be_checked_get_lane<mode>): New define_expand.
12519 * config/aarch64/aarch64-simd-builtins.def
12520 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
12521 New builtin definition.
12522 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
12523 Use new safe be builtin.
12525 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
12527 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
12529 (aarch64_be_st1<mode>): Likewise.
12530 (aarch_ld1<VALL:mode>): Define_expand modified.
12531 (aarch_st1<VALL:mode>): Likewise.
12532 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
12533 (UNSPEC_ST1): Likewise.
12535 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
12537 * config/microblaze/microblaze.md: Add trap insn and attribute
12539 2014-01-23 Dodji Seketeli <dodji@redhat.com>
12541 PR preprocessor/58580
12542 * input.h (location_get_source_line): Take an additional line_size
12544 (void diagnostics_file_cache_fini): Declare new function.
12545 * input.c (struct fcache): New type.
12546 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
12547 New static constants.
12548 (diagnostic_file_cache_init, total_lines_num)
12549 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
12550 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
12551 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
12552 (get_next_line, read_next_line, goto_next_line, read_line_num):
12553 New static function definitions.
12554 (diagnostic_file_cache_fini): New function.
12555 (location_get_source_line): Take an additional output line_len
12556 parameter. Re-write using lookup_or_add_file_to_cache_tab and
12558 * diagnostic.c (diagnostic_finish): Call
12559 diagnostic_file_cache_fini.
12560 (adjust_line): Take an additional input parameter for the length
12561 of the line, rather than calculating it with strlen.
12562 (diagnostic_show_locus): Adjust the use of
12563 location_get_source_line and adjust_line with respect to their new
12564 signature. While displaying a line now, do not stop at the first
12565 null byte. Rather, display the zero byte as a space and keep
12566 going until we reach the size of the line.
12567 * Makefile.in: Add vec.o to OBJS-libcommon
12569 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
12570 Ilya Tocar <ilya.tocar@intel.com>
12572 * config/i386/avx512fintrin.h (_mm512_kmov): New.
12573 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
12574 (__builtin_ia32_kmov16): Ditto.
12575 * config/i386/i386.md (UNSPEC_KMOV): New.
12578 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
12580 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
12581 (_mm512_storeu_si512): Ditto.
12583 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
12586 * rtl.h (get_referenced_operands): Declare.
12587 * recog.c (get_referenced_operands): New function.
12588 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
12589 operands have been referenced when recording LO_SUM references.
12591 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
12593 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
12595 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
12597 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
12598 Enable for generic and recent AMD targets.
12600 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
12602 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
12603 ARG_SIZE note when adjustment was eliminated.
12605 2014-01-22 Jeff Law <law@redhat.com>
12607 PR tree-optimization/59597
12608 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
12609 in file. Accept new argument REGISTERING and use it to modify
12610 dump output appropriately.
12611 (register_jump_thread): Corresponding changes.
12612 (mark_threaded_blocks): Reinstate code to cancel unprofitable
12613 thread paths involving joiner blocks. Add code to dump cancelled
12614 jump threading paths.
12616 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
12618 PR rtl-optimization/59477
12619 * lra-constraints.c (inherit_in_ebb): Process call for living hard
12620 regs. Update reloads_num and potential_reload_hard_regs for all insns.
12622 2014-01-22 Tom Tromey <tromey@redhat.com>
12624 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
12626 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
12628 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
12630 PR rtl-optimization/59896
12631 * lra-constraints.c (process_alt_operands): Check unused note for
12632 matched operands of insn with no output reloads.
12634 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
12636 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
12637 (mips_move_from_gpr_cost): Likewise.
12639 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
12641 PR rtl-optimization/59858
12642 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
12643 ira_class_hard_regs_num.
12644 (process_alt_operands): Increase reject for dying matched operand.
12646 2014-01-21 Jakub Jelinek <jakub@redhat.com>
12649 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
12650 smaller than size, perform several stores or loads and stores
12651 at dst + count - size to store or copy all of size bytes, rather
12652 than just last modesize bytes.
12654 2014-01-20 DJ Delorie <dj@redhat.com>
12656 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
12657 that CLOBBERs are REGs before propogating their values.
12659 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
12661 PR middle-end/59789
12662 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
12663 (cgraph_inline_failed_type): New function.
12664 * cgraph.h (DEFCIFCODE): Add type.
12665 (cgraph_inline_failed_type_t): New enum.
12666 (cgraph_inline_failed_type): New prototype.
12667 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
12668 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
12669 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
12670 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
12671 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
12672 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
12673 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
12674 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
12675 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
12676 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
12677 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
12678 OPTIMIZATION_MISMATCH.
12679 * tree-inline.c (expand_call_inline): Emit errors during
12680 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
12682 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
12685 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
12686 mode attribute in insn output.
12688 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
12690 * output.h (output_constant): Delete.
12691 * varasm.c (output_constant): Make private.
12693 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
12695 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
12697 2014-01-20 Jakub Jelinek <jakub@redhat.com>
12699 PR middle-end/59860
12700 * tree.h (fold_builtin_strcat): New prototype.
12701 * builtins.c (fold_builtin_strcat): No longer static. Add len
12702 argument, if non-NULL, don't call c_strlen. Optimize
12703 directly into __builtin_memcpy instead of __builtin_strcpy.
12704 (fold_builtin_2): Adjust fold_builtin_strcat caller.
12705 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
12707 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
12709 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
12710 for SImode_address_operand operands, having only a REG argument.
12712 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
12714 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
12715 loader name using mbig-endian.
12716 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
12718 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
12720 * doc/invoke.texi (-march): Clarify documentation for AArch64.
12721 (-mtune): Likewise.
12724 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
12726 * config/aarch64/aarch64-protos.h
12727 (aarch64_cannot_change_mode_class_ptr): Declare.
12728 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
12729 aarch64_cannot_change_mode_class_ptr): New.
12730 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
12731 backend hook aarch64_cannot_change_mode_class.
12733 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
12735 * common/config/aarch64/aarch64-common.c
12736 (aarch64_handle_option): Don't handle any option order logic here.
12737 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
12738 selected_cpu, warn on architecture version mismatch.
12739 (aarch64_override_options): Fix parsing order for option strings.
12741 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12742 Iain Sandoe <iain@codesourcery.com>
12745 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
12746 warning. Amend comment to reflect current functionality.
12748 2014-01-20 Richard Biener <rguenther@suse.de>
12750 PR middle-end/59860
12751 * builtins.c (fold_builtin_strcat): Remove case better handled
12752 by tree-ssa-strlen.c.
12754 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
12756 * config/aarch64/aarch64.opt
12757 (mcpu, march, mtune): Make case-insensitive.
12759 2014-01-20 Jakub Jelinek <jakub@redhat.com>
12762 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
12763 if operands[1] is a REG or ZERO_EXTEND of a REG.
12765 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
12767 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
12769 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
12771 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
12772 long non-pic millicode calls.
12774 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12776 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
12778 2014-01-19 Kito Cheng <kito@0xlab.org>
12780 * builtins.c (expand_movstr): Check movstr expand done or fail.
12782 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
12783 H.J. Lu <hongjiu.lu@intel.com>
12786 * config/i386/i386.md (*lea<mode>): Zero-extend return register
12787 to DImode for zero-extended addresses.
12789 2014-01-19 Jakub Jelinek <jakub@redhat.com>
12791 PR rtl-optimization/57763
12792 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
12793 on the new indirect jump_insn and increment LABEL_NUSES (label).
12795 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
12799 * config.gcc (x86_archs): New variable.
12800 (x86_64_archs): Likewise.
12801 (x86_cpus): Likewise.
12802 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
12803 --with-arch/--with-cpu= options.
12804 Support --with-arch=/--with-cpu={nehalem,westmere,
12805 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
12807 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
12809 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
12810 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
12812 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
12814 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
12816 2014-01-18 Jakub Jelinek <jakub@redhat.com>
12819 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
12820 clear cpp_get_options (parse_in)->warn_unused_macros for
12821 ix86_target_macros_internal with cpp_define.
12823 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
12825 * jump.c (delete_related_insns): Keep (use (insn))s.
12826 * reorg.c (redundant_insn): Check for barriers too.
12828 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
12830 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
12832 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
12834 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
12835 call to $$dyncall when TARGET_LONG_CALLS is true.
12837 2014-01-17 Jeff Law <law@redhat.com>
12839 * ree.c (combine_set_extension): Temporarily disable test for
12840 changing number of hard registers.
12842 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
12844 PR middle-end/58125
12845 * ipa-inline-analysis.c (inline_free_summary):
12846 Do not free summary of aliases.
12848 2014-01-17 Jakub Jelinek <jakub@redhat.com>
12850 PR middle-end/59706
12851 * gimplify.c (gimplify_expr): Use create_tmp_var
12852 instead of create_tmp_var_raw. If cond doesn't have
12853 integral type, don't add the IFN_ANNOTATE builtin at all.
12855 2014-01-17 Martin Jambor <mjambor@suse.cz>
12858 * ipa-cp.c (prev_edge_clone): New variable.
12859 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
12860 Also resize prev_edge_clone vector.
12861 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
12862 (ipcp_edge_removal_hook): New function.
12863 (ipcp_driver): Register ipcp_edge_removal_hook.
12865 2014-01-17 Andrew Pinski <apinski@cavium.com>
12866 Steve Ellcey <sellcey@mips.com>
12869 * config/mips/mips.c (mips_print_operand): Check operand mode instead
12872 2014-01-17 Jeff Law <law@redhat.com>
12874 PR middle-end/57904
12875 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
12876 so that pass_ccp runs first.
12878 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
12880 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
12881 (ix86_adjust_cost): Use !TARGET_XXX.
12882 (do_reorder_for_imul): Likewise.
12883 (swap_top_of_ready_list): Likewise.
12884 (ix86_sched_reorder): Likewise.
12886 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
12888 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
12889 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
12890 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
12891 (intel_memset): New. Duplicate slm_memset.
12892 (intel_cost): New. Duplicate slm_cost.
12893 (m_INTEL): New macro.
12894 (processor_target_table): Add "intel".
12895 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
12896 with PROCESSOR_INTEL for "intel".
12897 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
12898 PROCESSOR_SILVERMONT.
12899 (ix86_issue_rate): Likewise.
12900 (ix86_adjust_cost): Likewise.
12901 (ia32_multipass_dfa_lookahead): Likewise.
12902 (swap_top_of_ready_list): Likewise.
12903 (ix86_sched_reorder): Likewise.
12904 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
12905 instead of TARGET_OPT_AGU.
12906 * config/i386/i386.h (TARGET_INTEL): New.
12907 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
12908 (processor_type): Add PROCESSOR_INTEL.
12909 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
12910 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
12912 2014-01-17 Marek Polacek <polacek@redhat.com>
12915 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
12918 2014-01-17 Richard Biener <rguenther@suse.de>
12920 PR tree-optimization/46590
12921 * opts.c (default_options_table): Add entries for
12922 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
12923 all enabled at -O1 but not for -Og.
12924 * common.opt (fbranch-count-reg): Remove Init(1).
12925 (fmove-loop-invariants): Likewise.
12926 (ftree-pta): Likewise.
12928 2014-01-17 Jakub Jelinek <jakub@redhat.com>
12930 * config/i386/i386.c (ix86_data_alignment): For compatibility with
12931 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
12932 decls to at least the GCC 4.8 used alignments.
12935 * tree-nested.c (convert_nonlocal_reference_stmt,
12936 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
12937 of GIMPLE_BIND stmts, adjust associated decls.
12939 2014-01-17 Richard Biener <rguenther@suse.de>
12941 PR tree-optimization/46590
12942 * vec.h (vec<>::bseach): New member function implementing
12943 binary search according to C89 bsearch.
12944 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
12945 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
12946 bitmap pointer again. Make accesses_in_loop a flat array.
12947 (mem_ref_obstack): New global.
12948 (outermost_indep_loop): Adjust for mem_ref->stored changes.
12949 (mark_ref_stored): Likewise.
12950 (ref_indep_loop_p_2): Likewise.
12951 (set_ref_stored_in_loop): New helper function.
12952 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
12953 (memref_free): Adjust.
12954 (record_mem_ref_loc): Simplify.
12955 (gather_mem_refs_stmt): Adjust.
12956 (sort_locs_in_loop_postorder_cmp): New function.
12957 (analyze_memory_references): Sort accesses_in_loop after
12958 loop postorder number.
12959 (find_ref_loc_in_loop_cmp): New function.
12960 (for_all_locs_in_loop): Find relevant cluster of locs in
12961 accesses_in_loop and iterate without recursion.
12962 (execute_sm): Avoid uninit warning.
12963 (struct ref_always_accessed): Simplify.
12964 (ref_always_accessed::operator ()): Likewise.
12965 (ref_always_accessed_p): Likewise.
12966 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
12967 loop postorder numbers here.
12968 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
12971 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
12974 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
12975 on decls for which assemble_alias has been called.
12977 2014-01-17 Nick Clifton <nickc@redhat.com>
12979 * config/msp430/msp430.opt: (mcpu): New option.
12980 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
12981 (msp430_option_override): Parse target_cpu. If the MCU name
12982 matches a generic string, clear target_mcu.
12983 (msp430_attr): Allow numeric interrupt values up to 63.
12984 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
12985 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
12987 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
12989 * config/msp430/msp430.md (popm): Use %J rather than %I.
12990 (addsi3): Use msp430_nonimmediate_operand for operand 2.
12991 (addhi_cy_i): Use immediate_operand for operand 2.
12992 * doc/invoke.texi: Document -mcpu option.
12994 2014-01-17 Richard Biener <rguenther@suse.de>
12996 PR rtl-optimization/38518
12997 * df.h (df_analyze_loop): Declare.
12998 * df-core.c: Include cfgloop.h.
12999 (df_analyze_1): Split out main part of df_analyze.
13000 (df_analyze): Adjust.
13001 (loop_inverted_post_order_compute): New function.
13002 (loop_post_order_compute): Likewise.
13003 (df_analyze_loop): New function avoiding whole-function
13004 postorder computes.
13005 * loop-invariant.c (find_defs): Use df_analyze_loop.
13006 (find_invariants): Adjust.
13007 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
13009 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
13011 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
13012 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
13014 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
13016 * ipa-ref.c (ipa_remove_stmt_references): Fix references
13017 traversal when removing references.
13019 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
13022 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
13024 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
13026 PR middle-end/56791
13027 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
13028 pushing a reload for an autoinc when we had previously reloaded an
13029 inner part of the address.
13031 2014-01-16 Jakub Jelinek <jakub@redhat.com>
13033 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
13035 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
13036 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
13037 when not giving up or versioning for alias only because of
13039 (vect_analyze_data_ref_dependences): Set to true.
13040 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
13042 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
13043 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
13046 PR middle-end/58344
13047 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
13050 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
13051 operand 0 predicate for gathers, use a new pseudo as subtarget.
13053 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
13055 PR middle-end/59609
13056 * lra-constraints.c (process_alt_operands): Add printing debug info.
13057 Check absence of input/output reloads for matched operands too.
13059 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
13061 PR rtl-optimization/59835
13062 * ira.c (ira_init_register_move_cost): Increase cost for
13065 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
13067 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
13069 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
13072 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
13073 non-register objects. Use gen_(high/low)part more consistently.
13076 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
13079 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
13080 endian support, remove tests for WORDS_BIG_ENDIAN.
13081 (p8_mfvsrd_3_<mode>): Likewise.
13082 (reload_gpr_from_vsx<mode>): Likewise.
13083 (reload_gpr_from_vsxsf): Likewise.
13084 (p8_mfvsrd_4_disf): Likewise.
13086 2014-01-16 Richard Biener <rguenther@suse.de>
13088 PR rtl-optimization/46590
13089 * lcm.c (compute_antinout_edge): Use postorder iteration.
13090 (compute_laterin): Use inverted postorder iteration.
13092 2014-01-16 Nick Clifton <nickc@redhat.com>
13094 PR middle-end/28865
13095 * varasm.c (output_constant): Return the number of bytes actually
13097 (output_constructor_array_range): Update the field size with the
13098 number of bytes emitted by output_constant.
13099 (output_constructor_regular_field): Likewise. Also do not
13100 complain if the total number of bytes emitted is now greater
13101 than the expected fieldpos.
13102 * output.h (output_constant): Update prototype and descriptive comment.
13104 2014-01-16 Marek Polacek <polacek@redhat.com>
13106 PR middle-end/59827
13107 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
13108 it is error_mark_node.
13110 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
13112 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
13113 VALID_AVX256_REG_OR_OI_MODE.
13115 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
13117 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
13118 current procedure should be profiled.
13120 2014-01-15 Andrew Pinski <apinski@cavium.com>
13122 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
13123 of moving from/to the STACK_REG register class.
13125 2014-01-15 Richard Henderson <rth@redhat.com>
13128 * reginfo.c (global_regs_decl): Globalize.
13129 * rtl.h (global_regs_decl): Declare.
13130 * ira.c (do_reload): Diagnose frame_pointer_needed and it
13131 reserved via global_regs.
13133 2014-01-15 Teresa Johnson <tejohnson@google.com>
13135 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
13137 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
13139 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
13140 and vmulosh rather than call gen_vec_widen_smult_*.
13141 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
13142 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
13143 (vec_widen_smult_even_v16qi): Likewise.
13144 (vec_widen_umult_even_v8hi): Likewise.
13145 (vec_widen_smult_even_v8hi): Likewise.
13146 (vec_widen_umult_odd_v16qi): Likewise.
13147 (vec_widen_smult_odd_v16qi): Likewise.
13148 (vec_widen_umult_odd_v8hi): Likewise.
13149 (vec_widen_smult_odd_v8hi): Likewise.
13150 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
13151 vmuloub rather than call gen_vec_widen_umult_*.
13152 (vec_widen_umult_lo_v16qi): Likewise.
13153 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
13154 vmulosb rather than call gen_vec_widen_smult_*.
13155 (vec_widen_smult_lo_v16qi): Likewise.
13156 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
13157 rather than call gen_vec_widen_umult_*.
13158 (vec_widen_umult_lo_v8hi): Likewise.
13159 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
13160 rather than call gen_vec_widen_smult_*.
13161 (vec_widen_smult_lo_v8hi): Likewise.
13163 2014-01-15 Jeff Law <law@redhat.com>
13165 PR tree-optimization/59747
13166 * ree.c (find_and_remove_re): Properly handle case where a second
13167 eliminated extension requires widening a copy created for elimination
13168 of a prior extension.
13169 (combine_set_extension): Ensure that the number of hard regs needed
13170 for a destination register does not change when we widen it.
13172 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
13174 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
13175 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
13176 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
13177 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
13178 (avr-*-rtems*): Likewise.
13179 (bfin*-rtems*): Likewise.
13180 (moxie-*-rtems*): Likewise.
13181 (h8300-*-rtems*): Likewise.
13182 (i[34567]86-*-rtems*): Likewise.
13183 (lm32-*-rtems*): Likewise.
13184 (m32r-*-rtems*): Likewise.
13185 (m68k-*-rtems*): Likewise.
13186 (microblaze*-*-rtems*): Likewise.
13187 (mips*-*-rtems*): Likewise.
13188 (powerpc-*-rtems*): Likewise.
13189 (sh-*-rtems*): Likewise.
13190 (sparc-*-rtems*): Likewise.
13191 (sparc64-*-rtems*): Likewise.
13192 (v850-*-rtems*): Likewise.
13193 (m32c-*-rtems*): Likewise.
13195 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
13197 PR rtl-optimization/59511
13198 * ira.c (ira_init_register_move_cost): Use memory costs for some
13199 cases of register move cost calculations.
13200 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
13201 instead of BB frequency.
13202 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
13203 * lra-assigns.c (find_hard_regno_for): Ditto.
13205 2014-01-15 Richard Biener <rguenther@suse.de>
13207 PR tree-optimization/59822
13208 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
13209 (vectorizable_load): Use it to hoist defs of uses of invariant
13210 loads out of the loop.
13212 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
13213 Kugan Vivekanandarajah <kuganv@linaro.org>
13216 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
13219 2014-01-15 Richard Biener <rguenther@suse.de>
13221 PR rtl-optimization/59802
13222 * lcm.c (compute_available): Use inverted postorder to seed
13223 the initial worklist.
13225 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13228 * config/s390/s390.c (s390_preferred_reload_class): Don't return
13229 ADDR_REGS for invalid symrefs in non-PIC code.
13231 2014-01-15 Jakub Jelinek <jakub@redhat.com>
13234 * builtins.c (determine_block_size): Initialize *probable_max_size
13235 even if len_rtx is CONST_INT.
13237 2014-01-14 Andrew Pinski <apinski@cavium.com>
13239 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
13240 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
13241 (cortexa53_tunings): Likewise.
13242 (aarch64_sched_issue_rate): New function.
13243 (TARGET_SCHED_ISSUE_RATE): Define.
13245 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
13247 * ira-costs.c (find_costs_and_classes): Add missed
13248 ira_init_register_move_cost_if_necessary.
13250 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
13253 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
13255 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
13258 * config/i386/i386.c (type_natural_mode): Add a bool parameter
13259 to indicate if type is used for function return value. Warn ABI
13260 change if the vector mode isn't available for function return value.
13261 (ix86_function_arg_advance): Pass false to type_natural_mode.
13262 (ix86_function_arg): Likewise.
13263 (ix86_gimplify_va_arg): Likewise.
13264 (function_arg_32): Don't warn ABI change.
13265 (ix86_function_value): Pass true to type_natural_mode.
13266 (ix86_return_in_memory): Likewise.
13267 (ix86_struct_value_rtx): Removed.
13268 (TARGET_STRUCT_VALUE_RTX): Likewise.
13270 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
13272 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
13273 converting a conditional jump into a conditional return.
13275 2014-01-14 Richard Biener <rguenther@suse.de>
13277 PR tree-optimization/58921
13278 PR tree-optimization/59006
13279 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
13280 hoisting invariant stmts.
13281 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
13282 invariant loads on the preheader edge if possible.
13284 2014-01-14 Joey Ye <joey.ye@arm.com>
13286 * doc/plugin.texi (Building GCC plugins): Update to C++.
13288 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
13290 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
13291 (_mm_rcp28_round_ss): Ditto.
13292 (_mm_rsqrt28_round_sd): Ditto.
13293 (_mm_rsqrt28_round_ss): Ditto.
13294 (_mm_rcp28_sd): Ditto.
13295 (_mm_rcp28_ss): Ditto.
13296 (_mm_rsqrt28_sd): Ditto.
13297 (_mm_rsqrt28_ss): Ditto.
13298 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
13299 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
13300 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
13301 (IX86_BUILTIN_RCP28SD): Ditto.
13302 (IX86_BUILTIN_RCP28SS): Ditto.
13303 (IX86_BUILTIN_RSQRT28SD): Ditto.
13304 (IX86_BUILTIN_RSQRT28SS): Ditto.
13305 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
13306 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
13307 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
13308 (ix86_expand_special_args_builtin): Expand new FTYPE.
13309 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
13310 (srcp14<mode>): Make insn unary.
13311 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
13312 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
13313 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
13314 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
13315 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
13316 Fix rounding: make it SAE only.
13317 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
13319 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
13321 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
13322 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
13323 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
13324 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
13325 (round_saeonly_mask_scalar_operand4): Ditto.
13326 (round_saeonly_mask_scalar_op3): Ditto.
13327 (round_saeonly_mask_scalar_op4): Ditto.
13329 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13331 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
13332 Implement -maltivec=be for vec_insert and vec_extract.
13334 2014-01-10 DJ Delorie <dj@redhat.com>
13336 * config/msp430/msp430.md (call_internal): Don't allow memory
13337 references with SP as the base register.
13338 (call_value_internal): Likewise.
13339 * config/msp430/constraints.md (Yc): New. For memory references
13340 that don't use SP as a base register.
13342 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
13343 "an integer without a # prefix"
13344 * config/msp430/msp430.md (epilogue_helper): Use it.
13346 2014-01-13 Jakub Jelinek <jakub@redhat.com>
13349 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
13350 AVX512F gather builtins.
13351 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
13352 on gather decls with INTEGER_TYPE masktype.
13353 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
13354 directly into the builtin rather than hoisting it before loop.
13356 PR tree-optimization/59387
13357 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
13358 (scev_const_prop): If folded_casts and type has undefined overflow,
13359 use force_gimple_operand instead of force_gimple_operand_gsi and
13360 for each added stmt if it is assign with
13361 arith_code_with_undefined_signed_overflow, call
13362 rewrite_to_defined_overflow.
13363 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
13364 gimple-fold.h instead.
13365 (arith_code_with_undefined_signed_overflow,
13366 rewrite_to_defined_overflow): Moved to ...
13367 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
13368 rewrite_to_defined_overflow): ... here. No longer static.
13369 Include gimplify-me.h.
13370 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
13371 rewrite_to_defined_overflow): New prototypes.
13373 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13375 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
13377 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
13379 * builtins.c (get_object_alignment_2): Minor tweak.
13380 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
13382 2014-01-13 Christian Bruel <christian.bruel@st.com>
13384 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
13385 optimized non constant lengths.
13387 2014-01-13 Jakub Jelinek <jakub@redhat.com>
13390 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
13391 load as __atomic_load_N if possible.
13393 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
13395 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
13397 (rs6000_expand_builtin): Adjust call.
13399 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
13402 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
13403 * config/rs6000/rs6000.c: Include target-globals.h.
13404 (rs6000_set_current_function): Instead of doing target_reinit
13405 unconditionally, use save_target_globals_default_opts and
13406 restore_target_globals.
13408 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
13410 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
13411 (rs6000_expand_builtin): Handle mffs and mtfsf.
13412 (rs6000_init_builtins): Define mffs and mtfsf.
13413 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
13414 (rs6000_mffs): New pattern.
13415 (rs6000_mtfsf): New pattern.
13417 2014-01-11 Bin Cheng <bin.cheng@arm.com>
13419 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
13420 Start narrowing with START. Apply candidate-use pair
13421 and check overall cost in narrowing.
13422 (iv_ca_prune): Pass new argument.
13424 2014-01-10 Jeff Law <law@redhat.com>
13426 PR middle-end/59743
13427 * ree.c (combine_reaching_defs): Ensure the defining statement
13428 occurs before the extension when optimizing extensions with
13429 different source and destination hard registers.
13431 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
13434 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
13435 vtables into the type inheritance graph.
13437 2014-01-10 Jakub Jelinek <jakub@redhat.com>
13439 PR rtl-optimization/59754
13440 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
13441 modes in the REGNO != REGNO case.
13443 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13445 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
13447 2014-01-10 Jakub Jelinek <jakub@redhat.com>
13449 PR tree-optimization/59745
13450 * tree-predcom.c (tree_predictive_commoning_loop): Call
13451 free_affine_expand_cache if giving up because components is NULL.
13453 * target-globals.c (save_target_globals): Allocate < 4KB structs using
13454 GC in payload of target_globals struct instead of allocating them on
13455 the heap and the larger structs separately using GC.
13456 * target-globals.h (struct target_globals): Make regs, hard_regs,
13457 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
13458 of GTY((skip)) and change type to void *.
13459 (reset_target_globals): Cast loads from those fields to corresponding
13462 2014-01-10 Steve Ellcey <sellcey@mips.com>
13465 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
13466 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
13467 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
13469 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
13472 * aarch64-modes.def (CC_Zmode): New flags mode.
13473 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
13474 represents an equality.
13475 (aarch64_get_condition_code): Handle CC_Zmode.
13476 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
13478 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13480 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
13481 extraction in good case.
13483 2014-01-10 Richard Biener <rguenther@suse.de>
13485 PR tree-optimization/59374
13486 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
13487 checking after SLP discovery. Mark stmts not participating
13488 in any SLP instance properly.
13490 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13492 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
13493 when handling a SET rtx.
13495 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13497 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
13498 (cortex-a57): Likewise.
13499 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
13501 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13503 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
13504 non-iwmmxt builtins.
13506 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
13510 * ipa-devirt.c record_target_from_binfo): Take as argument
13511 stack of binfos and lookup matching one for virtual inheritance.
13512 (possible_polymorphic_call_targets_1): Update.
13514 2014-01-10 Huacai Chen <chenhc@lemote.com>
13516 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
13517 kernel strings for Loongson-2E/2F/3A.
13519 2014-01-10 Jakub Jelinek <jakub@redhat.com>
13521 PR middle-end/59670
13522 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
13523 is_gimple_call before calling gimple_call_internal_p.
13525 2014-01-09 Steve Ellcey <sellcey@mips.com>
13527 * Makefile.in (TREE_FLOW_H): Remove.
13528 (TREE_SSA_H): Add file names from tree-flow.h.
13529 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
13530 * tree.h: Remove tree-flow.h reference.
13531 * hash-table.h: Remove tree-flow.h reference.
13532 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
13533 reference with tree-ssa-loop.h.
13535 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13537 * doc/invoke.texi: Add -maltivec={be,le} options, and document
13538 default element-order behavior for -maltivec.
13539 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
13540 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
13541 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
13542 when targeting big endian, at least for now.
13543 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
13545 2014-01-09 Jakub Jelinek <jakub@redhat.com>
13547 PR middle-end/47735
13548 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
13549 var satisfies use_register_for_decl, just take into account type
13550 alignment, rather than decl alignment.
13552 PR tree-optimization/59622
13553 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
13554 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
13555 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
13556 Don't devirtualize for inplace at all. For targets.length () == 1,
13557 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
13559 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
13561 * config/i386/i386.md (cpu): Remove the unused btver1.
13563 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
13565 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
13567 2014-01-09 Jakub Jelinek <jakub@redhat.com>
13570 * tree-core.h (struct target_globals): New forward declaration.
13571 (struct tree_target_option): Add globals field.
13572 * tree.h (TREE_TARGET_GLOBALS): Define.
13573 (prepare_target_option_nodes_for_pch): New prototype.
13574 * target-globals.h (struct target_globals): Define even if
13575 !SWITCHABLE_TARGET.
13576 * tree.c (prepare_target_option_node_for_pch,
13577 prepare_target_option_nodes_for_pch): New functions.
13578 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
13579 * config/i386/i386.c: Include target-globals.h.
13580 (ix86_set_current_function): Instead of doing target_reinit
13581 unconditionally, use save_target_globals_default_opts and
13582 restore_target_globals.
13584 2014-01-09 Richard Biener <rguenther@suse.de>
13586 PR tree-optimization/59715
13587 * tree-cfg.h (split_critical_edges): Declare.
13588 * tree-cfg.c (split_critical_edges): Export.
13589 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
13591 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
13593 * cfgexpand.c (expand_stack_vars): Optionally disable
13594 asan stack protection.
13595 (expand_used_vars): Likewise.
13596 (partition_stack_vars): Likewise.
13597 * asan.c (asan_emit_stack_protection): Optionally disable
13598 after return stack usage.
13599 (instrument_derefs): Optionally disable memory access instrumentation.
13600 (instrument_builtin_call): Likewise.
13601 (instrument_strlen_call): Likewise.
13602 (asan_protect_global): Optionally disable global variables protection.
13603 * doc/invoke.texi: Added doc for new options.
13604 * params.def: Added new options.
13605 * params.h: Likewise.
13607 2014-01-09 Jakub Jelinek <jakub@redhat.com>
13609 PR rtl-optimization/59724
13610 * ifcvt.c (cond_exec_process_if_block): Don't call
13611 flow_find_head_matching_sequence with 0 longest_match.
13612 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
13613 non-active insns if !stop_after.
13614 (try_head_merge_bb): Revert 2014-01-07 changes.
13616 2014-01-08 Jeff Law <law@redhat.com>
13618 * ree.c (get_sub_rtx): New function, extracted from...
13619 (merge_def_and_ext): Here.
13620 (combine_reaching_defs): Use get_sub_rtx.
13622 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
13624 * cgraph.h (varpool_variable_node): Do not choke on null node.
13626 2014-01-08 Catherine Moore <clm@codesourcery.com>
13628 * config/mips/mips.md (simple_return): Attempt to use JRC
13630 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
13632 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
13634 PR rtl-optimization/59137
13635 * reorg.c (steal_delay_list_from_target): Call update_block for
13637 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
13639 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13641 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
13642 two duplicate entries.
13644 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
13647 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
13649 * config/mips/mips.c (mips_truncated_op_cost): New function.
13650 (mips_rtx_costs): Adjust test for BADDU.
13651 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
13653 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
13655 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
13656 (*baddu_si): ...this new pattern.
13658 2014-01-08 Jakub Jelinek <jakub@redhat.com>
13661 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
13663 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
13665 PR middle-end/57748
13666 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
13668 (expand_expr, expand_normal): Adjust.
13669 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
13670 inner_reference_p. Use inner_reference_p to expand inner references.
13671 (store_expr): Adjust.
13672 * cfgexpand.c (expand_call_stmt): Adjust.
13674 2014-01-08 Rong Xu <xur@google.com>
13676 * gcov-io.c (gcov_var): Move from gcov-io.h.
13677 (gcov_position): Ditto.
13678 (gcov_is_error): Ditto.
13679 (gcov_rewrite): Ditto.
13680 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
13681 only part to libgcc/libgcov.h.
13683 2014-01-08 Marek Polacek <polacek@redhat.com>
13685 PR middle-end/59669
13686 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
13688 2014-01-08 Marek Polacek <polacek@redhat.com>
13691 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
13693 2014-01-08 Jakub Jelinek <jakub@redhat.com>
13695 PR rtl-optimization/59649
13696 * stor-layout.c (get_mode_bounds): For BImode return
13697 0 and STORE_FLAG_VALUE.
13699 2014-01-08 Richard Biener <rguenther@suse.de>
13701 PR middle-end/59630
13702 * gimple.h (is_gimple_builtin_call): Remove.
13703 (gimple_builtin_call_types_compatible_p): New.
13704 (gimple_call_builtin_p): New overload.
13705 * gimple.c (is_gimple_builtin_call): Remove.
13706 (validate_call): Rename to ...
13707 (gimple_builtin_call_types_compatible_p): ... this and export. Also
13708 check return types.
13709 (validate_type): New static function.
13710 (gimple_call_builtin_p): New overload and adjust.
13711 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
13712 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
13713 (gimple_fold_stmt_to_constant_1): Likewise.
13714 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
13716 2014-01-08 Richard Biener <rguenther@suse.de>
13718 PR middle-end/59471
13719 * gimplify.c (gimplify_expr): Gimplify register-register type
13720 VIEW_CONVERT_EXPRs to separate stmts.
13722 2014-01-07 Jeff Law <law@redhat.com>
13724 PR middle-end/53623
13725 * ree.c (combine_set_extension): Handle case where source
13726 and destination registers in an extension insn are different.
13727 (combine_reaching_defs): Allow source and destination registers
13728 in extension to be different under limited circumstances.
13729 (add_removable_extension): Remove restriction that the
13730 source and destination registers in the extension are the same.
13731 (find_and_remove_re): Emit a copy from the extension's
13732 destination to its source after the defining insn if
13733 the source and destination registers are different.
13735 PR middle-end/59285
13736 * ifcvt.c (merge_if_block): If we are merging a block with more than
13737 one successor with a block with no successors, remove any BARRIER
13738 after the second block.
13740 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
13742 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
13744 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
13747 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
13748 for 14-bit register offsets when INT14_OK_STRICT is false.
13750 2014-01-07 Roland Stigge <stigge@antcom.de>
13751 Michael Meissner <meissner@linux.vnet.ibm.com>
13754 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
13755 Only check TFmode for SPE constants. Don't check TImode or TDmode.
13757 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
13759 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
13762 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
13764 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
13765 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
13766 rtx is const0_rtx or not.
13768 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
13771 * target-globals.c (save_target_globals): Remove this_fn_optab
13773 * toplev.c: Include optabs.h.
13774 (target_reinit): Temporarily restore the global options if another
13775 set of options are in force.
13777 2014-01-07 Jakub Jelinek <jakub@redhat.com>
13779 PR rtl-optimization/58668
13780 * cfgcleanup.c (flow_find_cross_jump): Don't count
13781 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
13782 to determine what is counted.
13783 (flow_find_head_matching_sequence): Use active_insn_p to determine
13785 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
13787 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
13788 determine what is counted.
13790 PR tree-optimization/59643
13791 * tree-predcom.c (split_data_refs_to_components): If one dr is
13792 read and one write, determine_offset fails and the write isn't
13793 in the bad component, just put the read into the bad component.
13795 2014-01-07 Mike Stump <mikestump@comcast.net>
13796 Jakub Jelinek <jakub@redhat.com>
13799 * tree-core.h (struct tree_optimization_option): Change optabs
13800 type from unsigned char * to void *.
13801 * optabs.c (init_tree_optimization_optabs): Adjust
13802 TREE_OPTIMIZATION_OPTABS initialization.
13804 2014-01-06 Jakub Jelinek <jakub@redhat.com>
13807 * config/i386/i386.h (struct machine_function): Add
13808 no_drap_save_restore field.
13809 * config/i386/i386.c (ix86_save_reg): Use
13810 !cfun->machine->no_drap_save_restore instead of
13811 crtl->stack_realign_needed.
13812 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
13813 this function clears frame_pointer_needed. Set
13814 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
13815 and DRAP reg is needed.
13817 2014-01-06 Marek Polacek <polacek@redhat.com>
13820 * doc/implement-c.texi: Mention that other integer types are
13821 permitted as bit-field types in strictly conforming mode.
13823 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
13825 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
13826 is newly allocated.
13828 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
13830 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
13832 2014-01-06 Martin Jambor <mjambor@suse.cz>
13835 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
13837 * ipa-prop.c (ipa_print_node_params): Fix indentation.
13839 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
13843 * var-tracking.c (add_stores): Preserve the value of the source even if
13844 we don't record the store.
13846 2014-01-06 Terry Guo <terry.guo@arm.com>
13848 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
13850 2014-01-05 Iain Sandoe <iain@codesourcery.com>
13853 * config/darwin.c (darwin_function_section): Adjust return values to
13854 correspond to optimisation changes made in r206070.
13856 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
13858 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
13859 from prefetch_block tune setting.
13860 (nocona_cost): Correct size of prefetch block to 64.
13862 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
13864 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
13865 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
13866 used to save the static chain register in the computation of the offset
13867 from which the FP registers need to be restored.
13869 2014-01-04 Jakub Jelinek <jakub@redhat.com>
13871 PR tree-optimization/59519
13872 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
13873 ICE if get_current_def (current_new_name) is already non-NULL, as long
13874 as it is a phi result of some other phi in *new_exit_bb that has
13877 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
13878 or vmovdqu* for misaligned_operand.
13879 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
13880 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
13881 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
13882 aligned_mem for AVX512F masked aligned load and store builtins and for
13883 non-temporal moves.
13885 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
13887 PR tree-optimization/59651
13888 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
13889 Address range for negative step should be added by TYPE_SIZE_UNIT.
13891 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
13893 * config/m68k/m68k.c (handle_move_double): Handle pushes with
13894 overlapping registers also for registers other than the stack pointer.
13896 2014-01-03 Marek Polacek <polacek@redhat.com>
13899 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
13902 2014-01-03 Jakub Jelinek <jakub@redhat.com>
13905 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
13908 * config/i386/i386.md (MODE_SIZE): New mode attribute.
13909 (push splitter): Use <P:MODE_SIZE> instead of
13910 GET_MODE_SIZE (<P:MODE>mode).
13911 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
13912 (mov -1, reg peephole2): Likewise.
13913 * config/i386/sse.md (*mov<mode>_internal,
13914 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
13915 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
13916 *<code><mode>3, *andnot<mode>3<mask_name>,
13917 <mask_codefor><code><mode>3<mask_name>): Likewise.
13918 * config/i386/subst.md (mask_mode512bit_condition,
13919 sd_mask_mode512bit_condition): Likewise.
13921 2014-01-02 Xinliang David Li <davidxl@google.com>
13923 PR tree-optimization/59303
13924 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
13925 (dump_predicates): Better output format.
13926 (pred_equal_p): New function.
13927 (is_neq_relop_p): Ditto.
13928 (is_neq_zero_form_p): Ditto.
13929 (pred_expr_equal_p): Ditto.
13930 (pred_neg_p): Ditto.
13931 (simplify_pred): Ditto.
13932 (simplify_preds_2): Ditto.
13933 (simplify_preds_3): Ditto.
13934 (simplify_preds_4): Ditto.
13935 (simplify_preds): Ditto.
13936 (push_pred): Ditto.
13937 (push_to_worklist): Ditto.
13938 (get_pred_info_from_cmp): Ditto.
13939 (is_degenerated_phi): Ditto.
13940 (normalize_one_pred_1): Ditto.
13941 (normalize_one_pred): Ditto.
13942 (normalize_one_pred_chain): Ditto.
13943 (normalize_preds): Ditto.
13944 (normalize_cond_1): Remove function.
13945 (normalize_cond): Ditto.
13946 (is_gcond_subset_of): Ditto.
13947 (is_subset_of_any): Ditto.
13948 (is_or_set_subset_of): Ditto.
13949 (is_and_set_subset_of): Ditto.
13950 (is_norm_cond_subset_of): Ditto.
13951 (pred_chain_length_cmp): Ditto.
13952 (convert_control_dep_chain_into_preds): Type change.
13953 (find_predicates): Ditto.
13954 (find_def_preds): Ditto.
13955 (destroy_predicates_vecs): Ditto.
13956 (find_matching_predicates_in_rest_chains): Ditto.
13957 (use_pred_not_overlap_with_undef_path_pred): Ditto.
13958 (is_pred_expr_subset): Ditto.
13959 (is_pred_chain_subset_of): Ditto.
13960 (is_included_in): Ditto.
13961 (is_superset_of): Ditto.
13963 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
13965 Update copyright years.
13967 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
13969 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
13970 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
13971 config/arc/arc.md, config/arc/arc.opt,
13972 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
13973 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
13974 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
13975 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
13976 config/linux-protos.h, config/linux.c, config/winnt-c.c,
13977 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
13978 vtable-verify.c, vtable-verify.h: Use the standard form for the
13981 2014-01-02 Tobias Burnus <burnus@net-b.de>
13983 * gcc.c (process_command): Update copyright notice dates.
13984 * gcov-dump.c: Ditto.
13986 * doc/cpp.texi: Bump @copying's copyright year.
13987 * doc/cppinternals.texi: Ditto.
13988 * doc/gcc.texi: Ditto.
13989 * doc/gccint.texi: Ditto.
13990 * doc/gcov.texi: Ditto.
13991 * doc/install.texi: Ditto.
13992 * doc/invoke.texi: Ditto.
13994 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
13996 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
13998 2014-01-01 Jakub Jelinek <jakub@redhat.com>
14000 * config/i386/sse.md (*mov<mode>_internal): Guard
14001 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
14003 PR rtl-optimization/59647
14004 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
14005 new_rtx into UNSIGNED_FLOAT rtxes.
14007 Copyright (C) 2014 Free Software Foundation, Inc.
14009 Copying and distribution of this file, with or without modification,
14010 are permitted in any medium without royalty provided the copyright
14011 notice and this notice are preserved.