1 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
3 * alias.c (record_alias_subset): Adjust.
4 * bitmap.c (bitmap_element_allocate): Likewise.
5 (bitmap_gc_alloc_stat): Likewise.
6 * cfg.c (init_flow): Likewise.
7 (alloc_block): Likewise.
8 (unchecked_make_edge): Likewise.
9 * cfgloop.c (alloc_loop): Likewise.
10 (flow_loops_find): Likewise.
11 (rescan_loop_exit): Likewise.
12 * cfgrtl.c (init_rtl_bb_info): Likewise.
13 * cgraph.c (insert_new_cgraph_node_version): Likewise.
14 (cgraph_allocate_node): Likewise.
15 (cgraph_create_edge_1): Likewise.
16 (cgraph_allocate_init_indirect_info): Likewise.
17 * cgraphclones.c (cgraph_clone_edge): Likewise.
18 * cgraphunit.c (add_asm_node): Likewise.
19 (init_lowered_empty_function): Likewise.
20 * config/aarch64/aarch64.c (aarch64_init_machine_status):
22 * config/alpha/alpha.c (alpha_init_machine_status): Likewise.
23 (alpha_use_linkage): Likewise.
24 * config/arc/arc.c (arc_init_machine_status): Likewise.
25 * config/arm/arm.c (arm_init_machine_status): Likewise.
26 * config/avr/avr.c (avr_init_machine_status): Likewise.
27 * config/bfin/bfin.c (bfin_init_machine_status): Likewise.
28 * config/c6x/c6x.c (c6x_init_machine_status): Likewise.
29 * config/cris/cris.c (cris_init_machine_status): Likewise.
30 * config/darwin.c (machopic_indirection_name): Likewise.
31 (darwin_build_constant_cfstring): Likewise.
32 (darwin_enter_string_into_cfstring_table): Likewise.
33 * config/epiphany/epiphany.c (epiphany_init_machine_status):
35 * config/frv/frv.c (frv_init_machine_status): Likewise.
36 * config/i386/i386.c (get_dllimport_decl): Likewise.
37 (ix86_init_machine_status): Likewise.
38 (assign_386_stack_local): Likewise.
39 * config/i386/winnt.c (i386_pe_record_external_function):
41 (i386_pe_maybe_record_exported_symbol): Likewise.
42 (i386_pe_record_stub): Likewise.
43 * config/ia64/ia64.c (ia64_init_machine_status): Likewise.
44 * config/iq2000/iq2000.c (iq2000_init_machine_status): Likewise.
45 * config/m32c/m32c.c (m32c_init_machine_status): Likewise.
46 (m32c_note_pragma_address): Likewise.
47 * config/mep/mep.c (mep_init_machine_status): Likewise.
48 (mep_note_pragma_flag): Likewise.
49 * config/mips/mips.c (mflip_mips16_use_mips16_p): Likewise.
50 (mips16_local_alias): Likewise.
51 (mips_init_machine_status): Likewise.
52 * config/mmix/mmix.c (mmix_init_machine_status): Likewise.
53 * config/moxie/moxie.c (moxie_init_machine_status): Likewise.
54 * config/msp430/msp430.c (msp430_init_machine_status): Likewise.
55 * config/nds32/nds32.c (nds32_init_machine_status): Likewise.
56 * config/nios2/nios2.c (nios2_init_machine_status): Likewise.
57 * config/pa/pa.c (pa_init_machine_status): Likewise.
58 (pa_get_deferred_plabel): Likewise.
59 * config/rl78/rl78.c (rl78_init_machine_status): Likewise.
60 * config/rs6000/rs6000.c (builtin_function_type): Likewise.
61 (rs6000_init_machine_status): Likewise.
62 (output_toc): Likewise.
63 * config/s390/s390.c (s390_init_machine_status): Likewise.
64 * config/score/score.c (score_output_external): Likewise.
65 * config/sparc/sparc.c (sparc_init_machine_status): Likewise.
66 * config/spu/spu.c (spu_init_machine_status): Likewise.
67 * config/tilegx/tilegx.c (tilegx_init_machine_status): Likewise.
68 * config/tilepro/tilepro.c (tilepro_init_machine_status):
70 * config/xtensa/xtensa.c (xtensa_init_machine_status): Likewise.
71 * coverage.c (coverage_end_function): Likewise.
72 * dbxout.c (dbxout_init): Likewise.
73 * doc/gty.texi: Don't mention variable_size attribute.
74 * dwarf2cfi.c (new_cfi): Adjust.
75 (new_cfi_row): Likewise.
76 (copy_cfi_row): Likewise.
77 (create_cie_data): Likewise.
78 * dwarf2out.c (dwarf2out_alloc_current_fde): Likewise.
79 (new_loc_descr): Likewise.
80 (find_AT_string_in_table): Likewise.
81 (add_addr_table_entry): Likewise.
83 (add_var_loc_to_decl): Likewise.
84 (clone_die): Likewise.
85 (clone_as_declaration): Likewise.
86 (break_out_comdat_types): Likewise.
87 (new_loc_list): Likewise.
88 (add_loc_descr_to_each): Likewise.
89 (add_location_or_const_value_attribute): Likewise.
90 (add_linkage_name): Likewise.
91 (lookup_filename): Likewise.
92 (dwarf2out_var_location): Likewise.
93 (new_line_info_table): Likewise.
94 (dwarf2out_init): Likewise.
95 (mem_loc_descriptor): Likewise.
96 (loc_descriptor): Likewise.
97 (add_const_value_attribute): Likewise.
98 (tree_add_const_value_attribute): Likewise.
99 (comp_dir_string): Likewise.
100 (dwarf2out_vms_debug_main_pointer): Likewise.
101 (string_cst_pool_decl): Likewise.
102 * emit-rtl.c (set_mem_attrs): Likewise.
103 (get_reg_attrs): Likewise.
104 (start_sequence): Likewise.
105 (init_emit): Likewise.
106 (init_emit_regs): Likewise.
107 * except.c (init_eh_for_function): Likewise.
108 (gen_eh_region): Likewise.
109 (gen_eh_region_catch): Likewise.
110 (gen_eh_landing_pad): Likewise.
111 (add_call_site): Likewise.
112 * function.c (add_frame_space): Likewise.
113 (insert_temp_slot_address): Likewise.
114 (assign_stack_temp_for_type): Likewise.
115 (get_hard_reg_initial_val): Likewise.
116 (allocate_struct_function): Likewise.
117 (prepare_function_start): Likewise.
118 (types_used_by_var_decl_insert): Likewise.
119 * gengtype.c (variable_size_p): Remove function.
120 (enum alloc_quantity): Remove enum.
121 (write_typed_alloc_def): Remove function.
122 (write_typed_struct_alloc_def): Likewise.
123 (write_typed_typedef_alloc_def): Likewise.
124 (write_typed_alloc_defns): Likewise.
126 * ggc-common.c (ggc_cleared_alloc_htab_ignore_args): Adjust.
127 (ggc_cleared_alloc_ptr_array_two_args): Likewise.
128 * ggc.h (ggc_alloc): new function.
129 (ggc_cleared_alloc): Likewise.
130 (ggc_vec_alloc): Template on type of vector element, and remove
131 element size argument.
132 (ggc_cleared_vec_alloc): Likewise.
133 * gimple.c (gimple_build_omp_for): Adjust.
134 (gimple_copy): Likewise.
135 * ipa-cp.c (get_replacement_map): Likewise.
136 (find_aggregate_values_for_callers_subset): Likewise.
137 (known_aggs_to_agg_replacement_list): Likewise.
138 * ipa-devirt.c (get_odr_type): Likewise.
139 * ipa-prop.c (ipa_node_duplication_hook): Likewise.
140 (read_agg_replacement_chain): Likewise.
141 * loop-iv.c (get_simple_loop_desc): Likewise.
142 * lto-cgraph.c (input_node_opt_summary): Likewise.
143 * lto-section-in.c (lto_new_in_decl_state): Likewise.
144 * lto-streamer-in.c (lto_input_eh_catch_list): Likewise.
145 (input_eh_region): Likewise.
146 (input_eh_lp): Likewise.
147 (input_cfg): Likewise.
148 * optabs.c (set_optab_libfunc): Likewise.
149 (init_tree_optimization_optabs): Likewise.
150 (set_conv_libfunc): Likewise.
151 * passes.c (do_per_function_toporder): Likewise.
152 * rtl.h: Don't use variable_size gty attribute.
153 * sese.c (if_region_set_false_region): Adjust.
154 * stringpool.c (gt_pch_save_stringpool): Likewise.
155 * target-globals.c (save_target_globals): Likewise.
156 * toplev.c (general_init): Likewise.
157 * trans-mem.c (record_tm_replacement): Likewise.
158 (split_bb_make_tm_edge): Likewise.
159 * tree-cfg.c (move_sese_region_to_fn): Likewise.
160 * tree-data-ref.h (lambda_vector_new): Likewise.
161 * tree-eh.c (add_stmt_to_eh_lp_fn): Likewise.
162 * tree-iterator.c (tsi_link_before): Likewise.
163 (tsi_link_after): Likewise.
164 * tree-scalar-evolution.c (new_scev_info_str): Likewise.
165 * tree-ssa-loop-niter.c (record_estimate): Likewise.
166 * tree-ssa-operands.c (ssa_operand_alloc): Likewise.
167 * tree-ssa-operands.h: Don't use variable_size gty attribute.
168 * tree-ssa.c (init_tree_ssa): Adjust.
169 * tree-ssanames.c (set_range_info): Likewise.
170 (get_ptr_info): Likewise.
171 (duplicate_ssa_name_ptr_info): Likewise.
172 (duplicate_ssa_name_range_info): Likewise.
173 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Likewise.
174 (unpack_ts_fixed_cst_value_fields): Likewise.
175 * tree.c (build_fixed): Likewise.
176 (build_real): Likewise.
177 (build_string): Likewise.
178 (decl_priority_info): Likewise.
179 (decl_debug_expr_insert): Likewise.
180 (decl_value_expr_insert): Likewise.
181 (decl_debug_args_insert): Likewise.
182 (type_hash_add): Likewise.
183 (build_omp_clause): Likewise.
184 * ubsan.c (decl_for_type_insert): Likewise.
185 * varasm.c (get_unnamed_section): Likewise.
186 (get_noswitch_section): Likewise.
187 (get_section): Likewise.
188 (get_block_for_section): Likewise.
189 (create_block_symbol): Likewise.
190 (build_constant_desc): Likewise.
191 (create_constant_pool): Likewise.
192 (force_const_mem): Likewise.
193 (record_tm_clone_pair): Likewise.
194 * varpool.c (varpool_create_empty_node): Likewise.
196 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
198 * dwarf2out.c (tree_add_const_value_attribute): Call
199 ggc_internal_cleared_alloc instead of ggc_alloc_cleared_atomic.
200 * gengtype.c (write_typed_alloc_def): Call ggc_internal_<x>alloc
201 instead of ggc_internal_<x>alloc_stat.
202 * ggc-common.c (ggc_internal_cleared_alloc): Drop _stat suffix.
203 (ggc_realloc): Likewise.
204 * ggc-none.c (ggc_internal_alloc): Likewise.
205 (ggc_internal_cleared_alloc): Likewise.
206 * ggc-page.c: Likewise.
207 * ggc.h (ggc_internal_alloc_stat): Likewise.
208 (ggc_internal_alloc): Remove macro.
209 (ggc_internal_cleared_alloc_stat): Drop _stat suffix.
210 (ggc_internal_cleared_alloc): Remove macro.
211 (GGC_RESIZEVEC): Adjust.
212 (ggc_resizevar): Remove macro.
213 (ggc_internal_vec_alloc_stat): Drop _stat suffix.
214 (ggc_internal_cleared_vec_alloc_stat): Likewise.
215 (ggc_internal_vec_cleared_alloc): Remove macro.
216 (ggc_alloc_atomic_stat): Drop _stat suffix.
217 (ggc_alloc_atomic): Remove macro.
218 (ggc_alloc_cleared_atomic): Remove macro.
219 (ggc_alloc_string_stat): Drop _stat suffix.
220 (ggc_alloc_string): Remove macro.
221 (ggc_alloc_rtx_def_stat): Adjust.
222 (ggc_alloc_tree_node_stat): Likewise.
223 (ggc_alloc_cleared_tree_node_stat): Likewise.
224 (ggc_alloc_cleared_gimple_statement_stat): Likewise.
225 (ggc_alloc_cleared_simd_clone_stat): Likewise.
226 * gimple.c (gimple_build_omp_for): Likewise.
227 (gimple_copy): Likewise.
228 * stringpool.c (ggc_alloc_string_stat): Drop _stat suffix.
229 * toplev.c (realloc_for_line_map): Adjust.
230 * tree-data-ref.h (lambda_vector_new): Likewise.
231 * tree-phinodes.c (allocate_phi_node): Likewise.
232 * tree.c (grow_tree_vec_stat): Likewise.
233 * vec.h (va_gc::reserve): Adjust.
235 2014-05-17 Ajit Agarwal <ajitkum@xilinx.com>
237 * config/microblaze/microblaze.c
238 (break_handler): New Declaration.
239 (microblaze_break_function_p,microblaze_is_break_handler): New.
240 (compute_frame_size): use of microblaze_break_function_p. Add
241 the test of break_handler.
242 (microblaze_function_prologue) : Add the test of variable
243 break_handler. Check the fnname by BREAK_HANDLER_NAME.
244 (microblaze_function_epilogue) : Add the test of break_handler.
245 (microblaze_globalize_label) : Add the test of break_handler. Check the
246 name by BREAK_HANDLER_NAME.
248 * config/microblaze/microblaze.h (BREAK_HANDLER_NAME) : New macro
250 * config/microblaze/microblaze.md :
251 (*<optab>,<optab>_internal): Add microblaze_is_break_handler () test.
252 (call_internal1,call_value_intern) : Use of microblaze_break_function_p.
253 Use of SYMBOL_REF_DECL.
255 * config/microblaze/microblaze-protos.h
256 (microblaze_break_function_p,microblaze_is_break_handler) : New Declaration.
258 * testsuite/gcc.target/microblaze/others/break_handler.c : New.
260 * doc/extend.texi (MicroBlaze break_handler Functions): Document new
261 MicroBlaze break_handler functions.
263 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
265 * doc/extend.texi (Size of an asm): Move node text according
266 to its @menu entry position.
268 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
270 PR tree-optimization/61140
271 PR tree-optimization/61150
272 PR tree-optimization/61197
273 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
275 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
277 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
279 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
281 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
282 __SIZEOF_INT128__ is defined.
284 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
286 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
287 (rs6000_delegitimize_address): Use it.
289 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
291 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
292 inplace argument. Store the new address in the original MEM when true.
293 * emit-rtl.c (change_address_1): Likewise.
294 (adjust_address_1, adjust_automodify_address_1, offset_address):
296 * rtl.h (plus_constant): Add an inplace argument.
297 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
298 when true. Avoid generating (plus X (const_int 0)).
299 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
300 in-place. Pass true to plus_constant.
301 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
303 2014-05-16 Dehao Chen <dehao@google.com>
305 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
307 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
310 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
312 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
314 2014-05-16 Dehao Chen <dehao@google.com>
316 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
317 optimize_function_for_size_p.
318 * regs.h (REG_FREQ_FROM_BB): Likewise.
320 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
323 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
324 negt_reg_operand cases.
325 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
327 * config/sh/predicates.md (cbranch_treg_value): Simplify.
329 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
331 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
334 2014-05-16 David Malcolm <dmalcolm@redhat.com>
337 2014-04-29 David Malcolm <dmalcolm@redhat.com>
339 * tree-cfg.c (dump_function_to_file): Dump the return type of
340 functions, in a line to itself before the function body, mimicking
341 the layout of a C function.
343 2014-05-16 Dehao Chen <dehao@google.com>
345 * cfghooks.c (make_forwarder_block): Use direct computation to
346 get fall-through edge's count and frequency.
348 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
350 * config/arc/arc.c (arc_init): Fix typo in error message.
351 * config/i386/i386.c (ix86_expand_builtin): Likewise.
352 (split_stack_prologue_scratch_regno): Likewise.
353 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
354 word from error message.
356 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
358 * ira-costs.c: Fix typo in comment.
360 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
362 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
364 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
366 * varpool.c (dump_varpool_node): Dump write-only flag.
367 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
369 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
370 write-only variables.
371 * ipa.c (process_references): New function.
372 (set_readonly_bit): New function.
373 (set_writeonly_bit): New function.
374 (clear_addressable_bit): New function.
375 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
376 fix handling of aliases.
377 * cgraph.h (struct varpool_node): Add writeonly flag.
379 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
381 PR rtl-optimization/60969
382 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
383 Calculate costs for this case.
385 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
387 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
388 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
390 2014-05-16 Richard Biener <rguenther@suse.de>
392 PR tree-optimization/61194
393 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
394 bool patterns ending in a COND_EXPR.
396 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
398 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
400 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
402 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
403 where we were unable to cost an RTX.
405 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
407 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
410 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
411 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
413 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
415 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
416 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
418 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
419 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
421 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
422 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
424 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
427 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
428 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
430 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
433 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
434 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
436 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
437 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
439 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
440 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
442 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
445 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
446 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
448 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
449 ZERO_EXTEND and SIGN_EXTEND better.
451 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
452 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
454 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
457 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
458 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
460 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
461 costs when costing loads and stores to memory.
463 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
464 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
466 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
469 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
471 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
473 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
474 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
476 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
478 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
480 (aarch64_rtx_mult_cost): New.
481 (aarch64_rtx_costs): Use it, refactor as appropriate.
483 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
484 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
486 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
487 emit instructions, return number of instructions which would
489 (aarch64_add_constant): Update call to aarch64_build_constant.
490 (aarch64_output_mi_thunk): Likewise.
491 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
494 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
496 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
497 (TARGET_RTX_COSTS): Call it.
499 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
501 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
502 (cortexa57_vector_cost): Likewise.
503 (cortexa57_tunings): Use them.
505 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
507 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
508 (cpu_addrcost_table): Use it.
509 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
510 (aarch64_address_cost): Rewrite using aarch64_classify_address,
513 2014-05-16 Richard Biener <rguenther@suse.de>
515 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
516 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
517 (visit_phi): Ignore edges marked as not executable.
518 (class cond_dom_walker): New.
519 (cond_dom_walker::before_dom_children): Value-number
520 control statements and mark successor edges as not
521 executable if possible.
522 (run_scc_vn): First walk all control statements in
523 dominator order, marking edges as not executable.
524 * tree-inline.c (copy_edges_for_bb): Be not confused
525 about random edge flags.
527 2014-05-16 Richard Biener <rguenther@suse.de>
529 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
531 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
534 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
535 (__TM_simple_begin): Use it.
536 (__TM_begin): Likewise.
538 2014-05-15 Martin Jambor <mjambor@suse.cz>
541 * ipa-prop.c (update_indirect_edges_after_inlining): Check
542 type_preserved flag when the indirect edge is polymorphic.
544 2014-05-15 Martin Jambor <mjambor@suse.cz>
546 PR tree-optimization/61090
547 * tree-sra.c (sra_modify_expr): Pass the current gsi to
550 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
552 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
553 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
555 2014-05-15 Jakub Jelinek <jakub@redhat.com>
557 PR tree-optimization/61158
558 * fold-const.c (fold_binary_loc): If X is zero-extended and
559 shiftc >= prec, make sure zerobits is all ones instead of
560 invoking undefined behavior.
562 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
564 * regcprop.h: New file.
565 * regcprop.c (skip_debug_insn_p): New decl.
566 (replace_oldest_value_reg): Check skip_debug_insn_p.
567 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
568 * shrink-wrap.c: Include regcprop.h.
569 (prepare_shrink_wrap): Call
570 copyprop_hardreg_forward_bb_without_debug_insn.
572 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
574 * shrink-wrap.h: Update comment.
575 * shrink-wrap.c: Update comment.
576 (next_block_for_reg): Rename to live_edge_for_reg.
577 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
578 (move_insn_for_shrink_wrap): Split live_edge.
579 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
581 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
583 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
585 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
586 * config/sparc/sparc.md (fptype_ut699): New attribute.
587 (in_branch_delay): Return false if -mfix-ut699 is specified and
588 fptype_ut699 is set to single.
589 (truncdfsf2): Add fptype_ut699 attribute.
590 (fix_truncdfsi2): Likewise.
591 (floatsisf2): Change fptype attribute.
592 (fix_truncsfsi2): Likewise.
593 (negtf2_notv9): Delete.
594 (negtf2_v9): Likewise.
595 (negtf2_hq): New instruction.
596 (negtf2): New instruction and splitter.
597 (negdf2_notv9): Rewrite.
598 (abstf2_notv9): Delete.
599 (abstf2_hq_v9): Likewise.
600 (abstf2_v9): Likewise.
601 (abstf2_hq): New instruction.
602 (abstf2): New instruction and splitter.
603 (absdf2_notv9): Rewrite.
605 2014-05-14 Cary Coutant <ccoutant@google.com>
608 * opts.c (common_handle_option): Don't special-case "-g".
609 (set_debug_level): Default to at least level 2 with "-g".
611 2014-05-14 DJ Delorie <dj@redhat.com>
613 * config/msp430/msp430.c (msp430_builtin): Add
614 MSP430_BUILTIN_DELAY_CYCLES.
615 (msp430_init_builtins): Register void __delay_cycles(long long).
616 (msp430_builtin_decl): Add it.
617 (cg_magic_constant): New.
618 (msp430_expand_delay_cycles): New.
619 (msp430_expand_builtin): Call it.
620 (msp430_print_operand_raw): Change integer printing from "int" to
622 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
623 (delay_cycles_start): New.
624 (delay_cycles_end): New.
625 (delay_cycles_32): New.
626 (delay_cycles_32x): New.
627 (delay_cycles_16): New.
628 (delay_cycles_16x): New.
629 (delay_cycles_2): New.
630 (delay_cycles_1): New.
631 * doc/extend.texi: Document __delay_cycles().
633 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
635 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
636 length attribute computation.
638 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
641 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
643 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
646 * config/sparc/sparc.md: Fix types of low and high in DI constant
647 splitter. Use gen_int_mode in some other splitters.
649 2014-05-14 Martin Jambor <mjambor@suse.cz>
652 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
654 2014-05-14 James Norris <jnorris@codesourcery.com>
656 * omp-low.c (expand_parallel_call): Remove shadow variable.
657 (expand_omp_taskreg): Likewise.
659 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
661 * common/config/i386/i386-common.c
662 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
663 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
664 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
665 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
666 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
667 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
668 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
669 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
670 xsavecintrin.h, xsavesintrin.h.
672 * config/i386/clflushoptintrin.h: New.
673 * config/i386/xsavecintrin.h: Ditto.
674 * config/i386/xsavesintrin.h: Ditto.
675 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
678 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
679 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
681 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
682 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
683 OPTION_MASK_ISA_XSAVES.
684 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
686 (PTA_CLFLUSHOPT) Define.
689 (ix86_option_override_internal): Handle new options.
690 (ix86_valid_target_attribute_inner_p): Ditto.
691 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
692 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
693 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
694 (bdesc_special_args): Add __builtin_ia32_xsaves,
695 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
696 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
697 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
698 (ix86_expand_builtin): Handle new builtins.
699 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
700 (TARGET_CLFLUSHOPT_P): Ditto.
701 (TARGET_XSAVEC): Ditto.
702 (TARGET_XSAVEC_P): Ditto.
703 (TARGET_XSAVES): Ditto.
704 (TARGET_XSAVES_P): Ditto.
705 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
706 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
707 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
709 (ANY_XRSTOR64): Ditto.
711 (xrstor): Change into <xrstor>.
712 (xrstor_rex64): Change into <xrstor>_rex64.
713 (xrstor64): Change into <xrstor>64
715 * config/i386/i386.opt (mclflushopt): New.
718 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
720 * doc/invoke.texi: Document new options.
722 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
724 PR rtl-optimization/60866
725 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
726 Default it to -1. Pass it down to init_simplejump_data.
727 (init_simplejump_data): New parameter old_seqno. Pass it down
728 to get_seqno_for_a_jump.
729 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
730 initializing new jump seqno as a last resort. Add comment.
731 (sel_redirect_edge_and_branch): Save old seqno of the conditional
732 jump and pass it down to sel_init_new_insn.
733 (sel_redirect_edge_and_branch_force): Likewise.
735 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
737 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
738 shifted values to avoid build warning.
740 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
742 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
743 * cfgrtl.c (rtl_merge_blocks): Fix comment.
744 (cfg_layout_merge_blocks): Likewise.
745 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
747 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
749 PR rtl-optimization/60901
750 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
751 bb predecessor belongs to the same scheduling region. Adjust comment.
753 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
755 * doc/sourcebuild.texi: (dfp_hw): Document.
756 (p8vector_hw): Likewise.
757 (powerpc_eabi_ok): Likewise.
758 (powerpc_elfv2): Likewise.
759 (powerpc_htm_ok): Likewise.
760 (ppc_recip_hw): Likewise.
763 2014-05-13 Cary Coutant <ccoutant@google.com>
765 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
767 2014-05-13 David Malcolm <dmalcolm@redhat.com>
769 * gengtype-parse.c (require3): Eliminate in favor of...
771 (require_template_declaration): Update to support optional single *
774 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
775 (create_user_defined_type): Handle a single level of explicit
776 pointerness within template arguments.
777 (struct write_types_data): Add field "kind".
778 (filter_type_name): Handle "*" character.
779 (write_user_func_for_structure_ptr): Require a write_types_data
780 rather than just a prefix string, so that we can look up the kind
781 of the wtd and use it as an index into wrote_user_func_for_ptr,
782 ensuring that such functions are written at most once. Support
783 subclasses by invoking the marking function of the ultimate base class.
784 (write_user_func_for_structure_body): Require a write_types_data
785 rather than just a prefix string, so that we can pass this to
786 write_user_func_for_structure_ptr.
787 (write_func_for_structure): Likewise.
788 (ggc_wtd): Add initializer of new "kind" field.
791 * gengtype.h (enum write_types_kinds): New.
792 (struct type): Add field wrote_user_func_for_ptr to the "s"
795 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
797 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
798 instead of const_binop.
799 (fold_binary_loc): Likewise.
801 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
803 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
804 calculation to match get_ref_base_and_extent.
806 2014-05-13 Catherine Moore <clm@codesourcery.com>
807 Sandra Loosemore <sandra@codesourcery.com>
809 * configure.ac: Fix assembly for explicit JALR relocation check.
810 * configure: Regenerate.
812 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
814 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
815 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
816 Remove associated type declarations and initialisations.
817 (arm_expand_neon_builtin): Likewise.
818 (neon_emit_pair_result_insn): Delete.
819 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
820 * config/arm/neon.md (neon_vtrn<mode>): Delete.
821 (neon_vzip<mode>): Likewise.
822 (neon_vuzp<mode>): Likewise.
824 2014-05-13 Richard Biener <rguenther@suse.de>
827 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
828 it needs revisiting whether the call still may be tail-called.
830 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
832 * rtl.def (SYMBOL_REF): Remove middle "0" field.
833 * rtl.h (block_symbol): Reduce number of fields to 2.
834 (rtx_def): Add u2.symbol_ref_flags.
835 (SYMBOL_REF_FLAGS): Use it.
836 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
837 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
838 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
839 Lower index of SYMBOL_REF_DATA.
840 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
841 Print SYMBOL_REF_FLAGS at the same time.
842 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
844 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
846 * rtl.def (VAR_LOCATION): Remove "i" field.
847 * rtl.h (rtx_def): Add u2.var_location_status.
848 (PAT_VAR_LOCATION_STATUS): Use it.
849 (gen_rtx_VAR_LOCATION): Declare.
850 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
851 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
852 * var-tracking.c (emit_note_insn_var_location): Remove casts.
854 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
856 * rtl.def (scratch): Fix outdated comment and remove "0" field.
857 * gengtype.c (adjust_field_rtx_def): Update accordingly.
859 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
861 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
862 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
863 * rtl.h (rtx_def): Add insn_uid to u2 field.
864 (RTX_FLAG_CHECK8): Delete in favor of...
865 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
866 (INSN_DELETED_P): Update accordingly.
867 (INSN_UID): Use u2.insn_uid.
868 (INSN_CHAIN_CODE_P): Define.
869 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
870 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
871 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
872 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
873 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
875 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
876 Update indices for insn-chain rtxes.
877 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
878 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
879 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
880 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
881 * combine.c (try_combine): Likewise.
882 * ira.c (setup_prohibited_mode_move_regs): Likewise.
884 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
886 * rtl.def (REG): Remove middle field.
887 * rtl.h (rtx_def): Add orignal_regno to u2.
888 (ORIGINAL_REGNO): Use it instead of field 1.
889 (REG_ATTRS): Lower field index accordingly.
890 * gengtype.c (adjust_field_rtx_def): Remove handling of
891 ORIGINAL_REGNO. Move REG_ATTRS index down.
892 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
893 code that prints the REGNO.
895 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
897 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
900 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
902 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
904 2014-05-13 Bin Cheng <bin.cheng@arm.com>
906 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
907 (alloc_iv): Lower base expressions containing ADDR_EXPR.
909 2014-05-13 Ian Bolton <ian.bolton@arm.com>
911 * config/aarch64/aarch64-protos.h
912 (aarch64_hard_regno_caller_save_mode): New prototype.
913 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
915 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
917 2014-05-13 Christian Bruel <christian.bruel@st.com>
919 * target.def (mode_switching): New hook vector.
920 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
921 (mode_exit, modepriority_to_mode): Likewise.
922 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
923 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
924 * target.h: Include tm.h and hard-reg-set.h.
925 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
926 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
927 * doc/tm.texi Regenerate.
928 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
929 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
930 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
931 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
932 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
933 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
934 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
935 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
936 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
937 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
938 (ix86_emit_mode_set): Hookify.
939 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
941 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
942 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
943 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
944 (epiphany_mode_priority_to_mode): Remove declaration.
945 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
946 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
947 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
949 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
950 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
951 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
953 2014-05-13 Jakub Jelinek <jakub@redhat.com>
956 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
957 is const0_rtx, return immediately. Don't test count == 0 when
960 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
962 * Makefile.in: add shrink-wrap.o.
963 * config/i386/i386.c: include "shrink-wrap.h"
964 * function.c: Likewise.
965 (requires_stack_frame_p, next_block_for_reg,
966 move_insn_for_shrink_wrap, prepare_shrink_wrap,
967 dup_block_and_redirect): Move to shrink-wrap.c
968 (thread_prologue_and_epilogue_insns): Extract three code segments
969 as functions in shrink-wrap.c
970 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
972 * shrink-wrap.c: New file.
973 * shrink-wrap.h: New file.
975 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
977 * doc/extend.texi: Reflect current numbers of pragmas. Remove
978 reference to Solaris.
980 2014-05-12 Mike Stump <mikestump@comcast.net>
983 * genattrtab.c (filename): Add.
984 (convert_set_attr_alternative): Improve error message.
985 (check_defs): Restore read_md_filename for error messages.
986 (gen_insn): Save filename.
988 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
990 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
991 -fno-local-ivars and -fivar-visibility.
992 * c-family/c.opt: Make -Wshadow also implicitly enable
995 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
997 * doc/tm.texi: Remove reference to deleted macro.
998 * doc/tm.texi.in: Likewise.
1000 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
1003 * config/avr/avr.c (avr_out_store_psi): Use correct constant
1006 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
1008 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
1009 * config/arm/aout.h (License): Same.
1010 * config/arm/bpabi.h (License): Same.
1011 * config/arm/elf.h (License): Same.
1012 * config/arm/linux-elf.h (License): Same.
1013 * config/arm/linux-gas.h (License): Same.
1014 * config/arm/netbsd-elf.h (License): Same.
1015 * config/arm/uclinux-eabi.h (License): Same.
1016 * config/arm/uclinux-elf.h (License): Same.
1017 * config/arm/vxworks.h (License): Same.
1019 2014-05-11 Jakub Jelinek <jakub@redhat.com>
1021 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
1022 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
1023 number of operands to 3.
1024 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
1025 * tree-nested.c (convert_nonlocal_omp_clauses,
1026 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
1027 * gimplify.c (gimplify_scan_omp_clauses): Handle
1028 OMP_CLAUSE_LINEAR_STMT.
1029 * omp-low.c (lower_rec_input_clauses): Fix typo.
1030 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
1031 cast between Fortran boolean_type_node and C _Bool if
1034 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
1036 PR tree-optimization/61136
1037 * wide-int.h (multiple_of_p): Define a version that doesn't return
1039 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
1040 integer_zerop/const_binop pair.
1041 (multiple_of_p): Likewise, converting both operands to widest_int
1044 2014-05-09 Teresa Johnson <tejohnson@google.com>
1046 * cgraphunit.c (analyze_functions): Use correct dump file.
1048 2014-05-09 Florian Weimer <fweimer@redhat.com>
1050 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
1052 (stack_protect_return_slot_p): New function.
1053 (expand_used_vars): Call stack_protect_decl_p and
1054 stack_protect_return_slot_p for -fstack-protector-strong.
1056 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
1059 * fold-const.c (fold_binary_loc): Changed width of mask.
1061 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
1063 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
1064 unsigned int initializers for regno_in, regno_out.
1066 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
1069 * config/avr/avr.md (cc): Add new attribute set_vzn.
1070 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
1071 Set cc insn attribute to set_vzn instead of set_zn for alternatives
1072 with INC, DEC or NEG.
1073 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
1074 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
1075 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
1077 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1080 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1082 * wide-int.cc (UTItype): Define.
1083 (UDWtype): Define for appropriate W_TYPE_SIZE.
1085 2014-05-09 Richard Biener <rguenther@suse.de>
1087 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
1088 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
1089 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
1090 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
1091 ssa_propagate): Adjust.
1093 2014-05-08 Jeff Law <law@redhat.com>
1095 PR tree-optimization/61009
1096 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
1097 tri-state rather than a boolean. When a block is too big to
1098 thread through, inform caller via negative return value.
1099 (thread_across_edge): If a block was too big for normal threading,
1100 then it's too big for a joiner too, so remove temporary equivalences
1101 and return immediately.
1103 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
1104 Matthias Klose <doko@ubuntu.com>
1107 * optc-gen.awk: Fix option handling for -Wunused-parameter.
1109 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
1112 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
1114 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
1117 * config/alpha/alpha.c: Include gimple-iterator.h.
1118 (alpha_gimple_fold_builtin): New function. Move
1119 ALPHA_BUILTIN_UMULH folding from ...
1120 (alpha_fold_builtin): ... here.
1121 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
1123 2014-05-08 Wei Mi <wmi@google.com>
1126 * config/i386/i386.c (ix86_compute_frame_layout): Update
1127 preferred_stack_boundary for call, expanded from tls descriptor.
1128 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
1129 to depend on SP register.
1130 (*tls_local_dynamic_base_32_gnu): Ditto.
1131 (*tls_local_dynamic_32_once): Ditto.
1132 (tls_global_dynamic_64_<mode>): Set
1133 ix86_tls_descriptor_calls_expanded_in_cfun.
1134 (tls_local_dynamic_base_64_<mode>): Ditto.
1135 (tls_global_dynamic_32): Set
1136 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
1137 to depend on SP register.
1138 (tls_local_dynamic_base_32): Ditto.
1140 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1142 * config/arm/arm_neon.h: Update comment.
1143 * config/arm/neon-docgen.ml: Delete.
1144 * config/arm/neon-gen.ml: Delete.
1145 * doc/arm-neon-intrinsics.texi: Update comment.
1147 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1149 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
1151 (vand, vorr, veor, vorn, vbic): Remove.
1152 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
1154 (neon_vsub_unspec): Likewise.
1155 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
1157 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1159 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
1160 (vadd_s16): Likewise.
1161 (vadd_s32): Likewise.
1162 (vadd_f32): Likewise.
1163 (vadd_u8): Likewise.
1164 (vadd_u16): Likewise.
1165 (vadd_u32): Likewise.
1166 (vadd_s64): Likewise.
1167 (vadd_u64): Likewise.
1168 (vaddq_s8): Likewise.
1169 (vaddq_s16): Likewise.
1170 (vaddq_s32): Likewise.
1171 (vaddq_s64): Likewise.
1172 (vaddq_f32): Likewise.
1173 (vaddq_u8): Likewise.
1174 (vaddq_u16): Likewise.
1175 (vaddq_u32): Likewise.
1176 (vaddq_u64): Likewise.
1177 (vmul_s8): Likewise.
1178 (vmul_s16): Likewise.
1179 (vmul_s32): Likewise.
1180 (vmul_f32): Likewise.
1181 (vmul_u8): Likewise.
1182 (vmul_u16): Likewise.
1183 (vmul_u32): Likewise.
1184 (vmul_p8): Likewise.
1185 (vmulq_s8): Likewise.
1186 (vmulq_s16): Likewise.
1187 (vmulq_s32): Likewise.
1188 (vmulq_f32): Likewise.
1189 (vmulq_u8): Likewise.
1190 (vmulq_u16): Likewise.
1191 (vmulq_u32): Likewise.
1192 (vsub_s8): Likewise.
1193 (vsub_s16): Likewise.
1194 (vsub_s32): Likewise.
1195 (vsub_f32): Likewise.
1196 (vsub_u8): Likewise.
1197 (vsub_u16): Likewise.
1198 (vsub_u32): Likewise.
1199 (vsub_s64): Likewise.
1200 (vsub_u64): Likewise.
1201 (vsubq_s8): Likewise.
1202 (vsubq_s16): Likewise.
1203 (vsubq_s32): Likewise.
1204 (vsubq_s64): Likewise.
1205 (vsubq_f32): Likewise.
1206 (vsubq_u8): Likewise.
1207 (vsubq_u16): Likewise.
1208 (vsubq_u32): Likewise.
1209 (vsubq_u64): Likewise.
1210 (vand_s8): Likewise.
1211 (vand_s16): Likewise.
1212 (vand_s32): Likewise.
1213 (vand_u8): Likewise.
1214 (vand_u16): Likewise.
1215 (vand_u32): Likewise.
1216 (vand_s64): Likewise.
1217 (vand_u64): Likewise.
1218 (vandq_s8): Likewise.
1219 (vandq_s16): Likewise.
1220 (vandq_s32): Likewise.
1221 (vandq_s64): Likewise.
1222 (vandq_u8): Likewise.
1223 (vandq_u16): Likewise.
1224 (vandq_u32): Likewise.
1225 (vandq_u64): Likewise.
1226 (vorr_s8): Likewise.
1227 (vorr_s16): Likewise.
1228 (vorr_s32): Likewise.
1229 (vorr_u8): Likewise.
1230 (vorr_u16): Likewise.
1231 (vorr_u32): Likewise.
1232 (vorr_s64): Likewise.
1233 (vorr_u64): Likewise.
1234 (vorrq_s8): Likewise.
1235 (vorrq_s16): Likewise.
1236 (vorrq_s32): Likewise.
1237 (vorrq_s64): Likewise.
1238 (vorrq_u8): Likewise.
1239 (vorrq_u16): Likewise.
1240 (vorrq_u32): Likewise.
1241 (vorrq_u64): Likewise.
1242 (veor_s8): Likewise.
1243 (veor_s16): Likewise.
1244 (veor_s32): Likewise.
1245 (veor_u8): Likewise.
1246 (veor_u16): Likewise.
1247 (veor_u32): Likewise.
1248 (veor_s64): Likewise.
1249 (veor_u64): Likewise.
1250 (veorq_s8): Likewise.
1251 (veorq_s16): Likewise.
1252 (veorq_s32): Likewise.
1253 (veorq_s64): Likewise.
1254 (veorq_u8): Likewise.
1255 (veorq_u16): Likewise.
1256 (veorq_u32): Likewise.
1257 (veorq_u64): Likewise.
1258 (vbic_s8): Likewise.
1259 (vbic_s16): Likewise.
1260 (vbic_s32): Likewise.
1261 (vbic_u8): Likewise.
1262 (vbic_u16): Likewise.
1263 (vbic_u32): Likewise.
1264 (vbic_s64): Likewise.
1265 (vbic_u64): Likewise.
1266 (vbicq_s8): Likewise.
1267 (vbicq_s16): Likewise.
1268 (vbicq_s32): Likewise.
1269 (vbicq_s64): Likewise.
1270 (vbicq_u8): Likewise.
1271 (vbicq_u16): Likewise.
1272 (vbicq_u32): Likewise.
1273 (vbicq_u64): Likewise.
1274 (vorn_s8): Likewise.
1275 (vorn_s16): Likewise.
1276 (vorn_s32): Likewise.
1277 (vorn_u8): Likewise.
1278 (vorn_u16): Likewise.
1279 (vorn_u32): Likewise.
1280 (vorn_s64): Likewise.
1281 (vorn_u64): Likewise.
1282 (vornq_s8): Likewise.
1283 (vornq_s16): Likewise.
1284 (vornq_s32): Likewise.
1285 (vornq_s64): Likewise.
1286 (vornq_u8): Likewise.
1287 (vornq_u16): Likewise.
1288 (vornq_u32): Likewise.
1289 (vornq_u64): Likewise.
1291 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1293 * wide-int.cc (UTItype): Define.
1294 (UDWtype): Define for appropriate W_TYPE_SIZE.
1296 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
1298 PR tree-optimization/59100
1299 * tree-ssa-phiopt.c: Include tree-inline.h.
1300 (neutral_element_p, absorbing_element_p): New functions.
1301 (value_replacement): Handle conditional binary operations with a
1302 neutral or absorbing element.
1304 2014-05-08 Richard Biener <rguenther@suse.de>
1306 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
1307 folding the expression.
1308 (valueize_expr): Remove.
1309 (visit_reference_op_load): Do not valueize the result of
1311 (simplify_binary_expression): Likewise.
1312 (simplify_unary_expression): Likewise.
1314 2014-05-08 Richard Biener <rguenther@suse.de>
1316 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
1317 looking at TYPE_ARG_TYPES.
1319 2014-05-08 Richard Biener <rguenther@suse.de>
1321 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
1322 pointer propagation special-case.
1324 2014-05-08 Bin Cheng <bin.cheng@arm.com>
1326 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
1327 core part of address expressions.
1329 2014-05-08 Alan Modra <amodra@gmail.com>
1332 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
1333 loads and stores when -mno-strict-align at any alignment.
1334 (expand_block_clear): Similarly. Also correct calculation of
1337 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
1340 * tree-complex.c (expand_complex_move): Keep line info when expanding
1342 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
1343 of complex expression. Use new argument to display correct location
1344 for values coming from phi statement.
1345 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
1346 (warn_uninitialized_phi): Pass location of phi argument to
1348 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
1349 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
1351 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
1353 * config/rs6000/predicates.md (indexed_address_mem): New.
1354 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
1355 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
1356 fpstore_ux, fpstore_u.
1357 (sign_extend, indexed, update): New.
1358 (cell_micro): Adjust.
1359 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
1360 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
1361 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
1362 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
1363 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
1364 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
1365 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
1366 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
1367 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
1368 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
1369 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
1370 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
1371 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
1372 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
1373 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
1375 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
1376 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
1377 *vsx_extract_<mode>_store): Adjust.
1378 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
1379 is_cracked_insn, insn_must_be_first_in_group,
1380 insn_must_be_last_in_group): Adjust.
1382 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
1384 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
1385 ppc440-fpstore): Adjust.
1386 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
1387 ppc476-fpstore): Adjust.
1388 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
1389 ppc601-fpstore): Adjust.
1390 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
1392 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
1394 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
1395 ppc7450-fpstore): Adjust.
1396 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
1397 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
1398 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
1400 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
1401 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
1402 cell-fpstore, cell-fpstore-update): Adjust.
1403 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
1404 ppce300c3_store, ppce300c3_fpstore): Adjust.
1405 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
1406 e500mc_fpstore): Adjust.
1407 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
1408 e500mc64_store, e500mc64_fpstore): Adjust.
1409 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
1410 e5500_fpstore): Adjust.
1411 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
1412 e6500_fpstore): Adjust.
1413 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
1415 * config/rs6000/power4.md (power4-load, power4-load-ext,
1416 power4-load-ext-update, power4-load-ext-update-indexed,
1417 power4-load-update-indexed, power4-load-update, power4-fpload,
1418 power4-fpload-update, power4-store, power4-store-update,
1419 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
1421 * config/rs6000/power5.md (power5-load, power5-load-ext,
1422 power5-load-ext-update, power5-load-ext-update-indexed,
1423 power5-load-update-indexed, power5-load-update, power5-fpload,
1424 power5-fpload-update, power5-store, power5-store-update,
1425 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
1427 * config/rs6000/power6.md (power6-load, power6-load-ext,
1428 power6-load-update, power6-load-update-indexed,
1429 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
1430 power6-fpload-update, power6-store, power6-store-update,
1431 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
1433 * config/rs6000/power7.md (power7-load, power7-load-ext,
1434 power7-load-update, power7-load-update-indexed,
1435 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
1436 power7-fpload-update, power7-store, power7-store-update,
1437 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
1439 * config/rs6000/power8.md (power8-load, power8-load-update,
1440 power8-load-ext, power8-load-ext-update, power8-fpload,
1441 power8-fpload-update, power8-store, power8-store-update-indexed,
1442 power8-fpstore, power8-fpstore-update): Adjust.
1443 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
1445 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
1446 titan_lsu_store, titan_lsu_fpstore): Adjust.
1447 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
1449 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
1452 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
1453 unrolled byte insns. Emit address increments after move insns.
1455 2014-05-07 David Malcolm <dmalcolm@redhat.com>
1457 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
1458 const_gimple, rather than a gimple.
1459 (gimple_call_builtin_p): Likewise, for the three variants.
1461 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
1462 (gimple_call_builtin_p): Likewise, for the three variants.
1464 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1466 PR tree-optimization/61095
1467 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
1469 2014-05-07 Richard Biener <rguenther@suse.de>
1471 PR tree-optimization/61034
1472 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
1473 (maybe_skip_until): Use translate to take into account
1474 lattices when trying to do disambiguations.
1475 (get_continuation_for_phi_1): Likewise.
1476 (get_continuation_for_phi): Adjust for added translate arguments.
1477 (walk_non_aliased_vuses): Likewise.
1478 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
1479 (walk_non_aliased_vuses): Likewise.
1480 (call_may_clobber_ref_p_1): Declare.
1481 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
1482 calls. Stop early if we are only supposed to disambiguate.
1483 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
1485 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
1487 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
1488 Emit an error when the function has arguments.
1490 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
1492 * cfgloop.h (unswitch_loops): Remove.
1493 * doc/passes.texi: Remove references to loop-unswitch.c
1494 * timevar.def (TV_LOOP_UNSWITCH): Remove.
1496 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
1498 * tree-vect-data-refs.c (vect_grouped_load_supported): New
1499 check for loads group of length 3.
1500 (vect_permute_load_chain): New permutations for loads group of
1502 * tree-vect-stmts.c (vect_model_load_cost): Change cost
1503 of vec_perm_shuffle for the new permutations.
1505 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
1507 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
1508 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
1509 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
1510 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
1511 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
1512 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
1513 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
1514 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
1516 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
1518 * loop-unswitch.c: Delete.
1520 2014-05-07 Richard Biener <rguenther@suse.de>
1522 * config.gcc: Always set need_64bit_hwint to yes.
1524 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
1526 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
1527 of using optimize_size.
1529 2014-05-06 Mike Stump <mikestump@comcast.net>
1531 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
1533 2014-05-06 Joseph Myers <joseph@codesourcery.com>
1535 * config/i386/sse.md (*mov<mode>_internal)
1536 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
1537 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
1538 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
1539 (*<code><mode>3, *andnot<mode>3<mask_name>)
1540 (<mask_codefor><code><mode>3<mask_name>): Only consider
1541 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
1543 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
1546 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1548 * lra-constraints.c (valid_address_p): Move earlier in file.
1549 Add a constraint argument to the address_info version.
1550 (satisfies_memory_constraint_p): New function.
1551 (satisfies_address_constraint_p): Likewise.
1552 (process_alt_operands, curr_insn_transform): Use them.
1553 (process_address): Pass the constraint to valid_address_p when
1554 checking address operands.
1556 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
1558 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
1559 to their respective blocks. Fix inadvertent use of "node".
1561 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
1563 * emit-rtl.c (init_derived_machine_modes): New functionm, split
1565 (init_emit_once): ...here.
1566 * rtl.h (init_derived_machine_modes): Declare.
1567 * toplev.c (do_compile): Call it even if no_backend.
1569 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
1570 Mike Stump <mikestump@comcast.net>
1571 Richard Sandiford <rdsandiford@googlemail.com>
1572 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1574 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
1575 (rtx_equal_for_memref_p): Update comment.
1576 (adjust_offset_for_component_ref): Use wide-int interfaces.
1577 * builtins.c (get_object_alignment_2): Likewise.
1578 (c_readstr): Likewise.
1579 (target_char_cast): Add comment.
1580 (determine_block_size): Use wide-int interfaces.
1581 (expand_builtin_signbit): Likewise.
1582 (fold_builtin_int_roundingfn): Likewise.
1583 (fold_builtin_bitop): Likewise.
1584 (fold_builtin_bswap): Likewise.
1585 (fold_builtin_logarithm): Use signop.
1586 (fold_builtin_pow): Likewise.
1587 (fold_builtin_memory_op): Use wide-int interfaces.
1588 (fold_builtin_object_size): Likewise.
1589 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
1590 nb_iterations_estimate.
1591 (record_niter_bound): Use wide-int interfaces.
1592 (get_estimated_loop_iterations_int): Likewise.
1593 (get_estimated_loop_iterations): Likewise.
1594 (get_max_loop_iterations): Likewise.
1595 * cfgloop.h: Include wide-int.h.
1596 (struct nb_iter_bound): Change bound to widest_int.
1597 (struct loop): Change nb_iterations_upper_bound and
1598 nb_iterations_estimate to widest_int.
1599 (record_niter_bound): Switch to use widest_int.
1600 (get_estimated_loop_iterations): Likewise.
1601 (get_max_loop_iterations): Likewise.
1602 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
1603 update for wide-int.
1604 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
1605 * combine.c (try_combine): Likewise.
1606 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
1607 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
1609 (aarch64_float_const_representable_p): Likewise.
1610 * config/arc/arc.c: Include wide-int.h.
1611 (arc_can_use_doloop_p): Use wide-int interfaces.
1612 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
1613 (vfp3_const_double_index): Likewise.
1614 * config/avr/avr.c (avr_out_round): Likewise.
1615 (avr_fold_builtin): Likewise.
1616 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
1617 (bfin_can_use_doloop_p): Likewise.
1618 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
1619 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
1620 * config/i386/i386.c: Include wide-int.h.
1621 (ix86_data_alignment): Use wide-int interfaces.
1622 (ix86_local_alignment): Likewise.
1623 (ix86_emit_swsqrtsf): Update real_from_integer.
1624 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
1625 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
1626 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
1627 (zero_constant): Likewise.
1628 (input_operand): Likewise.
1629 (splat_input_operand): Likewise.
1630 (non_logical_cint_operand): Change const_double to const_wide_int.
1631 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
1632 (easy_altivec_constant): Remove comment.
1633 (paired_expand_vector_init): Use CONSTANT_P.
1634 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
1635 (rs6000_emit_move): Update checks.
1636 (rs6000_aggregate_candidate): Use wide-int interfaces.
1637 (rs6000_expand_ternop_builtin): Likewise.
1638 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
1639 (rs6000_assemble_integer): Likewise.
1640 (rs6000_hash_constant): Likewise.
1641 (output_toc): Likewise.
1642 (rs6000_rtx_costs): Likewise.
1643 (rs6000_emit_swrsqrt); Update call to real_from_integer.
1644 * config/rs6000/rs6000-c.c: Include wide-int.h.
1645 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
1646 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
1647 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
1648 Handle CONST_WIDE_INT.
1649 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
1650 Use tree_fits_uhwi_p.
1651 * config/sparc/sparc.c: Include wide-int.h.
1652 (sparc_fold_builtin): Use wide-int interfaces.
1653 * config/vax/vax.c: Include wide-int.h.
1654 (vax_float_literal): Use real_from_integer.
1655 * coretypes.h (struct hwivec_def): New.
1657 (const_hwivec): New.
1658 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
1659 (equiv_constant): Handle CONST_WIDE_INT.
1660 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
1661 (cselib_hash_rtx): Handle CONST_WIDE_INT.
1662 * dbxout.c (stabstr_U): Use wide-int interfaces.
1663 (dbxout_type): Update to use cst_fits_shwi_p.
1664 * defaults.h (LOG2_BITS_PER_UNIT): Define.
1665 (TARGET_SUPPORTS_WIDE_INT): Add default.
1666 * dfp.c: Include wide-int.h.
1667 (decimal_real_to_integer2): Use wide-int interfaces and rename to
1668 decimal_real_to_integer.
1669 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
1670 decimal_real_to_integer.
1671 * doc/generic.texi (Constant expressions): Update for wide_int.
1672 * doc/rtl.texi (const_double): Likewise.
1673 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
1674 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
1675 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
1676 (REAL_VALUE_FROM_INT): Remove.
1677 (TARGET_SUPPORTS_WIDE_INT): New.
1678 * doc/tm.texi: Regenerate.
1679 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
1680 * double-int.h: Include wide-int.h.
1681 (struct wi::int_traits): New.
1682 * dwarf2out.c (get_full_len): New.
1683 (dw_val_equal_p): Add case dw_val_class_wide_int.
1684 (size_of_loc_descr): Likewise.
1685 (output_loc_operands): Likewise.
1686 (insert_double): Remove.
1687 (insert_wide_int): New.
1689 (print_die): Add case dw_val_class_wide_int.
1690 (attr_checksum): Likewise.
1691 (attr_checksum_ordered): Likewise.
1692 (same_dw_val_p): Likewise.
1693 (size_of_die): Likewise.
1694 (value_format): Likewise.
1695 (output_die): Likewise.
1696 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
1698 (clz_loc_descriptor): Use wide-int interfaces.
1699 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
1700 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
1701 (round_up_to_align): Use wide-int interfaces.
1702 (field_byte_offset): Likewise.
1703 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
1704 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
1705 CONST_DOUBLE handling. Use wide-int interfaces.
1706 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
1707 (gen_enumeration_type_die): Use add_AT_wide.
1708 (hash_loc_operands): Add case dw_val_class_wide_int.
1709 (compare_loc_operands): Likewise.
1710 * dwarf2out.h: Include wide-int.h.
1711 (wide_int_ptr): New.
1712 (enum dw_val_class): Add dw_val_class_wide_int.
1713 (struct dw_val_struct): Add val_wide.
1714 * emit-rtl.c (const_wide_int_htab): New.
1715 (const_wide_int_htab_hash): New.
1716 (const_wide_int_htab_eq): New.
1717 (lookup_const_wide_int): New.
1718 (const_double_htab_hash): Use wide-int interfaces.
1719 (const_double_htab_eq): Likewise.
1720 (rtx_to_double_int): Conditionally compile for wide-int.
1721 (immed_double_int_const): Rename to immed_wide_int_const and
1722 update for wide-int.
1723 (immed_double_const): Conditionally compile for wide-int.
1724 (init_emit_once): Use wide-int interfaces.
1725 * explow.c (plus_constant): Likewise.
1726 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
1727 (lshift_value): Use wide-int interfaces.
1728 (expand_mult): Likewise.
1729 (choose_multiplier): Likewise.
1730 (expand_smod_pow2): Likewise.
1731 (make_tree): Likewise.
1732 * expr.c (convert_modes): Consolidate handling of constants.
1733 Use wide-int interfaces.
1734 (emit_group_load_1): Add note.
1735 (store_expr): Update comment.
1736 (get_inner_reference): Use wide-int interfaces.
1737 (expand_constructor): Update comment.
1738 (expand_expr_real_2): Use wide-int interfaces.
1739 (expand_expr_real_1): Likewise.
1740 (reduce_to_bit_field_precision): Likewise.
1741 (const_vector_from_tree): Likewise.
1742 * final.c: Include wide-int-print.h.
1743 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
1744 * fixed-value.c: Include wide-int.h.
1745 (fixed_from_string): Use wide-int interfaces.
1746 (fixed_to_decimal): Likewise.
1747 (fixed_convert_from_real): Likewise.
1748 (real_convert_from_fixed): Likewise.
1749 * fold-const.h (mem_ref_offset): Return an offset_int.
1750 (div_if_zero_remainder): Remove code parameter.
1751 * fold-const.c (div_if_zero_remainder): Remove code parameter.
1752 Use wide-int interfaces.
1753 (may_negate_without_overflow_p): Use wide-int interfaces.
1754 (negate_expr_p): Likewise.
1755 (fold_negate_expr): Likewise.
1756 (int_const_binop_1): Likewise.
1757 (const_binop): Likewise.
1758 (fold_convert_const_int_from_int): Likewise.
1759 (fold_convert_const_int_from_real): Likewise.
1760 (fold_convert_const_int_from_fixed): Likewise.
1761 (fold_convert_const_fixed_from_int): Likewise.
1762 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
1763 (sign_bit_p): Use wide-int interfaces.
1764 (make_range_step): Likewise.
1765 (build_range_check): Likewise. Pass an integer of the correct type
1766 instead of using integer_one_node.
1767 (range_predecessor): Pass an integer of the correct type instead
1768 of using integer_one_node.
1769 (range_successor): Likewise.
1770 (merge_ranges): Likewise.
1771 (unextend): Use wide-int interfaces.
1772 (extract_muldiv_1): Likewise.
1773 (fold_div_compare): Likewise.
1774 (fold_single_bit_test): Likewise.
1775 (fold_sign_changed_comparison): Likewise.
1776 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
1777 (fold_plusminus_mult_expr): Use wide-int interfaces.
1778 (native_encode_int): Likewise.
1779 (native_interpret_int): Likewise.
1780 (fold_unary_loc): Likewise.
1781 (pointer_may_wrap_p): Likewise.
1782 (size_low_cst): Likewise.
1783 (mask_with_tz): Likewise.
1784 (fold_binary_loc): Likewise.
1785 (fold_ternary_loc): Likewise.
1786 (multiple_of_p): Likewise.
1787 (tree_call_nonnegative_warnv_p): Update calls to
1788 tree_int_cst_min_precision and real_from_integer.
1789 (fold_negate_const): Use wide-int interfaces.
1790 (fold_abs_const): Likewise.
1791 (fold_relational_const): Use tree_int_cst_lt.
1792 (round_up_loc): Use wide-int interfaces.
1793 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
1794 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
1795 * gengtype.c: Remove include of double-int.h.
1796 (do_typedef): Use wide-int interfaces.
1797 (open_base_files): Add wide-int.h.
1798 (main): Add offset_int and widest_int typedefs.
1799 * gengtype-lex.l: Handle "^".
1800 (CXX_KEYWORD): Add "static".
1801 * gengtype-parse.c (require3): New.
1802 (require_template_declaration): Handle constant template arguments
1803 and nested templates.
1804 * gengtype-state.c: Don't include "double-int.h".
1805 * genpreds.c (write_one_predicate_function): Update comment.
1806 (write_tm_constrs_h): Add check for hval and lval use in
1808 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
1809 (add_to_sequence): Likewise.
1810 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
1811 and const_double_operand.
1812 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
1814 * gimple-fold.c (get_base_constructor): Likewise.
1815 (fold_array_ctor_reference): Likewise.
1816 (fold_nonarray_ctor_reference): Likewise.
1817 (fold_const_aggregate_ref_1): Likewise.
1818 (gimple_val_nonnegative_real_p): Likewise.
1819 (gimple_fold_indirect_ref): Likewise.
1820 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
1821 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
1822 (struct slsr_cand_d): Change index to be widest_int.
1823 (struct incr_info_d): Change incr to be widest_int.
1824 (alloc_cand_and_find_basis): Use wide-int interfaces.
1825 (slsr_process_phi): Likewise.
1826 (backtrace_base_for_ref): Likewise. Return a widest_int.
1827 (restructure_reference): Take a widest_int instead of a double_int.
1828 (slsr_process_ref): Use wide-int interfaces.
1829 (create_mul_ssa_cand): Likewise.
1830 (create_mul_imm_cand): Likewise.
1831 (create_add_ssa_cand): Likewise.
1832 (create_add_imm_cand): Take a widest_int instead of a double_int.
1833 (slsr_process_add): Use wide-int interfaces.
1834 (slsr_process_cast): Likewise.
1835 (slsr_process_copy): Likewise.
1836 (dump_candidate): Likewise.
1837 (dump_incr_vec): Likewise.
1838 (replace_ref): Likewise.
1839 (cand_increment): Likewise. Return a widest_int.
1840 (cand_abs_increment): Likewise.
1841 (replace_mult_candidate): Take a widest_int instead of a double_int.
1842 (replace_unconditional_candidate): Use wide-int interfaces.
1843 (incr_vec_index): Take a widest_int instead of a double_int.
1844 (create_add_on_incoming_edge): Likewise.
1845 (create_phi_basis): Use wide-int interfaces.
1846 (replace_conditional_candidate): Likewise.
1847 (record_increment): Take a widest_int instead of a double_int.
1848 (record_phi_increments): Use wide-int interfaces.
1849 (phi_incr_cost): Take a widest_int instead of a double_int.
1850 (lowest_cost_path): Likewise.
1851 (total_savings): Likewise.
1852 (analyze_increments): Use wide-int interfaces.
1853 (ncd_with_phi): Take a widest_int instead of a double_int.
1854 (ncd_of_cand_and_phis): Likewise.
1855 (nearest_common_dominator_for_cands): Likewise.
1856 (insert_initializers): Use wide-int interfaces.
1857 (all_phi_incrs_profitable): Likewise.
1858 (replace_one_candidate): Likewise.
1859 (replace_profitable_candidates): Likewise.
1860 * godump.c: Include wide-int-print.h.
1861 (go_output_typedef): Use wide-int interfaces.
1862 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
1863 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
1864 (build_loop_iteration_domains): Likewise.
1865 * hooks.h: Include wide-int.h rather than double-int.h.
1866 (hook_bool_dint_dint_uint_bool_true): Delete.
1867 (hook_bool_wint_wint_uint_bool_true): Declare.
1868 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
1869 (hook_bool_wint_wint_uint_bool_true): New.
1870 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
1872 (ubsan_expand_si_overflow_mul_check): Likewise.
1873 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
1874 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
1875 (get_ancestor_addr_info): Likewise.
1876 (ipa_modify_call_arguments): Likewise.
1877 * loop-doloop.c (doloop_modify): Likewise.
1878 (doloop_optimize): Likewise.
1879 * loop-iv.c (iv_number_of_iterations): Likewise.
1880 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
1881 (unroll_loop_constant_iterations): Likewise.
1882 (decide_unroll_runtime_iterations): Likewise.
1883 (unroll_loop_runtime_iterations): Likewise.
1884 (decide_peel_simple): Likewise.
1885 (decide_unroll_stupid): Likewise.
1886 * lto-streamer-in.c (streamer_read_wi): Add.
1887 (input_cfg): Use wide-int interfaces.
1888 (lto_input_tree_1): Likewise.
1889 * lto-streamer-out.c (streamer_write_wi): Add.
1890 (hash_tree): Use wide-int interfaces.
1891 (output_cfg): Likewise.
1892 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
1893 (GTFILES): Add wide-int.h and signop.h.
1894 (TAGS): Look for .cc files too.
1895 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
1896 * optabs.c (expand_subword_shift): Likewise.
1897 (expand_doubleword_shift): Likewise.
1898 (expand_absneg_bit): Likewise.
1899 (expand_copysign_absneg): Likewise.
1900 (expand_copysign_bit): Likewise.
1901 * postreload.c (reload_cse_simplify_set): Likewise.
1902 * predict.c (predict_iv_comparison): Likewise.
1903 * pretty-print.h: Include wide-int-print.h.
1905 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
1906 * print-tree.c: Include wide-int-print.h.
1907 (print_node_brief): Use wide-int interfaces.
1908 (print_node): Likewise.
1909 * read-rtl.c (validate_const_wide_int): New.
1910 (read_rtx_code): Add CONST_WIDE_INT case.
1911 * real.c: Include wide-int.h.
1912 (real_to_integer2): Delete.
1913 (real_to_integer): New function, returning a wide_int.
1914 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
1915 (ten_to_ptwo): Update call to real_from_integer.
1916 (real_digit): Likewise.
1917 * real.h: Include signop.h, wide-int.h and insn-modes.h.
1918 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
1919 (REAL_VALUE_TO_INT): Delete.
1920 (real_to_integer): Declare a wide-int form.
1921 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
1922 * recog.c (const_int_operand): Improve comment.
1923 (const_scalar_int_operand): New.
1924 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
1925 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
1926 (split_double): Likewise.
1927 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
1928 (rtx_size): Likewise.
1929 (rtx_alloc_stat_v): New.
1930 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
1931 (cwi_output_hex): New.
1932 (iterative_hash_rtx): Handle CONST_WIDE_INT.
1933 (cwi_check_failed_bounds): New.
1934 * rtl.def (CONST_WIDE_INT): New.
1935 * rtl.h: Include <utility> and wide-int.h.
1936 (struct hwivec_def): New.
1937 (CWI_GET_NUM_ELEM): New.
1938 (CWI_PUT_NUM_ELEM): New.
1939 (struct rtx_def): Add num_elem and hwiv.
1940 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
1941 (CASE_CONST_UNIQUE): Likewise.
1942 (CASE_CONST_ANY): Likewise.
1943 (CONST_SCALAR_INT_P): Likewise.
1944 (CONST_WIDE_INT_P): New.
1946 (HWIVEC_CHECK): New.
1947 (cwi_check_failed_bounds): New.
1949 (HWIVEC_CHECK): New.
1950 (CONST_WIDE_INT_VEC) New.
1951 (CONST_WIDE_INT_NUNITS) New.
1952 (CONST_WIDE_INT_ELT) New.
1953 (rtx_mode_t): New type.
1954 (wi::int_traits <rtx_mode_t>): New.
1956 (wi::min_value): New.
1957 (wi::max_value): New.
1959 (const_wide_int_alloc): New.
1960 (immed_wide_int_const): New.
1961 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
1962 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
1963 * signop.h: New file.
1964 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
1965 (simplify_const_unary_operation): Use wide-int interfaces.
1966 (simplify_binary_operation_1): Likewise.
1967 (simplify_const_binary_operation): Likewise.
1968 (simplify_const_relational_operation): Likewise.
1969 (simplify_immed_subreg): Likewise.
1970 * stmt.c (expand_case): Likewise.
1971 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
1972 signop rather than a bool.
1973 * stor-layout.c (layout_type): Use wide-int interfaces.
1974 (initialize_sizetypes): Update calls to
1975 set_min_and_max_values_for_integral_type.
1976 (set_min_and_max_values_for_integral_type): Take a signop rather
1977 than a bool. Use wide-int interfaces.
1978 (fixup_signed_type): Update accordingly. Remove
1979 HOST_BITS_PER_DOUBLE_INT limit.
1980 (fixup_unsigned_type): Likewise.
1981 * system.h (STATIC_CONSTANT_P): New.
1982 (STATIC_ASSERT): New.
1983 * target.def (can_use_doloop_p): Take widest_ints rather than
1985 * target.h: Include wide-int.h rather than double-int.h.
1986 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
1988 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
1989 rather than INT_CST_LT_UNSIGNED.
1990 (can_use_doloop_if_innermost): Take widest_ints rather than
1992 * tree-affine.c: Include wide-int-print.h.
1993 (double_int_ext_for_comb): Delete.
1994 (wide_int_ext_for_comb): New.
1995 (aff_combination_zero): Use wide-int interfaces.
1996 (aff_combination_const): Take a widest_int instead of a double_int.
1997 (aff_combination_elt): Use wide-int interfaces.
1998 (aff_combination_scale): Take a widest_int instead of a double_int.
1999 (aff_combination_add_elt): Likewise.
2000 (aff_combination_add_cst): Likewise.
2001 (aff_combination_add): Use wide-int interfaces.
2002 (aff_combination_convert): Likewise.
2003 (tree_to_aff_combination): Likewise.
2004 (add_elt_to_tree): Take a widest_int instead of a double_int.
2005 (aff_combination_to_tree): Use wide-int interfaces.
2006 (aff_combination_remove_elt): Likewise.
2007 (aff_combination_add_product): Take a widest_int instead of
2009 (aff_combination_mult): Use wide-int interfaces.
2010 (aff_combination_expand): Likewise.
2011 (double_int_constant_multiple_p): Delete.
2012 (wide_int_constant_multiple_p): New.
2013 (aff_combination_constant_multiple_p): Take a widest_int pointer
2014 instead of a double_int pointer.
2015 (print_aff): Use wide-int interfaces.
2016 (get_inner_reference_aff): Take a widest_int pointer
2017 instead of a double_int pointer.
2018 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
2019 * tree-affine.h: Include wide-int.h.
2020 (struct aff_comb_elt): Change type of coef to widest_int.
2021 (struct affine_tree_combination): Change type of offset to widest_int.
2022 (double_int_ext_for_comb): Delete.
2023 (wide_int_ext_for_comb): New.
2024 (aff_combination_const): Use widest_int instead of double_int.
2025 (aff_combination_scale): Likewise.
2026 (aff_combination_add_elt): Likewise.
2027 (aff_combination_constant_multiple_p): Likewise.
2028 (get_inner_reference_aff): Likewise.
2029 (aff_comb_cannot_overlap_p): Likewise.
2030 (aff_combination_zero_p): Use wide-int interfaces.
2031 * tree.c: Include tree.h.
2032 (init_ttree): Use make_int_cst.
2033 (tree_code_size): Removed code for INTEGER_CST case.
2034 (tree_size): Add INTEGER_CST case.
2035 (make_node_stat): Update comment.
2036 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
2037 (build_int_cst_type): Use wide-int interfaces.
2038 (double_int_to_tree): Likewise.
2039 (double_int_fits_to_tree_p): Delete.
2040 (force_fit_type_double): Delete.
2041 (force_fit_type): New.
2042 (int_cst_hash_hash): Use wide-int interfaces.
2043 (int_cst_hash_eq): Likewise.
2044 (build_int_cst_wide): Delete.
2045 (wide_int_to_tree): New.
2046 (cache_integer_cst): Use wide-int interfaces.
2047 (build_low_bits_mask): Likewise.
2048 (cst_and_fits_in_hwi): Likewise.
2049 (real_value_from_int_cst): Likewise.
2050 (make_int_cst_stat): New.
2051 (integer_zerop): Use wide_int interfaces.
2052 (integer_onep): Likewise.
2053 (integer_all_onesp): Likewise.
2054 (integer_pow2p): Likewise.
2055 (integer_nonzerop): Likewise.
2056 (tree_log2): Likewise.
2057 (tree_floor_log2): Likewise.
2058 (tree_ctz): Likewise.
2059 (int_size_in_bytes): Likewise.
2060 (mem_ref_offset): Return an offset_int rather than a double_int.
2061 (build_type_attribute_qual_variant): Use wide_int interfaces.
2062 (type_hash_eq): Likewise
2063 (tree_int_cst_equal): Likewise.
2064 (tree_int_cst_lt): Delete.
2065 (tree_int_cst_compare): Likewise.
2066 (tree_fits_shwi_p): Use wide_int interfaces.
2067 (tree_fits_uhwi_p): Likewise.
2068 (tree_int_cst_sign_bit): Likewise.
2069 (tree_int_cst_sgn): Likewise.
2070 (tree_int_cst_min_precision): Take a signop rather than a bool.
2071 (simple_cst_equal): Use wide_int interfaces.
2072 (compare_tree_int): Likewise.
2073 (iterative_hash_expr): Likewise.
2074 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
2076 (get_type_static_bounds): Use wide_int interfaces.
2077 (tree_int_cst_elt_check_failed): New.
2078 (build_common_tree_nodes): Reordered to set prec before filling in
2080 (int_cst_value): Check cst_and_fits_in_hwi.
2081 (widest_int_cst_value): Use wide_int interfaces.
2082 (upper_bound_in_type): Likewise.
2083 (lower_bound_in_type): Likewise.
2084 (num_ending_zeros): Likewise.
2085 (drop_tree_overflow): Likewise.
2086 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
2087 (gen_conditions_for_pow_cst_base): Likewise.
2088 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
2089 (group_case_labels_stmt): Use wide-int interfaces.
2090 (verify_gimple_assign_binary): Likewise.
2091 (print_loop): Likewise.
2092 * tree-chrec.c (tree_fold_binomial): Likewise.
2093 * tree-core.h (struct tree_base): Add int_length.
2094 (struct tree_int_cst): Change rep of value.
2095 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
2096 (dr_may_alias_p): Likewise.
2097 (max_stmt_executions_tree): Likewise.
2098 * tree.def (INTEGER_CST): Update comment.
2099 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
2100 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
2101 * tree-dump.c: Include wide-int.h and wide-int-print.h.
2102 (dequeue_and_dump): Use wide-int interfaces.
2103 * tree.h: Include wide-int.h.
2104 (NULL_TREE): Moved to earlier loc in file.
2105 (TREE_INT_CST_ELT_CHECK): New.
2106 (tree_int_cst_elt_check_failed): New.
2108 (TREE_INT_CST): Delete.
2109 (TREE_INT_CST_LOW): Use wide-int interfaces.
2110 (TREE_INT_CST_HIGH): Delete.
2111 (TREE_INT_CST_NUNITS): New.
2112 (TREE_INT_CST_EXT_NUNITS): Likewise.
2113 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
2114 (TREE_INT_CST_ELT): Likewise.
2115 (INT_CST_LT): Delete.
2116 (tree_int_cst_elt_check): New (two forms).
2117 (type_code_size): Update comment.
2118 (make_int_cst_stat, make_int_cst): New.
2119 (tree_to_double_int): Delete.
2120 (double_int_fits_to_tree_p): Delete.
2121 (force_fit_type_double): Delete.
2122 (build_int_cstu): Replace with out-of-line function.
2123 (build_int_cst_wide): Delete.
2124 (tree_int_cst_lt): Define inline.
2125 (tree_int_cst_le): New.
2126 (tree_int_cst_compare): Define inline.
2127 (tree_int_cst_min_precision): Take a signop rather than a bool.
2128 (wi::int_traits <const_tree>): New.
2129 (wi::int_traits <tree>): New.
2130 (wi::extended_tree): New.
2131 (wi::int_traits <wi::extended_tree>): New.
2132 (wi::to_widest): New.
2133 (wi::to_offset): New.
2134 (wi::fits_to_tree_p): New.
2135 (wi::min_value): New.
2136 (wi::max_value): New.
2137 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
2138 (copy_tree_body_r): Likewise.
2139 * tree-object-size.c (compute_object_offset): Likewise.
2140 (addr_object_size): Likewise.
2141 * tree-predcom.c: Include wide-int-print.h.
2142 (struct dref_d): Change type of offset to widest_int.
2143 (dump_dref): Call wide-int printer.
2144 (aff_combination_dr_offset): Use wide-int interfaces.
2145 (determine_offset): Take a widest_int pointer rather than a
2147 (split_data_refs_to_components): Use wide-int interfaces.
2148 (suitable_component_p): Likewise.
2149 (order_drefs): Likewise.
2150 (add_ref_to_chain): Likewise.
2151 (valid_initializer_p): Likewise.
2152 (determine_roots_comp): Likewise.
2153 * tree-pretty-print.c: Include wide-int-print.h.
2154 (dump_generic_node): Use wide-int interfaces.
2155 * tree-sra.c (sra_ipa_modify_expr): Likewise.
2156 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
2157 (move_fixed_address_to_symbol): Likewise.
2158 (move_hint_to_base): Likewise.
2159 (move_pointer_to_base): Likewise.
2160 (move_variant_to_index): Likewise.
2161 (most_expensive_mult_to_index): Likewise.
2162 (addr_to_parts): Likewise.
2163 (copy_ref_info): Likewise.
2164 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
2165 (indirect_refs_may_alias_p): Likewise.
2166 (stmt_kills_ref_p_1): Likewise.
2167 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
2168 * tree-ssa-ccp.c: Update comment at top of file. Include
2170 (struct prop_value_d): Change type of mask to widest_int.
2171 (extend_mask): New function.
2172 (dump_lattice_value): Use wide-int interfaces.
2173 (get_default_value): Likewise.
2174 (set_constant_value): Likewise.
2175 (set_value_varying): Likewise.
2176 (valid_lattice_transition): Likewise.
2177 (set_lattice_value): Likewise.
2178 (value_to_double_int): Delete.
2179 (value_to_wide_int): New.
2180 (get_value_from_alignment): Use wide-int interfaces.
2181 (get_value_for_expr): Likewise.
2182 (do_dbg_cnt): Likewise.
2183 (ccp_finalize): Likewise.
2184 (ccp_lattice_meet): Likewise.
2185 (bit_value_unop_1): Use widest_ints rather than double_ints.
2186 (bit_value_binop_1): Likewise.
2187 (bit_value_unop): Use wide-int interfaces.
2188 (bit_value_binop): Likewise.
2189 (bit_value_assume_aligned): Likewise.
2190 (evaluate_stmt): Likewise.
2191 (ccp_fold_stmt): Likewise.
2192 (visit_cond_stmt): Likewise.
2193 (ccp_visit_stmt): Likewise.
2194 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
2195 (constant_pointer_difference): Likewise.
2196 (associate_pointerplus): Likewise.
2197 (combine_conversions): Likewise.
2198 * tree-ssa-loop.h: Include wide-int.h.
2199 (struct tree_niter_desc): Change type of max to widest_int.
2200 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
2201 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
2202 (remove_redundant_iv_tests): Likewise.
2203 (canonicalize_loop_induction_variables): Likewise.
2204 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
2205 (constant_multiple_of): Take a widest_int pointer instead of
2206 a double_int pointer.
2207 (get_computation_aff): Use wide-int interfaces.
2208 (ptr_difference_cost): Likewise.
2209 (difference_cost): Likewise.
2210 (get_loop_invariant_expr_id): Likewise.
2211 (get_computation_cost_at): Likewise.
2212 (iv_elimination_compare_lt): Likewise.
2213 (may_eliminate_iv): Likewise.
2214 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
2215 instead of double_int.
2216 (max_loop_iterations): Likewise.
2217 (max_stmt_executions): Likewise.
2218 (estimated_stmt_executions): Likewise.
2219 * tree-ssa-loop-niter.c: Include wide-int-print.h.
2220 (split_to_var_and_offset): Use wide-int interfaces.
2221 (determine_value_range): Likewise.
2222 (bound_difference_of_offsetted_base): Likewise.
2223 (bounds_add): Take a widest_int instead of a double_int.
2224 (number_of_iterations_ne_max): Use wide-int interfaces.
2225 (number_of_iterations_ne): Likewise.
2226 (number_of_iterations_lt_to_ne): Likewise.
2227 (assert_loop_rolls_lt): Likewise.
2228 (number_of_iterations_lt): Likewise.
2229 (number_of_iterations_le): Likewise.
2230 (number_of_iterations_cond): Likewise.
2231 (number_of_iterations_exit): Likewise.
2232 (finite_loop_p): Likewise.
2233 (derive_constant_upper_bound_assign): Likewise.
2234 (derive_constant_upper_bound): Return a widest_int.
2235 (derive_constant_upper_bound_ops): Likewise.
2236 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
2237 (record_estimate): Take a widest_int rather than a double_int.
2238 (record_nonwrapping_iv): Use wide-int interfaces.
2239 (double_int_cmp): Delete.
2240 (wide_int_cmp): New.
2241 (bound_index): Take a widest_int rather than a double_int.
2242 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
2243 (maybe_lower_iteration_bound): Likewise.
2244 (estimate_numbers_of_iterations_loop): Likewise.
2245 (estimated_loop_iterations): Take a widest_int pointer than than
2246 a double_int pointer.
2247 (estimated_loop_iterations_int): Use wide-int interfaces.
2248 (max_loop_iterations): Take a widest_int pointer than than
2249 a double_int pointer.
2250 (max_loop_iterations_int): Use wide-int interfaces.
2251 (max_stmt_executions): Take a widest_int pointer than than
2252 a double_int pointer.
2253 (estimated_stmt_executions): Likewise.
2254 (n_of_executions_at_most): Use wide-int interfaces.
2255 (scev_probably_wraps_p): Likewise.
2256 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
2258 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
2260 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
2261 double_ints. Adjust for trailing_wide_ints <3> representation.
2262 (set_nonzero_bits): Likewise.
2263 (get_range_info): Return wide_ints rather than double_ints.
2264 Adjust for trailing_wide_ints <3> representation.
2265 (get_nonzero_bits): Likewise.
2266 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
2268 * tree-ssanames.h (struct range_info_def): Replace min, max and
2269 nonzero_bits with a trailing_wide_ints <3>.
2270 (set_range_info): Use wide_int_refs rather than double_ints.
2271 (set_nonzero_bits): Likewise.
2272 (get_range_info): Return wide_ints rather than double_ints.
2273 (get_nonzero_bits): Likewise.
2274 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
2275 * tree-ssa-pre.c (phi_translate_1): Likewise.
2276 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
2277 (acceptable_pow_call): Likewise.
2278 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
2280 (vn_reference_fold_indirect): Likewise.
2281 (vn_reference_maybe_forwprop_address): Likewise.
2282 (valueize_refs_1): Likewise.
2283 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
2284 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
2285 tree_int_cst_lt and tree_int_cst_le.
2286 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
2288 (streamer_alloc_tree): Likewise.
2289 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
2290 (streamer_write_tree_header): Likewise.
2291 (streamer_write_integer_cst): Likewise.
2292 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
2293 (build_constructors): Likewise.
2294 (array_value_type): Likewise.
2295 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
2296 (vect_check_gather): Likewise.
2297 * tree-vect-generic.c (build_replicated_const): Likewise.
2298 (expand_vector_divmod): Likewise.
2299 * tree-vect-loop.c (vect_transform_loop): Likewise.
2300 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
2301 (vect_do_peeling_for_alignment): Likewise.
2302 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
2303 * tree-vrp.c: Include wide-int.h.
2304 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
2305 (extract_range_from_assert): Use wide-int interfaces.
2306 (vrp_int_const_binop): Likewise.
2307 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
2308 double_int pointers.
2309 (ranges_from_anti_range): Use wide-int interfaces.
2310 (quad_int_cmp): Delete.
2311 (quad_int_pair_sort): Likewise.
2312 (extract_range_from_binary_expr_1): Use wide-int interfaces.
2313 (extract_range_from_unary_expr_1): Likewise.
2314 (adjust_range_with_scev): Likewise.
2315 (masked_increment): Take and return wide_ints rather than double_ints.
2316 (register_edge_assert_for_2): Use wide-int interfaces.
2317 (check_array_ref): Likewise.
2318 (search_for_addr_array): Likewise.
2319 (maybe_set_nonzero_bits): Likewise.
2320 (union_ranges): Pass an integer of the correct type instead of
2321 using integer_one_node.
2322 (intersect_ranges): Likewise.
2323 (simplify_truth_ops_using_ranges): Likewise.
2324 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
2325 (range_fits_type_p): Likewise.
2326 (simplify_cond_using_ranges): Likewise. Take a signop rather than
2328 (simplify_conversion_using_ranges): Use wide-int interfaces.
2329 (simplify_float_conversion_using_ranges): Likewise.
2330 (vrp_finalize): Likewise.
2331 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
2332 (gimple_stringops_transform): Likewise.
2333 * varasm.c (decode_addr_const): Likewise.
2334 (const_hash_1): Likewise.
2335 (const_rtx_hash_1): Likewise
2336 (output_constant): Likewise.
2337 (array_size_for_constructor): Likewise.
2338 (output_constructor_regular_field): Likewise.
2339 (output_constructor_bitfield): Likewise.
2340 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
2341 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
2343 * gencheck.c: Define BITS_PER_UNIT.
2346 * wide-int-print.cc: New.
2347 * wide-int-print.h: New.
2349 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2351 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
2353 2014-05-06 Richard Biener <rguenther@suse.de>
2355 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
2356 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
2357 (TODO_verify_all): Adjust.
2358 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
2359 TODO_verify_stmts and TODO_verify_rtl_sharing.
2360 * bb-reorder.c: Likewise.
2361 * cfgexpand.c: Likewise.
2362 * cprop.c: Likewise.
2364 * function.c: Likewise.
2365 * fwprop.c: Likewise.
2367 * gimple-ssa-isolate-paths.c: Likewise.
2368 * gimple-ssa-strength-reduction.c: Likewise.
2369 * ipa-split.c: Likewise.
2370 * loop-init.c: Likewise.
2371 * loop-unroll.c: Likewise.
2372 * lower-subreg.c: Likewise.
2373 * modulo-sched.c: Likewise.
2374 * postreload-gcse.c: Likewise.
2375 * predict.c: Likewise.
2376 * recog.c: Likewise.
2377 * sched-rgn.c: Likewise.
2378 * store-motion.c: Likewise.
2379 * tracer.c: Likewise.
2380 * trans-mem.c: Likewise.
2381 * tree-call-cdce.c: Likewise.
2382 * tree-cfg.c: Likewise.
2383 * tree-cfgcleanup.c: Likewise.
2384 * tree-complex.c: Likewise.
2385 * tree-eh.c: Likewise.
2386 * tree-emutls.c: Likewise.
2387 * tree-if-conv.c: Likewise.
2388 * tree-into-ssa.c: Likewise.
2389 * tree-loop-distribution.c: Likewise.
2390 * tree-object-size.c: Likewise.
2391 * tree-parloops.c: Likewise.
2392 * tree-pass.h: Likewise.
2393 * tree-sra.c: Likewise.
2394 * tree-ssa-ccp.c: Likewise.
2395 * tree-ssa-copy.c: Likewise.
2396 * tree-ssa-copyrename.c: Likewise.
2397 * tree-ssa-dce.c: Likewise.
2398 * tree-ssa-dom.c: Likewise.
2399 * tree-ssa-dse.c: Likewise.
2400 * tree-ssa-forwprop.c: Likewise.
2401 * tree-ssa-ifcombine.c: Likewise.
2402 * tree-ssa-loop-ch.c: Likewise.
2403 * tree-ssa-loop-ivcanon.c: Likewise.
2404 * tree-ssa-loop.c: Likewise.
2405 * tree-ssa-math-opts.c: Likewise.
2406 * tree-ssa-phiopt.c: Likewise.
2407 * tree-ssa-phiprop.c: Likewise.
2408 * tree-ssa-pre.c: Likewise.
2409 * tree-ssa-reassoc.c: Likewise.
2410 * tree-ssa-sink.c: Likewise.
2411 * tree-ssa-strlen.c: Likewise.
2412 * tree-ssa-tail-merge.c: Likewise.
2413 * tree-ssa-uncprop.c: Likewise.
2414 * tree-switch-conversion.c: Likewise.
2415 * tree-tailcall.c: Likewise.
2416 * tree-vect-generic.c: Likewise.
2417 * tree-vectorizer.c: Likewise.
2418 * tree-vrp.c: Likewise.
2420 * var-tracking.c: Likewise.
2421 * bt-load.c: Likewise.
2422 * cfgcleanup.c: Likewise.
2423 * combine-stack-adj.c: Likewise.
2424 * combine.c: Likewise.
2425 * compare-elim.c: Likewise.
2426 * config/epiphany/resolve-sw-modes.c: Likewise.
2427 * config/i386/i386.c: Likewise.
2428 * config/mips/mips.c: Likewise.
2429 * config/s390/s390.c: Likewise.
2430 * config/sh/sh_treg_combine.cc: Likewise.
2431 * config/sparc/sparc.c: Likewise.
2434 * final.c: Likewise.
2435 * ifcvt.c: Likewise.
2436 * mode-switching.c: Likewise.
2437 * passes.c: Likewise.
2438 * postreload.c: Likewise.
2440 * reg-stack.c: Likewise.
2441 * regcprop.c: Likewise.
2442 * regrename.c: Likewise.
2445 2014-05-06 Richard Biener <rguenther@suse.de>
2448 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
2449 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
2451 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
2454 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
2456 2014-05-05 Radovan Obradovic <robradovic@mips.com>
2457 Tom de Vries <tom@codesourcery.com>
2459 * target.def (call_fusage_contains_non_callee_clobbers): New
2461 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
2463 (@node Miscellaneous Register Hooks): New node.
2464 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
2465 * doc/tm.texi: Regenerate.
2467 2014-05-05 Marek Polacek <polacek@redhat.com>
2470 * opts.c (common_handle_option): Call error_at instead of warning_at.
2472 2014-05-05 Richard Biener <rguenther@suse.de>
2474 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
2475 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
2476 under the TODO_verify_il umbrella.
2478 2014-05-05 Richard Biener <rguenther@suse.de>
2480 * passes.c (execute_function_todo): Move TODO_verify_flow under
2481 the TODO_verify_ul umbrella.
2483 2014-05-05 Richard Biener <rguenther@suse.de>
2486 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
2487 X & CST away from a CST that is the mask of a mode.
2489 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2491 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
2492 int argument to enum machine_mode.
2493 (picochip_class_max_nregs): Ditto.
2494 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
2495 (picochip_class_max_nregs): Ditto.
2497 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2499 * target.def: Add new target hook.
2500 * doc/tm.texi: Regenerate.
2501 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
2502 * targhooks.c (default_keep_leaf_when_profiled): New function.
2504 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
2505 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
2507 2014-05-05 Bin Cheng <bin.cheng@arm.com>
2509 PR tree-optimization/60363
2510 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
2511 (copy_phi_args): New parameters. Call get_value_locus_in_path.
2512 (update_destination_phis): New parameter.
2513 (create_edge_and_update_destination_phis): Ditto.
2514 (ssa_fix_duplicate_block_edges): Pass new arguments.
2515 (thread_single_edge): Ditto.
2517 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
2519 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
2520 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
2521 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
2522 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
2523 Use RS6000_BTM_HARD_FLOAT.
2524 (BU_MISC_2): Likewise.
2525 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
2526 RS6000_BTM_HARD_FLOAT.
2527 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
2529 (rs6000_invalid_builtin): Add hard floating builtin support.
2530 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
2531 hard float builtins.
2532 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
2534 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2536 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
2537 Add missing function* argument.
2539 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2541 * lra-constraints.c (valid_address_p): Move earlier in file.
2542 Add a constraint argument to the address_info version.
2543 (satisfies_memory_constraint_p): New function.
2544 (satisfies_address_constraint_p): Likewise.
2545 (process_alt_operands, curr_insn_transform): Use them.
2546 (process_address): Pass the constraint to valid_address_p when
2547 checking address operands.
2549 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2551 * config/mips/mips.c (mips_isa_rev): New variable.
2552 (mips_set_architecture): Set it.
2553 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
2555 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
2556 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
2557 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
2558 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
2559 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
2560 conditions in terms of mips_isa_rev.
2561 (mips_isa_rev): Declare.
2563 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2565 * config/sh/sh-mem.cc: Use tabs instead of spaces.
2566 (prob_unlikely, prob_likely): Make variables const.
2568 2014-05-03 Denis Chertykov <chertykov@gmail.com>
2570 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
2572 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2574 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
2576 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2578 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
2579 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
2580 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
2582 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
2583 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
2585 Replace usage of ROUND_REG with sh_round_reg.
2586 Use CEIL instead of ROUND_ADVANCE.
2588 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2591 * config/sh/sh.c: Include stdlib headers before everything else.
2593 2014-05-02 Jakub Jelinek <jakub@redhat.com>
2595 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
2596 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
2597 (gimplify_adjust_omp_clauses): Simd region is never
2598 directly nested in combined parallel. Instead, for linear
2599 with copyin/copyout, if in combined for simd loop, make decl
2600 firstprivate/lastprivate on OMP_FOR.
2601 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
2602 expand_omp_for_static_chunk): When setting endvar, also set
2603 fd->loop.v to the same value.
2605 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2607 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
2609 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
2611 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
2614 2014-05-02 Marek Polacek <polacek@redhat.com>
2616 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
2618 2014-05-02 Kito Cheng <kito@0xlab.org>
2620 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
2621 to a C expression marco.
2622 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
2623 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
2624 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
2625 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
2626 HONOR_REG_ALLOC_ORDER.
2627 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
2629 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2631 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
2633 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2635 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
2637 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
2639 * tree-if-conv.c (is_cond_scalar_reduction): New function.
2640 (convert_scalar_cond_reduction): Likewise.
2641 (predicate_scalar_phi): Add recognition and transformation
2642 of simple conditioanl reduction to be vectorizable.
2644 2014-05-01 Marek Polacek <polacek@redhat.com>
2647 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
2649 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
2651 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
2652 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
2653 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
2654 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
2655 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
2656 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
2657 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
2658 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
2660 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
2662 * config/arc/arc.opt (mlra): Move comment above option name
2663 to avoid mis-parsing as language options.
2665 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2667 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
2668 * config/sol2.h: ... here.
2669 * config/sol2-10.h: Remove.
2671 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
2672 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
2673 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
2674 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
2675 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
2676 * config/sol2.h: ... here.
2677 (SECTION_NAME_FORMAT): Don't redefine.
2678 (STARTFILE_ARCH32_SPEC): Rename to ...
2679 (STARTFILE_ARCH_SPEC): ... this.
2680 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
2681 * config/sparc/sol2.h: ... here.
2682 (SECTION_NAME_FORMAT): Don't undef.
2683 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
2684 (SUBTARGET_EXTRA_SPECS): Remove.
2685 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
2687 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
2688 (MD_STARTFILE_PREFIX): Remove.
2689 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
2690 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
2691 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
2692 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
2693 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
2694 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
2695 * config/i386/sol2.h: ... here.
2696 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
2697 * config/i386/sol2-bi.h: Remove.
2698 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
2699 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
2701 * config/i386/t-sol2-64: Rename to ...
2702 * config/i386/t-sol2: ... this.
2703 * config/sparc/t-sol2-64: Rename to ...
2704 * config/sparc/t-sol2: ... this.
2706 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
2707 sol2_tm_file_head, sol2_tm_file_tail.
2708 Include ${cpu_type}/sol2.h before sol2.h.
2710 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
2711 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
2712 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
2713 Reflect i386/t-sol2-64 renaming.
2714 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
2715 Reflect sparc/t-sol2-64 renaming.
2717 2014-04-30 Richard Biener <rguenther@suse.de>
2719 * passes.c (execute_function_todo): Move TODO_verify_stmts
2720 and TODO_verify_ssa under the TODO_verify_il umbrella.
2721 * tree-ssa.h (verify_ssa): Adjust prototype.
2722 * tree-ssa.c (verify_ssa): Add parameter to tell whether
2723 we should verify SSA operands.
2724 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
2725 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
2726 whether we should verify whether not throwing stmts have EH info.
2727 * graphite-scop-detection.c (create_sese_edges): Adjust.
2728 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
2729 * tree-eh.c (lower_try_finally_switch): Do not add the
2730 default case label twice.
2732 2014-04-30 Marek Polacek <polacek@redhat.com>
2734 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
2735 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
2736 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
2737 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
2739 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
2741 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
2742 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
2743 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
2744 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
2745 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
2746 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
2747 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
2748 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
2750 2014-04-29 David Malcolm <dmalcolm@redhat.com>
2752 * tree-cfg.c (dump_function_to_file): Dump the return type of
2753 functions, in a line to itself before the function body, mimicking
2754 the layout of a C function.
2756 2014-04-29 Jakub Jelinek <jakub@redhat.com>
2758 PR tree-optimization/60971
2759 * tree-tailcall.c (process_assignment): Reject conversions which
2762 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
2764 * calls.c (initialize_argument_information): Always treat
2765 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
2766 (expand_call): Likewise.
2767 (emit_library_call_calue_1): Likewise.
2768 * expr.c (PUSH_ARGS_REVERSED): Do not define.
2769 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
2772 2014-04-29 Nick Clifton <nickc@redhat.com>
2774 * config/msp430/msp430.md (umulsidi): Fix typo.
2775 (mulhisi3): Enable even inside interrupt handlers.
2776 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
2777 bigger return address pushed in large mode.
2779 2014-04-29 Nick Clifton <nickc@redhat.com>
2781 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
2782 (arc_init_reg_tables): Use a machine_mode enum to iterate over
2784 * config/m32r/m32r.c (init_reg_tables): Likewise.
2785 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
2786 enum to hold the modes.
2788 2014-04-29 Richard Biener <rguenther@suse.de>
2790 * dominance.c (free_dominance_info): Add overload with
2792 (dom_info_state): Likewise.
2793 (dom_info_available_p): Likewise.
2794 * basic-block.h (free_dominance_info, dom_info_state,
2795 dom_info_available_p): Declare overloads.
2796 * passes.c (execute_function_todo): Verify that verifiers
2797 don't change dominator info state. Drop dominator info
2798 for IPA pass invocations.
2799 * cgraph.c (release_function_body): Restore asserts that
2800 dominator information is released.
2802 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
2804 * doc/invoke.texi: Fix typo.
2805 * tree-vrp.c: Fix typos.
2806 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
2808 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2810 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
2812 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
2814 * config/aarch64/aarch64-builtins.c
2815 (aarch64_types_storestruct_lane_qualifiers): New.
2816 (TYPES_STORESTRUCT_LANE): Likewise.
2817 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
2818 (st3_lane): Likewise.
2819 (st4_lane): Likewise.
2820 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
2821 (vec_store_lanesci_lane<mode>): Likewise.
2822 (vec_store_lanesxi_lane<mode>): Likewise.
2823 (aarch64_st2_lane<VQ:mode>): Likewise.
2824 (aarch64_st3_lane<VQ:mode>): Likewise.
2825 (aarch64_st4_lane<VQ:mode>): Likewise.
2826 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
2827 * config/aarch64/arm_neon.h
2828 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
2829 use new macro arguments.
2830 (__ST3_LANE_FUNC): Likewise.
2831 (__ST4_LANE_FUNC): Likewise.
2832 * config/aarch64/iterators.md (V_TWO_ELEM): New.
2833 (V_THREE_ELEM): Likewise.
2834 (V_FOUR_ELEM): Likewise.
2836 2014-04-28 David Malcolm <dmalcolm@redhat.com>
2838 * doc/gimple.texi: Replace the description of the now-defunct
2839 union gimple_statement_d with a diagram showing the
2840 gimple_statement_base class hierarchy and its relationships to
2841 the GSS_ and GIMPLE_ enums.
2843 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
2845 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
2846 * config/aarch64/aarch64.c
2847 (aarch64_cannot_change_mode_class): Weaken conditions.
2848 (aarch64_modes_tieable_p): New.
2849 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
2851 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
2853 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2854 (loadsync_<mode>): Change mode.
2855 (load_quadpti, store_quadpti): New.
2856 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2857 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2859 2014-04-28 Martin Jambor <mjambor@suse.cz>
2861 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
2862 same alias type as the original statement.
2863 (subreplacement_assignment_data): New type.
2864 (handle_unscalarized_data_in_subtree): New type of parameter,
2865 generate new memory accesses with same alias type as the original
2867 (load_assign_lhs_subreplacements): Likewise.
2868 (sra_modify_constructor_assign): Generate new memory accesses with
2869 same alias type as the original statement.
2871 2014-04-28 Richard Biener <rguenther@suse.de>
2873 * tree-pass.h (TODO_verify_il): Define.
2874 (TODO_verify_all): Complete properly.
2875 * passes.c (execute_function_todo): Move existing loop-closed
2876 SSA verification under TODO_verify_il.
2877 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
2878 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
2879 Fix tree sharing issue.
2881 2014-04-28 Richard Biener <rguenther@suse.de>
2884 * builtins.def (DEF_C11_BUILTIN): Add.
2885 (BUILT_IN_ALIGNED_ALLOC): Likewise.
2886 * coretypes.h (enum function_class): Add function_c11_misc.
2887 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
2888 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
2889 (call_may_clobber_ref_p_1): Likewise.
2890 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
2891 (mark_all_reaching_defs_necessary_1): Likewise.
2892 (propagate_necessity): Likewise.
2893 (eliminate_unnecessary_stmts): Likewise.
2894 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
2896 2014-04-28 Richard Biener <rguenther@suse.de>
2898 * tree-vrp.c (vrp_var_may_overflow): Remove.
2899 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
2900 with overflow immediately bump to one before that value and
2901 let iteration figure out overflow status.
2903 2014-04-28 Richard Biener <rguenther@suse.de>
2905 * configure.ac: Do valgrind header checks unconditionally.
2906 Add --enable-valgrind-annotations.
2907 * system.h: Guard valgrind header inclusion with
2908 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
2909 * alloc-pool.c (pool_alloc, pool_free): Use
2910 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
2911 to guard possibly dead code.
2912 * config.in: Regenerated.
2913 * configure: Likewise.
2915 2014-04-28 Jeff Law <law@redhat.com>
2917 PR tree-optimization/60902
2918 * tree-ssa-threadedge.c
2919 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
2920 over real defs when invalidating outputs from statements that do not
2921 produce useful outputs for threading.
2923 2014-04-28 Richard Biener <rguenther@suse.de>
2925 PR tree-optimization/60979
2926 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
2927 SCOPs that end in a block with a successor with abnormal
2930 2014-04-28 Richard Biener <rguenther@suse.de>
2932 * tree-pass.h (execute_pass_list): Adjust prototype.
2933 * passes.c (pass_manager::execute_early_local_passes): Adjust.
2934 (do_per_function): Change callback signature, push all actual
2935 work to the callbals.
2936 (do_per_function_toporder): Likewise.
2937 (execute_function_dump): Adjust.
2938 (execute_function_todo): Likewise.
2939 (clear_last_verified): Likewise.
2940 (verify_curr_properties): Likewise.
2941 (update_properties_after_pass): Likewise.
2942 (execute_pass_list_1): Split out from ...
2943 (execute_pass_list): ... here. Adjust.
2944 (execute_ipa_pass_list): Likewise.
2945 * cgraphunit.c (cgraph_add_new_function): Adjust.
2946 (analyze_function): Likewise.
2947 (expand_function): Likewise.
2948 * cgraph.c (release_function_body): Free dominance info
2949 here instead of asserting it was magically freed elsewhere.
2951 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
2953 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
2954 * configure: Regenerate.
2955 * config/sparc/sparc.opt (muser-mode): New option.
2956 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
2958 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
2959 * doc/invoke.texi (SPARC options): Document -muser-mode.
2961 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
2963 * cselib.c (find_slot_memmode): Delete.
2964 (cselib_hasher): Change compare_type to a struct.
2965 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
2967 (preserve_constants_and_equivs): Adjust for new compare_type.
2968 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
2969 (wrap_constant): Delete.
2970 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
2972 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
2974 * doc/install.texi (Building with profile feedback): Remove
2977 2014-04-26 Tom de Vries <tom@codesourcery.com>
2979 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
2982 2014-04-25 Cary Coutant <ccoutant@google.com>
2985 * dwarf2out.c (should_move_die_to_comdat): A type definition
2986 can contain a subprogram definition, but don't move it to a
2988 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
2989 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
2991 (clone_tree_hash): Rename to...
2992 (clone_tree_partial): ...this; change callers. Copy
2993 DW_TAG_subprogram DIEs as declarations.
2994 (copy_decls_walk): Don't copy children of a declaration into a
2997 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
3000 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
3003 2014-04-25 Jiong Wang <jiong.wang@arm.com>
3005 * config/arm/predicates.md (call_insn_operand): Add long_call check.
3006 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
3008 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
3011 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3013 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
3015 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3017 PR tree-optimization/60930
3018 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
3019 creating a multiply candidate by folding two constant
3020 multiplicands when the result overflows.
3022 2014-04-25 Jakub Jelinek <jakub@redhat.com>
3024 PR tree-optimization/60960
3025 * tree-vect-generic.c (expand_vector_operation): Only call
3026 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
3028 2014-04-25 Tom de Vries <tom@codesourcery.com>
3030 * expr.c (clobber_reg_mode): New function.
3031 * expr.h (clobber_reg): New function.
3033 2014-04-25 Tom de Vries <tom@codesourcery.com>
3035 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
3038 2014-04-25 Radovan Obradovic <robradovic@mips.com>
3039 Tom de Vries <tom@codesourcery.com>
3041 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
3043 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
3044 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
3045 new argument to find_all_hard_reg_sets call.
3047 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3049 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
3050 Use HOST_WIDE_INT_C for mask literal.
3051 (aarch_rev16_shleft_mask_imm_p): Likewise.
3053 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
3056 * config/sparc/sparc.md (ashlsi3_extend): Delete.
3058 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
3060 PR preprocessor/56540
3061 * config/i386/i386-c.c (ix86_target_macros): Define
3062 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
3064 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3066 * configure.ac (tga_func): Remove.
3067 (LIB_TLS_SPEC): Remove.
3068 * configure: Regenerate.
3069 * config.in: Regenerate.
3070 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
3072 2014-04-25 Richard Biener <rguenther@suse.de>
3075 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
3076 call stmt use/clobber sets during stmt walk instead of
3077 walking the possibly incomplete set of caller edges.
3079 2014-04-25 Richard Biener <rguenther@suse.de>
3082 * passes.c (apply_ipa_transforms): Inline into only caller ...
3083 (execute_one_pass): ... here. Properly bring in function
3084 bodies for nodes we want to apply IPA transforms to.
3086 2014-04-24 Cong Hou <congh@google.com>
3088 PR tree-optimization/60896
3089 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
3090 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
3091 (vect_mark_pattern_stmts): Set the def type of all statements in
3092 PATTERN_DEF_SEQ as vect_internal_def.
3094 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
3096 * doc/extend.texi (PowerPC Built-in Functions): Document new
3097 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
3098 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
3100 * config/rs6000/predicates.md (const_0_to_3_operand): New
3101 predicate to match 0..3 integer constants.
3103 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
3104 to support adding miscellaneous builtin functions.
3105 (BU_DFP_MISC_2): Likewise.
3106 (BU_P7_MISC_1): Likewise.
3107 (BU_P7_MISC_2): Likewise.
3108 (BU_P8V_MISC_3): Likewise.
3109 (BU_MISC_1): Likewise.
3110 (BU_MISC_2): Likewise.
3111 (DIVWE): Add extended divide builtin functions.
3114 (DIVWEUO): Likewise.
3118 (DIVDEUO): Likewise.
3119 (DXEX): Add decimal floating-point builtin functions.
3122 (DDEDPDQ): Likewise.
3124 (DENBCDQ): Likewise.
3131 (CDTBCD): Add new BCD builtin functions.
3135 (BCDADD_LT): Likewise.
3136 (BCDADD_EQ): Likewise.
3137 (BCDADD_GT): Likewise.
3138 (BCDADD_OV): Likewise.
3140 (BCDSUB_LT): Likewise.
3141 (BCDSUB_EQ): Likewise.
3142 (BCDSUB_GT): Likewise.
3143 (BCDSUB_OV): Likewise.
3144 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
3145 (UNPACK_TD): Likewise.
3146 (PACK_TF): Likewise.
3147 (UNPACK_TF): Likewise.
3148 (UNPACK_TF_0): Likewise.
3149 (UNPACK_TF_1): Likewise.
3150 (PACK_V1TI): Likewise.
3151 (UNPACK_V1TI): Likewise.
3153 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
3154 support for decimal floating point builtin functions.
3155 (rs6000_expand_ternop_builtin): Add checks for the new builtin
3156 functions that take constant arguments.
3157 (rs6000_invalid_builtin): Add decimal floating point builtin support.
3158 (rs6000_init_builtins): Setup long double, _Decimal64, and
3159 _Decimal128 types for new builtin functions.
3160 (builtin_function_type): Set the unsigned flags appropriately for
3161 the new builtin functions.
3162 (rs6000_opt_masks): Add support for decimal floating point builtin
3165 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
3166 floating point builtin functions.
3167 (RS6000_BTM_COMMON): Likewise.
3168 (RS6000_BTI_long_double): Likewise.
3169 (RS6000_BTI_dfloat64): Likewise.
3170 (RS6000_BTI_dfloat128): Likewise.
3171 (long_double_type_internal_node): Likewise.
3172 (dfloat64_type_internal_node): Likewise.
3173 (dfloat128_type_internal_node): Likewise.
3175 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
3176 2.07 bcd arithmetic instructions.
3177 (UNSPEC_BCDSUB): Likewise.
3178 (UNSPEC_BCD_OVERFLOW): Likewise.
3179 (UNSPEC_BCD_ADD_SUB): Likewise.
3180 (bcd_add_sub): Likewise.
3181 (BCD_TEST): Likewise.
3182 (bcd<bcd_add_sub>): Likewise.
3183 (bcd<bcd_add_sub>_test): Likewise.
3184 (bcd<bcd_add_sub>_test2): Likewise.
3185 (bcd<bcd_add_sub>_<code>): Likewise.
3186 (peephole2 for combined bcd ops): Likewise.
3188 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
3189 decimal floating point builtin functions.
3190 (UNSPEC_DENBCD): Likewise.
3191 (UNSPEC_DXEX): Likewise.
3192 (UNSPEC_DIEX): Likewise.
3193 (UNSPEC_DSCLI): Likewise.
3194 (UNSPEC_DSCRI): Likewise.
3195 (D64_D128): Likewise.
3196 (dfp_suffix): Likewise.
3197 (dfp_ddedpd_<mode>): Likewise.
3198 (dfp_denbcd_<mode>): Likewise.
3199 (dfp_dxex_<mode>): Likewise.
3200 (dfp_diex_<mode>): Likewise.
3201 (dfp_dscli_<mode>): Likewise.
3202 (dfp_dscri_<mode>): Likewise.
3204 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
3206 (UNSPEC_CDTBCD): Likewise.
3207 (UNSPEC_CBCDTD): Likewise.
3208 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
3209 (UNSPEC_DIVEO): Likewise.
3210 (UNSPEC_DIVEU): Likewise.
3211 (UNSPEC_DIVEUO): Likewise.
3212 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
3213 pack/unpack 128-bit types.
3214 (UNSPEC_PACK_128BIT): Likewise.
3215 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
3216 (udiv<mode>3): Use idiv_ldiv mode attribute.
3217 (div<mode>3): Likewise.
3218 (addg6s): Add new BCD builtin functions.
3221 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
3222 (div_extend): Likewise.
3223 (div<div_extend>_<mode>"): Likewise.
3224 (FP128_64): Add support for new builtin functions to pack/unpack
3226 (unpack<mode>): Likewise.
3227 (unpacktf_0): Likewise.
3228 (unpacktf_1): Likewise.
3229 (unpack<mode>_dm): Likewise.
3230 (unpack<mode>_nodm): Likewise.
3231 (pack<mode>): Likewise.
3232 (unpackv1ti): Likewise.
3233 (packv1ti): Likewise.
3235 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
3237 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
3240 2014-04-24 Jakub Jelinek <jakub@redhat.com>
3242 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
3243 * gimplify.c (omp_is_private): Change last argument's type to int.
3244 Only diagnose lastprivate if the simd argument is 1, only diagnose
3245 linear if the simd argument is 2.
3246 (gimplify_omp_for): Adjust omp_is_private callers. When adding
3247 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
3248 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
3249 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
3250 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
3251 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
3252 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
3253 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
3254 * tree-nested.c (convert_nonlocal_omp_clauses,
3255 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
3257 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
3260 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
3263 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
3265 * flag-types.h (enum ivar_visibility): Add.
3267 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
3269 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
3270 function * argument.
3272 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
3274 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
3276 2014-04-24 Radovan Obradovic <robradovic@mips.com>
3277 Tom de Vries <tom@codesourcery.com>
3279 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
3280 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
3282 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
3283 * emit-rtl.c (try_split): Same.
3285 2014-04-24 Radovan Obradovic <robradovic@mips.com>
3286 Tom de Vries <tom@codesourcery.com>
3288 * common.opt (fuse-caller-save): New option.
3290 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
3292 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
3293 elements for big-endian.
3295 2014-04-24 Richard Biener <rguenther@suse.de>
3297 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
3298 during TER and instead use the sepops interface for expanding
3299 non-GIMPLE_SINGLE_RHS.
3301 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3303 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
3304 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
3306 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3308 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
3309 assembler 64-bit option.
3310 * configure: Regenerate.
3312 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3314 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
3315 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
3316 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
3317 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
3318 (TARGET_CRYPTO): Take TARGET_SIMD into account.
3320 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3322 * config/aarch64/aarch64-builtins.c
3323 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
3324 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
3325 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
3326 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
3328 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
3329 (Vrevsuff): New mode attribute.
3331 2014-04-24 Terry Guo <terry.guo@arm.com>
3333 * config/arm/arm.h (machine_function): Define variable
3334 after_arm_reorg here.
3335 * config/arm/arm.c (after_arm_reorg): Remove the definition.
3336 (arm_split_constant): Update the way to access variable
3339 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
3341 2014-04-23 Tom de Vries <tom@codesourcery.com>
3343 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
3345 2014-04-23 David Malcolm <dmalcolm@redhat.com>
3347 * is-a.h: Update comments to reflect the following changes to the
3348 "pointerness" of the API, making the template parameter match the
3349 return type, allowing use of is-a.h with typedefs of pointers.
3350 (is_a_helper::cast): Return a T rather then a pointer to a T, so
3351 that the return type matches the parameter to the is_a_helper.
3353 (dyn_cast): Likewise.
3355 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
3356 pointer from the is-a.h API.
3358 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
3359 (is_a_helper <cgraph_node *>::test): ...this, matching change to
3361 (is_a_helper <varpool_node>::test): Likewise, convert to...
3362 (is_a_helper <varpool_node *>::test): ...this.
3364 (varpool_first_variable): Update for removal of implicit pointer
3365 from the is-a.h API.
3366 (varpool_next_variable): Likewise.
3367 (varpool_first_static_initializer): Likewise.
3368 (varpool_next_static_initializer): Likewise.
3369 (varpool_first_defined_variable): Likewise.
3370 (varpool_next_defined_variable): Likewise.
3371 (cgraph_first_defined_function): Likewise.
3372 (cgraph_next_defined_function): Likewise.
3373 (cgraph_first_function): Likewise.
3374 (cgraph_next_function): Likewise.
3375 (cgraph_first_function_with_gimple_body): Likewise.
3376 (cgraph_next_function_with_gimple_body): Likewise.
3377 (cgraph_alias_target): Likewise.
3378 (varpool_alias_target): Likewise.
3379 (cgraph_function_or_thunk_node): Likewise.
3380 (varpool_variable_node): Likewise.
3381 (symtab_real_symbol_p): Likewise.
3382 * cgraphunit.c (referred_to_p): Likewise.
3383 (analyze_functions): Likewise.
3384 (handle_alias_pairs): Likewise.
3385 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
3386 * gimple-ssa.h (gimple_vuse_op): Likewise.
3387 (gimple_vdef_op): Likewise.
3388 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
3389 * gimple.c (gimple_build_asm_1): Likewise.
3390 (gimple_build_try): Likewise.
3391 (gimple_build_resx): Likewise.
3392 (gimple_build_eh_dispatch): Likewise.
3393 (gimple_build_omp_for): Likewise.
3394 (gimple_omp_for_set_clauses): Likewise.
3396 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
3397 (is_a_helper <gimple_statement_asm *>::test): ...this.
3398 (is_a_helper <gimple_statement_bind>::test): Convert to...
3399 (is_a_helper <gimple_statement_bind *>::test): ...this.
3400 (is_a_helper <gimple_statement_call>::test): Convert to...
3401 (is_a_helper <gimple_statement_call *>::test): ...this.
3402 (is_a_helper <gimple_statement_catch>::test): Convert to...
3403 (is_a_helper <gimple_statement_catch *>::test): ...this.
3404 (is_a_helper <gimple_statement_resx>::test): Convert to...
3405 (is_a_helper <gimple_statement_resx *>::test): ...this.
3406 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
3407 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
3408 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
3409 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
3410 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
3411 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
3412 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
3413 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
3414 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
3415 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
3416 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
3417 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
3418 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
3419 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
3420 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
3421 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
3422 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
3423 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
3424 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
3425 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
3426 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
3427 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
3428 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
3429 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
3430 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
3431 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
3432 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
3433 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
3434 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
3435 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
3436 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
3437 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
3438 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
3439 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
3440 (is_a_helper <gimple_statement_phi>::test): Convert to...
3441 (is_a_helper <gimple_statement_phi *>::test): ...this.
3442 (is_a_helper <gimple_statement_transaction>::test): Convert to...
3443 (is_a_helper <gimple_statement_transaction *>::test): ...this.
3444 (is_a_helper <gimple_statement_try>::test): Convert to...
3445 (is_a_helper <gimple_statement_try *>::test): ...this.
3446 (is_a_helper <gimple_statement_wce>::test): Convert to...
3447 (is_a_helper <gimple_statement_wce *>::test): ...this.
3448 (is_a_helper <const gimple_statement_asm>::test): Convert to...
3449 (is_a_helper <const gimple_statement_asm *>::test): ...this.
3450 (is_a_helper <const gimple_statement_bind>::test): Convert to...
3451 (is_a_helper <const gimple_statement_bind *>::test): ...this.
3452 (is_a_helper <const gimple_statement_call>::test): Convert to...
3453 (is_a_helper <const gimple_statement_call *>::test): ...this.
3454 (is_a_helper <const gimple_statement_catch>::test): Convert to...
3455 (is_a_helper <const gimple_statement_catch *>::test): ...this.
3456 (is_a_helper <const gimple_statement_resx>::test): Convert to...
3457 (is_a_helper <const gimple_statement_resx *>::test): ...this.
3458 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
3459 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
3460 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
3461 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
3462 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
3464 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
3466 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
3468 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
3470 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
3471 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
3472 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
3474 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
3475 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
3477 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
3478 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
3479 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
3480 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
3481 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
3482 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
3484 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
3485 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
3486 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
3487 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
3489 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
3490 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
3491 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
3492 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
3493 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
3494 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
3495 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
3496 (is_a_helper <const gimple_statement_phi>::test): Convert to...
3497 (is_a_helper <const gimple_statement_phi *>::test): ...this.
3498 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
3499 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
3500 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
3501 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
3502 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
3503 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
3504 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
3506 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
3508 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
3509 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
3511 (gimple_use_ops): Update for removal of implicit pointer from the
3513 (gimple_set_use_ops): Likewise.
3514 (gimple_vuse): Likewise.
3515 (gimple_vdef): Likewise.
3516 (gimple_vuse_ptr): Likewise.
3517 (gimple_vdef_ptr): Likewise.
3518 (gimple_set_vuse): Likewise.
3519 (gimple_set_vdef): Likewise.
3520 (gimple_omp_return_set_lhs): Likewise.
3521 (gimple_omp_return_lhs): Likewise.
3522 (gimple_omp_return_lhs_ptr): Likewise.
3523 (gimple_call_fntype): Likewise.
3524 (gimple_call_set_fntype): Likewise.
3525 (gimple_call_set_internal_fn): Likewise.
3526 (gimple_call_use_set): Likewise.
3527 (gimple_call_clobber_set): Likewise.
3528 (gimple_bind_vars): Likewise.
3529 (gimple_bind_set_vars): Likewise.
3530 (gimple_bind_body_ptr): Likewise.
3531 (gimple_bind_set_body): Likewise.
3532 (gimple_bind_add_stmt): Likewise.
3533 (gimple_bind_block): Likewise.
3534 (gimple_bind_set_block): Likewise.
3535 (gimple_asm_ninputs): Likewise.
3536 (gimple_asm_noutputs): Likewise.
3537 (gimple_asm_nclobbers): Likewise.
3538 (gimple_asm_nlabels): Likewise.
3539 (gimple_asm_input_op): Likewise.
3540 (gimple_asm_input_op_ptr): Likewise.
3541 (gimple_asm_output_op): Likewise.
3542 (gimple_asm_output_op_ptr): Likewise.
3543 (gimple_asm_set_output_op): Likewise.
3544 (gimple_asm_clobber_op): Likewise.
3545 (gimple_asm_set_clobber_op): Likewise.
3546 (gimple_asm_label_op): Likewise.
3547 (gimple_asm_set_label_op): Likewise.
3548 (gimple_asm_string): Likewise.
3549 (gimple_catch_types): Likewise.
3550 (gimple_catch_types_ptr): Likewise.
3551 (gimple_catch_handler_ptr): Likewise.
3552 (gimple_catch_set_types): Likewise.
3553 (gimple_catch_set_handler): Likewise.
3554 (gimple_eh_filter_types): Likewise.
3555 (gimple_eh_filter_types_ptr): Likewise.
3556 (gimple_eh_filter_failure_ptr): Likewise.
3557 (gimple_eh_filter_set_types): Likewise.
3558 (gimple_eh_filter_set_failure): Likewise.
3559 (gimple_eh_must_not_throw_fndecl): Likewise.
3560 (gimple_eh_must_not_throw_set_fndecl): Likewise.
3561 (gimple_eh_else_n_body_ptr): Likewise.
3562 (gimple_eh_else_e_body_ptr): Likewise.
3563 (gimple_eh_else_set_n_body): Likewise.
3564 (gimple_eh_else_set_e_body): Likewise.
3565 (gimple_try_eval_ptr): Likewise.
3566 (gimple_try_cleanup_ptr): Likewise.
3567 (gimple_try_set_eval): Likewise.
3568 (gimple_try_set_cleanup): Likewise.
3569 (gimple_wce_cleanup_ptr): Likewise.
3570 (gimple_wce_set_cleanup): Likewise.
3571 (gimple_phi_capacity): Likewise.
3572 (gimple_phi_num_args): Likewise.
3573 (gimple_phi_result): Likewise.
3574 (gimple_phi_result_ptr): Likewise.
3575 (gimple_phi_set_result): Likewise.
3576 (gimple_phi_arg): Likewise.
3577 (gimple_phi_set_arg): Likewise.
3578 (gimple_resx_region): Likewise.
3579 (gimple_resx_set_region): Likewise.
3580 (gimple_eh_dispatch_region): Likewise.
3581 (gimple_eh_dispatch_set_region): Likewise.
3582 (gimple_omp_critical_name): Likewise.
3583 (gimple_omp_critical_name_ptr): Likewise.
3584 (gimple_omp_critical_set_name): Likewise.
3585 (gimple_omp_for_clauses): Likewise.
3586 (gimple_omp_for_clauses_ptr): Likewise.
3587 (gimple_omp_for_set_clauses): Likewise.
3588 (gimple_omp_for_collapse): Likewise.
3589 (gimple_omp_for_index): Likewise.
3590 (gimple_omp_for_index_ptr): Likewise.
3591 (gimple_omp_for_set_index): Likewise.
3592 (gimple_omp_for_initial): Likewise.
3593 (gimple_omp_for_initial_ptr): Likewise.
3594 (gimple_omp_for_set_initial): Likewise.
3595 (gimple_omp_for_final): Likewise.
3596 (gimple_omp_for_final_ptr): Likewise.
3597 (gimple_omp_for_set_final): Likewise.
3598 (gimple_omp_for_incr): Likewise.
3599 (gimple_omp_for_incr_ptr): Likewise.
3600 (gimple_omp_for_set_incr): Likewise.
3601 (gimple_omp_for_pre_body_ptr): Likewise.
3602 (gimple_omp_for_set_pre_body): Likewise.
3603 (gimple_omp_parallel_clauses): Likewise.
3604 (gimple_omp_parallel_clauses_ptr): Likewise.
3605 (gimple_omp_parallel_set_clauses): Likewise.
3606 (gimple_omp_parallel_child_fn): Likewise.
3607 (gimple_omp_parallel_child_fn_ptr): Likewise.
3608 (gimple_omp_parallel_set_child_fn): Likewise.
3609 (gimple_omp_parallel_data_arg): Likewise.
3610 (gimple_omp_parallel_data_arg_ptr): Likewise.
3611 (gimple_omp_parallel_set_data_arg): Likewise.
3612 (gimple_omp_task_clauses): Likewise.
3613 (gimple_omp_task_clauses_ptr): Likewise.
3614 (gimple_omp_task_set_clauses): Likewise.
3615 (gimple_omp_task_child_fn): Likewise.
3616 (gimple_omp_task_child_fn_ptr): Likewise.
3617 (gimple_omp_task_set_child_fn): Likewise.
3618 (gimple_omp_task_data_arg): Likewise.
3619 (gimple_omp_task_data_arg_ptr): Likewise.
3620 (gimple_omp_task_set_data_arg): Likewise.
3621 (gimple_omp_taskreg_clauses): Likewise.
3622 (gimple_omp_taskreg_clauses_ptr): Likewise.
3623 (gimple_omp_taskreg_set_clauses): Likewise.
3624 (gimple_omp_taskreg_child_fn): Likewise.
3625 (gimple_omp_taskreg_child_fn_ptr): Likewise.
3626 (gimple_omp_taskreg_set_child_fn): Likewise.
3627 (gimple_omp_taskreg_data_arg): Likewise.
3628 (gimple_omp_taskreg_data_arg_ptr): Likewise.
3629 (gimple_omp_taskreg_set_data_arg): Likewise.
3630 (gimple_omp_task_copy_fn): Likewise.
3631 (gimple_omp_task_copy_fn_ptr): Likewise.
3632 (gimple_omp_task_set_copy_fn): Likewise.
3633 (gimple_omp_task_arg_size): Likewise.
3634 (gimple_omp_task_arg_size_ptr): Likewise.
3635 (gimple_omp_task_set_arg_size): Likewise.
3636 (gimple_omp_task_arg_align): Likewise.
3637 (gimple_omp_task_arg_align_ptr): Likewise.
3638 (gimple_omp_task_set_arg_align): Likewise.
3639 (gimple_omp_single_clauses): Likewise.
3640 (gimple_omp_single_clauses_ptr): Likewise.
3641 (gimple_omp_single_set_clauses): Likewise.
3642 (gimple_omp_target_clauses): Likewise.
3643 (gimple_omp_target_clauses_ptr): Likewise.
3644 (gimple_omp_target_set_clauses): Likewise.
3645 (gimple_omp_target_child_fn): Likewise.
3646 (gimple_omp_target_child_fn_ptr): Likewise.
3647 (gimple_omp_target_set_child_fn): Likewise.
3648 (gimple_omp_target_data_arg): Likewise.
3649 (gimple_omp_target_data_arg_ptr): Likewise.
3650 (gimple_omp_target_set_data_arg): Likewise.
3651 (gimple_omp_teams_clauses): Likewise.
3652 (gimple_omp_teams_clauses_ptr): Likewise.
3653 (gimple_omp_teams_set_clauses): Likewise.
3654 (gimple_omp_sections_clauses): Likewise.
3655 (gimple_omp_sections_clauses_ptr): Likewise.
3656 (gimple_omp_sections_set_clauses): Likewise.
3657 (gimple_omp_sections_control): Likewise.
3658 (gimple_omp_sections_control_ptr): Likewise.
3659 (gimple_omp_sections_set_control): Likewise.
3660 (gimple_omp_for_set_cond): Likewise.
3661 (gimple_omp_for_cond): Likewise.
3662 (gimple_omp_atomic_store_set_val): Likewise.
3663 (gimple_omp_atomic_store_val): Likewise.
3664 (gimple_omp_atomic_store_val_ptr): Likewise.
3665 (gimple_omp_atomic_load_set_lhs): Likewise.
3666 (gimple_omp_atomic_load_lhs): Likewise.
3667 (gimple_omp_atomic_load_lhs_ptr): Likewise.
3668 (gimple_omp_atomic_load_set_rhs): Likewise.
3669 (gimple_omp_atomic_load_rhs): Likewise.
3670 (gimple_omp_atomic_load_rhs_ptr): Likewise.
3671 (gimple_omp_continue_control_def): Likewise.
3672 (gimple_omp_continue_control_def_ptr): Likewise.
3673 (gimple_omp_continue_set_control_def): Likewise.
3674 (gimple_omp_continue_control_use): Likewise.
3675 (gimple_omp_continue_control_use_ptr): Likewise.
3676 (gimple_omp_continue_set_control_use): Likewise.
3677 (gimple_transaction_body_ptr): Likewise.
3678 (gimple_transaction_label): Likewise.
3679 (gimple_transaction_label_ptr): Likewise.
3680 (gimple_transaction_set_body): Likewise.
3681 (gimple_transaction_set_label): Likewise.
3683 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
3684 * ipa-inline-analysis.c (inline_write_summary): Likewise.
3685 * ipa-ref.c (ipa_record_reference): Likewise.
3686 * ipa-reference.c (analyze_function): Likewise.
3687 (ipa_reference_write_optimization_summary): Likewise.
3688 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
3689 (address_taken_from_non_vtable_p): Likewise.
3690 (comdat_can_be_unshared_p_1): Likewise.
3691 * lto-cgraph.c (lto_output_ref): Likewise.
3692 (add_references): Likewise.
3693 (compute_ltrans_boundary): Likewise.
3694 (output_symtab): Likewise.
3695 (input_ref): Likewise.
3696 (input_cgraph_1): Likewise.
3697 (output_cgraph_opt_summary): Likewise.
3698 * lto-streamer-out.c (lto_output): Likewise.
3699 (output_symbol_p): Likewise.
3700 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
3701 (lsei_start_function_in_partition): Likewise.
3702 (lsei_next_variable_in_partition): Likewise.
3703 (lsei_start_variable_in_partition): Likewise.
3704 * symtab.c (insert_to_assembler_name_hash): Likewise.
3705 (unlink_from_assembler_name_hash): Likewise.
3706 (symtab_unregister_node): Likewise.
3707 (symtab_remove_node): Likewise.
3708 (dump_symtab_node): Likewise.
3709 (verify_symtab_base): Likewise.
3710 (verify_symtab_node): Likewise.
3711 (symtab_make_decl_local): Likewise.
3712 (symtab_alias_ultimate_target): Likewise.
3713 (symtab_resolve_alias): Likewise.
3714 (symtab_get_symbol_partitioning_class): Likewise.
3715 * tree-phinodes.c (allocate_phi_node): Likewise.
3716 (reserve_phi_args_for_new_edge): Likewise.
3717 (remove_phi_args): Likewise.
3718 * varpool.c (varpool_node_for_asm): Likewise.
3719 (varpool_remove_unreferenced_decls): Likewise.
3721 2014-04-23 Jeff Law <law@redhat.com>
3723 PR tree-optimization/60902
3724 * tree-ssa-threadedge.c
3725 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
3726 invalidate outputs from statements that do not produce useful
3727 outputs for threading.
3729 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3731 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
3732 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
3733 machine descriptions for Stack Smashing Protector.
3735 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
3737 * aarch64.md (<optab>_rol<mode>3): New pattern.
3738 (<optab>_rolsi3_uxtw): Likewise.
3739 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
3741 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
3743 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
3744 (arm_cortex_a12_tune): Likewise.
3746 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3748 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
3750 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3752 * config/arm/arm.md (arm_rev16si2): New pattern.
3753 (arm_rev16si2_alt): Likewise.
3754 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
3756 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3758 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
3759 (rev16<mode>2_alt): Likewise.
3760 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
3761 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
3762 (aarch_rev16_shleft_mask_imm_p): Likewise.
3763 (aarch_rev16_p_1): Likewise.
3764 (aarch_rev16_p): Likewise.
3765 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
3766 (aarch_rev16_shright_mask_imm_p): Likewise.
3767 (aarch_rev16_shleft_mask_imm_p): Likewise.
3769 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3771 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
3772 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
3774 (cortex_a53_extra_costs): Likewise.
3775 (cortex_a57_extra_costs): Likewise.
3776 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
3777 (cortexa7_extra_costs): Likewise.
3778 (cortexa8_extra_costs): Likewise.
3779 (cortexa12_extra_costs): Likewise.
3780 (cortexa15_extra_costs): Likewise.
3781 (v7m_extra_costs): Likewise.
3782 (arm_new_rtx_costs): Handle BSWAP.
3784 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3786 * config/arm/arm.c (cortexa8_extra_costs): New table.
3787 (arm_cortex_a8_tune): New tuning struct.
3788 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
3790 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3792 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
3794 2014-04-23 Richard Biener <rguenther@suse.de>
3796 * Makefile.in (OBJS): Remove loop-unswitch.o.
3797 * tree-pass.h (make_pass_rtl_unswitch): Remove.
3798 * passes.def (pass_rtl_unswitch): Likewise.
3799 * loop-init.c (gate_rtl_unswitch): Likewise.
3800 (rtl_unswitch): Likewise.
3801 (pass_data_rtl_unswitch): Likewise.
3802 (pass_rtl_unswitch): Likewise.
3803 (make_pass_rtl_unswitch): Likewise.
3804 * rtl.h (reversed_condition): Likewise.
3805 (compare_and_jump_seq): Likewise.
3806 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
3808 * loop-unroll.c (compare_and_jump_seq): Likewise.
3810 2014-04-23 Richard Biener <rguenther@suse.de>
3812 PR tree-optimization/60903
3813 * tree-ssa-loop-im.c (analyze_memory_references): Remove
3814 commented code block.
3815 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
3816 loop flags to newly created BBs and edges.
3818 2014-04-23 Nick Clifton <nickc@redhat.com>
3820 * config/msp430/msp430.c (msp430_handle_option): Move function
3822 (msp430_option_override): Simplify mcu and mcpu option handling.
3823 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
3824 support for -mhwmult command line option.
3825 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
3826 -mhwmult command line option.
3827 (msp430_hwmult_enabled): Delete.
3828 (msp43o_output_labelref): Add support for -mhwmult command line option.
3829 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
3830 (umulsidi3): Likewise.
3831 * config/msp430/msp430.opt (mmcu): Add Report attribute.
3832 (mcpu, mlarge, msmall): Likewise.
3833 (mhwmult): New option.
3834 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
3836 (msp430_is_f5_mcu): Remove prototype.
3837 (msp430_use_f5_series_hwmult): Add prototype.
3838 * config/msp430/msp430-opts.h: New file.
3839 * common/config/msp430: New directory.
3840 * common/config/msp430/msp430-common.c: New file.
3841 * config.gcc (msp430): Remove target_has_targetm_common.
3842 * doc/invoke.texi: Document -mhwmult command line option.
3844 2014-04-23 Nick Clifton <nickc@redhat.com>
3846 * config/i386/cygwin.h (ENDFILE_SPEC): Include
3847 default-manifest.o if it can be found in the search path.
3848 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
3850 2014-04-23 Terry Guo <terry.guo@arm.com>
3852 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
3854 2014-04-23 Richard Biener <rguenther@suse.de>
3857 * tree-inline.c (declare_return_variable): Use mark_addressable.
3859 2014-04-23 Richard Biener <rguenther@suse.de>
3862 * loop-init.c (loop_optimizer_init): Make sure to apply
3863 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
3865 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3868 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
3870 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
3871 if flag_sanitize_undefined_trap_on_error.
3872 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
3873 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
3874 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
3875 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
3876 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
3877 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
3878 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
3879 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
3880 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
3881 * ubsan.c (ubsan_instrument_unreachable): Return
3882 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
3883 (ubsan_expand_null_ifn): Emit __builtin_trap ()
3884 if flag_sanitize_undefined_trap_on_error and
3885 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
3886 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
3887 instrument_bool_enum_load): Emit __builtin_trap () if
3888 flag_sanitize_undefined_trap_on_error and
3889 __builtin_handle_*_abort () if !flag_sanitize_recover.
3890 * doc/invoke.texi (-fsanitize-recover,
3891 -fsanitize-undefined-trap-on-error): Document.
3893 2014-04-22 Christian Bruel <christian.bruel@st.com>
3895 * config/sh/sh.md (mov<mode>): Replace movQIHI.
3896 Force immediates to SImode.
3898 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
3900 * config/nios2/nios2.md (UNSPEC_ROUND): New.
3902 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
3903 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
3904 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
3905 (nios2_fpu_insn): Add entry for round.
3906 (N2FPU_NO_ERRNO_P): Define.
3907 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
3909 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
3911 2014-04-22 Richard Henderson <rth@redhat.com>
3913 * config/aarch64/aarch64 (addti3, subti3): New expanders.
3914 (add<GPI>3_compare0): Remove leading * from name.
3915 (add<GPI>3_carryin): Likewise.
3916 (sub<GPI>3_compare0): Likewise.
3917 (sub<GPI>3_carryin): Likewise.
3918 (<su_optab>mulditi3): New expander.
3919 (multi3): New expander.
3920 (madd<GPI>): Remove leading * from name.
3922 2014-04-22 Martin Jambor <mjambor@suse.cz>
3924 * cgraphclones.c (cgraph_function_versioning): Copy
3925 ipa_transforms_to_apply instead of asserting it is empty.
3927 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
3930 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
3931 on count_exp to get mode.
3933 2014-04-22 Andrew Pinski <apinski@cavium.com>
3935 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
3936 Handle TLS for ILP32.
3937 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
3938 (tlsie_small_<mode>): this and handle PTR.
3939 (tlsie_small_sidi): New pattern.
3940 (tlsle_small): Change to an expand to handle ILP32.
3941 (tlsle_small_<mode>): New pattern.
3942 (tlsdesc_small): Rename to ...
3943 (tlsdesc_small_<mode>): this and handle PTR.
3945 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3947 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
3949 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3951 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
3952 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
3953 (aarch64_types_signed_poly_qualifiers): Likewise.
3954 (aarch64_types_unsigned_signed_qualifiers): Likewise.
3955 (aarch64_types_poly_signed_qualifiers): Likewise.
3956 (TYPES_REINTERP_SS): Type macro added.
3957 (TYPES_REINTERP_SU): Likewise.
3958 (TYPES_REINTERP_SP): Likewise.
3959 (TYPES_REINTERP_US): Likewise.
3960 (TYPES_REINTERP_PS): Likewise.
3961 (aarch64_fold_builtin): New expression folding added.
3962 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
3963 Declarations removed.
3964 (REINTERP_SS): Declarations added.
3965 (REINTERP_US): Likewise.
3966 (REINTERP_PS): Likewise.
3967 (REINTERP_SU): Likewise.
3968 (REINTERP_SP): Likewise.
3969 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
3970 (vreinterpretq_p8_f64): Likewise.
3971 (vreinterpret_p16_f64): Likewise.
3972 (vreinterpretq_p16_f64): Likewise.
3973 (vreinterpret_f32_f64): Likewise.
3974 (vreinterpretq_f32_f64): Likewise.
3975 (vreinterpret_f64_f32): Likewise.
3976 (vreinterpret_f64_p8): Likewise.
3977 (vreinterpret_f64_p16): Likewise.
3978 (vreinterpret_f64_s8): Likewise.
3979 (vreinterpret_f64_s16): Likewise.
3980 (vreinterpret_f64_s32): Likewise.
3981 (vreinterpret_f64_s64): Likewise.
3982 (vreinterpret_f64_u8): Likewise.
3983 (vreinterpret_f64_u16): Likewise.
3984 (vreinterpret_f64_u32): Likewise.
3985 (vreinterpret_f64_u64): Likewise.
3986 (vreinterpretq_f64_f32): Likewise.
3987 (vreinterpretq_f64_p8): Likewise.
3988 (vreinterpretq_f64_p16): Likewise.
3989 (vreinterpretq_f64_s8): Likewise.
3990 (vreinterpretq_f64_s16): Likewise.
3991 (vreinterpretq_f64_s32): Likewise.
3992 (vreinterpretq_f64_s64): Likewise.
3993 (vreinterpretq_f64_u8): Likewise.
3994 (vreinterpretq_f64_u16): Likewise.
3995 (vreinterpretq_f64_u32): Likewise.
3996 (vreinterpretq_f64_u64): Likewise.
3997 (vreinterpret_s64_f64): Likewise.
3998 (vreinterpretq_s64_f64): Likewise.
3999 (vreinterpret_u64_f64): Likewise.
4000 (vreinterpretq_u64_f64): Likewise.
4001 (vreinterpret_s8_f64): Likewise.
4002 (vreinterpretq_s8_f64): Likewise.
4003 (vreinterpret_s16_f64): Likewise.
4004 (vreinterpretq_s16_f64): Likewise.
4005 (vreinterpret_s32_f64): Likewise.
4006 (vreinterpretq_s32_f64): Likewise.
4007 (vreinterpret_u8_f64): Likewise.
4008 (vreinterpretq_u8_f64): Likewise.
4009 (vreinterpret_u16_f64): Likewise.
4010 (vreinterpretq_u16_f64): Likewise.
4011 (vreinterpret_u32_f64): Likewise.
4012 (vreinterpretq_u32_f64): Likewise.
4014 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4016 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
4017 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
4018 (vreinterpret_p8_s8): Likewise.
4019 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
4020 (vreinterpret_p8_s16): Likewise.
4021 (vreinterpret_p8_s32): Likewise.
4022 (vreinterpret_p8_s64): Likewise.
4023 (vreinterpret_p8_f32): Likewise.
4024 (vreinterpret_p8_u8): Likewise.
4025 (vreinterpret_p8_u16): Likewise.
4026 (vreinterpret_p8_u32): Likewise.
4027 (vreinterpret_p8_u64): Likewise.
4028 (vreinterpret_p8_p16): Likewise.
4029 (vreinterpretq_p8_s8): Likewise.
4030 (vreinterpretq_p8_s16): Likewise.
4031 (vreinterpretq_p8_s32): Likewise.
4032 (vreinterpretq_p8_s64): Likewise.
4033 (vreinterpretq_p8_f32): Likewise.
4034 (vreinterpretq_p8_u8): Likewise.
4035 (vreinterpretq_p8_u16): Likewise.
4036 (vreinterpretq_p8_u32): Likewise.
4037 (vreinterpretq_p8_u64): Likewise.
4038 (vreinterpretq_p8_p16): Likewise.
4039 (vreinterpret_p16_s8): Likewise.
4040 (vreinterpret_p16_s16): Likewise.
4041 (vreinterpret_p16_s32): Likewise.
4042 (vreinterpret_p16_s64): Likewise.
4043 (vreinterpret_p16_f32): Likewise.
4044 (vreinterpret_p16_u8): Likewise.
4045 (vreinterpret_p16_u16): Likewise.
4046 (vreinterpret_p16_u32): Likewise.
4047 (vreinterpret_p16_u64): Likewise.
4048 (vreinterpret_p16_p8): Likewise.
4049 (vreinterpretq_p16_s8): Likewise.
4050 (vreinterpretq_p16_s16): Likewise.
4051 (vreinterpretq_p16_s32): Likewise.
4052 (vreinterpretq_p16_s64): Likewise.
4053 (vreinterpretq_p16_f32): Likewise.
4054 (vreinterpretq_p16_u8): Likewise.
4055 (vreinterpretq_p16_u16): Likewise.
4056 (vreinterpretq_p16_u32): Likewise.
4057 (vreinterpretq_p16_u64): Likewise.
4058 (vreinterpretq_p16_p8): Likewise.
4059 (vreinterpret_f32_s8): Likewise.
4060 (vreinterpret_f32_s16): Likewise.
4061 (vreinterpret_f32_s32): Likewise.
4062 (vreinterpret_f32_s64): Likewise.
4063 (vreinterpret_f32_u8): Likewise.
4064 (vreinterpret_f32_u16): Likewise.
4065 (vreinterpret_f32_u32): Likewise.
4066 (vreinterpret_f32_u64): Likewise.
4067 (vreinterpret_f32_p8): Likewise.
4068 (vreinterpret_f32_p16): Likewise.
4069 (vreinterpretq_f32_s8): Likewise.
4070 (vreinterpretq_f32_s16): Likewise.
4071 (vreinterpretq_f32_s32): Likewise.
4072 (vreinterpretq_f32_s64): Likewise.
4073 (vreinterpretq_f32_u8): Likewise.
4074 (vreinterpretq_f32_u16): Likewise.
4075 (vreinterpretq_f32_u32): Likewise.
4076 (vreinterpretq_f32_u64): Likewise.
4077 (vreinterpretq_f32_p8): Likewise.
4078 (vreinterpretq_f32_p16): Likewise.
4079 (vreinterpret_s64_s8): Likewise.
4080 (vreinterpret_s64_s16): Likewise.
4081 (vreinterpret_s64_s32): Likewise.
4082 (vreinterpret_s64_f32): Likewise.
4083 (vreinterpret_s64_u8): Likewise.
4084 (vreinterpret_s64_u16): Likewise.
4085 (vreinterpret_s64_u32): Likewise.
4086 (vreinterpret_s64_u64): Likewise.
4087 (vreinterpret_s64_p8): Likewise.
4088 (vreinterpret_s64_p16): Likewise.
4089 (vreinterpretq_s64_s8): Likewise.
4090 (vreinterpretq_s64_s16): Likewise.
4091 (vreinterpretq_s64_s32): Likewise.
4092 (vreinterpretq_s64_f32): Likewise.
4093 (vreinterpretq_s64_u8): Likewise.
4094 (vreinterpretq_s64_u16): Likewise.
4095 (vreinterpretq_s64_u32): Likewise.
4096 (vreinterpretq_s64_u64): Likewise.
4097 (vreinterpretq_s64_p8): Likewise.
4098 (vreinterpretq_s64_p16): Likewise.
4099 (vreinterpret_u64_s8): Likewise.
4100 (vreinterpret_u64_s16): Likewise.
4101 (vreinterpret_u64_s32): Likewise.
4102 (vreinterpret_u64_s64): Likewise.
4103 (vreinterpret_u64_f32): Likewise.
4104 (vreinterpret_u64_u8): Likewise.
4105 (vreinterpret_u64_u16): Likewise.
4106 (vreinterpret_u64_u32): Likewise.
4107 (vreinterpret_u64_p8): Likewise.
4108 (vreinterpret_u64_p16): Likewise.
4109 (vreinterpretq_u64_s8): Likewise.
4110 (vreinterpretq_u64_s16): Likewise.
4111 (vreinterpretq_u64_s32): Likewise.
4112 (vreinterpretq_u64_s64): Likewise.
4113 (vreinterpretq_u64_f32): Likewise.
4114 (vreinterpretq_u64_u8): Likewise.
4115 (vreinterpretq_u64_u16): Likewise.
4116 (vreinterpretq_u64_u32): Likewise.
4117 (vreinterpretq_u64_p8): Likewise.
4118 (vreinterpretq_u64_p16): Likewise.
4119 (vreinterpret_s8_s16): Likewise.
4120 (vreinterpret_s8_s32): Likewise.
4121 (vreinterpret_s8_s64): Likewise.
4122 (vreinterpret_s8_f32): Likewise.
4123 (vreinterpret_s8_u8): Likewise.
4124 (vreinterpret_s8_u16): Likewise.
4125 (vreinterpret_s8_u32): Likewise.
4126 (vreinterpret_s8_u64): Likewise.
4127 (vreinterpret_s8_p8): Likewise.
4128 (vreinterpret_s8_p16): Likewise.
4129 (vreinterpretq_s8_s16): Likewise.
4130 (vreinterpretq_s8_s32): Likewise.
4131 (vreinterpretq_s8_s64): Likewise.
4132 (vreinterpretq_s8_f32): Likewise.
4133 (vreinterpretq_s8_u8): Likewise.
4134 (vreinterpretq_s8_u16): Likewise.
4135 (vreinterpretq_s8_u32): Likewise.
4136 (vreinterpretq_s8_u64): Likewise.
4137 (vreinterpretq_s8_p8): Likewise.
4138 (vreinterpretq_s8_p16): Likewise.
4139 (vreinterpret_s16_s8): Likewise.
4140 (vreinterpret_s16_s32): Likewise.
4141 (vreinterpret_s16_s64): Likewise.
4142 (vreinterpret_s16_f32): Likewise.
4143 (vreinterpret_s16_u8): Likewise.
4144 (vreinterpret_s16_u16): Likewise.
4145 (vreinterpret_s16_u32): Likewise.
4146 (vreinterpret_s16_u64): Likewise.
4147 (vreinterpret_s16_p8): Likewise.
4148 (vreinterpret_s16_p16): Likewise.
4149 (vreinterpretq_s16_s8): Likewise.
4150 (vreinterpretq_s16_s32): Likewise.
4151 (vreinterpretq_s16_s64): Likewise.
4152 (vreinterpretq_s16_f32): Likewise.
4153 (vreinterpretq_s16_u8): Likewise.
4154 (vreinterpretq_s16_u16): Likewise.
4155 (vreinterpretq_s16_u32): Likewise.
4156 (vreinterpretq_s16_u64): Likewise.
4157 (vreinterpretq_s16_p8): Likewise.
4158 (vreinterpretq_s16_p16): Likewise.
4159 (vreinterpret_s32_s8): Likewise.
4160 (vreinterpret_s32_s16): Likewise.
4161 (vreinterpret_s32_s64): Likewise.
4162 (vreinterpret_s32_f32): Likewise.
4163 (vreinterpret_s32_u8): Likewise.
4164 (vreinterpret_s32_u16): Likewise.
4165 (vreinterpret_s32_u32): Likewise.
4166 (vreinterpret_s32_u64): Likewise.
4167 (vreinterpret_s32_p8): Likewise.
4168 (vreinterpret_s32_p16): Likewise.
4169 (vreinterpretq_s32_s8): Likewise.
4170 (vreinterpretq_s32_s16): Likewise.
4171 (vreinterpretq_s32_s64): Likewise.
4172 (vreinterpretq_s32_f32): Likewise.
4173 (vreinterpretq_s32_u8): Likewise.
4174 (vreinterpretq_s32_u16): Likewise.
4175 (vreinterpretq_s32_u32): Likewise.
4176 (vreinterpretq_s32_u64): Likewise.
4177 (vreinterpretq_s32_p8): Likewise.
4178 (vreinterpretq_s32_p16): Likewise.
4179 (vreinterpret_u8_s8): Likewise.
4180 (vreinterpret_u8_s16): Likewise.
4181 (vreinterpret_u8_s32): Likewise.
4182 (vreinterpret_u8_s64): Likewise.
4183 (vreinterpret_u8_f32): Likewise.
4184 (vreinterpret_u8_u16): Likewise.
4185 (vreinterpret_u8_u32): Likewise.
4186 (vreinterpret_u8_u64): Likewise.
4187 (vreinterpret_u8_p8): Likewise.
4188 (vreinterpret_u8_p16): Likewise.
4189 (vreinterpretq_u8_s8): Likewise.
4190 (vreinterpretq_u8_s16): Likewise.
4191 (vreinterpretq_u8_s32): Likewise.
4192 (vreinterpretq_u8_s64): Likewise.
4193 (vreinterpretq_u8_f32): Likewise.
4194 (vreinterpretq_u8_u16): Likewise.
4195 (vreinterpretq_u8_u32): Likewise.
4196 (vreinterpretq_u8_u64): Likewise.
4197 (vreinterpretq_u8_p8): Likewise.
4198 (vreinterpretq_u8_p16): Likewise.
4199 (vreinterpret_u16_s8): Likewise.
4200 (vreinterpret_u16_s16): Likewise.
4201 (vreinterpret_u16_s32): Likewise.
4202 (vreinterpret_u16_s64): Likewise.
4203 (vreinterpret_u16_f32): Likewise.
4204 (vreinterpret_u16_u8): Likewise.
4205 (vreinterpret_u16_u32): Likewise.
4206 (vreinterpret_u16_u64): Likewise.
4207 (vreinterpret_u16_p8): Likewise.
4208 (vreinterpret_u16_p16): Likewise.
4209 (vreinterpretq_u16_s8): Likewise.
4210 (vreinterpretq_u16_s16): Likewise.
4211 (vreinterpretq_u16_s32): Likewise.
4212 (vreinterpretq_u16_s64): Likewise.
4213 (vreinterpretq_u16_f32): Likewise.
4214 (vreinterpretq_u16_u8): Likewise.
4215 (vreinterpretq_u16_u32): Likewise.
4216 (vreinterpretq_u16_u64): Likewise.
4217 (vreinterpretq_u16_p8): Likewise.
4218 (vreinterpretq_u16_p16): Likewise.
4219 (vreinterpret_u32_s8): Likewise.
4220 (vreinterpret_u32_s16): Likewise.
4221 (vreinterpret_u32_s32): Likewise.
4222 (vreinterpret_u32_s64): Likewise.
4223 (vreinterpret_u32_f32): Likewise.
4224 (vreinterpret_u32_u8): Likewise.
4225 (vreinterpret_u32_u16): Likewise.
4226 (vreinterpret_u32_u64): Likewise.
4227 (vreinterpret_u32_p8): Likewise.
4228 (vreinterpret_u32_p16): Likewise.
4229 (vreinterpretq_u32_s8): Likewise.
4230 (vreinterpretq_u32_s16): Likewise.
4231 (vreinterpretq_u32_s32): Likewise.
4232 (vreinterpretq_u32_s64): Likewise.
4233 (vreinterpretq_u32_f32): Likewise.
4234 (vreinterpretq_u32_u8): Likewise.
4235 (vreinterpretq_u32_u16): Likewise.
4236 (vreinterpretq_u32_u64): Likewise.
4237 (vreinterpretq_u32_p8): Likewise.
4238 (vreinterpretq_u32_p16): Likewise.
4240 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4242 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
4244 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
4246 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
4247 (vqnegd_s64): Likewise.
4248 (vqabs_s64): Likewise.
4249 (vqabsd_s64): Likewise.
4251 2014-04-22 Richard Henderson <rth@redhat.com>
4253 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
4254 computation to the top of the loop.
4256 2014-04-22 Renlin <renlin.li@arm.com>
4257 Jiong Wang <jiong.wang@arm.com>
4259 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
4260 * config/aarch64/aarch64.c (aarch64_layout_frame)
4261 (aarch64_initial_elimination_offset): Likewise.
4263 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
4265 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
4268 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
4270 * machmode.h (bitwise_mode_for_mode): Declare.
4271 * stor-layout.h (bitwise_type_for_mode): Likewise.
4272 * stor-layout.c (bitwise_mode_for_mode): New function.
4273 (bitwise_type_for_mode): Likewise.
4274 * builtins.c (fold_builtin_memory_op): Use it instead of
4275 int_mode_for_mode and build_nonstandard_integer_type.
4277 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4279 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
4280 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
4281 (*-*-solaris2*): Simplify.
4282 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
4283 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
4284 *-*-solaris2.9* handling.
4286 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
4288 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
4289 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
4291 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
4292 * configure: Regenerate.
4294 * config/i386/sol2-9.h: Remove.
4296 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
4297 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
4298 Remove Solaris 9 references.
4300 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
4302 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
4303 (floatuns<GPI:mode><GPF:mode>2): Remove.
4304 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
4305 and floatuns conversions.
4306 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
4307 and floatuns conversions.
4308 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
4309 (w1,w2): New mode attributes for inequal width conversions.
4311 2014-04-22 Renlin Li <Renlin.Li@arm.com>
4313 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
4314 the output asm format.
4316 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
4318 * config/aarch64/aarch64-simd.md
4319 (aarch64_cm<optab>di): Always split.
4320 (*aarch64_cm<optab>di): New.
4321 (aarch64_cmtstdi): Always split.
4322 (*aarch64_cmtstdi): New.
4324 2014-04-22 Jakub Jelinek <jakub@redhat.com>
4326 PR tree-optimization/60823
4327 * omp-low.c (ipa_simd_modify_function_body): Go through
4328 all SSA_NAMEs and for those refering to vector arguments
4329 which are going to be replaced adjust SSA_NAME_VAR and,
4330 if it is a default definition, change it into a non-default
4331 definition assigned at the beginning of function from new_decl.
4332 (ipa_simd_modify_stmt_ops): Rewritten.
4333 * tree-dfa.c (set_ssa_default_def): When removing default def,
4334 check for NULL loc instead of NULL *loc.
4336 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4338 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
4339 restrictions on core registers for DImode values in Thumb2.
4341 2014-04-22 Ian Bolton <ian.bolton@arm.com>
4343 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
4344 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
4346 2014-04-22 Ian Bolton <ian.bolton@arm.com>
4348 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
4349 (*iordi_notzesidi_di): Likewise.
4350 (*iordi_notsesidi_di): Likewise.
4352 2014-04-22 Ian Bolton <ian.bolton@arm.com>
4354 * config/arm/arm-protos.h (tune_params): New struct members.
4355 * config/arm/arm.c: Initialise tune_params per processor.
4356 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
4357 for speed, based on new tune_params.
4359 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4361 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
4362 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
4363 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
4364 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
4365 * config/aarch64/arm_neon.h (vrnd_f64): Added.
4366 (vrnda_f64): Likewise.
4367 (vrndi_f64): Likewise.
4368 (vrndm_f64): Likewise.
4369 (vrndn_f64): Likewise.
4370 (vrndp_f64): Likewise.
4371 (vrndx_f64): Likewise.
4373 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4375 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
4376 GET_MODE_SIZE argument is enum machine_mode.
4378 2014-04-22 Jakub Jelinek <jakub@redhat.com>
4381 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
4382 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
4384 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
4387 * asan.c (asan_emit_stack_protection): Force the base to align to
4388 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
4389 appropriate bits if STRICT_ALIGNMENT.
4390 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
4392 (expand_used_vars): Leave a space in the stack frame for alignment
4393 if STRICT_ALIGNMENT.
4395 2014-04-21 David Malcolm <dmalcolm@redhat.com>
4397 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
4399 (gimple_store_p): Likewise.
4400 (gimple_assign_load_p): Likewise.
4401 (gimple_assign_cast_p): Likewise.
4402 (gimple_clobber_p): Likewise.
4404 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
4405 rather than a gimple.
4406 (gimple_assign_cast_p): Likewise.
4408 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
4411 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
4412 If mode is DDmode and TARGET_E500_DOUBLE allow move.
4414 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
4415 more debug information for E500 if -mdebug=reg.
4417 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
4420 * config/i386/i386.c (ix86_expand_builtin)
4421 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
4422 register for target RTX.
4423 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
4425 2014-04-18 Cong Hou <congh@google.com>
4427 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
4428 the widen-mult pattern by handling two operands with different sizes,
4429 and operands whose size is smaller than half of the result type.
4431 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
4433 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
4434 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
4435 (do_estimate_edge_time): Compute it.
4436 * ipa-inline.c (want_inline_small_function_p): Bypass
4437 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
4439 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
4441 * ipa-inline.c (spec_rem): New static variable.
4442 (dump_overall_stats): New function.
4443 (dump_inline_stats): New function.
4445 2014-04-18 Richard Henderson <rth@redhat.com>
4447 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
4448 to GET_MODE_SIZE, not a reg_class_t.
4450 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4452 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
4453 (vsx_xxmrglw_<mode>): Likewise.
4455 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
4458 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
4459 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
4460 (rs6000_init_hard_regno_mode_ok): Likewise.
4462 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
4464 * ipa-inline.c (inline_small_functions): Account only non-cold
4466 * doc/invoke.texi (inline-unit-growth): Update documentation.
4468 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
4470 * config/rs6000/rs6000.md (addti3, subti3): New.
4472 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
4475 * config/i386/i386.c (ix86_expand_clear): Remove outdated
4476 comment. Check optimize_insn_for_size_p instead of
4477 optimize_insn_for_speed_p.
4479 2014-04-17 Martin Jambor <mjambor@suse.cz>
4481 * gimple-iterator.c (gsi_start_edge): New function.
4482 * gimple-iterator.h (gsi_start_edge): Declare.
4483 * tree-sra.c (single_non_eh_succ): New function.
4484 (disqualify_ops_if_throwing_stmt): Renamed to
4485 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
4486 having one non-EH successor BB.
4487 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
4488 generate loads into replacements.
4489 (sra_modify_assign): Likewise and and also use the simple path for
4491 (sra_modify_function_body): Commit statements on edges.
4493 2014-04-17 Richard Biener <rguenther@suse.de>
4496 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
4497 comparison results and add clarifying comment.
4499 2014-04-17 Jakub Jelinek <jakub@redhat.com>
4501 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
4502 (blank_mode): Initialize it.
4503 (emit_mode_size_inline, emit_mode_nunits_inline,
4504 emit_mode_inner_inline): New functions.
4505 (emit_insn_modes_h): Call them and surround their output with
4506 #if GCC_VERSION >= 4001 ... #endif.
4507 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
4508 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
4509 mode_* arrays if the argument is __builtin_constant_p.
4510 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
4511 is enum machine_mode.
4513 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4515 * passes.c (opt_pass::execute): Adjust.
4516 (pass_manager::execute_pass_mode_switching): Likewise.
4517 (early_local_passes::execute): Likewise.
4518 (execute_one_pass): Pass cfun to the pass's execute method.
4519 * tree-pass.h (opt_pass::execute): Add function * argument.
4520 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
4521 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
4522 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
4523 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4524 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
4525 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
4526 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
4527 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
4528 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
4529 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
4530 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
4531 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
4532 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
4533 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
4534 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
4535 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
4536 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
4537 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
4538 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
4539 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4540 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4541 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4542 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4543 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4544 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4545 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4546 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
4547 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
4548 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
4551 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4553 * passes.c (opt_pass::gate): Take function * argument.
4554 (gate_all_early_local_passes): Merge into
4555 (early_local_passes::gate): this.
4556 (gate_all_early_optimizations): Merge into
4557 (all_early_optimizations::gate): this.
4558 (gate_all_optimizations): Mege into
4559 (all_optimizations::gate): this.
4560 (gate_all_optimizations_g): Merge into
4561 (all_optimizations_g::gate): this.
4562 (gate_rest_of_compilation): Mege into
4563 (rest_of_compilation::gate): this.
4564 (gate_postreload): Merge into
4565 (postreload::gate): this.
4566 (dump_one_pass): Pass cfun to the pass's gate method.
4567 (execute_ipa_summary_passes): Likewise.
4568 (execute_one_pass): Likewise.
4569 (ipa_write_summaries_2): Likewise.
4570 (ipa_write_optimization_summaries_1): Likewise.
4571 (ipa_read_summaries_1): Likewise.
4572 (ipa_read_optimization_summaries_1): Likewise.
4573 (execute_ipa_stmt_fixups): Likewise.
4574 * tree-pass.h (opt_pass::gate): Add function * argument.
4575 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
4576 combine-stack-adj.c, combine.c, compare-elim.c,
4577 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4578 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
4579 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
4580 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
4581 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
4582 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
4583 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
4584 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
4585 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
4586 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
4587 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
4588 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
4589 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
4590 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
4591 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4592 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4593 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4594 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4595 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4596 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4597 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4598 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4599 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
4600 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
4601 var-tracking.c, vtable-verify.c, web.c: Adjust.
4603 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4605 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
4606 * configure: Regenerate.
4608 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4610 * passes.c (dump_one_pass): don't check pass->has_gate.
4611 (execute_ipa_summary_passes): Likewise.
4612 (execute_one_pass): Likewise.
4613 (ipa_write_summaries_2): Likewise.
4614 (ipa_write_optimization_summaries_1): Likewise.
4615 (ipa_read_optimization_summaries_1): Likewise.
4616 (execute_ipa_stmt_fixups): Likewise.
4617 * tree-pass.h (pass_data::has_gate): Remove.
4618 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
4619 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
4620 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
4621 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4622 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
4623 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
4624 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
4625 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
4626 gimple-low.c, gimple-ssa-isolate-paths.c,
4627 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
4628 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
4629 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
4630 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
4631 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
4632 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
4633 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
4634 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
4635 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
4636 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
4637 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
4638 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4639 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4640 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4641 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4642 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4643 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4644 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4645 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4646 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
4647 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
4648 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
4651 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4653 * pass_manager.h (pass_manager::register_dump_files_1): Remove
4655 * passes.c (pass_manager::register_dump_files_1): Merge into
4656 (pass_manager::register_dump_files): this, and remove its handling of
4657 properties since the pass always has the properties anyway.
4658 (pass_manager::pass_manager): Adjust.
4660 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4662 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
4663 * passes.c (pass_manager::register_dump_files_1): Remove dead code
4664 dealing with properties.
4665 (pass_manager::register_dump_files): Adjust.
4667 2014-03-20 Mark Wielaard <mjw@redhat.com>
4669 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
4670 then represent the bound as normal constant value.
4672 2014-04-17 Jakub Jelinek <jakub@redhat.com>
4675 Forward port from 4.8 branch
4676 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
4678 * config/i386/bmiintrin.h (_blsi_u32): New.
4682 (_blsmsk_u32): Ditto.
4683 (_blsmsk_u64): Ditto.
4684 (_tzcnt_u32): Ditto.
4685 (_tzcnt_u64): Ditto.
4687 2014-04-17 Kito Cheng <kito@0xlab.org>
4689 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
4691 2014-04-17 Richard Biener <rguenther@suse.de>
4694 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
4695 boolean results for comparisons.
4697 2014-04-17 Richard Biener <rguenther@suse.de>
4699 PR tree-optimization/60836
4700 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
4701 initial PHI args to be gimple values.
4703 2014-04-17 Richard Biener <rguenther@suse.de>
4705 PR tree-optimization/60841
4706 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
4707 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
4708 of stmts to SLP build.
4709 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
4710 (vect_analyze_slp): Likewise.
4711 (vect_analyze_slp_instance): Likewise.
4712 (vect_build_slp_tree): Limit overall SLP tree growth.
4713 * tree-vectorizer.h (vect_analyze_data_refs,
4714 vect_analyze_slp): Adjust prototypes.
4716 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4718 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
4721 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4723 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
4724 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
4725 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
4726 for TARGET_SLOW_PSHUFB
4728 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4730 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
4731 * config/i386/i386.c (intel_cost): Ditto.
4733 2014-04-17 Joey Ye <joey.ye@arm.com>
4735 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
4737 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4739 * opts.c (common_handle_option): Disable -fipa-reference coorectly
4742 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4744 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
4745 (type_all_derivations_known_p): New predicate.
4746 (type_all_ctors_visible_p): New predicate.
4747 (type_possibly_instantiated_p): New predicate.
4748 (get_odr_type): Compute all_derivations_known.
4749 (dump_odr_type): Dump the flag.
4750 (maybe_record_type): Cleanup.
4751 (record_target_from_binfo): Add bases_to_consider array;
4752 record bases for types w/o instances and skip CXX destructor.
4753 (possible_polymorphic_call_targets_1): Add bases_to_consider
4754 and consider_construction parameters; check if type may have instance.
4755 (get_polymorphic_call_info): Set maybe_in_construction to true
4756 when we know nothing.
4757 (record_targets_from_bases): Skip CXX destructors; they are
4758 never called for types in construction.
4759 (possible_polymorphic_call_targets): Do not record target when
4760 type may not have instance.
4762 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4765 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
4766 external aliases alive, too.
4768 2014-04-16 Andrew Pinski <apinski@cavium.com>
4770 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
4773 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
4775 * final.c (compute_alignments): Do not apply loop alignment to a block
4776 falling through to the exit.
4778 2014-04-16 Catherine Moore <clm@codesourcery.com>
4780 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
4781 Adjust constraints for microMIPS store patterns.
4783 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4785 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
4787 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
4789 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
4790 (append_use): Run at -O0.
4791 (append_vdef): Likewise.
4792 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
4793 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
4795 2014-04-16 Jakub Jelinek <jakub@redhat.com>
4797 PR tree-optimization/60844
4798 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
4799 (propagate_op_to_single_use, remove_visited_stmt_chain,
4800 linearize_expr, repropagate_negates, reassociate_bb): Use it
4801 instead of gsi_remove.
4803 2014-04-16 Martin Jambor <mjambor@suse.cz>
4805 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
4806 ipa_transforms_to_apply.
4807 (cgraph_function_versioning): Assert that old_node has empty
4808 ipa_transforms_to_apply.
4809 * trans-mem.c (ipa_tm_create_version): Likewise.
4810 * tree-inline.c (tree_function_versioning): Do not duplicate
4811 ipa_transforms_to_apply.
4813 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4816 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
4818 Pass necessary as flags on 64-bit Solaris/x86.
4819 Use lowercase relocs for x86_64-*-*.
4820 * configure: Regenerate.
4822 2014-04-15 Jan Hubicka <jh@suse.cz>
4824 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
4825 (maybe_record_node, likely_target_p): Use it.
4827 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4830 Revert following patch
4832 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4835 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4836 software floating point or no floating point registers, do not
4837 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4838 in GPRs that occurs after we tested for GPRs that would never be
4841 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4842 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4843 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4844 specifically allow DDmode, since that does not use the SPE SIMD
4847 2014-03-21 Mark Wielaard <mjw@redhat.com>
4849 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
4850 as unsigned or int depending on type and value used.
4852 2014-04-15 Richard Biener <rguenther@suse.de>
4854 PR rtl-optimization/56965
4855 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
4856 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
4858 * alias.c (true_dependence_1): Do not call
4859 nonoverlapping_component_refs_p.
4860 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
4861 nonoverlapping_component_refs_p.
4862 (indirect_refs_may_alias_p): Likewise.
4864 2014-04-15 Teresa Johnson <tejohnson@google.com>
4866 * cfg.c (dump_bb_info): Fix flags check.
4867 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
4869 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4871 PR rtl-optimization/60663
4872 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
4875 2014-04-15 Richard Biener <rguenther@suse.de>
4877 * lto-streamer.h (LTO_major_version): Bump to 4.
4879 2014-04-15 Richard Biener <rguenther@suse.de>
4881 * common.opt (lto_partition_model): New enum.
4882 (flto-partition=): Merge separate options with a single with argument,
4883 add -flto-partition=one support.
4884 * flag-types.h (enum lto_partition_model): Declare.
4885 * opts.c (finish_options): Remove duplicate -flto-partition=
4887 * lto-wrapper.c (run_gcc): Adjust.
4889 2014-04-15 Richard Biener <rguenther@suse.de>
4891 * alias.c (ncr_compar): New function.
4892 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
4894 2014-04-15 Richard Biener <rguenther@suse.de>
4896 * alias.c (record_component_aliases): Do not walk BINFOs.
4898 2014-04-15 Richard Biener <rguenther@suse.de>
4900 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
4901 Add struct function argument and adjust.
4902 (find_func_aliases_for_call): Likewise.
4903 (find_func_aliases): Likewise.
4904 (find_func_clobbers): Likewise.
4905 (intra_create_variable_infos): Likewise.
4906 (compute_points_to_sets): Likewise.
4907 (ipa_pta_execute): Adjust. Do not push/pop cfun.
4909 2014-04-15 Richard Biener <rguenther@suse.de>
4911 * tree.c (iterative_hash_expr): Use enum tree_code_class
4912 to store TREE_CODE_CLASS.
4913 (tree_block): Likewise.
4914 (tree_set_block): Likewise.
4915 * tree.h (fold_build_pointer_plus_loc): Use
4916 convert_to_ptrofftype_loc.
4918 2014-04-15 Jakub Jelinek <jakub@redhat.com>
4921 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
4924 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
4926 * cfgloop.h (struct loop): Move force_vectorize down.
4927 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
4928 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
4929 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
4930 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
4931 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
4932 * tree-core.h (enum annot_expr_kind): Add new kind values.
4933 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
4934 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
4936 * tree.def (ANNOTATE_EXPR): Tweak comment.
4938 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4940 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
4943 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
4945 * tree.h (TYPE_IDENTIFIER): Declare.
4946 * tree.c (subrange_type_for_debug_p): Use it.
4947 * godump.c (go_format_type): Likewise.
4948 * dwarf2out.c (is_cxx_auto, modified_type_die,
4949 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
4950 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
4952 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4955 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
4957 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
4959 * config/i386/i386.c (examine_argument): Return bool. Return true if
4960 parameter should be passed in memory.
4961 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
4962 (construct_container): Update calls to examine_argument.
4963 (function_arg_advance_64): Ditto.
4964 (return_in_memory_32): Merge with ix86_return_in_memory.
4965 (return_in_memory_64): Ditto.
4966 (return_in_memory_ms_64): Ditto.
4968 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4970 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
4971 * coverage.c (coverage_compute_profile_id): Handle externally visible
4974 2014-04-14 Martin Jambor <mjambor@suse.cz>
4976 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
4977 DECL_DISREGARD_INLINE_LIMITS functions.
4979 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
4982 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
4984 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
4987 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
4988 optimize_insn_for_speed_p instead of
4989 optimize_function_for_speed_p.
4991 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
4993 * doc/invoke.texi (free): Document AArch64.
4995 2014-04-14 Richard Biener <rguenther@suse.de>
4997 PR tree-optimization/60042
4998 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
4999 (insert_into_preds_of_block): Do not prevent PHI insertion
5000 for REFERENCE exprs here ...
5001 (eliminate_dom_walker::before_dom_children): ... but prevent
5002 their use here under similar conditions when applied to the
5003 IL after PRE optimizations.
5005 2014-04-14 Richard Biener <rguenther@suse.de>
5007 * passes.def: Move early points-to after early SRA.
5009 2014-04-14 Richard Biener <rguenther@suse.de>
5011 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
5012 check for which sign-changes we allow when forwarding
5013 a converted value into a switch.
5015 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
5017 * stor-layout.c (place_field): Finalize non-constant offset for the
5020 2014-04-14 Richard Biener <rguenther@suse.de>
5022 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
5024 (expand_switch_using_bit_tests_p): Likewise.
5025 (process_switch): Compute and pass on speed_p based on the
5027 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
5028 optimize_bb_for_speed_p.
5030 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
5032 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
5033 * function.h (struct function): Rename has_force_vect_loops into
5034 has_force_vectorize_loops.
5035 * lto-streamer-in.c (input_cfg): Adjust for renaming.
5036 (input_struct_function_base): Likewise.
5037 * lto-streamer-out.c (output_cfg): Likewise.
5038 (output_struct_function_base): Likewise.
5039 * omp-low.c (expand_omp_simd): Likewise.
5040 * tree-cfg.c (move_sese_region_to_fn): Likewise.
5041 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
5042 (version_loop_for_if_conversion): Likewise.
5043 (tree_if_conversion): Likewise.
5044 (main_tree_if_conversion): Likewise.
5045 (gate_tree_if_conversion): Likewise.
5046 * tree-inline.c (copy_loops): Likewise.
5047 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
5048 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
5049 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
5050 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
5051 * tree-vectorizer.c (vectorize_loops): Likewise.
5052 * tree-vectorizer.h (unlimited_cost_model): Likewise.
5054 2014-04-14 Richard Biener <rguenther@suse.de>
5057 * lto-streamer-out.c (wrap_refs): New function.
5058 (lto_output): Wrap symbol references in global initializes in
5059 type-preserving MEM_REFs.
5061 2014-04-14 Christian Bruel <christian.bruel@st.com>
5063 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
5065 2014-04-14 Christian Bruel <christian.bruel@st.com>
5067 * config/sh/sh.md (setmemqi): New expand pattern.
5068 * config/sh/sh.h (CLEAR_RATIO): Define.
5069 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
5070 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
5072 2014-04-14 Richard Biener <rguenther@suse.de>
5075 * fold-const.c (negate_expr_p): Don't negate directional rounding
5077 (fold_negate_expr): Likewise.
5079 2014-04-14 Richard Biener <rguenther@suse.de>
5081 PR tree-optimization/59817
5082 PR tree-optimization/60453
5083 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
5084 recursion to catch all CHRECs in the scalar evolution and restrict
5085 the predicate for the remains appropriately.
5087 2014-04-12 Catherine Moore <clm@codesourcery.com>
5089 * config/mips/constraints.md: Add new register constraint "kb".
5090 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
5091 (*movhi_internal): Likewise.
5092 (*movqi_internal): Likewise.
5093 * config/mips/mips.h (M16_STORE_REGS): New register class.
5094 (REG_CLASS_NAMES): Add M16_STORE_REGS.
5095 (REG_CLASS_CONTENTS): Likewise.
5096 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
5098 2014-04-11 Tobias Burnus <burnus@net-b.de>
5101 * doc/invoke.texi (-Wformat-signedness): Document it.
5102 (Wformat=2): Mention that this enables -Wformat-signedness.
5104 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
5106 * common/config/epiphany/epiphany-common.c
5107 (epiphany_option_optimization_table): Enable section anchors by
5108 default at -O1 or higher.
5109 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
5110 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
5111 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
5112 carries no extra cost.
5113 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
5114 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
5115 * config/epiphany/predicates.md (memclob_operand): New predicate.
5116 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
5117 Use memclob_operand predicate and X constraint for operand 3.
5119 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
5121 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
5122 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
5125 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
5127 PR rtl-optimization/60651
5128 * mode-switching.c (optimize_mode_switching): Make sure to emit
5129 sets of a lower numbered entity before sets of a higher numbered
5130 entity to a mode of the same or lower priority.
5131 When creating a seginfo for a basic block that starts with a code
5132 label, move the insertion point past the code label.
5133 (new_seginfo): Document and enforce requirement that
5134 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
5135 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
5136 * doc/tm.texi: Regenerate.
5138 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
5141 * config/arc/arc.c (arc_save_restore): Fix assert typo.
5143 2013-04-11 Jakub Jelinek <jakub@redhat.com>
5145 * BASE-VER: Set to 4.10.0.
5147 2014-04-11 Tobias Burnus <burnus@net-b.de>
5150 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
5151 * doc/gcc.texi (Service): Update description in the @menu
5152 * doc/invoke.texi (Option Summary): Remove misplaced and
5155 2014-04-11 Steve Ellcey <sellcey@mips.com>
5156 Jakub Jelinek <jakub@redhat.com>
5159 * expr.c (convert_move): Use emit_store_flag_force instead of
5160 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
5163 2014-04-11 Richard Biener <rguenther@suse.de>
5166 * varasm.c (assemble_alias): Avoid endless error reporting
5167 recursion by setting TREE_ASM_WRITTEN.
5169 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5171 * config/s390/s390.md: Add a splitter for NOT rtx.
5173 2014-04-11 Jakub Jelinek <jakub@redhat.com>
5175 PR rtl-optimization/60663
5176 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
5178 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
5179 Jakub Jelinek <jakub@redhat.com>
5182 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
5183 flag from decl_node to node.
5185 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5188 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
5189 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
5190 ameliorating the cases where it can be.
5192 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
5195 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
5197 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
5198 (loadsync_<mode>): Change mode.
5199 (load_quadpti, store_quadpti): New.
5200 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
5201 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
5202 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
5204 2014-04-09 Cong Hou <congh@google.com>
5207 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
5210 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5212 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
5213 instead of vnor to exploit possible fusion opportunity in the
5215 (altivec_expand_vec_perm_const_le): Likewise.
5217 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
5219 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
5220 (loadsync_<mode>): Change mode.
5221 (load_quadpti, store_quadpti): New.
5222 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
5223 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
5225 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
5228 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
5229 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
5230 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
5232 2014-04-08 Richard Biener <rguenther@suse.de>
5235 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
5236 a 64bit widest int print double-int similar to on HWI64 hosts.
5238 2014-04-08 Richard Biener <rguenther@suse.de>
5240 PR tree-optimization/60785
5241 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
5242 default defs properly.
5244 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
5246 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
5247 (Weffc++): Likewise.
5249 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
5251 * ipa-devirt.c (maybe_record_node): When node is not recorded,
5252 set completep to false rather than true.
5254 2014-04-07 Douglas B Rupp <rupp@adacore.com>
5257 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
5258 ARM_TARGET2_DWARF_FORMAT.
5260 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
5263 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
5264 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
5267 2014-04-07 Richard Biener <rguenther@suse.de>
5269 PR tree-optimization/60766
5270 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
5271 (may_eliminate_iv): Convert cand_value_at result to desired type.
5273 2014-04-07 Jason Merrill <jason@redhat.com>
5276 * common.opt (-fno-gnu-unique): Add.
5277 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
5279 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5281 * haifa-sched.c: Fix outdated function reference and minor
5282 grammar errors in introductory comment.
5284 2014-04-07 Richard Biener <rguenther@suse.de>
5287 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
5289 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
5291 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
5294 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
5295 size accounting for thunks.
5296 (pa_asm_output_mi_thunk): Use final_start_function() and
5297 final_end_function() to output function start and end directives.
5299 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
5301 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
5302 device specific ISA/ feature information. Remove short_sp and
5303 errata_skip ds. Add avr_device_specific_features enum to have device
5305 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
5306 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
5307 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
5308 updated device specific info.
5309 * config/avr/avr-mcus.def: Merge device specific details to
5310 dev_attribute field.
5311 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
5313 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
5314 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
5315 assembler if RMW isa supported by current device.
5316 * config/avr/genmultilib.awk: Update as device info structure changed.
5317 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
5319 2014-04-04 Cong Hou <congh@google.com>
5321 PR tree-optimization/60656
5322 * tree-vect-stmts.c (supportable_widening_operation):
5323 Fix a bug that elements in a vector with vect_used_by_reduction
5324 property are incorrectly reordered when the operation on it is not
5325 consistant with the one in reduction operation.
5327 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
5329 PR rtl-optimization/60155
5330 * gcse.c (record_set_data): New function.
5331 (single_set_gcse): New function.
5332 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
5333 (hoist_code): Likewise.
5334 (get_pressure_class_and_nregs): Likewise.
5336 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
5338 * explow.c (probe_stack_range): Emit a final optimization blockage.
5340 2014-04-04 Anthony Green <green@moxielogic.com>
5342 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
5345 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
5348 * lto-cgraph.c (input_overwrite_node): Check that partitioning
5349 flags are set only during streaming.
5350 * ipa.c (process_references, walk_polymorphic_call_targets,
5351 symtab_remove_unreachable_nodes): Drop bodies of always inline
5352 after early inlining.
5353 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
5355 2014-04-04 Jakub Jelinek <jakub@redhat.com>
5356 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5359 * dwarf2out.c (const_ok_for_output_1): Reject expressions
5362 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5365 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
5367 (cortex_a53_fdivd): Likewise.
5369 2014-04-04 Martin Jambor <mjambor@suse.cz>
5372 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
5374 * cgraph.c (clone_of_p): Also return true if thunks match.
5375 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
5376 cgraph_function_or_thunk_node and an obsolete comment.
5377 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
5379 (build_function_decl_skip_args): Likewise.
5380 (set_new_clone_decl_and_node_flags): New function.
5381 (duplicate_thunk_for_node): Likewise.
5382 (redirect_edge_duplicating_thunks): Likewise.
5383 (cgraph_clone_node): New parameter args_to_skip, pass it to
5384 redirect_edge_duplicating_thunks which is called instead of
5385 cgraph_redirect_edge_callee.
5386 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
5387 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
5389 2014-04-04 Jeff Law <law@redhat.com>
5392 * config/arm/predicates.md (const_int_I_operand): New predicate.
5393 (const_int_M_operand): Similarly.
5394 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
5396 (insv_t2, extv_reg, extzv_t2): Likewise.
5397 (load_multiple_with_writeback): Similarly for const_int_I_operand.
5398 (pop_multiple_with_writeback_and_return): Likewise.
5399 (vfp_pop_multiple_with_writeback): Likewise
5401 2014-04-04 Richard Biener <rguenther@suse.de>
5404 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
5405 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
5407 * gimplify.h (gimple_add_tmp_var_fn): Declare.
5408 * gimplify.c (gimple_add_tmp_var_fn): New function.
5409 * gimple-expr.h (create_tmp_reg_fn): Declare.
5410 * gimple-expr.c (create_tmp_reg_fn): New function.
5411 * gimple-low.c (record_vars_into): Don't change cfun.
5412 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
5413 code generation without cfun.
5415 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
5418 * Makefile.in (install-driver): Fix shell scripting.
5420 2014-04-03 Cong Hou <congh@google.com>
5422 PR tree-optimization/60505
5423 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
5424 threshold of number of iterations below which no vectorization
5426 * tree-vect-loop.c (new_loop_vec_info):
5427 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
5428 * tree-vect-loop.c (vect_analyze_loop_operations):
5429 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
5430 * tree-vect-loop.c (vect_transform_loop):
5431 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
5432 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
5433 of iterations of the loop and see if we should build the epilogue.
5435 2014-04-03 Richard Biener <rguenther@suse.de>
5437 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
5438 (streamer_tree_cache_create): Adjust.
5439 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
5440 to allow optional nodes array.
5441 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
5442 (streamer_tree_cache_append): Likewise.
5443 (streamer_tree_cache_create): Create nodes array optionally
5444 as specified by parameter.
5445 * lto-streamer-out.c (create_output_block): Avoid maintaining
5446 the node array in the writer cache.
5447 (DFS_write_tree): Remove assertion.
5448 (produce_asm_for_decls): Free the out decl state hash table early.
5449 * lto-streamer-in.c (lto_data_in_create): Adjust for
5450 streamer_tree_cache_create prototype change.
5452 2014-04-03 Richard Biener <rguenther@suse.de>
5454 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
5455 set TREE_CHAIN to NULL_TREE.
5457 2014-04-03 Richard Biener <rguenther@suse.de>
5459 PR tree-optimization/60740
5460 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
5461 over all GIMPLE_COND operands.
5463 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
5465 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
5466 (Weffc++): Remove Scott's numbering, merge lists and reference
5469 2014-04-03 Nick Clifton <nickc@redhat.com>
5471 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
5474 2014-04-03 Martin Jambor <mjambor@suse.cz>
5476 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
5477 mention gcc_unreachable before failing.
5478 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
5481 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
5484 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
5485 inconsistent code and instead mark the context inconsistent.
5486 (possible_polymorphic_call_targets): For inconsistent contexts
5487 return empty complete list.
5489 2014-04-02 Anthony Green <green@moxielogic.com>
5491 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
5492 (extendqisi2, extendhisi2): Define.
5493 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
5494 (WCHAR_TYPE): Change to unsigned int.
5496 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5498 PR tree-optimization/60733
5499 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
5500 insertion point for PHI candidates to be the end of the feeding
5501 block for the PHI argument.
5503 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
5505 PR rtl-optimization/60650
5506 * lra-constraints.c (process_alt_operands): Decrease reject for
5507 earlyclobber matching.
5509 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5511 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
5513 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5515 * config/spu/spu.c (pad_bb): Do not crash when the last
5516 insn is CODE_FOR_blockage.
5518 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5520 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
5521 lies outside the target mode.
5523 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5526 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
5527 software floating point or no floating point registers, do not
5528 allow any type in the FPRs. Eliminate a test for SPE SIMD types
5529 in GPRs that occurs after we tested for GPRs that would never be
5532 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
5533 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
5534 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
5535 specifically allow DDmode, since that does not use the SPE SIMD
5538 2014-04-02 Richard Biener <rguenther@suse.de>
5541 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
5542 MODE_INTs. Properly use negv_optab.
5543 (expand_abs): Likewise.
5545 2014-04-02 Richard Biener <rguenther@suse.de>
5548 * Makefile.in (install-driver): Guard extra installs with special
5551 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
5553 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5556 2014-04-01 Richard Henderson <rth@redhat.com>
5559 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
5560 alternative enabled before register allocation.
5562 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
5564 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
5565 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
5567 (nios2_large_got_address): Remove unneeded 'sym' parameter.
5568 (nios2_got_address): Update nios2_large_got_address call site.
5569 (nios2_delegitimize_address): New function.
5570 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
5571 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
5572 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
5574 2014-04-01 Martin Husemann <martin@duskware.de>
5576 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
5579 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
5581 PR rtl-optimization/60604
5582 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
5583 check from register_operand.
5584 (register_operand): Redefine in terms of general_operand.
5585 (nonmemory_operand): Use register_operand for the non-constant cases.
5587 2014-04-01 Richard Biener <rguenther@suse.de>
5589 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
5591 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
5593 * doc/invoke.texi (mapp-regs): Clarify.
5595 2014-03-31 Ulrich Drepper <drepper@gmail.com>
5597 * config/i386/avx512fintrin.h (__v32hi): Define type.
5598 (__v64qi): Likewise.
5599 (_mm512_set1_epi8): Define.
5600 (_mm512_set1_epi16): Define.
5601 (_mm512_set4_epi32): Define.
5602 (_mm512_set4_epi64): Define.
5603 (_mm512_set4_pd): Define.
5604 (_mm512_set4_ps): Define.
5605 (_mm512_setr4_epi64): Define.
5606 (_mm512_setr4_epi32): Define.
5607 (_mm512_setr4_pd): Define.
5608 (_mm512_setr4_ps): Define.
5609 (_mm512_setzero_epi32): Define.
5611 2014-03-31 Martin Jambor <mjambor@suse.cz>
5614 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
5615 callsite_arguments_match_p. Updated all callers. Also check types of
5616 corresponding formal parameters and actual arguments.
5617 (not_all_callers_have_enough_arguments_p) Renamed to
5618 some_callers_have_mismatched_arguments_p.
5620 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
5622 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
5624 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
5627 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
5630 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
5632 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
5633 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
5635 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
5636 Use FMAMODE_NOVF512 mode iterator.
5637 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
5638 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
5639 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
5640 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
5642 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
5643 Use VF_128_256 mode iterator.
5644 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
5647 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5649 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
5650 static chain if needed.
5652 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
5655 * lra-constraints.c (index_part_to_reg): New.
5656 (process_address): Use it.
5658 2014-03-27 Jeff Law <law@redhat.com>
5659 Jakub Jelinek <jakub@redhat.com>
5662 * expr.c (do_tablejump): Use simplify_gen_binary rather than
5663 gen_rtx_{PLUS,MULT} to build up the address expression.
5665 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
5666 creating non-canonical RTL.
5668 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5671 * ipa-inline.c (want_inline_small_function_p): Short circuit large
5672 functions; reorganize to make cheap checks first.
5673 (inline_small_functions): Do not estimate growth when dumping;
5675 * ipa-inline.h (inline_summary): Add min_size.
5676 (growth_likely_positive): New function.
5677 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
5678 (set_cond_stmt_execution_predicate): Cleanup.
5679 (estimate_edge_size_and_time): Compute min_size.
5680 (estimate_calls_size_and_time): Likewise.
5681 (estimate_node_size_and_time): Likewise.
5682 (inline_update_overall_summary): Update min_size.
5683 (do_estimate_edge_time): Likewise.
5684 (do_estimate_edge_size): Update.
5685 (do_estimate_edge_hints): Update.
5686 (growth_likely_positive): New function.
5688 2014-03-28 Jakub Jelinek <jakub@redhat.com>
5691 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
5692 also if addr has VOIDmode.
5694 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5696 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
5697 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
5699 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
5700 instructions as well as AdvancedSIMD loads.
5702 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5704 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
5705 Use crypto_aese type.
5706 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
5707 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
5708 crypto_aese, crypto_aesmc. Move to types.md.
5709 * config/arm/types.md (crypto_aes): Split into crypto_aese,
5711 * config/arm/iterators.md (crypto_type): Likewise.
5713 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5715 * cgraph.c: Include expr.h and tree-dfa.h.
5716 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
5719 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
5722 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
5723 regs from checking multi-reg pseudos.
5725 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5727 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
5729 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5731 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
5732 if it would clobber the stack pointer, even temporarily.
5734 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
5736 * mode-switching.c: Make small adjustments to the top comment.
5738 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
5740 * config/rs6000/constraints.md (wD constraint): New constraint to
5741 match the constant integer to get the top DImode/DFmode out of a
5742 vector in a VSX register.
5744 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
5745 match the constant integer to get the top DImode/DFmode out of a
5746 vector in a VSX register.
5748 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
5751 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
5754 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
5755 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
5757 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
5758 Optimize vec_extract of 64-bit values, where the value being
5759 extracted is in the top word, where we can use scalar
5760 instructions. Add direct move and store support. Combine the big
5761 endian/little endian vector select load support into a single insn.
5762 (vsx_extract_<mode>_internal1): Likewise.
5763 (vsx_extract_<mode>_internal2): Likewise.
5764 (vsx_extract_<mode>_load): Likewise.
5765 (vsx_extract_<mode>_store): Likewise.
5766 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
5767 combined into vsx_extract_<mode>_load.
5768 (vsx_extract_<mode>_one_le): Likewise.
5770 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
5771 define the top 64-bit vector element.
5773 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
5776 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5777 Document vec_vbpermq builtin.
5780 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
5781 enable use of xxsldwi and xxpermdi builtin functions.
5782 (vec_xxpermdi): Likewise.
5784 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5785 Document use of vec_xxsldwi and vec_xxpermdi builtins.
5787 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
5789 PR rtl-optimization/60650
5790 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
5792 (find_spills_for): New.
5793 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
5794 Spill all pseudos on the second iteration.
5796 2014-03-27 Marek Polacek <polacek@redhat.com>
5799 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
5802 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5804 * config/s390/s390.c (s390_can_use_return_insn): Check for
5805 call-saved FPRs on 31 bit.
5807 2014-03-27 Jakub Jelinek <jakub@redhat.com>
5810 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
5811 if they need regimplification, just drop them instead of
5812 calling gimple_regimplify_operands on them.
5814 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
5817 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
5818 (aarch64_frame_pointer_required): Adjust logic.
5819 (aarch64_can_eliminate): Adjust logic.
5820 (aarch64_override_options_after_change): Adjust logic.
5822 2014-03-27 Dehao Chen <dehao@google.com>
5824 * ipa-inline.c (early_inliner): Update node's inline info.
5826 2014-03-26 Dehao Chen <dehao@google.com>
5828 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
5829 compiler inserted conditional jumps for NAN float check.
5831 2014-03-26 Jakub Jelinek <jakub@redhat.com>
5833 * ubsan.h (ubsan_create_data): Change second argument's type
5834 to const location_t *.
5835 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
5837 (ubsan_create_data): Change second argument to const location_t *PLOC.
5838 Create Loc field whenever PLOC is non-NULL.
5839 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
5840 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
5844 * real.c (real_to_integer2): Change type of low to UHWI.
5846 2014-03-26 Tobias Burnus <burnus@net-b.de>
5848 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
5849 (CILK_SELF_SPECS): New define.
5850 (driver_self_specs): Use it.
5852 2014-03-26 Richard Biener <rguenther@suse.de>
5854 * tree-pretty-print.c (percent_K_format): Implement special
5855 case for LTO and its stripped down BLOCK tree.
5857 2014-03-26 Jakub Jelinek <jakub@redhat.com>
5860 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
5862 * tree-vrp.c (simplify_internal_call_using_ranges): If only
5863 one range is range_int_cst_p, but not both, at least optimize
5864 addition/subtraction of 0 and multiplication by 0 or 1.
5865 * gimple-fold.c (gimple_fold_call): Fold
5866 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
5867 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
5868 INTEGER_CSTs, try to fold at least x * 0 and y - y.
5870 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
5872 PR rtl-optimization/60452
5873 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
5874 <case REG>: Return 1 for invalid offsets from the frame pointer.
5876 2014-03-26 Marek Polacek <polacek@redhat.com>
5879 * doc/extend.texi (C Extensions): Mention variable-length arrays in
5882 2014-03-26 Marek Polacek <polacek@redhat.com>
5885 * doc/extend.texi (Designated Inits): Describe what happens to omitted
5888 2014-03-26 Marek Polacek <polacek@redhat.com>
5891 * ira-color.c (update_conflict_hard_regno_costs): Perform the
5892 multiplication in unsigned type.
5894 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
5896 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
5898 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
5900 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
5902 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
5905 * cif-code.def (UNREACHABLE) New code.
5906 * ipa-inline.c (inline_small_functions): Skip edges to
5907 __builtlin_unreachable.
5908 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
5909 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
5910 predicate to __bulitin_unreachable.
5911 (set_cond_stmt_execution_predicate): Fix issue when
5912 invert_tree_comparison returns ERROR_MARK.
5913 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
5914 propagate to inline clones.
5915 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
5917 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
5918 * cgraphclones.c (cgraph_clone_node): If call destination is already
5919 ureachable, do not redirect it back.
5920 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
5923 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
5925 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
5926 Do not modify inline clones.
5928 2014-03-25 Jakub Jelinek <jakub@redhat.com>
5930 * config/i386/i386.md (general_sext_operand): New mode attr.
5931 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
5932 don't generate (sign_extend (const_int)).
5933 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
5934 operands[2]. Use We constraint instead of <i> and
5935 <general_sext_operand> predicate instead of <general_operand>.
5936 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
5937 * config/i386/constraints.md (We): New constraint.
5938 * config/i386/predicates.md (x86_64_sext_operand,
5939 sext_operand): New predicates.
5941 2014-03-25 Martin Jambor <mjambor@suse.cz>
5944 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
5945 inconsistent devirtualizations to __builtin_unreachable.
5947 2014-03-25 Marek Polacek <polacek@redhat.com>
5950 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
5952 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
5954 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
5955 order of elements for big-endian.
5957 2014-03-25 Richard Biener <rguenther@suse.de>
5960 * gimplify-me.c (gimple_regimplify_operands): Update the
5963 2014-03-25 Martin Jambor <mjambor@suse.cz>
5966 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
5967 (lto_output_varpool_node): Likewise.
5968 (input_overwrite_node): Likewise.
5969 (input_varpool_node): Likewise.
5971 2014-03-25 Richard Biener <rguenther@suse.de>
5973 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
5974 (run_gcc): Likewise.
5976 2014-03-25 Jakub Jelinek <jakub@redhat.com>
5978 * combine.c (simplify_compare_const): Add MODE argument.
5979 Handle mode_width 0 as very large mode_width.
5980 (try_combine, simplify_comparison): Adjust callers.
5982 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
5983 type to avoid signed integer overflow.
5984 * explow.c (plus_constant): Likewise.
5986 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
5988 * doc/generic.texi: Correct typos.
5990 2014-03-24 Tobias Burnus <burnus@net-b.de>
5992 * doc/invoke.texi (-flto): Expand section about
5993 using static libraries with LTO.
5995 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5997 PR rtl-optimization/60501
5998 * optabs.def (addptr3_optab): New optab.
5999 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
6000 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
6001 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
6003 * lra.c (emit_add3_insn): Use the addptr pattern if available.
6005 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
6007 2014-03-24 Ulrich Drepper <drepper@gmail.com>
6009 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
6012 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
6013 (_mm256_undefined_ps): Define.
6014 (_mm256_undefined_pd): Define.
6015 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
6016 (_mm_undefined_pd): Define.
6017 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
6018 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
6019 (_mm512_undefined_ps): Define.
6020 (_mm512_undefined_pd): Define.
6021 Use _mm*_undefined_*.
6022 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
6024 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
6026 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
6027 (lshr_simd): DI mode added.
6028 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
6029 (aarch64_ushr_simddi): Likewise.
6030 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
6031 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
6032 (vshrd_n_u64): Likewise.
6034 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6036 * Makefile.in (s-macro_list): Depend on cc1.
6038 2014-03-23 Teresa Johnson <tejohnson@google.com>
6040 * ipa-utils.c (ipa_print_order): Use specified dump file.
6042 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
6044 PR rtl-optimization/60601
6045 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
6047 * gcc.c (eval_spec_function): Initialize save_growing_value.
6049 2014-03-22 Jakub Jelinek <jakub@redhat.com>
6052 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
6053 code == MINUS_EXPR, never swap op0 with op1.
6055 * toplev.c (init_local_tick): Avoid signed integer multiplication
6057 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
6058 shift by first operand's bitsize.
6060 2014-03-21 Jakub Jelinek <jakub@redhat.com>
6063 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
6065 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
6066 TARGET_ISA_64BIT_P(x).
6068 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6070 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
6071 pattern for vector nor instead of subtract from splat(-1).
6072 (altivec_expand_vec_perm_const_le): Likewise.
6074 2014-03-21 Richard Henderson <rth@twiddle.net>
6077 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
6078 related insns after epilogue_completed.
6080 2014-03-21 Martin Jambor <mjambor@suse.cz>
6083 * cgraph.h (symtab_node): New flag body_removed.
6084 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
6085 when removing bodies.
6086 * symtab.c (dump_symtab_base): Dump body_removed flag.
6087 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
6088 had their bodies removed.
6090 2014-03-21 Martin Jambor <mjambor@suse.cz>
6093 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
6096 2014-03-21 Richard Biener <rguenther@suse.de>
6098 PR tree-optimization/60577
6099 * tree-core.h (struct tree_base): Document nothrow_flag use
6100 in VAR_DECL_NONALIASED.
6101 * tree.h (VAR_DECL_NONALIASED): New.
6102 (may_be_aliased): Adjust.
6103 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
6105 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
6107 * expr.c (expand_expr_real_1): Remove outdated comment.
6109 2014-03-20 Jakub Jelinek <jakub@redhat.com>
6112 * ira.c (adjust_cleared_regs): Call copy_rtx on
6113 *reg_equiv[REGNO (loc)].src_p before passing it to
6114 simplify_replace_fn_rtx.
6117 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
6118 into CONST, put pic register as first operand of PLUS. Use
6119 gen_const_mem for both 32-bit and 64-bit PIC got loads.
6121 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6123 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
6125 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
6127 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
6128 around for store forwarding issue in the FPU on the UT699.
6129 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
6130 loads and operations if -mfix-ut699 is specified.
6131 (divtf3_hq): Tweak attribute.
6132 (sqrttf2_hq): Likewise.
6134 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
6136 * calls.c (store_one_arg): Remove incorrect const qualification on the
6137 type of the temporary.
6138 * cfgexpand.c (expand_return): Likewise.
6139 * expr.c (expand_constructor): Likewise.
6140 (expand_expr_real_1): Likewise.
6142 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6144 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
6147 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
6150 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
6152 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
6154 * config/arm/aarch-common-protos.h
6155 (alu_cost_table): Fix spelling of "extend".
6156 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
6158 2014-03-19 Richard Biener <rguenther@suse.de>
6161 * tree-core.h (tree_type_common): Re-order pointer members
6162 to reduce recursion depth during GC walks.
6164 2014-03-19 Marek Polacek <polacek@redhat.com>
6167 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
6168 before accessing it.
6170 2014-03-19 Richard Biener <rguenther@suse.de>
6173 * lto-streamer-in.c (input_function): In WPA stage do not drop
6176 2014-03-19 Jakub Jelinek <jakub@redhat.com>
6178 PR tree-optimization/60559
6179 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
6180 with build_zero_cst assignment.
6182 2014-03-18 Kai Tietz <ktietz@redhat.com>
6184 PR rtl-optimization/56356
6185 * sdbout.c (sdbout_parms): Verify that parms'
6186 incoming argument is valid.
6187 (sdbout_reg_parms): Likewise.
6189 2014-03-18 Richard Henderson <rth@redhat.com>
6192 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
6193 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
6194 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
6196 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
6198 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
6199 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
6200 Italicize plugin event names in description. Explain that
6201 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
6202 Remind that no GCC functions should be called after PLUGIN_FINISH.
6203 Explain what pragmas with expansion are.
6205 2014-03-18 Martin Liska <mliska@suse.cz>
6207 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
6208 gimple call statement is update.
6209 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
6210 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
6212 2014-03-18 Jakub Jelinek <jakub@redhat.com>
6215 * ubsan.c (ubsan_instrument_unreachable): Call
6216 initialize_sanitizer_builtins.
6217 (ubsan_pass): Likewise.
6220 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
6221 varpool_finalize_decl instead of rest_of_decl_compilation.
6223 2014-03-18 Richard Biener <rguenther@suse.de>
6225 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
6226 by using bitmap_and_compl instead of bitmap_and_compl_into.
6227 (df_rd_transfer_function): Likewise.
6229 2014-03-18 Richard Biener <rguenther@suse.de>
6231 * doc/lto.texi (fresolution): Fix typo.
6233 2014-03-18 Richard Biener <rguenther@suse.de>
6235 * doc/invoke.texi (flto): Update for changes in 4.9.
6237 2014-03-18 Richard Biener <rguenther@suse.de>
6239 * doc/loop.texi: Remove section on the removed lambda framework.
6240 Update loop docs with recent changes in preserving loop structure.
6242 2014-03-18 Richard Biener <rguenther@suse.de>
6244 * doc/lto.texi (-fresolution): Document.
6246 2014-03-18 Richard Biener <rguenther@suse.de>
6248 * doc/contrib.texi: Adjust my name.
6250 2014-03-18 Jakub Jelinek <jakub@redhat.com>
6253 * internal-fn.c: Include diagnostic-core.h.
6254 (expand_BUILTIN_EXPECT): New function.
6255 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
6256 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
6257 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
6258 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
6260 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
6261 Revert 3 argument __builtin_expect code.
6262 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
6263 * gimple-fold.c (gimple_fold_call): Likewise.
6264 * tree.h (fold_builtin_expect): New prototype.
6265 * builtins.c (build_builtin_expect_predicate): Add predictor
6266 argument, if non-NULL, create 3 argument __builtin_expect.
6267 (fold_builtin_expect): No longer static. Add ARG2 argument,
6268 pass it through to build_builtin_expect_predicate.
6269 (fold_builtin_2): Adjust caller.
6270 (fold_builtin_3): Handle BUILT_IN_EXPECT.
6271 * internal-fn.def (BUILTIN_EXPECT): New.
6273 2014-03-18 Tobias Burnus <burnus@net-b.de>
6276 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
6277 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
6278 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
6280 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
6283 * predict.c (combine_predictions_for_bb): Fix up formatting.
6284 (expr_expected_value_1, expr_expected_value): Add predictor argument,
6285 fill what it points to if non-NULL.
6286 (tree_predict_by_opcode): Adjust caller, use the predictor.
6287 * predict.def (PRED_COMPARE_AND_SWAP): Add.
6289 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
6291 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
6292 proper constant for the store mode.
6294 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
6296 * symtab.c (change_decl_assembler_name): Fix transparent alias
6299 2014-03-16 Renlin Li <Renlin.Li@arm.com>
6301 * config/aarch64/aarch64.c: Correct the comments about the
6302 aarch64 stack layout.
6304 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
6306 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
6307 check for GF_OMP_FOR_KIND_FOR.
6309 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
6311 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
6312 ymm and zmm register names.
6314 2014-03-17 Jakub Jelinek <jakub@redhat.com>
6317 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
6318 note creation for the 2010-08-31 changes.
6320 2014-03-17 Marek Polacek <polacek@redhat.com>
6323 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
6324 as -fno-tree-loop-vectorize.
6325 (expand_omp_simd): Likewise.
6327 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
6329 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
6330 (eligible_for_call_delay): New prototype.
6331 * config/sparc/sparc.c (tls_call_delay): Rename into...
6332 (eligible_for_call_delay): ...this. Return false if the instruction
6333 cannot be put in the delay slot of a branch.
6334 (eligible_for_restore_insn): Simplify.
6335 (eligible_for_return_delay): Return false if the instruction cannot be
6336 put in the delay slot of a branch and simplify.
6337 (eligible_for_sibcall_delay): Return false if the instruction cannot be
6338 put in the delay slot of a branch.
6339 * config/sparc/sparc.md (fix_ut699): New attribute.
6340 (tls_call_delay): Delete.
6341 (in_call_delay): Reimplement.
6342 (eligible_for_sibcall_delay): Rename into...
6343 (in_sibcall_delay): ...this.
6344 (eligible_for_return_delay): Rename into...
6345 (in_return_delay): ...this.
6346 (in_branch_delay): Reimplement.
6347 (in_uncond_branch_delay): Delete.
6348 (in_annul_branch_delay): Delete.
6350 2014-03-14 Richard Henderson <rth@redhat.com>
6353 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
6354 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
6355 (*floathi<X87MODEF>2_i387_with_temp): Remove.
6356 (floathi splitters): Remove.
6357 (float<SWI48x>xf2): New pattern.
6358 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
6359 code that tried to handle DImode for 32-bit, but which was excluded
6360 by the pattern's condition. Drop allocation of stack temporary.
6361 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
6362 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
6363 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
6364 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
6365 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
6366 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
6367 (*float<SWI48><MODEF>2_sse_interunit): Remove.
6368 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
6369 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
6370 (*float<SWI48x><X87MODEF>2_i387): Remove.
6371 (all float _with_temp splitters): Remove.
6372 (*float<SWI48x><MODEF>2_i387): New pattern.
6373 (*float<SWI48><MODEF>2_sse): New pattern.
6374 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
6375 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
6377 2014-03-14 Jakub Jelinek <jakub@redhat.com>
6378 Marek Polacek <polacek@redhat.com>
6381 * common.opt (dump_base_name_prefixed): New Variable.
6382 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
6383 if x_dump_base_name_prefixed is already set, set it at the end.
6385 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
6387 PR rtl-optimization/60508
6388 * lra-constraints.c (get_reload_reg): Add new parameter
6390 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
6391 Pass the new parameter values.
6393 2014-03-14 Richard Biener <rguenther@suse.de>
6395 * common.opt: Revert unintented changes from r205065.
6398 2014-03-14 Richard Biener <rguenther@suse.de>
6401 * cfghooks.c (split_block): Properly adjust all loops the
6402 block was a latch of.
6404 2014-03-14 Martin Jambor <mjambor@suse.cz>
6407 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
6410 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
6413 * config/avr/avr.c (avr_set_current_function): Pass function name
6414 through default_strip_name_encoding before sanity checking instead
6415 of skipping the first char of the assembler name.
6417 2014-03-13 Richard Henderson <rth@redhat.com>
6420 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
6421 (ix86_force_to_memory, ix86_free_from_memory): Remove.
6422 * config/i386/i386-protos.h: Likewise.
6423 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
6424 in the expander instead of a splitter.
6425 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
6426 any possibility of requiring a memory.
6427 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
6428 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
6429 (fp branch splitters): Update for ix86_split_fp_branch.
6430 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
6431 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
6432 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
6433 (*fop_<MODEF>_2_i387): Remove f/r alternative.
6434 (*fop_<MODEF>_3_i387): Likewise.
6435 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
6436 (splitters for the fop_* register patterns): Remove.
6437 (fscalexf4_i387): Rename from *fscalexf4_i387.
6438 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
6440 2014-03-13 Jakub Jelinek <jakub@redhat.com>
6442 PR tree-optimization/59779
6443 * tree-dfa.c (get_ref_base_and_extent): Use double_int
6444 type for bitsize and maxsize instead of HOST_WIDE_INT.
6446 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
6448 PR rtl-optimization/57320
6449 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
6450 the CFG after thread_prologue_and_epilogue_insns.
6452 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
6454 PR rtl-optimization/57189
6455 * lra-constraints.c (process_alt_operands): Disfavor spilling
6458 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
6460 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
6462 2014-03-13 Jakub Jelinek <jakub@redhat.com>
6464 PR tree-optimization/59025
6466 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
6467 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
6469 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
6472 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
6473 calls of avr_out_plus_1.
6475 2014-03-13 Bin Cheng <bin.cheng@arm.com>
6477 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
6478 BB's single pred and update the father loop's latch info later.
6480 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
6482 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
6486 (VEC_base): Likewise.
6487 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
6488 registers, we need to swap double words in little endian mode.
6490 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
6491 to be a container mode for 128-bit integer operations added in ISA
6492 2.07. Unlike TImode and PTImode, the preferred register set is
6493 the Altivec/VMX registers for the 128-bit operations.
6495 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
6497 (rs6000_split_128bit_ok_p): Likewise.
6499 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
6500 macros for creating ISA 2.07 normal and overloaded builtin
6501 functions with 3 arguments.
6502 (BU_P8V_OVERLOAD_3): Likewise.
6503 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
6504 for use as overloaded functions.
6505 (VPERM_1TI_UNS): Likewise.
6506 (VSEL_1TI): Likewise.
6507 (VSEL_1TI_UNS): Likewise.
6508 (ST_INTERNAL_1ti): Likewise.
6509 (LD_INTERNAL_1ti): Likewise.
6510 (XXSEL_1TI): Likewise.
6511 (XXSEL_1TI_UNS): Likewise.
6512 (VPERM_1TI): Likewise.
6513 (VPERM_1TI_UNS): Likewise.
6514 (XXPERMDI_1TI): Likewise.
6515 (SET_1TI): Likewise.
6516 (LXVD2X_V1TI): Likewise.
6517 (STXVD2X_V1TI): Likewise.
6518 (VEC_INIT_V1TI): Likewise.
6519 (VEC_SET_V1TI): Likewise.
6520 (VEC_EXT_V1TI): Likewise.
6521 (EQV_V1TI): Likewise.
6522 (NAND_V1TI): Likewise.
6523 (ORC_V1TI): Likewise.
6524 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
6525 added in ISA 2.07. Add both normal 'altivec' builtins, and the
6527 (VADDUQM): Likewise.
6528 (VSUBCUQ): Likewise.
6529 (VADDEUQM): Likewise.
6530 (VADDECUQ): Likewise.
6531 (VSUBEUQM): Likewise.
6532 (VSUBECUQ): Likewise.
6534 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
6535 __int128_t and __uint128_t types.
6536 (__uint128_type): Likewise.
6537 (altivec_categorize_keyword): Add support for vector __int128_t,
6538 vector __uint128_t, vector __int128, and vector unsigned __int128
6539 as a container type for TImode operations that need to be done in
6540 VSX/Altivec registers.
6541 (rs6000_macro_to_expand): Likewise.
6542 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
6543 to support 128-bit integer instructions vaddcuq, vadduqm,
6544 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
6545 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
6547 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
6548 for V1TImode, and set up preferences to use VSX/Altivec registers.
6549 Setup VSX reload handlers.
6550 (rs6000_debug_reg_global): Likewise.
6551 (rs6000_init_hard_regno_mode_ok): Likewise.
6552 (rs6000_preferred_simd_mode): Likewise.
6553 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
6554 (easy_altivec_constant): Likewise.
6555 (output_vec_const_move): Likewise.
6556 (rs6000_expand_vector_set): Convert V1TImode set and extract to
6558 (rs6000_expand_vector_extract): Likewise.
6559 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
6561 (rs6000_const_vec): Add support for V1TImode.
6562 (rs6000_emit_le_vsx_load): Swap double words when loading or
6563 storing TImode/V1TImode.
6564 (rs6000_emit_le_vsx_store): Likewise.
6565 (rs6000_emit_le_vsx_move): Likewise.
6566 (rs6000_emit_move): Add support for V1TImode.
6567 (altivec_expand_ld_builtin): Likewise.
6568 (altivec_expand_st_builtin): Likewise.
6569 (altivec_expand_vec_init_builtin): Likewise.
6570 (altivec_expand_builtin): Likewise.
6571 (rs6000_init_builtins): Add support for V1TImode type. Add
6572 support for ISA 2.07 128-bit integer builtins. Define type names
6573 for the VSX/Altivec vector types.
6574 (altivec_init_builtins): Add support for overloaded vector
6575 functions with V1TImode type.
6576 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
6577 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
6579 (rs6000_split_128bit_ok_p): Likewise.
6580 (rs6000_handle_altivec_attribute): Create V1TImode from vector
6581 __int128_t and vector __uint128_t.
6583 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
6584 and mode attributes.
6591 (VS_scalar): Likewise.
6592 (VS_double): Likewise.
6593 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
6595 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
6596 we support the ISA 2.07 128-bit integer arithmetic instructions.
6597 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
6598 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
6599 and TImode types for use with the builtin functions.
6600 (V1TI_type_node): Likewise.
6601 (unsigned_V1TI_type_node): Likewise.
6602 (intTI_type_internal_node): Likewise.
6603 (uintTI_type_internal_node): Likewise.
6605 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
6606 128-bit builtin functions.
6607 (UNSPEC_VADDEUQM): Likewise.
6608 (UNSPEC_VADDECUQ): Likewise.
6609 (UNSPEC_VSUBCUQ): Likewise.
6610 (UNSPEC_VSUBEUQM): Likewise.
6611 (UNSPEC_VSUBECUQ): Likewise.
6612 (VM): Add V1TImode to vector mode iterators.
6614 (VI_unit): Likewise.
6615 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
6616 (altivec_vaddcuq): Likewise.
6617 (altivec_vsubuqm): Likewise.
6618 (altivec_vsubcuq): Likewise.
6619 (altivec_vaddeuqm): Likewise.
6620 (altivec_vaddecuq): Likewise.
6621 (altivec_vsubeuqm): Likewise.
6622 (altivec_vsubecuq): Likewise.
6624 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
6626 (BOOL_128): Likewise.
6627 (BOOL_REGS_OUTPUT): Likewise.
6628 (BOOL_REGS_OP1): Likewise.
6629 (BOOL_REGS_OP2): Likewise.
6630 (BOOL_REGS_UNARY): Likewise.
6631 (BOOL_REGS_AND_CR0): Likewise.
6633 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
6634 128-bit integer builtin support.
6635 (vec_vadduqm): Likewise.
6636 (vec_vaddecuq): Likewise.
6637 (vec_vaddeuqm): Likewise.
6638 (vec_vsubecuq): Likewise.
6639 (vec_vsubeuqm): Likewise.
6640 (vec_vsubcuq): Likewise.
6641 (vec_vsubuqm): Likewise.
6643 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6644 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
6645 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
6646 128-bit integer add/subtract to ISA 2.07.
6648 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
6650 * config/arc/arc.c (arc_predicate_delay_insns):
6651 Fix third argument passed to conditionalize_nonjump.
6653 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
6655 * config/aarch64/aarch64-builtins.c
6656 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
6657 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
6658 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
6659 instead of __builtin_lfloor.
6660 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
6662 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6664 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
6665 (tree_ssa_ifcombine_bb_1): New function.
6666 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
6667 is an empty forwarder block to then_bb or vice versa and then_bb
6668 and else_bb are effectively swapped.
6670 2014-03-12 Christian Bruel <christian.bruel@st.com>
6673 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
6674 REG_CFA_DEF_CFA note.
6675 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
6676 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
6678 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
6680 PR tree-optimization/60454
6681 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
6683 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6685 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
6686 Do not define target_cpu_default2 to generic.
6687 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
6688 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
6689 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
6691 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6692 Marc Glisse <marc.glisse@inria.fr>
6694 PR tree-optimization/60502
6695 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
6696 instead of build_low_bits_mask.
6698 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6701 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
6702 if there are multiple uses, but op doesn't live on E edge.
6703 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
6704 clobber stmts before __builtin_unreachable.
6706 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
6708 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
6709 hard_frame_pointer_rtx.
6710 * cse.c (cse_insn): Remove volatile check.
6711 * cselib.c (cselib_process_insn): Likewise.
6712 * dse.c (scan_insn): Likewise.
6714 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
6716 * config/arc/arc.c (conditionalize_nonjump): New function,
6718 (arc_ifcvt): ... this.
6719 (arc_predicate_delay_insns): Use it.
6721 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
6723 * config/arc/predicates.md (extend_operand): During/after reload,
6724 allow const_int_operand.
6725 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
6726 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
6727 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
6729 (umulsi3_highpart_i): Likewise.
6731 2014-03-11 Richard Biener <rguenther@suse.de>
6733 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
6734 Add asserts to guard possible wrong-code bugs.
6736 2014-03-11 Richard Biener <rguenther@suse.de>
6738 PR tree-optimization/60429
6739 PR tree-optimization/60485
6740 * tree-ssa-structalias.c (set_union_with_increment): Properly
6741 take into account all fields that overlap the shifted vars.
6742 (do_sd_constraint): Likewise.
6743 (do_ds_constraint): Likewise.
6744 (get_constraint_for_ptr_offset): Likewise.
6746 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
6748 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
6749 (nios2_compute_frame_layout):
6750 Add calculation of cfun->machine->fp_save_offset.
6751 (nios2_expand_prologue): Correct setting of frame pointer register
6753 (nios2_expand_epilogue): Update recovery of stack pointer from
6754 frame pointer accordingly.
6755 (nios2_initial_elimination_offset): Update calculation of offset
6756 for eliminating to HARD_FRAME_POINTER_REGNUM.
6758 2014-03-10 Jakub Jelinek <jakub@redhat.com>
6761 * ipa.c (symtab_remove_unreachable_nodes): Don't call
6762 cgraph_get_create_node on VAR_DECLs.
6764 2014-03-10 Richard Biener <rguenther@suse.de>
6767 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
6769 2014-03-08 Douglas B Rupp <rupp@gnat.com>
6771 * config/vms/vms.opt (vms_float_format): New variable.
6773 2014-03-08 Tobias Burnus <burnus@net-b.de>
6775 * doc/invoke.texi (-fcilkplus): Update implementation status.
6777 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
6778 Richard Biener <rguenther@suse.de>
6780 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
6781 consistently accross all TUs.
6782 (run_gcc): Enable -fshort-double automatically at link at link-time
6783 and disallow override.
6785 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
6788 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
6789 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
6790 if they can't be used.
6792 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6794 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
6795 for Solaris 11/x86 ld.
6796 * configure: Regenerate.
6798 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6800 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
6801 (LIB_TLS_SPEC): Save as ld_tls_libs.
6802 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
6803 (HAVE_AS_IX86_TLSLDM): New test.
6804 * configure, config.in: Regenerate.
6805 * config/i386/i386.c (legitimize_tls_address): Fall back to
6806 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
6807 cannot support TLS_MODEL_LOCAL_DYNAMIC.
6808 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
6809 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
6811 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
6813 * common.opt (fira-loop-pressure): Mark as optimization.
6815 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
6817 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
6818 an OpenMP mappable type.
6820 2014-03-06 Matthias Klose <doko@ubuntu.com>
6822 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
6823 MULTILIB_OSDIRNAMES is not defined.
6825 2014-03-06 Jakub Jelinek <jakub@redhat.com>
6826 Meador Inge <meadori@codesourcery.com>
6829 * config/arm/arm.c (arm_tls_symbol_p): Remove.
6830 (arm_legitimize_address): Call legitimize_tls_address for any
6831 arm_tls_referenced_p expression, handle constant addend. Call it
6832 before testing for !TARGET_ARM.
6833 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
6835 2014-03-06 Richard Biener <rguenther@suse.de>
6841 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
6843 * tree-streamer.c (record_common_node): Assert we don't record
6844 nodes with type double.
6845 (preload_common_node): Skip type double, complex double and double
6846 pointer since it is now frontend dependent due to fshort-double option.
6848 2014-03-06 Richard Biener <rguenther@suse.de>
6850 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
6851 or -fno-lto is specified and the linker has full plugin support.
6852 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
6853 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
6854 * lto-wrapper.c (merge_and_complain): Merge compile-time
6855 optimization levels.
6856 (run_gcc): And pass it through to the link options.
6858 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
6862 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6864 * cselib.c (remove_useless_values): Skip to avoid quadratic
6865 behavior if the condition moved from...
6866 (cselib_process_insn): ... here holds.
6868 2014-03-05 Jakub Jelinek <jakub@redhat.com>
6871 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
6872 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
6875 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
6876 (TM_H): Add x86-tune.def.
6878 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6880 * config/aarch64/aarch64.c (generic_tunings):
6881 Use cortexa57_extra_costs.
6883 2014-03-05 Jakub Jelinek <jakub@redhat.com>
6886 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
6887 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
6888 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
6891 2014-03-04 Heiher <r@hev.cc>
6893 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
6894 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
6896 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
6898 * config/i386/predicates.md (const2356_operand): Change to ...
6899 (const2367_operand): ... this.
6900 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
6902 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6903 (*avx512pf_scatterpf<mode>sf): Ditto.
6904 (avx512pf_scatterpf<mode>df): Ditto.
6905 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6906 (*avx512pf_scatterpf<mode>df): Ditto.
6907 * config/i386/i386.c (ix86_expand_builtin): Update
6908 incorrect hint operand error message.
6910 2014-03-04 Richard Biener <rguenther@suse.de>
6912 * lto-section-in.c (lto_get_section_data): Fix const cast.
6914 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
6916 * tree-streamer.c (record_common_node): Assert we don't record
6917 nodes with type double.
6918 (preload_common_node): Skip type double, complex double and double
6919 pointer since it is now frontend dependent due to fshort-double option.
6921 2014-03-04 Richard Biener <rguenther@suse.de>
6924 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
6925 (lto_input_toplevel_asms): Likewise.
6926 * lto-section-in.c (lto_get_section_data): Instead do it here
6929 2014-03-04 Richard Biener <rguenther@suse.de>
6931 PR tree-optimization/60382
6932 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
6933 dead PHIs a reduction.
6935 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
6937 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
6939 (_mm_prefetch): Move out of GCC target("sse") pragma.
6940 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
6941 GCC target("prfchw") pragma.
6942 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
6944 * config/i386/i386.c (ix86_option_override_internal): Enable
6945 -mprfchw with -mprefetchwt1.
6947 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
6949 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
6952 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
6954 * opts.h (CL_PCH_IGNORE): Define.
6955 * targhooks.c (option_affects_pch_p):
6956 Return false for options that have CL_PCH_IGNORE set.
6957 * opt-functions.awk: Process PchIgnore.
6958 * doc/options.texi: Document PchIgnore.
6960 * config/arc/arc.opt (misize): Add PchIgnore property.
6962 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6964 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
6965 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
6966 constraint on constants to permit them being loaded into
6967 GENERAL_REGS or BASE_REGS.
6969 2014-03-03 Nick Clifton <nickc@redhat.com>
6971 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
6972 anti-cacnonical alternatives.
6973 (negandhi3_real): New pattern.
6974 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
6976 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6978 * config/avr/avr-mcus.def: Remove atxmega16x1.
6979 * config/avr/avr-tables.opt: Regenerate.
6980 * config/avr/t-multilib: Regenerate.
6981 * doc/avr-mmcu.texi: Regenerate.
6983 2014-03-03 Tobias Grosser <tobias@grosser.es>
6984 Mircea Namolaru <mircea.namolaru@inria.fr>
6986 PR tree-optimization/58028
6987 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
6990 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6992 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
6993 not handled by recognizers.
6995 2014-03-03 Jakub Jelinek <jakub@redhat.com>
6998 * function.c (expand_function_end): Don't emit
6999 clobber_return_register sequence if clobber_after is a BARRIER.
7000 * cfgexpand.c (construct_exit_block): Append instructions before
7001 return_label to prev_bb.
7003 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7005 * config/rs6000/constraints.md: Document reserved use of "wc".
7007 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
7010 * ipa.c (function_and_variable_visibility): When dissolving comdat
7011 group, also set all symbols to local.
7013 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
7018 2013-12-14 Jan Hubicka <jh@suse.cz>
7020 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
7022 2014-03-02 Jon Beniston <jon@beniston.com>
7028 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
7029 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
7030 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
7031 (simple_return, *simple_return): New patterns
7032 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
7033 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
7035 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
7037 * dwarf2out.c (gen_subprogram_die): Tidy.
7039 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
7042 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
7043 (*mov_t_msb_neg_negc): ... this new insn.
7045 2014-02-28 Jason Merrill <jason@redhat.com>
7048 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
7051 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
7054 * dwarf2out.c (decltype_auto_die): New static.
7055 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
7056 (gen_type_die_with_usage): Handle 'decltype(auto)'.
7057 (is_cxx_auto): Likewise.
7059 2014-02-28 Ian Bolton <ian.bolton@arm.com>
7061 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
7062 we are not using general regs only.
7064 2014-02-28 Richard Biener <rguenther@suse.de>
7067 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
7068 previous fix and only allow to remove trivial pre-headers
7069 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
7070 (remove_forwarder_block): Properly update the latch of a loop.
7072 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7075 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
7076 (cselib_preserved_hash_table): New.
7077 (preserve_constants_and_equivs): Move preserved vals to it.
7078 (cselib_find_slot): Look it up first.
7079 (cselib_init): Initialize it.
7080 (cselib_finish): Release it.
7081 (dump_cselib_table): Dump it.
7083 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7086 * cselib.c (remove_useless_values): Skip to avoid quadratic
7087 behavior if the condition moved from...
7088 (cselib_process_insn): ... here holds.
7090 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7093 * var-tracking.c (vt_initialize): Apply the same condition to
7094 preserve the CFA base value.
7096 2014-02-28 Joey Ye <joey.ye@arm.com>
7099 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
7100 if reload in progress or completed.
7102 2014-02-28 Tobias Burnus <burnus@net-b.de>
7105 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
7108 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
7110 * doc/tm.texi.in (Condition Code Status): Update documention for
7111 relative locations of cc0-setter and cc0-user.
7113 2014-02-27 Jeff Law <law@redhat.com>
7115 PR rtl-optimization/52714
7116 * combine.c (try_combine): When splitting an unrecognized PARALLEL
7117 into two independent simple sets, if I3 is a jump, ensure the
7118 pattern we place into I3 is a (set (pc) ...).
7120 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
7121 Jeff Law <law@redhat.com>
7123 PR rtl-optimization/49847
7124 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
7125 are in different blocks.
7126 * doc/tm.texi (Condition Code Status): Update documention for
7127 relative locations of cc0-setter and cc0-user.
7129 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
7132 * lra.c (lra_emit_add): Check SUBREG too.
7134 2014-02-27 Andreas Schwab <schwab@suse.de>
7136 * config/m68k/m68k.c (m68k_option_override): Disable
7137 -flive-range-shrinkage for classic m68k.
7138 (m68k_override_options_after_change): Likewise.
7140 2014-02-27 Marek Polacek <polacek@redhat.com>
7143 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
7144 -Wmaybe-uninitialized.
7146 2014-02-27 Alan Modra <amodra@gmail.com>
7149 * reload1.c (emit_input_reload_insns): When reload_override_in,
7150 set old to rl->in_reg when rl->in_reg is a subreg.
7152 2014-02-26 Richard Biener <rguenther@suse.de>
7155 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
7157 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
7159 * common/config/i386/predicates.md (const1256_operand): Remove.
7160 (const2356_operand): New.
7161 (const_1_to_2_operand): Remove.
7162 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7163 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7164 (*avx512pf_gatherpf<mode>sf): Ditto.
7165 (avx512pf_gatherpf<mode>df): Ditto.
7166 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7167 (*avx512pf_gatherpf<mode>df): Ditto.
7168 (avx512pf_scatterpf<mode>sf): Ditto.
7169 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7170 (*avx512pf_scatterpf<mode>sf): Ditto.
7171 (avx512pf_scatterpf<mode>df): Ditto.
7172 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7173 (*avx512pf_scatterpf<mode>df): Ditto.
7174 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
7176 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
7178 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
7179 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
7180 (_mm512_mask_testn_epi64_mask): Move to ...
7181 * config/i386/avx512cdintrin.h: Here.
7182 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
7183 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
7184 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
7185 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
7186 TARGET_AVX512F from TARGET_AVX512CD.
7188 2014-02-26 Richard Biener <rguenther@suse.de>
7191 * ipa.c (walk_polymorphic_call_targets): Properly guard
7192 call to inline_update_overall_summary.
7194 2014-02-26 Bin Cheng <bin.cheng@arm.com>
7197 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
7198 and latches only if requested. Fix latch if it is removed.
7199 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
7200 LOOPS_HAVE_PREHEADERS.
7202 2014-02-25 Andrew Pinski <apinski@cavium.com>
7204 * builtins.c (expand_builtin_thread_pointer): Create a new target
7205 when the target is NULL.
7207 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
7209 PR rtl-optimization/60317
7210 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
7211 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
7212 * lra-assigns.c: Include params.h.
7213 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
7214 other reload pseudos considerations.
7216 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7218 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
7219 to use canonical form for nor<mode>3.
7221 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7224 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
7227 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
7229 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
7230 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
7231 (ix86_handle_option): Handle OPT_mprefetchwt1.
7232 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
7233 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
7235 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7236 OPTION_MASK_ISA_PREFETCHWT1.
7237 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
7238 (PTA_PREFETCHWT1): New.
7239 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
7240 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
7241 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
7242 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
7243 (*prefetch_avx512pf_<mode>_: Change into ...
7244 (*prefetch_prefetchwt1_<mode>: This.
7245 * config/i386/i386.opt (mprefetchwt1): New.
7246 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
7247 (_mm_prefetch): Handle intent to write.
7248 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
7250 2014-02-25 Richard Biener <rguenther@suse.de>
7253 * emit-rtl.c (mem_attrs_htab): Remove.
7254 (mem_attrs_htab_hash): Likewise.
7255 (mem_attrs_htab_eq): Likewise.
7256 (set_mem_attrs): Always allocate new mem-attrs when something changed.
7257 (init_emit_once): Do not allocate mem_attrs_htab.
7259 2014-02-25 Richard Biener <rguenther@suse.de>
7262 * lto-opts.c (lto_write_options): Output non-explicit conservative
7263 -fwrapv, -fno-trapv and -fno-strict-overflow.
7264 * lto-wrapper.c (merge_and_complain): Handle merging those options.
7265 (run_gcc): And pass them through.
7267 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
7269 * sel-sched.c (calculate_new_fences): New parameter ptime.
7270 Calculate it as a maximum over all fence cycles.
7271 (sel_sched_region_2): Adjust the call to calculate_new_fences.
7272 Print the final schedule timing when sched_verbose.
7274 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
7276 PR rtl-optimization/60292
7277 * sel-sched.c (fill_vec_av_set): Do not reset target availability
7278 bit fot the fence instruction.
7280 2014-02-24 Alangi Derick <alangiderick@gmail.com>
7282 * calls.h: Fix typo in comment.
7284 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
7286 * config/pa/pa.c (pa_output_move_double): Don't valididate when
7287 adjusting offsetable addresses.
7289 2014-02-24 Guozhi Wei <carrot@google.com>
7291 * sparseset.h (sparseset_pop): Fix the wrong index.
7293 2014-02-24 Walter Lee <walt@tilera.com>
7295 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
7296 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
7298 * common/config/tilegx/tilegx-common.c
7299 (TARGET_DEFAULT_TARGET_FLAGS): Define.
7300 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
7302 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
7303 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
7304 (tilegx_gimplify_va_arg_expr): Handle big endian.
7305 (tilegx_expand_unaligned_load): Ditto.
7306 (tilegx_expand_unaligned_store): Ditto.
7307 (TARGET_RETURN_IN_MSB): New.
7308 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
7309 (TARGET_ENDIAN_DEFAULT): New.
7310 (TARGET_BIG_ENDIAN): Handle big endian.
7311 (BYTES_BIG_ENDIAN): Ditto.
7312 (WORDS_BIG_ENDIAN): Ditto.
7313 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
7316 * config/tilegx/tilegx.md (extv): Handle big endian.
7318 (insn_st<n>): Ditto.
7319 (insn_st<n>_add<bitsuffix>): Ditto.
7320 (insn_stnt<n>): Ditto.
7321 (insn_stnt<n>_add<bitsuffix>):Ditto.
7322 (vec_interleave_highv8qi): Handle big endian.
7323 (vec_interleave_highv8qi_be): New.
7324 (vec_interleave_highv8qi_le): New.
7325 (insn_v1int_h): Handle big endian.
7326 (vec_interleave_lowv8qi): Handle big endian.
7327 (vec_interleave_lowv8qi_be): New.
7328 (vec_interleave_lowv8qi_le): New.
7329 (insn_v1int_l): Handle big endian.
7330 (vec_interleave_highv4hi): Handle big endian.
7331 (vec_interleave_highv4hi_be): New.
7332 (vec_interleave_highv4hi_le): New.
7333 (insn_v2int_h): Handle big endian.
7334 (vec_interleave_lowv4hi): Handle big endian.
7335 (vec_interleave_lowv4hi_be): New.
7336 (vec_interleave_lowv4hi_le): New.
7337 (insn_v2int_l): Handle big endian.
7338 (vec_interleave_highv2si): Handle big endian.
7339 (vec_interleave_highv2si_be): New.
7340 (vec_interleave_highv2si_le): New.
7341 (insn_v4int_h): Handle big endian.
7342 (vec_interleave_lowv2si): Handle big endian.
7343 (vec_interleave_lowv2si_be): New.
7344 (vec_interleave_lowv2si_le): New.
7345 (insn_v4int_l): Handle big endian.
7346 * config/tilegx/tilegx.opt (mbig-endian): New option.
7347 (mlittle-endian): New option.
7348 * doc/install.texi: Document tilegxbe-linux.
7349 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
7351 2014-02-24 Martin Jambor <mjambor@suse.cz>
7354 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
7355 there are no parameter descriptors.
7357 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
7359 PR rtl-optimization/60268
7360 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
7361 initialization to ...
7362 (sched_rgn_init): ... here.
7363 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
7365 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
7367 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
7370 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
7372 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
7375 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
7377 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
7378 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
7380 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
7382 * config/microblaze/predicates.md: Add cmp_op predicate.
7383 * config/microblaze/microblaze.md: Add branch_compare instruction
7384 which uses cmp_op predicate and emits cmp insn before branch.
7385 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
7386 to microblaze_expand_conditional_branch and consolidate logic.
7387 (microblaze_expand_conditional_branch): emit branch_compare
7388 insn instead of handling cmp op separate from branch insn.
7390 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7392 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
7395 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7397 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
7398 define_insn with define_expand and new define_insn
7399 *altivec_lve<VI_char>x_internal.
7400 (altivec_stve<VI_char>x): Replace define_insn with define_expand
7401 and new define_insn *altivec_stve<VI_char>x_internal.
7402 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
7404 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
7406 (altivec_expand_stvex_be): New function.
7408 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
7410 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
7411 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
7412 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
7413 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
7415 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
7418 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
7419 instead of emit_move_insn.
7421 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7423 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
7425 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
7426 gen_altivec_vsumsws.
7428 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7430 * config/rs6000/altivec.md (altivec_lvxl): Rename as
7431 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
7432 (altivec_lvxl_<mode>): New define_expand incorporating
7433 -maltivec=be semantics where needed.
7434 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
7435 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
7436 semantics where needed.
7437 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
7438 (altivec_stvx_<mode>): New define_expand incorporating
7439 -maltivec=be semantics where needed.
7440 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
7441 VM2 iterator instead of V4SI.
7442 (altivec_stvxl_<mode>): New define_expand incorporating
7443 -maltivec=be semantics where needed.
7444 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
7445 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
7446 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
7447 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
7448 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
7449 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
7450 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
7451 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
7452 ALTIVEC_BUILTIN_STVXL.
7453 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
7454 (altivec_expand_stvx_be): Likewise.
7455 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
7456 (altivec_expand_lvx_be): Likewise.
7457 (altivec_expand_stvx_be): Likewise.
7458 (altivec_expand_builtin): Add cases for
7459 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
7460 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
7461 (altivec_init_builtins): Add definitions for
7462 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
7463 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
7465 2014-02-21 Catherine Moore <clm@codesourcery.com>
7467 * doc/invoke.texi (mvirt, mno-virt): Document.
7468 * config/mips/mips.opt (mvirt): New option.
7469 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
7471 2014-02-21 Richard Biener <rguenther@suse.de>
7473 PR tree-optimization/60276
7474 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
7475 (STMT_VINFO_MIN_NEG_DIST): New macro.
7476 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
7477 STMT_VINFO_MIN_NEG_DIST.
7478 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
7479 made for negative dependence distances still hold.
7481 2014-02-21 Richard Biener <rguenther@suse.de>
7484 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
7485 DECL_INITIAL for globals not in the current function context.
7487 2014-02-21 Jakub Jelinek <jakub@redhat.com>
7489 PR tree-optimization/56490
7490 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
7491 * tree-ssa-uninit.c: Include params.h.
7492 (compute_control_dep_chain): Add num_calls argument, return false
7493 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
7494 num_calls to recursive call.
7495 (find_predicates): Change dep_chain into normal array,
7496 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
7497 variable and adjust compute_control_dep_chain caller.
7498 (find_def_preds): Likewise.
7500 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
7502 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
7503 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
7505 2014-02-21 Nick Clifton <nickc@redhat.com>
7507 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
7508 (pushhi1): Likewise.
7509 (popqi1): Add mode to pre_dec.
7512 2014-02-21 Jakub Jelinek <jakub@redhat.com>
7514 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
7515 mode for mask of V8SFmode permutation.
7517 2014-02-20 Richard Henderson <rth@redhat.com>
7520 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
7521 a new pseudo for OLDVAL.
7523 2014-02-20 Jakub Jelinek <jakub@redhat.com>
7526 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
7527 gen_reg_rtx if d->testing_p.
7528 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
7529 if d->testing_p and we will certainly return true.
7530 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
7533 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
7535 * emit-rtl.c (gen_reg_rtx): Assert that
7536 crtl->emit.regno_pointer_align_length is non-zero.
7538 2014-02-20 Richard Henderson <rth@redhat.com>
7541 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
7542 on failure the store back into EXPECT.
7544 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
7545 Sandra Loosemore <sandra@codesourcery.com>
7547 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
7548 * config/nios2/nios2.c (nios2_function_profiler): Add
7549 -fPIC (flag_pic == 2) support.
7550 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
7551 (nios2_large_offset_p): New function.
7552 (nios2_unspec_reloc_p): Move up position, update to use
7553 nios2_large_offset_p.
7554 (nios2_unspec_address): Remove function.
7555 (nios2_unspec_offset): New function.
7556 (nios2_large_got_address): New function.
7557 (nios2_got_address): Add large offset support.
7558 (nios2_legitimize_tls_address): Update usage of removed and new
7560 (nios2_symbol_binds_local_p): New function.
7561 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
7562 (nios2_legitimize_address): Update to use nios2_large_offset_p.
7563 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
7564 (nios2_print_operand): Merge H/L processing, add hiadj/lo
7565 processing for (const (unspec ...)).
7566 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
7568 2014-02-20 Richard Biener <rguenther@suse.de>
7570 * tree-cfg.c (replace_uses_by): Mark altered BBs before
7571 doing the substitution.
7572 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
7574 2014-02-20 Martin Jambor <mjambor@suse.cz>
7577 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
7578 info when checking whether lattices are bottom.
7580 2014-02-20 Richard Biener <rguenther@suse.de>
7583 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
7586 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
7589 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
7590 parameter specifying the scaling.
7591 (inline_call): Update.
7592 (want_inline_recursively): Guard division by zero.
7593 (recursive_inlining): Update.
7594 * ipa-inline.h (clone_inlined_nodes): Update.
7596 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
7599 * config/i386/i386.c (classify_argument): Pass structures of size
7600 64 bytes or less in register.
7602 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
7603 Kirill Yukhin <kirill.yukhin@intel.com>
7605 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
7606 (_mm_rcp28_round_ss): Ditto.
7607 (_mm_rsqrt28_round_sd): Ditto.
7608 (_mm_rsqrt28_round_ss): Ditto.
7609 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
7610 (_mm_rcp14_round_ss): Ditto.
7611 (_mm_rsqrt14_round_sd): Ditto.
7612 (_mm_rsqrt14_round_ss): Ditto.
7613 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
7614 the first input operand, get rid of match_dup.
7615 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
7617 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7619 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
7620 operand as the first input operand, set type attribute.
7621 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7623 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
7624 operand as the first input operand, set type attribute.
7626 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7628 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
7631 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
7634 * config/i386/i386.c (construct_container): Remove TFmode check
7635 for X86_64_INTEGER_CLASS.
7637 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
7640 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
7641 only when -Wpsabi is enabled.
7643 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
7646 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
7647 passing arrays in registers are the same as for structs, so remove the
7648 special case for them.
7650 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
7652 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
7653 destination type, extract only the valid bits if the source type is not
7654 integral and has a different mode.
7656 2014-02-19 Richard Biener <rguenther@suse.de>
7659 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
7662 2014-02-19 Richard Biener <rguenther@suse.de>
7665 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
7666 (ipa_modify_call_arguments): Emit an argument load explicitely and
7667 preserve virtual SSA form there and for the replacement call.
7668 Do not update SSA form nor free dominance info.
7670 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7672 * ipa.c (function_and_variable_visibility): Also clear WEAK
7673 flag when disolving COMDAT_GROUP.
7675 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7677 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
7678 * ipa-prop.c (ipa_set_jf_known_type): Return early when
7680 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
7681 do more sanity checks.
7682 (detect_type_change): Return true when giving up early.
7683 (compute_complex_assign_jump_func): Fix type parameter of
7684 ipa_set_ancestor_jf.
7685 (compute_complex_ancestor_jump_func): Likewise.
7686 (update_jump_functions_after_inlining): Fix updating of
7688 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
7690 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7692 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
7693 inline clones when edge disappears.
7695 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
7698 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
7699 Split 64-bit moves into 2 patterns. Do not allow the use of
7700 direct move for TDmode in little endian, since the decimal value
7701 has little endian bytes within a word, but the 64-bit pieces are
7702 ordered in a big endian fashion, and normal subreg's of TDmode are
7704 (mov<mode>_64bit_dm): Likewise.
7705 (movtd_64bit_nodm): Likewise.
7707 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
7709 PR tree-optimization/60174
7710 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
7711 statement of an SSA_NAME that occurs in an abnormal PHI node.
7713 2014-02-18 Jakub Jelinek <jakub@redhat.com>
7716 * final.c (SEEN_BB): Remove.
7717 (SEEN_NOTE, SEEN_EMITTED): Renumber.
7718 (final_scan_insn): Don't force_source_line on second
7719 NOTE_INSN_BASIC_BLOCK.
7721 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
7724 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
7725 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
7726 (type_natural_mode): Warn ABI change when %zmm register is not
7727 available for AVX512F vector value passing.
7729 2014-02-18 Kai Tietz <ktietz@redhat.com>
7732 * config/i386/i386.c (ix86_expand_prologue): Use value in
7733 rax register as displacement when restoring %r10 or %rax.
7734 Fix wrong offset when restoring both registers.
7736 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
7738 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
7739 assertion with conditional return.
7741 2014-02-18 Jakub Jelinek <jakub@redhat.com>
7742 Uros Bizjak <ubizjak@gmail.com>
7745 * config/i386/driver-i386.c (host_detect_local_cpu): If
7746 YMM state is not saved by the OS, also clear has_f16c. Move
7747 CPUID 0x80000001 handling before YMM state saving checking.
7749 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
7751 PR rtl-optimization/58960
7752 * haifa-sched.c (alloc_global_sched_pressure_data): New,
7753 factored out from ...
7754 (sched_init): ... here.
7755 (free_global_sched_pressure_data): New, factored out from ...
7756 (sched_finish): ... here.
7757 * sched-int.h (free_global_sched_pressure_data): Declare.
7758 * sched-rgn.c (nr_regions_initial): New static global.
7759 (haifa_find_rgns): Initialize it.
7760 (schedule_region): Disable sched-pressure for the newly
7763 2014-02-17 Richard Biener <rguenther@suse.de>
7765 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
7766 release SSA defs of pattern stmts.
7768 2014-02-17 Richard Biener <rguenther@suse.de>
7770 * tree-inline.c (expand_call_inline): Release the virtual
7771 operand defined by the call we are about to inline.
7773 2014-02-17 Richard Biener <rguenther@suse.de>
7775 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
7777 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
7778 Ilya Tocar <ilya.tocar@intel.com>
7780 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
7781 arguments order in builtin.
7782 (_mm512_permutexvar_epi64): Ditto.
7783 (_mm512_mask_permutexvar_epi64): Ditto
7784 (_mm512_maskz_permutexvar_epi32): Ditto
7785 (_mm512_permutexvar_epi32): Ditto
7786 (_mm512_mask_permutexvar_epi32): Ditto
7788 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7790 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
7791 (p8_vmrgow): Likewise.
7793 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7795 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
7798 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
7801 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
7802 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
7803 into 64-bit and 32-bit moves. On 64-bit moves, add support for
7804 using direct move instructions on ISA 2.07. Also adjust
7805 instruction length for 64-bit.
7806 (mov<mode>_64bit, TFmode/TDmode): Likewise.
7807 (mov<mode>_32bit, TFmode/TDmode): Likewise.
7809 2014-02-15 Alan Modra <amodra@gmail.com>
7813 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
7814 find_replacement on parts of insn rtl that might be reloaded.
7816 2014-02-15 Richard Biener <rguenther@suse.de>
7818 PR tree-optimization/60183
7819 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
7820 (tree_ssa_phiprop): Calculate and free post-dominators.
7822 2014-02-14 Jeff Law <law@redhat.com>
7824 PR rtl-optimization/60131
7825 * ree.c (get_extended_src_reg): New function.
7826 (combine_reaching_defs): Use it rather than assuming location of REG.
7827 (find_and_remove_re): Verify first operand of extension is
7828 a REG before adding the insns to the copy list.
7830 2014-02-14 Roland McGrath <mcgrathr@google.com>
7832 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
7833 * configure: Regenerated.
7834 * config.in: Regenerated.
7835 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
7836 instead of ASM_SHORT.
7838 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
7839 Richard Earnshaw <rearnsha@arm.com>
7841 PR rtl-optimization/59535
7842 * lra-constraints.c (process_alt_operands): Encourage alternative
7843 when unassigned pseudo class is superset of the alternative class.
7844 (inherit_reload_reg): Don't inherit when optimizing for code size.
7845 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
7846 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
7847 modes not less than 4 for Thumb1.
7849 2014-02-14 Kyle McMartin <kyle@redhat.com>
7852 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
7854 2014-02-14 Richard Biener <rguenther@suse.de>
7856 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
7857 (get_frame_arg): Drop the assert with langhook types_compatible_p.
7858 Do not strip INDIRECT_REFs.
7860 2014-02-14 Richard Biener <rguenther@suse.de>
7863 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
7864 DECL_FUNCTION_SPECIFIC_TARGET.
7865 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
7866 * tree-streamer-out.c (pack_ts_target_option): Remove.
7867 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
7868 (write_ts_function_decl_tree_pointers): Do not stream
7869 DECL_FUNCTION_SPECIFIC_TARGET.
7870 * tree-streamer-in.c (unpack_ts_target_option): Remove.
7871 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
7872 (lto_input_ts_function_decl_tree_pointers): Do not stream
7873 DECL_FUNCTION_SPECIFIC_TARGET.
7875 2014-02-14 Jakub Jelinek <jakub@redhat.com>
7877 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
7878 (get_initial_def_for_induction, vectorizable_induction): Ignore
7879 debug stmts when looking for exit_phi.
7880 (vectorizable_live_operation): Fix up condition.
7882 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7884 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
7885 nreverse() because it changes the content of original tree list.
7887 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7889 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
7890 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
7892 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7894 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
7895 GNU coding standards.
7897 2014-02-13 Jakub Jelinek <jakub@redhat.com>
7900 * dwarf2out.c (gen_subprogram_die): Don't call
7901 add_calling_convention_attribute if subr_die is old_die.
7903 2014-02-13 Sharad Singhai <singhai@google.com>
7905 * doc/optinfo.texi: Fix order of nodes.
7907 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
7909 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
7910 operands[2], not operands[3].
7912 2014-02-13 Richard Biener <rguenther@suse.de>
7915 * doc/install.texi (ISL): Update recommended version to 0.12.2,
7916 mention the possibility of an in-tree build.
7917 (CLooG): Update recommended version to 0.18.1, mention the
7918 possibility of an in-tree build and clarify that the ISL
7919 bundled with CLooG does not work.
7921 2014-02-13 Jakub Jelinek <jakub@redhat.com>
7924 * expr.c (compress_float_constant): If x is a hard register,
7925 extend into a pseudo and then move to x.
7927 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
7929 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
7930 caused by bad second argument to warning_at() with -mhotpatch and
7931 nested functions (e.g. with gfortran).
7933 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
7935 * opts.c (option_name): Remove "enabled by default" rider.
7937 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
7939 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
7941 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
7942 Uros Bizjak <ubizjak@gmail.com>
7945 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
7946 * configure: Regenerated.
7948 2014-02-12 Richard Biener <rguenther@suse.de>
7950 * vec.c (vec_prefix::calculate_allocation): Move as
7951 inline variant to vec.h.
7952 (vec_prefix::calculate_allocation_1): New out-of-line version.
7953 * vec.h (vec_prefix::calculate_allocation_1): Declare.
7954 (vec_prefix::m_has_auto_buf): Rename to ...
7955 (vec_prefix::m_using_auto_storage): ... this.
7956 (vec_prefix::calculate_allocation): Inline the easy cases
7957 and dispatch to calculate_allocation_1 which doesn't need the
7959 (va_heap::reserve): Use gcc_checking_assert.
7960 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
7961 m_using_auto_storage.
7962 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
7964 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
7965 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
7966 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
7968 2014-02-12 Richard Biener <rguenther@suse.de>
7970 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
7971 when we found a dependence.
7973 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
7975 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
7977 (maybe_fold_stmt): ... into this new function.
7978 * omp-low.c (lower_omp): Update comment.
7980 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
7983 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
7986 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
7988 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
7989 identifiers in comments.
7990 (cortexa53_extra_costs): Likewise.
7991 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
7992 (cortexa7_extra_costs): Likewise.
7993 (cortexa12_extra_costs): Likewise.
7994 (cortexa15_extra_costs): Likewise.
7995 (v7m_extra_costs): Likewise.
7997 2014-02-12 Richard Biener <rguenther@suse.de>
8000 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
8001 of posix_memalign being successful.
8002 (lower_stmt): Restrict lowering of posix_memalign to when
8003 -ftree-bit-ccp is enabled.
8005 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
8007 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
8009 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
8011 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
8013 PR rtl-optimization/60116
8014 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
8015 other_insn once the combination has been validated.
8017 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
8020 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
8022 * ipa-devirt.c: Include demangle.h
8023 (odr_violation_reported): New static variable.
8024 (add_type_duplicate): Update odr_violations.
8025 (maybe_record_node): Add completep parameter; update it.
8026 (record_target_from_binfo): Add COMPLETEP parameter;
8027 update it as needed.
8028 (possible_polymorphic_call_targets_1): Likewise.
8029 (struct polymorphic_call_target_d): Add nonconstruction_targets;
8030 rename FINAL to COMPLETE.
8031 (record_targets_from_bases): Sanity check we found the binfo;
8032 fix COMPLETEP updating.
8033 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
8034 parameter, fix computing of COMPLETEP.
8035 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
8036 at LTO time do demangling.
8037 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
8038 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
8040 (gimple_get_virt_method_for_binfo): Likewise.
8041 * gimple-fold.h (gimple_get_virt_method_for_binfo,
8042 gimple_get_virt_method_for_vtable): Update prototypes.
8044 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
8047 * genautomata.c (add_presence_absence): Fix typo with
8048 {final_}presence_list.
8050 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
8053 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
8054 for VSX/Altivec vectors that land in GPR registers.
8056 2014-02-11 Richard Henderson <rth@redhat.com>
8057 Jakub Jelinek <jakub@redhat.com>
8060 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
8061 around drhs if type conversion to lacc->type is not useless.
8063 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8065 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
8067 (cortex-a57.cortex-a53): Likewise.
8068 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
8070 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8072 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
8075 2014-02-11 Renlin Li <Renlin.Li@arm.com>
8077 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
8078 add_options_for_arm_vfp3.
8080 2014-02-11 Jeff Law <law@redhat.com>
8083 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
8084 object with an undesirable mode.
8086 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8089 * config/i386/sol2-9.h: New file.
8090 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
8091 *-*-solaris2.9*): Use it.
8093 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
8095 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
8096 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
8098 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
8100 * config/microblaze/microblaze.c: Extend mcpu version format
8102 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
8104 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
8106 2014-02-10 Richard Henderson <rth@redhat.com>
8109 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
8110 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
8111 ms-abi vs -mno-accumulate-outgoing-args.
8112 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
8113 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
8116 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
8119 * cfgexpand.c (expand_asm_operands): Attach source location to
8120 ASM_INPUT rtx objects.
8121 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
8123 2014-02-10 Nick Clifton <nickc@redhat.com>
8125 * config/mn10300/mn10300.c (popcount): New function.
8126 (mn10300_expand_prologue): Include saved registers in stack usage
8129 2014-02-10 Jeff Law <law@redhat.com>
8132 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
8133 when changing the SET_DEST of a prior insn to avoid an input reload.
8135 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
8137 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
8138 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
8139 -mcall-openbsd, or -mcall-linux.
8140 (CC1_ENDIAN_BIG_SPEC): Remove.
8141 (CC1_ENDIAN_LITTLE_SPEC): Remove.
8142 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
8143 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
8144 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
8145 and %cc1_endian_default.
8146 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
8148 2014-02-10 Richard Biener <rguenther@suse.de>
8150 PR tree-optimization/60115
8151 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
8152 MEM_REF handling. Properly verify that the accesses are not
8153 out of the objects bound.
8155 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8157 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
8160 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
8162 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
8163 proper constants and fix formatting.
8164 (possible_polymorphic_call_targets): Fix formatting.
8166 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
8167 Ilya Tocar <ilya.tocar@intel.com>
8169 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
8170 (_mm512_loadu_epi32): Renamed into...
8171 (_mm512_loadu_si512): This.
8172 (_mm512_storeu_epi32): Renamed into...
8173 (_mm512_storeu_si512): This.
8174 (_mm512_maskz_ceil_ps): Removed.
8175 (_mm512_maskz_ceil_pd): Ditto.
8176 (_mm512_maskz_floor_ps): Ditto.
8177 (_mm512_maskz_floor_pd): Ditto.
8178 (_mm512_floor_round_ps): Ditto.
8179 (_mm512_floor_round_pd): Ditto.
8180 (_mm512_ceil_round_ps): Ditto.
8181 (_mm512_ceil_round_pd): Ditto.
8182 (_mm512_mask_floor_round_ps): Ditto.
8183 (_mm512_mask_floor_round_pd): Ditto.
8184 (_mm512_mask_ceil_round_ps): Ditto.
8185 (_mm512_mask_ceil_round_pd): Ditto.
8186 (_mm512_maskz_floor_round_ps): Ditto.
8187 (_mm512_maskz_floor_round_pd): Ditto.
8188 (_mm512_maskz_ceil_round_ps): Ditto.
8189 (_mm512_maskz_ceil_round_pd): Ditto.
8190 (_mm512_expand_pd): Ditto.
8191 (_mm512_expand_ps): Ditto.
8192 * config/i386/i386.c (ix86_builtins): Remove
8193 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
8194 (bdesc_args): Ditto.
8195 * config/i386/predicates.md (const1256_operand): New.
8196 (const_1_to_2_operand): Ditto.
8197 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
8198 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8199 (*avx512pf_gatherpf<mode>sf): Ditto.
8200 (avx512pf_gatherpf<mode>df): Ditto.
8201 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8202 (*avx512pf_gatherpf<mode>df): Ditto.
8203 (avx512pf_scatterpf<mode>sf): Ditto.
8204 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8205 (*avx512pf_scatterpf<mode>sf): Ditto.
8206 (avx512pf_scatterpf<mode>df): Ditto.
8207 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8208 (*avx512pf_scatterpf<mode>df): Ditto.
8209 (avx512f_expand<mode>): Removed.
8210 (<shift_insn><mode>3<mask_name>): Change predicate type.
8212 2014-02-08 Jakub Jelinek <jakub@redhat.com>
8214 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
8215 not at the end of datarefs vector use ordered_remove to avoid
8216 reordering datarefs vector.
8219 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
8220 mark local addressable non-static vars as GOVD_PRIVATE
8221 instead of GOVD_LOCAL.
8222 * omp-low.c (lower_omp_for): Move gimple_bind_vars
8223 and BLOCK_VARS of gimple_bind_block to new_stmt rather
8227 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
8228 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
8229 assume_aligned or alloc_align attributes.
8230 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
8231 arguments. Handle also assume_aligned and alloc_align attributes.
8232 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
8233 calls to functions with assume_aligned or alloc_align attributes.
8234 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
8236 2014-02-08 Terry Guo <terry.guo@arm.com>
8238 * doc/invoke.texi: Document ARM -march=armv7e-m.
8240 2014-02-08 Jakub Jelinek <jakub@redhat.com>
8242 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
8243 flag on __cilkrts_rethrow builtin.
8246 * ipa-cp.c (determine_versionability): Fail at -O0
8247 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
8248 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
8251 2014-02-04 Jakub Jelinek <jakub@redhat.com>
8254 * tree-inline.c (copy_forbidden): Fail for
8255 __attribute__((optimize (0))) functions.
8257 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
8259 * varpool.c: Include pointer-set.h.
8260 (varpool_remove_unreferenced_decls): Variables in other partitions
8261 will not be output; be however careful to not lose information
8264 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
8266 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
8267 lookup in the vtable constructor.
8269 2014-02-07 Jeff Law <law@redhat.com>
8272 * config/m68k/m68k.md (ashldi_extsi): Turn into a
8273 define_insn_and_split.
8275 * ipa-inline.c (inline_small_functions): Fix typos.
8277 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8279 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
8280 (s390_can_use_return_insn): Declare.
8281 * config/s390/s390.h (EPILOGUE_USES): Define.
8282 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
8284 (s390_chunkify_start): Handle return JUMP_LABELs.
8285 (s390_early_mach): Emit a main_pool instruction on the entry edge.
8286 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
8287 (s390_can_use_return_insn): New functions.
8288 (s390_fix_long_loop_prediction): Handle conditional returns.
8289 (TARGET_SET_UP_BY_PROLOGUE): Define.
8290 * config/s390/s390.md (ANY_RETURN): New code iterator.
8291 (*creturn, *csimple_return, return, simple_return): New patterns.
8293 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8295 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
8296 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
8297 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
8298 REG_CFA_RESTORE list when deciding not to restore a register.
8300 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8302 * config/s390/s390.c: Include tree-pass.h and context.h.
8303 (s390_early_mach): New function, split out from...
8304 (s390_emit_prologue): ...here.
8305 (pass_data_s390_early_mach): New pass structure.
8306 (pass_s390_early_mach): New class.
8307 (s390_option_override): Create and register early_mach pass.
8308 Move to end of file.
8310 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8312 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
8313 to match for the exit block.
8315 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8317 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
8318 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
8319 Reject misaligned operands.
8321 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8323 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
8325 2014-02-07 Richard Biener <rguenther@suse.de>
8328 * gimple-low.c (lower_builtin_posix_memalign): New function.
8329 (lower_stmt): Call it to lower posix_memalign in a way
8330 to make alignment info accessible.
8332 2014-02-07 Jakub Jelinek <jakub@redhat.com>
8335 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
8336 __builtin_setjmp_receiver.
8338 2014-02-07 Richard Biener <rguenther@suse.de>
8341 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
8342 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
8343 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
8344 Handle BUILT_IN_POSIX_MEMALIGN.
8345 (find_func_clobbers): Likewise.
8346 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
8347 (call_may_clobber_ref_p_1): Likewise.
8349 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8352 * ipa-devirt.c (record_target_from_binfo): Remove overactive
8355 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8358 * lto-cgraph.c (lto_output_node): Use
8359 symtab_get_symbol_partitioning_class.
8360 (lto_output_varpool_node): likewise.
8361 (symtab_get_symbol_partitioning_class): Move here from
8363 * cgraph.h (symbol_partitioning_class): Likewise.
8364 (symtab_get_symbol_partitioning_class): Declare.
8366 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8368 * ggc.h (ggc_internal_cleared_alloc): New macro.
8369 * vec.h (vec_safe_copy): Handle memory stats.
8370 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
8371 * target-globals.c (save_target_globals): Likewise.
8373 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8376 * expr.c (emit_move_resolve_push): Export; be bit more selective
8377 on when to clear alias set.
8378 * expr.h (emit_move_resolve_push): Declare.
8379 * function.h (struct function): Add tail_call_marked.
8380 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
8381 * config/i386/i386-protos.h (ix86_expand_push): Remove.
8382 * config/i386/i386.md (TImode move expander): De not call
8384 (FP push expanders): Preserve memory attributes.
8385 * config/i386/sse.md (push<mode>1): Remove.
8386 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
8387 (ix86_expand_push): Remove.
8388 * config/i386/mmx.md (push<mode>1): Remove.
8390 2014-02-06 Jakub Jelinek <jakub@redhat.com>
8392 PR rtl-optimization/60030
8393 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
8394 lopart with paradoxical subreg before shifting it up by hprec.
8396 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8398 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
8399 Remove extra newline at end of file.
8400 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
8401 (arm_issue_rate): Handle cortexa57.
8402 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
8403 (cortex-a57.cortex-a53): Likewise.
8405 2014-02-06 Jakub Jelinek <jakub@redhat.com>
8408 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
8409 don't record in REG_FRAME_RELATED_EXPR registers not set in that
8411 (arm_expand_prologue): Adjust all callers.
8412 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
8413 info, registers also at the lowest numbered registers side. Use
8414 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
8418 * var-tracking.c (adjust_mems): Before adding a SET to
8419 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
8421 2014-02-06 Alan Modra <amodra@gmail.com>
8424 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
8425 change SDmode to DDmode when lra_in_progress.
8427 2014-02-06 Jakub Jelinek <jakub@redhat.com>
8430 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
8431 free_data_ref on the dr first, and before goto again also set dr
8432 to the next dr. For simd_lane_access, free old datarefs[i] before
8433 overwriting it. For get_vectype_for_scalar_type failure, don't
8434 free_data_ref if simd_lane_access.
8436 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
8439 * tree.h (opts_for_fn): New inline function.
8440 (opt_for_fn): Define.
8441 * config/i386/i386.c (ix86_function_regparm): Use
8442 opt_for_fn (decl, optimize) instead of optimize.
8444 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
8446 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
8447 for SYMBOL_REF in large memory model.
8449 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8451 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
8453 (cortex-a57): Likewise.
8454 (cortex-a57.cortex-a53): Likewise.
8456 2014-02-06 Yury Gribov <y.gribov@samsung.com>
8457 Kugan Vivekanandarajah <kuganv@linaro.org>
8459 * config/arm/arm.c (arm_vector_alignment_reachable): Check
8461 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
8463 2014-02-06 Richard Biener <rguenther@suse.de>
8465 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
8466 set_loop_copy and initialize_original_copy_tables.
8468 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
8470 * config/aarch64/aarch64-simd.md
8471 (aarch64_ashr_simddi): Change QI to SI.
8473 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
8474 Jakub Jelinek <jakub@redhat.com>
8477 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
8480 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8482 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
8483 CODE_FOR_altivec_vpku[hw]um to
8484 CODE_FOR_altivec_vpku[hw]um_direct.
8485 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
8486 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
8487 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
8488 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
8490 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8492 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
8493 generation for -maltivec=be.
8494 (altivec_vsumsws): Simplify redundant test.
8496 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8498 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
8499 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
8500 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
8501 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
8502 gen_altivec_vpkuwum.
8503 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
8505 (altivec_vpks<VI_char>ss): Likewise.
8506 (altivec_vpks<VI_char>us): Likewise.
8507 (altivec_vpku<VI_char>us): Likewise.
8508 (altivec_vpku<VI_char>um): Likewise.
8509 (altivec_vpku<VI_char>um_direct): New (copy of
8510 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
8512 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
8513 target is little endian and -maltivec=be is not specified.
8514 (*altivec_vupkhs<VU_char>_direct): New (copy of
8515 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
8516 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
8517 target is little endian and -maltivec=be is not specified.
8518 (*altivec_vupkls<VU_char>_direct): New (copy of
8519 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
8520 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
8521 little endian and -maltivec=be is not specified.
8522 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
8523 little endian and -maltivec=be is not specified.
8525 2014-02-05 Richard Henderson <rth@redhat.com>
8528 * combine-stack-adj.c: Revert r206943.
8529 * sched-int.h (struct deps_desc): Add last_args_size.
8530 * sched-deps.c (init_deps): Initialize it.
8531 (sched_analyze_insn): Add OUTPUT dependencies between insns that
8532 contain REG_ARGS_SIZE notes.
8534 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
8536 * lto-cgraph.c (asm_nodes_output): Make global.
8537 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
8538 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
8539 (driver_handle_option): Handle OPT_fwpa.
8541 2014-02-05 Jakub Jelinek <jakub@redhat.com>
8544 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
8545 a comment typo and formatting issue. If odr_hash hasn't been
8546 created, return vNULL and set *completep to false.
8549 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
8550 bb with no successors.
8552 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
8555 * doc/invoke.texi (-march): Clarify documentation for ARM.
8559 2014-02-05 Richard Biener <rguenther@suse.de>
8561 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
8562 when not vectorizing because of too many alias checks.
8563 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
8564 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
8566 2014-02-05 Nick Clifton <nickc@redhat.com>
8568 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
8569 accept extended registers in any mode when compiling for the MN10300.
8571 2014-02-05 Yury Gribov <y.gribov@samsung.com>
8573 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
8574 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
8575 sanitization attributes.
8576 (can_inline_edge_p): Likewise.
8577 (sanitize_attrs_match_for_inline_p): New function.
8579 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8581 * ipa-prop.c (detect_type_change): Shor circuit testing of
8582 type changes on THIS pointer.
8584 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
8587 * config/pa/pa.c (legitimize_tls_address): Return original address
8588 if not passed a SYMBOL_REF rtx.
8589 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
8591 (pa_emit_move_sequence): Simplify TLS source operands.
8592 (pa_legitimate_constant_p): Reject all TLS constants.
8593 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
8594 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
8596 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8598 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
8599 groups when we know they are controlled by LTO.
8600 * varasm.c (default_binds_local_p_1): If object is in other partition,
8601 it will be resolved locally.
8603 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
8605 * config/host-linux.c (linux_gt_pch_use_address): Don't
8606 use SSIZE_MAX because it is not always defined.
8608 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
8611 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
8612 threshold for pseudo splitting.
8613 (update_ebb_live_info): Process call argument hard registers and
8614 hard registers from insn definition too.
8615 (max_small_class_regs_num): New constant.
8616 (inherit_in_ebb): Update live hard regs through EBBs. Update
8617 reloads_num only for small register classes. Don't split for
8620 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
8623 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
8626 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8628 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
8631 2014-02-04 Marek Polacek <polacek@redhat.com>
8633 * gdbinit.in (pel): Define.
8635 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
8637 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
8640 2014-02-04 Richard Biener <rguenther@suse.de>
8643 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
8644 in function context local.
8645 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
8646 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
8647 similar to LTO_imported_decl_ref.
8649 2014-02-04 Jakub Jelinek <jakub@redhat.com>
8651 PR tree-optimization/60002
8652 * cgraphclones.c (build_function_decl_skip_args): Clear
8655 PR tree-optimization/60023
8656 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
8657 false to gsi_replace.
8658 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
8659 has been in some EH region and vec_stmt could throw, add
8660 vec_stmt into the same EH region.
8661 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
8662 has no lhs, ignore it.
8663 * internal-fn.c (expand_MASK_LOAD): Likewise.
8666 * tree-inline.c (copy_forbidden): Fail for
8667 __attribute__((optimize (0))) functions.
8670 * omp-low.c (simd_clone_struct_copy): If from->inbranch
8671 is set, copy one less argument.
8672 (expand_simd_clones): Don't subtract clone_info->inbranch
8673 from simd_clone_struct_alloc argument.
8675 PR rtl-optimization/57915
8676 * recog.c (simplify_while_replacing): If all unary/binary/relational
8677 operation arguments are constant, attempt to simplify those.
8680 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
8681 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
8683 2014-02-04 Richard Biener <rguenther@suse.de>
8685 PR tree-optimization/60012
8686 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
8687 TBAA disambiguation to all DDRs.
8689 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8692 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
8693 (LINK_SPEC): Use it for -shared, -shared-libgcc.
8695 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8698 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
8700 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8702 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
8703 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
8705 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8708 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
8709 to figure out targets of polymorphic calls with known decl.
8710 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8711 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
8712 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
8713 (get_polymorphic_call_info): ... here.
8714 (get_polymorphic_call_info_from_invariant): New function.
8716 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8718 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
8719 lookup via vtable pointer; check for type consistency
8720 and turn inconsitent facts into UNREACHABLE.
8721 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8722 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
8723 type inconsistent querries; return UNREACHABLE instead.
8725 2014-02-03 Richard Henderson <rth@twiddle.net>
8728 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
8729 already processed this node.
8730 (normalize_one_pred_1): Pass along mark_set.
8731 (normalize_one_pred): Create and destroy a pointer_set_t.
8732 (normalize_one_pred_chain): Likewise.
8734 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
8736 PR gcov-profile/58602
8737 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
8739 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8742 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
8743 -fno-devirtualize; try to devirtualize by the knowledge of
8744 virtual table pointer given by aggregate propagation.
8745 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8746 (ipa_print_node_jump_functions): Dump also offset that
8747 is relevant for polymorphic calls.
8748 (determine_known_aggregate_parts): Add arg_type parameter; use it
8749 instead of determining the type from pointer type.
8750 (ipa_compute_jump_functions_for_edge): Update call of
8751 determine_known_aggregate_parts.
8752 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
8753 (gimple_get_virt_method_for_binfo): ... here; simplify using
8754 vtable_pointer_value_to_vtable.
8755 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
8756 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
8757 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
8758 (vtable_pointer_value_to_vtable): Break out from ...; handle also
8760 (vtable_pointer_value_to_binfo): ... here.
8761 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
8763 2014-02-03 Teresa Johnson <tejohnson@google.com>
8765 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
8766 redef of outer loop index variable.
8768 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
8772 * doc/extend.texi (Function Attributes): Typo.
8774 2014-02-03 Cong Hou <congh@google.com>
8776 PR tree-optimization/60000
8777 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
8778 if the vectorized statement is a store. A store statement can only
8779 appear at the end of pattern statements.
8781 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
8783 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
8784 (ix86_option_override_internal): Default long double to 64-bit for
8785 32-bit Bionic and to 128-bit for 64-bit Bionic.
8787 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
8788 TARGET_LONG_DOUBLE_128 is true.
8789 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
8791 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
8792 (mlong-double-64): Negate -mlong-double-128.
8793 (mlong-double-128): New option.
8795 * config/i386/i386-c.c (ix86_target_macros): Define
8796 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
8798 * doc/invoke.texi: Document -mlong-double-128.
8800 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
8802 PR rtl-optimization/60024
8803 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
8805 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
8807 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
8809 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
8811 PR rtl-optimization/57662
8812 * sel-sched.c (code_motion_path_driver): Do not mark already not
8813 existing blocks in the visiting bitmap.
8815 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
8817 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
8818 on the insn being emitted.
8820 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
8821 Will Deacon <will.deacon@arm.com>
8823 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
8825 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8827 * config/arm/arm-tables.opt: Regenerate.
8829 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8831 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
8832 for vector types other than V16QImode.
8833 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
8834 define_expand, and call altivec_expand_vec_perm_le when producing
8835 code with little endian element order.
8836 (*altivec_vperm_<mode>_internal): New insn having previous
8837 behavior of altivec_vperm_<mode>.
8838 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
8839 altivec_expand_vec_perm_le when producing code with little endian
8841 (*altivec_vperm_<mode>_uns_internal): New insn having previous
8842 behavior of altivec_vperm_<mode>_uns.
8844 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8846 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
8847 (altivec_vsumsws): Add handling for -maltivec=be with a little
8849 (altivec_vsumsws_direct): New.
8850 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
8851 gen_altivec_vsumsws.
8853 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
8855 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
8856 vtable_pointer_value_to_binfo): New functions.
8857 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
8858 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
8860 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
8862 * config/nios2/nios2.md (load_got_register): Initialize GOT
8863 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
8864 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
8866 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
8868 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
8869 preserverd by passthrough, do not propagate the type.
8871 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8873 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
8874 (mips_atomic_assign_expand_fenv): New function.
8875 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
8877 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8879 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
8880 (__builtin_mips_set_fcsr): Likewise.
8881 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
8882 MIPS_USI_FTYPE_VOID.
8883 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
8884 (mips16_expand_set_fcsr): Likewise.
8885 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
8886 (mips16_set_fcsr_stub): Likewise.
8887 (mips16_get_fcsr_one_only_stub): New class.
8888 (mips16_set_fcsr_one_only_stub): Likewise.
8889 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
8890 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
8891 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
8892 (hard_float): New availability predicate.
8893 (mips_builtins): Add get_fcsr and set_fcsr.
8894 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
8895 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
8896 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
8897 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
8898 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
8901 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8903 * config/mips/mips.c (mips_one_only_stub): New class.
8904 (mips_need_mips16_rdhwr_p): Replace with...
8905 (mips16_rdhwr_stub): ...this new variable.
8906 (mips16_stub_call_address): New function.
8907 (mips16_rdhwr_one_only_stub): New class.
8908 (mips_expand_thread_pointer): Use mips16_stub_call_address.
8909 (mips_output_mips16_rdhwr): Delete.
8910 (mips_finish_stub): New function.
8911 (mips_code_end): Use it to handle rdhwr stubs.
8913 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
8916 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
8917 when calculating size of integer atomic types.
8919 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
8921 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
8923 2014-02-01 Jakub Jelinek <jakub@redhat.com>
8925 PR tree-optimization/60003
8926 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
8927 * profile.c (branch_prob): Use gimple_call_builtin_p
8928 to check for BUILT_IN_SETJMP_RECEIVER.
8929 * tree-inline.c (copy_bb): Call notice_special_calls.
8931 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
8934 * lra-constraints.c (process_alt_operands): Update reload_sum only
8937 2014-01-31 Richard Henderson <rth@redhat.com>
8940 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
8941 until after else_eh is processed.
8943 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
8945 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
8946 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
8947 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
8948 in smmintrin.h, remove them.
8949 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
8950 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
8951 * config/i386/i386.md (ROUND_SAE): Fix value.
8952 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
8953 (const48_operand): New.
8954 * config/i386/subst.md (round), (round_expand): Use
8955 const_4_or_8_to_11_operand.
8956 (round_saeonly), (round_saeonly_expand): Use const48_operand.
8958 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
8960 * config/i386/constraints.md (Yk): Swap meaning with k.
8961 * config/i386/i386.md (movhi_internal): Change Yk to k.
8962 (movqi_internal): Ditto.
8963 (*k<logic><mode>): Ditto.
8966 (kandn<mode>): Ditto.
8967 (*<code>hi_1): Ditto.
8968 (*<code>qi_1): Ditto.
8969 (kxnor<mode>): Ditto.
8970 (kortestzhi): Ditto.
8971 (kortestchi): Ditto.
8973 (*one_cmplhi2_1): Ditto.
8974 (*one_cmplqi2_1): Ditto.
8975 * config/i386/sse.md (): Change k to Yk.
8976 (avx512f_load<mode>_mask): Ditto.
8977 (avx512f_blendm<mode>): Ditto.
8978 (avx512f_store<mode>_mask): Ditto.
8979 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
8980 (avx512f_storedqu<mode>_mask): Ditto.
8981 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
8983 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
8984 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
8985 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
8986 (avx512f_maskcmp<mode>3): Ditto.
8987 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
8988 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
8989 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
8990 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
8991 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
8992 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
8993 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
8994 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
8995 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
8996 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
8997 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
8998 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
8999 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
9000 (vec_extract_lo_<mode>_maskm): Ditto.
9001 (vec_extract_hi_<mode>_maskm): Ditto.
9002 (avx512f_vternlog<mode>_mask): Ditto.
9003 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
9004 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
9005 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
9006 (avx512f_<code>v8div16qi2_mask): Ditto.
9007 (avx512f_<code>v8div16qi2_mask_store): Ditto.
9008 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
9009 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
9010 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
9011 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
9012 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9013 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9014 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9015 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9016 (avx512cd_maskb_vec_dupv8di): Ditto.
9017 (avx512cd_maskw_vec_dupv16si): Ditto.
9018 (avx512f_vpermi2var<mode>3_maskz): Ditto.
9019 (avx512f_vpermi2var<mode>3_mask): Ditto.
9020 (avx512f_vpermi2var<mode>3_mask): Ditto.
9021 (avx512f_vpermt2var<mode>3_maskz): Ditto.
9022 (*avx512f_gathersi<mode>): Ditto.
9023 (*avx512f_gathersi<mode>_2): Ditto.
9024 (*avx512f_gatherdi<mode>): Ditto.
9025 (*avx512f_gatherdi<mode>_2): Ditto.
9026 (*avx512f_scattersi<mode>): Ditto.
9027 (*avx512f_scatterdi<mode>): Ditto.
9028 (avx512f_compress<mode>_mask): Ditto.
9029 (avx512f_compressstore<mode>_mask): Ditto.
9030 (avx512f_expand<mode>_mask): Ditto.
9031 * config/i386/subst.md (mask): Change k to Yk.
9032 (mask_scalar_merge): Ditto.
9035 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
9037 * doc/extend.texi (Vector Extensions): Document ?: in C++.
9039 2014-01-31 Richard Biener <rguenther@suse.de>
9042 * builtins.c (fold_builtin_memory_op): Make sure to not
9043 use a floating-point mode or a boolean or enumeral type for
9046 2014-01-30 DJ Delorie <dj@redhat.com>
9048 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
9049 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
9050 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
9051 whenever main() has an epilogue.
9053 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9055 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
9056 unused variable "field".
9057 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
9058 (vsx_mergeh_<mode>): Likewise.
9059 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
9060 (altivec_vmrghh): Likewise.
9061 (altivec_vmrghw): Likewise.
9062 (altivec_vmrglb): Likewise.
9063 (altivec_vmrglh): Likewise.
9064 (altivec_vmrglw): Likewise.
9065 (altivec_vspltb): Add missing uses.
9066 (altivec_vsplth): Likewise.
9067 (altivec_vspltw): Likewise.
9068 (altivec_vspltsf): Likewise.
9070 2014-01-30 Jakub Jelinek <jakub@redhat.com>
9073 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
9074 frame related instructions.
9076 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
9078 PR rtl-optimization/59959
9079 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
9080 any reload of register whose subreg is invalid.
9082 2014-01-30 Jakub Jelinek <jakub@redhat.com>
9084 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
9085 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
9086 Add missing return type - void.
9088 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9090 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
9091 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
9092 remove element index adjustment for endian (now handled in vsx.md
9094 (altivec_expand_vec_perm_const): Use
9095 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
9096 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
9097 (vsx_xxspltw_<mode>): Adjust element index for little endian.
9098 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
9099 define_expand and a new define_insn *altivec_vspltb_internal;
9100 adjust for -maltivec=be on a little endian target.
9101 (altivec_vspltb_direct): New.
9102 (altivec_vsplth): Divide into a define_expand and a new
9103 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
9104 little endian target.
9105 (altivec_vsplth_direct): New.
9106 (altivec_vspltw): Divide into a define_expand and a new
9107 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
9108 little endian target.
9109 (altivec_vspltw_direct): New.
9110 (altivec_vspltsf): Divide into a define_expand and a new
9111 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
9112 a little endian target.
9114 2014-01-30 Richard Biener <rguenther@suse.de>
9116 PR tree-optimization/59993
9117 * tree-ssa-forwprop.c (associate_pointerplus): Check we
9118 can propagate form the earlier stmt and avoid the transform
9119 when the intermediate result is needed.
9121 2014-01-30 Alangi Derick <alangiderick@gmail.com>
9123 * README.Portability: Fix typo.
9125 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
9127 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
9128 comparison_operator with ordered_comparison_operator.
9130 2014-01-30 Nick Clifton <nickc@redhat.com>
9132 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
9133 Rename to mn10300_store_multiple_regs.
9134 * config/mn10300/mn10300.c: Likewise.
9135 * config/mn10300/mn10300.md (store_movm): Fix typo: call
9136 store_multiple_regs.
9137 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
9138 Call mn10300_store_multiple_regs.
9140 2014-01-30 Nick Clifton <nickc@redhat.com>
9141 DJ Delorie <dj@redhat.com>
9143 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
9144 %fp 2 to keep registers after it properly word-aligned.
9145 (rl78_alloc_physical_registers_umul): Handle the case where both
9146 input operands are the same.
9148 2014-01-30 Richard Biener <rguenther@suse.de>
9150 PR tree-optimization/59903
9151 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
9154 2014-01-30 Jason Merrill <jason@redhat.com>
9157 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
9160 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
9162 2014-01-30 Richard Biener <rguenther@suse.de>
9164 PR tree-optimization/59951
9165 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
9167 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
9170 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
9171 SFmode to DFmode case.
9173 2014-01-29 DJ Delorie <dj@redhat.com>
9175 * config/msp430/msp430.opt (-minrt): New.
9176 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
9178 (ENDFILE_SPEC): Likewise.
9180 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
9182 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
9183 (estimate_function_body_sizes): Use it.
9185 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
9188 * dwarf2out.c (is_cxx_auto): New.
9189 (is_base_type): Use it.
9190 (gen_type_die_with_usage): Likewise.
9192 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9194 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
9195 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
9196 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
9197 -maltivec=be with LE targets.
9198 (vsx_mergeh_<mode>): Likewise.
9199 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
9200 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
9201 (altivec_vmrghb): Replace with define_expand and new
9202 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
9203 (altivec_vmrghb_direct): New define_insn.
9204 (altivec_vmrghh): Replace with define_expand and new
9205 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
9206 (altivec_vmrghh_direct): New define_insn.
9207 (altivec_vmrghw): Replace with define_expand and new
9208 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
9209 (altivec_vmrghw_direct): New define_insn.
9210 (*altivec_vmrghsf): Adjust for endianness.
9211 (altivec_vmrglb): Replace with define_expand and new
9212 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
9213 (altivec_vmrglb_direct): New define_insn.
9214 (altivec_vmrglh): Replace with define_expand and new
9215 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
9216 (altivec_vmrglh_direct): New define_insn.
9217 (altivec_vmrglw): Replace with define_expand and new
9218 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
9219 (altivec_vmrglw_direct): New define_insn.
9220 (*altivec_vmrglsf): Adjust for endianness.
9221 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
9222 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
9223 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
9224 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
9225 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
9226 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
9227 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
9228 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
9230 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
9232 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
9233 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
9236 2014-01-29 Richard Biener <rguenther@suse.de>
9238 PR tree-optimization/58742
9239 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
9240 associate_pointerplus_align.
9241 (associate_pointerplus_diff): New function.
9242 (associate_pointerplus): Likewise. Call associate_pointerplus_align
9243 and associate_pointerplus_diff.
9245 2014-01-29 Richard Biener <rguenther@suse.de>
9247 * lto-streamer.h (LTO_major_version): Bump to 3.
9248 (LTO_minor_version): Reset to 0.
9250 2014-01-29 Renlin Li <Renlin.Li@arm.com>
9252 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
9253 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
9254 (arm_file_start): Generate correct asm header for armv7ve.
9255 * config/arm/bpabi.h: Add multilib support for armv7ve.
9256 * config/arm/driver-arm.c: Change the architectures of cortex-a7
9257 and cortex-a15 to armv7ve.
9258 * config/arm/t-aprofile: Add multilib support for armv7ve.
9259 * doc/invoke.texi: Document -march=armv7ve.
9261 2014-01-29 Richard Biener <rguenther@suse.de>
9263 PR tree-optimization/58742
9264 * tree-ssa-forwprop.c (associate_plusminus): Return true
9265 if we changed sth, defer EH cleanup to ...
9266 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
9267 (simplify_mult): New function.
9269 2014-01-29 Jakub Jelinek <jakub@redhat.com>
9272 PR tree-optimization/59920
9273 * tree.c (build_common_builtin_nodes): Remove
9274 __builtin_setjmp_dispatcher initialization.
9275 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
9276 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
9277 instead of gsi_after_labels + manually skipping debug stmts.
9278 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
9279 ignore bbs with IFN_ABNORMAL_DISPATCHER.
9280 * tree-inline.c (copy_edges_for_bb): Remove
9281 can_make_abnormal_goto argument, instead add abnormal_goto_dest
9282 argument. Ignore computed_goto_p stmts. Don't call
9283 make_abnormal_goto_edges. If a call might need abnormal edges
9284 for non-local gotos, see if it already has an edge to
9285 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
9286 with true argument, don't do anything then, otherwise add
9287 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
9288 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
9290 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
9291 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
9292 (lower_stmt): Don't set data->calls_builtin_setjmp.
9293 (lower_builtin_setjmp): Adjust comment.
9294 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
9295 * tree-cfg.c (found_computed_goto): Remove.
9296 (factor_computed_gotos): Remove.
9297 (make_goto_expr_edges): Return bool, true for computed gotos.
9298 Don't call make_abnormal_goto_edges.
9299 (build_gimple_cfg): Don't set found_computed_goto, don't call
9300 factor_computed_gotos.
9301 (computed_goto_p): No longer static.
9302 (make_blocks): Don't set found_computed_goto.
9303 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
9304 (make_edges): If make_goto_expr_edges returns true, push bb
9305 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
9306 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
9307 vector. Record mapping between bbs and OpenMP regions if there
9308 are any, adjust make_gimple_omp_edges caller. Call
9309 handle_abnormal_edges.
9310 (make_abnormal_goto_edges): Remove.
9311 * tree-cfg.h (make_abnormal_goto_edges): Remove.
9312 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
9313 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
9314 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
9315 * internal-fn.def (ABNORMAL_DISPATCHER): New.
9316 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
9317 filling *region also set *region_idx to (*region)->entry->index.
9320 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
9321 For REGs set ORIGINAL_REGNO.
9323 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
9325 * doc/md.texi: Mention that a target shouldn't implement
9326 vec_widen_(s|u)mul_even/odd pair if it is less efficient
9329 2014-01-29 Jakub Jelinek <jakub@redhat.com>
9331 PR tree-optimization/59594
9332 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
9333 a copy of the datarefs vector rather than the vector itself.
9335 2014-01-28 Jason Merrill <jason@redhat.com>
9338 * dwarf2out.c (auto_die): New static.
9339 (gen_type_die_with_usage): Handle C++1y 'auto'.
9340 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
9343 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
9346 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
9347 (SPEC_X32): Likewise.
9348 (SPEC_64): Likewise.
9349 * config/i386/i386.c (ix86_option_override_internal): Turn off
9350 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
9352 (x86_file_start): Output .code16gcc for TARGET_16BIT.
9353 * config/i386/i386.h (TARGET_16BIT): New macro.
9354 (TARGET_16BIT_P): Likewise.
9355 * config/i386/i386.opt: Add m16.
9356 * doc/invoke.texi: Document -m16.
9358 2014-01-28 Jakub Jelinek <jakub@redhat.com>
9360 PR preprocessor/59935
9361 * input.c (location_get_source_line): Bail out on when line number
9362 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
9364 2014-01-28 Richard Biener <rguenther@suse.de>
9366 PR tree-optimization/58742
9367 * tree-ssa-forwprop.c (associate_plusminus): Handle
9368 pointer subtraction of the form (T)(P + A) - (T)P.
9370 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9372 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
9375 2014-01-28 Richard Biener <rguenther@suse.de>
9378 2014-01-28 Richard Biener <rguenther@suse.de>
9380 PR rtl-optimization/45364
9381 PR rtl-optimization/59890
9382 * var-tracking.c (local_get_addr_clear_given_value): Handle
9383 already cleared slot.
9384 (val_reset): Handle not allocated local_get_addr_cache.
9385 (vt_find_locations): Use post-order on the inverted CFG.
9387 2014-01-28 Richard Biener <rguenther@suse.de>
9389 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
9391 2014-01-28 Richard Biener <rguenther@suse.de>
9393 PR rtl-optimization/45364
9394 PR rtl-optimization/59890
9395 * var-tracking.c (local_get_addr_clear_given_value): Handle
9396 already cleared slot.
9397 (val_reset): Handle not allocated local_get_addr_cache.
9398 (vt_find_locations): Use post-order on the inverted CFG.
9400 2014-01-28 Alan Modra <amodra@gmail.com>
9402 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
9403 * configure.ac <recursive call for build != host>: Define
9404 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
9405 and LD_FOR_BUILD too.
9406 * configure: Regenerate.
9408 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
9410 * config/i386/i386.c (get_builtin_code_for_version): Separate
9411 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
9412 Broadwell from Haswell.
9414 2014-01-27 Steve Ellcey <sellcey@mips.com>
9416 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
9417 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
9418 * config/mips/mips.c (mips_option_override): Change setting
9420 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
9421 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
9422 Change from Mask to Var.
9424 2014-01-27 Jeff Law <law@redhat.com>
9426 * ipa-inline.c (inline_small_functions): Fix typo.
9428 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
9430 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
9431 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
9432 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
9433 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
9434 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
9435 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
9436 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
9437 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
9438 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
9439 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
9440 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
9441 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
9442 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
9443 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
9444 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
9445 (_mm512_storeu_epi64): Ditto.
9446 (_mm512_cmpge_epi32_mask): Ditto.
9447 (_mm512_cmpge_epu32_mask): Ditto.
9448 (_mm512_cmpge_epi64_mask): Ditto.
9449 (_mm512_cmpge_epu64_mask): Ditto.
9450 (_mm512_cmple_epi32_mask): Ditto.
9451 (_mm512_cmple_epu32_mask): Ditto.
9452 (_mm512_cmple_epi64_mask): Ditto.
9453 (_mm512_cmple_epu64_mask): Ditto.
9454 (_mm512_cmplt_epi32_mask): Ditto.
9455 (_mm512_cmplt_epu32_mask): Ditto.
9456 (_mm512_cmplt_epi64_mask): Ditto.
9457 (_mm512_cmplt_epu64_mask): Ditto.
9458 (_mm512_cmpneq_epi32_mask): Ditto.
9459 (_mm512_cmpneq_epu32_mask): Ditto.
9460 (_mm512_cmpneq_epi64_mask): Ditto.
9461 (_mm512_cmpneq_epu64_mask): Ditto.
9462 (_mm512_expand_pd): Ditto.
9463 (_mm512_expand_ps): Ditto.
9464 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
9465 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
9466 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
9467 * config/i386/i386.c (ix86_builtins): Add
9468 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
9469 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
9470 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
9471 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
9472 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
9473 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
9474 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
9475 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
9476 IX86_BUILTIN_PMOVUSQW512_MEM.
9477 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
9478 __builtin_ia32_pmovsqd512mem_mask,
9479 __builtin_ia32_pmovqd512mem_mask,
9480 __builtin_ia32_pmovusqw512mem_mask,
9481 __builtin_ia32_pmovsqw512mem_mask,
9482 __builtin_ia32_pmovqw512mem_mask,
9483 __builtin_ia32_pmovusdw512mem_mask,
9484 __builtin_ia32_pmovsdw512mem_mask,
9485 __builtin_ia32_pmovdw512mem_mask,
9486 __builtin_ia32_pmovqb512mem_mask,
9487 __builtin_ia32_pmovusqb512mem_mask,
9488 __builtin_ia32_pmovsqb512mem_mask,
9489 __builtin_ia32_pmovusdb512mem_mask,
9490 __builtin_ia32_pmovsdb512mem_mask,
9491 __builtin_ia32_pmovdb512mem_mask.
9492 (bdesc_args): Add __builtin_ia32_expanddf512,
9493 __builtin_ia32_expandsf512.
9494 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
9495 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
9496 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
9497 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
9498 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
9499 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
9500 (avx512f_<code>v8div16qi2_mask_store): This.
9501 (avx512f_expand<mode>): New.
9503 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
9505 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
9507 (_mm512_mask_prefetch_i64gather_pd): Ditto.
9508 (_mm512_prefetch_i32scatter_pd): Ditto.
9509 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
9510 (_mm512_prefetch_i64scatter_pd): Ditto.
9511 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
9512 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
9513 (_mm512_mask_prefetch_i64gather_ps): Ditto.
9514 (_mm512_prefetch_i32scatter_ps): Ditto.
9515 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
9516 (_mm512_prefetch_i64scatter_ps): Ditto.
9517 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
9518 * config/i386/i386-builtin-types.def: Define
9519 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
9520 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
9521 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
9522 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
9523 IX86_BUILTIN_SCATTERPFQPD.
9524 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
9525 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
9526 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
9527 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
9528 __builtin_ia32_scatterpfqps.
9529 (ix86_expand_builtin): Expand new built-ins.
9530 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
9531 fix memory access data type.
9532 (*avx512pf_gatherpf<mode>_mask): Ditto.
9533 (*avx512pf_gatherpf<mode>): Ditto.
9534 (avx512pf_scatterpf<mode>): Ditto.
9535 (*avx512pf_scatterpf<mode>_mask): Ditto.
9536 (*avx512pf_scatterpf<mode>): Ditto.
9537 (GATHER_SCATTER_SF_MEM_MODE): New.
9538 (avx512pf_gatherpf<mode>df): Ditto.
9539 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9540 (*avx512pf_scatterpf<mode>df): Ditto.
9542 2014-01-27 Jakub Jelinek <jakub@redhat.com>
9545 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
9546 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
9549 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
9551 * common/config/arm/arm-common.c
9552 (arm_rewrite_mcpu): Handle multiple names.
9554 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
9556 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
9558 * gimple-builder.h (create_gimple_tmp): Delete.
9560 2014-01-27 Christian Bruel <christian.bruel@st.com>
9562 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
9565 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
9567 * config/pa/pa.md (call): Generate indirect long calls to non-local
9568 functions when outputing 32-bit code.
9569 (call_value): Likewise except for special call to buggy powf function.
9571 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
9572 portable runtime and PIC indirect calls.
9573 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
9574 and PIC call sequences. Use ldo instead of blr to set return register
9575 in PIC call sequence.
9577 2014-01-25 Walter Lee <walt@tilera.com>
9579 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
9580 avoid clobbering a live register.
9582 2014-01-25 Walter Lee <walt@tilera.com>
9584 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
9585 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
9586 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
9587 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
9589 2014-01-25 Walter Lee <walt@tilera.com>
9591 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
9592 arguments on even registers.
9593 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
9595 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
9596 (BIGGEST_ALIGNMENT): Ditto.
9597 (BIGGEST_FIELD_ALIGNMENT): Ditto.
9599 2014-01-25 Walter Lee <walt@tilera.com>
9601 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
9602 insns before bundling.
9603 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
9605 2014-01-25 Walter Lee <walt@tilera.com>
9607 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
9608 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
9609 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
9611 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
9613 * config/mips/constraints.md (kl): Delete.
9614 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
9615 define expands, using...
9616 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
9617 instructions for MIPS16.
9618 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
9619 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
9621 2014-01-25 Walter Lee <walt@tilera.com>
9623 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
9627 2014-01-25 Walter Lee <walt@tilera.com>
9629 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
9630 (TARGET_EXPAND_TO_RTL_HOOK): Define.
9632 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
9634 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
9637 2014-01-25 Jakub Jelinek <jakub@redhat.com>
9639 * print-rtl.c (in_call_function_usage): New var.
9640 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
9641 EXPR_LIST mode as mode and not as reg note name.
9644 * cfgloopmanip.c (copy_loop_info): If
9645 loop->warned_aggressive_loop_optimizations, make sure
9646 the flag is set in target loop too.
9648 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
9650 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
9652 * builtins.def: Likewise.
9653 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
9654 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
9655 * ira.c (ira_setup_eliminable_regset): Likewise.
9656 * omp-low.c (gate_expand_omp): Likewise.
9657 (execute_lower_omp): Likewise.
9658 (diagnose_sb_0): Likewise.
9659 (gate_diagnose_omp_blocks): Likewise.
9660 (simd_clone_clauses_extract): Likewise.
9663 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9665 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
9666 correction for little endian...
9667 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
9670 2014-01-24 Jeff Law <law@redhat.com>
9672 PR tree-optimization/59919
9673 * tree-vrp.c (find_assert_locations_1): Do not register asserts
9674 for non-returning calls.
9676 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
9678 * common/config/aarch64/aarch64-common.c
9679 (aarch64_rewrite_mcpu): Handle multiple names.
9680 * config/aarch64/aarch64.h
9681 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
9683 2014-01-24 Dodji Seketeli <dodji@redhat.com>
9685 * input.c (add_file_to_cache_tab): Handle the case where fopen
9688 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
9691 * config/i386/i386.md (pushsf splitter): Get stack adjustment
9692 from push operand if code of push isn't PRE_DEC.
9694 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
9697 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
9698 -mquad-memory-atomic. Update -mquad-memory documentation to say
9699 it is only used for non-atomic loads/stores.
9701 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
9702 -mquad-memory or -mquad-memory-atomic switches.
9704 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
9705 -mquad-memory-atomic to ISA 2.07 support.
9707 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
9708 to separate support of normal quad word memory operations (ldq, stq)
9709 from the atomic quad word memory operations.
9711 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
9712 support to separate non-atomic quad word operations from atomic
9713 quad word operations. Disable non-atomic quad word operations in
9714 little endian mode so that we don't have to swap words after the
9715 load and before the store.
9716 (quad_load_store_p): Add comment about atomic quad word support.
9717 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
9718 options printed with -mdebug=reg.
9720 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
9721 -mquad-memory-atomic as the test for whether we have quad word
9722 atomic instructions.
9723 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
9724 or -mp8-vector are used, allow byte/half-word atomic operations.
9726 * config/rs6000/sync.md (load_lockedti): Insure that the address
9727 is a proper indexed or indirect address for the lqarx instruction.
9728 On little endian systems, swap the hi/lo registers after the lqarx
9730 (load_lockedpti): Use indexed_or_indirect_operand predicate to
9731 insure the address is valid for the lqarx instruction.
9732 (store_conditionalti): Insure that the address is a proper indexed
9733 or indirect address for the stqcrx. instruction. On little endian
9734 systems, swap the hi/lo registers before doing the stqcrx.
9736 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
9737 insure the address is valid for the stqcrx. instruction.
9739 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
9740 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
9741 type of quad memory support is available.
9743 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
9746 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
9747 there is a danger of looping.
9749 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
9751 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
9752 force flag_ira_loop_pressure if set via command line.
9754 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
9756 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
9757 (ashr_simd): New builtin handling DI mode.
9758 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
9759 (aarch64_sshr_simddi): New match pattern.
9760 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
9761 (vshrd_n_s64): Likewise.
9762 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
9764 2014-01-23 Nick Clifton <nickc@redhat.com>
9766 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
9767 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
9768 favour of mcu specific scripts.
9769 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
9772 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
9773 Alex Velenko <Alex.Velenko@arm.com>
9775 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
9776 (vaddv_s16): Likewise.
9777 (vaddv_s32): Likewise.
9778 (vaddv_u8): Likewise.
9779 (vaddv_u16): Likewise.
9780 (vaddv_u32): Likewise.
9781 (vaddvq_s8): Likewise.
9782 (vaddvq_s16): Likewise.
9783 (vaddvq_s32): Likewise.
9784 (vaddvq_s64): Likewise.
9785 (vaddvq_u8): Likewise.
9786 (vaddvq_u16): Likewise.
9787 (vaddvq_u32): Likewise.
9788 (vaddvq_u64): Likewise.
9789 (vaddv_f32): Likewise.
9790 (vaddvq_f32): Likewise.
9791 (vaddvq_f64): Likewise.
9792 (vmaxv_f32): Likewise.
9793 (vmaxv_s8): Likewise.
9794 (vmaxv_s16): Likewise.
9795 (vmaxv_s32): Likewise.
9796 (vmaxv_u8): Likewise.
9797 (vmaxv_u16): Likewise.
9798 (vmaxv_u32): Likewise.
9799 (vmaxvq_f32): Likewise.
9800 (vmaxvq_f64): Likewise.
9801 (vmaxvq_s8): Likewise.
9802 (vmaxvq_s16): Likewise.
9803 (vmaxvq_s32): Likewise.
9804 (vmaxvq_u8): Likewise.
9805 (vmaxvq_u16): Likewise.
9806 (vmaxvq_u32): Likewise.
9807 (vmaxnmv_f32): Likewise.
9808 (vmaxnmvq_f32): Likewise.
9809 (vmaxnmvq_f64): Likewise.
9810 (vminv_f32): Likewise.
9811 (vminv_s8): Likewise.
9812 (vminv_s16): Likewise.
9813 (vminv_s32): Likewise.
9814 (vminv_u8): Likewise.
9815 (vminv_u16): Likewise.
9816 (vminv_u32): Likewise.
9817 (vminvq_f32): Likewise.
9818 (vminvq_f64): Likewise.
9819 (vminvq_s8): Likewise.
9820 (vminvq_s16): Likewise.
9821 (vminvq_s32): Likewise.
9822 (vminvq_u8): Likewise.
9823 (vminvq_u16): Likewise.
9824 (vminvq_u32): Likewise.
9825 (vminnmv_f32): Likewise.
9826 (vminnmvq_f32): Likewise.
9827 (vminnmvq_f64): Likewise.
9829 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
9831 * config/aarch64/aarch64-simd.md
9832 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
9833 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
9834 (*aarch64_mul3_elt<mode>): Likewise.
9835 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
9836 (*aarch64_mul3_elt_to_64v2df): Likewise.
9837 (*aarch64_mla_elt<mode>): Likewise.
9838 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
9839 (*aarch64_mls_elt<mode>): Likewise.
9840 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
9841 (*aarch64_fma4_elt<mode>): Likewise.
9842 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
9843 (*aarch64_fma4_elt_to_64v2df): Likewise.
9844 (*aarch64_fnma4_elt<mode>): Likewise.
9845 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
9846 (*aarch64_fnma4_elt_to_64v2df): Likewise.
9847 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
9848 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
9849 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
9850 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
9851 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
9852 (aarch64_sqdmull_lane<mode>_internal): Likewise.
9853 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
9855 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
9857 * config/aarch64/aarch64-simd.md
9858 (aarch64_be_checked_get_lane<mode>): New define_expand.
9859 * config/aarch64/aarch64-simd-builtins.def
9860 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
9861 New builtin definition.
9862 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
9863 Use new safe be builtin.
9865 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
9867 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
9869 (aarch64_be_st1<mode>): Likewise.
9870 (aarch_ld1<VALL:mode>): Define_expand modified.
9871 (aarch_st1<VALL:mode>): Likewise.
9872 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
9873 (UNSPEC_ST1): Likewise.
9875 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
9877 * config/microblaze/microblaze.md: Add trap insn and attribute
9879 2014-01-23 Dodji Seketeli <dodji@redhat.com>
9881 PR preprocessor/58580
9882 * input.h (location_get_source_line): Take an additional line_size
9884 (void diagnostics_file_cache_fini): Declare new function.
9885 * input.c (struct fcache): New type.
9886 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
9887 New static constants.
9888 (diagnostic_file_cache_init, total_lines_num)
9889 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
9890 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
9891 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
9892 (get_next_line, read_next_line, goto_next_line, read_line_num):
9893 New static function definitions.
9894 (diagnostic_file_cache_fini): New function.
9895 (location_get_source_line): Take an additional output line_len
9896 parameter. Re-write using lookup_or_add_file_to_cache_tab and
9898 * diagnostic.c (diagnostic_finish): Call
9899 diagnostic_file_cache_fini.
9900 (adjust_line): Take an additional input parameter for the length
9901 of the line, rather than calculating it with strlen.
9902 (diagnostic_show_locus): Adjust the use of
9903 location_get_source_line and adjust_line with respect to their new
9904 signature. While displaying a line now, do not stop at the first
9905 null byte. Rather, display the zero byte as a space and keep
9906 going until we reach the size of the line.
9907 * Makefile.in: Add vec.o to OBJS-libcommon
9909 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
9910 Ilya Tocar <ilya.tocar@intel.com>
9912 * config/i386/avx512fintrin.h (_mm512_kmov): New.
9913 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
9914 (__builtin_ia32_kmov16): Ditto.
9915 * config/i386/i386.md (UNSPEC_KMOV): New.
9918 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
9920 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
9921 (_mm512_storeu_si512): Ditto.
9923 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
9926 * rtl.h (get_referenced_operands): Declare.
9927 * recog.c (get_referenced_operands): New function.
9928 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
9929 operands have been referenced when recording LO_SUM references.
9931 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
9933 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
9935 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
9937 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
9938 Enable for generic and recent AMD targets.
9940 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
9942 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
9943 ARG_SIZE note when adjustment was eliminated.
9945 2014-01-22 Jeff Law <law@redhat.com>
9947 PR tree-optimization/59597
9948 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
9949 in file. Accept new argument REGISTERING and use it to modify
9950 dump output appropriately.
9951 (register_jump_thread): Corresponding changes.
9952 (mark_threaded_blocks): Reinstate code to cancel unprofitable
9953 thread paths involving joiner blocks. Add code to dump cancelled
9954 jump threading paths.
9956 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
9958 PR rtl-optimization/59477
9959 * lra-constraints.c (inherit_in_ebb): Process call for living hard
9960 regs. Update reloads_num and potential_reload_hard_regs for all insns.
9962 2014-01-22 Tom Tromey <tromey@redhat.com>
9964 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
9966 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
9968 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
9970 PR rtl-optimization/59896
9971 * lra-constraints.c (process_alt_operands): Check unused note for
9972 matched operands of insn with no output reloads.
9974 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
9976 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
9977 (mips_move_from_gpr_cost): Likewise.
9979 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
9981 PR rtl-optimization/59858
9982 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
9983 ira_class_hard_regs_num.
9984 (process_alt_operands): Increase reject for dying matched operand.
9986 2014-01-21 Jakub Jelinek <jakub@redhat.com>
9989 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
9990 smaller than size, perform several stores or loads and stores
9991 at dst + count - size to store or copy all of size bytes, rather
9992 than just last modesize bytes.
9994 2014-01-20 DJ Delorie <dj@redhat.com>
9996 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
9997 that CLOBBERs are REGs before propogating their values.
9999 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
10001 PR middle-end/59789
10002 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
10003 (cgraph_inline_failed_type): New function.
10004 * cgraph.h (DEFCIFCODE): Add type.
10005 (cgraph_inline_failed_type_t): New enum.
10006 (cgraph_inline_failed_type): New prototype.
10007 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
10008 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
10009 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
10010 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
10011 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
10012 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
10013 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
10014 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
10015 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
10016 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
10017 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
10018 OPTIMIZATION_MISMATCH.
10019 * tree-inline.c (expand_call_inline): Emit errors during
10020 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
10022 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
10025 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
10026 mode attribute in insn output.
10028 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
10030 * output.h (output_constant): Delete.
10031 * varasm.c (output_constant): Make private.
10033 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
10035 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
10037 2014-01-20 Jakub Jelinek <jakub@redhat.com>
10039 PR middle-end/59860
10040 * tree.h (fold_builtin_strcat): New prototype.
10041 * builtins.c (fold_builtin_strcat): No longer static. Add len
10042 argument, if non-NULL, don't call c_strlen. Optimize
10043 directly into __builtin_memcpy instead of __builtin_strcpy.
10044 (fold_builtin_2): Adjust fold_builtin_strcat caller.
10045 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
10047 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
10049 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
10050 for SImode_address_operand operands, having only a REG argument.
10052 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
10054 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
10055 loader name using mbig-endian.
10056 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
10058 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
10060 * doc/invoke.texi (-march): Clarify documentation for AArch64.
10061 (-mtune): Likewise.
10064 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
10066 * config/aarch64/aarch64-protos.h
10067 (aarch64_cannot_change_mode_class_ptr): Declare.
10068 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
10069 aarch64_cannot_change_mode_class_ptr): New.
10070 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
10071 backend hook aarch64_cannot_change_mode_class.
10073 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
10075 * common/config/aarch64/aarch64-common.c
10076 (aarch64_handle_option): Don't handle any option order logic here.
10077 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
10078 selected_cpu, warn on architecture version mismatch.
10079 (aarch64_override_options): Fix parsing order for option strings.
10081 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10082 Iain Sandoe <iain@codesourcery.com>
10085 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
10086 warning. Amend comment to reflect current functionality.
10088 2014-01-20 Richard Biener <rguenther@suse.de>
10090 PR middle-end/59860
10091 * builtins.c (fold_builtin_strcat): Remove case better handled
10092 by tree-ssa-strlen.c.
10094 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
10096 * config/aarch64/aarch64.opt
10097 (mcpu, march, mtune): Make case-insensitive.
10099 2014-01-20 Jakub Jelinek <jakub@redhat.com>
10102 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
10103 if operands[1] is a REG or ZERO_EXTEND of a REG.
10105 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
10107 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
10109 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
10111 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
10112 long non-pic millicode calls.
10114 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10116 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
10118 2014-01-19 Kito Cheng <kito@0xlab.org>
10120 * builtins.c (expand_movstr): Check movstr expand done or fail.
10122 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
10123 H.J. Lu <hongjiu.lu@intel.com>
10126 * config/i386/i386.md (*lea<mode>): Zero-extend return register
10127 to DImode for zero-extended addresses.
10129 2014-01-19 Jakub Jelinek <jakub@redhat.com>
10131 PR rtl-optimization/57763
10132 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
10133 on the new indirect jump_insn and increment LABEL_NUSES (label).
10135 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
10139 * config.gcc (x86_archs): New variable.
10140 (x86_64_archs): Likewise.
10141 (x86_cpus): Likewise.
10142 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
10143 --with-arch/--with-cpu= options.
10144 Support --with-arch=/--with-cpu={nehalem,westmere,
10145 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
10147 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
10149 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
10150 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
10152 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
10154 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
10156 2014-01-18 Jakub Jelinek <jakub@redhat.com>
10159 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
10160 clear cpp_get_options (parse_in)->warn_unused_macros for
10161 ix86_target_macros_internal with cpp_define.
10163 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
10165 * jump.c (delete_related_insns): Keep (use (insn))s.
10166 * reorg.c (redundant_insn): Check for barriers too.
10168 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
10170 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
10172 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
10174 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
10175 call to $$dyncall when TARGET_LONG_CALLS is true.
10177 2014-01-17 Jeff Law <law@redhat.com>
10179 * ree.c (combine_set_extension): Temporarily disable test for
10180 changing number of hard registers.
10182 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
10184 PR middle-end/58125
10185 * ipa-inline-analysis.c (inline_free_summary):
10186 Do not free summary of aliases.
10188 2014-01-17 Jakub Jelinek <jakub@redhat.com>
10190 PR middle-end/59706
10191 * gimplify.c (gimplify_expr): Use create_tmp_var
10192 instead of create_tmp_var_raw. If cond doesn't have
10193 integral type, don't add the IFN_ANNOTATE builtin at all.
10195 2014-01-17 Martin Jambor <mjambor@suse.cz>
10198 * ipa-cp.c (prev_edge_clone): New variable.
10199 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
10200 Also resize prev_edge_clone vector.
10201 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
10202 (ipcp_edge_removal_hook): New function.
10203 (ipcp_driver): Register ipcp_edge_removal_hook.
10205 2014-01-17 Andrew Pinski <apinski@cavium.com>
10206 Steve Ellcey <sellcey@mips.com>
10209 * config/mips/mips.c (mips_print_operand): Check operand mode instead
10212 2014-01-17 Jeff Law <law@redhat.com>
10214 PR middle-end/57904
10215 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
10216 so that pass_ccp runs first.
10218 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
10220 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
10221 (ix86_adjust_cost): Use !TARGET_XXX.
10222 (do_reorder_for_imul): Likewise.
10223 (swap_top_of_ready_list): Likewise.
10224 (ix86_sched_reorder): Likewise.
10226 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
10228 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
10229 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
10230 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
10231 (intel_memset): New. Duplicate slm_memset.
10232 (intel_cost): New. Duplicate slm_cost.
10233 (m_INTEL): New macro.
10234 (processor_target_table): Add "intel".
10235 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
10236 with PROCESSOR_INTEL for "intel".
10237 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
10238 PROCESSOR_SILVERMONT.
10239 (ix86_issue_rate): Likewise.
10240 (ix86_adjust_cost): Likewise.
10241 (ia32_multipass_dfa_lookahead): Likewise.
10242 (swap_top_of_ready_list): Likewise.
10243 (ix86_sched_reorder): Likewise.
10244 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
10245 instead of TARGET_OPT_AGU.
10246 * config/i386/i386.h (TARGET_INTEL): New.
10247 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
10248 (processor_type): Add PROCESSOR_INTEL.
10249 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
10250 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
10252 2014-01-17 Marek Polacek <polacek@redhat.com>
10255 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
10258 2014-01-17 Richard Biener <rguenther@suse.de>
10260 PR tree-optimization/46590
10261 * opts.c (default_options_table): Add entries for
10262 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
10263 all enabled at -O1 but not for -Og.
10264 * common.opt (fbranch-count-reg): Remove Init(1).
10265 (fmove-loop-invariants): Likewise.
10266 (ftree-pta): Likewise.
10268 2014-01-17 Jakub Jelinek <jakub@redhat.com>
10270 * config/i386/i386.c (ix86_data_alignment): For compatibility with
10271 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
10272 decls to at least the GCC 4.8 used alignments.
10275 * tree-nested.c (convert_nonlocal_reference_stmt,
10276 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
10277 of GIMPLE_BIND stmts, adjust associated decls.
10279 2014-01-17 Richard Biener <rguenther@suse.de>
10281 PR tree-optimization/46590
10282 * vec.h (vec<>::bseach): New member function implementing
10283 binary search according to C89 bsearch.
10284 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
10285 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
10286 bitmap pointer again. Make accesses_in_loop a flat array.
10287 (mem_ref_obstack): New global.
10288 (outermost_indep_loop): Adjust for mem_ref->stored changes.
10289 (mark_ref_stored): Likewise.
10290 (ref_indep_loop_p_2): Likewise.
10291 (set_ref_stored_in_loop): New helper function.
10292 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
10293 (memref_free): Adjust.
10294 (record_mem_ref_loc): Simplify.
10295 (gather_mem_refs_stmt): Adjust.
10296 (sort_locs_in_loop_postorder_cmp): New function.
10297 (analyze_memory_references): Sort accesses_in_loop after
10298 loop postorder number.
10299 (find_ref_loc_in_loop_cmp): New function.
10300 (for_all_locs_in_loop): Find relevant cluster of locs in
10301 accesses_in_loop and iterate without recursion.
10302 (execute_sm): Avoid uninit warning.
10303 (struct ref_always_accessed): Simplify.
10304 (ref_always_accessed::operator ()): Likewise.
10305 (ref_always_accessed_p): Likewise.
10306 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
10307 loop postorder numbers here.
10308 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
10311 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
10314 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
10315 on decls for which assemble_alias has been called.
10317 2014-01-17 Nick Clifton <nickc@redhat.com>
10319 * config/msp430/msp430.opt: (mcpu): New option.
10320 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
10321 (msp430_option_override): Parse target_cpu. If the MCU name
10322 matches a generic string, clear target_mcu.
10323 (msp430_attr): Allow numeric interrupt values up to 63.
10324 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
10325 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
10327 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
10329 * config/msp430/msp430.md (popm): Use %J rather than %I.
10330 (addsi3): Use msp430_nonimmediate_operand for operand 2.
10331 (addhi_cy_i): Use immediate_operand for operand 2.
10332 * doc/invoke.texi: Document -mcpu option.
10334 2014-01-17 Richard Biener <rguenther@suse.de>
10336 PR rtl-optimization/38518
10337 * df.h (df_analyze_loop): Declare.
10338 * df-core.c: Include cfgloop.h.
10339 (df_analyze_1): Split out main part of df_analyze.
10340 (df_analyze): Adjust.
10341 (loop_inverted_post_order_compute): New function.
10342 (loop_post_order_compute): Likewise.
10343 (df_analyze_loop): New function avoiding whole-function
10344 postorder computes.
10345 * loop-invariant.c (find_defs): Use df_analyze_loop.
10346 (find_invariants): Adjust.
10347 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
10349 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
10351 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
10352 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
10354 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
10356 * ipa-ref.c (ipa_remove_stmt_references): Fix references
10357 traversal when removing references.
10359 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
10362 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
10364 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
10366 PR middle-end/56791
10367 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
10368 pushing a reload for an autoinc when we had previously reloaded an
10369 inner part of the address.
10371 2014-01-16 Jakub Jelinek <jakub@redhat.com>
10373 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
10375 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
10376 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
10377 when not giving up or versioning for alias only because of
10379 (vect_analyze_data_ref_dependences): Set to true.
10380 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
10382 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
10383 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
10386 PR middle-end/58344
10387 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
10390 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
10391 operand 0 predicate for gathers, use a new pseudo as subtarget.
10393 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
10395 PR middle-end/59609
10396 * lra-constraints.c (process_alt_operands): Add printing debug info.
10397 Check absence of input/output reloads for matched operands too.
10399 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
10401 PR rtl-optimization/59835
10402 * ira.c (ira_init_register_move_cost): Increase cost for
10405 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
10407 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
10409 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
10412 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
10413 non-register objects. Use gen_(high/low)part more consistently.
10416 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
10419 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
10420 endian support, remove tests for WORDS_BIG_ENDIAN.
10421 (p8_mfvsrd_3_<mode>): Likewise.
10422 (reload_gpr_from_vsx<mode>): Likewise.
10423 (reload_gpr_from_vsxsf): Likewise.
10424 (p8_mfvsrd_4_disf): Likewise.
10426 2014-01-16 Richard Biener <rguenther@suse.de>
10428 PR rtl-optimization/46590
10429 * lcm.c (compute_antinout_edge): Use postorder iteration.
10430 (compute_laterin): Use inverted postorder iteration.
10432 2014-01-16 Nick Clifton <nickc@redhat.com>
10434 PR middle-end/28865
10435 * varasm.c (output_constant): Return the number of bytes actually
10437 (output_constructor_array_range): Update the field size with the
10438 number of bytes emitted by output_constant.
10439 (output_constructor_regular_field): Likewise. Also do not
10440 complain if the total number of bytes emitted is now greater
10441 than the expected fieldpos.
10442 * output.h (output_constant): Update prototype and descriptive comment.
10444 2014-01-16 Marek Polacek <polacek@redhat.com>
10446 PR middle-end/59827
10447 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
10448 it is error_mark_node.
10450 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
10452 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
10453 VALID_AVX256_REG_OR_OI_MODE.
10455 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
10457 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
10458 current procedure should be profiled.
10460 2014-01-15 Andrew Pinski <apinski@cavium.com>
10462 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
10463 of moving from/to the STACK_REG register class.
10465 2014-01-15 Richard Henderson <rth@redhat.com>
10468 * reginfo.c (global_regs_decl): Globalize.
10469 * rtl.h (global_regs_decl): Declare.
10470 * ira.c (do_reload): Diagnose frame_pointer_needed and it
10471 reserved via global_regs.
10473 2014-01-15 Teresa Johnson <tejohnson@google.com>
10475 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
10477 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10479 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
10480 and vmulosh rather than call gen_vec_widen_smult_*.
10481 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
10482 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
10483 (vec_widen_smult_even_v16qi): Likewise.
10484 (vec_widen_umult_even_v8hi): Likewise.
10485 (vec_widen_smult_even_v8hi): Likewise.
10486 (vec_widen_umult_odd_v16qi): Likewise.
10487 (vec_widen_smult_odd_v16qi): Likewise.
10488 (vec_widen_umult_odd_v8hi): Likewise.
10489 (vec_widen_smult_odd_v8hi): Likewise.
10490 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
10491 vmuloub rather than call gen_vec_widen_umult_*.
10492 (vec_widen_umult_lo_v16qi): Likewise.
10493 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
10494 vmulosb rather than call gen_vec_widen_smult_*.
10495 (vec_widen_smult_lo_v16qi): Likewise.
10496 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
10497 rather than call gen_vec_widen_umult_*.
10498 (vec_widen_umult_lo_v8hi): Likewise.
10499 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
10500 rather than call gen_vec_widen_smult_*.
10501 (vec_widen_smult_lo_v8hi): Likewise.
10503 2014-01-15 Jeff Law <law@redhat.com>
10505 PR tree-optimization/59747
10506 * ree.c (find_and_remove_re): Properly handle case where a second
10507 eliminated extension requires widening a copy created for elimination
10508 of a prior extension.
10509 (combine_set_extension): Ensure that the number of hard regs needed
10510 for a destination register does not change when we widen it.
10512 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
10514 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
10515 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
10516 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
10517 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
10518 (avr-*-rtems*): Likewise.
10519 (bfin*-rtems*): Likewise.
10520 (moxie-*-rtems*): Likewise.
10521 (h8300-*-rtems*): Likewise.
10522 (i[34567]86-*-rtems*): Likewise.
10523 (lm32-*-rtems*): Likewise.
10524 (m32r-*-rtems*): Likewise.
10525 (m68k-*-rtems*): Likewise.
10526 (microblaze*-*-rtems*): Likewise.
10527 (mips*-*-rtems*): Likewise.
10528 (powerpc-*-rtems*): Likewise.
10529 (sh-*-rtems*): Likewise.
10530 (sparc-*-rtems*): Likewise.
10531 (sparc64-*-rtems*): Likewise.
10532 (v850-*-rtems*): Likewise.
10533 (m32c-*-rtems*): Likewise.
10535 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
10537 PR rtl-optimization/59511
10538 * ira.c (ira_init_register_move_cost): Use memory costs for some
10539 cases of register move cost calculations.
10540 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
10541 instead of BB frequency.
10542 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
10543 * lra-assigns.c (find_hard_regno_for): Ditto.
10545 2014-01-15 Richard Biener <rguenther@suse.de>
10547 PR tree-optimization/59822
10548 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
10549 (vectorizable_load): Use it to hoist defs of uses of invariant
10550 loads out of the loop.
10552 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
10553 Kugan Vivekanandarajah <kuganv@linaro.org>
10556 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
10559 2014-01-15 Richard Biener <rguenther@suse.de>
10561 PR rtl-optimization/59802
10562 * lcm.c (compute_available): Use inverted postorder to seed
10563 the initial worklist.
10565 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10568 * config/s390/s390.c (s390_preferred_reload_class): Don't return
10569 ADDR_REGS for invalid symrefs in non-PIC code.
10571 2014-01-15 Jakub Jelinek <jakub@redhat.com>
10574 * builtins.c (determine_block_size): Initialize *probable_max_size
10575 even if len_rtx is CONST_INT.
10577 2014-01-14 Andrew Pinski <apinski@cavium.com>
10579 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
10580 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
10581 (cortexa53_tunings): Likewise.
10582 (aarch64_sched_issue_rate): New function.
10583 (TARGET_SCHED_ISSUE_RATE): Define.
10585 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
10587 * ira-costs.c (find_costs_and_classes): Add missed
10588 ira_init_register_move_cost_if_necessary.
10590 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
10593 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
10595 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
10598 * config/i386/i386.c (type_natural_mode): Add a bool parameter
10599 to indicate if type is used for function return value. Warn ABI
10600 change if the vector mode isn't available for function return value.
10601 (ix86_function_arg_advance): Pass false to type_natural_mode.
10602 (ix86_function_arg): Likewise.
10603 (ix86_gimplify_va_arg): Likewise.
10604 (function_arg_32): Don't warn ABI change.
10605 (ix86_function_value): Pass true to type_natural_mode.
10606 (ix86_return_in_memory): Likewise.
10607 (ix86_struct_value_rtx): Removed.
10608 (TARGET_STRUCT_VALUE_RTX): Likewise.
10610 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10612 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
10613 converting a conditional jump into a conditional return.
10615 2014-01-14 Richard Biener <rguenther@suse.de>
10617 PR tree-optimization/58921
10618 PR tree-optimization/59006
10619 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
10620 hoisting invariant stmts.
10621 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
10622 invariant loads on the preheader edge if possible.
10624 2014-01-14 Joey Ye <joey.ye@arm.com>
10626 * doc/plugin.texi (Building GCC plugins): Update to C++.
10628 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
10630 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
10631 (_mm_rcp28_round_ss): Ditto.
10632 (_mm_rsqrt28_round_sd): Ditto.
10633 (_mm_rsqrt28_round_ss): Ditto.
10634 (_mm_rcp28_sd): Ditto.
10635 (_mm_rcp28_ss): Ditto.
10636 (_mm_rsqrt28_sd): Ditto.
10637 (_mm_rsqrt28_ss): Ditto.
10638 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
10639 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
10640 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
10641 (IX86_BUILTIN_RCP28SD): Ditto.
10642 (IX86_BUILTIN_RCP28SS): Ditto.
10643 (IX86_BUILTIN_RSQRT28SD): Ditto.
10644 (IX86_BUILTIN_RSQRT28SS): Ditto.
10645 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
10646 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
10647 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
10648 (ix86_expand_special_args_builtin): Expand new FTYPE.
10649 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
10650 (srcp14<mode>): Make insn unary.
10651 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
10652 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
10653 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
10654 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
10655 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
10656 Fix rounding: make it SAE only.
10657 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
10659 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
10661 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
10662 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
10663 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
10664 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
10665 (round_saeonly_mask_scalar_operand4): Ditto.
10666 (round_saeonly_mask_scalar_op3): Ditto.
10667 (round_saeonly_mask_scalar_op4): Ditto.
10669 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10671 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
10672 Implement -maltivec=be for vec_insert and vec_extract.
10674 2014-01-10 DJ Delorie <dj@redhat.com>
10676 * config/msp430/msp430.md (call_internal): Don't allow memory
10677 references with SP as the base register.
10678 (call_value_internal): Likewise.
10679 * config/msp430/constraints.md (Yc): New. For memory references
10680 that don't use SP as a base register.
10682 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
10683 "an integer without a # prefix"
10684 * config/msp430/msp430.md (epilogue_helper): Use it.
10686 2014-01-13 Jakub Jelinek <jakub@redhat.com>
10689 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
10690 AVX512F gather builtins.
10691 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
10692 on gather decls with INTEGER_TYPE masktype.
10693 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
10694 directly into the builtin rather than hoisting it before loop.
10696 PR tree-optimization/59387
10697 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
10698 (scev_const_prop): If folded_casts and type has undefined overflow,
10699 use force_gimple_operand instead of force_gimple_operand_gsi and
10700 for each added stmt if it is assign with
10701 arith_code_with_undefined_signed_overflow, call
10702 rewrite_to_defined_overflow.
10703 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
10704 gimple-fold.h instead.
10705 (arith_code_with_undefined_signed_overflow,
10706 rewrite_to_defined_overflow): Moved to ...
10707 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
10708 rewrite_to_defined_overflow): ... here. No longer static.
10709 Include gimplify-me.h.
10710 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
10711 rewrite_to_defined_overflow): New prototypes.
10713 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10715 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
10717 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
10719 * builtins.c (get_object_alignment_2): Minor tweak.
10720 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
10722 2014-01-13 Christian Bruel <christian.bruel@st.com>
10724 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
10725 optimized non constant lengths.
10727 2014-01-13 Jakub Jelinek <jakub@redhat.com>
10730 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
10731 load as __atomic_load_N if possible.
10733 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
10735 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
10737 (rs6000_expand_builtin): Adjust call.
10739 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
10742 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
10743 * config/rs6000/rs6000.c: Include target-globals.h.
10744 (rs6000_set_current_function): Instead of doing target_reinit
10745 unconditionally, use save_target_globals_default_opts and
10746 restore_target_globals.
10748 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
10750 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
10751 (rs6000_expand_builtin): Handle mffs and mtfsf.
10752 (rs6000_init_builtins): Define mffs and mtfsf.
10753 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
10754 (rs6000_mffs): New pattern.
10755 (rs6000_mtfsf): New pattern.
10757 2014-01-11 Bin Cheng <bin.cheng@arm.com>
10759 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
10760 Start narrowing with START. Apply candidate-use pair
10761 and check overall cost in narrowing.
10762 (iv_ca_prune): Pass new argument.
10764 2014-01-10 Jeff Law <law@redhat.com>
10766 PR middle-end/59743
10767 * ree.c (combine_reaching_defs): Ensure the defining statement
10768 occurs before the extension when optimizing extensions with
10769 different source and destination hard registers.
10771 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
10774 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
10775 vtables into the type inheritance graph.
10777 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10779 PR rtl-optimization/59754
10780 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
10781 modes in the REGNO != REGNO case.
10783 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10785 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
10787 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10789 PR tree-optimization/59745
10790 * tree-predcom.c (tree_predictive_commoning_loop): Call
10791 free_affine_expand_cache if giving up because components is NULL.
10793 * target-globals.c (save_target_globals): Allocate < 4KB structs using
10794 GC in payload of target_globals struct instead of allocating them on
10795 the heap and the larger structs separately using GC.
10796 * target-globals.h (struct target_globals): Make regs, hard_regs,
10797 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
10798 of GTY((skip)) and change type to void *.
10799 (reset_target_globals): Cast loads from those fields to corresponding
10802 2014-01-10 Steve Ellcey <sellcey@mips.com>
10805 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
10806 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
10807 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
10809 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
10812 * aarch64-modes.def (CC_Zmode): New flags mode.
10813 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
10814 represents an equality.
10815 (aarch64_get_condition_code): Handle CC_Zmode.
10816 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
10818 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10820 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
10821 extraction in good case.
10823 2014-01-10 Richard Biener <rguenther@suse.de>
10825 PR tree-optimization/59374
10826 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
10827 checking after SLP discovery. Mark stmts not participating
10828 in any SLP instance properly.
10830 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10832 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
10833 when handling a SET rtx.
10835 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10837 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
10838 (cortex-a57): Likewise.
10839 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
10841 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10843 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
10844 non-iwmmxt builtins.
10846 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
10850 * ipa-devirt.c record_target_from_binfo): Take as argument
10851 stack of binfos and lookup matching one for virtual inheritance.
10852 (possible_polymorphic_call_targets_1): Update.
10854 2014-01-10 Huacai Chen <chenhc@lemote.com>
10856 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
10857 kernel strings for Loongson-2E/2F/3A.
10859 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10861 PR middle-end/59670
10862 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
10863 is_gimple_call before calling gimple_call_internal_p.
10865 2014-01-09 Steve Ellcey <sellcey@mips.com>
10867 * Makefile.in (TREE_FLOW_H): Remove.
10868 (TREE_SSA_H): Add file names from tree-flow.h.
10869 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
10870 * tree.h: Remove tree-flow.h reference.
10871 * hash-table.h: Remove tree-flow.h reference.
10872 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
10873 reference with tree-ssa-loop.h.
10875 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10877 * doc/invoke.texi: Add -maltivec={be,le} options, and document
10878 default element-order behavior for -maltivec.
10879 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
10880 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
10881 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
10882 when targeting big endian, at least for now.
10883 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
10885 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10887 PR middle-end/47735
10888 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
10889 var satisfies use_register_for_decl, just take into account type
10890 alignment, rather than decl alignment.
10892 PR tree-optimization/59622
10893 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
10894 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
10895 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
10896 Don't devirtualize for inplace at all. For targets.length () == 1,
10897 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
10899 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
10901 * config/i386/i386.md (cpu): Remove the unused btver1.
10903 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
10905 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
10907 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10910 * tree-core.h (struct target_globals): New forward declaration.
10911 (struct tree_target_option): Add globals field.
10912 * tree.h (TREE_TARGET_GLOBALS): Define.
10913 (prepare_target_option_nodes_for_pch): New prototype.
10914 * target-globals.h (struct target_globals): Define even if
10915 !SWITCHABLE_TARGET.
10916 * tree.c (prepare_target_option_node_for_pch,
10917 prepare_target_option_nodes_for_pch): New functions.
10918 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
10919 * config/i386/i386.c: Include target-globals.h.
10920 (ix86_set_current_function): Instead of doing target_reinit
10921 unconditionally, use save_target_globals_default_opts and
10922 restore_target_globals.
10924 2014-01-09 Richard Biener <rguenther@suse.de>
10926 PR tree-optimization/59715
10927 * tree-cfg.h (split_critical_edges): Declare.
10928 * tree-cfg.c (split_critical_edges): Export.
10929 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
10931 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
10933 * cfgexpand.c (expand_stack_vars): Optionally disable
10934 asan stack protection.
10935 (expand_used_vars): Likewise.
10936 (partition_stack_vars): Likewise.
10937 * asan.c (asan_emit_stack_protection): Optionally disable
10938 after return stack usage.
10939 (instrument_derefs): Optionally disable memory access instrumentation.
10940 (instrument_builtin_call): Likewise.
10941 (instrument_strlen_call): Likewise.
10942 (asan_protect_global): Optionally disable global variables protection.
10943 * doc/invoke.texi: Added doc for new options.
10944 * params.def: Added new options.
10945 * params.h: Likewise.
10947 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10949 PR rtl-optimization/59724
10950 * ifcvt.c (cond_exec_process_if_block): Don't call
10951 flow_find_head_matching_sequence with 0 longest_match.
10952 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
10953 non-active insns if !stop_after.
10954 (try_head_merge_bb): Revert 2014-01-07 changes.
10956 2014-01-08 Jeff Law <law@redhat.com>
10958 * ree.c (get_sub_rtx): New function, extracted from...
10959 (merge_def_and_ext): Here.
10960 (combine_reaching_defs): Use get_sub_rtx.
10962 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
10964 * cgraph.h (varpool_variable_node): Do not choke on null node.
10966 2014-01-08 Catherine Moore <clm@codesourcery.com>
10968 * config/mips/mips.md (simple_return): Attempt to use JRC
10970 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
10972 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
10974 PR rtl-optimization/59137
10975 * reorg.c (steal_delay_list_from_target): Call update_block for
10977 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
10979 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10981 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
10982 two duplicate entries.
10984 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
10987 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
10989 * config/mips/mips.c (mips_truncated_op_cost): New function.
10990 (mips_rtx_costs): Adjust test for BADDU.
10991 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
10993 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
10995 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
10996 (*baddu_si): ...this new pattern.
10998 2014-01-08 Jakub Jelinek <jakub@redhat.com>
11001 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
11003 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
11005 PR middle-end/57748
11006 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
11008 (expand_expr, expand_normal): Adjust.
11009 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
11010 inner_reference_p. Use inner_reference_p to expand inner references.
11011 (store_expr): Adjust.
11012 * cfgexpand.c (expand_call_stmt): Adjust.
11014 2014-01-08 Rong Xu <xur@google.com>
11016 * gcov-io.c (gcov_var): Move from gcov-io.h.
11017 (gcov_position): Ditto.
11018 (gcov_is_error): Ditto.
11019 (gcov_rewrite): Ditto.
11020 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
11021 only part to libgcc/libgcov.h.
11023 2014-01-08 Marek Polacek <polacek@redhat.com>
11025 PR middle-end/59669
11026 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
11028 2014-01-08 Marek Polacek <polacek@redhat.com>
11031 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
11033 2014-01-08 Jakub Jelinek <jakub@redhat.com>
11035 PR rtl-optimization/59649
11036 * stor-layout.c (get_mode_bounds): For BImode return
11037 0 and STORE_FLAG_VALUE.
11039 2014-01-08 Richard Biener <rguenther@suse.de>
11041 PR middle-end/59630
11042 * gimple.h (is_gimple_builtin_call): Remove.
11043 (gimple_builtin_call_types_compatible_p): New.
11044 (gimple_call_builtin_p): New overload.
11045 * gimple.c (is_gimple_builtin_call): Remove.
11046 (validate_call): Rename to ...
11047 (gimple_builtin_call_types_compatible_p): ... this and export. Also
11048 check return types.
11049 (validate_type): New static function.
11050 (gimple_call_builtin_p): New overload and adjust.
11051 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
11052 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
11053 (gimple_fold_stmt_to_constant_1): Likewise.
11054 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
11056 2014-01-08 Richard Biener <rguenther@suse.de>
11058 PR middle-end/59471
11059 * gimplify.c (gimplify_expr): Gimplify register-register type
11060 VIEW_CONVERT_EXPRs to separate stmts.
11062 2014-01-07 Jeff Law <law@redhat.com>
11064 PR middle-end/53623
11065 * ree.c (combine_set_extension): Handle case where source
11066 and destination registers in an extension insn are different.
11067 (combine_reaching_defs): Allow source and destination registers
11068 in extension to be different under limited circumstances.
11069 (add_removable_extension): Remove restriction that the
11070 source and destination registers in the extension are the same.
11071 (find_and_remove_re): Emit a copy from the extension's
11072 destination to its source after the defining insn if
11073 the source and destination registers are different.
11075 PR middle-end/59285
11076 * ifcvt.c (merge_if_block): If we are merging a block with more than
11077 one successor with a block with no successors, remove any BARRIER
11078 after the second block.
11080 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
11082 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
11084 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
11087 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
11088 for 14-bit register offsets when INT14_OK_STRICT is false.
11090 2014-01-07 Roland Stigge <stigge@antcom.de>
11091 Michael Meissner <meissner@linux.vnet.ibm.com>
11094 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
11095 Only check TFmode for SPE constants. Don't check TImode or TDmode.
11097 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
11099 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
11102 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
11104 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
11105 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
11106 rtx is const0_rtx or not.
11108 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
11111 * target-globals.c (save_target_globals): Remove this_fn_optab
11113 * toplev.c: Include optabs.h.
11114 (target_reinit): Temporarily restore the global options if another
11115 set of options are in force.
11117 2014-01-07 Jakub Jelinek <jakub@redhat.com>
11119 PR rtl-optimization/58668
11120 * cfgcleanup.c (flow_find_cross_jump): Don't count
11121 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
11122 to determine what is counted.
11123 (flow_find_head_matching_sequence): Use active_insn_p to determine
11125 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
11127 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
11128 determine what is counted.
11130 PR tree-optimization/59643
11131 * tree-predcom.c (split_data_refs_to_components): If one dr is
11132 read and one write, determine_offset fails and the write isn't
11133 in the bad component, just put the read into the bad component.
11135 2014-01-07 Mike Stump <mikestump@comcast.net>
11136 Jakub Jelinek <jakub@redhat.com>
11139 * tree-core.h (struct tree_optimization_option): Change optabs
11140 type from unsigned char * to void *.
11141 * optabs.c (init_tree_optimization_optabs): Adjust
11142 TREE_OPTIMIZATION_OPTABS initialization.
11144 2014-01-06 Jakub Jelinek <jakub@redhat.com>
11147 * config/i386/i386.h (struct machine_function): Add
11148 no_drap_save_restore field.
11149 * config/i386/i386.c (ix86_save_reg): Use
11150 !cfun->machine->no_drap_save_restore instead of
11151 crtl->stack_realign_needed.
11152 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
11153 this function clears frame_pointer_needed. Set
11154 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
11155 and DRAP reg is needed.
11157 2014-01-06 Marek Polacek <polacek@redhat.com>
11160 * doc/implement-c.texi: Mention that other integer types are
11161 permitted as bit-field types in strictly conforming mode.
11163 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
11165 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
11166 is newly allocated.
11168 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
11170 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
11172 2014-01-06 Martin Jambor <mjambor@suse.cz>
11175 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
11177 * ipa-prop.c (ipa_print_node_params): Fix indentation.
11179 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
11183 * var-tracking.c (add_stores): Preserve the value of the source even if
11184 we don't record the store.
11186 2014-01-06 Terry Guo <terry.guo@arm.com>
11188 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
11190 2014-01-05 Iain Sandoe <iain@codesourcery.com>
11193 * config/darwin.c (darwin_function_section): Adjust return values to
11194 correspond to optimisation changes made in r206070.
11196 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
11198 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
11199 from prefetch_block tune setting.
11200 (nocona_cost): Correct size of prefetch block to 64.
11202 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
11204 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
11205 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
11206 used to save the static chain register in the computation of the offset
11207 from which the FP registers need to be restored.
11209 2014-01-04 Jakub Jelinek <jakub@redhat.com>
11211 PR tree-optimization/59519
11212 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
11213 ICE if get_current_def (current_new_name) is already non-NULL, as long
11214 as it is a phi result of some other phi in *new_exit_bb that has
11217 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
11218 or vmovdqu* for misaligned_operand.
11219 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
11220 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
11221 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
11222 aligned_mem for AVX512F masked aligned load and store builtins and for
11223 non-temporal moves.
11225 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
11227 PR tree-optimization/59651
11228 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
11229 Address range for negative step should be added by TYPE_SIZE_UNIT.
11231 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
11233 * config/m68k/m68k.c (handle_move_double): Handle pushes with
11234 overlapping registers also for registers other than the stack pointer.
11236 2014-01-03 Marek Polacek <polacek@redhat.com>
11239 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
11242 2014-01-03 Jakub Jelinek <jakub@redhat.com>
11245 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
11248 * config/i386/i386.md (MODE_SIZE): New mode attribute.
11249 (push splitter): Use <P:MODE_SIZE> instead of
11250 GET_MODE_SIZE (<P:MODE>mode).
11251 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
11252 (mov -1, reg peephole2): Likewise.
11253 * config/i386/sse.md (*mov<mode>_internal,
11254 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
11255 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
11256 *<code><mode>3, *andnot<mode>3<mask_name>,
11257 <mask_codefor><code><mode>3<mask_name>): Likewise.
11258 * config/i386/subst.md (mask_mode512bit_condition,
11259 sd_mask_mode512bit_condition): Likewise.
11261 2014-01-02 Xinliang David Li <davidxl@google.com>
11263 PR tree-optimization/59303
11264 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
11265 (dump_predicates): Better output format.
11266 (pred_equal_p): New function.
11267 (is_neq_relop_p): Ditto.
11268 (is_neq_zero_form_p): Ditto.
11269 (pred_expr_equal_p): Ditto.
11270 (pred_neg_p): Ditto.
11271 (simplify_pred): Ditto.
11272 (simplify_preds_2): Ditto.
11273 (simplify_preds_3): Ditto.
11274 (simplify_preds_4): Ditto.
11275 (simplify_preds): Ditto.
11276 (push_pred): Ditto.
11277 (push_to_worklist): Ditto.
11278 (get_pred_info_from_cmp): Ditto.
11279 (is_degenerated_phi): Ditto.
11280 (normalize_one_pred_1): Ditto.
11281 (normalize_one_pred): Ditto.
11282 (normalize_one_pred_chain): Ditto.
11283 (normalize_preds): Ditto.
11284 (normalize_cond_1): Remove function.
11285 (normalize_cond): Ditto.
11286 (is_gcond_subset_of): Ditto.
11287 (is_subset_of_any): Ditto.
11288 (is_or_set_subset_of): Ditto.
11289 (is_and_set_subset_of): Ditto.
11290 (is_norm_cond_subset_of): Ditto.
11291 (pred_chain_length_cmp): Ditto.
11292 (convert_control_dep_chain_into_preds): Type change.
11293 (find_predicates): Ditto.
11294 (find_def_preds): Ditto.
11295 (destroy_predicates_vecs): Ditto.
11296 (find_matching_predicates_in_rest_chains): Ditto.
11297 (use_pred_not_overlap_with_undef_path_pred): Ditto.
11298 (is_pred_expr_subset): Ditto.
11299 (is_pred_chain_subset_of): Ditto.
11300 (is_included_in): Ditto.
11301 (is_superset_of): Ditto.
11303 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
11305 Update copyright years.
11307 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
11309 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
11310 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
11311 config/arc/arc.md, config/arc/arc.opt,
11312 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
11313 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
11314 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
11315 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
11316 config/linux-protos.h, config/linux.c, config/winnt-c.c,
11317 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
11318 vtable-verify.c, vtable-verify.h: Use the standard form for the
11321 2014-01-02 Tobias Burnus <burnus@net-b.de>
11323 * gcc.c (process_command): Update copyright notice dates.
11324 * gcov-dump.c: Ditto.
11326 * doc/cpp.texi: Bump @copying's copyright year.
11327 * doc/cppinternals.texi: Ditto.
11328 * doc/gcc.texi: Ditto.
11329 * doc/gccint.texi: Ditto.
11330 * doc/gcov.texi: Ditto.
11331 * doc/install.texi: Ditto.
11332 * doc/invoke.texi: Ditto.
11334 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11336 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
11338 2014-01-01 Jakub Jelinek <jakub@redhat.com>
11340 * config/i386/sse.md (*mov<mode>_internal): Guard
11341 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
11343 PR rtl-optimization/59647
11344 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
11345 new_rtx into UNSIGNED_FLOAT rtxes.
11347 Copyright (C) 2014 Free Software Foundation, Inc.
11349 Copying and distribution of this file, with or without modification,
11350 are permitted in any medium without royalty provided the copyright
11351 notice and this notice are preserved.