1 2014-05-07 Richard Biener <rguenther@suse.de>
3 * config.gcc: Always set need_64bit_hwint to yes.
5 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
7 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
8 of using optimize_size.
10 2014-05-06 Mike Stump <mikestump@comcast.net>
12 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
14 2014-05-06 Joseph Myers <joseph@codesourcery.com>
16 * config/i386/sse.md (*mov<mode>_internal)
17 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
18 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
19 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
20 (*<code><mode>3, *andnot<mode>3<mask_name>)
21 (<mask_codefor><code><mode>3<mask_name>): Only consider
22 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
24 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
27 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
29 * lra-constraints.c (valid_address_p): Move earlier in file.
30 Add a constraint argument to the address_info version.
31 (satisfies_memory_constraint_p): New function.
32 (satisfies_address_constraint_p): Likewise.
33 (process_alt_operands, curr_insn_transform): Use them.
34 (process_address): Pass the constraint to valid_address_p when
35 checking address operands.
37 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
39 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
40 to their respective blocks. Fix inadvertent use of "node".
42 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
44 * emit-rtl.c (init_derived_machine_modes): New functionm, split
46 (init_emit_once): ...here.
47 * rtl.h (init_derived_machine_modes): Declare.
48 * toplev.c (do_compile): Call it even if no_backend.
50 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
51 Mike Stump <mikestump@comcast.net>
52 Richard Sandiford <rdsandiford@googlemail.com>
53 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
55 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
56 (rtx_equal_for_memref_p): Update comment.
57 (adjust_offset_for_component_ref): Use wide-int interfaces.
58 * builtins.c (get_object_alignment_2): Likewise.
59 (c_readstr): Likewise.
60 (target_char_cast): Add comment.
61 (determine_block_size): Use wide-int interfaces.
62 (expand_builtin_signbit): Likewise.
63 (fold_builtin_int_roundingfn): Likewise.
64 (fold_builtin_bitop): Likewise.
65 (fold_builtin_bswap): Likewise.
66 (fold_builtin_logarithm): Use signop.
67 (fold_builtin_pow): Likewise.
68 (fold_builtin_memory_op): Use wide-int interfaces.
69 (fold_builtin_object_size): Likewise.
70 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
71 nb_iterations_estimate.
72 (record_niter_bound): Use wide-int interfaces.
73 (get_estimated_loop_iterations_int): Likewise.
74 (get_estimated_loop_iterations): Likewise.
75 (get_max_loop_iterations): Likewise.
76 * cfgloop.h: Include wide-int.h.
77 (struct nb_iter_bound): Change bound to widest_int.
78 (struct loop): Change nb_iterations_upper_bound and
79 nb_iterations_estimate to widest_int.
80 (record_niter_bound): Switch to use widest_int.
81 (get_estimated_loop_iterations): Likewise.
82 (get_max_loop_iterations): Likewise.
83 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
85 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
86 * combine.c (try_combine): Likewise.
87 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
88 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
90 (aarch64_float_const_representable_p): Likewise.
91 * config/arc/arc.c: Include wide-int.h.
92 (arc_can_use_doloop_p): Use wide-int interfaces.
93 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
94 (vfp3_const_double_index): Likewise.
95 * config/avr/avr.c (avr_out_round): Likewise.
96 (avr_fold_builtin): Likewise.
97 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
98 (bfin_can_use_doloop_p): Likewise.
99 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
100 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
101 * config/i386/i386.c: Include wide-int.h.
102 (ix86_data_alignment): Use wide-int interfaces.
103 (ix86_local_alignment): Likewise.
104 (ix86_emit_swsqrtsf): Update real_from_integer.
105 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
106 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
107 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
108 (zero_constant): Likewise.
109 (input_operand): Likewise.
110 (splat_input_operand): Likewise.
111 (non_logical_cint_operand): Change const_double to const_wide_int.
112 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
113 (easy_altivec_constant): Remove comment.
114 (paired_expand_vector_init): Use CONSTANT_P.
115 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
116 (rs6000_emit_move): Update checks.
117 (rs6000_aggregate_candidate): Use wide-int interfaces.
118 (rs6000_expand_ternop_builtin): Likewise.
119 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
120 (rs6000_assemble_integer): Likewise.
121 (rs6000_hash_constant): Likewise.
122 (output_toc): Likewise.
123 (rs6000_rtx_costs): Likewise.
124 (rs6000_emit_swrsqrt); Update call to real_from_integer.
125 * config/rs6000/rs6000-c.c: Include wide-int.h.
126 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
127 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
128 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
129 Handle CONST_WIDE_INT.
130 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
131 Use tree_fits_uhwi_p.
132 * config/sparc/sparc.c: Include wide-int.h.
133 (sparc_fold_builtin): Use wide-int interfaces.
134 * config/vax/vax.c: Include wide-int.h.
135 (vax_float_literal): Use real_from_integer.
136 * coretypes.h (struct hwivec_def): New.
139 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
140 (equiv_constant): Handle CONST_WIDE_INT.
141 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
142 (cselib_hash_rtx): Handle CONST_WIDE_INT.
143 * dbxout.c (stabstr_U): Use wide-int interfaces.
144 (dbxout_type): Update to use cst_fits_shwi_p.
145 * defaults.h (LOG2_BITS_PER_UNIT): Define.
146 (TARGET_SUPPORTS_WIDE_INT): Add default.
147 * dfp.c: Include wide-int.h.
148 (decimal_real_to_integer2): Use wide-int interfaces and rename to
149 decimal_real_to_integer.
150 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
151 decimal_real_to_integer.
152 * doc/generic.texi (Constant expressions): Update for wide_int.
153 * doc/rtl.texi (const_double): Likewise.
154 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
155 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
156 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
157 (REAL_VALUE_FROM_INT): Remove.
158 (TARGET_SUPPORTS_WIDE_INT): New.
159 * doc/tm.texi: Regenerate.
160 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
161 * double-int.h: Include wide-int.h.
162 (struct wi::int_traits): New.
163 * dwarf2out.c (get_full_len): New.
164 (dw_val_equal_p): Add case dw_val_class_wide_int.
165 (size_of_loc_descr): Likewise.
166 (output_loc_operands): Likewise.
167 (insert_double): Remove.
168 (insert_wide_int): New.
170 (print_die): Add case dw_val_class_wide_int.
171 (attr_checksum): Likewise.
172 (attr_checksum_ordered): Likewise.
173 (same_dw_val_p): Likewise.
174 (size_of_die): Likewise.
175 (value_format): Likewise.
176 (output_die): Likewise.
177 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
179 (clz_loc_descriptor): Use wide-int interfaces.
180 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
181 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
182 (round_up_to_align): Use wide-int interfaces.
183 (field_byte_offset): Likewise.
184 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
185 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
186 CONST_DOUBLE handling. Use wide-int interfaces.
187 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
188 (gen_enumeration_type_die): Use add_AT_wide.
189 (hash_loc_operands): Add case dw_val_class_wide_int.
190 (compare_loc_operands): Likewise.
191 * dwarf2out.h: Include wide-int.h.
193 (enum dw_val_class): Add dw_val_class_wide_int.
194 (struct dw_val_struct): Add val_wide.
195 * emit-rtl.c (const_wide_int_htab): New.
196 (const_wide_int_htab_hash): New.
197 (const_wide_int_htab_eq): New.
198 (lookup_const_wide_int): New.
199 (const_double_htab_hash): Use wide-int interfaces.
200 (const_double_htab_eq): Likewise.
201 (rtx_to_double_int): Conditionally compile for wide-int.
202 (immed_double_int_const): Rename to immed_wide_int_const and
204 (immed_double_const): Conditionally compile for wide-int.
205 (init_emit_once): Use wide-int interfaces.
206 * explow.c (plus_constant): Likewise.
207 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
208 (lshift_value): Use wide-int interfaces.
209 (expand_mult): Likewise.
210 (choose_multiplier): Likewise.
211 (expand_smod_pow2): Likewise.
212 (make_tree): Likewise.
213 * expr.c (convert_modes): Consolidate handling of constants.
214 Use wide-int interfaces.
215 (emit_group_load_1): Add note.
216 (store_expr): Update comment.
217 (get_inner_reference): Use wide-int interfaces.
218 (expand_constructor): Update comment.
219 (expand_expr_real_2): Use wide-int interfaces.
220 (expand_expr_real_1): Likewise.
221 (reduce_to_bit_field_precision): Likewise.
222 (const_vector_from_tree): Likewise.
223 * final.c: Include wide-int-print.h.
224 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
225 * fixed-value.c: Include wide-int.h.
226 (fixed_from_string): Use wide-int interfaces.
227 (fixed_to_decimal): Likewise.
228 (fixed_convert_from_real): Likewise.
229 (real_convert_from_fixed): Likewise.
230 * fold-const.h (mem_ref_offset): Return an offset_int.
231 (div_if_zero_remainder): Remove code parameter.
232 * fold-const.c (div_if_zero_remainder): Remove code parameter.
233 Use wide-int interfaces.
234 (may_negate_without_overflow_p): Use wide-int interfaces.
235 (negate_expr_p): Likewise.
236 (fold_negate_expr): Likewise.
237 (int_const_binop_1): Likewise.
238 (const_binop): Likewise.
239 (fold_convert_const_int_from_int): Likewise.
240 (fold_convert_const_int_from_real): Likewise.
241 (fold_convert_const_int_from_fixed): Likewise.
242 (fold_convert_const_fixed_from_int): Likewise.
243 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
244 (sign_bit_p): Use wide-int interfaces.
245 (make_range_step): Likewise.
246 (build_range_check): Likewise. Pass an integer of the correct type
247 instead of using integer_one_node.
248 (range_predecessor): Pass an integer of the correct type instead
249 of using integer_one_node.
250 (range_successor): Likewise.
251 (merge_ranges): Likewise.
252 (unextend): Use wide-int interfaces.
253 (extract_muldiv_1): Likewise.
254 (fold_div_compare): Likewise.
255 (fold_single_bit_test): Likewise.
256 (fold_sign_changed_comparison): Likewise.
257 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
258 (fold_plusminus_mult_expr): Use wide-int interfaces.
259 (native_encode_int): Likewise.
260 (native_interpret_int): Likewise.
261 (fold_unary_loc): Likewise.
262 (pointer_may_wrap_p): Likewise.
263 (size_low_cst): Likewise.
264 (mask_with_tz): Likewise.
265 (fold_binary_loc): Likewise.
266 (fold_ternary_loc): Likewise.
267 (multiple_of_p): Likewise.
268 (tree_call_nonnegative_warnv_p): Update calls to
269 tree_int_cst_min_precision and real_from_integer.
270 (fold_negate_const): Use wide-int interfaces.
271 (fold_abs_const): Likewise.
272 (fold_relational_const): Use tree_int_cst_lt.
273 (round_up_loc): Use wide-int interfaces.
274 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
275 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
276 * gengtype.c: Remove include of double-int.h.
277 (do_typedef): Use wide-int interfaces.
278 (open_base_files): Add wide-int.h.
279 (main): Add offset_int and widest_int typedefs.
280 * gengtype-lex.l: Handle "^".
281 (CXX_KEYWORD): Add "static".
282 * gengtype-parse.c (require3): New.
283 (require_template_declaration): Handle constant template arguments
284 and nested templates.
285 * gengtype-state.c: Don't include "double-int.h".
286 * genpreds.c (write_one_predicate_function): Update comment.
287 (write_tm_constrs_h): Add check for hval and lval use in
289 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
290 (add_to_sequence): Likewise.
291 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
292 and const_double_operand.
293 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
295 * gimple-fold.c (get_base_constructor): Likewise.
296 (fold_array_ctor_reference): Likewise.
297 (fold_nonarray_ctor_reference): Likewise.
298 (fold_const_aggregate_ref_1): Likewise.
299 (gimple_val_nonnegative_real_p): Likewise.
300 (gimple_fold_indirect_ref): Likewise.
301 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
302 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
303 (struct slsr_cand_d): Change index to be widest_int.
304 (struct incr_info_d): Change incr to be widest_int.
305 (alloc_cand_and_find_basis): Use wide-int interfaces.
306 (slsr_process_phi): Likewise.
307 (backtrace_base_for_ref): Likewise. Return a widest_int.
308 (restructure_reference): Take a widest_int instead of a double_int.
309 (slsr_process_ref): Use wide-int interfaces.
310 (create_mul_ssa_cand): Likewise.
311 (create_mul_imm_cand): Likewise.
312 (create_add_ssa_cand): Likewise.
313 (create_add_imm_cand): Take a widest_int instead of a double_int.
314 (slsr_process_add): Use wide-int interfaces.
315 (slsr_process_cast): Likewise.
316 (slsr_process_copy): Likewise.
317 (dump_candidate): Likewise.
318 (dump_incr_vec): Likewise.
319 (replace_ref): Likewise.
320 (cand_increment): Likewise. Return a widest_int.
321 (cand_abs_increment): Likewise.
322 (replace_mult_candidate): Take a widest_int instead of a double_int.
323 (replace_unconditional_candidate): Use wide-int interfaces.
324 (incr_vec_index): Take a widest_int instead of a double_int.
325 (create_add_on_incoming_edge): Likewise.
326 (create_phi_basis): Use wide-int interfaces.
327 (replace_conditional_candidate): Likewise.
328 (record_increment): Take a widest_int instead of a double_int.
329 (record_phi_increments): Use wide-int interfaces.
330 (phi_incr_cost): Take a widest_int instead of a double_int.
331 (lowest_cost_path): Likewise.
332 (total_savings): Likewise.
333 (analyze_increments): Use wide-int interfaces.
334 (ncd_with_phi): Take a widest_int instead of a double_int.
335 (ncd_of_cand_and_phis): Likewise.
336 (nearest_common_dominator_for_cands): Likewise.
337 (insert_initializers): Use wide-int interfaces.
338 (all_phi_incrs_profitable): Likewise.
339 (replace_one_candidate): Likewise.
340 (replace_profitable_candidates): Likewise.
341 * godump.c: Include wide-int-print.h.
342 (go_output_typedef): Use wide-int interfaces.
343 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
344 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
345 (build_loop_iteration_domains): Likewise.
346 * hooks.h: Include wide-int.h rather than double-int.h.
347 (hook_bool_dint_dint_uint_bool_true): Delete.
348 (hook_bool_wint_wint_uint_bool_true): Declare.
349 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
350 (hook_bool_wint_wint_uint_bool_true): New.
351 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
353 (ubsan_expand_si_overflow_mul_check): Likewise.
354 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
355 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
356 (get_ancestor_addr_info): Likewise.
357 (ipa_modify_call_arguments): Likewise.
358 * loop-doloop.c (doloop_modify): Likewise.
359 (doloop_optimize): Likewise.
360 * loop-iv.c (iv_number_of_iterations): Likewise.
361 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
362 (unroll_loop_constant_iterations): Likewise.
363 (decide_unroll_runtime_iterations): Likewise.
364 (unroll_loop_runtime_iterations): Likewise.
365 (decide_peel_simple): Likewise.
366 (decide_unroll_stupid): Likewise.
367 * lto-streamer-in.c (streamer_read_wi): Add.
368 (input_cfg): Use wide-int interfaces.
369 (lto_input_tree_1): Likewise.
370 * lto-streamer-out.c (streamer_write_wi): Add.
371 (hash_tree): Use wide-int interfaces.
372 (output_cfg): Likewise.
373 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
374 (GTFILES): Add wide-int.h and signop.h.
375 (TAGS): Look for .cc files too.
376 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
377 * optabs.c (expand_subword_shift): Likewise.
378 (expand_doubleword_shift): Likewise.
379 (expand_absneg_bit): Likewise.
380 (expand_copysign_absneg): Likewise.
381 (expand_copysign_bit): Likewise.
382 * postreload.c (reload_cse_simplify_set): Likewise.
383 * predict.c (predict_iv_comparison): Likewise.
384 * pretty-print.h: Include wide-int-print.h.
386 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
387 * print-tree.c: Include wide-int-print.h.
388 (print_node_brief): Use wide-int interfaces.
389 (print_node): Likewise.
390 * read-rtl.c (validate_const_wide_int): New.
391 (read_rtx_code): Add CONST_WIDE_INT case.
392 * real.c: Include wide-int.h.
393 (real_to_integer2): Delete.
394 (real_to_integer): New function, returning a wide_int.
395 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
396 (ten_to_ptwo): Update call to real_from_integer.
397 (real_digit): Likewise.
398 * real.h: Include signop.h, wide-int.h and insn-modes.h.
399 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
400 (REAL_VALUE_TO_INT): Delete.
401 (real_to_integer): Declare a wide-int form.
402 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
403 * recog.c (const_int_operand): Improve comment.
404 (const_scalar_int_operand): New.
405 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
406 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
407 (split_double): Likewise.
408 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
409 (rtx_size): Likewise.
410 (rtx_alloc_stat_v): New.
411 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
412 (cwi_output_hex): New.
413 (iterative_hash_rtx): Handle CONST_WIDE_INT.
414 (cwi_check_failed_bounds): New.
415 * rtl.def (CONST_WIDE_INT): New.
416 * rtl.h: Include <utility> and wide-int.h.
417 (struct hwivec_def): New.
418 (CWI_GET_NUM_ELEM): New.
419 (CWI_PUT_NUM_ELEM): New.
420 (struct rtx_def): Add num_elem and hwiv.
421 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
422 (CASE_CONST_UNIQUE): Likewise.
423 (CASE_CONST_ANY): Likewise.
424 (CONST_SCALAR_INT_P): Likewise.
425 (CONST_WIDE_INT_P): New.
428 (cwi_check_failed_bounds): New.
431 (CONST_WIDE_INT_VEC) New.
432 (CONST_WIDE_INT_NUNITS) New.
433 (CONST_WIDE_INT_ELT) New.
434 (rtx_mode_t): New type.
435 (wi::int_traits <rtx_mode_t>): New.
437 (wi::min_value): New.
438 (wi::max_value): New.
440 (const_wide_int_alloc): New.
441 (immed_wide_int_const): New.
442 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
443 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
444 * signop.h: New file.
445 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
446 (simplify_const_unary_operation): Use wide-int interfaces.
447 (simplify_binary_operation_1): Likewise.
448 (simplify_const_binary_operation): Likewise.
449 (simplify_const_relational_operation): Likewise.
450 (simplify_immed_subreg): Likewise.
451 * stmt.c (expand_case): Likewise.
452 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
453 signop rather than a bool.
454 * stor-layout.c (layout_type): Use wide-int interfaces.
455 (initialize_sizetypes): Update calls to
456 set_min_and_max_values_for_integral_type.
457 (set_min_and_max_values_for_integral_type): Take a signop rather
458 than a bool. Use wide-int interfaces.
459 (fixup_signed_type): Update accordingly. Remove
460 HOST_BITS_PER_DOUBLE_INT limit.
461 (fixup_unsigned_type): Likewise.
462 * system.h (STATIC_CONSTANT_P): New.
463 (STATIC_ASSERT): New.
464 * target.def (can_use_doloop_p): Take widest_ints rather than
466 * target.h: Include wide-int.h rather than double-int.h.
467 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
469 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
470 rather than INT_CST_LT_UNSIGNED.
471 (can_use_doloop_if_innermost): Take widest_ints rather than
473 * tree-affine.c: Include wide-int-print.h.
474 (double_int_ext_for_comb): Delete.
475 (wide_int_ext_for_comb): New.
476 (aff_combination_zero): Use wide-int interfaces.
477 (aff_combination_const): Take a widest_int instead of a double_int.
478 (aff_combination_elt): Use wide-int interfaces.
479 (aff_combination_scale): Take a widest_int instead of a double_int.
480 (aff_combination_add_elt): Likewise.
481 (aff_combination_add_cst): Likewise.
482 (aff_combination_add): Use wide-int interfaces.
483 (aff_combination_convert): Likewise.
484 (tree_to_aff_combination): Likewise.
485 (add_elt_to_tree): Take a widest_int instead of a double_int.
486 (aff_combination_to_tree): Use wide-int interfaces.
487 (aff_combination_remove_elt): Likewise.
488 (aff_combination_add_product): Take a widest_int instead of
490 (aff_combination_mult): Use wide-int interfaces.
491 (aff_combination_expand): Likewise.
492 (double_int_constant_multiple_p): Delete.
493 (wide_int_constant_multiple_p): New.
494 (aff_combination_constant_multiple_p): Take a widest_int pointer
495 instead of a double_int pointer.
496 (print_aff): Use wide-int interfaces.
497 (get_inner_reference_aff): Take a widest_int pointer
498 instead of a double_int pointer.
499 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
500 * tree-affine.h: Include wide-int.h.
501 (struct aff_comb_elt): Change type of coef to widest_int.
502 (struct affine_tree_combination): Change type of offset to widest_int.
503 (double_int_ext_for_comb): Delete.
504 (wide_int_ext_for_comb): New.
505 (aff_combination_const): Use widest_int instead of double_int.
506 (aff_combination_scale): Likewise.
507 (aff_combination_add_elt): Likewise.
508 (aff_combination_constant_multiple_p): Likewise.
509 (get_inner_reference_aff): Likewise.
510 (aff_comb_cannot_overlap_p): Likewise.
511 (aff_combination_zero_p): Use wide-int interfaces.
512 * tree.c: Include tree.h.
513 (init_ttree): Use make_int_cst.
514 (tree_code_size): Removed code for INTEGER_CST case.
515 (tree_size): Add INTEGER_CST case.
516 (make_node_stat): Update comment.
517 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
518 (build_int_cst_type): Use wide-int interfaces.
519 (double_int_to_tree): Likewise.
520 (double_int_fits_to_tree_p): Delete.
521 (force_fit_type_double): Delete.
522 (force_fit_type): New.
523 (int_cst_hash_hash): Use wide-int interfaces.
524 (int_cst_hash_eq): Likewise.
525 (build_int_cst_wide): Delete.
526 (wide_int_to_tree): New.
527 (cache_integer_cst): Use wide-int interfaces.
528 (build_low_bits_mask): Likewise.
529 (cst_and_fits_in_hwi): Likewise.
530 (real_value_from_int_cst): Likewise.
531 (make_int_cst_stat): New.
532 (integer_zerop): Use wide_int interfaces.
533 (integer_onep): Likewise.
534 (integer_all_onesp): Likewise.
535 (integer_pow2p): Likewise.
536 (integer_nonzerop): Likewise.
537 (tree_log2): Likewise.
538 (tree_floor_log2): Likewise.
539 (tree_ctz): Likewise.
540 (int_size_in_bytes): Likewise.
541 (mem_ref_offset): Return an offset_int rather than a double_int.
542 (build_type_attribute_qual_variant): Use wide_int interfaces.
543 (type_hash_eq): Likewise
544 (tree_int_cst_equal): Likewise.
545 (tree_int_cst_lt): Delete.
546 (tree_int_cst_compare): Likewise.
547 (tree_fits_shwi_p): Use wide_int interfaces.
548 (tree_fits_uhwi_p): Likewise.
549 (tree_int_cst_sign_bit): Likewise.
550 (tree_int_cst_sgn): Likewise.
551 (tree_int_cst_min_precision): Take a signop rather than a bool.
552 (simple_cst_equal): Use wide_int interfaces.
553 (compare_tree_int): Likewise.
554 (iterative_hash_expr): Likewise.
555 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
557 (get_type_static_bounds): Use wide_int interfaces.
558 (tree_int_cst_elt_check_failed): New.
559 (build_common_tree_nodes): Reordered to set prec before filling in
561 (int_cst_value): Check cst_and_fits_in_hwi.
562 (widest_int_cst_value): Use wide_int interfaces.
563 (upper_bound_in_type): Likewise.
564 (lower_bound_in_type): Likewise.
565 (num_ending_zeros): Likewise.
566 (drop_tree_overflow): Likewise.
567 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
568 (gen_conditions_for_pow_cst_base): Likewise.
569 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
570 (group_case_labels_stmt): Use wide-int interfaces.
571 (verify_gimple_assign_binary): Likewise.
572 (print_loop): Likewise.
573 * tree-chrec.c (tree_fold_binomial): Likewise.
574 * tree-core.h (struct tree_base): Add int_length.
575 (struct tree_int_cst): Change rep of value.
576 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
577 (dr_may_alias_p): Likewise.
578 (max_stmt_executions_tree): Likewise.
579 * tree.def (INTEGER_CST): Update comment.
580 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
581 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
582 * tree-dump.c: Include wide-int.h and wide-int-print.h.
583 (dequeue_and_dump): Use wide-int interfaces.
584 * tree.h: Include wide-int.h.
585 (NULL_TREE): Moved to earlier loc in file.
586 (TREE_INT_CST_ELT_CHECK): New.
587 (tree_int_cst_elt_check_failed): New.
589 (TREE_INT_CST): Delete.
590 (TREE_INT_CST_LOW): Use wide-int interfaces.
591 (TREE_INT_CST_HIGH): Delete.
592 (TREE_INT_CST_NUNITS): New.
593 (TREE_INT_CST_EXT_NUNITS): Likewise.
594 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
595 (TREE_INT_CST_ELT): Likewise.
596 (INT_CST_LT): Delete.
597 (tree_int_cst_elt_check): New (two forms).
598 (type_code_size): Update comment.
599 (make_int_cst_stat, make_int_cst): New.
600 (tree_to_double_int): Delete.
601 (double_int_fits_to_tree_p): Delete.
602 (force_fit_type_double): Delete.
603 (build_int_cstu): Replace with out-of-line function.
604 (build_int_cst_wide): Delete.
605 (tree_int_cst_lt): Define inline.
606 (tree_int_cst_le): New.
607 (tree_int_cst_compare): Define inline.
608 (tree_int_cst_min_precision): Take a signop rather than a bool.
609 (wi::int_traits <const_tree>): New.
610 (wi::int_traits <tree>): New.
611 (wi::extended_tree): New.
612 (wi::int_traits <wi::extended_tree>): New.
613 (wi::to_widest): New.
614 (wi::to_offset): New.
615 (wi::fits_to_tree_p): New.
616 (wi::min_value): New.
617 (wi::max_value): New.
618 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
619 (copy_tree_body_r): Likewise.
620 * tree-object-size.c (compute_object_offset): Likewise.
621 (addr_object_size): Likewise.
622 * tree-predcom.c: Include wide-int-print.h.
623 (struct dref_d): Change type of offset to widest_int.
624 (dump_dref): Call wide-int printer.
625 (aff_combination_dr_offset): Use wide-int interfaces.
626 (determine_offset): Take a widest_int pointer rather than a
628 (split_data_refs_to_components): Use wide-int interfaces.
629 (suitable_component_p): Likewise.
630 (order_drefs): Likewise.
631 (add_ref_to_chain): Likewise.
632 (valid_initializer_p): Likewise.
633 (determine_roots_comp): Likewise.
634 * tree-pretty-print.c: Include wide-int-print.h.
635 (dump_generic_node): Use wide-int interfaces.
636 * tree-sra.c (sra_ipa_modify_expr): Likewise.
637 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
638 (move_fixed_address_to_symbol): Likewise.
639 (move_hint_to_base): Likewise.
640 (move_pointer_to_base): Likewise.
641 (move_variant_to_index): Likewise.
642 (most_expensive_mult_to_index): Likewise.
643 (addr_to_parts): Likewise.
644 (copy_ref_info): Likewise.
645 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
646 (indirect_refs_may_alias_p): Likewise.
647 (stmt_kills_ref_p_1): Likewise.
648 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
649 * tree-ssa-ccp.c: Update comment at top of file. Include
651 (struct prop_value_d): Change type of mask to widest_int.
652 (extend_mask): New function.
653 (dump_lattice_value): Use wide-int interfaces.
654 (get_default_value): Likewise.
655 (set_constant_value): Likewise.
656 (set_value_varying): Likewise.
657 (valid_lattice_transition): Likewise.
658 (set_lattice_value): Likewise.
659 (value_to_double_int): Delete.
660 (value_to_wide_int): New.
661 (get_value_from_alignment): Use wide-int interfaces.
662 (get_value_for_expr): Likewise.
663 (do_dbg_cnt): Likewise.
664 (ccp_finalize): Likewise.
665 (ccp_lattice_meet): Likewise.
666 (bit_value_unop_1): Use widest_ints rather than double_ints.
667 (bit_value_binop_1): Likewise.
668 (bit_value_unop): Use wide-int interfaces.
669 (bit_value_binop): Likewise.
670 (bit_value_assume_aligned): Likewise.
671 (evaluate_stmt): Likewise.
672 (ccp_fold_stmt): Likewise.
673 (visit_cond_stmt): Likewise.
674 (ccp_visit_stmt): Likewise.
675 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
676 (constant_pointer_difference): Likewise.
677 (associate_pointerplus): Likewise.
678 (combine_conversions): Likewise.
679 * tree-ssa-loop.h: Include wide-int.h.
680 (struct tree_niter_desc): Change type of max to widest_int.
681 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
682 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
683 (remove_redundant_iv_tests): Likewise.
684 (canonicalize_loop_induction_variables): Likewise.
685 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
686 (constant_multiple_of): Take a widest_int pointer instead of
687 a double_int pointer.
688 (get_computation_aff): Use wide-int interfaces.
689 (ptr_difference_cost): Likewise.
690 (difference_cost): Likewise.
691 (get_loop_invariant_expr_id): Likewise.
692 (get_computation_cost_at): Likewise.
693 (iv_elimination_compare_lt): Likewise.
694 (may_eliminate_iv): Likewise.
695 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
696 instead of double_int.
697 (max_loop_iterations): Likewise.
698 (max_stmt_executions): Likewise.
699 (estimated_stmt_executions): Likewise.
700 * tree-ssa-loop-niter.c: Include wide-int-print.h.
701 (split_to_var_and_offset): Use wide-int interfaces.
702 (determine_value_range): Likewise.
703 (bound_difference_of_offsetted_base): Likewise.
704 (bounds_add): Take a widest_int instead of a double_int.
705 (number_of_iterations_ne_max): Use wide-int interfaces.
706 (number_of_iterations_ne): Likewise.
707 (number_of_iterations_lt_to_ne): Likewise.
708 (assert_loop_rolls_lt): Likewise.
709 (number_of_iterations_lt): Likewise.
710 (number_of_iterations_le): Likewise.
711 (number_of_iterations_cond): Likewise.
712 (number_of_iterations_exit): Likewise.
713 (finite_loop_p): Likewise.
714 (derive_constant_upper_bound_assign): Likewise.
715 (derive_constant_upper_bound): Return a widest_int.
716 (derive_constant_upper_bound_ops): Likewise.
717 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
718 (record_estimate): Take a widest_int rather than a double_int.
719 (record_nonwrapping_iv): Use wide-int interfaces.
720 (double_int_cmp): Delete.
722 (bound_index): Take a widest_int rather than a double_int.
723 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
724 (maybe_lower_iteration_bound): Likewise.
725 (estimate_numbers_of_iterations_loop): Likewise.
726 (estimated_loop_iterations): Take a widest_int pointer than than
727 a double_int pointer.
728 (estimated_loop_iterations_int): Use wide-int interfaces.
729 (max_loop_iterations): Take a widest_int pointer than than
730 a double_int pointer.
731 (max_loop_iterations_int): Use wide-int interfaces.
732 (max_stmt_executions): Take a widest_int pointer than than
733 a double_int pointer.
734 (estimated_stmt_executions): Likewise.
735 (n_of_executions_at_most): Use wide-int interfaces.
736 (scev_probably_wraps_p): Likewise.
737 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
739 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
741 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
742 double_ints. Adjust for trailing_wide_ints <3> representation.
743 (set_nonzero_bits): Likewise.
744 (get_range_info): Return wide_ints rather than double_ints.
745 Adjust for trailing_wide_ints <3> representation.
746 (get_nonzero_bits): Likewise.
747 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
749 * tree-ssanames.h (struct range_info_def): Replace min, max and
750 nonzero_bits with a trailing_wide_ints <3>.
751 (set_range_info): Use wide_int_refs rather than double_ints.
752 (set_nonzero_bits): Likewise.
753 (get_range_info): Return wide_ints rather than double_ints.
754 (get_nonzero_bits): Likewise.
755 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
756 * tree-ssa-pre.c (phi_translate_1): Likewise.
757 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
758 (acceptable_pow_call): Likewise.
759 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
761 (vn_reference_fold_indirect): Likewise.
762 (vn_reference_maybe_forwprop_address): Likewise.
763 (valueize_refs_1): Likewise.
764 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
765 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
766 tree_int_cst_lt and tree_int_cst_le.
767 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
769 (streamer_alloc_tree): Likewise.
770 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
771 (streamer_write_tree_header): Likewise.
772 (streamer_write_integer_cst): Likewise.
773 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
774 (build_constructors): Likewise.
775 (array_value_type): Likewise.
776 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
777 (vect_check_gather): Likewise.
778 * tree-vect-generic.c (build_replicated_const): Likewise.
779 (expand_vector_divmod): Likewise.
780 * tree-vect-loop.c (vect_transform_loop): Likewise.
781 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
782 (vect_do_peeling_for_alignment): Likewise.
783 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
784 * tree-vrp.c: Include wide-int.h.
785 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
786 (extract_range_from_assert): Use wide-int interfaces.
787 (vrp_int_const_binop): Likewise.
788 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
790 (ranges_from_anti_range): Use wide-int interfaces.
791 (quad_int_cmp): Delete.
792 (quad_int_pair_sort): Likewise.
793 (extract_range_from_binary_expr_1): Use wide-int interfaces.
794 (extract_range_from_unary_expr_1): Likewise.
795 (adjust_range_with_scev): Likewise.
796 (masked_increment): Take and return wide_ints rather than double_ints.
797 (register_edge_assert_for_2): Use wide-int interfaces.
798 (check_array_ref): Likewise.
799 (search_for_addr_array): Likewise.
800 (maybe_set_nonzero_bits): Likewise.
801 (union_ranges): Pass an integer of the correct type instead of
802 using integer_one_node.
803 (intersect_ranges): Likewise.
804 (simplify_truth_ops_using_ranges): Likewise.
805 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
806 (range_fits_type_p): Likewise.
807 (simplify_cond_using_ranges): Likewise. Take a signop rather than
809 (simplify_conversion_using_ranges): Use wide-int interfaces.
810 (simplify_float_conversion_using_ranges): Likewise.
811 (vrp_finalize): Likewise.
812 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
813 (gimple_stringops_transform): Likewise.
814 * varasm.c (decode_addr_const): Likewise.
815 (const_hash_1): Likewise.
816 (const_rtx_hash_1): Likewise
817 (output_constant): Likewise.
818 (array_size_for_constructor): Likewise.
819 (output_constructor_regular_field): Likewise.
820 (output_constructor_bitfield): Likewise.
821 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
822 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
824 * gencheck.c: Define BITS_PER_UNIT.
827 * wide-int-print.cc: New.
828 * wide-int-print.h: New.
830 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
832 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
834 2014-05-06 Richard Biener <rguenther@suse.de>
836 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
837 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
838 (TODO_verify_all): Adjust.
839 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
840 TODO_verify_stmts and TODO_verify_rtl_sharing.
841 * bb-reorder.c: Likewise.
842 * cfgexpand.c: Likewise.
845 * function.c: Likewise.
846 * fwprop.c: Likewise.
848 * gimple-ssa-isolate-paths.c: Likewise.
849 * gimple-ssa-strength-reduction.c: Likewise.
850 * ipa-split.c: Likewise.
851 * loop-init.c: Likewise.
852 * loop-unroll.c: Likewise.
853 * lower-subreg.c: Likewise.
854 * modulo-sched.c: Likewise.
855 * postreload-gcse.c: Likewise.
856 * predict.c: Likewise.
858 * sched-rgn.c: Likewise.
859 * store-motion.c: Likewise.
860 * tracer.c: Likewise.
861 * trans-mem.c: Likewise.
862 * tree-call-cdce.c: Likewise.
863 * tree-cfg.c: Likewise.
864 * tree-cfgcleanup.c: Likewise.
865 * tree-complex.c: Likewise.
866 * tree-eh.c: Likewise.
867 * tree-emutls.c: Likewise.
868 * tree-if-conv.c: Likewise.
869 * tree-into-ssa.c: Likewise.
870 * tree-loop-distribution.c: Likewise.
871 * tree-object-size.c: Likewise.
872 * tree-parloops.c: Likewise.
873 * tree-pass.h: Likewise.
874 * tree-sra.c: Likewise.
875 * tree-ssa-ccp.c: Likewise.
876 * tree-ssa-copy.c: Likewise.
877 * tree-ssa-copyrename.c: Likewise.
878 * tree-ssa-dce.c: Likewise.
879 * tree-ssa-dom.c: Likewise.
880 * tree-ssa-dse.c: Likewise.
881 * tree-ssa-forwprop.c: Likewise.
882 * tree-ssa-ifcombine.c: Likewise.
883 * tree-ssa-loop-ch.c: Likewise.
884 * tree-ssa-loop-ivcanon.c: Likewise.
885 * tree-ssa-loop.c: Likewise.
886 * tree-ssa-math-opts.c: Likewise.
887 * tree-ssa-phiopt.c: Likewise.
888 * tree-ssa-phiprop.c: Likewise.
889 * tree-ssa-pre.c: Likewise.
890 * tree-ssa-reassoc.c: Likewise.
891 * tree-ssa-sink.c: Likewise.
892 * tree-ssa-strlen.c: Likewise.
893 * tree-ssa-tail-merge.c: Likewise.
894 * tree-ssa-uncprop.c: Likewise.
895 * tree-switch-conversion.c: Likewise.
896 * tree-tailcall.c: Likewise.
897 * tree-vect-generic.c: Likewise.
898 * tree-vectorizer.c: Likewise.
899 * tree-vrp.c: Likewise.
901 * var-tracking.c: Likewise.
902 * bt-load.c: Likewise.
903 * cfgcleanup.c: Likewise.
904 * combine-stack-adj.c: Likewise.
905 * combine.c: Likewise.
906 * compare-elim.c: Likewise.
907 * config/epiphany/resolve-sw-modes.c: Likewise.
908 * config/i386/i386.c: Likewise.
909 * config/mips/mips.c: Likewise.
910 * config/s390/s390.c: Likewise.
911 * config/sh/sh_treg_combine.cc: Likewise.
912 * config/sparc/sparc.c: Likewise.
917 * mode-switching.c: Likewise.
918 * passes.c: Likewise.
919 * postreload.c: Likewise.
921 * reg-stack.c: Likewise.
922 * regcprop.c: Likewise.
923 * regrename.c: Likewise.
926 2014-05-06 Richard Biener <rguenther@suse.de>
929 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
930 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
932 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
935 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
937 2014-05-05 Radovan Obradovic <robradovic@mips.com>
938 Tom de Vries <tom@codesourcery.com>
940 * target.def (call_fusage_contains_non_callee_clobbers): New DEFHOOKPOD.
941 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
943 (@node Miscellaneous Register Hooks): New node.
944 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
945 * doc/tm.texi: Regenerate.
947 2014-05-05 Marek Polacek <polacek@redhat.com>
950 * opts.c (common_handle_option): Call error_at instead of warning_at.
952 2014-05-05 Richard Biener <rguenther@suse.de>
954 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
955 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
956 under the TODO_verify_il umbrella.
958 2014-05-05 Richard Biener <rguenther@suse.de>
960 * passes.c (execute_function_todo): Move TODO_verify_flow under
961 the TODO_verify_ul umbrella.
963 2014-05-05 Richard Biener <rguenther@suse.de>
966 * fold-const.c (fold_binary_loc): Consistently avoid
967 canonicalizing X & CST away from a CST that is the mask
970 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
972 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
973 int argument to enum machine_mode.
974 (picochip_class_max_nregs): Ditto.
975 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
976 (picochip_class_max_nregs): Ditto.
978 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
980 * target.def: Add new target hook.
981 * doc/tm.texi: Regenerate.
982 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
983 * targhooks.c (default_keep_leaf_when_profiled): New function.
985 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
986 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
988 2014-05-05 Bin Cheng <bin.cheng@arm.com>
990 PR tree-optimization/60363
991 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
992 (copy_phi_args): New parameters. Call get_value_locus_in_path.
993 (update_destination_phis): New parameter.
994 (create_edge_and_update_destination_phis): Ditto.
995 (ssa_fix_duplicate_block_edges): Pass new arguments.
996 (thread_single_edge): Ditto.
998 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
1000 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
1001 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
1002 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
1003 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
1004 Use RS6000_BTM_HARD_FLOAT.
1005 (BU_MISC_2): Likewise.
1006 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
1007 RS6000_BTM_HARD_FLOAT.
1008 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
1010 (rs6000_invalid_builtin): Add hard floating builtin support.
1011 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
1012 hard float builtins.
1013 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
1015 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1017 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
1018 Add missing function* argument.
1020 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1022 * lra-constraints.c (valid_address_p): Move earlier in file.
1023 Add a constraint argument to the address_info version.
1024 (satisfies_memory_constraint_p): New function.
1025 (satisfies_address_constraint_p): Likewise.
1026 (process_alt_operands, curr_insn_transform): Use them.
1027 (process_address): Pass the constraint to valid_address_p when
1028 checking address operands.
1030 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1032 * config/mips/mips.c (mips_isa_rev): New variable.
1033 (mips_set_architecture): Set it.
1034 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1036 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1037 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1038 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1039 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1040 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
1041 conditions in terms of mips_isa_rev.
1042 (mips_isa_rev): Declare.
1044 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1046 * config/sh/sh-mem.cc: Use tabs instead of spaces.
1047 (prob_unlikely, prob_likely): Make variables const.
1049 2014-05-03 Denis Chertykov <chertykov@gmail.com>
1051 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
1053 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1055 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
1057 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1059 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
1060 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
1061 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
1063 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
1064 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
1066 Replace usage of ROUND_REG with sh_round_reg.
1067 Use CEIL instead of ROUND_ADVANCE.
1069 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1072 * config/sh/sh.c: Include stdlib headers before everything else.
1074 2014-05-02 Jakub Jelinek <jakub@redhat.com>
1076 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
1077 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
1078 (gimplify_adjust_omp_clauses): Simd region is never
1079 directly nested in combined parallel. Instead, for linear
1080 with copyin/copyout, if in combined for simd loop, make decl
1081 firstprivate/lastprivate on OMP_FOR.
1082 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
1083 expand_omp_for_static_chunk): When setting endvar, also set
1084 fd->loop.v to the same value.
1086 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1088 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
1090 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
1092 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
1095 2014-05-02 Marek Polacek <polacek@redhat.com>
1097 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
1099 2014-05-02 Kito Cheng <kito@0xlab.org>
1101 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
1102 to a C expression marco.
1103 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
1104 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
1105 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
1106 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
1107 HONOR_REG_ALLOC_ORDER.
1108 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
1110 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1112 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
1114 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1116 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
1118 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
1120 * tree-if-conv.c (is_cond_scalar_reduction): New function.
1121 (convert_scalar_cond_reduction): Likewise.
1122 (predicate_scalar_phi): Add recognition and transformation
1123 of simple conditioanl reduction to be vectorizable.
1125 2014-05-01 Marek Polacek <polacek@redhat.com>
1128 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
1130 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
1132 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
1133 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
1134 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
1135 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
1136 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
1137 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
1138 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
1139 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
1141 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
1143 * config/arc/arc.opt (mlra): Move comment above option name
1144 to avoid mis-parsing as language options.
1146 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1148 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
1149 * config/sol2.h: ... here.
1150 * config/sol2-10.h: Remove.
1152 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
1153 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
1154 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
1155 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
1156 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
1157 * config/sol2.h: ... here.
1158 (SECTION_NAME_FORMAT): Don't redefine.
1159 (STARTFILE_ARCH32_SPEC): Rename to ...
1160 (STARTFILE_ARCH_SPEC): ... this.
1161 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
1162 * config/sparc/sol2.h: ... here.
1163 (SECTION_NAME_FORMAT): Don't undef.
1164 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
1165 (SUBTARGET_EXTRA_SPECS): Remove.
1166 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
1168 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
1169 (MD_STARTFILE_PREFIX): Remove.
1170 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
1171 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
1172 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
1173 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
1174 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
1175 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
1176 * config/i386/sol2.h: ... here.
1177 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
1178 * config/i386/sol2-bi.h: Remove.
1179 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
1180 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
1182 * config/i386/t-sol2-64: Rename to ...
1183 * config/i386/t-sol2: ... this.
1184 * config/sparc/t-sol2-64: Rename to ...
1185 * config/sparc/t-sol2: ... this.
1187 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
1188 sol2_tm_file_head, sol2_tm_file_tail.
1189 Include ${cpu_type}/sol2.h before sol2.h.
1191 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
1192 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
1193 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
1194 Reflect i386/t-sol2-64 renaming.
1195 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
1196 Reflect sparc/t-sol2-64 renaming.
1198 2014-04-30 Richard Biener <rguenther@suse.de>
1200 * passes.c (execute_function_todo): Move TODO_verify_stmts
1201 and TODO_verify_ssa under the TODO_verify_il umbrella.
1202 * tree-ssa.h (verify_ssa): Adjust prototype.
1203 * tree-ssa.c (verify_ssa): Add parameter to tell whether
1204 we should verify SSA operands.
1205 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
1206 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
1207 whether we should verify whether not throwing stmts have EH info.
1208 * graphite-scop-detection.c (create_sese_edges): Adjust.
1209 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
1210 * tree-eh.c (lower_try_finally_switch): Do not add the
1211 default case label twice.
1213 2014-04-30 Marek Polacek <polacek@redhat.com>
1215 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
1216 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
1217 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
1218 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
1220 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
1222 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
1223 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
1224 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
1225 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
1226 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
1227 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
1228 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
1229 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
1231 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1233 * tree-cfg.c (dump_function_to_file): Dump the return type of
1234 functions, in a line to itself before the function body, mimicking
1235 the layout of a C function.
1237 2014-04-29 Jakub Jelinek <jakub@redhat.com>
1239 PR tree-optimization/60971
1240 * tree-tailcall.c (process_assignment): Reject conversions which
1243 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
1245 * calls.c (initialize_argument_information): Always treat
1246 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
1247 (expand_call): Likewise.
1248 (emit_library_call_calue_1): Likewise.
1249 * expr.c (PUSH_ARGS_REVERSED): Do not define.
1250 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
1253 2014-04-29 Nick Clifton <nickc@redhat.com>
1255 * config/msp430/msp430.md (umulsidi): Fix typo.
1256 (mulhisi3): Enable even inside interrupt handlers.
1257 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
1258 bigger return address pushed in large mode.
1260 2014-04-29 Nick Clifton <nickc@redhat.com>
1262 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
1263 (arc_init_reg_tables): Use a machine_mode enum to iterate over
1265 * config/m32r/m32r.c (init_reg_tables): Likewise.
1266 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
1267 enum to hold the modes.
1269 2014-04-29 Richard Biener <rguenther@suse.de>
1271 * dominance.c (free_dominance_info): Add overload with
1273 (dom_info_state): Likewise.
1274 (dom_info_available_p): Likewise.
1275 * basic-block.h (free_dominance_info, dom_info_state,
1276 dom_info_available_p): Declare overloads.
1277 * passes.c (execute_function_todo): Verify that verifiers
1278 don't change dominator info state. Drop dominator info
1279 for IPA pass invocations.
1280 * cgraph.c (release_function_body): Restore asserts that
1281 dominator information is released.
1283 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
1285 * doc/invoke.texi: Fix typo.
1286 * tree-vrp.c: Fix typos.
1287 * gimple.c (infer_nonnull_range): Reorder operands of an &&
1290 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1292 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
1294 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1296 * config/aarch64/aarch64-builtins.c
1297 (aarch64_types_storestruct_lane_qualifiers): New.
1298 (TYPES_STORESTRUCT_LANE): Likewise.
1299 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
1300 (st3_lane): Likewise.
1301 (st4_lane): Likewise.
1302 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
1303 (vec_store_lanesci_lane<mode>): Likewise.
1304 (vec_store_lanesxi_lane<mode>): Likewise.
1305 (aarch64_st2_lane<VQ:mode>): Likewise.
1306 (aarch64_st3_lane<VQ:mode>): Likewise.
1307 (aarch64_st4_lane<VQ:mode>): Likewise.
1308 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
1309 * config/aarch64/arm_neon.h
1310 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
1311 use new macro arguments.
1312 (__ST3_LANE_FUNC): Likewise.
1313 (__ST4_LANE_FUNC): Likewise.
1314 * config/aarch64/iterators.md (V_TWO_ELEM): New.
1315 (V_THREE_ELEM): Likewise.
1316 (V_FOUR_ELEM): Likewise.
1318 2014-04-28 David Malcolm <dmalcolm@redhat.com>
1320 * doc/gimple.texi: Replace the description of the now-defunct
1321 union gimple_statement_d with a diagram showing the
1322 gimple_statement_base class hierarchy and its relationships to
1323 the GSS_ and GIMPLE_ enums.
1325 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1327 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
1328 * config/aarch64/aarch64.c
1329 (aarch64_cannot_change_mode_class): Weaken conditions.
1330 (aarch64_modes_tieable_p): New.
1331 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
1333 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
1335 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
1336 (loadsync_<mode>): Change mode.
1337 (load_quadpti, store_quadpti): New.
1338 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
1339 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
1341 2014-04-28 Martin Jambor <mjambor@suse.cz>
1343 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
1344 same alias type as the original statement.
1345 (subreplacement_assignment_data): New type.
1346 (handle_unscalarized_data_in_subtree): New type of parameter,
1347 generate new memory accesses with same alias type as the original
1349 (load_assign_lhs_subreplacements): Likewise.
1350 (sra_modify_constructor_assign): Generate new memory accesses with
1351 same alias type as the original statement.
1353 2014-04-28 Richard Biener <rguenther@suse.de>
1355 * tree-pass.h (TODO_verify_il): Define.
1356 (TODO_verify_all): Complete properly.
1357 * passes.c (execute_function_todo): Move existing loop-closed
1358 SSA verification under TODO_verify_il.
1359 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
1360 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
1361 Fix tree sharing issue.
1363 2014-04-28 Richard Biener <rguenther@suse.de>
1366 * builtins.def (DEF_C11_BUILTIN): Add.
1367 (BUILT_IN_ALIGNED_ALLOC): Likewise.
1368 * coretypes.h (enum function_class): Add function_c11_misc.
1369 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1370 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
1371 (call_may_clobber_ref_p_1): Likewise.
1372 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
1373 (mark_all_reaching_defs_necessary_1): Likewise.
1374 (propagate_necessity): Likewise.
1375 (eliminate_unnecessary_stmts): Likewise.
1376 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
1378 2014-04-28 Richard Biener <rguenther@suse.de>
1380 * tree-vrp.c (vrp_var_may_overflow): Remove.
1381 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
1382 with overflow immediately bump to one before that value and
1383 let iteration figure out overflow status.
1385 2014-04-28 Richard Biener <rguenther@suse.de>
1387 * configure.ac: Do valgrind header checks unconditionally.
1388 Add --enable-valgrind-annotations.
1389 * system.h: Guard valgrind header inclusion with
1390 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
1391 * alloc-pool.c (pool_alloc, pool_free): Use
1392 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
1393 to guard possibly dead code.
1394 * config.in: Regenerated.
1395 * configure: Likewise.
1397 2014-04-28 Jeff Law <law@redhat.com>
1399 PR tree-optimization/60902
1400 * tree-ssa-threadedge.c
1401 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
1402 over real defs when invalidating outputs from statements that do not
1403 produce useful outputs for threading.
1405 2014-04-28 Richard Biener <rguenther@suse.de>
1407 PR tree-optimization/60979
1408 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
1409 SCOPs that end in a block with a successor with abnormal
1412 2014-04-28 Richard Biener <rguenther@suse.de>
1414 * tree-pass.h (execute_pass_list): Adjust prototype.
1415 * passes.c (pass_manager::execute_early_local_passes):
1417 (do_per_function): Change callback signature, push all actual
1418 work to the callbals.
1419 (do_per_function_toporder): Likewise.
1420 (execute_function_dump): Adjust.
1421 (execute_function_todo): Likewise.
1422 (clear_last_verified): Likewise.
1423 (verify_curr_properties): Likewise.
1424 (update_properties_after_pass): Likewise.
1425 (execute_pass_list_1): Split out from ...
1426 (execute_pass_list): ... here. Adjust.
1427 (execute_ipa_pass_list): Likewise.
1428 * cgraphunit.c (cgraph_add_new_function): Adjust.
1429 (analyze_function): Likewise.
1430 (expand_function): Likewise.
1431 * cgraph.c (release_function_body): Free dominance info
1432 here instead of asserting it was magically freed elsewhere.
1434 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
1436 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
1437 * configure: Regenerate.
1438 * config/sparc/sparc.opt (muser-mode): New option.
1439 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
1441 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
1442 * doc/invoke.texi (SPARC options): Document -muser-mode.
1444 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
1446 * cselib.c (find_slot_memmode): Delete.
1447 (cselib_hasher): Change compare_type to a struct.
1448 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
1450 (preserve_constants_and_equivs): Adjust for new compare_type.
1451 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
1452 (wrap_constant): Delete.
1453 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
1455 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
1457 * doc/install.texi (Building with profile feedback): Remove
1460 2014-04-26 Tom de Vries <tom@codesourcery.com>
1462 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
1465 2014-04-25 Cary Coutant <ccoutant@google.com>
1468 * dwarf2out.c (should_move_die_to_comdat): A type definition
1469 can contain a subprogram definition, but don't move it to a
1471 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
1472 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
1474 (clone_tree_hash): Rename to...
1475 (clone_tree_partial): ...this; change callers. Copy
1476 DW_TAG_subprogram DIEs as declarations.
1477 (copy_decls_walk): Don't copy children of a declaration into a
1480 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
1483 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
1486 2014-04-25 Jiong Wang <jiong.wang@arm.com>
1488 * config/arm/predicates.md (call_insn_operand): Add long_call check.
1489 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
1491 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
1494 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1496 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
1499 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1501 PR tree-optimization/60930
1502 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
1503 creating a multiply candidate by folding two constant
1504 multiplicands when the result overflows.
1506 2014-04-25 Jakub Jelinek <jakub@redhat.com>
1508 PR tree-optimization/60960
1509 * tree-vect-generic.c (expand_vector_operation): Only call
1510 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
1512 2014-04-25 Tom de Vries <tom@codesourcery.com>
1514 * expr.c (clobber_reg_mode): New function.
1515 * expr.h (clobber_reg): New function.
1517 2014-04-25 Tom de Vries <tom@codesourcery.com>
1519 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
1522 2014-04-25 Radovan Obradovic <robradovic@mips.com>
1523 Tom de Vries <tom@codesourcery.com>
1525 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
1527 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
1528 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
1529 new argument to find_all_hard_reg_sets call.
1531 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1533 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
1534 Use HOST_WIDE_INT_C for mask literal.
1535 (aarch_rev16_shleft_mask_imm_p): Likewise.
1537 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
1540 * config/sparc/sparc.md (ashlsi3_extend): Delete.
1542 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
1544 PR preprocessor/56540
1545 * config/i386/i386-c.c (ix86_target_macros): Define
1546 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
1548 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1550 * configure.ac (tga_func): Remove.
1551 (LIB_TLS_SPEC): Remove.
1552 * configure: Regenerate.
1553 * config.in: Regenerate.
1554 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
1556 2014-04-25 Richard Biener <rguenther@suse.de>
1559 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
1560 call stmt use/clobber sets during stmt walk instead of
1561 walking the possibly incomplete set of caller edges.
1563 2014-04-25 Richard Biener <rguenther@suse.de>
1566 * passes.c (apply_ipa_transforms): Inline into only caller ...
1567 (execute_one_pass): ... here. Properly bring in function
1568 bodies for nodes we want to apply IPA transforms to.
1570 2014-04-24 Cong Hou <congh@google.com>
1572 PR tree-optimization/60896
1573 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
1574 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
1575 (vect_mark_pattern_stmts): Set the def type of all statements in
1576 PATTERN_DEF_SEQ as vect_internal_def.
1578 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
1580 * doc/extend.texi (PowerPC Built-in Functions): Document new
1581 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
1583 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
1585 * config/rs6000/predicates.md (const_0_to_3_operand): New
1586 predicate to match 0..3 integer constants.
1588 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
1589 to support adding miscellaneous builtin functions.
1590 (BU_DFP_MISC_2): Likewise.
1591 (BU_P7_MISC_1): Likewise.
1592 (BU_P7_MISC_2): Likewise.
1593 (BU_P8V_MISC_3): Likewise.
1594 (BU_MISC_1): Likewise.
1595 (BU_MISC_2): Likewise.
1596 (DIVWE): Add extended divide builtin functions.
1599 (DIVWEUO): Likewise.
1603 (DIVDEUO): Likewise.
1604 (DXEX): Add decimal floating-point builtin functions.
1607 (DDEDPDQ): Likewise.
1609 (DENBCDQ): Likewise.
1616 (CDTBCD): Add new BCD builtin functions.
1620 (BCDADD_LT): Likewise.
1621 (BCDADD_EQ): Likewise.
1622 (BCDADD_GT): Likewise.
1623 (BCDADD_OV): Likewise.
1625 (BCDSUB_LT): Likewise.
1626 (BCDSUB_EQ): Likewise.
1627 (BCDSUB_GT): Likewise.
1628 (BCDSUB_OV): Likewise.
1629 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
1630 (UNPACK_TD): Likewise.
1631 (PACK_TF): Likewise.
1632 (UNPACK_TF): Likewise.
1633 (UNPACK_TF_0): Likewise.
1634 (UNPACK_TF_1): Likewise.
1635 (PACK_V1TI): Likewise.
1636 (UNPACK_V1TI): Likewise.
1638 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
1639 support for decimal floating point builtin functions.
1640 (rs6000_expand_ternop_builtin): Add checks for the new builtin
1641 functions that take constant arguments.
1642 (rs6000_invalid_builtin): Add decimal floating point builtin
1644 (rs6000_init_builtins): Setup long double, _Decimal64, and
1645 _Decimal128 types for new builtin functions.
1646 (builtin_function_type): Set the unsigned flags appropriately for
1647 the new builtin functions.
1648 (rs6000_opt_masks): Add support for decimal floating point builtin
1651 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
1652 floating point builtin functions.
1653 (RS6000_BTM_COMMON): Likewise.
1654 (RS6000_BTI_long_double): Likewise.
1655 (RS6000_BTI_dfloat64): Likewise.
1656 (RS6000_BTI_dfloat128): Likewise.
1657 (long_double_type_internal_node): Likewise.
1658 (dfloat64_type_internal_node): Likewise.
1659 (dfloat128_type_internal_node): Likewise.
1661 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
1662 2.07 bcd arithmetic instructions.
1663 (UNSPEC_BCDSUB): Likewise.
1664 (UNSPEC_BCD_OVERFLOW): Likewise.
1665 (UNSPEC_BCD_ADD_SUB): Likewise.
1666 (bcd_add_sub): Likewise.
1667 (BCD_TEST): Likewise.
1668 (bcd<bcd_add_sub>): Likewise.
1669 (bcd<bcd_add_sub>_test): Likewise.
1670 (bcd<bcd_add_sub>_test2): Likewise.
1671 (bcd<bcd_add_sub>_<code>): Likewise.
1672 (peephole2 for combined bcd ops): Likewise.
1674 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
1675 decimal floating point builtin functions.
1676 (UNSPEC_DENBCD): Likewise.
1677 (UNSPEC_DXEX): Likewise.
1678 (UNSPEC_DIEX): Likewise.
1679 (UNSPEC_DSCLI): Likewise.
1680 (UNSPEC_DSCRI): Likewise.
1681 (D64_D128): Likewise.
1682 (dfp_suffix): Likewise.
1683 (dfp_ddedpd_<mode>): Likewise.
1684 (dfp_denbcd_<mode>): Likewise.
1685 (dfp_dxex_<mode>): Likewise.
1686 (dfp_diex_<mode>): Likewise.
1687 (dfp_dscli_<mode>): Likewise.
1688 (dfp_dscri_<mode>): Likewise.
1690 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
1692 (UNSPEC_CDTBCD): Likewise.
1693 (UNSPEC_CBCDTD): Likewise.
1694 (UNSPEC_DIVE): Add support for new extended divide builtin
1696 (UNSPEC_DIVEO): Likewise.
1697 (UNSPEC_DIVEU): Likewise.
1698 (UNSPEC_DIVEUO): Likewise.
1699 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
1700 pack/unpack 128-bit types.
1701 (UNSPEC_PACK_128BIT): Likewise.
1702 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
1703 (udiv<mode>3): Use idiv_ldiv mode attribute.
1704 (div<mode>3): Likewise.
1705 (addg6s): Add new BCD builtin functions.
1708 (UNSPEC_DIV_EXTEND): Add support for new extended divide
1710 (div_extend): Likewise.
1711 (div<div_extend>_<mode>"): Likewise.
1712 (FP128_64): Add support for new builtin functions to pack/unpack
1714 (unpack<mode>): Likewise.
1715 (unpacktf_0): Likewise.
1716 (unpacktf_1): Likewise.
1717 (unpack<mode>_dm): Likewise.
1718 (unpack<mode>_nodm): Likewise.
1719 (pack<mode>): Likewise.
1720 (unpackv1ti): Likewise.
1721 (packv1ti): Likewise.
1723 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
1725 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
1728 2014-04-24 Jakub Jelinek <jakub@redhat.com>
1730 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
1731 * gimplify.c (omp_is_private): Change last argument's type to int.
1732 Only diagnose lastprivate if the simd argument is 1, only diagnose
1733 linear if the simd argument is 2.
1734 (gimplify_omp_for): Adjust omp_is_private callers. When adding
1735 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
1736 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
1737 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
1738 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
1739 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
1740 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
1741 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
1742 * tree-nested.c (convert_nonlocal_omp_clauses,
1743 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
1745 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
1748 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
1751 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
1753 * flag-types.h (enum ivar_visibility): Add.
1755 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
1757 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
1758 function * argument.
1760 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
1762 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
1764 2014-04-24 Radovan Obradovic <robradovic@mips.com>
1765 Tom de Vries <tom@codesourcery.com>
1767 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
1768 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
1770 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
1771 * emit-rtl.c (try_split): Same.
1773 2014-04-24 Radovan Obradovic <robradovic@mips.com>
1774 Tom de Vries <tom@codesourcery.com>
1776 * common.opt (fuse-caller-save): New option.
1778 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
1780 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
1783 2014-04-24 Richard Biener <rguenther@suse.de>
1785 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
1786 during TER and instead use the sepops interface for expanding
1787 non-GIMPLE_SINGLE_RHS.
1789 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1791 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
1792 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
1794 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1796 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
1797 assembler 64-bit option.
1798 * configure: Regenerate.
1800 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1802 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
1803 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
1804 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
1805 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
1806 (TARGET_CRYPTO): Take TARGET_SIMD into account.
1808 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1810 * config/aarch64/aarch64-builtins.c
1811 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
1812 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
1813 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
1814 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
1816 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
1817 (Vrevsuff): New mode attribute.
1819 2014-04-24 Terry Guo <terry.guo@arm.com>
1821 * config/arm/arm.h (machine_function): Define variable
1822 after_arm_reorg here.
1823 * config/arm/arm.c (after_arm_reorg): Remove the definition.
1824 (arm_split_constant): Update the way to access variable
1827 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
1829 2014-04-23 Tom de Vries <tom@codesourcery.com>
1831 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
1833 2014-04-23 David Malcolm <dmalcolm@redhat.com>
1835 * is-a.h: Update comments to reflect the following changes to the
1836 "pointerness" of the API, making the template parameter match the
1837 return type, allowing use of is-a.h with typedefs of pointers.
1838 (is_a_helper::cast): Return a T rather then a pointer to a T, so
1839 that the return type matches the parameter to the is_a_helper.
1841 (dyn_cast): Likewise.
1843 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
1844 pointer from the is-a.h API.
1846 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
1847 (is_a_helper <cgraph_node *>::test): ...this, matching change to
1849 (is_a_helper <varpool_node>::test): Likewise, convert to...
1850 (is_a_helper <varpool_node *>::test): ...this.
1852 (varpool_first_variable): Update for removal of implicit pointer
1853 from the is-a.h API.
1854 (varpool_next_variable): Likewise.
1855 (varpool_first_static_initializer): Likewise.
1856 (varpool_next_static_initializer): Likewise.
1857 (varpool_first_defined_variable): Likewise.
1858 (varpool_next_defined_variable): Likewise.
1859 (cgraph_first_defined_function): Likewise.
1860 (cgraph_next_defined_function): Likewise.
1861 (cgraph_first_function): Likewise.
1862 (cgraph_next_function): Likewise.
1863 (cgraph_first_function_with_gimple_body): Likewise.
1864 (cgraph_next_function_with_gimple_body): Likewise.
1865 (cgraph_alias_target): Likewise.
1866 (varpool_alias_target): Likewise.
1867 (cgraph_function_or_thunk_node): Likewise.
1868 (varpool_variable_node): Likewise.
1869 (symtab_real_symbol_p): Likewise.
1870 * cgraphunit.c (referred_to_p): Likewise.
1871 (analyze_functions): Likewise.
1872 (handle_alias_pairs): Likewise.
1873 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1874 * gimple-ssa.h (gimple_vuse_op): Likewise.
1875 (gimple_vdef_op): Likewise.
1876 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
1877 * gimple.c (gimple_build_asm_1): Likewise.
1878 (gimple_build_try): Likewise.
1879 (gimple_build_resx): Likewise.
1880 (gimple_build_eh_dispatch): Likewise.
1881 (gimple_build_omp_for): Likewise.
1882 (gimple_omp_for_set_clauses): Likewise.
1884 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
1885 (is_a_helper <gimple_statement_asm *>::test): ...this.
1886 (is_a_helper <gimple_statement_bind>::test): Convert to...
1887 (is_a_helper <gimple_statement_bind *>::test): ...this.
1888 (is_a_helper <gimple_statement_call>::test): Convert to...
1889 (is_a_helper <gimple_statement_call *>::test): ...this.
1890 (is_a_helper <gimple_statement_catch>::test): Convert to...
1891 (is_a_helper <gimple_statement_catch *>::test): ...this.
1892 (is_a_helper <gimple_statement_resx>::test): Convert to...
1893 (is_a_helper <gimple_statement_resx *>::test): ...this.
1894 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
1895 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
1896 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
1897 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
1898 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
1899 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
1900 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
1901 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
1902 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
1903 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
1904 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
1905 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
1906 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
1907 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
1908 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
1909 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
1910 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
1911 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
1912 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
1913 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
1914 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
1915 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
1916 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
1917 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
1918 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
1919 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
1920 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
1921 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
1922 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
1923 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
1924 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
1925 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
1926 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
1927 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
1928 (is_a_helper <gimple_statement_phi>::test): Convert to...
1929 (is_a_helper <gimple_statement_phi *>::test): ...this.
1930 (is_a_helper <gimple_statement_transaction>::test): Convert to...
1931 (is_a_helper <gimple_statement_transaction *>::test): ...this.
1932 (is_a_helper <gimple_statement_try>::test): Convert to...
1933 (is_a_helper <gimple_statement_try *>::test): ...this.
1934 (is_a_helper <gimple_statement_wce>::test): Convert to...
1935 (is_a_helper <gimple_statement_wce *>::test): ...this.
1936 (is_a_helper <const gimple_statement_asm>::test): Convert to...
1937 (is_a_helper <const gimple_statement_asm *>::test): ...this.
1938 (is_a_helper <const gimple_statement_bind>::test): Convert to...
1939 (is_a_helper <const gimple_statement_bind *>::test): ...this.
1940 (is_a_helper <const gimple_statement_call>::test): Convert to...
1941 (is_a_helper <const gimple_statement_call *>::test): ...this.
1942 (is_a_helper <const gimple_statement_catch>::test): Convert to...
1943 (is_a_helper <const gimple_statement_catch *>::test): ...this.
1944 (is_a_helper <const gimple_statement_resx>::test): Convert to...
1945 (is_a_helper <const gimple_statement_resx *>::test): ...this.
1946 (is_a_helper <const gimple_statement_eh_dispatch>::test):
1948 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
1950 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
1952 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
1953 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
1955 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
1957 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
1959 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
1961 (is_a_helper <const gimple_statement_omp_return>::test): Convert
1963 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
1964 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
1966 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
1967 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
1969 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
1970 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
1971 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
1972 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
1974 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
1975 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
1977 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
1978 (is_a_helper <const gimple_statement_omp_target>::test): Convert
1980 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
1981 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
1983 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
1984 (is_a_helper <const gimple_statement_omp_single>::test): Convert
1986 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
1987 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
1989 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
1990 (is_a_helper <const gimple_statement_omp_task>::test): Convert
1992 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
1993 (is_a_helper <const gimple_statement_phi>::test): Convert to...
1994 (is_a_helper <const gimple_statement_phi *>::test): ...this.
1995 (is_a_helper <const gimple_statement_transaction>::test): Convert
1997 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
1998 (is_a_helper <const gimple_statement_with_ops>::test): Convert
2000 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
2001 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
2002 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
2003 (is_a_helper <const gimple_statement_with_memory_ops>::test):
2005 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
2007 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
2009 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
2011 (gimple_use_ops): Update for removal of implicit pointer from the
2013 (gimple_set_use_ops): Likewise.
2014 (gimple_vuse): Likewise.
2015 (gimple_vdef): Likewise.
2016 (gimple_vuse_ptr): Likewise.
2017 (gimple_vdef_ptr): Likewise.
2018 (gimple_set_vuse): Likewise.
2019 (gimple_set_vdef): Likewise.
2020 (gimple_omp_return_set_lhs): Likewise.
2021 (gimple_omp_return_lhs): Likewise.
2022 (gimple_omp_return_lhs_ptr): Likewise.
2023 (gimple_call_fntype): Likewise.
2024 (gimple_call_set_fntype): Likewise.
2025 (gimple_call_set_internal_fn): Likewise.
2026 (gimple_call_use_set): Likewise.
2027 (gimple_call_clobber_set): Likewise.
2028 (gimple_bind_vars): Likewise.
2029 (gimple_bind_set_vars): Likewise.
2030 (gimple_bind_body_ptr): Likewise.
2031 (gimple_bind_set_body): Likewise.
2032 (gimple_bind_add_stmt): Likewise.
2033 (gimple_bind_block): Likewise.
2034 (gimple_bind_set_block): Likewise.
2035 (gimple_asm_ninputs): Likewise.
2036 (gimple_asm_noutputs): Likewise.
2037 (gimple_asm_nclobbers): Likewise.
2038 (gimple_asm_nlabels): Likewise.
2039 (gimple_asm_input_op): Likewise.
2040 (gimple_asm_input_op_ptr): Likewise.
2041 (gimple_asm_output_op): Likewise.
2042 (gimple_asm_output_op_ptr): Likewise.
2043 (gimple_asm_set_output_op): Likewise.
2044 (gimple_asm_clobber_op): Likewise.
2045 (gimple_asm_set_clobber_op): Likewise.
2046 (gimple_asm_label_op): Likewise.
2047 (gimple_asm_set_label_op): Likewise.
2048 (gimple_asm_string): Likewise.
2049 (gimple_catch_types): Likewise.
2050 (gimple_catch_types_ptr): Likewise.
2051 (gimple_catch_handler_ptr): Likewise.
2052 (gimple_catch_set_types): Likewise.
2053 (gimple_catch_set_handler): Likewise.
2054 (gimple_eh_filter_types): Likewise.
2055 (gimple_eh_filter_types_ptr): Likewise.
2056 (gimple_eh_filter_failure_ptr): Likewise.
2057 (gimple_eh_filter_set_types): Likewise.
2058 (gimple_eh_filter_set_failure): Likewise.
2059 (gimple_eh_must_not_throw_fndecl): Likewise.
2060 (gimple_eh_must_not_throw_set_fndecl): Likewise.
2061 (gimple_eh_else_n_body_ptr): Likewise.
2062 (gimple_eh_else_e_body_ptr): Likewise.
2063 (gimple_eh_else_set_n_body): Likewise.
2064 (gimple_eh_else_set_e_body): Likewise.
2065 (gimple_try_eval_ptr): Likewise.
2066 (gimple_try_cleanup_ptr): Likewise.
2067 (gimple_try_set_eval): Likewise.
2068 (gimple_try_set_cleanup): Likewise.
2069 (gimple_wce_cleanup_ptr): Likewise.
2070 (gimple_wce_set_cleanup): Likewise.
2071 (gimple_phi_capacity): Likewise.
2072 (gimple_phi_num_args): Likewise.
2073 (gimple_phi_result): Likewise.
2074 (gimple_phi_result_ptr): Likewise.
2075 (gimple_phi_set_result): Likewise.
2076 (gimple_phi_arg): Likewise.
2077 (gimple_phi_set_arg): Likewise.
2078 (gimple_resx_region): Likewise.
2079 (gimple_resx_set_region): Likewise.
2080 (gimple_eh_dispatch_region): Likewise.
2081 (gimple_eh_dispatch_set_region): Likewise.
2082 (gimple_omp_critical_name): Likewise.
2083 (gimple_omp_critical_name_ptr): Likewise.
2084 (gimple_omp_critical_set_name): Likewise.
2085 (gimple_omp_for_clauses): Likewise.
2086 (gimple_omp_for_clauses_ptr): Likewise.
2087 (gimple_omp_for_set_clauses): Likewise.
2088 (gimple_omp_for_collapse): Likewise.
2089 (gimple_omp_for_index): Likewise.
2090 (gimple_omp_for_index_ptr): Likewise.
2091 (gimple_omp_for_set_index): Likewise.
2092 (gimple_omp_for_initial): Likewise.
2093 (gimple_omp_for_initial_ptr): Likewise.
2094 (gimple_omp_for_set_initial): Likewise.
2095 (gimple_omp_for_final): Likewise.
2096 (gimple_omp_for_final_ptr): Likewise.
2097 (gimple_omp_for_set_final): Likewise.
2098 (gimple_omp_for_incr): Likewise.
2099 (gimple_omp_for_incr_ptr): Likewise.
2100 (gimple_omp_for_set_incr): Likewise.
2101 (gimple_omp_for_pre_body_ptr): Likewise.
2102 (gimple_omp_for_set_pre_body): Likewise.
2103 (gimple_omp_parallel_clauses): Likewise.
2104 (gimple_omp_parallel_clauses_ptr): Likewise.
2105 (gimple_omp_parallel_set_clauses): Likewise.
2106 (gimple_omp_parallel_child_fn): Likewise.
2107 (gimple_omp_parallel_child_fn_ptr): Likewise.
2108 (gimple_omp_parallel_set_child_fn): Likewise.
2109 (gimple_omp_parallel_data_arg): Likewise.
2110 (gimple_omp_parallel_data_arg_ptr): Likewise.
2111 (gimple_omp_parallel_set_data_arg): Likewise.
2112 (gimple_omp_task_clauses): Likewise.
2113 (gimple_omp_task_clauses_ptr): Likewise.
2114 (gimple_omp_task_set_clauses): Likewise.
2115 (gimple_omp_task_child_fn): Likewise.
2116 (gimple_omp_task_child_fn_ptr): Likewise.
2117 (gimple_omp_task_set_child_fn): Likewise.
2118 (gimple_omp_task_data_arg): Likewise.
2119 (gimple_omp_task_data_arg_ptr): Likewise.
2120 (gimple_omp_task_set_data_arg): Likewise.
2121 (gimple_omp_taskreg_clauses): Likewise.
2122 (gimple_omp_taskreg_clauses_ptr): Likewise.
2123 (gimple_omp_taskreg_set_clauses): Likewise.
2124 (gimple_omp_taskreg_child_fn): Likewise.
2125 (gimple_omp_taskreg_child_fn_ptr): Likewise.
2126 (gimple_omp_taskreg_set_child_fn): Likewise.
2127 (gimple_omp_taskreg_data_arg): Likewise.
2128 (gimple_omp_taskreg_data_arg_ptr): Likewise.
2129 (gimple_omp_taskreg_set_data_arg): Likewise.
2130 (gimple_omp_task_copy_fn): Likewise.
2131 (gimple_omp_task_copy_fn_ptr): Likewise.
2132 (gimple_omp_task_set_copy_fn): Likewise.
2133 (gimple_omp_task_arg_size): Likewise.
2134 (gimple_omp_task_arg_size_ptr): Likewise.
2135 (gimple_omp_task_set_arg_size): Likewise.
2136 (gimple_omp_task_arg_align): Likewise.
2137 (gimple_omp_task_arg_align_ptr): Likewise.
2138 (gimple_omp_task_set_arg_align): Likewise.
2139 (gimple_omp_single_clauses): Likewise.
2140 (gimple_omp_single_clauses_ptr): Likewise.
2141 (gimple_omp_single_set_clauses): Likewise.
2142 (gimple_omp_target_clauses): Likewise.
2143 (gimple_omp_target_clauses_ptr): Likewise.
2144 (gimple_omp_target_set_clauses): Likewise.
2145 (gimple_omp_target_child_fn): Likewise.
2146 (gimple_omp_target_child_fn_ptr): Likewise.
2147 (gimple_omp_target_set_child_fn): Likewise.
2148 (gimple_omp_target_data_arg): Likewise.
2149 (gimple_omp_target_data_arg_ptr): Likewise.
2150 (gimple_omp_target_set_data_arg): Likewise.
2151 (gimple_omp_teams_clauses): Likewise.
2152 (gimple_omp_teams_clauses_ptr): Likewise.
2153 (gimple_omp_teams_set_clauses): Likewise.
2154 (gimple_omp_sections_clauses): Likewise.
2155 (gimple_omp_sections_clauses_ptr): Likewise.
2156 (gimple_omp_sections_set_clauses): Likewise.
2157 (gimple_omp_sections_control): Likewise.
2158 (gimple_omp_sections_control_ptr): Likewise.
2159 (gimple_omp_sections_set_control): Likewise.
2160 (gimple_omp_for_set_cond): Likewise.
2161 (gimple_omp_for_cond): Likewise.
2162 (gimple_omp_atomic_store_set_val): Likewise.
2163 (gimple_omp_atomic_store_val): Likewise.
2164 (gimple_omp_atomic_store_val_ptr): Likewise.
2165 (gimple_omp_atomic_load_set_lhs): Likewise.
2166 (gimple_omp_atomic_load_lhs): Likewise.
2167 (gimple_omp_atomic_load_lhs_ptr): Likewise.
2168 (gimple_omp_atomic_load_set_rhs): Likewise.
2169 (gimple_omp_atomic_load_rhs): Likewise.
2170 (gimple_omp_atomic_load_rhs_ptr): Likewise.
2171 (gimple_omp_continue_control_def): Likewise.
2172 (gimple_omp_continue_control_def_ptr): Likewise.
2173 (gimple_omp_continue_set_control_def): Likewise.
2174 (gimple_omp_continue_control_use): Likewise.
2175 (gimple_omp_continue_control_use_ptr): Likewise.
2176 (gimple_omp_continue_set_control_use): Likewise.
2177 (gimple_transaction_body_ptr): Likewise.
2178 (gimple_transaction_label): Likewise.
2179 (gimple_transaction_label_ptr): Likewise.
2180 (gimple_transaction_set_body): Likewise.
2181 (gimple_transaction_set_label): Likewise.
2183 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
2184 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2185 * ipa-ref.c (ipa_record_reference): Likewise.
2186 * ipa-reference.c (analyze_function): Likewise.
2187 (ipa_reference_write_optimization_summary): Likewise.
2188 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
2189 (address_taken_from_non_vtable_p): Likewise.
2190 (comdat_can_be_unshared_p_1): Likewise.
2191 * lto-cgraph.c (lto_output_ref): Likewise.
2192 (add_references): Likewise.
2193 (compute_ltrans_boundary): Likewise.
2194 (output_symtab): Likewise.
2195 (input_ref): Likewise.
2196 (input_cgraph_1): Likewise.
2197 (output_cgraph_opt_summary): Likewise.
2198 * lto-streamer-out.c (lto_output): Likewise.
2199 (output_symbol_p): Likewise.
2200 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
2201 (lsei_start_function_in_partition): Likewise.
2202 (lsei_next_variable_in_partition): Likewise.
2203 (lsei_start_variable_in_partition): Likewise.
2204 * symtab.c (insert_to_assembler_name_hash): Likewise.
2205 (unlink_from_assembler_name_hash): Likewise.
2206 (symtab_unregister_node): Likewise.
2207 (symtab_remove_node): Likewise.
2208 (dump_symtab_node): Likewise.
2209 (verify_symtab_base): Likewise.
2210 (verify_symtab_node): Likewise.
2211 (symtab_make_decl_local): Likewise.
2212 (symtab_alias_ultimate_target): Likewise.
2213 (symtab_resolve_alias): Likewise.
2214 (symtab_get_symbol_partitioning_class): Likewise.
2215 * tree-phinodes.c (allocate_phi_node): Likewise.
2216 (reserve_phi_args_for_new_edge): Likewise.
2217 (remove_phi_args): Likewise.
2218 * varpool.c (varpool_node_for_asm): Likewise.
2219 (varpool_remove_unreferenced_decls): Likewise.
2221 2014-04-23 Jeff Law <law@redhat.com>
2223 PR tree-optimization/60902
2224 * tree-ssa-threadedge.c
2225 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
2226 invalidate outputs from statements that do not produce useful
2227 outputs for threading.
2229 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2231 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
2232 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
2233 machine descriptions for Stack Smashing Protector.
2235 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
2237 * aarch64.md (<optab>_rol<mode>3): New pattern.
2238 (<optab>_rolsi3_uxtw): Likewise.
2239 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
2241 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
2243 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
2244 (arm_cortex_a12_tune): Likewise.
2246 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2248 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
2250 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2252 * config/arm/arm.md (arm_rev16si2): New pattern.
2253 (arm_rev16si2_alt): Likewise.
2254 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
2256 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2258 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
2259 (rev16<mode>2_alt): Likewise.
2260 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
2261 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
2262 (aarch_rev16_shleft_mask_imm_p): Likewise.
2263 (aarch_rev16_p_1): Likewise.
2264 (aarch_rev16_p): Likewise.
2265 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
2266 (aarch_rev16_shright_mask_imm_p): Likewise.
2267 (aarch_rev16_shleft_mask_imm_p): Likewise.
2269 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2271 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
2272 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
2274 (cortex_a53_extra_costs): Likewise.
2275 (cortex_a57_extra_costs): Likewise.
2276 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
2277 (cortexa7_extra_costs): Likewise.
2278 (cortexa8_extra_costs): Likewise.
2279 (cortexa12_extra_costs): Likewise.
2280 (cortexa15_extra_costs): Likewise.
2281 (v7m_extra_costs): Likewise.
2282 (arm_new_rtx_costs): Handle BSWAP.
2284 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2286 * config/arm/arm.c (cortexa8_extra_costs): New table.
2287 (arm_cortex_a8_tune): New tuning struct.
2288 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
2290 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2292 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
2294 2014-04-23 Richard Biener <rguenther@suse.de>
2296 * Makefile.in (OBJS): Remove loop-unswitch.o.
2297 * loop-unswitch.c: Delete.
2298 * tree-pass.h (make_pass_rtl_unswitch): Remove.
2299 * passes.def (pass_rtl_unswitch): Likewise.
2300 * loop-init.c (gate_rtl_unswitch): Likewise.
2301 (rtl_unswitch): Likewise.
2302 (pass_data_rtl_unswitch): Likewise.
2303 (pass_rtl_unswitch): Likewise.
2304 (make_pass_rtl_unswitch): Likewise.
2305 * rtl.h (reversed_condition): Likewise.
2306 (compare_and_jump_seq): Likewise.
2307 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
2309 * loop-unroll.c (compare_and_jump_seq): Likewise.
2311 2014-04-23 Richard Biener <rguenther@suse.de>
2313 PR tree-optimization/60903
2314 * tree-ssa-loop-im.c (analyze_memory_references): Remove
2315 commented code block.
2316 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
2317 loop flags to newly created BBs and edges.
2319 2014-04-23 Nick Clifton <nickc@redhat.com>
2321 * config/msp430/msp430.c (msp430_handle_option): Move function
2323 (msp430_option_override): Simplify mcu and mcpu option handling.
2324 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
2325 support for -mhwmult command line option.
2326 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
2327 -mhwmult command line option.
2328 (msp430_hwmult_enabled): Delete.
2329 (msp43o_output_labelref): Add support for -mhwmult command line
2331 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
2332 (umulsidi3): Likewise.
2333 * config/msp430/msp430.opt (mmcu): Add Report attribute.
2334 (mcpu, mlarge, msmall): Likewise.
2335 (mhwmult): New option.
2336 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
2338 (msp430_is_f5_mcu): Remove prototype.
2339 (msp430_use_f5_series_hwmult): Add prototype.
2340 * config/msp430/msp430-opts.h: New file.
2341 * common/config/msp430: New directory.
2342 * common/config/msp430/msp430-common.c: New file.
2343 * config.gcc (msp430): Remove target_has_targetm_common.
2344 * doc/invoke.texi: Document -mhwmult command line option.
2346 2014-04-23 Nick Clifton <nickc@redhat.com>
2348 * config/i386/cygwin.h (ENDFILE_SPEC): Include
2349 default-manifest.o if it can be found in the search path.
2350 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
2352 2014-04-23 Terry Guo <terry.guo@arm.com>
2354 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
2356 2014-04-23 Richard Biener <rguenther@suse.de>
2359 * tree-inline.c (declare_return_variable): Use mark_addressable.
2361 2014-04-23 Richard Biener <rguenther@suse.de>
2364 * loop-init.c (loop_optimizer_init): Make sure to apply
2365 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
2367 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2370 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
2372 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
2373 if flag_sanitize_undefined_trap_on_error.
2374 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
2375 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
2376 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
2377 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
2378 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
2379 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
2380 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
2381 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
2382 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
2383 * ubsan.c (ubsan_instrument_unreachable): Return
2384 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
2385 (ubsan_expand_null_ifn): Emit __builtin_trap ()
2386 if flag_sanitize_undefined_trap_on_error and
2387 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
2388 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
2389 instrument_bool_enum_load): Emit __builtin_trap () if
2390 flag_sanitize_undefined_trap_on_error and
2391 __builtin_handle_*_abort () if !flag_sanitize_recover.
2392 * doc/invoke.texi (-fsanitize-recover,
2393 -fsanitize-undefined-trap-on-error): Document.
2395 2014-04-22 Christian Bruel <christian.bruel@st.com>
2397 * config/sh/sh.md (mov<mode>): Replace movQIHI.
2398 Force immediates to SImode.
2400 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
2402 * config/nios2/nios2.md (UNSPEC_ROUND): New.
2404 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
2405 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
2406 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
2407 (nios2_fpu_insn): Add entry for round.
2408 (N2FPU_NO_ERRNO_P): Define.
2409 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
2411 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
2413 2014-04-22 Richard Henderson <rth@redhat.com>
2415 * config/aarch64/aarch64 (addti3, subti3): New expanders.
2416 (add<GPI>3_compare0): Remove leading * from name.
2417 (add<GPI>3_carryin): Likewise.
2418 (sub<GPI>3_compare0): Likewise.
2419 (sub<GPI>3_carryin): Likewise.
2420 (<su_optab>mulditi3): New expander.
2421 (multi3): New expander.
2422 (madd<GPI>): Remove leading * from name.
2424 2014-04-22 Martin Jambor <mjambor@suse.cz>
2426 * cgraphclones.c (cgraph_function_versioning): Copy
2427 ipa_transforms_to_apply instead of asserting it is empty.
2429 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
2432 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
2433 on count_exp to get mode.
2435 2014-04-22 Andrew Pinski <apinski@cavium.com>
2437 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
2438 Handle TLS for ILP32.
2439 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
2440 (tlsie_small_<mode>): this and handle PTR.
2441 (tlsie_small_sidi): New pattern.
2442 (tlsle_small): Change to an expand to handle ILP32.
2443 (tlsle_small_<mode>): New pattern.
2444 (tlsdesc_small): Rename to ...
2445 (tlsdesc_small_<mode>): this and handle PTR.
2447 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2449 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
2451 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2453 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2454 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
2455 (aarch64_types_signed_poly_qualifiers): Likewise.
2456 (aarch64_types_unsigned_signed_qualifiers): Likewise.
2457 (aarch64_types_poly_signed_qualifiers): Likewise.
2458 (TYPES_REINTERP_SS): Type macro added.
2459 (TYPES_REINTERP_SU): Likewise.
2460 (TYPES_REINTERP_SP): Likewise.
2461 (TYPES_REINTERP_US): Likewise.
2462 (TYPES_REINTERP_PS): Likewise.
2463 (aarch64_fold_builtin): New expression folding added.
2464 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
2465 Declarations removed.
2466 (REINTERP_SS): Declarations added.
2467 (REINTERP_US): Likewise.
2468 (REINTERP_PS): Likewise.
2469 (REINTERP_SU): Likewise.
2470 (REINTERP_SP): Likewise.
2471 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
2472 (vreinterpretq_p8_f64): Likewise.
2473 (vreinterpret_p16_f64): Likewise.
2474 (vreinterpretq_p16_f64): Likewise.
2475 (vreinterpret_f32_f64): Likewise.
2476 (vreinterpretq_f32_f64): Likewise.
2477 (vreinterpret_f64_f32): Likewise.
2478 (vreinterpret_f64_p8): Likewise.
2479 (vreinterpret_f64_p16): Likewise.
2480 (vreinterpret_f64_s8): Likewise.
2481 (vreinterpret_f64_s16): Likewise.
2482 (vreinterpret_f64_s32): Likewise.
2483 (vreinterpret_f64_s64): Likewise.
2484 (vreinterpret_f64_u8): Likewise.
2485 (vreinterpret_f64_u16): Likewise.
2486 (vreinterpret_f64_u32): Likewise.
2487 (vreinterpret_f64_u64): Likewise.
2488 (vreinterpretq_f64_f32): Likewise.
2489 (vreinterpretq_f64_p8): Likewise.
2490 (vreinterpretq_f64_p16): Likewise.
2491 (vreinterpretq_f64_s8): Likewise.
2492 (vreinterpretq_f64_s16): Likewise.
2493 (vreinterpretq_f64_s32): Likewise.
2494 (vreinterpretq_f64_s64): Likewise.
2495 (vreinterpretq_f64_u8): Likewise.
2496 (vreinterpretq_f64_u16): Likewise.
2497 (vreinterpretq_f64_u32): Likewise.
2498 (vreinterpretq_f64_u64): Likewise.
2499 (vreinterpret_s64_f64): Likewise.
2500 (vreinterpretq_s64_f64): Likewise.
2501 (vreinterpret_u64_f64): Likewise.
2502 (vreinterpretq_u64_f64): Likewise.
2503 (vreinterpret_s8_f64): Likewise.
2504 (vreinterpretq_s8_f64): Likewise.
2505 (vreinterpret_s16_f64): Likewise.
2506 (vreinterpretq_s16_f64): Likewise.
2507 (vreinterpret_s32_f64): Likewise.
2508 (vreinterpretq_s32_f64): Likewise.
2509 (vreinterpret_u8_f64): Likewise.
2510 (vreinterpretq_u8_f64): Likewise.
2511 (vreinterpret_u16_f64): Likewise.
2512 (vreinterpretq_u16_f64): Likewise.
2513 (vreinterpret_u32_f64): Likewise.
2514 (vreinterpretq_u32_f64): Likewise.
2516 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2518 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2519 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
2520 (vreinterpret_p8_s8): Likewise.
2521 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
2522 (vreinterpret_p8_s16): Likewise.
2523 (vreinterpret_p8_s32): Likewise.
2524 (vreinterpret_p8_s64): Likewise.
2525 (vreinterpret_p8_f32): Likewise.
2526 (vreinterpret_p8_u8): Likewise.
2527 (vreinterpret_p8_u16): Likewise.
2528 (vreinterpret_p8_u32): Likewise.
2529 (vreinterpret_p8_u64): Likewise.
2530 (vreinterpret_p8_p16): Likewise.
2531 (vreinterpretq_p8_s8): Likewise.
2532 (vreinterpretq_p8_s16): Likewise.
2533 (vreinterpretq_p8_s32): Likewise.
2534 (vreinterpretq_p8_s64): Likewise.
2535 (vreinterpretq_p8_f32): Likewise.
2536 (vreinterpretq_p8_u8): Likewise.
2537 (vreinterpretq_p8_u16): Likewise.
2538 (vreinterpretq_p8_u32): Likewise.
2539 (vreinterpretq_p8_u64): Likewise.
2540 (vreinterpretq_p8_p16): Likewise.
2541 (vreinterpret_p16_s8): Likewise.
2542 (vreinterpret_p16_s16): Likewise.
2543 (vreinterpret_p16_s32): Likewise.
2544 (vreinterpret_p16_s64): Likewise.
2545 (vreinterpret_p16_f32): Likewise.
2546 (vreinterpret_p16_u8): Likewise.
2547 (vreinterpret_p16_u16): Likewise.
2548 (vreinterpret_p16_u32): Likewise.
2549 (vreinterpret_p16_u64): Likewise.
2550 (vreinterpret_p16_p8): Likewise.
2551 (vreinterpretq_p16_s8): Likewise.
2552 (vreinterpretq_p16_s16): Likewise.
2553 (vreinterpretq_p16_s32): Likewise.
2554 (vreinterpretq_p16_s64): Likewise.
2555 (vreinterpretq_p16_f32): Likewise.
2556 (vreinterpretq_p16_u8): Likewise.
2557 (vreinterpretq_p16_u16): Likewise.
2558 (vreinterpretq_p16_u32): Likewise.
2559 (vreinterpretq_p16_u64): Likewise.
2560 (vreinterpretq_p16_p8): Likewise.
2561 (vreinterpret_f32_s8): Likewise.
2562 (vreinterpret_f32_s16): Likewise.
2563 (vreinterpret_f32_s32): Likewise.
2564 (vreinterpret_f32_s64): Likewise.
2565 (vreinterpret_f32_u8): Likewise.
2566 (vreinterpret_f32_u16): Likewise.
2567 (vreinterpret_f32_u32): Likewise.
2568 (vreinterpret_f32_u64): Likewise.
2569 (vreinterpret_f32_p8): Likewise.
2570 (vreinterpret_f32_p16): Likewise.
2571 (vreinterpretq_f32_s8): Likewise.
2572 (vreinterpretq_f32_s16): Likewise.
2573 (vreinterpretq_f32_s32): Likewise.
2574 (vreinterpretq_f32_s64): Likewise.
2575 (vreinterpretq_f32_u8): Likewise.
2576 (vreinterpretq_f32_u16): Likewise.
2577 (vreinterpretq_f32_u32): Likewise.
2578 (vreinterpretq_f32_u64): Likewise.
2579 (vreinterpretq_f32_p8): Likewise.
2580 (vreinterpretq_f32_p16): Likewise.
2581 (vreinterpret_s64_s8): Likewise.
2582 (vreinterpret_s64_s16): Likewise.
2583 (vreinterpret_s64_s32): Likewise.
2584 (vreinterpret_s64_f32): Likewise.
2585 (vreinterpret_s64_u8): Likewise.
2586 (vreinterpret_s64_u16): Likewise.
2587 (vreinterpret_s64_u32): Likewise.
2588 (vreinterpret_s64_u64): Likewise.
2589 (vreinterpret_s64_p8): Likewise.
2590 (vreinterpret_s64_p16): Likewise.
2591 (vreinterpretq_s64_s8): Likewise.
2592 (vreinterpretq_s64_s16): Likewise.
2593 (vreinterpretq_s64_s32): Likewise.
2594 (vreinterpretq_s64_f32): Likewise.
2595 (vreinterpretq_s64_u8): Likewise.
2596 (vreinterpretq_s64_u16): Likewise.
2597 (vreinterpretq_s64_u32): Likewise.
2598 (vreinterpretq_s64_u64): Likewise.
2599 (vreinterpretq_s64_p8): Likewise.
2600 (vreinterpretq_s64_p16): Likewise.
2601 (vreinterpret_u64_s8): Likewise.
2602 (vreinterpret_u64_s16): Likewise.
2603 (vreinterpret_u64_s32): Likewise.
2604 (vreinterpret_u64_s64): Likewise.
2605 (vreinterpret_u64_f32): Likewise.
2606 (vreinterpret_u64_u8): Likewise.
2607 (vreinterpret_u64_u16): Likewise.
2608 (vreinterpret_u64_u32): Likewise.
2609 (vreinterpret_u64_p8): Likewise.
2610 (vreinterpret_u64_p16): Likewise.
2611 (vreinterpretq_u64_s8): Likewise.
2612 (vreinterpretq_u64_s16): Likewise.
2613 (vreinterpretq_u64_s32): Likewise.
2614 (vreinterpretq_u64_s64): Likewise.
2615 (vreinterpretq_u64_f32): Likewise.
2616 (vreinterpretq_u64_u8): Likewise.
2617 (vreinterpretq_u64_u16): Likewise.
2618 (vreinterpretq_u64_u32): Likewise.
2619 (vreinterpretq_u64_p8): Likewise.
2620 (vreinterpretq_u64_p16): Likewise.
2621 (vreinterpret_s8_s16): Likewise.
2622 (vreinterpret_s8_s32): Likewise.
2623 (vreinterpret_s8_s64): Likewise.
2624 (vreinterpret_s8_f32): Likewise.
2625 (vreinterpret_s8_u8): Likewise.
2626 (vreinterpret_s8_u16): Likewise.
2627 (vreinterpret_s8_u32): Likewise.
2628 (vreinterpret_s8_u64): Likewise.
2629 (vreinterpret_s8_p8): Likewise.
2630 (vreinterpret_s8_p16): Likewise.
2631 (vreinterpretq_s8_s16): Likewise.
2632 (vreinterpretq_s8_s32): Likewise.
2633 (vreinterpretq_s8_s64): Likewise.
2634 (vreinterpretq_s8_f32): Likewise.
2635 (vreinterpretq_s8_u8): Likewise.
2636 (vreinterpretq_s8_u16): Likewise.
2637 (vreinterpretq_s8_u32): Likewise.
2638 (vreinterpretq_s8_u64): Likewise.
2639 (vreinterpretq_s8_p8): Likewise.
2640 (vreinterpretq_s8_p16): Likewise.
2641 (vreinterpret_s16_s8): Likewise.
2642 (vreinterpret_s16_s32): Likewise.
2643 (vreinterpret_s16_s64): Likewise.
2644 (vreinterpret_s16_f32): Likewise.
2645 (vreinterpret_s16_u8): Likewise.
2646 (vreinterpret_s16_u16): Likewise.
2647 (vreinterpret_s16_u32): Likewise.
2648 (vreinterpret_s16_u64): Likewise.
2649 (vreinterpret_s16_p8): Likewise.
2650 (vreinterpret_s16_p16): Likewise.
2651 (vreinterpretq_s16_s8): Likewise.
2652 (vreinterpretq_s16_s32): Likewise.
2653 (vreinterpretq_s16_s64): Likewise.
2654 (vreinterpretq_s16_f32): Likewise.
2655 (vreinterpretq_s16_u8): Likewise.
2656 (vreinterpretq_s16_u16): Likewise.
2657 (vreinterpretq_s16_u32): Likewise.
2658 (vreinterpretq_s16_u64): Likewise.
2659 (vreinterpretq_s16_p8): Likewise.
2660 (vreinterpretq_s16_p16): Likewise.
2661 (vreinterpret_s32_s8): Likewise.
2662 (vreinterpret_s32_s16): Likewise.
2663 (vreinterpret_s32_s64): Likewise.
2664 (vreinterpret_s32_f32): Likewise.
2665 (vreinterpret_s32_u8): Likewise.
2666 (vreinterpret_s32_u16): Likewise.
2667 (vreinterpret_s32_u32): Likewise.
2668 (vreinterpret_s32_u64): Likewise.
2669 (vreinterpret_s32_p8): Likewise.
2670 (vreinterpret_s32_p16): Likewise.
2671 (vreinterpretq_s32_s8): Likewise.
2672 (vreinterpretq_s32_s16): Likewise.
2673 (vreinterpretq_s32_s64): Likewise.
2674 (vreinterpretq_s32_f32): Likewise.
2675 (vreinterpretq_s32_u8): Likewise.
2676 (vreinterpretq_s32_u16): Likewise.
2677 (vreinterpretq_s32_u32): Likewise.
2678 (vreinterpretq_s32_u64): Likewise.
2679 (vreinterpretq_s32_p8): Likewise.
2680 (vreinterpretq_s32_p16): Likewise.
2681 (vreinterpret_u8_s8): Likewise.
2682 (vreinterpret_u8_s16): Likewise.
2683 (vreinterpret_u8_s32): Likewise.
2684 (vreinterpret_u8_s64): Likewise.
2685 (vreinterpret_u8_f32): Likewise.
2686 (vreinterpret_u8_u16): Likewise.
2687 (vreinterpret_u8_u32): Likewise.
2688 (vreinterpret_u8_u64): Likewise.
2689 (vreinterpret_u8_p8): Likewise.
2690 (vreinterpret_u8_p16): Likewise.
2691 (vreinterpretq_u8_s8): Likewise.
2692 (vreinterpretq_u8_s16): Likewise.
2693 (vreinterpretq_u8_s32): Likewise.
2694 (vreinterpretq_u8_s64): Likewise.
2695 (vreinterpretq_u8_f32): Likewise.
2696 (vreinterpretq_u8_u16): Likewise.
2697 (vreinterpretq_u8_u32): Likewise.
2698 (vreinterpretq_u8_u64): Likewise.
2699 (vreinterpretq_u8_p8): Likewise.
2700 (vreinterpretq_u8_p16): Likewise.
2701 (vreinterpret_u16_s8): Likewise.
2702 (vreinterpret_u16_s16): Likewise.
2703 (vreinterpret_u16_s32): Likewise.
2704 (vreinterpret_u16_s64): Likewise.
2705 (vreinterpret_u16_f32): Likewise.
2706 (vreinterpret_u16_u8): Likewise.
2707 (vreinterpret_u16_u32): Likewise.
2708 (vreinterpret_u16_u64): Likewise.
2709 (vreinterpret_u16_p8): Likewise.
2710 (vreinterpret_u16_p16): Likewise.
2711 (vreinterpretq_u16_s8): Likewise.
2712 (vreinterpretq_u16_s16): Likewise.
2713 (vreinterpretq_u16_s32): Likewise.
2714 (vreinterpretq_u16_s64): Likewise.
2715 (vreinterpretq_u16_f32): Likewise.
2716 (vreinterpretq_u16_u8): Likewise.
2717 (vreinterpretq_u16_u32): Likewise.
2718 (vreinterpretq_u16_u64): Likewise.
2719 (vreinterpretq_u16_p8): Likewise.
2720 (vreinterpretq_u16_p16): Likewise.
2721 (vreinterpret_u32_s8): Likewise.
2722 (vreinterpret_u32_s16): Likewise.
2723 (vreinterpret_u32_s32): Likewise.
2724 (vreinterpret_u32_s64): Likewise.
2725 (vreinterpret_u32_f32): Likewise.
2726 (vreinterpret_u32_u8): Likewise.
2727 (vreinterpret_u32_u16): Likewise.
2728 (vreinterpret_u32_u64): Likewise.
2729 (vreinterpret_u32_p8): Likewise.
2730 (vreinterpret_u32_p16): Likewise.
2731 (vreinterpretq_u32_s8): Likewise.
2732 (vreinterpretq_u32_s16): Likewise.
2733 (vreinterpretq_u32_s32): Likewise.
2734 (vreinterpretq_u32_s64): Likewise.
2735 (vreinterpretq_u32_f32): Likewise.
2736 (vreinterpretq_u32_u8): Likewise.
2737 (vreinterpretq_u32_u16): Likewise.
2738 (vreinterpretq_u32_u64): Likewise.
2739 (vreinterpretq_u32_p8): Likewise.
2740 (vreinterpretq_u32_p16): Likewise.
2742 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2744 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
2746 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
2749 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
2750 (vqnegd_s64): Likewise.
2751 (vqabs_s64): Likewise.
2752 (vqabsd_s64): Likewise.
2754 2014-04-22 Richard Henderson <rth@redhat.com>
2756 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
2757 computation to the top of the loop.
2759 2014-04-22 Renlin <renlin.li@arm.com>
2760 Jiong Wang <jiong.wang@arm.com>
2762 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
2763 * config/aarch64/aarch64.c (aarch64_layout_frame)
2764 (aarch64_initial_elimination_offset): Likewise.
2766 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
2768 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
2771 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
2773 * machmode.h (bitwise_mode_for_mode): Declare.
2774 * stor-layout.h (bitwise_type_for_mode): Likewise.
2775 * stor-layout.c (bitwise_mode_for_mode): New function.
2776 (bitwise_type_for_mode): Likewise.
2777 * builtins.c (fold_builtin_memory_op): Use it instead of
2778 int_mode_for_mode and build_nonstandard_integer_type.
2780 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2782 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
2783 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
2784 (*-*-solaris2*): Simplify.
2785 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
2786 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
2787 *-*-solaris2.9* handling.
2789 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
2791 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
2792 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
2794 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
2795 * configure: Regenerate.
2797 * config/i386/sol2-9.h: Remove.
2799 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
2800 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
2801 Remove Solaris 9 references.
2803 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
2805 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
2806 (floatuns<GPI:mode><GPF:mode>2): Remove.
2807 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
2808 and floatuns conversions.
2809 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
2810 and floatuns conversions.
2811 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
2812 (w1,w2): New mode attributes for inequal width conversions.
2814 2014-04-22 Renlin Li <Renlin.Li@arm.com>
2816 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
2817 the output asm format.
2819 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
2821 * config/aarch64/aarch64-simd.md
2822 (aarch64_cm<optab>di): Always split.
2823 (*aarch64_cm<optab>di): New.
2824 (aarch64_cmtstdi): Always split.
2825 (*aarch64_cmtstdi): New.
2827 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2829 PR tree-optimization/60823
2830 * omp-low.c (ipa_simd_modify_function_body): Go through
2831 all SSA_NAMEs and for those refering to vector arguments
2832 which are going to be replaced adjust SSA_NAME_VAR and,
2833 if it is a default definition, change it into a non-default
2834 definition assigned at the beginning of function from new_decl.
2835 (ipa_simd_modify_stmt_ops): Rewritten.
2836 * tree-dfa.c (set_ssa_default_def): When removing default def,
2837 check for NULL loc instead of NULL *loc.
2839 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2841 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
2842 restrictions on core registers for DImode values in Thumb2.
2844 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2846 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
2847 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
2849 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2851 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
2852 (*iordi_notzesidi_di): Likewise.
2853 (*iordi_notsesidi_di): Likewise.
2855 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2857 * config/arm/arm-protos.h (tune_params): New struct members.
2858 * config/arm/arm.c: Initialise tune_params per processor.
2859 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
2860 for speed, based on new tune_params.
2862 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2864 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
2866 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
2868 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
2870 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
2871 * config/aarch64/arm_neon.h (vrnd_f64): Added.
2872 (vrnda_f64): Likewise.
2873 (vrndi_f64): Likewise.
2874 (vrndm_f64): Likewise.
2875 (vrndn_f64): Likewise.
2876 (vrndp_f64): Likewise.
2877 (vrndx_f64): Likewise.
2879 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2881 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
2882 GET_MODE_SIZE argument is enum machine_mode.
2884 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2887 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
2888 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
2890 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
2893 * asan.c (asan_emit_stack_protection): Force the base to align to
2894 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
2895 appropriate bits if STRICT_ALIGNMENT.
2896 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
2898 (expand_used_vars): Leave a space in the stack frame for alignment
2899 if STRICT_ALIGNMENT.
2901 2014-04-21 David Malcolm <dmalcolm@redhat.com>
2903 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
2905 (gimple_store_p): Likewise.
2906 (gimple_assign_load_p): Likewise.
2907 (gimple_assign_cast_p): Likewise.
2908 (gimple_clobber_p): Likewise.
2910 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
2911 rather than a gimple.
2912 (gimple_assign_cast_p): Likewise.
2914 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
2917 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
2918 If mode is DDmode and TARGET_E500_DOUBLE allow move.
2920 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
2921 more debug information for E500 if -mdebug=reg.
2923 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
2926 * config/i386/i386.c (ix86_expand_builtin)
2927 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
2928 register for target RTX.
2929 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
2931 2014-04-18 Cong Hou <congh@google.com>
2933 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
2934 the widen-mult pattern by handling two operands with different sizes,
2935 and operands whose size is smaller than half of the result type.
2937 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
2939 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
2940 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
2941 (do_estimate_edge_time): Compute it.
2942 * ipa-inline.c (want_inline_small_function_p): Bypass
2943 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
2945 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
2947 * ipa-inline.c (spec_rem): New static variable.
2948 (dump_overall_stats): New function.
2949 (dump_inline_stats): New function.
2951 2014-04-18 Richard Henderson <rth@redhat.com>
2953 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
2954 to GET_MODE_SIZE, not a reg_class_t.
2956 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2958 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
2959 (vsx_xxmrglw_<mode>): Likewise.
2961 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2964 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
2965 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
2966 (rs6000_init_hard_regno_mode_ok): Likewise.
2968 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
2970 * ipa-inline.c (inline_small_functions): Account only non-cold
2972 * doc/invoke.texi (inline-unit-growth): Update documentation.
2974 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
2976 * config/rs6000/rs6000.md (addti3, subti3): New.
2978 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
2981 * config/i386/i386.c (ix86_expand_clear): Remove outdated
2982 comment. Check optimize_insn_for_size_p instead of
2983 optimize_insn_for_speed_p.
2985 2014-04-17 Martin Jambor <mjambor@suse.cz>
2987 * gimple-iterator.c (gsi_start_edge): New function.
2988 * gimple-iterator.h (gsi_start_edge): Declare.
2989 * tree-sra.c (single_non_eh_succ): New function.
2990 (disqualify_ops_if_throwing_stmt): Renamed to
2991 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
2992 having one non-EH successor BB.
2993 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
2994 generate loads into replacements.
2995 (sra_modify_assign): Likewise and and also use the simple path for
2997 (sra_modify_function_body): Commit statements on edges.
2999 2014-04-17 Richard Biener <rguenther@suse.de>
3002 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
3003 comparison results and add clarifying comment.
3005 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3007 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
3008 (blank_mode): Initialize it.
3009 (emit_mode_size_inline, emit_mode_nunits_inline,
3010 emit_mode_inner_inline): New functions.
3011 (emit_insn_modes_h): Call them and surround their output with
3012 #if GCC_VERSION >= 4001 ... #endif.
3013 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
3014 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
3015 mode_* arrays if the argument is __builtin_constant_p.
3016 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
3017 is enum machine_mode.
3019 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3021 * passes.c (opt_pass::execute): Adjust.
3022 (pass_manager::execute_pass_mode_switching): Likewise.
3023 (early_local_passes::execute): Likewise.
3024 (execute_one_pass): Pass cfun to the pass's execute method.
3025 * tree-pass.h (opt_pass::execute): Add function * argument.
3026 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3027 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3028 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3029 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3030 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3031 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
3032 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
3033 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3034 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3035 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3036 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3037 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3038 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3039 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3040 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3041 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3042 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3043 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3044 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3045 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3046 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3047 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3048 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3049 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3050 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3051 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3052 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3053 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3054 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3057 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3059 * passes.c (opt_pass::gate): Take function * argument.
3060 (gate_all_early_local_passes): Merge into
3061 (early_local_passes::gate): this.
3062 (gate_all_early_optimizations): Merge into
3063 (all_early_optimizations::gate): this.
3064 (gate_all_optimizations): Mege into
3065 (all_optimizations::gate): this.
3066 (gate_all_optimizations_g): Merge into
3067 (all_optimizations_g::gate): this.
3068 (gate_rest_of_compilation): Mege into
3069 (rest_of_compilation::gate): this.
3070 (gate_postreload): Merge into
3071 (postreload::gate): this.
3072 (dump_one_pass): Pass cfun to the pass's gate method.
3073 (execute_ipa_summary_passes): Likewise.
3074 (execute_one_pass): Likewise.
3075 (ipa_write_summaries_2): Likewise.
3076 (ipa_write_optimization_summaries_1): Likewise.
3077 (ipa_read_summaries_1): Likewise.
3078 (ipa_read_optimization_summaries_1): Likewise.
3079 (execute_ipa_stmt_fixups): Likewise.
3080 * tree-pass.h (opt_pass::gate): Add function * argument.
3081 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
3082 combine-stack-adj.c, combine.c, compare-elim.c,
3083 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3084 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
3085 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
3086 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
3087 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3088 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3089 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3090 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
3091 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
3092 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
3093 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3094 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3095 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
3096 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3097 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3098 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3099 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3100 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3101 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3102 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3103 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3104 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3105 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3106 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
3107 var-tracking.c, vtable-verify.c, web.c: Adjust.
3109 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3111 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
3112 * configure: Regenerate.
3114 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3116 * passes.c (dump_one_pass): don't check pass->has_gate.
3117 (execute_ipa_summary_passes): Likewise.
3118 (execute_one_pass): Likewise.
3119 (ipa_write_summaries_2): Likewise.
3120 (ipa_write_optimization_summaries_1): Likewise.
3121 (ipa_read_optimization_summaries_1): Likewise.
3122 (execute_ipa_stmt_fixups): Likewise.
3123 * tree-pass.h (pass_data::has_gate): Remove.
3124 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3125 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3126 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3127 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3128 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3129 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
3130 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
3131 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
3132 gimple-low.c, gimple-ssa-isolate-paths.c,
3133 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
3134 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3135 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3136 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
3137 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
3138 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
3139 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
3140 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3141 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
3142 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3143 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3144 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3145 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3146 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3147 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3148 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3149 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3150 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3151 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3152 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3153 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3154 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3157 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3159 * pass_manager.h (pass_manager::register_dump_files_1): Remove
3161 * passes.c (pass_manager::register_dump_files_1): Merge into
3162 (pass_manager::register_dump_files): this, and remove its handling of
3163 properties since the pass always has the properties anyway.
3164 (pass_manager::pass_manager): Adjust.
3166 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3168 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
3169 * passes.c (pass_manager::register_dump_files_1): Remove dead code
3170 dealing with properties.
3171 (pass_manager::register_dump_files): Adjust.
3173 2014-03-20 Mark Wielaard <mjw@redhat.com>
3175 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
3176 then represent the bound as normal constant value.
3178 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3181 Forward port from 4.8 branch
3182 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
3184 * config/i386/bmiintrin.h (_blsi_u32): New.
3188 (_blsmsk_u32): Ditto.
3189 (_blsmsk_u64): Ditto.
3190 (_tzcnt_u32): Ditto.
3191 (_tzcnt_u64): Ditto.
3193 2014-04-17 Kito Cheng <kito@0xlab.org>
3195 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
3197 2014-04-17 Richard Biener <rguenther@suse.de>
3200 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
3201 boolean results for comparisons.
3203 2014-04-17 Richard Biener <rguenther@suse.de>
3205 PR tree-optimization/60836
3206 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
3207 initial PHI args to be gimple values.
3209 2014-04-17 Richard Biener <rguenther@suse.de>
3211 PR tree-optimization/60841
3212 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
3213 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
3214 of stmts to SLP build.
3215 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
3216 (vect_analyze_slp): Likewise.
3217 (vect_analyze_slp_instance): Likewise.
3218 (vect_build_slp_tree): Limit overall SLP tree growth.
3219 * tree-vectorizer.h (vect_analyze_data_refs,
3220 vect_analyze_slp): Adjust prototypes.
3222 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3224 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
3227 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3229 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
3230 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
3231 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
3232 for TARGET_SLOW_PSHUFB
3234 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3236 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
3237 * config/i386/i386.c (intel_cost): Ditto.
3239 2014-04-17 Joey Ye <joey.ye@arm.com>
3241 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
3243 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3245 * opts.c (common_handle_option): Disable -fipa-reference coorectly
3248 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3250 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
3251 (type_all_derivations_known_p): New predicate.
3252 (type_all_ctors_visible_p): New predicate.
3253 (type_possibly_instantiated_p): New predicate.
3254 (get_odr_type): Compute all_derivations_known.
3255 (dump_odr_type): Dump the flag.
3256 (maybe_record_type): Cleanup.
3257 (record_target_from_binfo): Add bases_to_consider array;
3258 record bases for types w/o instances and skip CXX destructor.
3259 (possible_polymorphic_call_targets_1): Add bases_to_consider
3260 and consider_construction parameters; check if type may have instance.
3261 (get_polymorphic_call_info): Set maybe_in_construction to true
3262 when we know nothing.
3263 (record_targets_from_bases): Skip CXX destructors; they are
3264 never called for types in construction.
3265 (possible_polymorphic_call_targets): Do not record target when
3266 type may not have instance.
3268 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3271 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
3272 external aliases alive, too.
3274 2014-04-16 Andrew Pinski <apinski@cavium.com>
3276 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
3279 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3281 * final.c (compute_alignments): Do not apply loop alignment to a block
3282 falling through to the exit.
3284 2014-04-16 Catherine Moore <clm@codesourcery.com>
3286 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
3287 Adjust constraints for microMIPS store patterns.
3289 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3291 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
3293 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3295 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
3296 (append_use): Run at -O0.
3297 (append_vdef): Likewise.
3298 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
3299 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
3301 2014-04-16 Jakub Jelinek <jakub@redhat.com>
3303 PR tree-optimization/60844
3304 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
3305 (propagate_op_to_single_use, remove_visited_stmt_chain,
3306 linearize_expr, repropagate_negates, reassociate_bb): Use it
3307 instead of gsi_remove.
3309 2014-04-16 Martin Jambor <mjambor@suse.cz>
3311 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
3312 ipa_transforms_to_apply.
3313 (cgraph_function_versioning): Assert that old_node has empty
3314 ipa_transforms_to_apply.
3315 * trans-mem.c (ipa_tm_create_version): Likewise.
3316 * tree-inline.c (tree_function_versioning): Do not duplicate
3317 ipa_transforms_to_apply.
3319 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3322 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
3324 Pass necessary as flags on 64-bit Solaris/x86.
3325 Use lowercase relocs for x86_64-*-*.
3326 * configure: Regenerate.
3328 2014-04-15 Jan Hubicka <jh@suse.cz>
3330 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
3331 (maybe_record_node, likely_target_p): Use it.
3333 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3336 Revert following patch
3338 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3341 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3342 software floating point or no floating point registers, do not
3343 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3344 in GPRs that occurs after we tested for GPRs that would never be
3347 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3348 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3349 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3350 specifically allow DDmode, since that does not use the SPE SIMD
3353 2014-03-21 Mark Wielaard <mjw@redhat.com>
3355 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
3356 as unsigned or int depending on type and value used.
3358 2014-04-15 Richard Biener <rguenther@suse.de>
3360 PR rtl-optimization/56965
3361 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
3362 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
3364 * alias.c (true_dependence_1): Do not call
3365 nonoverlapping_component_refs_p.
3366 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
3367 nonoverlapping_component_refs_p.
3368 (indirect_refs_may_alias_p): Likewise.
3370 2014-04-15 Teresa Johnson <tejohnson@google.com>
3372 * cfg.c (dump_bb_info): Fix flags check.
3373 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
3375 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3377 PR rtl-optimization/60663
3378 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
3381 2014-04-15 Richard Biener <rguenther@suse.de>
3383 * lto-streamer.h (LTO_major_version): Bump to 4.
3385 2014-04-15 Richard Biener <rguenther@suse.de>
3387 * common.opt (lto_partition_model): New enum.
3388 (flto-partition=): Merge separate options with a single with argument,
3389 add -flto-partition=one support.
3390 * flag-types.h (enum lto_partition_model): Declare.
3391 * opts.c (finish_options): Remove duplicate -flto-partition=
3393 * lto-wrapper.c (run_gcc): Adjust.
3395 2014-04-15 Richard Biener <rguenther@suse.de>
3397 * alias.c (ncr_compar): New function.
3398 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
3400 2014-04-15 Richard Biener <rguenther@suse.de>
3402 * alias.c (record_component_aliases): Do not walk BINFOs.
3404 2014-04-15 Richard Biener <rguenther@suse.de>
3406 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
3407 Add struct function argument and adjust.
3408 (find_func_aliases_for_call): Likewise.
3409 (find_func_aliases): Likewise.
3410 (find_func_clobbers): Likewise.
3411 (intra_create_variable_infos): Likewise.
3412 (compute_points_to_sets): Likewise.
3413 (ipa_pta_execute): Adjust. Do not push/pop cfun.
3415 2014-04-15 Richard Biener <rguenther@suse.de>
3417 * tree.c (iterative_hash_expr): Use enum tree_code_class
3418 to store TREE_CODE_CLASS.
3419 (tree_block): Likewise.
3420 (tree_set_block): Likewise.
3421 * tree.h (fold_build_pointer_plus_loc): Use
3422 convert_to_ptrofftype_loc.
3424 2014-04-15 Jakub Jelinek <jakub@redhat.com>
3427 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
3430 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
3432 * cfgloop.h (struct loop): Move force_vectorize down.
3433 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
3434 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
3435 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
3436 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
3437 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
3438 * tree-core.h (enum annot_expr_kind): Add new kind values.
3439 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
3440 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
3442 * tree.def (ANNOTATE_EXPR): Tweak comment.
3444 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3446 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
3449 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
3451 * tree.h (TYPE_IDENTIFIER): Declare.
3452 * tree.c (subrange_type_for_debug_p): Use it.
3453 * godump.c (go_format_type): Likewise.
3454 * dwarf2out.c (is_cxx_auto, modified_type_die,
3455 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
3456 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
3458 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3461 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
3463 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
3465 * config/i386/i386.c (examine_argument): Return bool. Return true if
3466 parameter should be passed in memory.
3467 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
3468 (construct_container): Update calls to examine_argument.
3469 (function_arg_advance_64): Ditto.
3470 (return_in_memory_32): Merge with ix86_return_in_memory.
3471 (return_in_memory_64): Ditto.
3472 (return_in_memory_ms_64): Ditto.
3474 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3476 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
3477 * coverage.c (coverage_compute_profile_id): Handle externally visible
3480 2014-04-14 Martin Jambor <mjambor@suse.cz>
3482 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
3483 DECL_DISREGARD_INLINE_LIMITS functions.
3485 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3488 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
3490 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3493 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
3494 optimize_insn_for_speed_p instead of
3495 optimize_function_for_speed_p.
3497 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
3499 * doc/invoke.texi (free): Document AArch64.
3501 2014-04-14 Richard Biener <rguenther@suse.de>
3503 PR tree-optimization/60042
3504 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
3505 (insert_into_preds_of_block): Do not prevent PHI insertion
3506 for REFERENCE exprs here ...
3507 (eliminate_dom_walker::before_dom_children): ... but prevent
3508 their use here under similar conditions when applied to the
3509 IL after PRE optimizations.
3511 2014-04-14 Richard Biener <rguenther@suse.de>
3513 * passes.def: Move early points-to after early SRA.
3515 2014-04-14 Richard Biener <rguenther@suse.de>
3517 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
3518 check for which sign-changes we allow when forwarding
3519 a converted value into a switch.
3521 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3523 * stor-layout.c (place_field): Finalize non-constant offset for the
3526 2014-04-14 Richard Biener <rguenther@suse.de>
3528 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
3530 (expand_switch_using_bit_tests_p): Likewise.
3531 (process_switch): Compute and pass on speed_p based on the
3533 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
3534 optimize_bb_for_speed_p.
3536 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3538 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
3539 * function.h (struct function): Rename has_force_vect_loops into
3540 has_force_vectorize_loops.
3541 * lto-streamer-in.c (input_cfg): Adjust for renaming.
3542 (input_struct_function_base): Likewise.
3543 * lto-streamer-out.c (output_cfg): Likewise.
3544 (output_struct_function_base): Likewise.
3545 * omp-low.c (expand_omp_simd): Likewise.
3546 * tree-cfg.c (move_sese_region_to_fn): Likewise.
3547 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
3548 (version_loop_for_if_conversion): Likewise.
3549 (tree_if_conversion): Likewise.
3550 (main_tree_if_conversion): Likewise.
3551 (gate_tree_if_conversion): Likewise.
3552 * tree-inline.c (copy_loops): Likewise.
3553 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
3554 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
3555 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
3556 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
3557 * tree-vectorizer.c (vectorize_loops): Likewise.
3558 * tree-vectorizer.h (unlimited_cost_model): Likewise.
3560 2014-04-14 Richard Biener <rguenther@suse.de>
3563 * lto-streamer-out.c (wrap_refs): New function.
3564 (lto_output): Wrap symbol references in global initializes in
3565 type-preserving MEM_REFs.
3567 2014-04-14 Christian Bruel <christian.bruel@st.com>
3569 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
3571 2014-04-14 Christian Bruel <christian.bruel@st.com>
3573 * config/sh/sh.md (setmemqi): New expand pattern.
3574 * config/sh/sh.h (CLEAR_RATIO): Define.
3575 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
3576 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
3578 2014-04-14 Richard Biener <rguenther@suse.de>
3581 * fold-const.c (negate_expr_p): Don't negate directional rounding
3583 (fold_negate_expr): Likewise.
3585 2014-04-14 Richard Biener <rguenther@suse.de>
3587 PR tree-optimization/59817
3588 PR tree-optimization/60453
3589 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
3590 recursion to catch all CHRECs in the scalar evolution and restrict
3591 the predicate for the remains appropriately.
3593 2014-04-12 Catherine Moore <clm@codesourcery.com>
3595 * config/mips/constraints.md: Add new register constraint "kb".
3596 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
3597 (*movhi_internal): Likewise.
3598 (*movqi_internal): Likewise.
3599 * config/mips/mips.h (M16_STORE_REGS): New register class.
3600 (REG_CLASS_NAMES): Add M16_STORE_REGS.
3601 (REG_CLASS_CONTENTS): Likewise.
3602 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
3604 2014-04-11 Tobias Burnus <burnus@net-b.de>
3607 * doc/invoke.texi (-Wformat-signedness): Document it.
3608 (Wformat=2): Mention that this enables -Wformat-signedness.
3610 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3612 * common/config/epiphany/epiphany-common.c
3613 (epiphany_option_optimization_table): Enable section anchors by
3614 default at -O1 or higher.
3615 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
3616 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
3617 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
3618 carries no extra cost.
3619 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
3620 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
3621 * config/epiphany/predicates.md (memclob_operand): New predicate.
3622 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
3623 Use memclob_operand predicate and X constraint for operand 3.
3625 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3627 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
3628 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
3631 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3633 PR rtl-optimization/60651
3634 * mode-switching.c (optimize_mode_switching): Make sure to emit
3635 sets of a lower numbered entity before sets of a higher numbered
3636 entity to a mode of the same or lower priority.
3637 When creating a seginfo for a basic block that starts with a code
3638 label, move the insertion point past the code label.
3639 (new_seginfo): Document and enforce requirement that
3640 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
3641 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
3642 * doc/tm.texi: Regenerate.
3644 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
3647 * config/arc/arc.c (arc_save_restore): Fix assert typo.
3649 2013-04-11 Jakub Jelinek <jakub@redhat.com>
3651 * BASE-VER: Set to 4.10.0.
3653 2014-04-11 Tobias Burnus <burnus@net-b.de>
3656 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
3657 * doc/gcc.texi (Service): Update description in the @menu
3658 * doc/invoke.texi (Option Summary): Remove misplaced and
3661 2014-04-11 Steve Ellcey <sellcey@mips.com>
3662 Jakub Jelinek <jakub@redhat.com>
3665 * expr.c (convert_move): Use emit_store_flag_force instead of
3666 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
3669 2014-04-11 Richard Biener <rguenther@suse.de>
3672 * varasm.c (assemble_alias): Avoid endless error reporting
3673 recursion by setting TREE_ASM_WRITTEN.
3675 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3677 * config/s390/s390.md: Add a splitter for NOT rtx.
3679 2014-04-11 Jakub Jelinek <jakub@redhat.com>
3681 PR rtl-optimization/60663
3682 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
3684 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
3685 Jakub Jelinek <jakub@redhat.com>
3688 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
3689 flag from decl_node to node.
3691 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3694 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
3695 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
3696 ameliorating the cases where it can be.
3698 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
3701 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
3703 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3704 (loadsync_<mode>): Change mode.
3705 (load_quadpti, store_quadpti): New.
3706 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3707 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3708 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
3710 2014-04-09 Cong Hou <congh@google.com>
3713 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
3716 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3718 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
3719 instead of vnor to exploit possible fusion opportunity in the
3721 (altivec_expand_vec_perm_const_le): Likewise.
3723 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
3725 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3726 (loadsync_<mode>): Change mode.
3727 (load_quadpti, store_quadpti): New.
3728 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3729 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3731 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
3734 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
3735 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
3736 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
3738 2014-04-08 Richard Biener <rguenther@suse.de>
3741 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
3742 a 64bit widest int print double-int similar to on HWI64 hosts.
3744 2014-04-08 Richard Biener <rguenther@suse.de>
3746 PR tree-optimization/60785
3747 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
3748 default defs properly.
3750 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
3752 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
3753 (Weffc++): Likewise.
3755 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
3757 * ipa-devirt.c (maybe_record_node): When node is not recorded,
3758 set completep to false rather than true.
3760 2014-04-07 Douglas B Rupp <rupp@adacore.com>
3763 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
3764 ARM_TARGET2_DWARF_FORMAT.
3766 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
3769 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
3770 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
3773 2014-04-07 Richard Biener <rguenther@suse.de>
3775 PR tree-optimization/60766
3776 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
3777 (may_eliminate_iv): Convert cand_value_at result to desired type.
3779 2014-04-07 Jason Merrill <jason@redhat.com>
3782 * common.opt (-fno-gnu-unique): Add.
3783 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
3785 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3787 * haifa-sched.c: Fix outdated function reference and minor
3788 grammar errors in introductory comment.
3790 2014-04-07 Richard Biener <rguenther@suse.de>
3793 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
3795 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
3797 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
3800 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
3801 size accounting for thunks.
3802 (pa_asm_output_mi_thunk): Use final_start_function() and
3803 final_end_function() to output function start and end directives.
3805 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3807 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
3808 device specific ISA/ feature information. Remove short_sp and
3809 errata_skip ds. Add avr_device_specific_features enum to have device
3811 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
3812 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
3813 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
3814 updated device specific info.
3815 * config/avr/avr-mcus.def: Merge device specific details to
3816 dev_attribute field.
3817 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
3819 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
3820 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
3821 assembler if RMW isa supported by current device.
3822 * config/avr/genmultilib.awk: Update as device info structure changed.
3823 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
3825 2014-04-04 Cong Hou <congh@google.com>
3827 PR tree-optimization/60656
3828 * tree-vect-stmts.c (supportable_widening_operation):
3829 Fix a bug that elements in a vector with vect_used_by_reduction
3830 property are incorrectly reordered when the operation on it is not
3831 consistant with the one in reduction operation.
3833 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
3835 PR rtl-optimization/60155
3836 * gcse.c (record_set_data): New function.
3837 (single_set_gcse): New function.
3838 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
3839 (hoist_code): Likewise.
3840 (get_pressure_class_and_nregs): Likewise.
3842 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
3844 * explow.c (probe_stack_range): Emit a final optimization blockage.
3846 2014-04-04 Anthony Green <green@moxielogic.com>
3848 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
3851 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
3854 * lto-cgraph.c (input_overwrite_node): Check that partitioning
3855 flags are set only during streaming.
3856 * ipa.c (process_references, walk_polymorphic_call_targets,
3857 symtab_remove_unreachable_nodes): Drop bodies of always inline
3858 after early inlining.
3859 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
3861 2014-04-04 Jakub Jelinek <jakub@redhat.com>
3862 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3865 * dwarf2out.c (const_ok_for_output_1): Reject expressions
3868 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3871 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
3873 (cortex_a53_fdivd): Likewise.
3875 2014-04-04 Martin Jambor <mjambor@suse.cz>
3878 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
3880 * cgraph.c (clone_of_p): Also return true if thunks match.
3881 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
3882 cgraph_function_or_thunk_node and an obsolete comment.
3883 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
3885 (build_function_decl_skip_args): Likewise.
3886 (set_new_clone_decl_and_node_flags): New function.
3887 (duplicate_thunk_for_node): Likewise.
3888 (redirect_edge_duplicating_thunks): Likewise.
3889 (cgraph_clone_node): New parameter args_to_skip, pass it to
3890 redirect_edge_duplicating_thunks which is called instead of
3891 cgraph_redirect_edge_callee.
3892 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
3893 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
3895 2014-04-04 Jeff Law <law@redhat.com>
3898 * config/arm/predicates.md (const_int_I_operand): New predicate.
3899 (const_int_M_operand): Similarly.
3900 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
3902 (insv_t2, extv_reg, extzv_t2): Likewise.
3903 (load_multiple_with_writeback): Similarly for const_int_I_operand.
3904 (pop_multiple_with_writeback_and_return): Likewise.
3905 (vfp_pop_multiple_with_writeback): Likewise
3907 2014-04-04 Richard Biener <rguenther@suse.de>
3910 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
3911 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
3913 * gimplify.h (gimple_add_tmp_var_fn): Declare.
3914 * gimplify.c (gimple_add_tmp_var_fn): New function.
3915 * gimple-expr.h (create_tmp_reg_fn): Declare.
3916 * gimple-expr.c (create_tmp_reg_fn): New function.
3917 * gimple-low.c (record_vars_into): Don't change cfun.
3918 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
3919 code generation without cfun.
3921 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
3924 * Makefile.in (install-driver): Fix shell scripting.
3926 2014-04-03 Cong Hou <congh@google.com>
3928 PR tree-optimization/60505
3929 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
3930 threshold of number of iterations below which no vectorization
3932 * tree-vect-loop.c (new_loop_vec_info):
3933 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
3934 * tree-vect-loop.c (vect_analyze_loop_operations):
3935 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
3936 * tree-vect-loop.c (vect_transform_loop):
3937 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
3938 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
3939 of iterations of the loop and see if we should build the epilogue.
3941 2014-04-03 Richard Biener <rguenther@suse.de>
3943 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
3944 (streamer_tree_cache_create): Adjust.
3945 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
3946 to allow optional nodes array.
3947 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
3948 (streamer_tree_cache_append): Likewise.
3949 (streamer_tree_cache_create): Create nodes array optionally
3950 as specified by parameter.
3951 * lto-streamer-out.c (create_output_block): Avoid maintaining
3952 the node array in the writer cache.
3953 (DFS_write_tree): Remove assertion.
3954 (produce_asm_for_decls): Free the out decl state hash table early.
3955 * lto-streamer-in.c (lto_data_in_create): Adjust for
3956 streamer_tree_cache_create prototype change.
3958 2014-04-03 Richard Biener <rguenther@suse.de>
3960 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
3961 set TREE_CHAIN to NULL_TREE.
3963 2014-04-03 Richard Biener <rguenther@suse.de>
3965 PR tree-optimization/60740
3966 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
3967 over all GIMPLE_COND operands.
3969 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
3971 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
3972 (Weffc++): Remove Scott's numbering, merge lists and reference
3975 2014-04-03 Nick Clifton <nickc@redhat.com>
3977 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
3980 2014-04-03 Martin Jambor <mjambor@suse.cz>
3982 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
3983 mention gcc_unreachable before failing.
3984 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
3987 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
3990 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
3991 inconsistent code and instead mark the context inconsistent.
3992 (possible_polymorphic_call_targets): For inconsistent contexts
3993 return empty complete list.
3995 2014-04-02 Anthony Green <green@moxielogic.com>
3997 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
3998 (extendqisi2, extendhisi2): Define.
3999 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
4000 (WCHAR_TYPE): Change to unsigned int.
4002 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4004 PR tree-optimization/60733
4005 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
4006 insertion point for PHI candidates to be the end of the feeding
4007 block for the PHI argument.
4009 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
4011 PR rtl-optimization/60650
4012 * lra-constraints.c (process_alt_operands): Decrease reject for
4013 earlyclobber matching.
4015 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4017 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
4019 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4021 * config/spu/spu.c (pad_bb): Do not crash when the last
4022 insn is CODE_FOR_blockage.
4024 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4026 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
4027 lies outside the target mode.
4029 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4032 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4033 software floating point or no floating point registers, do not
4034 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4035 in GPRs that occurs after we tested for GPRs that would never be
4038 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4039 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4040 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4041 specifically allow DDmode, since that does not use the SPE SIMD
4044 2014-04-02 Richard Biener <rguenther@suse.de>
4047 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4048 MODE_INTs. Properly use negv_optab.
4049 (expand_abs): Likewise.
4051 2014-04-02 Richard Biener <rguenther@suse.de>
4054 * Makefile.in (install-driver): Guard extra installs with special
4057 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4059 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4062 2014-04-01 Richard Henderson <rth@redhat.com>
4065 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
4066 alternative enabled before register allocation.
4068 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
4070 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
4071 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
4073 (nios2_large_got_address): Remove unneeded 'sym' parameter.
4074 (nios2_got_address): Update nios2_large_got_address call site.
4075 (nios2_delegitimize_address): New function.
4076 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
4077 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
4078 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
4080 2014-04-01 Martin Husemann <martin@duskware.de>
4082 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
4085 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
4087 PR rtl-optimization/60604
4088 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
4089 check from register_operand.
4090 (register_operand): Redefine in terms of general_operand.
4091 (nonmemory_operand): Use register_operand for the non-constant cases.
4093 2014-04-01 Richard Biener <rguenther@suse.de>
4095 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
4097 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
4099 * doc/invoke.texi (mapp-regs): Clarify.
4101 2014-03-31 Ulrich Drepper <drepper@gmail.com>
4103 * config/i386/avx512fintrin.h (__v32hi): Define type.
4104 (__v64qi): Likewise.
4105 (_mm512_set1_epi8): Define.
4106 (_mm512_set1_epi16): Define.
4107 (_mm512_set4_epi32): Define.
4108 (_mm512_set4_epi64): Define.
4109 (_mm512_set4_pd): Define.
4110 (_mm512_set4_ps): Define.
4111 (_mm512_setr4_epi64): Define.
4112 (_mm512_setr4_epi32): Define.
4113 (_mm512_setr4_pd): Define.
4114 (_mm512_setr4_ps): Define.
4115 (_mm512_setzero_epi32): Define.
4117 2014-03-31 Martin Jambor <mjambor@suse.cz>
4120 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
4121 callsite_arguments_match_p. Updated all callers. Also check types of
4122 corresponding formal parameters and actual arguments.
4123 (not_all_callers_have_enough_arguments_p) Renamed to
4124 some_callers_have_mismatched_arguments_p.
4126 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
4128 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
4130 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
4133 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
4136 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
4138 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
4139 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
4141 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
4142 Use FMAMODE_NOVF512 mode iterator.
4143 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
4144 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
4145 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
4146 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
4148 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
4149 Use VF_128_256 mode iterator.
4150 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
4153 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4155 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
4156 static chain if needed.
4158 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4161 * lra-constraints.c (index_part_to_reg): New.
4162 (process_address): Use it.
4164 2014-03-27 Jeff Law <law@redhat.com>
4165 Jakub Jelinek <jakub@redhat.com>
4168 * expr.c (do_tablejump): Use simplify_gen_binary rather than
4169 gen_rtx_{PLUS,MULT} to build up the address expression.
4171 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
4172 creating non-canonical RTL.
4174 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4177 * ipa-inline.c (want_inline_small_function_p): Short circuit large
4178 functions; reorganize to make cheap checks first.
4179 (inline_small_functions): Do not estimate growth when dumping;
4181 * ipa-inline.h (inline_summary): Add min_size.
4182 (growth_likely_positive): New function.
4183 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
4184 (set_cond_stmt_execution_predicate): Cleanup.
4185 (estimate_edge_size_and_time): Compute min_size.
4186 (estimate_calls_size_and_time): Likewise.
4187 (estimate_node_size_and_time): Likewise.
4188 (inline_update_overall_summary): Update min_size.
4189 (do_estimate_edge_time): Likewise.
4190 (do_estimate_edge_size): Update.
4191 (do_estimate_edge_hints): Update.
4192 (growth_likely_positive): New function.
4194 2014-03-28 Jakub Jelinek <jakub@redhat.com>
4197 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
4198 also if addr has VOIDmode.
4200 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4202 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
4203 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
4205 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
4206 instructions as well as AdvancedSIMD loads.
4208 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4210 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
4211 Use crypto_aese type.
4212 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
4213 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
4214 crypto_aese, crypto_aesmc. Move to types.md.
4215 * config/arm/types.md (crypto_aes): Split into crypto_aese,
4217 * config/arm/iterators.md (crypto_type): Likewise.
4219 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4221 * cgraph.c: Include expr.h and tree-dfa.h.
4222 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
4225 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4228 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
4229 regs from checking multi-reg pseudos.
4231 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4233 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
4235 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4237 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
4238 if it would clobber the stack pointer, even temporarily.
4240 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
4242 * mode-switching.c: Make small adjustments to the top comment.
4244 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4246 * config/rs6000/constraints.md (wD constraint): New constraint to
4247 match the constant integer to get the top DImode/DFmode out of a
4248 vector in a VSX register.
4250 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
4251 match the constant integer to get the top DImode/DFmode out of a
4252 vector in a VSX register.
4254 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
4257 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
4260 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
4261 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
4263 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
4264 Optimize vec_extract of 64-bit values, where the value being
4265 extracted is in the top word, where we can use scalar
4266 instructions. Add direct move and store support. Combine the big
4267 endian/little endian vector select load support into a single insn.
4268 (vsx_extract_<mode>_internal1): Likewise.
4269 (vsx_extract_<mode>_internal2): Likewise.
4270 (vsx_extract_<mode>_load): Likewise.
4271 (vsx_extract_<mode>_store): Likewise.
4272 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
4273 combined into vsx_extract_<mode>_load.
4274 (vsx_extract_<mode>_one_le): Likewise.
4276 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
4277 define the top 64-bit vector element.
4279 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
4282 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4283 Document vec_vbpermq builtin.
4286 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
4287 enable use of xxsldwi and xxpermdi builtin functions.
4288 (vec_xxpermdi): Likewise.
4290 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4291 Document use of vec_xxsldwi and vec_xxpermdi builtins.
4293 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
4295 PR rtl-optimization/60650
4296 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
4298 (find_spills_for): New.
4299 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
4300 Spill all pseudos on the second iteration.
4302 2014-03-27 Marek Polacek <polacek@redhat.com>
4305 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
4308 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4310 * config/s390/s390.c (s390_can_use_return_insn): Check for
4311 call-saved FPRs on 31 bit.
4313 2014-03-27 Jakub Jelinek <jakub@redhat.com>
4316 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
4317 if they need regimplification, just drop them instead of
4318 calling gimple_regimplify_operands on them.
4320 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
4323 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
4324 (aarch64_frame_pointer_required): Adjust logic.
4325 (aarch64_can_eliminate): Adjust logic.
4326 (aarch64_override_options_after_change): Adjust logic.
4328 2014-03-27 Dehao Chen <dehao@google.com>
4330 * ipa-inline.c (early_inliner): Update node's inline info.
4332 2014-03-26 Dehao Chen <dehao@google.com>
4334 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
4335 compiler inserted conditional jumps for NAN float check.
4337 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4339 * ubsan.h (ubsan_create_data): Change second argument's type
4340 to const location_t *.
4341 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
4343 (ubsan_create_data): Change second argument to const location_t *PLOC.
4344 Create Loc field whenever PLOC is non-NULL.
4345 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
4346 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
4350 * real.c (real_to_integer2): Change type of low to UHWI.
4352 2014-03-26 Tobias Burnus <burnus@net-b.de>
4354 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
4355 (CILK_SELF_SPECS): New define.
4356 (driver_self_specs): Use it.
4358 2014-03-26 Richard Biener <rguenther@suse.de>
4360 * tree-pretty-print.c (percent_K_format): Implement special
4361 case for LTO and its stripped down BLOCK tree.
4363 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4366 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
4368 * tree-vrp.c (simplify_internal_call_using_ranges): If only
4369 one range is range_int_cst_p, but not both, at least optimize
4370 addition/subtraction of 0 and multiplication by 0 or 1.
4371 * gimple-fold.c (gimple_fold_call): Fold
4372 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
4373 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
4374 INTEGER_CSTs, try to fold at least x * 0 and y - y.
4376 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
4378 PR rtl-optimization/60452
4379 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
4380 <case REG>: Return 1 for invalid offsets from the frame pointer.
4382 2014-03-26 Marek Polacek <polacek@redhat.com>
4385 * doc/extend.texi (C Extensions): Mention variable-length arrays in
4388 2014-03-26 Marek Polacek <polacek@redhat.com>
4391 * doc/extend.texi (Designated Inits): Describe what happens to omitted
4394 2014-03-26 Marek Polacek <polacek@redhat.com>
4397 * ira-color.c (update_conflict_hard_regno_costs): Perform the
4398 multiplication in unsigned type.
4400 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4402 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
4404 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4406 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
4408 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4411 * cif-code.def (UNREACHABLE) New code.
4412 * ipa-inline.c (inline_small_functions): Skip edges to
4413 __builtlin_unreachable.
4414 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
4415 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
4416 predicate to __bulitin_unreachable.
4417 (set_cond_stmt_execution_predicate): Fix issue when
4418 invert_tree_comparison returns ERROR_MARK.
4419 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
4420 propagate to inline clones.
4421 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
4423 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
4424 * cgraphclones.c (cgraph_clone_node): If call destination is already
4425 ureachable, do not redirect it back.
4426 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
4429 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4431 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
4432 Do not modify inline clones.
4434 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4436 * config/i386/i386.md (general_sext_operand): New mode attr.
4437 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
4438 don't generate (sign_extend (const_int)).
4439 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
4440 operands[2]. Use We constraint instead of <i> and
4441 <general_sext_operand> predicate instead of <general_operand>.
4442 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
4443 * config/i386/constraints.md (We): New constraint.
4444 * config/i386/predicates.md (x86_64_sext_operand,
4445 sext_operand): New predicates.
4447 2014-03-25 Martin Jambor <mjambor@suse.cz>
4450 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
4451 inconsistent devirtualizations to __builtin_unreachable.
4453 2014-03-25 Marek Polacek <polacek@redhat.com>
4456 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
4458 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
4460 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
4461 order of elements for big-endian.
4463 2014-03-25 Richard Biener <rguenther@suse.de>
4466 * gimplify-me.c (gimple_regimplify_operands): Update the
4469 2014-03-25 Martin Jambor <mjambor@suse.cz>
4472 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
4473 (lto_output_varpool_node): Likewise.
4474 (input_overwrite_node): Likewise.
4475 (input_varpool_node): Likewise.
4477 2014-03-25 Richard Biener <rguenther@suse.de>
4479 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
4480 (run_gcc): Likewise.
4482 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4484 * combine.c (simplify_compare_const): Add MODE argument.
4485 Handle mode_width 0 as very large mode_width.
4486 (try_combine, simplify_comparison): Adjust callers.
4488 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
4489 type to avoid signed integer overflow.
4490 * explow.c (plus_constant): Likewise.
4492 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
4494 * doc/generic.texi: Correct typos.
4496 2014-03-24 Tobias Burnus <burnus@net-b.de>
4498 * doc/invoke.texi (-flto): Expand section about
4499 using static libraries with LTO.
4501 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4503 PR rtl-optimization/60501
4504 * optabs.def (addptr3_optab): New optab.
4505 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
4506 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
4507 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
4509 * lra.c (emit_add3_insn): Use the addptr pattern if available.
4511 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
4513 2014-03-24 Ulrich Drepper <drepper@gmail.com>
4515 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
4518 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
4519 (_mm256_undefined_ps): Define.
4520 (_mm256_undefined_pd): Define.
4521 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
4522 (_mm_undefined_pd): Define.
4523 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
4524 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
4525 (_mm512_undefined_ps): Define.
4526 (_mm512_undefined_pd): Define.
4527 Use _mm*_undefined_*.
4528 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
4530 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
4532 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
4533 (lshr_simd): DI mode added.
4534 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
4535 (aarch64_ushr_simddi): Likewise.
4536 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
4537 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
4538 (vshrd_n_u64): Likewise.
4540 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4542 * Makefile.in (s-macro_list): Depend on cc1.
4544 2014-03-23 Teresa Johnson <tejohnson@google.com>
4546 * ipa-utils.c (ipa_print_order): Use specified dump file.
4548 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
4550 PR rtl-optimization/60601
4551 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
4553 * gcc.c (eval_spec_function): Initialize save_growing_value.
4555 2014-03-22 Jakub Jelinek <jakub@redhat.com>
4558 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
4559 code == MINUS_EXPR, never swap op0 with op1.
4561 * toplev.c (init_local_tick): Avoid signed integer multiplication
4563 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
4564 shift by first operand's bitsize.
4566 2014-03-21 Jakub Jelinek <jakub@redhat.com>
4569 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
4571 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
4572 TARGET_ISA_64BIT_P(x).
4574 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4576 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
4577 pattern for vector nor instead of subtract from splat(-1).
4578 (altivec_expand_vec_perm_const_le): Likewise.
4580 2014-03-21 Richard Henderson <rth@twiddle.net>
4583 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
4584 related insns after epilogue_completed.
4586 2014-03-21 Martin Jambor <mjambor@suse.cz>
4589 * cgraph.h (symtab_node): New flag body_removed.
4590 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
4591 when removing bodies.
4592 * symtab.c (dump_symtab_base): Dump body_removed flag.
4593 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
4594 had their bodies removed.
4596 2014-03-21 Martin Jambor <mjambor@suse.cz>
4599 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
4602 2014-03-21 Richard Biener <rguenther@suse.de>
4604 PR tree-optimization/60577
4605 * tree-core.h (struct tree_base): Document nothrow_flag use
4606 in VAR_DECL_NONALIASED.
4607 * tree.h (VAR_DECL_NONALIASED): New.
4608 (may_be_aliased): Adjust.
4609 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
4611 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4613 * expr.c (expand_expr_real_1): Remove outdated comment.
4615 2014-03-20 Jakub Jelinek <jakub@redhat.com>
4618 * ira.c (adjust_cleared_regs): Call copy_rtx on
4619 *reg_equiv[REGNO (loc)].src_p before passing it to
4620 simplify_replace_fn_rtx.
4623 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
4624 into CONST, put pic register as first operand of PLUS. Use
4625 gen_const_mem for both 32-bit and 64-bit PIC got loads.
4627 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4629 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
4631 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4633 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
4634 around for store forwarding issue in the FPU on the UT699.
4635 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
4636 loads and operations if -mfix-ut699 is specified.
4637 (divtf3_hq): Tweak attribute.
4638 (sqrttf2_hq): Likewise.
4640 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4642 * calls.c (store_one_arg): Remove incorrect const qualification on the
4643 type of the temporary.
4644 * cfgexpand.c (expand_return): Likewise.
4645 * expr.c (expand_constructor): Likewise.
4646 (expand_expr_real_1): Likewise.
4648 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4650 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
4653 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
4656 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
4658 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
4660 * config/arm/aarch-common-protos.h
4661 (alu_cost_table): Fix spelling of "extend".
4662 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
4664 2014-03-19 Richard Biener <rguenther@suse.de>
4667 * tree-core.h (tree_type_common): Re-order pointer members
4668 to reduce recursion depth during GC walks.
4670 2014-03-19 Marek Polacek <polacek@redhat.com>
4673 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
4674 before accessing it.
4676 2014-03-19 Richard Biener <rguenther@suse.de>
4679 * lto-streamer-in.c (input_function): In WPA stage do not drop
4682 2014-03-19 Jakub Jelinek <jakub@redhat.com>
4684 PR tree-optimization/60559
4685 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
4686 with build_zero_cst assignment.
4688 2014-03-18 Kai Tietz <ktietz@redhat.com>
4690 PR rtl-optimization/56356
4691 * sdbout.c (sdbout_parms): Verify that parms'
4692 incoming argument is valid.
4693 (sdbout_reg_parms): Likewise.
4695 2014-03-18 Richard Henderson <rth@redhat.com>
4698 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
4699 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
4700 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
4702 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
4704 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
4705 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
4706 Italicize plugin event names in description. Explain that
4707 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
4708 Remind that no GCC functions should be called after PLUGIN_FINISH.
4709 Explain what pragmas with expansion are.
4711 2014-03-18 Martin Liska <mliska@suse.cz>
4713 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
4714 gimple call statement is update.
4715 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
4716 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
4718 2014-03-18 Jakub Jelinek <jakub@redhat.com>
4721 * ubsan.c (ubsan_instrument_unreachable): Call
4722 initialize_sanitizer_builtins.
4723 (ubsan_pass): Likewise.
4726 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
4727 varpool_finalize_decl instead of rest_of_decl_compilation.
4729 2014-03-18 Richard Biener <rguenther@suse.de>
4731 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
4732 by using bitmap_and_compl instead of bitmap_and_compl_into.
4733 (df_rd_transfer_function): Likewise.
4735 2014-03-18 Richard Biener <rguenther@suse.de>
4737 * doc/lto.texi (fresolution): Fix typo.
4739 2014-03-18 Richard Biener <rguenther@suse.de>
4741 * doc/invoke.texi (flto): Update for changes in 4.9.
4743 2014-03-18 Richard Biener <rguenther@suse.de>
4745 * doc/loop.texi: Remove section on the removed lambda framework.
4746 Update loop docs with recent changes in preserving loop structure.
4748 2014-03-18 Richard Biener <rguenther@suse.de>
4750 * doc/lto.texi (-fresolution): Document.
4752 2014-03-18 Richard Biener <rguenther@suse.de>
4754 * doc/contrib.texi: Adjust my name.
4756 2014-03-18 Jakub Jelinek <jakub@redhat.com>
4759 * internal-fn.c: Include diagnostic-core.h.
4760 (expand_BUILTIN_EXPECT): New function.
4761 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
4762 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
4763 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
4764 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
4766 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
4767 Revert 3 argument __builtin_expect code.
4768 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
4769 * gimple-fold.c (gimple_fold_call): Likewise.
4770 * tree.h (fold_builtin_expect): New prototype.
4771 * builtins.c (build_builtin_expect_predicate): Add predictor
4772 argument, if non-NULL, create 3 argument __builtin_expect.
4773 (fold_builtin_expect): No longer static. Add ARG2 argument,
4774 pass it through to build_builtin_expect_predicate.
4775 (fold_builtin_2): Adjust caller.
4776 (fold_builtin_3): Handle BUILT_IN_EXPECT.
4777 * internal-fn.def (BUILTIN_EXPECT): New.
4779 2014-03-18 Tobias Burnus <burnus@net-b.de>
4782 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
4783 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
4784 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
4786 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
4789 * predict.c (combine_predictions_for_bb): Fix up formatting.
4790 (expr_expected_value_1, expr_expected_value): Add predictor argument,
4791 fill what it points to if non-NULL.
4792 (tree_predict_by_opcode): Adjust caller, use the predictor.
4793 * predict.def (PRED_COMPARE_AND_SWAP): Add.
4795 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
4797 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
4798 proper constant for the store mode.
4800 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
4802 * symtab.c (change_decl_assembler_name): Fix transparent alias
4805 2014-03-16 Renlin Li <Renlin.Li@arm.com>
4807 * config/aarch64/aarch64.c: Correct the comments about the
4808 aarch64 stack layout.
4810 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
4812 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
4813 check for GF_OMP_FOR_KIND_FOR.
4815 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
4817 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
4818 ymm and zmm register names.
4820 2014-03-17 Jakub Jelinek <jakub@redhat.com>
4823 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
4824 note creation for the 2010-08-31 changes.
4826 2014-03-17 Marek Polacek <polacek@redhat.com>
4829 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
4830 as -fno-tree-loop-vectorize.
4831 (expand_omp_simd): Likewise.
4833 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
4835 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
4836 (eligible_for_call_delay): New prototype.
4837 * config/sparc/sparc.c (tls_call_delay): Rename into...
4838 (eligible_for_call_delay): ...this. Return false if the instruction
4839 cannot be put in the delay slot of a branch.
4840 (eligible_for_restore_insn): Simplify.
4841 (eligible_for_return_delay): Return false if the instruction cannot be
4842 put in the delay slot of a branch and simplify.
4843 (eligible_for_sibcall_delay): Return false if the instruction cannot be
4844 put in the delay slot of a branch.
4845 * config/sparc/sparc.md (fix_ut699): New attribute.
4846 (tls_call_delay): Delete.
4847 (in_call_delay): Reimplement.
4848 (eligible_for_sibcall_delay): Rename into...
4849 (in_sibcall_delay): ...this.
4850 (eligible_for_return_delay): Rename into...
4851 (in_return_delay): ...this.
4852 (in_branch_delay): Reimplement.
4853 (in_uncond_branch_delay): Delete.
4854 (in_annul_branch_delay): Delete.
4856 2014-03-14 Richard Henderson <rth@redhat.com>
4859 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
4860 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
4861 (*floathi<X87MODEF>2_i387_with_temp): Remove.
4862 (floathi splitters): Remove.
4863 (float<SWI48x>xf2): New pattern.
4864 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
4865 code that tried to handle DImode for 32-bit, but which was excluded
4866 by the pattern's condition. Drop allocation of stack temporary.
4867 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
4868 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
4869 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
4870 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
4871 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
4872 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
4873 (*float<SWI48><MODEF>2_sse_interunit): Remove.
4874 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
4875 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
4876 (*float<SWI48x><X87MODEF>2_i387): Remove.
4877 (all float _with_temp splitters): Remove.
4878 (*float<SWI48x><MODEF>2_i387): New pattern.
4879 (*float<SWI48><MODEF>2_sse): New pattern.
4880 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
4881 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
4883 2014-03-14 Jakub Jelinek <jakub@redhat.com>
4884 Marek Polacek <polacek@redhat.com>
4887 * common.opt (dump_base_name_prefixed): New Variable.
4888 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
4889 if x_dump_base_name_prefixed is already set, set it at the end.
4891 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
4893 PR rtl-optimization/60508
4894 * lra-constraints.c (get_reload_reg): Add new parameter
4896 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
4897 Pass the new parameter values.
4899 2014-03-14 Richard Biener <rguenther@suse.de>
4901 * common.opt: Revert unintented changes from r205065.
4904 2014-03-14 Richard Biener <rguenther@suse.de>
4907 * cfghooks.c (split_block): Properly adjust all loops the
4908 block was a latch of.
4910 2014-03-14 Martin Jambor <mjambor@suse.cz>
4913 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
4916 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
4919 * config/avr/avr.c (avr_set_current_function): Pass function name
4920 through default_strip_name_encoding before sanity checking instead
4921 of skipping the first char of the assembler name.
4923 2014-03-13 Richard Henderson <rth@redhat.com>
4926 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
4927 (ix86_force_to_memory, ix86_free_from_memory): Remove.
4928 * config/i386/i386-protos.h: Likewise.
4929 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
4930 in the expander instead of a splitter.
4931 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
4932 any possibility of requiring a memory.
4933 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
4934 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
4935 (fp branch splitters): Update for ix86_split_fp_branch.
4936 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
4937 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
4938 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
4939 (*fop_<MODEF>_2_i387): Remove f/r alternative.
4940 (*fop_<MODEF>_3_i387): Likewise.
4941 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
4942 (splitters for the fop_* register patterns): Remove.
4943 (fscalexf4_i387): Rename from *fscalexf4_i387.
4944 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
4946 2014-03-13 Jakub Jelinek <jakub@redhat.com>
4948 PR tree-optimization/59779
4949 * tree-dfa.c (get_ref_base_and_extent): Use double_int
4950 type for bitsize and maxsize instead of HOST_WIDE_INT.
4952 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
4954 PR rtl-optimization/57320
4955 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
4956 the CFG after thread_prologue_and_epilogue_insns.
4958 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
4960 PR rtl-optimization/57189
4961 * lra-constraints.c (process_alt_operands): Disfavor spilling
4964 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
4966 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
4968 2014-03-13 Jakub Jelinek <jakub@redhat.com>
4970 PR tree-optimization/59025
4972 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
4973 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
4975 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
4978 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
4979 calls of avr_out_plus_1.
4981 2014-03-13 Bin Cheng <bin.cheng@arm.com>
4983 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
4984 BB's single pred and update the father loop's latch info later.
4986 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
4988 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
4992 (VEC_base): Likewise.
4993 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
4994 registers, we need to swap double words in little endian mode.
4996 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
4997 to be a container mode for 128-bit integer operations added in ISA
4998 2.07. Unlike TImode and PTImode, the preferred register set is
4999 the Altivec/VMX registers for the 128-bit operations.
5001 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
5003 (rs6000_split_128bit_ok_p): Likewise.
5005 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
5006 macros for creating ISA 2.07 normal and overloaded builtin
5007 functions with 3 arguments.
5008 (BU_P8V_OVERLOAD_3): Likewise.
5009 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
5010 for use as overloaded functions.
5011 (VPERM_1TI_UNS): Likewise.
5012 (VSEL_1TI): Likewise.
5013 (VSEL_1TI_UNS): Likewise.
5014 (ST_INTERNAL_1ti): Likewise.
5015 (LD_INTERNAL_1ti): Likewise.
5016 (XXSEL_1TI): Likewise.
5017 (XXSEL_1TI_UNS): Likewise.
5018 (VPERM_1TI): Likewise.
5019 (VPERM_1TI_UNS): Likewise.
5020 (XXPERMDI_1TI): Likewise.
5021 (SET_1TI): Likewise.
5022 (LXVD2X_V1TI): Likewise.
5023 (STXVD2X_V1TI): Likewise.
5024 (VEC_INIT_V1TI): Likewise.
5025 (VEC_SET_V1TI): Likewise.
5026 (VEC_EXT_V1TI): Likewise.
5027 (EQV_V1TI): Likewise.
5028 (NAND_V1TI): Likewise.
5029 (ORC_V1TI): Likewise.
5030 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
5031 added in ISA 2.07. Add both normal 'altivec' builtins, and the
5033 (VADDUQM): Likewise.
5034 (VSUBCUQ): Likewise.
5035 (VADDEUQM): Likewise.
5036 (VADDECUQ): Likewise.
5037 (VSUBEUQM): Likewise.
5038 (VSUBECUQ): Likewise.
5040 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5041 __int128_t and __uint128_t types.
5042 (__uint128_type): Likewise.
5043 (altivec_categorize_keyword): Add support for vector __int128_t,
5044 vector __uint128_t, vector __int128, and vector unsigned __int128
5045 as a container type for TImode operations that need to be done in
5046 VSX/Altivec registers.
5047 (rs6000_macro_to_expand): Likewise.
5048 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5049 to support 128-bit integer instructions vaddcuq, vadduqm,
5050 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5051 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5053 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5054 for V1TImode, and set up preferences to use VSX/Altivec registers.
5055 Setup VSX reload handlers.
5056 (rs6000_debug_reg_global): Likewise.
5057 (rs6000_init_hard_regno_mode_ok): Likewise.
5058 (rs6000_preferred_simd_mode): Likewise.
5059 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5060 (easy_altivec_constant): Likewise.
5061 (output_vec_const_move): Likewise.
5062 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5064 (rs6000_expand_vector_extract): Likewise.
5065 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
5067 (rs6000_const_vec): Add support for V1TImode.
5068 (rs6000_emit_le_vsx_load): Swap double words when loading or
5069 storing TImode/V1TImode.
5070 (rs6000_emit_le_vsx_store): Likewise.
5071 (rs6000_emit_le_vsx_move): Likewise.
5072 (rs6000_emit_move): Add support for V1TImode.
5073 (altivec_expand_ld_builtin): Likewise.
5074 (altivec_expand_st_builtin): Likewise.
5075 (altivec_expand_vec_init_builtin): Likewise.
5076 (altivec_expand_builtin): Likewise.
5077 (rs6000_init_builtins): Add support for V1TImode type. Add
5078 support for ISA 2.07 128-bit integer builtins. Define type names
5079 for the VSX/Altivec vector types.
5080 (altivec_init_builtins): Add support for overloaded vector
5081 functions with V1TImode type.
5082 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
5083 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
5085 (rs6000_split_128bit_ok_p): Likewise.
5086 (rs6000_handle_altivec_attribute): Create V1TImode from vector
5087 __int128_t and vector __uint128_t.
5089 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
5090 and mode attributes.
5097 (VS_scalar): Likewise.
5098 (VS_double): Likewise.
5099 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
5101 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
5102 we support the ISA 2.07 128-bit integer arithmetic instructions.
5103 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
5104 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
5105 and TImode types for use with the builtin functions.
5106 (V1TI_type_node): Likewise.
5107 (unsigned_V1TI_type_node): Likewise.
5108 (intTI_type_internal_node): Likewise.
5109 (uintTI_type_internal_node): Likewise.
5111 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
5112 128-bit builtin functions.
5113 (UNSPEC_VADDEUQM): Likewise.
5114 (UNSPEC_VADDECUQ): Likewise.
5115 (UNSPEC_VSUBCUQ): Likewise.
5116 (UNSPEC_VSUBEUQM): Likewise.
5117 (UNSPEC_VSUBECUQ): Likewise.
5118 (VM): Add V1TImode to vector mode iterators.
5120 (VI_unit): Likewise.
5121 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
5122 (altivec_vaddcuq): Likewise.
5123 (altivec_vsubuqm): Likewise.
5124 (altivec_vsubcuq): Likewise.
5125 (altivec_vaddeuqm): Likewise.
5126 (altivec_vaddecuq): Likewise.
5127 (altivec_vsubeuqm): Likewise.
5128 (altivec_vsubecuq): Likewise.
5130 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
5132 (BOOL_128): Likewise.
5133 (BOOL_REGS_OUTPUT): Likewise.
5134 (BOOL_REGS_OP1): Likewise.
5135 (BOOL_REGS_OP2): Likewise.
5136 (BOOL_REGS_UNARY): Likewise.
5137 (BOOL_REGS_AND_CR0): Likewise.
5139 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
5140 128-bit integer builtin support.
5141 (vec_vadduqm): Likewise.
5142 (vec_vaddecuq): Likewise.
5143 (vec_vaddeuqm): Likewise.
5144 (vec_vsubecuq): Likewise.
5145 (vec_vsubeuqm): Likewise.
5146 (vec_vsubcuq): Likewise.
5147 (vec_vsubuqm): Likewise.
5149 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5150 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
5151 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
5152 128-bit integer add/subtract to ISA 2.07.
5154 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
5156 * config/arc/arc.c (arc_predicate_delay_insns):
5157 Fix third argument passed to conditionalize_nonjump.
5159 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
5161 * config/aarch64/aarch64-builtins.c
5162 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
5163 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
5164 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
5165 instead of __builtin_lfloor.
5166 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
5168 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5170 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
5171 (tree_ssa_ifcombine_bb_1): New function.
5172 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
5173 is an empty forwarder block to then_bb or vice versa and then_bb
5174 and else_bb are effectively swapped.
5176 2014-03-12 Christian Bruel <christian.bruel@st.com>
5179 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
5180 REG_CFA_DEF_CFA note.
5181 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
5182 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
5184 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
5186 PR tree-optimization/60454
5187 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
5189 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5191 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
5192 Do not define target_cpu_default2 to generic.
5193 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
5194 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
5195 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
5197 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5198 Marc Glisse <marc.glisse@inria.fr>
5200 PR tree-optimization/60502
5201 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
5202 instead of build_low_bits_mask.
5204 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5207 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
5208 if there are multiple uses, but op doesn't live on E edge.
5209 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
5210 clobber stmts before __builtin_unreachable.
5212 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
5214 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
5215 hard_frame_pointer_rtx.
5216 * cse.c (cse_insn): Remove volatile check.
5217 * cselib.c (cselib_process_insn): Likewise.
5218 * dse.c (scan_insn): Likewise.
5220 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5222 * config/arc/arc.c (conditionalize_nonjump): New function,
5224 (arc_ifcvt): ... this.
5225 (arc_predicate_delay_insns): Use it.
5227 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5229 * config/arc/predicates.md (extend_operand): During/after reload,
5230 allow const_int_operand.
5231 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
5232 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
5233 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
5235 (umulsi3_highpart_i): Likewise.
5237 2014-03-11 Richard Biener <rguenther@suse.de>
5239 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
5240 Add asserts to guard possible wrong-code bugs.
5242 2014-03-11 Richard Biener <rguenther@suse.de>
5244 PR tree-optimization/60429
5245 PR tree-optimization/60485
5246 * tree-ssa-structalias.c (set_union_with_increment): Properly
5247 take into account all fields that overlap the shifted vars.
5248 (do_sd_constraint): Likewise.
5249 (do_ds_constraint): Likewise.
5250 (get_constraint_for_ptr_offset): Likewise.
5252 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
5254 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
5255 (nios2_compute_frame_layout):
5256 Add calculation of cfun->machine->fp_save_offset.
5257 (nios2_expand_prologue): Correct setting of frame pointer register
5259 (nios2_expand_epilogue): Update recovery of stack pointer from
5260 frame pointer accordingly.
5261 (nios2_initial_elimination_offset): Update calculation of offset
5262 for eliminating to HARD_FRAME_POINTER_REGNUM.
5264 2014-03-10 Jakub Jelinek <jakub@redhat.com>
5267 * ipa.c (symtab_remove_unreachable_nodes): Don't call
5268 cgraph_get_create_node on VAR_DECLs.
5270 2014-03-10 Richard Biener <rguenther@suse.de>
5273 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
5275 2014-03-08 Douglas B Rupp <rupp@gnat.com>
5277 * config/vms/vms.opt (vms_float_format): New variable.
5279 2014-03-08 Tobias Burnus <burnus@net-b.de>
5281 * doc/invoke.texi (-fcilkplus): Update implementation status.
5283 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
5284 Richard Biener <rguenther@suse.de>
5286 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
5287 consistently accross all TUs.
5288 (run_gcc): Enable -fshort-double automatically at link at link-time
5289 and disallow override.
5291 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
5294 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
5295 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
5296 if they can't be used.
5298 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5300 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
5301 for Solaris 11/x86 ld.
5302 * configure: Regenerate.
5304 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5306 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
5307 (LIB_TLS_SPEC): Save as ld_tls_libs.
5308 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
5309 (HAVE_AS_IX86_TLSLDM): New test.
5310 * configure, config.in: Regenerate.
5311 * config/i386/i386.c (legitimize_tls_address): Fall back to
5312 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
5313 cannot support TLS_MODEL_LOCAL_DYNAMIC.
5314 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
5315 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
5317 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
5319 * common.opt (fira-loop-pressure): Mark as optimization.
5321 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
5323 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
5324 an OpenMP mappable type.
5326 2014-03-06 Matthias Klose <doko@ubuntu.com>
5328 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
5329 MULTILIB_OSDIRNAMES is not defined.
5331 2014-03-06 Jakub Jelinek <jakub@redhat.com>
5332 Meador Inge <meadori@codesourcery.com>
5335 * config/arm/arm.c (arm_tls_symbol_p): Remove.
5336 (arm_legitimize_address): Call legitimize_tls_address for any
5337 arm_tls_referenced_p expression, handle constant addend. Call it
5338 before testing for !TARGET_ARM.
5339 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
5341 2014-03-06 Richard Biener <rguenther@suse.de>
5347 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5349 * tree-streamer.c (record_common_node): Assert we don't record
5350 nodes with type double.
5351 (preload_common_node): Skip type double, complex double and double
5352 pointer since it is now frontend dependent due to fshort-double option.
5354 2014-03-06 Richard Biener <rguenther@suse.de>
5356 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
5357 or -fno-lto is specified and the linker has full plugin support.
5358 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
5359 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
5360 * lto-wrapper.c (merge_and_complain): Merge compile-time
5361 optimization levels.
5362 (run_gcc): And pass it through to the link options.
5364 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
5368 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5370 * cselib.c (remove_useless_values): Skip to avoid quadratic
5371 behavior if the condition moved from...
5372 (cselib_process_insn): ... here holds.
5374 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5377 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
5378 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
5381 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
5382 (TM_H): Add x86-tune.def.
5384 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5386 * config/aarch64/aarch64.c (generic_tunings):
5387 Use cortexa57_extra_costs.
5389 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5392 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
5393 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
5394 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
5397 2014-03-04 Heiher <r@hev.cc>
5399 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
5400 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
5402 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
5404 * config/i386/predicates.md (const2356_operand): Change to ...
5405 (const2367_operand): ... this.
5406 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
5408 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5409 (*avx512pf_scatterpf<mode>sf): Ditto.
5410 (avx512pf_scatterpf<mode>df): Ditto.
5411 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5412 (*avx512pf_scatterpf<mode>df): Ditto.
5413 * config/i386/i386.c (ix86_expand_builtin): Update
5414 incorrect hint operand error message.
5416 2014-03-04 Richard Biener <rguenther@suse.de>
5418 * lto-section-in.c (lto_get_section_data): Fix const cast.
5420 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5422 * tree-streamer.c (record_common_node): Assert we don't record
5423 nodes with type double.
5424 (preload_common_node): Skip type double, complex double and double
5425 pointer since it is now frontend dependent due to fshort-double option.
5427 2014-03-04 Richard Biener <rguenther@suse.de>
5430 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
5431 (lto_input_toplevel_asms): Likewise.
5432 * lto-section-in.c (lto_get_section_data): Instead do it here
5435 2014-03-04 Richard Biener <rguenther@suse.de>
5437 PR tree-optimization/60382
5438 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
5439 dead PHIs a reduction.
5441 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
5443 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
5445 (_mm_prefetch): Move out of GCC target("sse") pragma.
5446 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
5447 GCC target("prfchw") pragma.
5448 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
5450 * config/i386/i386.c (ix86_option_override_internal): Enable
5451 -mprfchw with -mprefetchwt1.
5453 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5455 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
5458 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5460 * opts.h (CL_PCH_IGNORE): Define.
5461 * targhooks.c (option_affects_pch_p):
5462 Return false for options that have CL_PCH_IGNORE set.
5463 * opt-functions.awk: Process PchIgnore.
5464 * doc/options.texi: Document PchIgnore.
5466 * config/arc/arc.opt (misize): Add PchIgnore property.
5468 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5470 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
5471 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
5472 constraint on constants to permit them being loaded into
5473 GENERAL_REGS or BASE_REGS.
5475 2014-03-03 Nick Clifton <nickc@redhat.com>
5477 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
5478 anti-cacnonical alternatives.
5479 (negandhi3_real): New pattern.
5480 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
5482 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5484 * config/avr/avr-mcus.def: Remove atxmega16x1.
5485 * config/avr/avr-tables.opt: Regenerate.
5486 * config/avr/t-multilib: Regenerate.
5487 * doc/avr-mmcu.texi: Regenerate.
5489 2014-03-03 Tobias Grosser <tobias@grosser.es>
5490 Mircea Namolaru <mircea.namolaru@inria.fr>
5492 PR tree-optimization/58028
5493 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
5496 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5498 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
5499 not handled by recognizers.
5501 2014-03-03 Jakub Jelinek <jakub@redhat.com>
5504 * function.c (expand_function_end): Don't emit
5505 clobber_return_register sequence if clobber_after is a BARRIER.
5506 * cfgexpand.c (construct_exit_block): Append instructions before
5507 return_label to prev_bb.
5509 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5511 * config/rs6000/constraints.md: Document reserved use of "wc".
5513 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5516 * ipa.c (function_and_variable_visibility): When dissolving comdat
5517 group, also set all symbols to local.
5519 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5524 2013-12-14 Jan Hubicka <jh@suse.cz>
5526 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
5528 2014-03-02 Jon Beniston <jon@beniston.com>
5534 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
5535 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
5536 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
5537 (simple_return, *simple_return): New patterns
5538 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
5539 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
5541 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
5543 * dwarf2out.c (gen_subprogram_die): Tidy.
5545 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
5548 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
5549 (*mov_t_msb_neg_negc): ... this new insn.
5551 2014-02-28 Jason Merrill <jason@redhat.com>
5554 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
5557 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
5560 * dwarf2out.c (decltype_auto_die): New static.
5561 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
5562 (gen_type_die_with_usage): Handle 'decltype(auto)'.
5563 (is_cxx_auto): Likewise.
5565 2014-02-28 Ian Bolton <ian.bolton@arm.com>
5567 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
5568 we are not using general regs only.
5570 2014-02-28 Richard Biener <rguenther@suse.de>
5573 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
5574 previous fix and only allow to remove trivial pre-headers
5575 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
5576 (remove_forwarder_block): Properly update the latch of a loop.
5578 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5581 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
5582 (cselib_preserved_hash_table): New.
5583 (preserve_constants_and_equivs): Move preserved vals to it.
5584 (cselib_find_slot): Look it up first.
5585 (cselib_init): Initialize it.
5586 (cselib_finish): Release it.
5587 (dump_cselib_table): Dump it.
5589 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5592 * cselib.c (remove_useless_values): Skip to avoid quadratic
5593 behavior if the condition moved from...
5594 (cselib_process_insn): ... here holds.
5596 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5599 * var-tracking.c (vt_initialize): Apply the same condition to
5600 preserve the CFA base value.
5602 2014-02-28 Joey Ye <joey.ye@arm.com>
5605 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
5606 if reload in progress or completed.
5608 2014-02-28 Tobias Burnus <burnus@net-b.de>
5611 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
5614 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
5616 * doc/tm.texi.in (Condition Code Status): Update documention for
5617 relative locations of cc0-setter and cc0-user.
5619 2014-02-27 Jeff Law <law@redhat.com>
5621 PR rtl-optimization/52714
5622 * combine.c (try_combine): When splitting an unrecognized PARALLEL
5623 into two independent simple sets, if I3 is a jump, ensure the
5624 pattern we place into I3 is a (set (pc) ...).
5626 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
5627 Jeff Law <law@redhat.com>
5629 PR rtl-optimization/49847
5630 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
5631 are in different blocks.
5632 * doc/tm.texi (Condition Code Status): Update documention for
5633 relative locations of cc0-setter and cc0-user.
5635 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
5638 * lra.c (lra_emit_add): Check SUBREG too.
5640 2014-02-27 Andreas Schwab <schwab@suse.de>
5642 * config/m68k/m68k.c (m68k_option_override): Disable
5643 -flive-range-shrinkage for classic m68k.
5644 (m68k_override_options_after_change): Likewise.
5646 2014-02-27 Marek Polacek <polacek@redhat.com>
5649 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
5650 -Wmaybe-uninitialized.
5652 2014-02-27 Alan Modra <amodra@gmail.com>
5655 * reload1.c (emit_input_reload_insns): When reload_override_in,
5656 set old to rl->in_reg when rl->in_reg is a subreg.
5658 2014-02-26 Richard Biener <rguenther@suse.de>
5661 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
5663 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
5665 * common/config/i386/predicates.md (const1256_operand): Remove.
5666 (const2356_operand): New.
5667 (const_1_to_2_operand): Remove.
5668 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5669 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5670 (*avx512pf_gatherpf<mode>sf): Ditto.
5671 (avx512pf_gatherpf<mode>df): Ditto.
5672 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5673 (*avx512pf_gatherpf<mode>df): Ditto.
5674 (avx512pf_scatterpf<mode>sf): Ditto.
5675 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5676 (*avx512pf_scatterpf<mode>sf): Ditto.
5677 (avx512pf_scatterpf<mode>df): Ditto.
5678 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5679 (*avx512pf_scatterpf<mode>df): Ditto.
5680 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
5682 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
5684 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
5685 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
5686 (_mm512_mask_testn_epi64_mask): Move to ...
5687 * config/i386/avx512cdintrin.h: Here.
5688 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
5689 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
5690 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
5691 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
5692 TARGET_AVX512F from TARGET_AVX512CD.
5694 2014-02-26 Richard Biener <rguenther@suse.de>
5697 * ipa.c (walk_polymorphic_call_targets): Properly guard
5698 call to inline_update_overall_summary.
5700 2014-02-26 Bin Cheng <bin.cheng@arm.com>
5703 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
5704 and latches only if requested. Fix latch if it is removed.
5705 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
5706 LOOPS_HAVE_PREHEADERS.
5708 2014-02-25 Andrew Pinski <apinski@cavium.com>
5710 * builtins.c (expand_builtin_thread_pointer): Create a new target
5711 when the target is NULL.
5713 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
5715 PR rtl-optimization/60317
5716 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
5717 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
5718 * lra-assigns.c: Include params.h.
5719 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
5720 other reload pseudos considerations.
5722 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5724 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
5725 to use canonical form for nor<mode>3.
5727 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5730 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
5733 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
5735 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
5736 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
5737 (ix86_handle_option): Handle OPT_mprefetchwt1.
5738 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
5739 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
5741 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
5742 OPTION_MASK_ISA_PREFETCHWT1.
5743 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
5744 (PTA_PREFETCHWT1): New.
5745 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
5746 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
5747 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
5748 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
5749 (*prefetch_avx512pf_<mode>_: Change into ...
5750 (*prefetch_prefetchwt1_<mode>: This.
5751 * config/i386/i386.opt (mprefetchwt1): New.
5752 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
5753 (_mm_prefetch): Handle intent to write.
5754 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
5756 2014-02-25 Richard Biener <rguenther@suse.de>
5759 * emit-rtl.c (mem_attrs_htab): Remove.
5760 (mem_attrs_htab_hash): Likewise.
5761 (mem_attrs_htab_eq): Likewise.
5762 (set_mem_attrs): Always allocate new mem-attrs when something changed.
5763 (init_emit_once): Do not allocate mem_attrs_htab.
5765 2014-02-25 Richard Biener <rguenther@suse.de>
5768 * lto-opts.c (lto_write_options): Output non-explicit conservative
5769 -fwrapv, -fno-trapv and -fno-strict-overflow.
5770 * lto-wrapper.c (merge_and_complain): Handle merging those options.
5771 (run_gcc): And pass them through.
5773 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
5775 * sel-sched.c (calculate_new_fences): New parameter ptime.
5776 Calculate it as a maximum over all fence cycles.
5777 (sel_sched_region_2): Adjust the call to calculate_new_fences.
5778 Print the final schedule timing when sched_verbose.
5780 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
5782 PR rtl-optimization/60292
5783 * sel-sched.c (fill_vec_av_set): Do not reset target availability
5784 bit fot the fence instruction.
5786 2014-02-24 Alangi Derick <alangiderick@gmail.com>
5788 * calls.h: Fix typo in comment.
5790 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
5792 * config/pa/pa.c (pa_output_move_double): Don't valididate when
5793 adjusting offsetable addresses.
5795 2014-02-24 Guozhi Wei <carrot@google.com>
5797 * sparseset.h (sparseset_pop): Fix the wrong index.
5799 2014-02-24 Walter Lee <walt@tilera.com>
5801 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
5802 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
5804 * common/config/tilegx/tilegx-common.c
5805 (TARGET_DEFAULT_TARGET_FLAGS): Define.
5806 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
5808 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
5809 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
5810 (tilegx_gimplify_va_arg_expr): Handle big endian.
5811 (tilegx_expand_unaligned_load): Ditto.
5812 (tilegx_expand_unaligned_store): Ditto.
5813 (TARGET_RETURN_IN_MSB): New.
5814 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
5815 (TARGET_ENDIAN_DEFAULT): New.
5816 (TARGET_BIG_ENDIAN): Handle big endian.
5817 (BYTES_BIG_ENDIAN): Ditto.
5818 (WORDS_BIG_ENDIAN): Ditto.
5819 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
5822 * config/tilegx/tilegx.md (extv): Handle big endian.
5824 (insn_st<n>): Ditto.
5825 (insn_st<n>_add<bitsuffix>): Ditto.
5826 (insn_stnt<n>): Ditto.
5827 (insn_stnt<n>_add<bitsuffix>):Ditto.
5828 (vec_interleave_highv8qi): Handle big endian.
5829 (vec_interleave_highv8qi_be): New.
5830 (vec_interleave_highv8qi_le): New.
5831 (insn_v1int_h): Handle big endian.
5832 (vec_interleave_lowv8qi): Handle big endian.
5833 (vec_interleave_lowv8qi_be): New.
5834 (vec_interleave_lowv8qi_le): New.
5835 (insn_v1int_l): Handle big endian.
5836 (vec_interleave_highv4hi): Handle big endian.
5837 (vec_interleave_highv4hi_be): New.
5838 (vec_interleave_highv4hi_le): New.
5839 (insn_v2int_h): Handle big endian.
5840 (vec_interleave_lowv4hi): Handle big endian.
5841 (vec_interleave_lowv4hi_be): New.
5842 (vec_interleave_lowv4hi_le): New.
5843 (insn_v2int_l): Handle big endian.
5844 (vec_interleave_highv2si): Handle big endian.
5845 (vec_interleave_highv2si_be): New.
5846 (vec_interleave_highv2si_le): New.
5847 (insn_v4int_h): Handle big endian.
5848 (vec_interleave_lowv2si): Handle big endian.
5849 (vec_interleave_lowv2si_be): New.
5850 (vec_interleave_lowv2si_le): New.
5851 (insn_v4int_l): Handle big endian.
5852 * config/tilegx/tilegx.opt (mbig-endian): New option.
5853 (mlittle-endian): New option.
5854 * doc/install.texi: Document tilegxbe-linux.
5855 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
5857 2014-02-24 Martin Jambor <mjambor@suse.cz>
5860 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
5861 there are no parameter descriptors.
5863 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
5865 PR rtl-optimization/60268
5866 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
5867 initialization to ...
5868 (sched_rgn_init): ... here.
5869 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
5871 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5873 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
5876 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
5878 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
5881 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5883 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
5884 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
5886 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5888 * config/microblaze/predicates.md: Add cmp_op predicate.
5889 * config/microblaze/microblaze.md: Add branch_compare instruction
5890 which uses cmp_op predicate and emits cmp insn before branch.
5891 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
5892 to microblaze_expand_conditional_branch and consolidate logic.
5893 (microblaze_expand_conditional_branch): emit branch_compare
5894 insn instead of handling cmp op separate from branch insn.
5896 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5898 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
5901 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5903 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
5904 define_insn with define_expand and new define_insn
5905 *altivec_lve<VI_char>x_internal.
5906 (altivec_stve<VI_char>x): Replace define_insn with define_expand
5907 and new define_insn *altivec_stve<VI_char>x_internal.
5908 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
5910 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
5912 (altivec_expand_stvex_be): New function.
5914 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
5916 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
5917 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
5918 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
5919 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
5921 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
5924 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
5925 instead of emit_move_insn.
5927 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5929 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
5931 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
5932 gen_altivec_vsumsws.
5934 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5936 * config/rs6000/altivec.md (altivec_lvxl): Rename as
5937 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
5938 (altivec_lvxl_<mode>): New define_expand incorporating
5939 -maltivec=be semantics where needed.
5940 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
5941 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
5942 semantics where needed.
5943 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
5944 (altivec_stvx_<mode>): New define_expand incorporating
5945 -maltivec=be semantics where needed.
5946 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
5947 VM2 iterator instead of V4SI.
5948 (altivec_stvxl_<mode>): New define_expand incorporating
5949 -maltivec=be semantics where needed.
5950 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
5951 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
5952 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
5953 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
5954 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
5955 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
5956 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
5957 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
5958 ALTIVEC_BUILTIN_STVXL.
5959 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
5960 (altivec_expand_stvx_be): Likewise.
5961 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
5962 (altivec_expand_lvx_be): Likewise.
5963 (altivec_expand_stvx_be): Likewise.
5964 (altivec_expand_builtin): Add cases for
5965 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
5966 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
5967 (altivec_init_builtins): Add definitions for
5968 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
5969 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
5971 2014-02-21 Catherine Moore <clm@codesourcery.com>
5973 * doc/invoke.texi (mvirt, mno-virt): Document.
5974 * config/mips/mips.opt (mvirt): New option.
5975 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
5977 2014-02-21 Richard Biener <rguenther@suse.de>
5979 PR tree-optimization/60276
5980 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
5981 (STMT_VINFO_MIN_NEG_DIST): New macro.
5982 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
5983 STMT_VINFO_MIN_NEG_DIST.
5984 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
5985 made for negative dependence distances still hold.
5987 2014-02-21 Richard Biener <rguenther@suse.de>
5990 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
5991 DECL_INITIAL for globals not in the current function context.
5993 2014-02-21 Jakub Jelinek <jakub@redhat.com>
5995 PR tree-optimization/56490
5996 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
5997 * tree-ssa-uninit.c: Include params.h.
5998 (compute_control_dep_chain): Add num_calls argument, return false
5999 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
6000 num_calls to recursive call.
6001 (find_predicates): Change dep_chain into normal array,
6002 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
6003 variable and adjust compute_control_dep_chain caller.
6004 (find_def_preds): Likewise.
6006 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
6008 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
6009 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
6011 2014-02-21 Nick Clifton <nickc@redhat.com>
6013 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
6014 (pushhi1): Likewise.
6015 (popqi1): Add mode to pre_dec.
6018 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6020 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
6021 mode for mask of V8SFmode permutation.
6023 2014-02-20 Richard Henderson <rth@redhat.com>
6026 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
6027 a new pseudo for OLDVAL.
6029 2014-02-20 Jakub Jelinek <jakub@redhat.com>
6032 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
6033 gen_reg_rtx if d->testing_p.
6034 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
6035 if d->testing_p and we will certainly return true.
6036 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6039 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6041 * emit-rtl.c (gen_reg_rtx): Assert that
6042 crtl->emit.regno_pointer_align_length is non-zero.
6044 2014-02-20 Richard Henderson <rth@redhat.com>
6047 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6048 on failure the store back into EXPECT.
6050 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6051 Sandra Loosemore <sandra@codesourcery.com>
6053 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6054 * config/nios2/nios2.c (nios2_function_profiler): Add
6055 -fPIC (flag_pic == 2) support.
6056 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6057 (nios2_large_offset_p): New function.
6058 (nios2_unspec_reloc_p): Move up position, update to use
6059 nios2_large_offset_p.
6060 (nios2_unspec_address): Remove function.
6061 (nios2_unspec_offset): New function.
6062 (nios2_large_got_address): New function.
6063 (nios2_got_address): Add large offset support.
6064 (nios2_legitimize_tls_address): Update usage of removed and new
6066 (nios2_symbol_binds_local_p): New function.
6067 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
6068 (nios2_legitimize_address): Update to use nios2_large_offset_p.
6069 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
6070 (nios2_print_operand): Merge H/L processing, add hiadj/lo
6071 processing for (const (unspec ...)).
6072 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
6074 2014-02-20 Richard Biener <rguenther@suse.de>
6076 * tree-cfg.c (replace_uses_by): Mark altered BBs before
6077 doing the substitution.
6078 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
6080 2014-02-20 Martin Jambor <mjambor@suse.cz>
6083 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
6084 info when checking whether lattices are bottom.
6086 2014-02-20 Richard Biener <rguenther@suse.de>
6089 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
6092 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
6095 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
6096 parameter specifying the scaling.
6097 (inline_call): Update.
6098 (want_inline_recursively): Guard division by zero.
6099 (recursive_inlining): Update.
6100 * ipa-inline.h (clone_inlined_nodes): Update.
6102 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6105 * config/i386/i386.c (classify_argument): Pass structures of size
6106 64 bytes or less in register.
6108 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6109 Kirill Yukhin <kirill.yukhin@intel.com>
6111 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
6112 (_mm_rcp28_round_ss): Ditto.
6113 (_mm_rsqrt28_round_sd): Ditto.
6114 (_mm_rsqrt28_round_ss): Ditto.
6115 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
6116 (_mm_rcp14_round_ss): Ditto.
6117 (_mm_rsqrt14_round_sd): Ditto.
6118 (_mm_rsqrt14_round_ss): Ditto.
6119 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
6120 the first input operand, get rid of match_dup.
6121 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
6123 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
6125 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
6126 operand as the first input operand, set type attribute.
6127 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
6129 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
6130 operand as the first input operand, set type attribute.
6132 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6134 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
6137 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
6140 * config/i386/i386.c (construct_container): Remove TFmode check
6141 for X86_64_INTEGER_CLASS.
6143 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
6146 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
6147 only when -Wpsabi is enabled.
6149 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
6152 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
6153 passing arrays in registers are the same as for structs, so remove the
6154 special case for them.
6156 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
6158 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
6159 destination type, extract only the valid bits if the source type is not
6160 integral and has a different mode.
6162 2014-02-19 Richard Biener <rguenther@suse.de>
6165 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
6168 2014-02-19 Richard Biener <rguenther@suse.de>
6171 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
6172 (ipa_modify_call_arguments): Emit an argument load explicitely and
6173 preserve virtual SSA form there and for the replacement call.
6174 Do not update SSA form nor free dominance info.
6176 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6178 * ipa.c (function_and_variable_visibility): Also clear WEAK
6179 flag when disolving COMDAT_GROUP.
6181 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6183 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
6184 * ipa-prop.c (ipa_set_jf_known_type): Return early when
6186 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
6187 do more sanity checks.
6188 (detect_type_change): Return true when giving up early.
6189 (compute_complex_assign_jump_func): Fix type parameter of
6190 ipa_set_ancestor_jf.
6191 (compute_complex_ancestor_jump_func): Likewise.
6192 (update_jump_functions_after_inlining): Fix updating of
6194 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
6196 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6198 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
6199 inline clones when edge disappears.
6201 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6204 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
6205 Split 64-bit moves into 2 patterns. Do not allow the use of
6206 direct move for TDmode in little endian, since the decimal value
6207 has little endian bytes within a word, but the 64-bit pieces are
6208 ordered in a big endian fashion, and normal subreg's of TDmode are
6210 (mov<mode>_64bit_dm): Likewise.
6211 (movtd_64bit_nodm): Likewise.
6213 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6215 PR tree-optimization/60174
6216 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
6217 statement of an SSA_NAME that occurs in an abnormal PHI node.
6219 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6222 * final.c (SEEN_BB): Remove.
6223 (SEEN_NOTE, SEEN_EMITTED): Renumber.
6224 (final_scan_insn): Don't force_source_line on second
6225 NOTE_INSN_BASIC_BLOCK.
6227 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
6230 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
6231 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
6232 (type_natural_mode): Warn ABI change when %zmm register is not
6233 available for AVX512F vector value passing.
6235 2014-02-18 Kai Tietz <ktietz@redhat.com>
6238 * config/i386/i386.c (ix86_expand_prologue): Use value in
6239 rax register as displacement when restoring %r10 or %rax.
6240 Fix wrong offset when restoring both registers.
6242 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6244 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
6245 assertion with conditional return.
6247 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6248 Uros Bizjak <ubizjak@gmail.com>
6251 * config/i386/driver-i386.c (host_detect_local_cpu): If
6252 YMM state is not saved by the OS, also clear has_f16c. Move
6253 CPUID 0x80000001 handling before YMM state saving checking.
6255 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
6257 PR rtl-optimization/58960
6258 * haifa-sched.c (alloc_global_sched_pressure_data): New,
6259 factored out from ...
6260 (sched_init): ... here.
6261 (free_global_sched_pressure_data): New, factored out from ...
6262 (sched_finish): ... here.
6263 * sched-int.h (free_global_sched_pressure_data): Declare.
6264 * sched-rgn.c (nr_regions_initial): New static global.
6265 (haifa_find_rgns): Initialize it.
6266 (schedule_region): Disable sched-pressure for the newly
6269 2014-02-17 Richard Biener <rguenther@suse.de>
6271 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
6272 release SSA defs of pattern stmts.
6274 2014-02-17 Richard Biener <rguenther@suse.de>
6276 * tree-inline.c (expand_call_inline): Release the virtual
6277 operand defined by the call we are about to inline.
6279 2014-02-17 Richard Biener <rguenther@suse.de>
6281 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
6283 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
6284 Ilya Tocar <ilya.tocar@intel.com>
6286 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
6287 arguments order in builtin.
6288 (_mm512_permutexvar_epi64): Ditto.
6289 (_mm512_mask_permutexvar_epi64): Ditto
6290 (_mm512_maskz_permutexvar_epi32): Ditto
6291 (_mm512_permutexvar_epi32): Ditto
6292 (_mm512_mask_permutexvar_epi32): Ditto
6294 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6296 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
6297 (p8_vmrgow): Likewise.
6299 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6301 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
6304 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
6307 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
6308 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
6309 into 64-bit and 32-bit moves. On 64-bit moves, add support for
6310 using direct move instructions on ISA 2.07. Also adjust
6311 instruction length for 64-bit.
6312 (mov<mode>_64bit, TFmode/TDmode): Likewise.
6313 (mov<mode>_32bit, TFmode/TDmode): Likewise.
6315 2014-02-15 Alan Modra <amodra@gmail.com>
6319 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
6320 find_replacement on parts of insn rtl that might be reloaded.
6322 2014-02-15 Richard Biener <rguenther@suse.de>
6324 PR tree-optimization/60183
6325 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
6326 (tree_ssa_phiprop): Calculate and free post-dominators.
6328 2014-02-14 Jeff Law <law@redhat.com>
6330 PR rtl-optimization/60131
6331 * ree.c (get_extended_src_reg): New function.
6332 (combine_reaching_defs): Use it rather than assuming location of REG.
6333 (find_and_remove_re): Verify first operand of extension is
6334 a REG before adding the insns to the copy list.
6336 2014-02-14 Roland McGrath <mcgrathr@google.com>
6338 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
6339 * configure: Regenerated.
6340 * config.in: Regenerated.
6341 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
6342 instead of ASM_SHORT.
6344 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
6345 Richard Earnshaw <rearnsha@arm.com>
6347 PR rtl-optimization/59535
6348 * lra-constraints.c (process_alt_operands): Encourage alternative
6349 when unassigned pseudo class is superset of the alternative class.
6350 (inherit_reload_reg): Don't inherit when optimizing for code size.
6351 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
6352 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
6353 modes not less than 4 for Thumb1.
6355 2014-02-14 Kyle McMartin <kyle@redhat.com>
6358 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
6360 2014-02-14 Richard Biener <rguenther@suse.de>
6362 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
6363 (get_frame_arg): Drop the assert with langhook types_compatible_p.
6364 Do not strip INDIRECT_REFs.
6366 2014-02-14 Richard Biener <rguenther@suse.de>
6369 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
6370 DECL_FUNCTION_SPECIFIC_TARGET.
6371 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
6372 * tree-streamer-out.c (pack_ts_target_option): Remove.
6373 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
6374 (write_ts_function_decl_tree_pointers): Do not stream
6375 DECL_FUNCTION_SPECIFIC_TARGET.
6376 * tree-streamer-in.c (unpack_ts_target_option): Remove.
6377 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
6378 (lto_input_ts_function_decl_tree_pointers): Do not stream
6379 DECL_FUNCTION_SPECIFIC_TARGET.
6381 2014-02-14 Jakub Jelinek <jakub@redhat.com>
6383 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
6384 (get_initial_def_for_induction, vectorizable_induction): Ignore
6385 debug stmts when looking for exit_phi.
6386 (vectorizable_live_operation): Fix up condition.
6388 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6390 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
6391 nreverse() because it changes the content of original tree list.
6393 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6395 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
6396 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
6398 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6400 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
6401 GNU coding standards.
6403 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6406 * dwarf2out.c (gen_subprogram_die): Don't call
6407 add_calling_convention_attribute if subr_die is old_die.
6409 2014-02-13 Sharad Singhai <singhai@google.com>
6411 * doc/optinfo.texi: Fix order of nodes.
6413 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
6415 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
6416 operands[2], not operands[3].
6418 2014-02-13 Richard Biener <rguenther@suse.de>
6421 * doc/install.texi (ISL): Update recommended version to 0.12.2,
6422 mention the possibility of an in-tree build.
6423 (CLooG): Update recommended version to 0.18.1, mention the
6424 possibility of an in-tree build and clarify that the ISL
6425 bundled with CLooG does not work.
6427 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6430 * expr.c (compress_float_constant): If x is a hard register,
6431 extend into a pseudo and then move to x.
6433 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
6435 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
6436 caused by bad second argument to warning_at() with -mhotpatch and
6437 nested functions (e.g. with gfortran).
6439 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
6441 * opts.c (option_name): Remove "enabled by default" rider.
6443 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
6445 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
6447 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
6448 Uros Bizjak <ubizjak@gmail.com>
6451 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
6452 * configure: Regenerated.
6454 2014-02-12 Richard Biener <rguenther@suse.de>
6456 * vec.c (vec_prefix::calculate_allocation): Move as
6457 inline variant to vec.h.
6458 (vec_prefix::calculate_allocation_1): New out-of-line version.
6459 * vec.h (vec_prefix::calculate_allocation_1): Declare.
6460 (vec_prefix::m_has_auto_buf): Rename to ...
6461 (vec_prefix::m_using_auto_storage): ... this.
6462 (vec_prefix::calculate_allocation): Inline the easy cases
6463 and dispatch to calculate_allocation_1 which doesn't need the
6465 (va_heap::reserve): Use gcc_checking_assert.
6466 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
6467 m_using_auto_storage.
6468 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
6470 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
6471 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
6472 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
6474 2014-02-12 Richard Biener <rguenther@suse.de>
6476 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
6477 when we found a dependence.
6479 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
6481 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
6483 (maybe_fold_stmt): ... into this new function.
6484 * omp-low.c (lower_omp): Update comment.
6486 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
6489 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
6492 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
6494 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
6495 identifiers in comments.
6496 (cortexa53_extra_costs): Likewise.
6497 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
6498 (cortexa7_extra_costs): Likewise.
6499 (cortexa12_extra_costs): Likewise.
6500 (cortexa15_extra_costs): Likewise.
6501 (v7m_extra_costs): Likewise.
6503 2014-02-12 Richard Biener <rguenther@suse.de>
6506 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
6507 of posix_memalign being successful.
6508 (lower_stmt): Restrict lowering of posix_memalign to when
6509 -ftree-bit-ccp is enabled.
6511 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6513 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
6515 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
6517 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
6519 PR rtl-optimization/60116
6520 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
6521 other_insn once the combination has been validated.
6523 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
6526 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
6528 * ipa-devirt.c: Include demangle.h
6529 (odr_violation_reported): New static variable.
6530 (add_type_duplicate): Update odr_violations.
6531 (maybe_record_node): Add completep parameter; update it.
6532 (record_target_from_binfo): Add COMPLETEP parameter;
6533 update it as needed.
6534 (possible_polymorphic_call_targets_1): Likewise.
6535 (struct polymorphic_call_target_d): Add nonconstruction_targets;
6536 rename FINAL to COMPLETE.
6537 (record_targets_from_bases): Sanity check we found the binfo;
6538 fix COMPLETEP updating.
6539 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
6540 parameter, fix computing of COMPLETEP.
6541 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
6542 at LTO time do demangling.
6543 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
6544 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
6546 (gimple_get_virt_method_for_binfo): Likewise.
6547 * gimple-fold.h (gimple_get_virt_method_for_binfo,
6548 gimple_get_virt_method_for_vtable): Update prototypes.
6550 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
6553 * genautomata.c (add_presence_absence): Fix typo with
6554 {final_}presence_list.
6556 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
6559 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
6560 for VSX/Altivec vectors that land in GPR registers.
6562 2014-02-11 Richard Henderson <rth@redhat.com>
6563 Jakub Jelinek <jakub@redhat.com>
6566 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
6567 around drhs if type conversion to lacc->type is not useless.
6569 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6571 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
6573 (cortex-a57.cortex-a53): Likewise.
6574 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
6576 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6578 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
6581 2014-02-11 Renlin Li <Renlin.Li@arm.com>
6583 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
6584 add_options_for_arm_vfp3.
6586 2014-02-11 Jeff Law <law@redhat.com>
6589 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
6590 object with an undesirable mode.
6592 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6595 * config/i386/sol2-9.h: New file.
6596 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
6597 *-*-solaris2.9*): Use it.
6599 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6601 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
6602 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
6604 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6606 * config/microblaze/microblaze.c: Extend mcpu version format
6608 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
6610 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
6612 2014-02-10 Richard Henderson <rth@redhat.com>
6615 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
6616 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
6617 ms-abi vs -mno-accumulate-outgoing-args.
6618 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
6619 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
6622 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
6625 * cfgexpand.c (expand_asm_operands): Attach source location to
6626 ASM_INPUT rtx objects.
6627 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
6629 2014-02-10 Nick Clifton <nickc@redhat.com>
6631 * config/mn10300/mn10300.c (popcount): New function.
6632 (mn10300_expand_prologue): Include saved registers in stack usage
6635 2014-02-10 Jeff Law <law@redhat.com>
6638 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
6639 when changing the SET_DEST of a prior insn to avoid an input reload.
6641 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6643 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
6644 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
6645 -mcall-openbsd, or -mcall-linux.
6646 (CC1_ENDIAN_BIG_SPEC): Remove.
6647 (CC1_ENDIAN_LITTLE_SPEC): Remove.
6648 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
6649 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
6650 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
6651 and %cc1_endian_default.
6652 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
6654 2014-02-10 Richard Biener <rguenther@suse.de>
6656 PR tree-optimization/60115
6657 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
6658 MEM_REF handling. Properly verify that the accesses are not
6659 out of the objects bound.
6661 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6663 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
6666 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
6668 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
6669 proper constants and fix formatting.
6670 (possible_polymorphic_call_targets): Fix formatting.
6672 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
6673 Ilya Tocar <ilya.tocar@intel.com>
6675 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
6676 (_mm512_loadu_epi32): Renamed into...
6677 (_mm512_loadu_si512): This.
6678 (_mm512_storeu_epi32): Renamed into...
6679 (_mm512_storeu_si512): This.
6680 (_mm512_maskz_ceil_ps): Removed.
6681 (_mm512_maskz_ceil_pd): Ditto.
6682 (_mm512_maskz_floor_ps): Ditto.
6683 (_mm512_maskz_floor_pd): Ditto.
6684 (_mm512_floor_round_ps): Ditto.
6685 (_mm512_floor_round_pd): Ditto.
6686 (_mm512_ceil_round_ps): Ditto.
6687 (_mm512_ceil_round_pd): Ditto.
6688 (_mm512_mask_floor_round_ps): Ditto.
6689 (_mm512_mask_floor_round_pd): Ditto.
6690 (_mm512_mask_ceil_round_ps): Ditto.
6691 (_mm512_mask_ceil_round_pd): Ditto.
6692 (_mm512_maskz_floor_round_ps): Ditto.
6693 (_mm512_maskz_floor_round_pd): Ditto.
6694 (_mm512_maskz_ceil_round_ps): Ditto.
6695 (_mm512_maskz_ceil_round_pd): Ditto.
6696 (_mm512_expand_pd): Ditto.
6697 (_mm512_expand_ps): Ditto.
6698 * config/i386/i386.c (ix86_builtins): Remove
6699 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
6700 (bdesc_args): Ditto.
6701 * config/i386/predicates.md (const1256_operand): New.
6702 (const_1_to_2_operand): Ditto.
6703 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6704 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6705 (*avx512pf_gatherpf<mode>sf): Ditto.
6706 (avx512pf_gatherpf<mode>df): Ditto.
6707 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6708 (*avx512pf_gatherpf<mode>df): Ditto.
6709 (avx512pf_scatterpf<mode>sf): Ditto.
6710 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6711 (*avx512pf_scatterpf<mode>sf): Ditto.
6712 (avx512pf_scatterpf<mode>df): Ditto.
6713 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6714 (*avx512pf_scatterpf<mode>df): Ditto.
6715 (avx512f_expand<mode>): Removed.
6716 (<shift_insn><mode>3<mask_name>): Change predicate type.
6718 2014-02-08 Jakub Jelinek <jakub@redhat.com>
6720 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
6721 not at the end of datarefs vector use ordered_remove to avoid
6722 reordering datarefs vector.
6725 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
6726 mark local addressable non-static vars as GOVD_PRIVATE
6727 instead of GOVD_LOCAL.
6728 * omp-low.c (lower_omp_for): Move gimple_bind_vars
6729 and BLOCK_VARS of gimple_bind_block to new_stmt rather
6733 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
6734 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
6735 assume_aligned or alloc_align attributes.
6736 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
6737 arguments. Handle also assume_aligned and alloc_align attributes.
6738 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
6739 calls to functions with assume_aligned or alloc_align attributes.
6740 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
6742 2014-02-08 Terry Guo <terry.guo@arm.com>
6744 * doc/invoke.texi: Document ARM -march=armv7e-m.
6746 2014-02-08 Jakub Jelinek <jakub@redhat.com>
6748 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
6749 flag on __cilkrts_rethrow builtin.
6752 * ipa-cp.c (determine_versionability): Fail at -O0
6753 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
6754 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
6757 2014-02-04 Jakub Jelinek <jakub@redhat.com>
6760 * tree-inline.c (copy_forbidden): Fail for
6761 __attribute__((optimize (0))) functions.
6763 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
6765 * varpool.c: Include pointer-set.h.
6766 (varpool_remove_unreferenced_decls): Variables in other partitions
6767 will not be output; be however careful to not lose information
6770 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
6772 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
6773 lookup in the vtable constructor.
6775 2014-02-07 Jeff Law <law@redhat.com>
6778 * config/m68k/m68k.md (ashldi_extsi): Turn into a
6779 define_insn_and_split.
6781 * ipa-inline.c (inline_small_functions): Fix typos.
6783 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6785 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
6786 (s390_can_use_return_insn): Declare.
6787 * config/s390/s390.h (EPILOGUE_USES): Define.
6788 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
6790 (s390_chunkify_start): Handle return JUMP_LABELs.
6791 (s390_early_mach): Emit a main_pool instruction on the entry edge.
6792 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
6793 (s390_can_use_return_insn): New functions.
6794 (s390_fix_long_loop_prediction): Handle conditional returns.
6795 (TARGET_SET_UP_BY_PROLOGUE): Define.
6796 * config/s390/s390.md (ANY_RETURN): New code iterator.
6797 (*creturn, *csimple_return, return, simple_return): New patterns.
6799 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6801 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
6802 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
6803 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
6804 REG_CFA_RESTORE list when deciding not to restore a register.
6806 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6808 * config/s390/s390.c: Include tree-pass.h and context.h.
6809 (s390_early_mach): New function, split out from...
6810 (s390_emit_prologue): ...here.
6811 (pass_data_s390_early_mach): New pass structure.
6812 (pass_s390_early_mach): New class.
6813 (s390_option_override): Create and register early_mach pass.
6814 Move to end of file.
6816 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6818 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
6819 to match for the exit block.
6821 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6823 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
6824 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
6825 Reject misaligned operands.
6827 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6829 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
6831 2014-02-07 Richard Biener <rguenther@suse.de>
6834 * gimple-low.c (lower_builtin_posix_memalign): New function.
6835 (lower_stmt): Call it to lower posix_memalign in a way
6836 to make alignment info accessible.
6838 2014-02-07 Jakub Jelinek <jakub@redhat.com>
6841 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
6842 __builtin_setjmp_receiver.
6844 2014-02-07 Richard Biener <rguenther@suse.de>
6847 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
6848 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
6849 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
6850 Handle BUILT_IN_POSIX_MEMALIGN.
6851 (find_func_clobbers): Likewise.
6852 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
6853 (call_may_clobber_ref_p_1): Likewise.
6855 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6858 * ipa-devirt.c (record_target_from_binfo): Remove overactive
6861 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6864 * lto-cgraph.c (lto_output_node): Use
6865 symtab_get_symbol_partitioning_class.
6866 (lto_output_varpool_node): likewise.
6867 (symtab_get_symbol_partitioning_class): Move here from
6869 * cgraph.h (symbol_partitioning_class): Likewise.
6870 (symtab_get_symbol_partitioning_class): Declare.
6872 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6874 * ggc.h (ggc_internal_cleared_alloc): New macro.
6875 * vec.h (vec_safe_copy): Handle memory stats.
6876 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
6877 * target-globals.c (save_target_globals): Likewise.
6879 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6882 * expr.c (emit_move_resolve_push): Export; be bit more selective
6883 on when to clear alias set.
6884 * expr.h (emit_move_resolve_push): Declare.
6885 * function.h (struct function): Add tail_call_marked.
6886 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
6887 * config/i386/i386-protos.h (ix86_expand_push): Remove.
6888 * config/i386/i386.md (TImode move expander): De not call
6890 (FP push expanders): Preserve memory attributes.
6891 * config/i386/sse.md (push<mode>1): Remove.
6892 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
6893 (ix86_expand_push): Remove.
6894 * config/i386/mmx.md (push<mode>1): Remove.
6896 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6898 PR rtl-optimization/60030
6899 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
6900 lopart with paradoxical subreg before shifting it up by hprec.
6902 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6904 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
6905 Remove extra newline at end of file.
6906 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
6907 (arm_issue_rate): Handle cortexa57.
6908 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
6909 (cortex-a57.cortex-a53): Likewise.
6911 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6914 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
6915 don't record in REG_FRAME_RELATED_EXPR registers not set in that
6917 (arm_expand_prologue): Adjust all callers.
6918 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
6919 info, registers also at the lowest numbered registers side. Use
6920 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
6924 * var-tracking.c (adjust_mems): Before adding a SET to
6925 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
6927 2014-02-06 Alan Modra <amodra@gmail.com>
6930 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
6931 change SDmode to DDmode when lra_in_progress.
6933 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6936 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
6937 free_data_ref on the dr first, and before goto again also set dr
6938 to the next dr. For simd_lane_access, free old datarefs[i] before
6939 overwriting it. For get_vectype_for_scalar_type failure, don't
6940 free_data_ref if simd_lane_access.
6942 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
6945 * tree.h (opts_for_fn): New inline function.
6946 (opt_for_fn): Define.
6947 * config/i386/i386.c (ix86_function_regparm): Use
6948 opt_for_fn (decl, optimize) instead of optimize.
6950 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
6952 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
6953 for SYMBOL_REF in large memory model.
6955 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6957 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
6959 (cortex-a57): Likewise.
6960 (cortex-a57.cortex-a53): Likewise.
6962 2014-02-06 Yury Gribov <y.gribov@samsung.com>
6963 Kugan Vivekanandarajah <kuganv@linaro.org>
6965 * config/arm/arm.c (arm_vector_alignment_reachable): Check
6967 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
6969 2014-02-06 Richard Biener <rguenther@suse.de>
6971 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
6972 set_loop_copy and initialize_original_copy_tables.
6974 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
6976 * config/aarch64/aarch64-simd.md
6977 (aarch64_ashr_simddi): Change QI to SI.
6979 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
6980 Jakub Jelinek <jakub@redhat.com>
6983 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
6986 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6988 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
6989 CODE_FOR_altivec_vpku[hw]um to
6990 CODE_FOR_altivec_vpku[hw]um_direct.
6991 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
6992 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
6993 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
6994 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
6996 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6998 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
6999 generation for -maltivec=be.
7000 (altivec_vsumsws): Simplify redundant test.
7002 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7004 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
7005 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
7006 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
7007 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
7008 gen_altivec_vpkuwum.
7009 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
7011 (altivec_vpks<VI_char>ss): Likewise.
7012 (altivec_vpks<VI_char>us): Likewise.
7013 (altivec_vpku<VI_char>us): Likewise.
7014 (altivec_vpku<VI_char>um): Likewise.
7015 (altivec_vpku<VI_char>um_direct): New (copy of
7016 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
7018 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
7019 target is little endian and -maltivec=be is not specified.
7020 (*altivec_vupkhs<VU_char>_direct): New (copy of
7021 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
7022 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
7023 target is little endian and -maltivec=be is not specified.
7024 (*altivec_vupkls<VU_char>_direct): New (copy of
7025 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
7026 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
7027 little endian and -maltivec=be is not specified.
7028 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
7029 little endian and -maltivec=be is not specified.
7031 2014-02-05 Richard Henderson <rth@redhat.com>
7034 * combine-stack-adj.c: Revert r206943.
7035 * sched-int.h (struct deps_desc): Add last_args_size.
7036 * sched-deps.c (init_deps): Initialize it.
7037 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7038 contain REG_ARGS_SIZE notes.
7040 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7042 * lto-cgraph.c (asm_nodes_output): Make global.
7043 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7044 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7045 (driver_handle_option): Handle OPT_fwpa.
7047 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7050 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7051 a comment typo and formatting issue. If odr_hash hasn't been
7052 created, return vNULL and set *completep to false.
7055 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7056 bb with no successors.
7058 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7061 * doc/invoke.texi (-march): Clarify documentation for ARM.
7065 2014-02-05 Richard Biener <rguenther@suse.de>
7067 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
7068 when not vectorizing because of too many alias checks.
7069 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
7070 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
7072 2014-02-05 Nick Clifton <nickc@redhat.com>
7074 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
7075 accept extended registers in any mode when compiling for the MN10300.
7077 2014-02-05 Yury Gribov <y.gribov@samsung.com>
7079 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
7080 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
7081 sanitization attributes.
7082 (can_inline_edge_p): Likewise.
7083 (sanitize_attrs_match_for_inline_p): New function.
7085 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7087 * ipa-prop.c (detect_type_change): Shor circuit testing of
7088 type changes on THIS pointer.
7090 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
7093 * config/pa/pa.c (legitimize_tls_address): Return original address
7094 if not passed a SYMBOL_REF rtx.
7095 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
7097 (pa_emit_move_sequence): Simplify TLS source operands.
7098 (pa_legitimate_constant_p): Reject all TLS constants.
7099 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
7100 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
7102 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7104 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
7105 groups when we know they are controlled by LTO.
7106 * varasm.c (default_binds_local_p_1): If object is in other partition,
7107 it will be resolved locally.
7109 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7111 * config/host-linux.c (linux_gt_pch_use_address): Don't
7112 use SSIZE_MAX because it is not always defined.
7114 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
7117 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
7118 threshold for pseudo splitting.
7119 (update_ebb_live_info): Process call argument hard registers and
7120 hard registers from insn definition too.
7121 (max_small_class_regs_num): New constant.
7122 (inherit_in_ebb): Update live hard regs through EBBs. Update
7123 reloads_num only for small register classes. Don't split for
7126 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
7129 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
7132 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7134 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
7137 2014-02-04 Marek Polacek <polacek@redhat.com>
7139 * gdbinit.in (pel): Define.
7141 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7143 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
7146 2014-02-04 Richard Biener <rguenther@suse.de>
7149 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
7150 in function context local.
7151 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
7152 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
7153 similar to LTO_imported_decl_ref.
7155 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7157 PR tree-optimization/60002
7158 * cgraphclones.c (build_function_decl_skip_args): Clear
7161 PR tree-optimization/60023
7162 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
7163 false to gsi_replace.
7164 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
7165 has been in some EH region and vec_stmt could throw, add
7166 vec_stmt into the same EH region.
7167 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
7168 has no lhs, ignore it.
7169 * internal-fn.c (expand_MASK_LOAD): Likewise.
7172 * tree-inline.c (copy_forbidden): Fail for
7173 __attribute__((optimize (0))) functions.
7176 * omp-low.c (simd_clone_struct_copy): If from->inbranch
7177 is set, copy one less argument.
7178 (expand_simd_clones): Don't subtract clone_info->inbranch
7179 from simd_clone_struct_alloc argument.
7181 PR rtl-optimization/57915
7182 * recog.c (simplify_while_replacing): If all unary/binary/relational
7183 operation arguments are constant, attempt to simplify those.
7186 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
7187 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
7189 2014-02-04 Richard Biener <rguenther@suse.de>
7191 PR tree-optimization/60012
7192 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
7193 TBAA disambiguation to all DDRs.
7195 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7198 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
7199 (LINK_SPEC): Use it for -shared, -shared-libgcc.
7201 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7204 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
7206 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7208 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
7209 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
7211 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7214 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
7215 to figure out targets of polymorphic calls with known decl.
7216 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7217 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
7218 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
7219 (get_polymorphic_call_info): ... here.
7220 (get_polymorphic_call_info_from_invariant): New function.
7222 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7224 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
7225 lookup via vtable pointer; check for type consistency
7226 and turn inconsitent facts into UNREACHABLE.
7227 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7228 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
7229 type inconsistent querries; return UNREACHABLE instead.
7231 2014-02-03 Richard Henderson <rth@twiddle.net>
7234 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
7235 already processed this node.
7236 (normalize_one_pred_1): Pass along mark_set.
7237 (normalize_one_pred): Create and destroy a pointer_set_t.
7238 (normalize_one_pred_chain): Likewise.
7240 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
7242 PR gcov-profile/58602
7243 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
7245 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7248 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
7249 -fno-devirtualize; try to devirtualize by the knowledge of
7250 virtual table pointer given by aggregate propagation.
7251 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7252 (ipa_print_node_jump_functions): Dump also offset that
7253 is relevant for polymorphic calls.
7254 (determine_known_aggregate_parts): Add arg_type parameter; use it
7255 instead of determining the type from pointer type.
7256 (ipa_compute_jump_functions_for_edge): Update call of
7257 determine_known_aggregate_parts.
7258 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
7259 (gimple_get_virt_method_for_binfo): ... here; simplify using
7260 vtable_pointer_value_to_vtable.
7261 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
7262 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
7263 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
7264 (vtable_pointer_value_to_vtable): Break out from ...; handle also
7266 (vtable_pointer_value_to_binfo): ... here.
7267 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
7269 2014-02-03 Teresa Johnson <tejohnson@google.com>
7271 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
7272 redef of outer loop index variable.
7274 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
7278 * doc/extend.texi (Function Attributes): Typo.
7280 2014-02-03 Cong Hou <congh@google.com>
7282 PR tree-optimization/60000
7283 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
7284 if the vectorized statement is a store. A store statement can only
7285 appear at the end of pattern statements.
7287 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7289 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
7290 (ix86_option_override_internal): Default long double to 64-bit for
7291 32-bit Bionic and to 128-bit for 64-bit Bionic.
7293 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
7294 TARGET_LONG_DOUBLE_128 is true.
7295 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
7297 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
7298 (mlong-double-64): Negate -mlong-double-128.
7299 (mlong-double-128): New option.
7301 * config/i386/i386-c.c (ix86_target_macros): Define
7302 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
7304 * doc/invoke.texi: Document -mlong-double-128.
7306 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7308 PR rtl-optimization/60024
7309 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
7311 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
7313 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
7315 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7317 PR rtl-optimization/57662
7318 * sel-sched.c (code_motion_path_driver): Do not mark already not
7319 existing blocks in the visiting bitmap.
7321 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7323 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
7324 on the insn being emitted.
7326 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
7327 Will Deacon <will.deacon@arm.com>
7329 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
7331 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7333 * config/arm/arm-tables.opt: Regenerate.
7335 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7337 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
7338 for vector types other than V16QImode.
7339 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
7340 define_expand, and call altivec_expand_vec_perm_le when producing
7341 code with little endian element order.
7342 (*altivec_vperm_<mode>_internal): New insn having previous
7343 behavior of altivec_vperm_<mode>.
7344 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
7345 altivec_expand_vec_perm_le when producing code with little endian
7347 (*altivec_vperm_<mode>_uns_internal): New insn having previous
7348 behavior of altivec_vperm_<mode>_uns.
7350 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7352 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
7353 (altivec_vsumsws): Add handling for -maltivec=be with a little
7355 (altivec_vsumsws_direct): New.
7356 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
7357 gen_altivec_vsumsws.
7359 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7361 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
7362 vtable_pointer_value_to_binfo): New functions.
7363 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
7364 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
7366 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
7368 * config/nios2/nios2.md (load_got_register): Initialize GOT
7369 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
7370 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
7372 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7374 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
7375 preserverd by passthrough, do not propagate the type.
7377 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7379 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
7380 (mips_atomic_assign_expand_fenv): New function.
7381 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
7383 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7385 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
7386 (__builtin_mips_set_fcsr): Likewise.
7387 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
7388 MIPS_USI_FTYPE_VOID.
7389 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
7390 (mips16_expand_set_fcsr): Likewise.
7391 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
7392 (mips16_set_fcsr_stub): Likewise.
7393 (mips16_get_fcsr_one_only_stub): New class.
7394 (mips16_set_fcsr_one_only_stub): Likewise.
7395 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
7396 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
7397 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
7398 (hard_float): New availability predicate.
7399 (mips_builtins): Add get_fcsr and set_fcsr.
7400 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
7401 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
7402 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
7403 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
7404 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
7407 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7409 * config/mips/mips.c (mips_one_only_stub): New class.
7410 (mips_need_mips16_rdhwr_p): Replace with...
7411 (mips16_rdhwr_stub): ...this new variable.
7412 (mips16_stub_call_address): New function.
7413 (mips16_rdhwr_one_only_stub): New class.
7414 (mips_expand_thread_pointer): Use mips16_stub_call_address.
7415 (mips_output_mips16_rdhwr): Delete.
7416 (mips_finish_stub): New function.
7417 (mips_code_end): Use it to handle rdhwr stubs.
7419 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
7422 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
7423 when calculating size of integer atomic types.
7425 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
7427 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
7429 2014-02-01 Jakub Jelinek <jakub@redhat.com>
7431 PR tree-optimization/60003
7432 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
7433 * profile.c (branch_prob): Use gimple_call_builtin_p
7434 to check for BUILT_IN_SETJMP_RECEIVER.
7435 * tree-inline.c (copy_bb): Call notice_special_calls.
7437 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
7440 * lra-constraints.c (process_alt_operands): Update reload_sum only
7443 2014-01-31 Richard Henderson <rth@redhat.com>
7446 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
7447 until after else_eh is processed.
7449 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7451 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
7452 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
7453 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
7454 in smmintrin.h, remove them.
7455 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
7456 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
7457 * config/i386/i386.md (ROUND_SAE): Fix value.
7458 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
7459 (const48_operand): New.
7460 * config/i386/subst.md (round), (round_expand): Use
7461 const_4_or_8_to_11_operand.
7462 (round_saeonly), (round_saeonly_expand): Use const48_operand.
7464 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7466 * config/i386/constraints.md (Yk): Swap meaning with k.
7467 * config/i386/i386.md (movhi_internal): Change Yk to k.
7468 (movqi_internal): Ditto.
7469 (*k<logic><mode>): Ditto.
7472 (kandn<mode>): Ditto.
7473 (*<code>hi_1): Ditto.
7474 (*<code>qi_1): Ditto.
7475 (kxnor<mode>): Ditto.
7476 (kortestzhi): Ditto.
7477 (kortestchi): Ditto.
7479 (*one_cmplhi2_1): Ditto.
7480 (*one_cmplqi2_1): Ditto.
7481 * config/i386/sse.md (): Change k to Yk.
7482 (avx512f_load<mode>_mask): Ditto.
7483 (avx512f_blendm<mode>): Ditto.
7484 (avx512f_store<mode>_mask): Ditto.
7485 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
7486 (avx512f_storedqu<mode>_mask): Ditto.
7487 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
7489 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
7490 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
7491 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
7492 (avx512f_maskcmp<mode>3): Ditto.
7493 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
7494 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
7495 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
7496 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
7497 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
7498 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
7499 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
7500 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
7501 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
7502 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
7503 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
7504 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
7505 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
7506 (vec_extract_lo_<mode>_maskm): Ditto.
7507 (vec_extract_hi_<mode>_maskm): Ditto.
7508 (avx512f_vternlog<mode>_mask): Ditto.
7509 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
7510 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
7511 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
7512 (avx512f_<code>v8div16qi2_mask): Ditto.
7513 (avx512f_<code>v8div16qi2_mask_store): Ditto.
7514 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
7515 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
7516 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
7517 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
7518 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7519 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7520 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7521 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7522 (avx512cd_maskb_vec_dupv8di): Ditto.
7523 (avx512cd_maskw_vec_dupv16si): Ditto.
7524 (avx512f_vpermi2var<mode>3_maskz): Ditto.
7525 (avx512f_vpermi2var<mode>3_mask): Ditto.
7526 (avx512f_vpermi2var<mode>3_mask): Ditto.
7527 (avx512f_vpermt2var<mode>3_maskz): Ditto.
7528 (*avx512f_gathersi<mode>): Ditto.
7529 (*avx512f_gathersi<mode>_2): Ditto.
7530 (*avx512f_gatherdi<mode>): Ditto.
7531 (*avx512f_gatherdi<mode>_2): Ditto.
7532 (*avx512f_scattersi<mode>): Ditto.
7533 (*avx512f_scatterdi<mode>): Ditto.
7534 (avx512f_compress<mode>_mask): Ditto.
7535 (avx512f_compressstore<mode>_mask): Ditto.
7536 (avx512f_expand<mode>_mask): Ditto.
7537 * config/i386/subst.md (mask): Change k to Yk.
7538 (mask_scalar_merge): Ditto.
7541 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
7543 * doc/extend.texi (Vector Extensions): Document ?: in C++.
7545 2014-01-31 Richard Biener <rguenther@suse.de>
7548 * builtins.c (fold_builtin_memory_op): Make sure to not
7549 use a floating-point mode or a boolean or enumeral type for
7552 2014-01-30 DJ Delorie <dj@redhat.com>
7554 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
7555 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
7556 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
7557 whenever main() has an epilogue.
7559 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7561 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
7562 unused variable "field".
7563 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
7564 (vsx_mergeh_<mode>): Likewise.
7565 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
7566 (altivec_vmrghh): Likewise.
7567 (altivec_vmrghw): Likewise.
7568 (altivec_vmrglb): Likewise.
7569 (altivec_vmrglh): Likewise.
7570 (altivec_vmrglw): Likewise.
7571 (altivec_vspltb): Add missing uses.
7572 (altivec_vsplth): Likewise.
7573 (altivec_vspltw): Likewise.
7574 (altivec_vspltsf): Likewise.
7576 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7579 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
7580 frame related instructions.
7582 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
7584 PR rtl-optimization/59959
7585 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
7586 any reload of register whose subreg is invalid.
7588 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7590 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
7591 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
7592 Add missing return type - void.
7594 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7596 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
7597 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
7598 remove element index adjustment for endian (now handled in vsx.md
7600 (altivec_expand_vec_perm_const): Use
7601 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
7602 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
7603 (vsx_xxspltw_<mode>): Adjust element index for little endian.
7604 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
7605 define_expand and a new define_insn *altivec_vspltb_internal;
7606 adjust for -maltivec=be on a little endian target.
7607 (altivec_vspltb_direct): New.
7608 (altivec_vsplth): Divide into a define_expand and a new
7609 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
7610 little endian target.
7611 (altivec_vsplth_direct): New.
7612 (altivec_vspltw): Divide into a define_expand and a new
7613 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
7614 little endian target.
7615 (altivec_vspltw_direct): New.
7616 (altivec_vspltsf): Divide into a define_expand and a new
7617 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
7618 a little endian target.
7620 2014-01-30 Richard Biener <rguenther@suse.de>
7622 PR tree-optimization/59993
7623 * tree-ssa-forwprop.c (associate_pointerplus): Check we
7624 can propagate form the earlier stmt and avoid the transform
7625 when the intermediate result is needed.
7627 2014-01-30 Alangi Derick <alangiderick@gmail.com>
7629 * README.Portability: Fix typo.
7631 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
7633 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
7634 comparison_operator with ordered_comparison_operator.
7636 2014-01-30 Nick Clifton <nickc@redhat.com>
7638 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
7639 Rename to mn10300_store_multiple_regs.
7640 * config/mn10300/mn10300.c: Likewise.
7641 * config/mn10300/mn10300.md (store_movm): Fix typo: call
7642 store_multiple_regs.
7643 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
7644 Call mn10300_store_multiple_regs.
7646 2014-01-30 Nick Clifton <nickc@redhat.com>
7647 DJ Delorie <dj@redhat.com>
7649 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
7650 %fp 2 to keep registers after it properly word-aligned.
7651 (rl78_alloc_physical_registers_umul): Handle the case where both
7652 input operands are the same.
7654 2014-01-30 Richard Biener <rguenther@suse.de>
7656 PR tree-optimization/59903
7657 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
7660 2014-01-30 Jason Merrill <jason@redhat.com>
7663 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
7666 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
7668 2014-01-30 Richard Biener <rguenther@suse.de>
7670 PR tree-optimization/59951
7671 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
7673 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
7676 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
7677 SFmode to DFmode case.
7679 2014-01-29 DJ Delorie <dj@redhat.com>
7681 * config/msp430/msp430.opt (-minrt): New.
7682 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
7684 (ENDFILE_SPEC): Likewise.
7686 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
7688 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
7689 (estimate_function_body_sizes): Use it.
7691 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
7694 * dwarf2out.c (is_cxx_auto): New.
7695 (is_base_type): Use it.
7696 (gen_type_die_with_usage): Likewise.
7698 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7700 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
7701 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
7702 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
7703 -maltivec=be with LE targets.
7704 (vsx_mergeh_<mode>): Likewise.
7705 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
7706 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
7707 (altivec_vmrghb): Replace with define_expand and new
7708 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
7709 (altivec_vmrghb_direct): New define_insn.
7710 (altivec_vmrghh): Replace with define_expand and new
7711 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
7712 (altivec_vmrghh_direct): New define_insn.
7713 (altivec_vmrghw): Replace with define_expand and new
7714 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
7715 (altivec_vmrghw_direct): New define_insn.
7716 (*altivec_vmrghsf): Adjust for endianness.
7717 (altivec_vmrglb): Replace with define_expand and new
7718 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
7719 (altivec_vmrglb_direct): New define_insn.
7720 (altivec_vmrglh): Replace with define_expand and new
7721 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
7722 (altivec_vmrglh_direct): New define_insn.
7723 (altivec_vmrglw): Replace with define_expand and new
7724 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
7725 (altivec_vmrglw_direct): New define_insn.
7726 (*altivec_vmrglsf): Adjust for endianness.
7727 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
7728 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
7729 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
7730 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
7731 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
7732 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
7733 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
7734 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
7736 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
7738 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
7739 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
7742 2014-01-29 Richard Biener <rguenther@suse.de>
7744 PR tree-optimization/58742
7745 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
7746 associate_pointerplus_align.
7747 (associate_pointerplus_diff): New function.
7748 (associate_pointerplus): Likewise. Call associate_pointerplus_align
7749 and associate_pointerplus_diff.
7751 2014-01-29 Richard Biener <rguenther@suse.de>
7753 * lto-streamer.h (LTO_major_version): Bump to 3.
7754 (LTO_minor_version): Reset to 0.
7756 2014-01-29 Renlin Li <Renlin.Li@arm.com>
7758 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
7759 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
7760 (arm_file_start): Generate correct asm header for armv7ve.
7761 * config/arm/bpabi.h: Add multilib support for armv7ve.
7762 * config/arm/driver-arm.c: Change the architectures of cortex-a7
7763 and cortex-a15 to armv7ve.
7764 * config/arm/t-aprofile: Add multilib support for armv7ve.
7765 * doc/invoke.texi: Document -march=armv7ve.
7767 2014-01-29 Richard Biener <rguenther@suse.de>
7769 PR tree-optimization/58742
7770 * tree-ssa-forwprop.c (associate_plusminus): Return true
7771 if we changed sth, defer EH cleanup to ...
7772 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
7773 (simplify_mult): New function.
7775 2014-01-29 Jakub Jelinek <jakub@redhat.com>
7778 PR tree-optimization/59920
7779 * tree.c (build_common_builtin_nodes): Remove
7780 __builtin_setjmp_dispatcher initialization.
7781 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
7782 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
7783 instead of gsi_after_labels + manually skipping debug stmts.
7784 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
7785 ignore bbs with IFN_ABNORMAL_DISPATCHER.
7786 * tree-inline.c (copy_edges_for_bb): Remove
7787 can_make_abnormal_goto argument, instead add abnormal_goto_dest
7788 argument. Ignore computed_goto_p stmts. Don't call
7789 make_abnormal_goto_edges. If a call might need abnormal edges
7790 for non-local gotos, see if it already has an edge to
7791 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
7792 with true argument, don't do anything then, otherwise add
7793 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
7794 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
7796 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
7797 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
7798 (lower_stmt): Don't set data->calls_builtin_setjmp.
7799 (lower_builtin_setjmp): Adjust comment.
7800 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
7801 * tree-cfg.c (found_computed_goto): Remove.
7802 (factor_computed_gotos): Remove.
7803 (make_goto_expr_edges): Return bool, true for computed gotos.
7804 Don't call make_abnormal_goto_edges.
7805 (build_gimple_cfg): Don't set found_computed_goto, don't call
7806 factor_computed_gotos.
7807 (computed_goto_p): No longer static.
7808 (make_blocks): Don't set found_computed_goto.
7809 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
7810 (make_edges): If make_goto_expr_edges returns true, push bb
7811 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
7812 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
7813 vector. Record mapping between bbs and OpenMP regions if there
7814 are any, adjust make_gimple_omp_edges caller. Call
7815 handle_abnormal_edges.
7816 (make_abnormal_goto_edges): Remove.
7817 * tree-cfg.h (make_abnormal_goto_edges): Remove.
7818 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
7819 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
7820 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
7821 * internal-fn.def (ABNORMAL_DISPATCHER): New.
7822 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
7823 filling *region also set *region_idx to (*region)->entry->index.
7826 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
7827 For REGs set ORIGINAL_REGNO.
7829 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
7831 * doc/md.texi: Mention that a target shouldn't implement
7832 vec_widen_(s|u)mul_even/odd pair if it is less efficient
7835 2014-01-29 Jakub Jelinek <jakub@redhat.com>
7837 PR tree-optimization/59594
7838 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
7839 a copy of the datarefs vector rather than the vector itself.
7841 2014-01-28 Jason Merrill <jason@redhat.com>
7844 * dwarf2out.c (auto_die): New static.
7845 (gen_type_die_with_usage): Handle C++1y 'auto'.
7846 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
7849 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
7852 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
7853 (SPEC_X32): Likewise.
7854 (SPEC_64): Likewise.
7855 * config/i386/i386.c (ix86_option_override_internal): Turn off
7856 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
7858 (x86_file_start): Output .code16gcc for TARGET_16BIT.
7859 * config/i386/i386.h (TARGET_16BIT): New macro.
7860 (TARGET_16BIT_P): Likewise.
7861 * config/i386/i386.opt: Add m16.
7862 * doc/invoke.texi: Document -m16.
7864 2014-01-28 Jakub Jelinek <jakub@redhat.com>
7866 PR preprocessor/59935
7867 * input.c (location_get_source_line): Bail out on when line number
7868 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
7870 2014-01-28 Richard Biener <rguenther@suse.de>
7872 PR tree-optimization/58742
7873 * tree-ssa-forwprop.c (associate_plusminus): Handle
7874 pointer subtraction of the form (T)(P + A) - (T)P.
7876 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7878 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
7881 2014-01-28 Richard Biener <rguenther@suse.de>
7884 2014-01-28 Richard Biener <rguenther@suse.de>
7886 PR rtl-optimization/45364
7887 PR rtl-optimization/59890
7888 * var-tracking.c (local_get_addr_clear_given_value): Handle
7889 already cleared slot.
7890 (val_reset): Handle not allocated local_get_addr_cache.
7891 (vt_find_locations): Use post-order on the inverted CFG.
7893 2014-01-28 Richard Biener <rguenther@suse.de>
7895 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
7897 2014-01-28 Richard Biener <rguenther@suse.de>
7899 PR rtl-optimization/45364
7900 PR rtl-optimization/59890
7901 * var-tracking.c (local_get_addr_clear_given_value): Handle
7902 already cleared slot.
7903 (val_reset): Handle not allocated local_get_addr_cache.
7904 (vt_find_locations): Use post-order on the inverted CFG.
7906 2014-01-28 Alan Modra <amodra@gmail.com>
7908 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
7909 * configure.ac <recursive call for build != host>: Define
7910 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
7911 and LD_FOR_BUILD too.
7912 * configure: Regenerate.
7914 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
7916 * config/i386/i386.c (get_builtin_code_for_version): Separate
7917 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
7918 Broadwell from Haswell.
7920 2014-01-27 Steve Ellcey <sellcey@mips.com>
7922 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
7923 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
7924 * config/mips/mips.c (mips_option_override): Change setting
7926 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
7927 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
7928 Change from Mask to Var.
7930 2014-01-27 Jeff Law <law@redhat.com>
7932 * ipa-inline.c (inline_small_functions): Fix typo.
7934 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
7936 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
7937 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
7938 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
7939 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
7940 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
7941 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
7942 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
7943 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
7944 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
7945 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
7946 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
7947 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
7948 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
7949 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
7950 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
7951 (_mm512_storeu_epi64): Ditto.
7952 (_mm512_cmpge_epi32_mask): Ditto.
7953 (_mm512_cmpge_epu32_mask): Ditto.
7954 (_mm512_cmpge_epi64_mask): Ditto.
7955 (_mm512_cmpge_epu64_mask): Ditto.
7956 (_mm512_cmple_epi32_mask): Ditto.
7957 (_mm512_cmple_epu32_mask): Ditto.
7958 (_mm512_cmple_epi64_mask): Ditto.
7959 (_mm512_cmple_epu64_mask): Ditto.
7960 (_mm512_cmplt_epi32_mask): Ditto.
7961 (_mm512_cmplt_epu32_mask): Ditto.
7962 (_mm512_cmplt_epi64_mask): Ditto.
7963 (_mm512_cmplt_epu64_mask): Ditto.
7964 (_mm512_cmpneq_epi32_mask): Ditto.
7965 (_mm512_cmpneq_epu32_mask): Ditto.
7966 (_mm512_cmpneq_epi64_mask): Ditto.
7967 (_mm512_cmpneq_epu64_mask): Ditto.
7968 (_mm512_expand_pd): Ditto.
7969 (_mm512_expand_ps): Ditto.
7970 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
7971 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
7972 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
7973 * config/i386/i386.c (ix86_builtins): Add
7974 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
7975 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
7976 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
7977 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
7978 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
7979 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
7980 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
7981 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
7982 IX86_BUILTIN_PMOVUSQW512_MEM.
7983 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
7984 __builtin_ia32_pmovsqd512mem_mask,
7985 __builtin_ia32_pmovqd512mem_mask,
7986 __builtin_ia32_pmovusqw512mem_mask,
7987 __builtin_ia32_pmovsqw512mem_mask,
7988 __builtin_ia32_pmovqw512mem_mask,
7989 __builtin_ia32_pmovusdw512mem_mask,
7990 __builtin_ia32_pmovsdw512mem_mask,
7991 __builtin_ia32_pmovdw512mem_mask,
7992 __builtin_ia32_pmovqb512mem_mask,
7993 __builtin_ia32_pmovusqb512mem_mask,
7994 __builtin_ia32_pmovsqb512mem_mask,
7995 __builtin_ia32_pmovusdb512mem_mask,
7996 __builtin_ia32_pmovsdb512mem_mask,
7997 __builtin_ia32_pmovdb512mem_mask.
7998 (bdesc_args): Add __builtin_ia32_expanddf512,
7999 __builtin_ia32_expandsf512.
8000 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
8001 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
8002 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
8003 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
8004 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
8005 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
8006 (avx512f_<code>v8div16qi2_mask_store): This.
8007 (avx512f_expand<mode>): New.
8009 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
8011 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
8013 (_mm512_mask_prefetch_i64gather_pd): Ditto.
8014 (_mm512_prefetch_i32scatter_pd): Ditto.
8015 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
8016 (_mm512_prefetch_i64scatter_pd): Ditto.
8017 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
8018 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
8019 (_mm512_mask_prefetch_i64gather_ps): Ditto.
8020 (_mm512_prefetch_i32scatter_ps): Ditto.
8021 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
8022 (_mm512_prefetch_i64scatter_ps): Ditto.
8023 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
8024 * config/i386/i386-builtin-types.def: Define
8025 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
8026 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
8027 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
8028 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
8029 IX86_BUILTIN_SCATTERPFQPD.
8030 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
8031 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
8032 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
8033 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
8034 __builtin_ia32_scatterpfqps.
8035 (ix86_expand_builtin): Expand new built-ins.
8036 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8037 fix memory access data type.
8038 (*avx512pf_gatherpf<mode>_mask): Ditto.
8039 (*avx512pf_gatherpf<mode>): Ditto.
8040 (avx512pf_scatterpf<mode>): Ditto.
8041 (*avx512pf_scatterpf<mode>_mask): Ditto.
8042 (*avx512pf_scatterpf<mode>): Ditto.
8043 (GATHER_SCATTER_SF_MEM_MODE): New.
8044 (avx512pf_gatherpf<mode>df): Ditto.
8045 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8046 (*avx512pf_scatterpf<mode>df): Ditto.
8048 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8051 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8052 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8055 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8057 * common/config/arm/arm-common.c
8058 (arm_rewrite_mcpu): Handle multiple names.
8060 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8062 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8064 * gimple-builder.h (create_gimple_tmp): Delete.
8066 2014-01-27 Christian Bruel <christian.bruel@st.com>
8068 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
8071 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
8073 * config/pa/pa.md (call): Generate indirect long calls to non-local
8074 functions when outputing 32-bit code.
8075 (call_value): Likewise except for special call to buggy powf function.
8077 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
8078 portable runtime and PIC indirect calls.
8079 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
8080 and PIC call sequences. Use ldo instead of blr to set return register
8081 in PIC call sequence.
8083 2014-01-25 Walter Lee <walt@tilera.com>
8085 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
8086 avoid clobbering a live register.
8088 2014-01-25 Walter Lee <walt@tilera.com>
8090 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
8091 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
8092 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
8093 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
8095 2014-01-25 Walter Lee <walt@tilera.com>
8097 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
8098 arguments on even registers.
8099 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
8101 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
8102 (BIGGEST_ALIGNMENT): Ditto.
8103 (BIGGEST_FIELD_ALIGNMENT): Ditto.
8105 2014-01-25 Walter Lee <walt@tilera.com>
8107 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
8108 insns before bundling.
8109 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8111 2014-01-25 Walter Lee <walt@tilera.com>
8113 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
8114 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
8115 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
8117 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8119 * config/mips/constraints.md (kl): Delete.
8120 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
8121 define expands, using...
8122 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
8123 instructions for MIPS16.
8124 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
8125 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
8127 2014-01-25 Walter Lee <walt@tilera.com>
8129 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
8133 2014-01-25 Walter Lee <walt@tilera.com>
8135 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
8136 (TARGET_EXPAND_TO_RTL_HOOK): Define.
8138 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8140 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
8143 2014-01-25 Jakub Jelinek <jakub@redhat.com>
8145 * print-rtl.c (in_call_function_usage): New var.
8146 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
8147 EXPR_LIST mode as mode and not as reg note name.
8150 * cfgloopmanip.c (copy_loop_info): If
8151 loop->warned_aggressive_loop_optimizations, make sure
8152 the flag is set in target loop too.
8154 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
8156 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
8158 * builtins.def: Likewise.
8159 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
8160 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
8161 * ira.c (ira_setup_eliminable_regset): Likewise.
8162 * omp-low.c (gate_expand_omp): Likewise.
8163 (execute_lower_omp): Likewise.
8164 (diagnose_sb_0): Likewise.
8165 (gate_diagnose_omp_blocks): Likewise.
8166 (simd_clone_clauses_extract): Likewise.
8169 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8171 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
8172 correction for little endian...
8173 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
8176 2014-01-24 Jeff Law <law@redhat.com>
8178 PR tree-optimization/59919
8179 * tree-vrp.c (find_assert_locations_1): Do not register asserts
8180 for non-returning calls.
8182 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
8184 * common/config/aarch64/aarch64-common.c
8185 (aarch64_rewrite_mcpu): Handle multiple names.
8186 * config/aarch64/aarch64.h
8187 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8189 2014-01-24 Dodji Seketeli <dodji@redhat.com>
8191 * input.c (add_file_to_cache_tab): Handle the case where fopen
8194 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
8197 * config/i386/i386.md (pushsf splitter): Get stack adjustment
8198 from push operand if code of push isn't PRE_DEC.
8200 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8203 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
8204 -mquad-memory-atomic. Update -mquad-memory documentation to say
8205 it is only used for non-atomic loads/stores.
8207 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
8208 -mquad-memory or -mquad-memory-atomic switches.
8210 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
8211 -mquad-memory-atomic to ISA 2.07 support.
8213 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
8214 to separate support of normal quad word memory operations (ldq, stq)
8215 from the atomic quad word memory operations.
8217 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
8218 support to separate non-atomic quad word operations from atomic
8219 quad word operations. Disable non-atomic quad word operations in
8220 little endian mode so that we don't have to swap words after the
8221 load and before the store.
8222 (quad_load_store_p): Add comment about atomic quad word support.
8223 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
8224 options printed with -mdebug=reg.
8226 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
8227 -mquad-memory-atomic as the test for whether we have quad word
8228 atomic instructions.
8229 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
8230 or -mp8-vector are used, allow byte/half-word atomic operations.
8232 * config/rs6000/sync.md (load_lockedti): Insure that the address
8233 is a proper indexed or indirect address for the lqarx instruction.
8234 On little endian systems, swap the hi/lo registers after the lqarx
8236 (load_lockedpti): Use indexed_or_indirect_operand predicate to
8237 insure the address is valid for the lqarx instruction.
8238 (store_conditionalti): Insure that the address is a proper indexed
8239 or indirect address for the stqcrx. instruction. On little endian
8240 systems, swap the hi/lo registers before doing the stqcrx.
8242 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
8243 insure the address is valid for the stqcrx. instruction.
8245 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8246 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
8247 type of quad memory support is available.
8249 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
8252 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
8253 there is a danger of looping.
8255 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
8257 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8258 force flag_ira_loop_pressure if set via command line.
8260 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8262 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
8263 (ashr_simd): New builtin handling DI mode.
8264 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
8265 (aarch64_sshr_simddi): New match pattern.
8266 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
8267 (vshrd_n_s64): Likewise.
8268 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
8270 2014-01-23 Nick Clifton <nickc@redhat.com>
8272 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
8273 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
8274 favour of mcu specific scripts.
8275 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
8278 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8279 Alex Velenko <Alex.Velenko@arm.com>
8281 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
8282 (vaddv_s16): Likewise.
8283 (vaddv_s32): Likewise.
8284 (vaddv_u8): Likewise.
8285 (vaddv_u16): Likewise.
8286 (vaddv_u32): Likewise.
8287 (vaddvq_s8): Likewise.
8288 (vaddvq_s16): Likewise.
8289 (vaddvq_s32): Likewise.
8290 (vaddvq_s64): Likewise.
8291 (vaddvq_u8): Likewise.
8292 (vaddvq_u16): Likewise.
8293 (vaddvq_u32): Likewise.
8294 (vaddvq_u64): Likewise.
8295 (vaddv_f32): Likewise.
8296 (vaddvq_f32): Likewise.
8297 (vaddvq_f64): Likewise.
8298 (vmaxv_f32): Likewise.
8299 (vmaxv_s8): Likewise.
8300 (vmaxv_s16): Likewise.
8301 (vmaxv_s32): Likewise.
8302 (vmaxv_u8): Likewise.
8303 (vmaxv_u16): Likewise.
8304 (vmaxv_u32): Likewise.
8305 (vmaxvq_f32): Likewise.
8306 (vmaxvq_f64): Likewise.
8307 (vmaxvq_s8): Likewise.
8308 (vmaxvq_s16): Likewise.
8309 (vmaxvq_s32): Likewise.
8310 (vmaxvq_u8): Likewise.
8311 (vmaxvq_u16): Likewise.
8312 (vmaxvq_u32): Likewise.
8313 (vmaxnmv_f32): Likewise.
8314 (vmaxnmvq_f32): Likewise.
8315 (vmaxnmvq_f64): Likewise.
8316 (vminv_f32): Likewise.
8317 (vminv_s8): Likewise.
8318 (vminv_s16): Likewise.
8319 (vminv_s32): Likewise.
8320 (vminv_u8): Likewise.
8321 (vminv_u16): Likewise.
8322 (vminv_u32): Likewise.
8323 (vminvq_f32): Likewise.
8324 (vminvq_f64): Likewise.
8325 (vminvq_s8): Likewise.
8326 (vminvq_s16): Likewise.
8327 (vminvq_s32): Likewise.
8328 (vminvq_u8): Likewise.
8329 (vminvq_u16): Likewise.
8330 (vminvq_u32): Likewise.
8331 (vminnmv_f32): Likewise.
8332 (vminnmvq_f32): Likewise.
8333 (vminnmvq_f64): Likewise.
8335 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8337 * config/aarch64/aarch64-simd.md
8338 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
8339 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
8340 (*aarch64_mul3_elt<mode>): Likewise.
8341 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
8342 (*aarch64_mul3_elt_to_64v2df): Likewise.
8343 (*aarch64_mla_elt<mode>): Likewise.
8344 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8345 (*aarch64_mls_elt<mode>): Likewise.
8346 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8347 (*aarch64_fma4_elt<mode>): Likewise.
8348 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8349 (*aarch64_fma4_elt_to_64v2df): Likewise.
8350 (*aarch64_fnma4_elt<mode>): Likewise.
8351 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8352 (*aarch64_fnma4_elt_to_64v2df): Likewise.
8353 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
8354 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
8355 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8356 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8357 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
8358 (aarch64_sqdmull_lane<mode>_internal): Likewise.
8359 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
8361 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
8363 * config/aarch64/aarch64-simd.md
8364 (aarch64_be_checked_get_lane<mode>): New define_expand.
8365 * config/aarch64/aarch64-simd-builtins.def
8366 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
8367 New builtin definition.
8368 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
8369 Use new safe be builtin.
8371 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8373 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
8375 (aarch64_be_st1<mode>): Likewise.
8376 (aarch_ld1<VALL:mode>): Define_expand modified.
8377 (aarch_st1<VALL:mode>): Likewise.
8378 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
8379 (UNSPEC_ST1): Likewise.
8381 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
8383 * config/microblaze/microblaze.md: Add trap insn and attribute
8385 2014-01-23 Dodji Seketeli <dodji@redhat.com>
8387 PR preprocessor/58580
8388 * input.h (location_get_source_line): Take an additional line_size
8390 (void diagnostics_file_cache_fini): Declare new function.
8391 * input.c (struct fcache): New type.
8392 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
8393 New static constants.
8394 (diagnostic_file_cache_init, total_lines_num)
8395 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
8396 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
8397 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
8398 (get_next_line, read_next_line, goto_next_line, read_line_num):
8399 New static function definitions.
8400 (diagnostic_file_cache_fini): New function.
8401 (location_get_source_line): Take an additional output line_len
8402 parameter. Re-write using lookup_or_add_file_to_cache_tab and
8404 * diagnostic.c (diagnostic_finish): Call
8405 diagnostic_file_cache_fini.
8406 (adjust_line): Take an additional input parameter for the length
8407 of the line, rather than calculating it with strlen.
8408 (diagnostic_show_locus): Adjust the use of
8409 location_get_source_line and adjust_line with respect to their new
8410 signature. While displaying a line now, do not stop at the first
8411 null byte. Rather, display the zero byte as a space and keep
8412 going until we reach the size of the line.
8413 * Makefile.in: Add vec.o to OBJS-libcommon
8415 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8416 Ilya Tocar <ilya.tocar@intel.com>
8418 * config/i386/avx512fintrin.h (_mm512_kmov): New.
8419 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
8420 (__builtin_ia32_kmov16): Ditto.
8421 * config/i386/i386.md (UNSPEC_KMOV): New.
8424 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8426 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
8427 (_mm512_storeu_si512): Ditto.
8429 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
8432 * rtl.h (get_referenced_operands): Declare.
8433 * recog.c (get_referenced_operands): New function.
8434 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
8435 operands have been referenced when recording LO_SUM references.
8437 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
8439 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
8441 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8443 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
8444 Enable for generic and recent AMD targets.
8446 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8448 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
8449 ARG_SIZE note when adjustment was eliminated.
8451 2014-01-22 Jeff Law <law@redhat.com>
8453 PR tree-optimization/59597
8454 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
8455 in file. Accept new argument REGISTERING and use it to modify
8456 dump output appropriately.
8457 (register_jump_thread): Corresponding changes.
8458 (mark_threaded_blocks): Reinstate code to cancel unprofitable
8459 thread paths involving joiner blocks. Add code to dump cancelled
8460 jump threading paths.
8462 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
8464 PR rtl-optimization/59477
8465 * lra-constraints.c (inherit_in_ebb): Process call for living hard
8466 regs. Update reloads_num and potential_reload_hard_regs for all insns.
8468 2014-01-22 Tom Tromey <tromey@redhat.com>
8470 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
8472 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
8474 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8476 PR rtl-optimization/59896
8477 * lra-constraints.c (process_alt_operands): Check unused note for
8478 matched operands of insn with no output reloads.
8480 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
8482 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
8483 (mips_move_from_gpr_cost): Likewise.
8485 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8487 PR rtl-optimization/59858
8488 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
8489 ira_class_hard_regs_num.
8490 (process_alt_operands): Increase reject for dying matched operand.
8492 2014-01-21 Jakub Jelinek <jakub@redhat.com>
8495 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
8496 smaller than size, perform several stores or loads and stores
8497 at dst + count - size to store or copy all of size bytes, rather
8498 than just last modesize bytes.
8500 2014-01-20 DJ Delorie <dj@redhat.com>
8502 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
8503 that CLOBBERs are REGs before propogating their values.
8505 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
8508 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
8509 (cgraph_inline_failed_type): New function.
8510 * cgraph.h (DEFCIFCODE): Add type.
8511 (cgraph_inline_failed_type_t): New enum.
8512 (cgraph_inline_failed_type): New prototype.
8513 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
8514 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
8515 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
8516 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
8517 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
8518 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
8519 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
8520 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
8521 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
8522 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
8523 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
8524 OPTIMIZATION_MISMATCH.
8525 * tree-inline.c (expand_call_inline): Emit errors during
8526 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
8528 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8531 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
8532 mode attribute in insn output.
8534 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
8536 * output.h (output_constant): Delete.
8537 * varasm.c (output_constant): Make private.
8539 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
8541 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
8543 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8546 * tree.h (fold_builtin_strcat): New prototype.
8547 * builtins.c (fold_builtin_strcat): No longer static. Add len
8548 argument, if non-NULL, don't call c_strlen. Optimize
8549 directly into __builtin_memcpy instead of __builtin_strcpy.
8550 (fold_builtin_2): Adjust fold_builtin_strcat caller.
8551 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
8553 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8555 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8556 for SImode_address_operand operands, having only a REG argument.
8558 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
8560 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
8561 loader name using mbig-endian.
8562 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
8564 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8566 * doc/invoke.texi (-march): Clarify documentation for AArch64.
8570 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
8572 * config/aarch64/aarch64-protos.h
8573 (aarch64_cannot_change_mode_class_ptr): Declare.
8574 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
8575 aarch64_cannot_change_mode_class_ptr): New.
8576 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
8577 backend hook aarch64_cannot_change_mode_class.
8579 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8581 * common/config/aarch64/aarch64-common.c
8582 (aarch64_handle_option): Don't handle any option order logic here.
8583 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
8584 selected_cpu, warn on architecture version mismatch.
8585 (aarch64_override_options): Fix parsing order for option strings.
8587 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8588 Iain Sandoe <iain@codesourcery.com>
8591 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
8592 warning. Amend comment to reflect current functionality.
8594 2014-01-20 Richard Biener <rguenther@suse.de>
8597 * builtins.c (fold_builtin_strcat): Remove case better handled
8598 by tree-ssa-strlen.c.
8600 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
8602 * config/aarch64/aarch64.opt
8603 (mcpu, march, mtune): Make case-insensitive.
8605 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8608 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8609 if operands[1] is a REG or ZERO_EXTEND of a REG.
8611 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
8613 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
8615 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
8617 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
8618 long non-pic millicode calls.
8620 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8622 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
8624 2014-01-19 Kito Cheng <kito@0xlab.org>
8626 * builtins.c (expand_movstr): Check movstr expand done or fail.
8628 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8629 H.J. Lu <hongjiu.lu@intel.com>
8632 * config/i386/i386.md (*lea<mode>): Zero-extend return register
8633 to DImode for zero-extended addresses.
8635 2014-01-19 Jakub Jelinek <jakub@redhat.com>
8637 PR rtl-optimization/57763
8638 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
8639 on the new indirect jump_insn and increment LABEL_NUSES (label).
8641 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
8645 * config.gcc (x86_archs): New variable.
8646 (x86_64_archs): Likewise.
8647 (x86_cpus): Likewise.
8648 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
8649 --with-arch/--with-cpu= options.
8650 Support --with-arch=/--with-cpu={nehalem,westmere,
8651 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
8653 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8655 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
8656 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
8658 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8660 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
8662 2014-01-18 Jakub Jelinek <jakub@redhat.com>
8665 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
8666 clear cpp_get_options (parse_in)->warn_unused_macros for
8667 ix86_target_macros_internal with cpp_define.
8669 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
8671 * jump.c (delete_related_insns): Keep (use (insn))s.
8672 * reorg.c (redundant_insn): Check for barriers too.
8674 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8676 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
8678 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
8680 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
8681 call to $$dyncall when TARGET_LONG_CALLS is true.
8683 2014-01-17 Jeff Law <law@redhat.com>
8685 * ree.c (combine_set_extension): Temporarily disable test for
8686 changing number of hard registers.
8688 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
8691 * ipa-inline-analysis.c (inline_free_summary):
8692 Do not free summary of aliases.
8694 2014-01-17 Jakub Jelinek <jakub@redhat.com>
8697 * gimplify.c (gimplify_expr): Use create_tmp_var
8698 instead of create_tmp_var_raw. If cond doesn't have
8699 integral type, don't add the IFN_ANNOTATE builtin at all.
8701 2014-01-17 Martin Jambor <mjambor@suse.cz>
8704 * ipa-cp.c (prev_edge_clone): New variable.
8705 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
8706 Also resize prev_edge_clone vector.
8707 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
8708 (ipcp_edge_removal_hook): New function.
8709 (ipcp_driver): Register ipcp_edge_removal_hook.
8711 2014-01-17 Andrew Pinski <apinski@cavium.com>
8712 Steve Ellcey <sellcey@mips.com>
8715 * config/mips/mips.c (mips_print_operand): Check operand mode instead
8718 2014-01-17 Jeff Law <law@redhat.com>
8721 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
8722 so that pass_ccp runs first.
8724 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8726 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
8727 (ix86_adjust_cost): Use !TARGET_XXX.
8728 (do_reorder_for_imul): Likewise.
8729 (swap_top_of_ready_list): Likewise.
8730 (ix86_sched_reorder): Likewise.
8732 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8734 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
8735 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
8736 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
8737 (intel_memset): New. Duplicate slm_memset.
8738 (intel_cost): New. Duplicate slm_cost.
8739 (m_INTEL): New macro.
8740 (processor_target_table): Add "intel".
8741 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
8742 with PROCESSOR_INTEL for "intel".
8743 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
8744 PROCESSOR_SILVERMONT.
8745 (ix86_issue_rate): Likewise.
8746 (ix86_adjust_cost): Likewise.
8747 (ia32_multipass_dfa_lookahead): Likewise.
8748 (swap_top_of_ready_list): Likewise.
8749 (ix86_sched_reorder): Likewise.
8750 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
8751 instead of TARGET_OPT_AGU.
8752 * config/i386/i386.h (TARGET_INTEL): New.
8753 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
8754 (processor_type): Add PROCESSOR_INTEL.
8755 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
8756 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
8758 2014-01-17 Marek Polacek <polacek@redhat.com>
8761 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
8764 2014-01-17 Richard Biener <rguenther@suse.de>
8766 PR tree-optimization/46590
8767 * opts.c (default_options_table): Add entries for
8768 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
8769 all enabled at -O1 but not for -Og.
8770 * common.opt (fbranch-count-reg): Remove Init(1).
8771 (fmove-loop-invariants): Likewise.
8772 (ftree-pta): Likewise.
8774 2014-01-17 Jakub Jelinek <jakub@redhat.com>
8776 * config/i386/i386.c (ix86_data_alignment): For compatibility with
8777 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
8778 decls to at least the GCC 4.8 used alignments.
8781 * tree-nested.c (convert_nonlocal_reference_stmt,
8782 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
8783 of GIMPLE_BIND stmts, adjust associated decls.
8785 2014-01-17 Richard Biener <rguenther@suse.de>
8787 PR tree-optimization/46590
8788 * vec.h (vec<>::bseach): New member function implementing
8789 binary search according to C89 bsearch.
8790 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
8791 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
8792 bitmap pointer again. Make accesses_in_loop a flat array.
8793 (mem_ref_obstack): New global.
8794 (outermost_indep_loop): Adjust for mem_ref->stored changes.
8795 (mark_ref_stored): Likewise.
8796 (ref_indep_loop_p_2): Likewise.
8797 (set_ref_stored_in_loop): New helper function.
8798 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
8799 (memref_free): Adjust.
8800 (record_mem_ref_loc): Simplify.
8801 (gather_mem_refs_stmt): Adjust.
8802 (sort_locs_in_loop_postorder_cmp): New function.
8803 (analyze_memory_references): Sort accesses_in_loop after
8804 loop postorder number.
8805 (find_ref_loc_in_loop_cmp): New function.
8806 (for_all_locs_in_loop): Find relevant cluster of locs in
8807 accesses_in_loop and iterate without recursion.
8808 (execute_sm): Avoid uninit warning.
8809 (struct ref_always_accessed): Simplify.
8810 (ref_always_accessed::operator ()): Likewise.
8811 (ref_always_accessed_p): Likewise.
8812 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
8813 loop postorder numbers here.
8814 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
8817 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
8820 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
8821 on decls for which assemble_alias has been called.
8823 2014-01-17 Nick Clifton <nickc@redhat.com>
8825 * config/msp430/msp430.opt: (mcpu): New option.
8826 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
8827 (msp430_option_override): Parse target_cpu. If the MCU name
8828 matches a generic string, clear target_mcu.
8829 (msp430_attr): Allow numeric interrupt values up to 63.
8830 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
8831 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
8833 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
8835 * config/msp430/msp430.md (popm): Use %J rather than %I.
8836 (addsi3): Use msp430_nonimmediate_operand for operand 2.
8837 (addhi_cy_i): Use immediate_operand for operand 2.
8838 * doc/invoke.texi: Document -mcpu option.
8840 2014-01-17 Richard Biener <rguenther@suse.de>
8842 PR rtl-optimization/38518
8843 * df.h (df_analyze_loop): Declare.
8844 * df-core.c: Include cfgloop.h.
8845 (df_analyze_1): Split out main part of df_analyze.
8846 (df_analyze): Adjust.
8847 (loop_inverted_post_order_compute): New function.
8848 (loop_post_order_compute): Likewise.
8849 (df_analyze_loop): New function avoiding whole-function
8851 * loop-invariant.c (find_defs): Use df_analyze_loop.
8852 (find_invariants): Adjust.
8853 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
8855 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
8857 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
8858 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
8860 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
8862 * ipa-ref.c (ipa_remove_stmt_references): Fix references
8863 traversal when removing references.
8865 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
8868 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
8870 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
8873 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
8874 pushing a reload for an autoinc when we had previously reloaded an
8875 inner part of the address.
8877 2014-01-16 Jakub Jelinek <jakub@redhat.com>
8879 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
8881 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
8882 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
8883 when not giving up or versioning for alias only because of
8885 (vect_analyze_data_ref_dependences): Set to true.
8886 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
8888 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
8889 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
8893 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
8896 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
8897 operand 0 predicate for gathers, use a new pseudo as subtarget.
8899 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
8902 * lra-constraints.c (process_alt_operands): Add printing debug info.
8903 Check absence of input/output reloads for matched operands too.
8905 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
8907 PR rtl-optimization/59835
8908 * ira.c (ira_init_register_move_cost): Increase cost for
8911 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
8913 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
8915 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
8918 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
8919 non-register objects. Use gen_(high/low)part more consistently.
8922 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8925 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
8926 endian support, remove tests for WORDS_BIG_ENDIAN.
8927 (p8_mfvsrd_3_<mode>): Likewise.
8928 (reload_gpr_from_vsx<mode>): Likewise.
8929 (reload_gpr_from_vsxsf): Likewise.
8930 (p8_mfvsrd_4_disf): Likewise.
8932 2014-01-16 Richard Biener <rguenther@suse.de>
8934 PR rtl-optimization/46590
8935 * lcm.c (compute_antinout_edge): Use postorder iteration.
8936 (compute_laterin): Use inverted postorder iteration.
8938 2014-01-16 Nick Clifton <nickc@redhat.com>
8941 * varasm.c (output_constant): Return the number of bytes actually
8943 (output_constructor_array_range): Update the field size with the
8944 number of bytes emitted by output_constant.
8945 (output_constructor_regular_field): Likewise. Also do not
8946 complain if the total number of bytes emitted is now greater
8947 than the expected fieldpos.
8948 * output.h (output_constant): Update prototype and descriptive comment.
8950 2014-01-16 Marek Polacek <polacek@redhat.com>
8953 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
8954 it is error_mark_node.
8956 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
8958 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
8959 VALID_AVX256_REG_OR_OI_MODE.
8961 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
8963 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
8964 current procedure should be profiled.
8966 2014-01-15 Andrew Pinski <apinski@cavium.com>
8968 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
8969 of moving from/to the STACK_REG register class.
8971 2014-01-15 Richard Henderson <rth@redhat.com>
8974 * reginfo.c (global_regs_decl): Globalize.
8975 * rtl.h (global_regs_decl): Declare.
8976 * ira.c (do_reload): Diagnose frame_pointer_needed and it
8977 reserved via global_regs.
8979 2014-01-15 Teresa Johnson <tejohnson@google.com>
8981 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
8983 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8985 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
8986 and vmulosh rather than call gen_vec_widen_smult_*.
8987 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
8988 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
8989 (vec_widen_smult_even_v16qi): Likewise.
8990 (vec_widen_umult_even_v8hi): Likewise.
8991 (vec_widen_smult_even_v8hi): Likewise.
8992 (vec_widen_umult_odd_v16qi): Likewise.
8993 (vec_widen_smult_odd_v16qi): Likewise.
8994 (vec_widen_umult_odd_v8hi): Likewise.
8995 (vec_widen_smult_odd_v8hi): Likewise.
8996 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
8997 vmuloub rather than call gen_vec_widen_umult_*.
8998 (vec_widen_umult_lo_v16qi): Likewise.
8999 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
9000 vmulosb rather than call gen_vec_widen_smult_*.
9001 (vec_widen_smult_lo_v16qi): Likewise.
9002 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
9003 rather than call gen_vec_widen_umult_*.
9004 (vec_widen_umult_lo_v8hi): Likewise.
9005 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
9006 rather than call gen_vec_widen_smult_*.
9007 (vec_widen_smult_lo_v8hi): Likewise.
9009 2014-01-15 Jeff Law <law@redhat.com>
9011 PR tree-optimization/59747
9012 * ree.c (find_and_remove_re): Properly handle case where a second
9013 eliminated extension requires widening a copy created for elimination
9014 of a prior extension.
9015 (combine_set_extension): Ensure that the number of hard regs needed
9016 for a destination register does not change when we widen it.
9018 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
9020 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
9021 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
9022 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
9023 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
9024 (avr-*-rtems*): Likewise.
9025 (bfin*-rtems*): Likewise.
9026 (moxie-*-rtems*): Likewise.
9027 (h8300-*-rtems*): Likewise.
9028 (i[34567]86-*-rtems*): Likewise.
9029 (lm32-*-rtems*): Likewise.
9030 (m32r-*-rtems*): Likewise.
9031 (m68k-*-rtems*): Likewise.
9032 (microblaze*-*-rtems*): Likewise.
9033 (mips*-*-rtems*): Likewise.
9034 (powerpc-*-rtems*): Likewise.
9035 (sh-*-rtems*): Likewise.
9036 (sparc-*-rtems*): Likewise.
9037 (sparc64-*-rtems*): Likewise.
9038 (v850-*-rtems*): Likewise.
9039 (m32c-*-rtems*): Likewise.
9041 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9043 PR rtl-optimization/59511
9044 * ira.c (ira_init_register_move_cost): Use memory costs for some
9045 cases of register move cost calculations.
9046 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9047 instead of BB frequency.
9048 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9049 * lra-assigns.c (find_hard_regno_for): Ditto.
9051 2014-01-15 Richard Biener <rguenther@suse.de>
9053 PR tree-optimization/59822
9054 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9055 (vectorizable_load): Use it to hoist defs of uses of invariant
9056 loads out of the loop.
9058 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9059 Kugan Vivekanandarajah <kuganv@linaro.org>
9062 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9065 2014-01-15 Richard Biener <rguenther@suse.de>
9067 PR rtl-optimization/59802
9068 * lcm.c (compute_available): Use inverted postorder to seed
9069 the initial worklist.
9071 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9074 * config/s390/s390.c (s390_preferred_reload_class): Don't return
9075 ADDR_REGS for invalid symrefs in non-PIC code.
9077 2014-01-15 Jakub Jelinek <jakub@redhat.com>
9080 * builtins.c (determine_block_size): Initialize *probable_max_size
9081 even if len_rtx is CONST_INT.
9083 2014-01-14 Andrew Pinski <apinski@cavium.com>
9085 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
9086 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
9087 (cortexa53_tunings): Likewise.
9088 (aarch64_sched_issue_rate): New function.
9089 (TARGET_SCHED_ISSUE_RATE): Define.
9091 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9093 * ira-costs.c (find_costs_and_classes): Add missed
9094 ira_init_register_move_cost_if_necessary.
9096 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9099 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
9101 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
9104 * config/i386/i386.c (type_natural_mode): Add a bool parameter
9105 to indicate if type is used for function return value. Warn ABI
9106 change if the vector mode isn't available for function return value.
9107 (ix86_function_arg_advance): Pass false to type_natural_mode.
9108 (ix86_function_arg): Likewise.
9109 (ix86_gimplify_va_arg): Likewise.
9110 (function_arg_32): Don't warn ABI change.
9111 (ix86_function_value): Pass true to type_natural_mode.
9112 (ix86_return_in_memory): Likewise.
9113 (ix86_struct_value_rtx): Removed.
9114 (TARGET_STRUCT_VALUE_RTX): Likewise.
9116 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9118 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
9119 converting a conditional jump into a conditional return.
9121 2014-01-14 Richard Biener <rguenther@suse.de>
9123 PR tree-optimization/58921
9124 PR tree-optimization/59006
9125 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
9126 hoisting invariant stmts.
9127 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
9128 invariant loads on the preheader edge if possible.
9130 2014-01-14 Joey Ye <joey.ye@arm.com>
9132 * doc/plugin.texi (Building GCC plugins): Update to C++.
9134 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
9136 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
9137 (_mm_rcp28_round_ss): Ditto.
9138 (_mm_rsqrt28_round_sd): Ditto.
9139 (_mm_rsqrt28_round_ss): Ditto.
9140 (_mm_rcp28_sd): Ditto.
9141 (_mm_rcp28_ss): Ditto.
9142 (_mm_rsqrt28_sd): Ditto.
9143 (_mm_rsqrt28_ss): Ditto.
9144 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
9145 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
9146 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
9147 (IX86_BUILTIN_RCP28SD): Ditto.
9148 (IX86_BUILTIN_RCP28SS): Ditto.
9149 (IX86_BUILTIN_RSQRT28SD): Ditto.
9150 (IX86_BUILTIN_RSQRT28SS): Ditto.
9151 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
9152 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
9153 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
9154 (ix86_expand_special_args_builtin): Expand new FTYPE.
9155 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
9156 (srcp14<mode>): Make insn unary.
9157 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
9158 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
9159 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
9160 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
9161 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
9162 Fix rounding: make it SAE only.
9163 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9165 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9167 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
9168 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
9169 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
9170 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
9171 (round_saeonly_mask_scalar_operand4): Ditto.
9172 (round_saeonly_mask_scalar_op3): Ditto.
9173 (round_saeonly_mask_scalar_op4): Ditto.
9175 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9177 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9178 Implement -maltivec=be for vec_insert and vec_extract.
9180 2014-01-10 DJ Delorie <dj@redhat.com>
9182 * config/msp430/msp430.md (call_internal): Don't allow memory
9183 references with SP as the base register.
9184 (call_value_internal): Likewise.
9185 * config/msp430/constraints.md (Yc): New. For memory references
9186 that don't use SP as a base register.
9188 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
9189 "an integer without a # prefix"
9190 * config/msp430/msp430.md (epilogue_helper): Use it.
9192 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9195 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
9196 AVX512F gather builtins.
9197 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
9198 on gather decls with INTEGER_TYPE masktype.
9199 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
9200 directly into the builtin rather than hoisting it before loop.
9202 PR tree-optimization/59387
9203 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
9204 (scev_const_prop): If folded_casts and type has undefined overflow,
9205 use force_gimple_operand instead of force_gimple_operand_gsi and
9206 for each added stmt if it is assign with
9207 arith_code_with_undefined_signed_overflow, call
9208 rewrite_to_defined_overflow.
9209 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
9210 gimple-fold.h instead.
9211 (arith_code_with_undefined_signed_overflow,
9212 rewrite_to_defined_overflow): Moved to ...
9213 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
9214 rewrite_to_defined_overflow): ... here. No longer static.
9215 Include gimplify-me.h.
9216 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
9217 rewrite_to_defined_overflow): New prototypes.
9219 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9221 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
9223 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
9225 * builtins.c (get_object_alignment_2): Minor tweak.
9226 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
9228 2014-01-13 Christian Bruel <christian.bruel@st.com>
9230 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
9231 optimized non constant lengths.
9233 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9236 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
9237 load as __atomic_load_N if possible.
9239 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9241 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
9243 (rs6000_expand_builtin): Adjust call.
9245 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9248 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
9249 * config/rs6000/rs6000.c: Include target-globals.h.
9250 (rs6000_set_current_function): Instead of doing target_reinit
9251 unconditionally, use save_target_globals_default_opts and
9252 restore_target_globals.
9254 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
9256 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
9257 (rs6000_expand_builtin): Handle mffs and mtfsf.
9258 (rs6000_init_builtins): Define mffs and mtfsf.
9259 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
9260 (rs6000_mffs): New pattern.
9261 (rs6000_mtfsf): New pattern.
9263 2014-01-11 Bin Cheng <bin.cheng@arm.com>
9265 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
9266 Start narrowing with START. Apply candidate-use pair
9267 and check overall cost in narrowing.
9268 (iv_ca_prune): Pass new argument.
9270 2014-01-10 Jeff Law <law@redhat.com>
9273 * ree.c (combine_reaching_defs): Ensure the defining statement
9274 occurs before the extension when optimizing extensions with
9275 different source and destination hard registers.
9277 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9280 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
9281 vtables into the type inheritance graph.
9283 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9285 PR rtl-optimization/59754
9286 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
9287 modes in the REGNO != REGNO case.
9289 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9291 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
9293 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9295 PR tree-optimization/59745
9296 * tree-predcom.c (tree_predictive_commoning_loop): Call
9297 free_affine_expand_cache if giving up because components is NULL.
9299 * target-globals.c (save_target_globals): Allocate < 4KB structs using
9300 GC in payload of target_globals struct instead of allocating them on
9301 the heap and the larger structs separately using GC.
9302 * target-globals.h (struct target_globals): Make regs, hard_regs,
9303 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
9304 of GTY((skip)) and change type to void *.
9305 (reset_target_globals): Cast loads from those fields to corresponding
9308 2014-01-10 Steve Ellcey <sellcey@mips.com>
9311 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
9312 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
9313 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
9315 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
9318 * aarch64-modes.def (CC_Zmode): New flags mode.
9319 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
9320 represents an equality.
9321 (aarch64_get_condition_code): Handle CC_Zmode.
9322 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
9324 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9326 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
9327 extraction in good case.
9329 2014-01-10 Richard Biener <rguenther@suse.de>
9331 PR tree-optimization/59374
9332 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
9333 checking after SLP discovery. Mark stmts not participating
9334 in any SLP instance properly.
9336 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9338 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
9339 when handling a SET rtx.
9341 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9343 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
9344 (cortex-a57): Likewise.
9345 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
9347 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9349 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
9350 non-iwmmxt builtins.
9352 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9356 * ipa-devirt.c record_target_from_binfo): Take as argument
9357 stack of binfos and lookup matching one for virtual inheritance.
9358 (possible_polymorphic_call_targets_1): Update.
9360 2014-01-10 Huacai Chen <chenhc@lemote.com>
9362 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
9363 kernel strings for Loongson-2E/2F/3A.
9365 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9368 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
9369 is_gimple_call before calling gimple_call_internal_p.
9371 2014-01-09 Steve Ellcey <sellcey@mips.com>
9373 * Makefile.in (TREE_FLOW_H): Remove.
9374 (TREE_SSA_H): Add file names from tree-flow.h.
9375 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
9376 * tree.h: Remove tree-flow.h reference.
9377 * hash-table.h: Remove tree-flow.h reference.
9378 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
9379 reference with tree-ssa-loop.h.
9381 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9383 * doc/invoke.texi: Add -maltivec={be,le} options, and document
9384 default element-order behavior for -maltivec.
9385 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
9386 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
9387 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
9388 when targeting big endian, at least for now.
9389 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
9391 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9394 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
9395 var satisfies use_register_for_decl, just take into account type
9396 alignment, rather than decl alignment.
9398 PR tree-optimization/59622
9399 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
9400 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
9401 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
9402 Don't devirtualize for inplace at all. For targets.length () == 1,
9403 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
9405 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9407 * config/i386/i386.md (cpu): Remove the unused btver1.
9409 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9411 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
9413 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9416 * tree-core.h (struct target_globals): New forward declaration.
9417 (struct tree_target_option): Add globals field.
9418 * tree.h (TREE_TARGET_GLOBALS): Define.
9419 (prepare_target_option_nodes_for_pch): New prototype.
9420 * target-globals.h (struct target_globals): Define even if
9422 * tree.c (prepare_target_option_node_for_pch,
9423 prepare_target_option_nodes_for_pch): New functions.
9424 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
9425 * config/i386/i386.c: Include target-globals.h.
9426 (ix86_set_current_function): Instead of doing target_reinit
9427 unconditionally, use save_target_globals_default_opts and
9428 restore_target_globals.
9430 2014-01-09 Richard Biener <rguenther@suse.de>
9432 PR tree-optimization/59715
9433 * tree-cfg.h (split_critical_edges): Declare.
9434 * tree-cfg.c (split_critical_edges): Export.
9435 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
9437 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
9439 * cfgexpand.c (expand_stack_vars): Optionally disable
9440 asan stack protection.
9441 (expand_used_vars): Likewise.
9442 (partition_stack_vars): Likewise.
9443 * asan.c (asan_emit_stack_protection): Optionally disable
9444 after return stack usage.
9445 (instrument_derefs): Optionally disable memory access instrumentation.
9446 (instrument_builtin_call): Likewise.
9447 (instrument_strlen_call): Likewise.
9448 (asan_protect_global): Optionally disable global variables protection.
9449 * doc/invoke.texi: Added doc for new options.
9450 * params.def: Added new options.
9451 * params.h: Likewise.
9453 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9455 PR rtl-optimization/59724
9456 * ifcvt.c (cond_exec_process_if_block): Don't call
9457 flow_find_head_matching_sequence with 0 longest_match.
9458 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
9459 non-active insns if !stop_after.
9460 (try_head_merge_bb): Revert 2014-01-07 changes.
9462 2014-01-08 Jeff Law <law@redhat.com>
9464 * ree.c (get_sub_rtx): New function, extracted from...
9465 (merge_def_and_ext): Here.
9466 (combine_reaching_defs): Use get_sub_rtx.
9468 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
9470 * cgraph.h (varpool_variable_node): Do not choke on null node.
9472 2014-01-08 Catherine Moore <clm@codesourcery.com>
9474 * config/mips/mips.md (simple_return): Attempt to use JRC
9476 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
9478 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9480 PR rtl-optimization/59137
9481 * reorg.c (steal_delay_list_from_target): Call update_block for
9483 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
9485 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9487 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
9488 two duplicate entries.
9490 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9493 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
9495 * config/mips/mips.c (mips_truncated_op_cost): New function.
9496 (mips_rtx_costs): Adjust test for BADDU.
9497 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
9499 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
9501 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
9502 (*baddu_si): ...this new pattern.
9504 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9507 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
9509 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9512 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
9514 (expand_expr, expand_normal): Adjust.
9515 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
9516 inner_reference_p. Use inner_reference_p to expand inner references.
9517 (store_expr): Adjust.
9518 * cfgexpand.c (expand_call_stmt): Adjust.
9520 2014-01-08 Rong Xu <xur@google.com>
9522 * gcov-io.c (gcov_var): Move from gcov-io.h.
9523 (gcov_position): Ditto.
9524 (gcov_is_error): Ditto.
9525 (gcov_rewrite): Ditto.
9526 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
9527 only part to libgcc/libgcov.h.
9529 2014-01-08 Marek Polacek <polacek@redhat.com>
9532 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
9534 2014-01-08 Marek Polacek <polacek@redhat.com>
9537 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
9539 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9541 PR rtl-optimization/59649
9542 * stor-layout.c (get_mode_bounds): For BImode return
9543 0 and STORE_FLAG_VALUE.
9545 2014-01-08 Richard Biener <rguenther@suse.de>
9548 * gimple.h (is_gimple_builtin_call): Remove.
9549 (gimple_builtin_call_types_compatible_p): New.
9550 (gimple_call_builtin_p): New overload.
9551 * gimple.c (is_gimple_builtin_call): Remove.
9552 (validate_call): Rename to ...
9553 (gimple_builtin_call_types_compatible_p): ... this and export. Also
9555 (validate_type): New static function.
9556 (gimple_call_builtin_p): New overload and adjust.
9557 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
9558 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
9559 (gimple_fold_stmt_to_constant_1): Likewise.
9560 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
9562 2014-01-08 Richard Biener <rguenther@suse.de>
9565 * gimplify.c (gimplify_expr): Gimplify register-register type
9566 VIEW_CONVERT_EXPRs to separate stmts.
9568 2014-01-07 Jeff Law <law@redhat.com>
9571 * ree.c (combine_set_extension): Handle case where source
9572 and destination registers in an extension insn are different.
9573 (combine_reaching_defs): Allow source and destination registers
9574 in extension to be different under limited circumstances.
9575 (add_removable_extension): Remove restriction that the
9576 source and destination registers in the extension are the same.
9577 (find_and_remove_re): Emit a copy from the extension's
9578 destination to its source after the defining insn if
9579 the source and destination registers are different.
9582 * ifcvt.c (merge_if_block): If we are merging a block with more than
9583 one successor with a block with no successors, remove any BARRIER
9584 after the second block.
9586 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
9588 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
9590 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
9593 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
9594 for 14-bit register offsets when INT14_OK_STRICT is false.
9596 2014-01-07 Roland Stigge <stigge@antcom.de>
9597 Michael Meissner <meissner@linux.vnet.ibm.com>
9600 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
9601 Only check TFmode for SPE constants. Don't check TImode or TDmode.
9603 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
9605 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
9608 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
9610 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
9611 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
9612 rtx is const0_rtx or not.
9614 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
9617 * target-globals.c (save_target_globals): Remove this_fn_optab
9619 * toplev.c: Include optabs.h.
9620 (target_reinit): Temporarily restore the global options if another
9621 set of options are in force.
9623 2014-01-07 Jakub Jelinek <jakub@redhat.com>
9625 PR rtl-optimization/58668
9626 * cfgcleanup.c (flow_find_cross_jump): Don't count
9627 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
9628 to determine what is counted.
9629 (flow_find_head_matching_sequence): Use active_insn_p to determine
9631 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
9633 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
9634 determine what is counted.
9636 PR tree-optimization/59643
9637 * tree-predcom.c (split_data_refs_to_components): If one dr is
9638 read and one write, determine_offset fails and the write isn't
9639 in the bad component, just put the read into the bad component.
9641 2014-01-07 Mike Stump <mikestump@comcast.net>
9642 Jakub Jelinek <jakub@redhat.com>
9645 * tree-core.h (struct tree_optimization_option): Change optabs
9646 type from unsigned char * to void *.
9647 * optabs.c (init_tree_optimization_optabs): Adjust
9648 TREE_OPTIMIZATION_OPTABS initialization.
9650 2014-01-06 Jakub Jelinek <jakub@redhat.com>
9653 * config/i386/i386.h (struct machine_function): Add
9654 no_drap_save_restore field.
9655 * config/i386/i386.c (ix86_save_reg): Use
9656 !cfun->machine->no_drap_save_restore instead of
9657 crtl->stack_realign_needed.
9658 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
9659 this function clears frame_pointer_needed. Set
9660 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
9661 and DRAP reg is needed.
9663 2014-01-06 Marek Polacek <polacek@redhat.com>
9666 * doc/implement-c.texi: Mention that other integer types are
9667 permitted as bit-field types in strictly conforming mode.
9669 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
9671 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
9674 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
9676 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
9678 2014-01-06 Martin Jambor <mjambor@suse.cz>
9681 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
9683 * ipa-prop.c (ipa_print_node_params): Fix indentation.
9685 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
9689 * var-tracking.c (add_stores): Preserve the value of the source even if
9690 we don't record the store.
9692 2014-01-06 Terry Guo <terry.guo@arm.com>
9694 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
9696 2014-01-05 Iain Sandoe <iain@codesourcery.com>
9699 * config/darwin.c (darwin_function_section): Adjust return values to
9700 correspond to optimisation changes made in r206070.
9702 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
9704 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
9705 from prefetch_block tune setting.
9706 (nocona_cost): Correct size of prefetch block to 64.
9708 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
9710 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
9711 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
9712 used to save the static chain register in the computation of the offset
9713 from which the FP registers need to be restored.
9715 2014-01-04 Jakub Jelinek <jakub@redhat.com>
9717 PR tree-optimization/59519
9718 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
9719 ICE if get_current_def (current_new_name) is already non-NULL, as long
9720 as it is a phi result of some other phi in *new_exit_bb that has
9723 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
9724 or vmovdqu* for misaligned_operand.
9725 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
9726 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
9727 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
9728 aligned_mem for AVX512F masked aligned load and store builtins and for
9731 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
9733 PR tree-optimization/59651
9734 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
9735 Address range for negative step should be added by TYPE_SIZE_UNIT.
9737 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
9739 * config/m68k/m68k.c (handle_move_double): Handle pushes with
9740 overlapping registers also for registers other than the stack pointer.
9742 2014-01-03 Marek Polacek <polacek@redhat.com>
9745 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
9748 2014-01-03 Jakub Jelinek <jakub@redhat.com>
9751 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
9754 * config/i386/i386.md (MODE_SIZE): New mode attribute.
9755 (push splitter): Use <P:MODE_SIZE> instead of
9756 GET_MODE_SIZE (<P:MODE>mode).
9757 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
9758 (mov -1, reg peephole2): Likewise.
9759 * config/i386/sse.md (*mov<mode>_internal,
9760 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
9761 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
9762 *<code><mode>3, *andnot<mode>3<mask_name>,
9763 <mask_codefor><code><mode>3<mask_name>): Likewise.
9764 * config/i386/subst.md (mask_mode512bit_condition,
9765 sd_mask_mode512bit_condition): Likewise.
9767 2014-01-02 Xinliang David Li <davidxl@google.com>
9769 PR tree-optimization/59303
9770 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
9771 (dump_predicates): Better output format.
9772 (pred_equal_p): New function.
9773 (is_neq_relop_p): Ditto.
9774 (is_neq_zero_form_p): Ditto.
9775 (pred_expr_equal_p): Ditto.
9776 (pred_neg_p): Ditto.
9777 (simplify_pred): Ditto.
9778 (simplify_preds_2): Ditto.
9779 (simplify_preds_3): Ditto.
9780 (simplify_preds_4): Ditto.
9781 (simplify_preds): Ditto.
9783 (push_to_worklist): Ditto.
9784 (get_pred_info_from_cmp): Ditto.
9785 (is_degenerated_phi): Ditto.
9786 (normalize_one_pred_1): Ditto.
9787 (normalize_one_pred): Ditto.
9788 (normalize_one_pred_chain): Ditto.
9789 (normalize_preds): Ditto.
9790 (normalize_cond_1): Remove function.
9791 (normalize_cond): Ditto.
9792 (is_gcond_subset_of): Ditto.
9793 (is_subset_of_any): Ditto.
9794 (is_or_set_subset_of): Ditto.
9795 (is_and_set_subset_of): Ditto.
9796 (is_norm_cond_subset_of): Ditto.
9797 (pred_chain_length_cmp): Ditto.
9798 (convert_control_dep_chain_into_preds): Type change.
9799 (find_predicates): Ditto.
9800 (find_def_preds): Ditto.
9801 (destroy_predicates_vecs): Ditto.
9802 (find_matching_predicates_in_rest_chains): Ditto.
9803 (use_pred_not_overlap_with_undef_path_pred): Ditto.
9804 (is_pred_expr_subset): Ditto.
9805 (is_pred_chain_subset_of): Ditto.
9806 (is_included_in): Ditto.
9807 (is_superset_of): Ditto.
9809 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
9811 Update copyright years.
9813 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
9815 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
9816 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
9817 config/arc/arc.md, config/arc/arc.opt,
9818 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
9819 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
9820 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
9821 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
9822 config/linux-protos.h, config/linux.c, config/winnt-c.c,
9823 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
9824 vtable-verify.c, vtable-verify.h: Use the standard form for the
9827 2014-01-02 Tobias Burnus <burnus@net-b.de>
9829 * gcc.c (process_command): Update copyright notice dates.
9830 * gcov-dump.c: Ditto.
9832 * doc/cpp.texi: Bump @copying's copyright year.
9833 * doc/cppinternals.texi: Ditto.
9834 * doc/gcc.texi: Ditto.
9835 * doc/gccint.texi: Ditto.
9836 * doc/gcov.texi: Ditto.
9837 * doc/install.texi: Ditto.
9838 * doc/invoke.texi: Ditto.
9840 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9842 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
9844 2014-01-01 Jakub Jelinek <jakub@redhat.com>
9846 * config/i386/sse.md (*mov<mode>_internal): Guard
9847 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
9849 PR rtl-optimization/59647
9850 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
9851 new_rtx into UNSIGNED_FLOAT rtxes.
9853 Copyright (C) 2014 Free Software Foundation, Inc.
9855 Copying and distribution of this file, with or without modification,
9856 are permitted in any medium without royalty provided the copyright
9857 notice and this notice are preserved.