1 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
3 * loop-unswitch.c: Delete.
5 2014-05-07 Richard Biener <rguenther@suse.de>
7 * config.gcc: Always set need_64bit_hwint to yes.
9 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
11 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
12 of using optimize_size.
14 2014-05-06 Mike Stump <mikestump@comcast.net>
16 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
18 2014-05-06 Joseph Myers <joseph@codesourcery.com>
20 * config/i386/sse.md (*mov<mode>_internal)
21 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
22 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
23 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
24 (*<code><mode>3, *andnot<mode>3<mask_name>)
25 (<mask_codefor><code><mode>3<mask_name>): Only consider
26 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
28 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
31 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
33 * lra-constraints.c (valid_address_p): Move earlier in file.
34 Add a constraint argument to the address_info version.
35 (satisfies_memory_constraint_p): New function.
36 (satisfies_address_constraint_p): Likewise.
37 (process_alt_operands, curr_insn_transform): Use them.
38 (process_address): Pass the constraint to valid_address_p when
39 checking address operands.
41 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
43 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
44 to their respective blocks. Fix inadvertent use of "node".
46 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
48 * emit-rtl.c (init_derived_machine_modes): New functionm, split
50 (init_emit_once): ...here.
51 * rtl.h (init_derived_machine_modes): Declare.
52 * toplev.c (do_compile): Call it even if no_backend.
54 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
55 Mike Stump <mikestump@comcast.net>
56 Richard Sandiford <rdsandiford@googlemail.com>
57 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
59 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
60 (rtx_equal_for_memref_p): Update comment.
61 (adjust_offset_for_component_ref): Use wide-int interfaces.
62 * builtins.c (get_object_alignment_2): Likewise.
63 (c_readstr): Likewise.
64 (target_char_cast): Add comment.
65 (determine_block_size): Use wide-int interfaces.
66 (expand_builtin_signbit): Likewise.
67 (fold_builtin_int_roundingfn): Likewise.
68 (fold_builtin_bitop): Likewise.
69 (fold_builtin_bswap): Likewise.
70 (fold_builtin_logarithm): Use signop.
71 (fold_builtin_pow): Likewise.
72 (fold_builtin_memory_op): Use wide-int interfaces.
73 (fold_builtin_object_size): Likewise.
74 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
75 nb_iterations_estimate.
76 (record_niter_bound): Use wide-int interfaces.
77 (get_estimated_loop_iterations_int): Likewise.
78 (get_estimated_loop_iterations): Likewise.
79 (get_max_loop_iterations): Likewise.
80 * cfgloop.h: Include wide-int.h.
81 (struct nb_iter_bound): Change bound to widest_int.
82 (struct loop): Change nb_iterations_upper_bound and
83 nb_iterations_estimate to widest_int.
84 (record_niter_bound): Switch to use widest_int.
85 (get_estimated_loop_iterations): Likewise.
86 (get_max_loop_iterations): Likewise.
87 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
89 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
90 * combine.c (try_combine): Likewise.
91 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
92 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
94 (aarch64_float_const_representable_p): Likewise.
95 * config/arc/arc.c: Include wide-int.h.
96 (arc_can_use_doloop_p): Use wide-int interfaces.
97 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
98 (vfp3_const_double_index): Likewise.
99 * config/avr/avr.c (avr_out_round): Likewise.
100 (avr_fold_builtin): Likewise.
101 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
102 (bfin_can_use_doloop_p): Likewise.
103 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
104 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
105 * config/i386/i386.c: Include wide-int.h.
106 (ix86_data_alignment): Use wide-int interfaces.
107 (ix86_local_alignment): Likewise.
108 (ix86_emit_swsqrtsf): Update real_from_integer.
109 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
110 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
111 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
112 (zero_constant): Likewise.
113 (input_operand): Likewise.
114 (splat_input_operand): Likewise.
115 (non_logical_cint_operand): Change const_double to const_wide_int.
116 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
117 (easy_altivec_constant): Remove comment.
118 (paired_expand_vector_init): Use CONSTANT_P.
119 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
120 (rs6000_emit_move): Update checks.
121 (rs6000_aggregate_candidate): Use wide-int interfaces.
122 (rs6000_expand_ternop_builtin): Likewise.
123 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
124 (rs6000_assemble_integer): Likewise.
125 (rs6000_hash_constant): Likewise.
126 (output_toc): Likewise.
127 (rs6000_rtx_costs): Likewise.
128 (rs6000_emit_swrsqrt); Update call to real_from_integer.
129 * config/rs6000/rs6000-c.c: Include wide-int.h.
130 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
131 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
132 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
133 Handle CONST_WIDE_INT.
134 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
135 Use tree_fits_uhwi_p.
136 * config/sparc/sparc.c: Include wide-int.h.
137 (sparc_fold_builtin): Use wide-int interfaces.
138 * config/vax/vax.c: Include wide-int.h.
139 (vax_float_literal): Use real_from_integer.
140 * coretypes.h (struct hwivec_def): New.
143 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
144 (equiv_constant): Handle CONST_WIDE_INT.
145 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
146 (cselib_hash_rtx): Handle CONST_WIDE_INT.
147 * dbxout.c (stabstr_U): Use wide-int interfaces.
148 (dbxout_type): Update to use cst_fits_shwi_p.
149 * defaults.h (LOG2_BITS_PER_UNIT): Define.
150 (TARGET_SUPPORTS_WIDE_INT): Add default.
151 * dfp.c: Include wide-int.h.
152 (decimal_real_to_integer2): Use wide-int interfaces and rename to
153 decimal_real_to_integer.
154 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
155 decimal_real_to_integer.
156 * doc/generic.texi (Constant expressions): Update for wide_int.
157 * doc/rtl.texi (const_double): Likewise.
158 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
159 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
160 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
161 (REAL_VALUE_FROM_INT): Remove.
162 (TARGET_SUPPORTS_WIDE_INT): New.
163 * doc/tm.texi: Regenerate.
164 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
165 * double-int.h: Include wide-int.h.
166 (struct wi::int_traits): New.
167 * dwarf2out.c (get_full_len): New.
168 (dw_val_equal_p): Add case dw_val_class_wide_int.
169 (size_of_loc_descr): Likewise.
170 (output_loc_operands): Likewise.
171 (insert_double): Remove.
172 (insert_wide_int): New.
174 (print_die): Add case dw_val_class_wide_int.
175 (attr_checksum): Likewise.
176 (attr_checksum_ordered): Likewise.
177 (same_dw_val_p): Likewise.
178 (size_of_die): Likewise.
179 (value_format): Likewise.
180 (output_die): Likewise.
181 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
183 (clz_loc_descriptor): Use wide-int interfaces.
184 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
185 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
186 (round_up_to_align): Use wide-int interfaces.
187 (field_byte_offset): Likewise.
188 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
189 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
190 CONST_DOUBLE handling. Use wide-int interfaces.
191 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
192 (gen_enumeration_type_die): Use add_AT_wide.
193 (hash_loc_operands): Add case dw_val_class_wide_int.
194 (compare_loc_operands): Likewise.
195 * dwarf2out.h: Include wide-int.h.
197 (enum dw_val_class): Add dw_val_class_wide_int.
198 (struct dw_val_struct): Add val_wide.
199 * emit-rtl.c (const_wide_int_htab): New.
200 (const_wide_int_htab_hash): New.
201 (const_wide_int_htab_eq): New.
202 (lookup_const_wide_int): New.
203 (const_double_htab_hash): Use wide-int interfaces.
204 (const_double_htab_eq): Likewise.
205 (rtx_to_double_int): Conditionally compile for wide-int.
206 (immed_double_int_const): Rename to immed_wide_int_const and
208 (immed_double_const): Conditionally compile for wide-int.
209 (init_emit_once): Use wide-int interfaces.
210 * explow.c (plus_constant): Likewise.
211 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
212 (lshift_value): Use wide-int interfaces.
213 (expand_mult): Likewise.
214 (choose_multiplier): Likewise.
215 (expand_smod_pow2): Likewise.
216 (make_tree): Likewise.
217 * expr.c (convert_modes): Consolidate handling of constants.
218 Use wide-int interfaces.
219 (emit_group_load_1): Add note.
220 (store_expr): Update comment.
221 (get_inner_reference): Use wide-int interfaces.
222 (expand_constructor): Update comment.
223 (expand_expr_real_2): Use wide-int interfaces.
224 (expand_expr_real_1): Likewise.
225 (reduce_to_bit_field_precision): Likewise.
226 (const_vector_from_tree): Likewise.
227 * final.c: Include wide-int-print.h.
228 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
229 * fixed-value.c: Include wide-int.h.
230 (fixed_from_string): Use wide-int interfaces.
231 (fixed_to_decimal): Likewise.
232 (fixed_convert_from_real): Likewise.
233 (real_convert_from_fixed): Likewise.
234 * fold-const.h (mem_ref_offset): Return an offset_int.
235 (div_if_zero_remainder): Remove code parameter.
236 * fold-const.c (div_if_zero_remainder): Remove code parameter.
237 Use wide-int interfaces.
238 (may_negate_without_overflow_p): Use wide-int interfaces.
239 (negate_expr_p): Likewise.
240 (fold_negate_expr): Likewise.
241 (int_const_binop_1): Likewise.
242 (const_binop): Likewise.
243 (fold_convert_const_int_from_int): Likewise.
244 (fold_convert_const_int_from_real): Likewise.
245 (fold_convert_const_int_from_fixed): Likewise.
246 (fold_convert_const_fixed_from_int): Likewise.
247 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
248 (sign_bit_p): Use wide-int interfaces.
249 (make_range_step): Likewise.
250 (build_range_check): Likewise. Pass an integer of the correct type
251 instead of using integer_one_node.
252 (range_predecessor): Pass an integer of the correct type instead
253 of using integer_one_node.
254 (range_successor): Likewise.
255 (merge_ranges): Likewise.
256 (unextend): Use wide-int interfaces.
257 (extract_muldiv_1): Likewise.
258 (fold_div_compare): Likewise.
259 (fold_single_bit_test): Likewise.
260 (fold_sign_changed_comparison): Likewise.
261 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
262 (fold_plusminus_mult_expr): Use wide-int interfaces.
263 (native_encode_int): Likewise.
264 (native_interpret_int): Likewise.
265 (fold_unary_loc): Likewise.
266 (pointer_may_wrap_p): Likewise.
267 (size_low_cst): Likewise.
268 (mask_with_tz): Likewise.
269 (fold_binary_loc): Likewise.
270 (fold_ternary_loc): Likewise.
271 (multiple_of_p): Likewise.
272 (tree_call_nonnegative_warnv_p): Update calls to
273 tree_int_cst_min_precision and real_from_integer.
274 (fold_negate_const): Use wide-int interfaces.
275 (fold_abs_const): Likewise.
276 (fold_relational_const): Use tree_int_cst_lt.
277 (round_up_loc): Use wide-int interfaces.
278 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
279 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
280 * gengtype.c: Remove include of double-int.h.
281 (do_typedef): Use wide-int interfaces.
282 (open_base_files): Add wide-int.h.
283 (main): Add offset_int and widest_int typedefs.
284 * gengtype-lex.l: Handle "^".
285 (CXX_KEYWORD): Add "static".
286 * gengtype-parse.c (require3): New.
287 (require_template_declaration): Handle constant template arguments
288 and nested templates.
289 * gengtype-state.c: Don't include "double-int.h".
290 * genpreds.c (write_one_predicate_function): Update comment.
291 (write_tm_constrs_h): Add check for hval and lval use in
293 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
294 (add_to_sequence): Likewise.
295 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
296 and const_double_operand.
297 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
299 * gimple-fold.c (get_base_constructor): Likewise.
300 (fold_array_ctor_reference): Likewise.
301 (fold_nonarray_ctor_reference): Likewise.
302 (fold_const_aggregate_ref_1): Likewise.
303 (gimple_val_nonnegative_real_p): Likewise.
304 (gimple_fold_indirect_ref): Likewise.
305 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
306 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
307 (struct slsr_cand_d): Change index to be widest_int.
308 (struct incr_info_d): Change incr to be widest_int.
309 (alloc_cand_and_find_basis): Use wide-int interfaces.
310 (slsr_process_phi): Likewise.
311 (backtrace_base_for_ref): Likewise. Return a widest_int.
312 (restructure_reference): Take a widest_int instead of a double_int.
313 (slsr_process_ref): Use wide-int interfaces.
314 (create_mul_ssa_cand): Likewise.
315 (create_mul_imm_cand): Likewise.
316 (create_add_ssa_cand): Likewise.
317 (create_add_imm_cand): Take a widest_int instead of a double_int.
318 (slsr_process_add): Use wide-int interfaces.
319 (slsr_process_cast): Likewise.
320 (slsr_process_copy): Likewise.
321 (dump_candidate): Likewise.
322 (dump_incr_vec): Likewise.
323 (replace_ref): Likewise.
324 (cand_increment): Likewise. Return a widest_int.
325 (cand_abs_increment): Likewise.
326 (replace_mult_candidate): Take a widest_int instead of a double_int.
327 (replace_unconditional_candidate): Use wide-int interfaces.
328 (incr_vec_index): Take a widest_int instead of a double_int.
329 (create_add_on_incoming_edge): Likewise.
330 (create_phi_basis): Use wide-int interfaces.
331 (replace_conditional_candidate): Likewise.
332 (record_increment): Take a widest_int instead of a double_int.
333 (record_phi_increments): Use wide-int interfaces.
334 (phi_incr_cost): Take a widest_int instead of a double_int.
335 (lowest_cost_path): Likewise.
336 (total_savings): Likewise.
337 (analyze_increments): Use wide-int interfaces.
338 (ncd_with_phi): Take a widest_int instead of a double_int.
339 (ncd_of_cand_and_phis): Likewise.
340 (nearest_common_dominator_for_cands): Likewise.
341 (insert_initializers): Use wide-int interfaces.
342 (all_phi_incrs_profitable): Likewise.
343 (replace_one_candidate): Likewise.
344 (replace_profitable_candidates): Likewise.
345 * godump.c: Include wide-int-print.h.
346 (go_output_typedef): Use wide-int interfaces.
347 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
348 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
349 (build_loop_iteration_domains): Likewise.
350 * hooks.h: Include wide-int.h rather than double-int.h.
351 (hook_bool_dint_dint_uint_bool_true): Delete.
352 (hook_bool_wint_wint_uint_bool_true): Declare.
353 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
354 (hook_bool_wint_wint_uint_bool_true): New.
355 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
357 (ubsan_expand_si_overflow_mul_check): Likewise.
358 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
359 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
360 (get_ancestor_addr_info): Likewise.
361 (ipa_modify_call_arguments): Likewise.
362 * loop-doloop.c (doloop_modify): Likewise.
363 (doloop_optimize): Likewise.
364 * loop-iv.c (iv_number_of_iterations): Likewise.
365 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
366 (unroll_loop_constant_iterations): Likewise.
367 (decide_unroll_runtime_iterations): Likewise.
368 (unroll_loop_runtime_iterations): Likewise.
369 (decide_peel_simple): Likewise.
370 (decide_unroll_stupid): Likewise.
371 * lto-streamer-in.c (streamer_read_wi): Add.
372 (input_cfg): Use wide-int interfaces.
373 (lto_input_tree_1): Likewise.
374 * lto-streamer-out.c (streamer_write_wi): Add.
375 (hash_tree): Use wide-int interfaces.
376 (output_cfg): Likewise.
377 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
378 (GTFILES): Add wide-int.h and signop.h.
379 (TAGS): Look for .cc files too.
380 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
381 * optabs.c (expand_subword_shift): Likewise.
382 (expand_doubleword_shift): Likewise.
383 (expand_absneg_bit): Likewise.
384 (expand_copysign_absneg): Likewise.
385 (expand_copysign_bit): Likewise.
386 * postreload.c (reload_cse_simplify_set): Likewise.
387 * predict.c (predict_iv_comparison): Likewise.
388 * pretty-print.h: Include wide-int-print.h.
390 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
391 * print-tree.c: Include wide-int-print.h.
392 (print_node_brief): Use wide-int interfaces.
393 (print_node): Likewise.
394 * read-rtl.c (validate_const_wide_int): New.
395 (read_rtx_code): Add CONST_WIDE_INT case.
396 * real.c: Include wide-int.h.
397 (real_to_integer2): Delete.
398 (real_to_integer): New function, returning a wide_int.
399 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
400 (ten_to_ptwo): Update call to real_from_integer.
401 (real_digit): Likewise.
402 * real.h: Include signop.h, wide-int.h and insn-modes.h.
403 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
404 (REAL_VALUE_TO_INT): Delete.
405 (real_to_integer): Declare a wide-int form.
406 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
407 * recog.c (const_int_operand): Improve comment.
408 (const_scalar_int_operand): New.
409 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
410 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
411 (split_double): Likewise.
412 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
413 (rtx_size): Likewise.
414 (rtx_alloc_stat_v): New.
415 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
416 (cwi_output_hex): New.
417 (iterative_hash_rtx): Handle CONST_WIDE_INT.
418 (cwi_check_failed_bounds): New.
419 * rtl.def (CONST_WIDE_INT): New.
420 * rtl.h: Include <utility> and wide-int.h.
421 (struct hwivec_def): New.
422 (CWI_GET_NUM_ELEM): New.
423 (CWI_PUT_NUM_ELEM): New.
424 (struct rtx_def): Add num_elem and hwiv.
425 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
426 (CASE_CONST_UNIQUE): Likewise.
427 (CASE_CONST_ANY): Likewise.
428 (CONST_SCALAR_INT_P): Likewise.
429 (CONST_WIDE_INT_P): New.
432 (cwi_check_failed_bounds): New.
435 (CONST_WIDE_INT_VEC) New.
436 (CONST_WIDE_INT_NUNITS) New.
437 (CONST_WIDE_INT_ELT) New.
438 (rtx_mode_t): New type.
439 (wi::int_traits <rtx_mode_t>): New.
441 (wi::min_value): New.
442 (wi::max_value): New.
444 (const_wide_int_alloc): New.
445 (immed_wide_int_const): New.
446 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
447 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
448 * signop.h: New file.
449 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
450 (simplify_const_unary_operation): Use wide-int interfaces.
451 (simplify_binary_operation_1): Likewise.
452 (simplify_const_binary_operation): Likewise.
453 (simplify_const_relational_operation): Likewise.
454 (simplify_immed_subreg): Likewise.
455 * stmt.c (expand_case): Likewise.
456 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
457 signop rather than a bool.
458 * stor-layout.c (layout_type): Use wide-int interfaces.
459 (initialize_sizetypes): Update calls to
460 set_min_and_max_values_for_integral_type.
461 (set_min_and_max_values_for_integral_type): Take a signop rather
462 than a bool. Use wide-int interfaces.
463 (fixup_signed_type): Update accordingly. Remove
464 HOST_BITS_PER_DOUBLE_INT limit.
465 (fixup_unsigned_type): Likewise.
466 * system.h (STATIC_CONSTANT_P): New.
467 (STATIC_ASSERT): New.
468 * target.def (can_use_doloop_p): Take widest_ints rather than
470 * target.h: Include wide-int.h rather than double-int.h.
471 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
473 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
474 rather than INT_CST_LT_UNSIGNED.
475 (can_use_doloop_if_innermost): Take widest_ints rather than
477 * tree-affine.c: Include wide-int-print.h.
478 (double_int_ext_for_comb): Delete.
479 (wide_int_ext_for_comb): New.
480 (aff_combination_zero): Use wide-int interfaces.
481 (aff_combination_const): Take a widest_int instead of a double_int.
482 (aff_combination_elt): Use wide-int interfaces.
483 (aff_combination_scale): Take a widest_int instead of a double_int.
484 (aff_combination_add_elt): Likewise.
485 (aff_combination_add_cst): Likewise.
486 (aff_combination_add): Use wide-int interfaces.
487 (aff_combination_convert): Likewise.
488 (tree_to_aff_combination): Likewise.
489 (add_elt_to_tree): Take a widest_int instead of a double_int.
490 (aff_combination_to_tree): Use wide-int interfaces.
491 (aff_combination_remove_elt): Likewise.
492 (aff_combination_add_product): Take a widest_int instead of
494 (aff_combination_mult): Use wide-int interfaces.
495 (aff_combination_expand): Likewise.
496 (double_int_constant_multiple_p): Delete.
497 (wide_int_constant_multiple_p): New.
498 (aff_combination_constant_multiple_p): Take a widest_int pointer
499 instead of a double_int pointer.
500 (print_aff): Use wide-int interfaces.
501 (get_inner_reference_aff): Take a widest_int pointer
502 instead of a double_int pointer.
503 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
504 * tree-affine.h: Include wide-int.h.
505 (struct aff_comb_elt): Change type of coef to widest_int.
506 (struct affine_tree_combination): Change type of offset to widest_int.
507 (double_int_ext_for_comb): Delete.
508 (wide_int_ext_for_comb): New.
509 (aff_combination_const): Use widest_int instead of double_int.
510 (aff_combination_scale): Likewise.
511 (aff_combination_add_elt): Likewise.
512 (aff_combination_constant_multiple_p): Likewise.
513 (get_inner_reference_aff): Likewise.
514 (aff_comb_cannot_overlap_p): Likewise.
515 (aff_combination_zero_p): Use wide-int interfaces.
516 * tree.c: Include tree.h.
517 (init_ttree): Use make_int_cst.
518 (tree_code_size): Removed code for INTEGER_CST case.
519 (tree_size): Add INTEGER_CST case.
520 (make_node_stat): Update comment.
521 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
522 (build_int_cst_type): Use wide-int interfaces.
523 (double_int_to_tree): Likewise.
524 (double_int_fits_to_tree_p): Delete.
525 (force_fit_type_double): Delete.
526 (force_fit_type): New.
527 (int_cst_hash_hash): Use wide-int interfaces.
528 (int_cst_hash_eq): Likewise.
529 (build_int_cst_wide): Delete.
530 (wide_int_to_tree): New.
531 (cache_integer_cst): Use wide-int interfaces.
532 (build_low_bits_mask): Likewise.
533 (cst_and_fits_in_hwi): Likewise.
534 (real_value_from_int_cst): Likewise.
535 (make_int_cst_stat): New.
536 (integer_zerop): Use wide_int interfaces.
537 (integer_onep): Likewise.
538 (integer_all_onesp): Likewise.
539 (integer_pow2p): Likewise.
540 (integer_nonzerop): Likewise.
541 (tree_log2): Likewise.
542 (tree_floor_log2): Likewise.
543 (tree_ctz): Likewise.
544 (int_size_in_bytes): Likewise.
545 (mem_ref_offset): Return an offset_int rather than a double_int.
546 (build_type_attribute_qual_variant): Use wide_int interfaces.
547 (type_hash_eq): Likewise
548 (tree_int_cst_equal): Likewise.
549 (tree_int_cst_lt): Delete.
550 (tree_int_cst_compare): Likewise.
551 (tree_fits_shwi_p): Use wide_int interfaces.
552 (tree_fits_uhwi_p): Likewise.
553 (tree_int_cst_sign_bit): Likewise.
554 (tree_int_cst_sgn): Likewise.
555 (tree_int_cst_min_precision): Take a signop rather than a bool.
556 (simple_cst_equal): Use wide_int interfaces.
557 (compare_tree_int): Likewise.
558 (iterative_hash_expr): Likewise.
559 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
561 (get_type_static_bounds): Use wide_int interfaces.
562 (tree_int_cst_elt_check_failed): New.
563 (build_common_tree_nodes): Reordered to set prec before filling in
565 (int_cst_value): Check cst_and_fits_in_hwi.
566 (widest_int_cst_value): Use wide_int interfaces.
567 (upper_bound_in_type): Likewise.
568 (lower_bound_in_type): Likewise.
569 (num_ending_zeros): Likewise.
570 (drop_tree_overflow): Likewise.
571 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
572 (gen_conditions_for_pow_cst_base): Likewise.
573 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
574 (group_case_labels_stmt): Use wide-int interfaces.
575 (verify_gimple_assign_binary): Likewise.
576 (print_loop): Likewise.
577 * tree-chrec.c (tree_fold_binomial): Likewise.
578 * tree-core.h (struct tree_base): Add int_length.
579 (struct tree_int_cst): Change rep of value.
580 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
581 (dr_may_alias_p): Likewise.
582 (max_stmt_executions_tree): Likewise.
583 * tree.def (INTEGER_CST): Update comment.
584 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
585 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
586 * tree-dump.c: Include wide-int.h and wide-int-print.h.
587 (dequeue_and_dump): Use wide-int interfaces.
588 * tree.h: Include wide-int.h.
589 (NULL_TREE): Moved to earlier loc in file.
590 (TREE_INT_CST_ELT_CHECK): New.
591 (tree_int_cst_elt_check_failed): New.
593 (TREE_INT_CST): Delete.
594 (TREE_INT_CST_LOW): Use wide-int interfaces.
595 (TREE_INT_CST_HIGH): Delete.
596 (TREE_INT_CST_NUNITS): New.
597 (TREE_INT_CST_EXT_NUNITS): Likewise.
598 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
599 (TREE_INT_CST_ELT): Likewise.
600 (INT_CST_LT): Delete.
601 (tree_int_cst_elt_check): New (two forms).
602 (type_code_size): Update comment.
603 (make_int_cst_stat, make_int_cst): New.
604 (tree_to_double_int): Delete.
605 (double_int_fits_to_tree_p): Delete.
606 (force_fit_type_double): Delete.
607 (build_int_cstu): Replace with out-of-line function.
608 (build_int_cst_wide): Delete.
609 (tree_int_cst_lt): Define inline.
610 (tree_int_cst_le): New.
611 (tree_int_cst_compare): Define inline.
612 (tree_int_cst_min_precision): Take a signop rather than a bool.
613 (wi::int_traits <const_tree>): New.
614 (wi::int_traits <tree>): New.
615 (wi::extended_tree): New.
616 (wi::int_traits <wi::extended_tree>): New.
617 (wi::to_widest): New.
618 (wi::to_offset): New.
619 (wi::fits_to_tree_p): New.
620 (wi::min_value): New.
621 (wi::max_value): New.
622 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
623 (copy_tree_body_r): Likewise.
624 * tree-object-size.c (compute_object_offset): Likewise.
625 (addr_object_size): Likewise.
626 * tree-predcom.c: Include wide-int-print.h.
627 (struct dref_d): Change type of offset to widest_int.
628 (dump_dref): Call wide-int printer.
629 (aff_combination_dr_offset): Use wide-int interfaces.
630 (determine_offset): Take a widest_int pointer rather than a
632 (split_data_refs_to_components): Use wide-int interfaces.
633 (suitable_component_p): Likewise.
634 (order_drefs): Likewise.
635 (add_ref_to_chain): Likewise.
636 (valid_initializer_p): Likewise.
637 (determine_roots_comp): Likewise.
638 * tree-pretty-print.c: Include wide-int-print.h.
639 (dump_generic_node): Use wide-int interfaces.
640 * tree-sra.c (sra_ipa_modify_expr): Likewise.
641 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
642 (move_fixed_address_to_symbol): Likewise.
643 (move_hint_to_base): Likewise.
644 (move_pointer_to_base): Likewise.
645 (move_variant_to_index): Likewise.
646 (most_expensive_mult_to_index): Likewise.
647 (addr_to_parts): Likewise.
648 (copy_ref_info): Likewise.
649 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
650 (indirect_refs_may_alias_p): Likewise.
651 (stmt_kills_ref_p_1): Likewise.
652 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
653 * tree-ssa-ccp.c: Update comment at top of file. Include
655 (struct prop_value_d): Change type of mask to widest_int.
656 (extend_mask): New function.
657 (dump_lattice_value): Use wide-int interfaces.
658 (get_default_value): Likewise.
659 (set_constant_value): Likewise.
660 (set_value_varying): Likewise.
661 (valid_lattice_transition): Likewise.
662 (set_lattice_value): Likewise.
663 (value_to_double_int): Delete.
664 (value_to_wide_int): New.
665 (get_value_from_alignment): Use wide-int interfaces.
666 (get_value_for_expr): Likewise.
667 (do_dbg_cnt): Likewise.
668 (ccp_finalize): Likewise.
669 (ccp_lattice_meet): Likewise.
670 (bit_value_unop_1): Use widest_ints rather than double_ints.
671 (bit_value_binop_1): Likewise.
672 (bit_value_unop): Use wide-int interfaces.
673 (bit_value_binop): Likewise.
674 (bit_value_assume_aligned): Likewise.
675 (evaluate_stmt): Likewise.
676 (ccp_fold_stmt): Likewise.
677 (visit_cond_stmt): Likewise.
678 (ccp_visit_stmt): Likewise.
679 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
680 (constant_pointer_difference): Likewise.
681 (associate_pointerplus): Likewise.
682 (combine_conversions): Likewise.
683 * tree-ssa-loop.h: Include wide-int.h.
684 (struct tree_niter_desc): Change type of max to widest_int.
685 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
686 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
687 (remove_redundant_iv_tests): Likewise.
688 (canonicalize_loop_induction_variables): Likewise.
689 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
690 (constant_multiple_of): Take a widest_int pointer instead of
691 a double_int pointer.
692 (get_computation_aff): Use wide-int interfaces.
693 (ptr_difference_cost): Likewise.
694 (difference_cost): Likewise.
695 (get_loop_invariant_expr_id): Likewise.
696 (get_computation_cost_at): Likewise.
697 (iv_elimination_compare_lt): Likewise.
698 (may_eliminate_iv): Likewise.
699 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
700 instead of double_int.
701 (max_loop_iterations): Likewise.
702 (max_stmt_executions): Likewise.
703 (estimated_stmt_executions): Likewise.
704 * tree-ssa-loop-niter.c: Include wide-int-print.h.
705 (split_to_var_and_offset): Use wide-int interfaces.
706 (determine_value_range): Likewise.
707 (bound_difference_of_offsetted_base): Likewise.
708 (bounds_add): Take a widest_int instead of a double_int.
709 (number_of_iterations_ne_max): Use wide-int interfaces.
710 (number_of_iterations_ne): Likewise.
711 (number_of_iterations_lt_to_ne): Likewise.
712 (assert_loop_rolls_lt): Likewise.
713 (number_of_iterations_lt): Likewise.
714 (number_of_iterations_le): Likewise.
715 (number_of_iterations_cond): Likewise.
716 (number_of_iterations_exit): Likewise.
717 (finite_loop_p): Likewise.
718 (derive_constant_upper_bound_assign): Likewise.
719 (derive_constant_upper_bound): Return a widest_int.
720 (derive_constant_upper_bound_ops): Likewise.
721 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
722 (record_estimate): Take a widest_int rather than a double_int.
723 (record_nonwrapping_iv): Use wide-int interfaces.
724 (double_int_cmp): Delete.
726 (bound_index): Take a widest_int rather than a double_int.
727 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
728 (maybe_lower_iteration_bound): Likewise.
729 (estimate_numbers_of_iterations_loop): Likewise.
730 (estimated_loop_iterations): Take a widest_int pointer than than
731 a double_int pointer.
732 (estimated_loop_iterations_int): Use wide-int interfaces.
733 (max_loop_iterations): Take a widest_int pointer than than
734 a double_int pointer.
735 (max_loop_iterations_int): Use wide-int interfaces.
736 (max_stmt_executions): Take a widest_int pointer than than
737 a double_int pointer.
738 (estimated_stmt_executions): Likewise.
739 (n_of_executions_at_most): Use wide-int interfaces.
740 (scev_probably_wraps_p): Likewise.
741 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
743 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
745 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
746 double_ints. Adjust for trailing_wide_ints <3> representation.
747 (set_nonzero_bits): Likewise.
748 (get_range_info): Return wide_ints rather than double_ints.
749 Adjust for trailing_wide_ints <3> representation.
750 (get_nonzero_bits): Likewise.
751 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
753 * tree-ssanames.h (struct range_info_def): Replace min, max and
754 nonzero_bits with a trailing_wide_ints <3>.
755 (set_range_info): Use wide_int_refs rather than double_ints.
756 (set_nonzero_bits): Likewise.
757 (get_range_info): Return wide_ints rather than double_ints.
758 (get_nonzero_bits): Likewise.
759 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
760 * tree-ssa-pre.c (phi_translate_1): Likewise.
761 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
762 (acceptable_pow_call): Likewise.
763 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
765 (vn_reference_fold_indirect): Likewise.
766 (vn_reference_maybe_forwprop_address): Likewise.
767 (valueize_refs_1): Likewise.
768 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
769 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
770 tree_int_cst_lt and tree_int_cst_le.
771 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
773 (streamer_alloc_tree): Likewise.
774 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
775 (streamer_write_tree_header): Likewise.
776 (streamer_write_integer_cst): Likewise.
777 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
778 (build_constructors): Likewise.
779 (array_value_type): Likewise.
780 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
781 (vect_check_gather): Likewise.
782 * tree-vect-generic.c (build_replicated_const): Likewise.
783 (expand_vector_divmod): Likewise.
784 * tree-vect-loop.c (vect_transform_loop): Likewise.
785 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
786 (vect_do_peeling_for_alignment): Likewise.
787 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
788 * tree-vrp.c: Include wide-int.h.
789 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
790 (extract_range_from_assert): Use wide-int interfaces.
791 (vrp_int_const_binop): Likewise.
792 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
794 (ranges_from_anti_range): Use wide-int interfaces.
795 (quad_int_cmp): Delete.
796 (quad_int_pair_sort): Likewise.
797 (extract_range_from_binary_expr_1): Use wide-int interfaces.
798 (extract_range_from_unary_expr_1): Likewise.
799 (adjust_range_with_scev): Likewise.
800 (masked_increment): Take and return wide_ints rather than double_ints.
801 (register_edge_assert_for_2): Use wide-int interfaces.
802 (check_array_ref): Likewise.
803 (search_for_addr_array): Likewise.
804 (maybe_set_nonzero_bits): Likewise.
805 (union_ranges): Pass an integer of the correct type instead of
806 using integer_one_node.
807 (intersect_ranges): Likewise.
808 (simplify_truth_ops_using_ranges): Likewise.
809 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
810 (range_fits_type_p): Likewise.
811 (simplify_cond_using_ranges): Likewise. Take a signop rather than
813 (simplify_conversion_using_ranges): Use wide-int interfaces.
814 (simplify_float_conversion_using_ranges): Likewise.
815 (vrp_finalize): Likewise.
816 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
817 (gimple_stringops_transform): Likewise.
818 * varasm.c (decode_addr_const): Likewise.
819 (const_hash_1): Likewise.
820 (const_rtx_hash_1): Likewise
821 (output_constant): Likewise.
822 (array_size_for_constructor): Likewise.
823 (output_constructor_regular_field): Likewise.
824 (output_constructor_bitfield): Likewise.
825 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
826 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
828 * gencheck.c: Define BITS_PER_UNIT.
831 * wide-int-print.cc: New.
832 * wide-int-print.h: New.
834 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
836 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
838 2014-05-06 Richard Biener <rguenther@suse.de>
840 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
841 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
842 (TODO_verify_all): Adjust.
843 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
844 TODO_verify_stmts and TODO_verify_rtl_sharing.
845 * bb-reorder.c: Likewise.
846 * cfgexpand.c: Likewise.
849 * function.c: Likewise.
850 * fwprop.c: Likewise.
852 * gimple-ssa-isolate-paths.c: Likewise.
853 * gimple-ssa-strength-reduction.c: Likewise.
854 * ipa-split.c: Likewise.
855 * loop-init.c: Likewise.
856 * loop-unroll.c: Likewise.
857 * lower-subreg.c: Likewise.
858 * modulo-sched.c: Likewise.
859 * postreload-gcse.c: Likewise.
860 * predict.c: Likewise.
862 * sched-rgn.c: Likewise.
863 * store-motion.c: Likewise.
864 * tracer.c: Likewise.
865 * trans-mem.c: Likewise.
866 * tree-call-cdce.c: Likewise.
867 * tree-cfg.c: Likewise.
868 * tree-cfgcleanup.c: Likewise.
869 * tree-complex.c: Likewise.
870 * tree-eh.c: Likewise.
871 * tree-emutls.c: Likewise.
872 * tree-if-conv.c: Likewise.
873 * tree-into-ssa.c: Likewise.
874 * tree-loop-distribution.c: Likewise.
875 * tree-object-size.c: Likewise.
876 * tree-parloops.c: Likewise.
877 * tree-pass.h: Likewise.
878 * tree-sra.c: Likewise.
879 * tree-ssa-ccp.c: Likewise.
880 * tree-ssa-copy.c: Likewise.
881 * tree-ssa-copyrename.c: Likewise.
882 * tree-ssa-dce.c: Likewise.
883 * tree-ssa-dom.c: Likewise.
884 * tree-ssa-dse.c: Likewise.
885 * tree-ssa-forwprop.c: Likewise.
886 * tree-ssa-ifcombine.c: Likewise.
887 * tree-ssa-loop-ch.c: Likewise.
888 * tree-ssa-loop-ivcanon.c: Likewise.
889 * tree-ssa-loop.c: Likewise.
890 * tree-ssa-math-opts.c: Likewise.
891 * tree-ssa-phiopt.c: Likewise.
892 * tree-ssa-phiprop.c: Likewise.
893 * tree-ssa-pre.c: Likewise.
894 * tree-ssa-reassoc.c: Likewise.
895 * tree-ssa-sink.c: Likewise.
896 * tree-ssa-strlen.c: Likewise.
897 * tree-ssa-tail-merge.c: Likewise.
898 * tree-ssa-uncprop.c: Likewise.
899 * tree-switch-conversion.c: Likewise.
900 * tree-tailcall.c: Likewise.
901 * tree-vect-generic.c: Likewise.
902 * tree-vectorizer.c: Likewise.
903 * tree-vrp.c: Likewise.
905 * var-tracking.c: Likewise.
906 * bt-load.c: Likewise.
907 * cfgcleanup.c: Likewise.
908 * combine-stack-adj.c: Likewise.
909 * combine.c: Likewise.
910 * compare-elim.c: Likewise.
911 * config/epiphany/resolve-sw-modes.c: Likewise.
912 * config/i386/i386.c: Likewise.
913 * config/mips/mips.c: Likewise.
914 * config/s390/s390.c: Likewise.
915 * config/sh/sh_treg_combine.cc: Likewise.
916 * config/sparc/sparc.c: Likewise.
921 * mode-switching.c: Likewise.
922 * passes.c: Likewise.
923 * postreload.c: Likewise.
925 * reg-stack.c: Likewise.
926 * regcprop.c: Likewise.
927 * regrename.c: Likewise.
930 2014-05-06 Richard Biener <rguenther@suse.de>
933 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
934 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
936 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
939 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
941 2014-05-05 Radovan Obradovic <robradovic@mips.com>
942 Tom de Vries <tom@codesourcery.com>
944 * target.def (call_fusage_contains_non_callee_clobbers): New DEFHOOKPOD.
945 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
947 (@node Miscellaneous Register Hooks): New node.
948 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
949 * doc/tm.texi: Regenerate.
951 2014-05-05 Marek Polacek <polacek@redhat.com>
954 * opts.c (common_handle_option): Call error_at instead of warning_at.
956 2014-05-05 Richard Biener <rguenther@suse.de>
958 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
959 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
960 under the TODO_verify_il umbrella.
962 2014-05-05 Richard Biener <rguenther@suse.de>
964 * passes.c (execute_function_todo): Move TODO_verify_flow under
965 the TODO_verify_ul umbrella.
967 2014-05-05 Richard Biener <rguenther@suse.de>
970 * fold-const.c (fold_binary_loc): Consistently avoid
971 canonicalizing X & CST away from a CST that is the mask
974 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
976 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
977 int argument to enum machine_mode.
978 (picochip_class_max_nregs): Ditto.
979 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
980 (picochip_class_max_nregs): Ditto.
982 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
984 * target.def: Add new target hook.
985 * doc/tm.texi: Regenerate.
986 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
987 * targhooks.c (default_keep_leaf_when_profiled): New function.
989 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
990 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
992 2014-05-05 Bin Cheng <bin.cheng@arm.com>
994 PR tree-optimization/60363
995 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
996 (copy_phi_args): New parameters. Call get_value_locus_in_path.
997 (update_destination_phis): New parameter.
998 (create_edge_and_update_destination_phis): Ditto.
999 (ssa_fix_duplicate_block_edges): Pass new arguments.
1000 (thread_single_edge): Ditto.
1002 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
1004 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
1005 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
1006 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
1007 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
1008 Use RS6000_BTM_HARD_FLOAT.
1009 (BU_MISC_2): Likewise.
1010 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
1011 RS6000_BTM_HARD_FLOAT.
1012 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
1014 (rs6000_invalid_builtin): Add hard floating builtin support.
1015 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
1016 hard float builtins.
1017 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
1019 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1021 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
1022 Add missing function* argument.
1024 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1026 * lra-constraints.c (valid_address_p): Move earlier in file.
1027 Add a constraint argument to the address_info version.
1028 (satisfies_memory_constraint_p): New function.
1029 (satisfies_address_constraint_p): Likewise.
1030 (process_alt_operands, curr_insn_transform): Use them.
1031 (process_address): Pass the constraint to valid_address_p when
1032 checking address operands.
1034 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1036 * config/mips/mips.c (mips_isa_rev): New variable.
1037 (mips_set_architecture): Set it.
1038 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1040 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1041 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1042 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1043 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1044 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
1045 conditions in terms of mips_isa_rev.
1046 (mips_isa_rev): Declare.
1048 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1050 * config/sh/sh-mem.cc: Use tabs instead of spaces.
1051 (prob_unlikely, prob_likely): Make variables const.
1053 2014-05-03 Denis Chertykov <chertykov@gmail.com>
1055 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
1057 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1059 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
1061 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1063 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
1064 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
1065 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
1067 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
1068 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
1070 Replace usage of ROUND_REG with sh_round_reg.
1071 Use CEIL instead of ROUND_ADVANCE.
1073 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1076 * config/sh/sh.c: Include stdlib headers before everything else.
1078 2014-05-02 Jakub Jelinek <jakub@redhat.com>
1080 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
1081 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
1082 (gimplify_adjust_omp_clauses): Simd region is never
1083 directly nested in combined parallel. Instead, for linear
1084 with copyin/copyout, if in combined for simd loop, make decl
1085 firstprivate/lastprivate on OMP_FOR.
1086 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
1087 expand_omp_for_static_chunk): When setting endvar, also set
1088 fd->loop.v to the same value.
1090 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1092 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
1094 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
1096 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
1099 2014-05-02 Marek Polacek <polacek@redhat.com>
1101 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
1103 2014-05-02 Kito Cheng <kito@0xlab.org>
1105 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
1106 to a C expression marco.
1107 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
1108 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
1109 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
1110 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
1111 HONOR_REG_ALLOC_ORDER.
1112 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
1114 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1116 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
1118 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1120 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
1122 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
1124 * tree-if-conv.c (is_cond_scalar_reduction): New function.
1125 (convert_scalar_cond_reduction): Likewise.
1126 (predicate_scalar_phi): Add recognition and transformation
1127 of simple conditioanl reduction to be vectorizable.
1129 2014-05-01 Marek Polacek <polacek@redhat.com>
1132 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
1134 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
1136 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
1137 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
1138 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
1139 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
1140 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
1141 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
1142 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
1143 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
1145 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
1147 * config/arc/arc.opt (mlra): Move comment above option name
1148 to avoid mis-parsing as language options.
1150 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1152 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
1153 * config/sol2.h: ... here.
1154 * config/sol2-10.h: Remove.
1156 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
1157 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
1158 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
1159 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
1160 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
1161 * config/sol2.h: ... here.
1162 (SECTION_NAME_FORMAT): Don't redefine.
1163 (STARTFILE_ARCH32_SPEC): Rename to ...
1164 (STARTFILE_ARCH_SPEC): ... this.
1165 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
1166 * config/sparc/sol2.h: ... here.
1167 (SECTION_NAME_FORMAT): Don't undef.
1168 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
1169 (SUBTARGET_EXTRA_SPECS): Remove.
1170 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
1172 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
1173 (MD_STARTFILE_PREFIX): Remove.
1174 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
1175 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
1176 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
1177 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
1178 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
1179 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
1180 * config/i386/sol2.h: ... here.
1181 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
1182 * config/i386/sol2-bi.h: Remove.
1183 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
1184 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
1186 * config/i386/t-sol2-64: Rename to ...
1187 * config/i386/t-sol2: ... this.
1188 * config/sparc/t-sol2-64: Rename to ...
1189 * config/sparc/t-sol2: ... this.
1191 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
1192 sol2_tm_file_head, sol2_tm_file_tail.
1193 Include ${cpu_type}/sol2.h before sol2.h.
1195 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
1196 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
1197 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
1198 Reflect i386/t-sol2-64 renaming.
1199 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
1200 Reflect sparc/t-sol2-64 renaming.
1202 2014-04-30 Richard Biener <rguenther@suse.de>
1204 * passes.c (execute_function_todo): Move TODO_verify_stmts
1205 and TODO_verify_ssa under the TODO_verify_il umbrella.
1206 * tree-ssa.h (verify_ssa): Adjust prototype.
1207 * tree-ssa.c (verify_ssa): Add parameter to tell whether
1208 we should verify SSA operands.
1209 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
1210 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
1211 whether we should verify whether not throwing stmts have EH info.
1212 * graphite-scop-detection.c (create_sese_edges): Adjust.
1213 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
1214 * tree-eh.c (lower_try_finally_switch): Do not add the
1215 default case label twice.
1217 2014-04-30 Marek Polacek <polacek@redhat.com>
1219 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
1220 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
1221 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
1222 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
1224 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
1226 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
1227 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
1228 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
1229 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
1230 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
1231 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
1232 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
1233 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
1235 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1237 * tree-cfg.c (dump_function_to_file): Dump the return type of
1238 functions, in a line to itself before the function body, mimicking
1239 the layout of a C function.
1241 2014-04-29 Jakub Jelinek <jakub@redhat.com>
1243 PR tree-optimization/60971
1244 * tree-tailcall.c (process_assignment): Reject conversions which
1247 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
1249 * calls.c (initialize_argument_information): Always treat
1250 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
1251 (expand_call): Likewise.
1252 (emit_library_call_calue_1): Likewise.
1253 * expr.c (PUSH_ARGS_REVERSED): Do not define.
1254 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
1257 2014-04-29 Nick Clifton <nickc@redhat.com>
1259 * config/msp430/msp430.md (umulsidi): Fix typo.
1260 (mulhisi3): Enable even inside interrupt handlers.
1261 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
1262 bigger return address pushed in large mode.
1264 2014-04-29 Nick Clifton <nickc@redhat.com>
1266 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
1267 (arc_init_reg_tables): Use a machine_mode enum to iterate over
1269 * config/m32r/m32r.c (init_reg_tables): Likewise.
1270 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
1271 enum to hold the modes.
1273 2014-04-29 Richard Biener <rguenther@suse.de>
1275 * dominance.c (free_dominance_info): Add overload with
1277 (dom_info_state): Likewise.
1278 (dom_info_available_p): Likewise.
1279 * basic-block.h (free_dominance_info, dom_info_state,
1280 dom_info_available_p): Declare overloads.
1281 * passes.c (execute_function_todo): Verify that verifiers
1282 don't change dominator info state. Drop dominator info
1283 for IPA pass invocations.
1284 * cgraph.c (release_function_body): Restore asserts that
1285 dominator information is released.
1287 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
1289 * doc/invoke.texi: Fix typo.
1290 * tree-vrp.c: Fix typos.
1291 * gimple.c (infer_nonnull_range): Reorder operands of an &&
1294 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1296 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
1298 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1300 * config/aarch64/aarch64-builtins.c
1301 (aarch64_types_storestruct_lane_qualifiers): New.
1302 (TYPES_STORESTRUCT_LANE): Likewise.
1303 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
1304 (st3_lane): Likewise.
1305 (st4_lane): Likewise.
1306 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
1307 (vec_store_lanesci_lane<mode>): Likewise.
1308 (vec_store_lanesxi_lane<mode>): Likewise.
1309 (aarch64_st2_lane<VQ:mode>): Likewise.
1310 (aarch64_st3_lane<VQ:mode>): Likewise.
1311 (aarch64_st4_lane<VQ:mode>): Likewise.
1312 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
1313 * config/aarch64/arm_neon.h
1314 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
1315 use new macro arguments.
1316 (__ST3_LANE_FUNC): Likewise.
1317 (__ST4_LANE_FUNC): Likewise.
1318 * config/aarch64/iterators.md (V_TWO_ELEM): New.
1319 (V_THREE_ELEM): Likewise.
1320 (V_FOUR_ELEM): Likewise.
1322 2014-04-28 David Malcolm <dmalcolm@redhat.com>
1324 * doc/gimple.texi: Replace the description of the now-defunct
1325 union gimple_statement_d with a diagram showing the
1326 gimple_statement_base class hierarchy and its relationships to
1327 the GSS_ and GIMPLE_ enums.
1329 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1331 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
1332 * config/aarch64/aarch64.c
1333 (aarch64_cannot_change_mode_class): Weaken conditions.
1334 (aarch64_modes_tieable_p): New.
1335 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
1337 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
1339 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
1340 (loadsync_<mode>): Change mode.
1341 (load_quadpti, store_quadpti): New.
1342 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
1343 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
1345 2014-04-28 Martin Jambor <mjambor@suse.cz>
1347 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
1348 same alias type as the original statement.
1349 (subreplacement_assignment_data): New type.
1350 (handle_unscalarized_data_in_subtree): New type of parameter,
1351 generate new memory accesses with same alias type as the original
1353 (load_assign_lhs_subreplacements): Likewise.
1354 (sra_modify_constructor_assign): Generate new memory accesses with
1355 same alias type as the original statement.
1357 2014-04-28 Richard Biener <rguenther@suse.de>
1359 * tree-pass.h (TODO_verify_il): Define.
1360 (TODO_verify_all): Complete properly.
1361 * passes.c (execute_function_todo): Move existing loop-closed
1362 SSA verification under TODO_verify_il.
1363 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
1364 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
1365 Fix tree sharing issue.
1367 2014-04-28 Richard Biener <rguenther@suse.de>
1370 * builtins.def (DEF_C11_BUILTIN): Add.
1371 (BUILT_IN_ALIGNED_ALLOC): Likewise.
1372 * coretypes.h (enum function_class): Add function_c11_misc.
1373 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1374 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
1375 (call_may_clobber_ref_p_1): Likewise.
1376 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
1377 (mark_all_reaching_defs_necessary_1): Likewise.
1378 (propagate_necessity): Likewise.
1379 (eliminate_unnecessary_stmts): Likewise.
1380 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
1382 2014-04-28 Richard Biener <rguenther@suse.de>
1384 * tree-vrp.c (vrp_var_may_overflow): Remove.
1385 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
1386 with overflow immediately bump to one before that value and
1387 let iteration figure out overflow status.
1389 2014-04-28 Richard Biener <rguenther@suse.de>
1391 * configure.ac: Do valgrind header checks unconditionally.
1392 Add --enable-valgrind-annotations.
1393 * system.h: Guard valgrind header inclusion with
1394 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
1395 * alloc-pool.c (pool_alloc, pool_free): Use
1396 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
1397 to guard possibly dead code.
1398 * config.in: Regenerated.
1399 * configure: Likewise.
1401 2014-04-28 Jeff Law <law@redhat.com>
1403 PR tree-optimization/60902
1404 * tree-ssa-threadedge.c
1405 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
1406 over real defs when invalidating outputs from statements that do not
1407 produce useful outputs for threading.
1409 2014-04-28 Richard Biener <rguenther@suse.de>
1411 PR tree-optimization/60979
1412 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
1413 SCOPs that end in a block with a successor with abnormal
1416 2014-04-28 Richard Biener <rguenther@suse.de>
1418 * tree-pass.h (execute_pass_list): Adjust prototype.
1419 * passes.c (pass_manager::execute_early_local_passes):
1421 (do_per_function): Change callback signature, push all actual
1422 work to the callbals.
1423 (do_per_function_toporder): Likewise.
1424 (execute_function_dump): Adjust.
1425 (execute_function_todo): Likewise.
1426 (clear_last_verified): Likewise.
1427 (verify_curr_properties): Likewise.
1428 (update_properties_after_pass): Likewise.
1429 (execute_pass_list_1): Split out from ...
1430 (execute_pass_list): ... here. Adjust.
1431 (execute_ipa_pass_list): Likewise.
1432 * cgraphunit.c (cgraph_add_new_function): Adjust.
1433 (analyze_function): Likewise.
1434 (expand_function): Likewise.
1435 * cgraph.c (release_function_body): Free dominance info
1436 here instead of asserting it was magically freed elsewhere.
1438 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
1440 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
1441 * configure: Regenerate.
1442 * config/sparc/sparc.opt (muser-mode): New option.
1443 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
1445 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
1446 * doc/invoke.texi (SPARC options): Document -muser-mode.
1448 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
1450 * cselib.c (find_slot_memmode): Delete.
1451 (cselib_hasher): Change compare_type to a struct.
1452 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
1454 (preserve_constants_and_equivs): Adjust for new compare_type.
1455 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
1456 (wrap_constant): Delete.
1457 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
1459 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
1461 * doc/install.texi (Building with profile feedback): Remove
1464 2014-04-26 Tom de Vries <tom@codesourcery.com>
1466 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
1469 2014-04-25 Cary Coutant <ccoutant@google.com>
1472 * dwarf2out.c (should_move_die_to_comdat): A type definition
1473 can contain a subprogram definition, but don't move it to a
1475 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
1476 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
1478 (clone_tree_hash): Rename to...
1479 (clone_tree_partial): ...this; change callers. Copy
1480 DW_TAG_subprogram DIEs as declarations.
1481 (copy_decls_walk): Don't copy children of a declaration into a
1484 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
1487 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
1490 2014-04-25 Jiong Wang <jiong.wang@arm.com>
1492 * config/arm/predicates.md (call_insn_operand): Add long_call check.
1493 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
1495 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
1498 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1500 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
1503 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1505 PR tree-optimization/60930
1506 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
1507 creating a multiply candidate by folding two constant
1508 multiplicands when the result overflows.
1510 2014-04-25 Jakub Jelinek <jakub@redhat.com>
1512 PR tree-optimization/60960
1513 * tree-vect-generic.c (expand_vector_operation): Only call
1514 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
1516 2014-04-25 Tom de Vries <tom@codesourcery.com>
1518 * expr.c (clobber_reg_mode): New function.
1519 * expr.h (clobber_reg): New function.
1521 2014-04-25 Tom de Vries <tom@codesourcery.com>
1523 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
1526 2014-04-25 Radovan Obradovic <robradovic@mips.com>
1527 Tom de Vries <tom@codesourcery.com>
1529 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
1531 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
1532 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
1533 new argument to find_all_hard_reg_sets call.
1535 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1537 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
1538 Use HOST_WIDE_INT_C for mask literal.
1539 (aarch_rev16_shleft_mask_imm_p): Likewise.
1541 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
1544 * config/sparc/sparc.md (ashlsi3_extend): Delete.
1546 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
1548 PR preprocessor/56540
1549 * config/i386/i386-c.c (ix86_target_macros): Define
1550 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
1552 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1554 * configure.ac (tga_func): Remove.
1555 (LIB_TLS_SPEC): Remove.
1556 * configure: Regenerate.
1557 * config.in: Regenerate.
1558 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
1560 2014-04-25 Richard Biener <rguenther@suse.de>
1563 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
1564 call stmt use/clobber sets during stmt walk instead of
1565 walking the possibly incomplete set of caller edges.
1567 2014-04-25 Richard Biener <rguenther@suse.de>
1570 * passes.c (apply_ipa_transforms): Inline into only caller ...
1571 (execute_one_pass): ... here. Properly bring in function
1572 bodies for nodes we want to apply IPA transforms to.
1574 2014-04-24 Cong Hou <congh@google.com>
1576 PR tree-optimization/60896
1577 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
1578 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
1579 (vect_mark_pattern_stmts): Set the def type of all statements in
1580 PATTERN_DEF_SEQ as vect_internal_def.
1582 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
1584 * doc/extend.texi (PowerPC Built-in Functions): Document new
1585 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
1587 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
1589 * config/rs6000/predicates.md (const_0_to_3_operand): New
1590 predicate to match 0..3 integer constants.
1592 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
1593 to support adding miscellaneous builtin functions.
1594 (BU_DFP_MISC_2): Likewise.
1595 (BU_P7_MISC_1): Likewise.
1596 (BU_P7_MISC_2): Likewise.
1597 (BU_P8V_MISC_3): Likewise.
1598 (BU_MISC_1): Likewise.
1599 (BU_MISC_2): Likewise.
1600 (DIVWE): Add extended divide builtin functions.
1603 (DIVWEUO): Likewise.
1607 (DIVDEUO): Likewise.
1608 (DXEX): Add decimal floating-point builtin functions.
1611 (DDEDPDQ): Likewise.
1613 (DENBCDQ): Likewise.
1620 (CDTBCD): Add new BCD builtin functions.
1624 (BCDADD_LT): Likewise.
1625 (BCDADD_EQ): Likewise.
1626 (BCDADD_GT): Likewise.
1627 (BCDADD_OV): Likewise.
1629 (BCDSUB_LT): Likewise.
1630 (BCDSUB_EQ): Likewise.
1631 (BCDSUB_GT): Likewise.
1632 (BCDSUB_OV): Likewise.
1633 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
1634 (UNPACK_TD): Likewise.
1635 (PACK_TF): Likewise.
1636 (UNPACK_TF): Likewise.
1637 (UNPACK_TF_0): Likewise.
1638 (UNPACK_TF_1): Likewise.
1639 (PACK_V1TI): Likewise.
1640 (UNPACK_V1TI): Likewise.
1642 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
1643 support for decimal floating point builtin functions.
1644 (rs6000_expand_ternop_builtin): Add checks for the new builtin
1645 functions that take constant arguments.
1646 (rs6000_invalid_builtin): Add decimal floating point builtin
1648 (rs6000_init_builtins): Setup long double, _Decimal64, and
1649 _Decimal128 types for new builtin functions.
1650 (builtin_function_type): Set the unsigned flags appropriately for
1651 the new builtin functions.
1652 (rs6000_opt_masks): Add support for decimal floating point builtin
1655 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
1656 floating point builtin functions.
1657 (RS6000_BTM_COMMON): Likewise.
1658 (RS6000_BTI_long_double): Likewise.
1659 (RS6000_BTI_dfloat64): Likewise.
1660 (RS6000_BTI_dfloat128): Likewise.
1661 (long_double_type_internal_node): Likewise.
1662 (dfloat64_type_internal_node): Likewise.
1663 (dfloat128_type_internal_node): Likewise.
1665 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
1666 2.07 bcd arithmetic instructions.
1667 (UNSPEC_BCDSUB): Likewise.
1668 (UNSPEC_BCD_OVERFLOW): Likewise.
1669 (UNSPEC_BCD_ADD_SUB): Likewise.
1670 (bcd_add_sub): Likewise.
1671 (BCD_TEST): Likewise.
1672 (bcd<bcd_add_sub>): Likewise.
1673 (bcd<bcd_add_sub>_test): Likewise.
1674 (bcd<bcd_add_sub>_test2): Likewise.
1675 (bcd<bcd_add_sub>_<code>): Likewise.
1676 (peephole2 for combined bcd ops): Likewise.
1678 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
1679 decimal floating point builtin functions.
1680 (UNSPEC_DENBCD): Likewise.
1681 (UNSPEC_DXEX): Likewise.
1682 (UNSPEC_DIEX): Likewise.
1683 (UNSPEC_DSCLI): Likewise.
1684 (UNSPEC_DSCRI): Likewise.
1685 (D64_D128): Likewise.
1686 (dfp_suffix): Likewise.
1687 (dfp_ddedpd_<mode>): Likewise.
1688 (dfp_denbcd_<mode>): Likewise.
1689 (dfp_dxex_<mode>): Likewise.
1690 (dfp_diex_<mode>): Likewise.
1691 (dfp_dscli_<mode>): Likewise.
1692 (dfp_dscri_<mode>): Likewise.
1694 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
1696 (UNSPEC_CDTBCD): Likewise.
1697 (UNSPEC_CBCDTD): Likewise.
1698 (UNSPEC_DIVE): Add support for new extended divide builtin
1700 (UNSPEC_DIVEO): Likewise.
1701 (UNSPEC_DIVEU): Likewise.
1702 (UNSPEC_DIVEUO): Likewise.
1703 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
1704 pack/unpack 128-bit types.
1705 (UNSPEC_PACK_128BIT): Likewise.
1706 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
1707 (udiv<mode>3): Use idiv_ldiv mode attribute.
1708 (div<mode>3): Likewise.
1709 (addg6s): Add new BCD builtin functions.
1712 (UNSPEC_DIV_EXTEND): Add support for new extended divide
1714 (div_extend): Likewise.
1715 (div<div_extend>_<mode>"): Likewise.
1716 (FP128_64): Add support for new builtin functions to pack/unpack
1718 (unpack<mode>): Likewise.
1719 (unpacktf_0): Likewise.
1720 (unpacktf_1): Likewise.
1721 (unpack<mode>_dm): Likewise.
1722 (unpack<mode>_nodm): Likewise.
1723 (pack<mode>): Likewise.
1724 (unpackv1ti): Likewise.
1725 (packv1ti): Likewise.
1727 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
1729 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
1732 2014-04-24 Jakub Jelinek <jakub@redhat.com>
1734 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
1735 * gimplify.c (omp_is_private): Change last argument's type to int.
1736 Only diagnose lastprivate if the simd argument is 1, only diagnose
1737 linear if the simd argument is 2.
1738 (gimplify_omp_for): Adjust omp_is_private callers. When adding
1739 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
1740 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
1741 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
1742 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
1743 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
1744 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
1745 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
1746 * tree-nested.c (convert_nonlocal_omp_clauses,
1747 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
1749 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
1752 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
1755 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
1757 * flag-types.h (enum ivar_visibility): Add.
1759 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
1761 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
1762 function * argument.
1764 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
1766 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
1768 2014-04-24 Radovan Obradovic <robradovic@mips.com>
1769 Tom de Vries <tom@codesourcery.com>
1771 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
1772 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
1774 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
1775 * emit-rtl.c (try_split): Same.
1777 2014-04-24 Radovan Obradovic <robradovic@mips.com>
1778 Tom de Vries <tom@codesourcery.com>
1780 * common.opt (fuse-caller-save): New option.
1782 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
1784 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
1787 2014-04-24 Richard Biener <rguenther@suse.de>
1789 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
1790 during TER and instead use the sepops interface for expanding
1791 non-GIMPLE_SINGLE_RHS.
1793 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1795 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
1796 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
1798 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1800 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
1801 assembler 64-bit option.
1802 * configure: Regenerate.
1804 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1806 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
1807 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
1808 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
1809 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
1810 (TARGET_CRYPTO): Take TARGET_SIMD into account.
1812 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1814 * config/aarch64/aarch64-builtins.c
1815 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
1816 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
1817 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
1818 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
1820 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
1821 (Vrevsuff): New mode attribute.
1823 2014-04-24 Terry Guo <terry.guo@arm.com>
1825 * config/arm/arm.h (machine_function): Define variable
1826 after_arm_reorg here.
1827 * config/arm/arm.c (after_arm_reorg): Remove the definition.
1828 (arm_split_constant): Update the way to access variable
1831 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
1833 2014-04-23 Tom de Vries <tom@codesourcery.com>
1835 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
1837 2014-04-23 David Malcolm <dmalcolm@redhat.com>
1839 * is-a.h: Update comments to reflect the following changes to the
1840 "pointerness" of the API, making the template parameter match the
1841 return type, allowing use of is-a.h with typedefs of pointers.
1842 (is_a_helper::cast): Return a T rather then a pointer to a T, so
1843 that the return type matches the parameter to the is_a_helper.
1845 (dyn_cast): Likewise.
1847 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
1848 pointer from the is-a.h API.
1850 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
1851 (is_a_helper <cgraph_node *>::test): ...this, matching change to
1853 (is_a_helper <varpool_node>::test): Likewise, convert to...
1854 (is_a_helper <varpool_node *>::test): ...this.
1856 (varpool_first_variable): Update for removal of implicit pointer
1857 from the is-a.h API.
1858 (varpool_next_variable): Likewise.
1859 (varpool_first_static_initializer): Likewise.
1860 (varpool_next_static_initializer): Likewise.
1861 (varpool_first_defined_variable): Likewise.
1862 (varpool_next_defined_variable): Likewise.
1863 (cgraph_first_defined_function): Likewise.
1864 (cgraph_next_defined_function): Likewise.
1865 (cgraph_first_function): Likewise.
1866 (cgraph_next_function): Likewise.
1867 (cgraph_first_function_with_gimple_body): Likewise.
1868 (cgraph_next_function_with_gimple_body): Likewise.
1869 (cgraph_alias_target): Likewise.
1870 (varpool_alias_target): Likewise.
1871 (cgraph_function_or_thunk_node): Likewise.
1872 (varpool_variable_node): Likewise.
1873 (symtab_real_symbol_p): Likewise.
1874 * cgraphunit.c (referred_to_p): Likewise.
1875 (analyze_functions): Likewise.
1876 (handle_alias_pairs): Likewise.
1877 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1878 * gimple-ssa.h (gimple_vuse_op): Likewise.
1879 (gimple_vdef_op): Likewise.
1880 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
1881 * gimple.c (gimple_build_asm_1): Likewise.
1882 (gimple_build_try): Likewise.
1883 (gimple_build_resx): Likewise.
1884 (gimple_build_eh_dispatch): Likewise.
1885 (gimple_build_omp_for): Likewise.
1886 (gimple_omp_for_set_clauses): Likewise.
1888 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
1889 (is_a_helper <gimple_statement_asm *>::test): ...this.
1890 (is_a_helper <gimple_statement_bind>::test): Convert to...
1891 (is_a_helper <gimple_statement_bind *>::test): ...this.
1892 (is_a_helper <gimple_statement_call>::test): Convert to...
1893 (is_a_helper <gimple_statement_call *>::test): ...this.
1894 (is_a_helper <gimple_statement_catch>::test): Convert to...
1895 (is_a_helper <gimple_statement_catch *>::test): ...this.
1896 (is_a_helper <gimple_statement_resx>::test): Convert to...
1897 (is_a_helper <gimple_statement_resx *>::test): ...this.
1898 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
1899 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
1900 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
1901 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
1902 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
1903 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
1904 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
1905 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
1906 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
1907 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
1908 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
1909 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
1910 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
1911 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
1912 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
1913 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
1914 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
1915 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
1916 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
1917 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
1918 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
1919 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
1920 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
1921 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
1922 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
1923 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
1924 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
1925 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
1926 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
1927 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
1928 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
1929 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
1930 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
1931 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
1932 (is_a_helper <gimple_statement_phi>::test): Convert to...
1933 (is_a_helper <gimple_statement_phi *>::test): ...this.
1934 (is_a_helper <gimple_statement_transaction>::test): Convert to...
1935 (is_a_helper <gimple_statement_transaction *>::test): ...this.
1936 (is_a_helper <gimple_statement_try>::test): Convert to...
1937 (is_a_helper <gimple_statement_try *>::test): ...this.
1938 (is_a_helper <gimple_statement_wce>::test): Convert to...
1939 (is_a_helper <gimple_statement_wce *>::test): ...this.
1940 (is_a_helper <const gimple_statement_asm>::test): Convert to...
1941 (is_a_helper <const gimple_statement_asm *>::test): ...this.
1942 (is_a_helper <const gimple_statement_bind>::test): Convert to...
1943 (is_a_helper <const gimple_statement_bind *>::test): ...this.
1944 (is_a_helper <const gimple_statement_call>::test): Convert to...
1945 (is_a_helper <const gimple_statement_call *>::test): ...this.
1946 (is_a_helper <const gimple_statement_catch>::test): Convert to...
1947 (is_a_helper <const gimple_statement_catch *>::test): ...this.
1948 (is_a_helper <const gimple_statement_resx>::test): Convert to...
1949 (is_a_helper <const gimple_statement_resx *>::test): ...this.
1950 (is_a_helper <const gimple_statement_eh_dispatch>::test):
1952 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
1954 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
1956 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
1957 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
1959 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
1961 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
1963 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
1965 (is_a_helper <const gimple_statement_omp_return>::test): Convert
1967 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
1968 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
1970 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
1971 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
1973 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
1974 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
1975 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
1976 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
1978 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
1979 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
1981 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
1982 (is_a_helper <const gimple_statement_omp_target>::test): Convert
1984 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
1985 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
1987 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
1988 (is_a_helper <const gimple_statement_omp_single>::test): Convert
1990 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
1991 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
1993 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
1994 (is_a_helper <const gimple_statement_omp_task>::test): Convert
1996 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
1997 (is_a_helper <const gimple_statement_phi>::test): Convert to...
1998 (is_a_helper <const gimple_statement_phi *>::test): ...this.
1999 (is_a_helper <const gimple_statement_transaction>::test): Convert
2001 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
2002 (is_a_helper <const gimple_statement_with_ops>::test): Convert
2004 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
2005 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
2006 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
2007 (is_a_helper <const gimple_statement_with_memory_ops>::test):
2009 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
2011 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
2013 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
2015 (gimple_use_ops): Update for removal of implicit pointer from the
2017 (gimple_set_use_ops): Likewise.
2018 (gimple_vuse): Likewise.
2019 (gimple_vdef): Likewise.
2020 (gimple_vuse_ptr): Likewise.
2021 (gimple_vdef_ptr): Likewise.
2022 (gimple_set_vuse): Likewise.
2023 (gimple_set_vdef): Likewise.
2024 (gimple_omp_return_set_lhs): Likewise.
2025 (gimple_omp_return_lhs): Likewise.
2026 (gimple_omp_return_lhs_ptr): Likewise.
2027 (gimple_call_fntype): Likewise.
2028 (gimple_call_set_fntype): Likewise.
2029 (gimple_call_set_internal_fn): Likewise.
2030 (gimple_call_use_set): Likewise.
2031 (gimple_call_clobber_set): Likewise.
2032 (gimple_bind_vars): Likewise.
2033 (gimple_bind_set_vars): Likewise.
2034 (gimple_bind_body_ptr): Likewise.
2035 (gimple_bind_set_body): Likewise.
2036 (gimple_bind_add_stmt): Likewise.
2037 (gimple_bind_block): Likewise.
2038 (gimple_bind_set_block): Likewise.
2039 (gimple_asm_ninputs): Likewise.
2040 (gimple_asm_noutputs): Likewise.
2041 (gimple_asm_nclobbers): Likewise.
2042 (gimple_asm_nlabels): Likewise.
2043 (gimple_asm_input_op): Likewise.
2044 (gimple_asm_input_op_ptr): Likewise.
2045 (gimple_asm_output_op): Likewise.
2046 (gimple_asm_output_op_ptr): Likewise.
2047 (gimple_asm_set_output_op): Likewise.
2048 (gimple_asm_clobber_op): Likewise.
2049 (gimple_asm_set_clobber_op): Likewise.
2050 (gimple_asm_label_op): Likewise.
2051 (gimple_asm_set_label_op): Likewise.
2052 (gimple_asm_string): Likewise.
2053 (gimple_catch_types): Likewise.
2054 (gimple_catch_types_ptr): Likewise.
2055 (gimple_catch_handler_ptr): Likewise.
2056 (gimple_catch_set_types): Likewise.
2057 (gimple_catch_set_handler): Likewise.
2058 (gimple_eh_filter_types): Likewise.
2059 (gimple_eh_filter_types_ptr): Likewise.
2060 (gimple_eh_filter_failure_ptr): Likewise.
2061 (gimple_eh_filter_set_types): Likewise.
2062 (gimple_eh_filter_set_failure): Likewise.
2063 (gimple_eh_must_not_throw_fndecl): Likewise.
2064 (gimple_eh_must_not_throw_set_fndecl): Likewise.
2065 (gimple_eh_else_n_body_ptr): Likewise.
2066 (gimple_eh_else_e_body_ptr): Likewise.
2067 (gimple_eh_else_set_n_body): Likewise.
2068 (gimple_eh_else_set_e_body): Likewise.
2069 (gimple_try_eval_ptr): Likewise.
2070 (gimple_try_cleanup_ptr): Likewise.
2071 (gimple_try_set_eval): Likewise.
2072 (gimple_try_set_cleanup): Likewise.
2073 (gimple_wce_cleanup_ptr): Likewise.
2074 (gimple_wce_set_cleanup): Likewise.
2075 (gimple_phi_capacity): Likewise.
2076 (gimple_phi_num_args): Likewise.
2077 (gimple_phi_result): Likewise.
2078 (gimple_phi_result_ptr): Likewise.
2079 (gimple_phi_set_result): Likewise.
2080 (gimple_phi_arg): Likewise.
2081 (gimple_phi_set_arg): Likewise.
2082 (gimple_resx_region): Likewise.
2083 (gimple_resx_set_region): Likewise.
2084 (gimple_eh_dispatch_region): Likewise.
2085 (gimple_eh_dispatch_set_region): Likewise.
2086 (gimple_omp_critical_name): Likewise.
2087 (gimple_omp_critical_name_ptr): Likewise.
2088 (gimple_omp_critical_set_name): Likewise.
2089 (gimple_omp_for_clauses): Likewise.
2090 (gimple_omp_for_clauses_ptr): Likewise.
2091 (gimple_omp_for_set_clauses): Likewise.
2092 (gimple_omp_for_collapse): Likewise.
2093 (gimple_omp_for_index): Likewise.
2094 (gimple_omp_for_index_ptr): Likewise.
2095 (gimple_omp_for_set_index): Likewise.
2096 (gimple_omp_for_initial): Likewise.
2097 (gimple_omp_for_initial_ptr): Likewise.
2098 (gimple_omp_for_set_initial): Likewise.
2099 (gimple_omp_for_final): Likewise.
2100 (gimple_omp_for_final_ptr): Likewise.
2101 (gimple_omp_for_set_final): Likewise.
2102 (gimple_omp_for_incr): Likewise.
2103 (gimple_omp_for_incr_ptr): Likewise.
2104 (gimple_omp_for_set_incr): Likewise.
2105 (gimple_omp_for_pre_body_ptr): Likewise.
2106 (gimple_omp_for_set_pre_body): Likewise.
2107 (gimple_omp_parallel_clauses): Likewise.
2108 (gimple_omp_parallel_clauses_ptr): Likewise.
2109 (gimple_omp_parallel_set_clauses): Likewise.
2110 (gimple_omp_parallel_child_fn): Likewise.
2111 (gimple_omp_parallel_child_fn_ptr): Likewise.
2112 (gimple_omp_parallel_set_child_fn): Likewise.
2113 (gimple_omp_parallel_data_arg): Likewise.
2114 (gimple_omp_parallel_data_arg_ptr): Likewise.
2115 (gimple_omp_parallel_set_data_arg): Likewise.
2116 (gimple_omp_task_clauses): Likewise.
2117 (gimple_omp_task_clauses_ptr): Likewise.
2118 (gimple_omp_task_set_clauses): Likewise.
2119 (gimple_omp_task_child_fn): Likewise.
2120 (gimple_omp_task_child_fn_ptr): Likewise.
2121 (gimple_omp_task_set_child_fn): Likewise.
2122 (gimple_omp_task_data_arg): Likewise.
2123 (gimple_omp_task_data_arg_ptr): Likewise.
2124 (gimple_omp_task_set_data_arg): Likewise.
2125 (gimple_omp_taskreg_clauses): Likewise.
2126 (gimple_omp_taskreg_clauses_ptr): Likewise.
2127 (gimple_omp_taskreg_set_clauses): Likewise.
2128 (gimple_omp_taskreg_child_fn): Likewise.
2129 (gimple_omp_taskreg_child_fn_ptr): Likewise.
2130 (gimple_omp_taskreg_set_child_fn): Likewise.
2131 (gimple_omp_taskreg_data_arg): Likewise.
2132 (gimple_omp_taskreg_data_arg_ptr): Likewise.
2133 (gimple_omp_taskreg_set_data_arg): Likewise.
2134 (gimple_omp_task_copy_fn): Likewise.
2135 (gimple_omp_task_copy_fn_ptr): Likewise.
2136 (gimple_omp_task_set_copy_fn): Likewise.
2137 (gimple_omp_task_arg_size): Likewise.
2138 (gimple_omp_task_arg_size_ptr): Likewise.
2139 (gimple_omp_task_set_arg_size): Likewise.
2140 (gimple_omp_task_arg_align): Likewise.
2141 (gimple_omp_task_arg_align_ptr): Likewise.
2142 (gimple_omp_task_set_arg_align): Likewise.
2143 (gimple_omp_single_clauses): Likewise.
2144 (gimple_omp_single_clauses_ptr): Likewise.
2145 (gimple_omp_single_set_clauses): Likewise.
2146 (gimple_omp_target_clauses): Likewise.
2147 (gimple_omp_target_clauses_ptr): Likewise.
2148 (gimple_omp_target_set_clauses): Likewise.
2149 (gimple_omp_target_child_fn): Likewise.
2150 (gimple_omp_target_child_fn_ptr): Likewise.
2151 (gimple_omp_target_set_child_fn): Likewise.
2152 (gimple_omp_target_data_arg): Likewise.
2153 (gimple_omp_target_data_arg_ptr): Likewise.
2154 (gimple_omp_target_set_data_arg): Likewise.
2155 (gimple_omp_teams_clauses): Likewise.
2156 (gimple_omp_teams_clauses_ptr): Likewise.
2157 (gimple_omp_teams_set_clauses): Likewise.
2158 (gimple_omp_sections_clauses): Likewise.
2159 (gimple_omp_sections_clauses_ptr): Likewise.
2160 (gimple_omp_sections_set_clauses): Likewise.
2161 (gimple_omp_sections_control): Likewise.
2162 (gimple_omp_sections_control_ptr): Likewise.
2163 (gimple_omp_sections_set_control): Likewise.
2164 (gimple_omp_for_set_cond): Likewise.
2165 (gimple_omp_for_cond): Likewise.
2166 (gimple_omp_atomic_store_set_val): Likewise.
2167 (gimple_omp_atomic_store_val): Likewise.
2168 (gimple_omp_atomic_store_val_ptr): Likewise.
2169 (gimple_omp_atomic_load_set_lhs): Likewise.
2170 (gimple_omp_atomic_load_lhs): Likewise.
2171 (gimple_omp_atomic_load_lhs_ptr): Likewise.
2172 (gimple_omp_atomic_load_set_rhs): Likewise.
2173 (gimple_omp_atomic_load_rhs): Likewise.
2174 (gimple_omp_atomic_load_rhs_ptr): Likewise.
2175 (gimple_omp_continue_control_def): Likewise.
2176 (gimple_omp_continue_control_def_ptr): Likewise.
2177 (gimple_omp_continue_set_control_def): Likewise.
2178 (gimple_omp_continue_control_use): Likewise.
2179 (gimple_omp_continue_control_use_ptr): Likewise.
2180 (gimple_omp_continue_set_control_use): Likewise.
2181 (gimple_transaction_body_ptr): Likewise.
2182 (gimple_transaction_label): Likewise.
2183 (gimple_transaction_label_ptr): Likewise.
2184 (gimple_transaction_set_body): Likewise.
2185 (gimple_transaction_set_label): Likewise.
2187 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
2188 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2189 * ipa-ref.c (ipa_record_reference): Likewise.
2190 * ipa-reference.c (analyze_function): Likewise.
2191 (ipa_reference_write_optimization_summary): Likewise.
2192 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
2193 (address_taken_from_non_vtable_p): Likewise.
2194 (comdat_can_be_unshared_p_1): Likewise.
2195 * lto-cgraph.c (lto_output_ref): Likewise.
2196 (add_references): Likewise.
2197 (compute_ltrans_boundary): Likewise.
2198 (output_symtab): Likewise.
2199 (input_ref): Likewise.
2200 (input_cgraph_1): Likewise.
2201 (output_cgraph_opt_summary): Likewise.
2202 * lto-streamer-out.c (lto_output): Likewise.
2203 (output_symbol_p): Likewise.
2204 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
2205 (lsei_start_function_in_partition): Likewise.
2206 (lsei_next_variable_in_partition): Likewise.
2207 (lsei_start_variable_in_partition): Likewise.
2208 * symtab.c (insert_to_assembler_name_hash): Likewise.
2209 (unlink_from_assembler_name_hash): Likewise.
2210 (symtab_unregister_node): Likewise.
2211 (symtab_remove_node): Likewise.
2212 (dump_symtab_node): Likewise.
2213 (verify_symtab_base): Likewise.
2214 (verify_symtab_node): Likewise.
2215 (symtab_make_decl_local): Likewise.
2216 (symtab_alias_ultimate_target): Likewise.
2217 (symtab_resolve_alias): Likewise.
2218 (symtab_get_symbol_partitioning_class): Likewise.
2219 * tree-phinodes.c (allocate_phi_node): Likewise.
2220 (reserve_phi_args_for_new_edge): Likewise.
2221 (remove_phi_args): Likewise.
2222 * varpool.c (varpool_node_for_asm): Likewise.
2223 (varpool_remove_unreferenced_decls): Likewise.
2225 2014-04-23 Jeff Law <law@redhat.com>
2227 PR tree-optimization/60902
2228 * tree-ssa-threadedge.c
2229 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
2230 invalidate outputs from statements that do not produce useful
2231 outputs for threading.
2233 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2235 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
2236 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
2237 machine descriptions for Stack Smashing Protector.
2239 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
2241 * aarch64.md (<optab>_rol<mode>3): New pattern.
2242 (<optab>_rolsi3_uxtw): Likewise.
2243 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
2245 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
2247 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
2248 (arm_cortex_a12_tune): Likewise.
2250 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2252 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
2254 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2256 * config/arm/arm.md (arm_rev16si2): New pattern.
2257 (arm_rev16si2_alt): Likewise.
2258 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
2260 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2262 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
2263 (rev16<mode>2_alt): Likewise.
2264 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
2265 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
2266 (aarch_rev16_shleft_mask_imm_p): Likewise.
2267 (aarch_rev16_p_1): Likewise.
2268 (aarch_rev16_p): Likewise.
2269 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
2270 (aarch_rev16_shright_mask_imm_p): Likewise.
2271 (aarch_rev16_shleft_mask_imm_p): Likewise.
2273 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2275 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
2276 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
2278 (cortex_a53_extra_costs): Likewise.
2279 (cortex_a57_extra_costs): Likewise.
2280 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
2281 (cortexa7_extra_costs): Likewise.
2282 (cortexa8_extra_costs): Likewise.
2283 (cortexa12_extra_costs): Likewise.
2284 (cortexa15_extra_costs): Likewise.
2285 (v7m_extra_costs): Likewise.
2286 (arm_new_rtx_costs): Handle BSWAP.
2288 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2290 * config/arm/arm.c (cortexa8_extra_costs): New table.
2291 (arm_cortex_a8_tune): New tuning struct.
2292 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
2294 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2296 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
2298 2014-04-23 Richard Biener <rguenther@suse.de>
2300 * Makefile.in (OBJS): Remove loop-unswitch.o.
2301 * tree-pass.h (make_pass_rtl_unswitch): Remove.
2302 * passes.def (pass_rtl_unswitch): Likewise.
2303 * loop-init.c (gate_rtl_unswitch): Likewise.
2304 (rtl_unswitch): Likewise.
2305 (pass_data_rtl_unswitch): Likewise.
2306 (pass_rtl_unswitch): Likewise.
2307 (make_pass_rtl_unswitch): Likewise.
2308 * rtl.h (reversed_condition): Likewise.
2309 (compare_and_jump_seq): Likewise.
2310 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
2312 * loop-unroll.c (compare_and_jump_seq): Likewise.
2314 2014-04-23 Richard Biener <rguenther@suse.de>
2316 PR tree-optimization/60903
2317 * tree-ssa-loop-im.c (analyze_memory_references): Remove
2318 commented code block.
2319 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
2320 loop flags to newly created BBs and edges.
2322 2014-04-23 Nick Clifton <nickc@redhat.com>
2324 * config/msp430/msp430.c (msp430_handle_option): Move function
2326 (msp430_option_override): Simplify mcu and mcpu option handling.
2327 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
2328 support for -mhwmult command line option.
2329 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
2330 -mhwmult command line option.
2331 (msp430_hwmult_enabled): Delete.
2332 (msp43o_output_labelref): Add support for -mhwmult command line
2334 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
2335 (umulsidi3): Likewise.
2336 * config/msp430/msp430.opt (mmcu): Add Report attribute.
2337 (mcpu, mlarge, msmall): Likewise.
2338 (mhwmult): New option.
2339 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
2341 (msp430_is_f5_mcu): Remove prototype.
2342 (msp430_use_f5_series_hwmult): Add prototype.
2343 * config/msp430/msp430-opts.h: New file.
2344 * common/config/msp430: New directory.
2345 * common/config/msp430/msp430-common.c: New file.
2346 * config.gcc (msp430): Remove target_has_targetm_common.
2347 * doc/invoke.texi: Document -mhwmult command line option.
2349 2014-04-23 Nick Clifton <nickc@redhat.com>
2351 * config/i386/cygwin.h (ENDFILE_SPEC): Include
2352 default-manifest.o if it can be found in the search path.
2353 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
2355 2014-04-23 Terry Guo <terry.guo@arm.com>
2357 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
2359 2014-04-23 Richard Biener <rguenther@suse.de>
2362 * tree-inline.c (declare_return_variable): Use mark_addressable.
2364 2014-04-23 Richard Biener <rguenther@suse.de>
2367 * loop-init.c (loop_optimizer_init): Make sure to apply
2368 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
2370 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2373 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
2375 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
2376 if flag_sanitize_undefined_trap_on_error.
2377 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
2378 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
2379 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
2380 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
2381 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
2382 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
2383 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
2384 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
2385 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
2386 * ubsan.c (ubsan_instrument_unreachable): Return
2387 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
2388 (ubsan_expand_null_ifn): Emit __builtin_trap ()
2389 if flag_sanitize_undefined_trap_on_error and
2390 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
2391 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
2392 instrument_bool_enum_load): Emit __builtin_trap () if
2393 flag_sanitize_undefined_trap_on_error and
2394 __builtin_handle_*_abort () if !flag_sanitize_recover.
2395 * doc/invoke.texi (-fsanitize-recover,
2396 -fsanitize-undefined-trap-on-error): Document.
2398 2014-04-22 Christian Bruel <christian.bruel@st.com>
2400 * config/sh/sh.md (mov<mode>): Replace movQIHI.
2401 Force immediates to SImode.
2403 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
2405 * config/nios2/nios2.md (UNSPEC_ROUND): New.
2407 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
2408 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
2409 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
2410 (nios2_fpu_insn): Add entry for round.
2411 (N2FPU_NO_ERRNO_P): Define.
2412 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
2414 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
2416 2014-04-22 Richard Henderson <rth@redhat.com>
2418 * config/aarch64/aarch64 (addti3, subti3): New expanders.
2419 (add<GPI>3_compare0): Remove leading * from name.
2420 (add<GPI>3_carryin): Likewise.
2421 (sub<GPI>3_compare0): Likewise.
2422 (sub<GPI>3_carryin): Likewise.
2423 (<su_optab>mulditi3): New expander.
2424 (multi3): New expander.
2425 (madd<GPI>): Remove leading * from name.
2427 2014-04-22 Martin Jambor <mjambor@suse.cz>
2429 * cgraphclones.c (cgraph_function_versioning): Copy
2430 ipa_transforms_to_apply instead of asserting it is empty.
2432 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
2435 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
2436 on count_exp to get mode.
2438 2014-04-22 Andrew Pinski <apinski@cavium.com>
2440 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
2441 Handle TLS for ILP32.
2442 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
2443 (tlsie_small_<mode>): this and handle PTR.
2444 (tlsie_small_sidi): New pattern.
2445 (tlsle_small): Change to an expand to handle ILP32.
2446 (tlsle_small_<mode>): New pattern.
2447 (tlsdesc_small): Rename to ...
2448 (tlsdesc_small_<mode>): this and handle PTR.
2450 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2452 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
2454 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2456 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2457 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
2458 (aarch64_types_signed_poly_qualifiers): Likewise.
2459 (aarch64_types_unsigned_signed_qualifiers): Likewise.
2460 (aarch64_types_poly_signed_qualifiers): Likewise.
2461 (TYPES_REINTERP_SS): Type macro added.
2462 (TYPES_REINTERP_SU): Likewise.
2463 (TYPES_REINTERP_SP): Likewise.
2464 (TYPES_REINTERP_US): Likewise.
2465 (TYPES_REINTERP_PS): Likewise.
2466 (aarch64_fold_builtin): New expression folding added.
2467 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
2468 Declarations removed.
2469 (REINTERP_SS): Declarations added.
2470 (REINTERP_US): Likewise.
2471 (REINTERP_PS): Likewise.
2472 (REINTERP_SU): Likewise.
2473 (REINTERP_SP): Likewise.
2474 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
2475 (vreinterpretq_p8_f64): Likewise.
2476 (vreinterpret_p16_f64): Likewise.
2477 (vreinterpretq_p16_f64): Likewise.
2478 (vreinterpret_f32_f64): Likewise.
2479 (vreinterpretq_f32_f64): Likewise.
2480 (vreinterpret_f64_f32): Likewise.
2481 (vreinterpret_f64_p8): Likewise.
2482 (vreinterpret_f64_p16): Likewise.
2483 (vreinterpret_f64_s8): Likewise.
2484 (vreinterpret_f64_s16): Likewise.
2485 (vreinterpret_f64_s32): Likewise.
2486 (vreinterpret_f64_s64): Likewise.
2487 (vreinterpret_f64_u8): Likewise.
2488 (vreinterpret_f64_u16): Likewise.
2489 (vreinterpret_f64_u32): Likewise.
2490 (vreinterpret_f64_u64): Likewise.
2491 (vreinterpretq_f64_f32): Likewise.
2492 (vreinterpretq_f64_p8): Likewise.
2493 (vreinterpretq_f64_p16): Likewise.
2494 (vreinterpretq_f64_s8): Likewise.
2495 (vreinterpretq_f64_s16): Likewise.
2496 (vreinterpretq_f64_s32): Likewise.
2497 (vreinterpretq_f64_s64): Likewise.
2498 (vreinterpretq_f64_u8): Likewise.
2499 (vreinterpretq_f64_u16): Likewise.
2500 (vreinterpretq_f64_u32): Likewise.
2501 (vreinterpretq_f64_u64): Likewise.
2502 (vreinterpret_s64_f64): Likewise.
2503 (vreinterpretq_s64_f64): Likewise.
2504 (vreinterpret_u64_f64): Likewise.
2505 (vreinterpretq_u64_f64): Likewise.
2506 (vreinterpret_s8_f64): Likewise.
2507 (vreinterpretq_s8_f64): Likewise.
2508 (vreinterpret_s16_f64): Likewise.
2509 (vreinterpretq_s16_f64): Likewise.
2510 (vreinterpret_s32_f64): Likewise.
2511 (vreinterpretq_s32_f64): Likewise.
2512 (vreinterpret_u8_f64): Likewise.
2513 (vreinterpretq_u8_f64): Likewise.
2514 (vreinterpret_u16_f64): Likewise.
2515 (vreinterpretq_u16_f64): Likewise.
2516 (vreinterpret_u32_f64): Likewise.
2517 (vreinterpretq_u32_f64): Likewise.
2519 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2521 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2522 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
2523 (vreinterpret_p8_s8): Likewise.
2524 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
2525 (vreinterpret_p8_s16): Likewise.
2526 (vreinterpret_p8_s32): Likewise.
2527 (vreinterpret_p8_s64): Likewise.
2528 (vreinterpret_p8_f32): Likewise.
2529 (vreinterpret_p8_u8): Likewise.
2530 (vreinterpret_p8_u16): Likewise.
2531 (vreinterpret_p8_u32): Likewise.
2532 (vreinterpret_p8_u64): Likewise.
2533 (vreinterpret_p8_p16): Likewise.
2534 (vreinterpretq_p8_s8): Likewise.
2535 (vreinterpretq_p8_s16): Likewise.
2536 (vreinterpretq_p8_s32): Likewise.
2537 (vreinterpretq_p8_s64): Likewise.
2538 (vreinterpretq_p8_f32): Likewise.
2539 (vreinterpretq_p8_u8): Likewise.
2540 (vreinterpretq_p8_u16): Likewise.
2541 (vreinterpretq_p8_u32): Likewise.
2542 (vreinterpretq_p8_u64): Likewise.
2543 (vreinterpretq_p8_p16): Likewise.
2544 (vreinterpret_p16_s8): Likewise.
2545 (vreinterpret_p16_s16): Likewise.
2546 (vreinterpret_p16_s32): Likewise.
2547 (vreinterpret_p16_s64): Likewise.
2548 (vreinterpret_p16_f32): Likewise.
2549 (vreinterpret_p16_u8): Likewise.
2550 (vreinterpret_p16_u16): Likewise.
2551 (vreinterpret_p16_u32): Likewise.
2552 (vreinterpret_p16_u64): Likewise.
2553 (vreinterpret_p16_p8): Likewise.
2554 (vreinterpretq_p16_s8): Likewise.
2555 (vreinterpretq_p16_s16): Likewise.
2556 (vreinterpretq_p16_s32): Likewise.
2557 (vreinterpretq_p16_s64): Likewise.
2558 (vreinterpretq_p16_f32): Likewise.
2559 (vreinterpretq_p16_u8): Likewise.
2560 (vreinterpretq_p16_u16): Likewise.
2561 (vreinterpretq_p16_u32): Likewise.
2562 (vreinterpretq_p16_u64): Likewise.
2563 (vreinterpretq_p16_p8): Likewise.
2564 (vreinterpret_f32_s8): Likewise.
2565 (vreinterpret_f32_s16): Likewise.
2566 (vreinterpret_f32_s32): Likewise.
2567 (vreinterpret_f32_s64): Likewise.
2568 (vreinterpret_f32_u8): Likewise.
2569 (vreinterpret_f32_u16): Likewise.
2570 (vreinterpret_f32_u32): Likewise.
2571 (vreinterpret_f32_u64): Likewise.
2572 (vreinterpret_f32_p8): Likewise.
2573 (vreinterpret_f32_p16): Likewise.
2574 (vreinterpretq_f32_s8): Likewise.
2575 (vreinterpretq_f32_s16): Likewise.
2576 (vreinterpretq_f32_s32): Likewise.
2577 (vreinterpretq_f32_s64): Likewise.
2578 (vreinterpretq_f32_u8): Likewise.
2579 (vreinterpretq_f32_u16): Likewise.
2580 (vreinterpretq_f32_u32): Likewise.
2581 (vreinterpretq_f32_u64): Likewise.
2582 (vreinterpretq_f32_p8): Likewise.
2583 (vreinterpretq_f32_p16): Likewise.
2584 (vreinterpret_s64_s8): Likewise.
2585 (vreinterpret_s64_s16): Likewise.
2586 (vreinterpret_s64_s32): Likewise.
2587 (vreinterpret_s64_f32): Likewise.
2588 (vreinterpret_s64_u8): Likewise.
2589 (vreinterpret_s64_u16): Likewise.
2590 (vreinterpret_s64_u32): Likewise.
2591 (vreinterpret_s64_u64): Likewise.
2592 (vreinterpret_s64_p8): Likewise.
2593 (vreinterpret_s64_p16): Likewise.
2594 (vreinterpretq_s64_s8): Likewise.
2595 (vreinterpretq_s64_s16): Likewise.
2596 (vreinterpretq_s64_s32): Likewise.
2597 (vreinterpretq_s64_f32): Likewise.
2598 (vreinterpretq_s64_u8): Likewise.
2599 (vreinterpretq_s64_u16): Likewise.
2600 (vreinterpretq_s64_u32): Likewise.
2601 (vreinterpretq_s64_u64): Likewise.
2602 (vreinterpretq_s64_p8): Likewise.
2603 (vreinterpretq_s64_p16): Likewise.
2604 (vreinterpret_u64_s8): Likewise.
2605 (vreinterpret_u64_s16): Likewise.
2606 (vreinterpret_u64_s32): Likewise.
2607 (vreinterpret_u64_s64): Likewise.
2608 (vreinterpret_u64_f32): Likewise.
2609 (vreinterpret_u64_u8): Likewise.
2610 (vreinterpret_u64_u16): Likewise.
2611 (vreinterpret_u64_u32): Likewise.
2612 (vreinterpret_u64_p8): Likewise.
2613 (vreinterpret_u64_p16): Likewise.
2614 (vreinterpretq_u64_s8): Likewise.
2615 (vreinterpretq_u64_s16): Likewise.
2616 (vreinterpretq_u64_s32): Likewise.
2617 (vreinterpretq_u64_s64): Likewise.
2618 (vreinterpretq_u64_f32): Likewise.
2619 (vreinterpretq_u64_u8): Likewise.
2620 (vreinterpretq_u64_u16): Likewise.
2621 (vreinterpretq_u64_u32): Likewise.
2622 (vreinterpretq_u64_p8): Likewise.
2623 (vreinterpretq_u64_p16): Likewise.
2624 (vreinterpret_s8_s16): Likewise.
2625 (vreinterpret_s8_s32): Likewise.
2626 (vreinterpret_s8_s64): Likewise.
2627 (vreinterpret_s8_f32): Likewise.
2628 (vreinterpret_s8_u8): Likewise.
2629 (vreinterpret_s8_u16): Likewise.
2630 (vreinterpret_s8_u32): Likewise.
2631 (vreinterpret_s8_u64): Likewise.
2632 (vreinterpret_s8_p8): Likewise.
2633 (vreinterpret_s8_p16): Likewise.
2634 (vreinterpretq_s8_s16): Likewise.
2635 (vreinterpretq_s8_s32): Likewise.
2636 (vreinterpretq_s8_s64): Likewise.
2637 (vreinterpretq_s8_f32): Likewise.
2638 (vreinterpretq_s8_u8): Likewise.
2639 (vreinterpretq_s8_u16): Likewise.
2640 (vreinterpretq_s8_u32): Likewise.
2641 (vreinterpretq_s8_u64): Likewise.
2642 (vreinterpretq_s8_p8): Likewise.
2643 (vreinterpretq_s8_p16): Likewise.
2644 (vreinterpret_s16_s8): Likewise.
2645 (vreinterpret_s16_s32): Likewise.
2646 (vreinterpret_s16_s64): Likewise.
2647 (vreinterpret_s16_f32): Likewise.
2648 (vreinterpret_s16_u8): Likewise.
2649 (vreinterpret_s16_u16): Likewise.
2650 (vreinterpret_s16_u32): Likewise.
2651 (vreinterpret_s16_u64): Likewise.
2652 (vreinterpret_s16_p8): Likewise.
2653 (vreinterpret_s16_p16): Likewise.
2654 (vreinterpretq_s16_s8): Likewise.
2655 (vreinterpretq_s16_s32): Likewise.
2656 (vreinterpretq_s16_s64): Likewise.
2657 (vreinterpretq_s16_f32): Likewise.
2658 (vreinterpretq_s16_u8): Likewise.
2659 (vreinterpretq_s16_u16): Likewise.
2660 (vreinterpretq_s16_u32): Likewise.
2661 (vreinterpretq_s16_u64): Likewise.
2662 (vreinterpretq_s16_p8): Likewise.
2663 (vreinterpretq_s16_p16): Likewise.
2664 (vreinterpret_s32_s8): Likewise.
2665 (vreinterpret_s32_s16): Likewise.
2666 (vreinterpret_s32_s64): Likewise.
2667 (vreinterpret_s32_f32): Likewise.
2668 (vreinterpret_s32_u8): Likewise.
2669 (vreinterpret_s32_u16): Likewise.
2670 (vreinterpret_s32_u32): Likewise.
2671 (vreinterpret_s32_u64): Likewise.
2672 (vreinterpret_s32_p8): Likewise.
2673 (vreinterpret_s32_p16): Likewise.
2674 (vreinterpretq_s32_s8): Likewise.
2675 (vreinterpretq_s32_s16): Likewise.
2676 (vreinterpretq_s32_s64): Likewise.
2677 (vreinterpretq_s32_f32): Likewise.
2678 (vreinterpretq_s32_u8): Likewise.
2679 (vreinterpretq_s32_u16): Likewise.
2680 (vreinterpretq_s32_u32): Likewise.
2681 (vreinterpretq_s32_u64): Likewise.
2682 (vreinterpretq_s32_p8): Likewise.
2683 (vreinterpretq_s32_p16): Likewise.
2684 (vreinterpret_u8_s8): Likewise.
2685 (vreinterpret_u8_s16): Likewise.
2686 (vreinterpret_u8_s32): Likewise.
2687 (vreinterpret_u8_s64): Likewise.
2688 (vreinterpret_u8_f32): Likewise.
2689 (vreinterpret_u8_u16): Likewise.
2690 (vreinterpret_u8_u32): Likewise.
2691 (vreinterpret_u8_u64): Likewise.
2692 (vreinterpret_u8_p8): Likewise.
2693 (vreinterpret_u8_p16): Likewise.
2694 (vreinterpretq_u8_s8): Likewise.
2695 (vreinterpretq_u8_s16): Likewise.
2696 (vreinterpretq_u8_s32): Likewise.
2697 (vreinterpretq_u8_s64): Likewise.
2698 (vreinterpretq_u8_f32): Likewise.
2699 (vreinterpretq_u8_u16): Likewise.
2700 (vreinterpretq_u8_u32): Likewise.
2701 (vreinterpretq_u8_u64): Likewise.
2702 (vreinterpretq_u8_p8): Likewise.
2703 (vreinterpretq_u8_p16): Likewise.
2704 (vreinterpret_u16_s8): Likewise.
2705 (vreinterpret_u16_s16): Likewise.
2706 (vreinterpret_u16_s32): Likewise.
2707 (vreinterpret_u16_s64): Likewise.
2708 (vreinterpret_u16_f32): Likewise.
2709 (vreinterpret_u16_u8): Likewise.
2710 (vreinterpret_u16_u32): Likewise.
2711 (vreinterpret_u16_u64): Likewise.
2712 (vreinterpret_u16_p8): Likewise.
2713 (vreinterpret_u16_p16): Likewise.
2714 (vreinterpretq_u16_s8): Likewise.
2715 (vreinterpretq_u16_s16): Likewise.
2716 (vreinterpretq_u16_s32): Likewise.
2717 (vreinterpretq_u16_s64): Likewise.
2718 (vreinterpretq_u16_f32): Likewise.
2719 (vreinterpretq_u16_u8): Likewise.
2720 (vreinterpretq_u16_u32): Likewise.
2721 (vreinterpretq_u16_u64): Likewise.
2722 (vreinterpretq_u16_p8): Likewise.
2723 (vreinterpretq_u16_p16): Likewise.
2724 (vreinterpret_u32_s8): Likewise.
2725 (vreinterpret_u32_s16): Likewise.
2726 (vreinterpret_u32_s32): Likewise.
2727 (vreinterpret_u32_s64): Likewise.
2728 (vreinterpret_u32_f32): Likewise.
2729 (vreinterpret_u32_u8): Likewise.
2730 (vreinterpret_u32_u16): Likewise.
2731 (vreinterpret_u32_u64): Likewise.
2732 (vreinterpret_u32_p8): Likewise.
2733 (vreinterpret_u32_p16): Likewise.
2734 (vreinterpretq_u32_s8): Likewise.
2735 (vreinterpretq_u32_s16): Likewise.
2736 (vreinterpretq_u32_s32): Likewise.
2737 (vreinterpretq_u32_s64): Likewise.
2738 (vreinterpretq_u32_f32): Likewise.
2739 (vreinterpretq_u32_u8): Likewise.
2740 (vreinterpretq_u32_u16): Likewise.
2741 (vreinterpretq_u32_u64): Likewise.
2742 (vreinterpretq_u32_p8): Likewise.
2743 (vreinterpretq_u32_p16): Likewise.
2745 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2747 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
2749 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
2752 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
2753 (vqnegd_s64): Likewise.
2754 (vqabs_s64): Likewise.
2755 (vqabsd_s64): Likewise.
2757 2014-04-22 Richard Henderson <rth@redhat.com>
2759 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
2760 computation to the top of the loop.
2762 2014-04-22 Renlin <renlin.li@arm.com>
2763 Jiong Wang <jiong.wang@arm.com>
2765 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
2766 * config/aarch64/aarch64.c (aarch64_layout_frame)
2767 (aarch64_initial_elimination_offset): Likewise.
2769 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
2771 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
2774 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
2776 * machmode.h (bitwise_mode_for_mode): Declare.
2777 * stor-layout.h (bitwise_type_for_mode): Likewise.
2778 * stor-layout.c (bitwise_mode_for_mode): New function.
2779 (bitwise_type_for_mode): Likewise.
2780 * builtins.c (fold_builtin_memory_op): Use it instead of
2781 int_mode_for_mode and build_nonstandard_integer_type.
2783 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2785 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
2786 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
2787 (*-*-solaris2*): Simplify.
2788 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
2789 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
2790 *-*-solaris2.9* handling.
2792 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
2794 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
2795 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
2797 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
2798 * configure: Regenerate.
2800 * config/i386/sol2-9.h: Remove.
2802 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
2803 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
2804 Remove Solaris 9 references.
2806 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
2808 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
2809 (floatuns<GPI:mode><GPF:mode>2): Remove.
2810 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
2811 and floatuns conversions.
2812 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
2813 and floatuns conversions.
2814 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
2815 (w1,w2): New mode attributes for inequal width conversions.
2817 2014-04-22 Renlin Li <Renlin.Li@arm.com>
2819 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
2820 the output asm format.
2822 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
2824 * config/aarch64/aarch64-simd.md
2825 (aarch64_cm<optab>di): Always split.
2826 (*aarch64_cm<optab>di): New.
2827 (aarch64_cmtstdi): Always split.
2828 (*aarch64_cmtstdi): New.
2830 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2832 PR tree-optimization/60823
2833 * omp-low.c (ipa_simd_modify_function_body): Go through
2834 all SSA_NAMEs and for those refering to vector arguments
2835 which are going to be replaced adjust SSA_NAME_VAR and,
2836 if it is a default definition, change it into a non-default
2837 definition assigned at the beginning of function from new_decl.
2838 (ipa_simd_modify_stmt_ops): Rewritten.
2839 * tree-dfa.c (set_ssa_default_def): When removing default def,
2840 check for NULL loc instead of NULL *loc.
2842 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2844 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
2845 restrictions on core registers for DImode values in Thumb2.
2847 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2849 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
2850 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
2852 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2854 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
2855 (*iordi_notzesidi_di): Likewise.
2856 (*iordi_notsesidi_di): Likewise.
2858 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2860 * config/arm/arm-protos.h (tune_params): New struct members.
2861 * config/arm/arm.c: Initialise tune_params per processor.
2862 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
2863 for speed, based on new tune_params.
2865 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2867 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
2869 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
2871 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
2873 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
2874 * config/aarch64/arm_neon.h (vrnd_f64): Added.
2875 (vrnda_f64): Likewise.
2876 (vrndi_f64): Likewise.
2877 (vrndm_f64): Likewise.
2878 (vrndn_f64): Likewise.
2879 (vrndp_f64): Likewise.
2880 (vrndx_f64): Likewise.
2882 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2884 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
2885 GET_MODE_SIZE argument is enum machine_mode.
2887 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2890 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
2891 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
2893 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
2896 * asan.c (asan_emit_stack_protection): Force the base to align to
2897 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
2898 appropriate bits if STRICT_ALIGNMENT.
2899 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
2901 (expand_used_vars): Leave a space in the stack frame for alignment
2902 if STRICT_ALIGNMENT.
2904 2014-04-21 David Malcolm <dmalcolm@redhat.com>
2906 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
2908 (gimple_store_p): Likewise.
2909 (gimple_assign_load_p): Likewise.
2910 (gimple_assign_cast_p): Likewise.
2911 (gimple_clobber_p): Likewise.
2913 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
2914 rather than a gimple.
2915 (gimple_assign_cast_p): Likewise.
2917 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
2920 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
2921 If mode is DDmode and TARGET_E500_DOUBLE allow move.
2923 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
2924 more debug information for E500 if -mdebug=reg.
2926 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
2929 * config/i386/i386.c (ix86_expand_builtin)
2930 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
2931 register for target RTX.
2932 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
2934 2014-04-18 Cong Hou <congh@google.com>
2936 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
2937 the widen-mult pattern by handling two operands with different sizes,
2938 and operands whose size is smaller than half of the result type.
2940 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
2942 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
2943 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
2944 (do_estimate_edge_time): Compute it.
2945 * ipa-inline.c (want_inline_small_function_p): Bypass
2946 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
2948 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
2950 * ipa-inline.c (spec_rem): New static variable.
2951 (dump_overall_stats): New function.
2952 (dump_inline_stats): New function.
2954 2014-04-18 Richard Henderson <rth@redhat.com>
2956 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
2957 to GET_MODE_SIZE, not a reg_class_t.
2959 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2961 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
2962 (vsx_xxmrglw_<mode>): Likewise.
2964 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2967 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
2968 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
2969 (rs6000_init_hard_regno_mode_ok): Likewise.
2971 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
2973 * ipa-inline.c (inline_small_functions): Account only non-cold
2975 * doc/invoke.texi (inline-unit-growth): Update documentation.
2977 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
2979 * config/rs6000/rs6000.md (addti3, subti3): New.
2981 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
2984 * config/i386/i386.c (ix86_expand_clear): Remove outdated
2985 comment. Check optimize_insn_for_size_p instead of
2986 optimize_insn_for_speed_p.
2988 2014-04-17 Martin Jambor <mjambor@suse.cz>
2990 * gimple-iterator.c (gsi_start_edge): New function.
2991 * gimple-iterator.h (gsi_start_edge): Declare.
2992 * tree-sra.c (single_non_eh_succ): New function.
2993 (disqualify_ops_if_throwing_stmt): Renamed to
2994 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
2995 having one non-EH successor BB.
2996 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
2997 generate loads into replacements.
2998 (sra_modify_assign): Likewise and and also use the simple path for
3000 (sra_modify_function_body): Commit statements on edges.
3002 2014-04-17 Richard Biener <rguenther@suse.de>
3005 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
3006 comparison results and add clarifying comment.
3008 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3010 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
3011 (blank_mode): Initialize it.
3012 (emit_mode_size_inline, emit_mode_nunits_inline,
3013 emit_mode_inner_inline): New functions.
3014 (emit_insn_modes_h): Call them and surround their output with
3015 #if GCC_VERSION >= 4001 ... #endif.
3016 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
3017 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
3018 mode_* arrays if the argument is __builtin_constant_p.
3019 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
3020 is enum machine_mode.
3022 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3024 * passes.c (opt_pass::execute): Adjust.
3025 (pass_manager::execute_pass_mode_switching): Likewise.
3026 (early_local_passes::execute): Likewise.
3027 (execute_one_pass): Pass cfun to the pass's execute method.
3028 * tree-pass.h (opt_pass::execute): Add function * argument.
3029 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3030 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3031 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3032 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3033 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3034 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
3035 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
3036 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3037 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3038 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3039 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3040 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3041 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3042 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3043 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3044 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3045 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3046 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3047 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3048 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3049 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3050 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3051 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3052 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3053 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3054 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3055 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3056 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3057 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3060 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3062 * passes.c (opt_pass::gate): Take function * argument.
3063 (gate_all_early_local_passes): Merge into
3064 (early_local_passes::gate): this.
3065 (gate_all_early_optimizations): Merge into
3066 (all_early_optimizations::gate): this.
3067 (gate_all_optimizations): Mege into
3068 (all_optimizations::gate): this.
3069 (gate_all_optimizations_g): Merge into
3070 (all_optimizations_g::gate): this.
3071 (gate_rest_of_compilation): Mege into
3072 (rest_of_compilation::gate): this.
3073 (gate_postreload): Merge into
3074 (postreload::gate): this.
3075 (dump_one_pass): Pass cfun to the pass's gate method.
3076 (execute_ipa_summary_passes): Likewise.
3077 (execute_one_pass): Likewise.
3078 (ipa_write_summaries_2): Likewise.
3079 (ipa_write_optimization_summaries_1): Likewise.
3080 (ipa_read_summaries_1): Likewise.
3081 (ipa_read_optimization_summaries_1): Likewise.
3082 (execute_ipa_stmt_fixups): Likewise.
3083 * tree-pass.h (opt_pass::gate): Add function * argument.
3084 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
3085 combine-stack-adj.c, combine.c, compare-elim.c,
3086 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3087 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
3088 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
3089 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
3090 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3091 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3092 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3093 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
3094 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
3095 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
3096 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3097 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3098 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
3099 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3100 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3101 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3102 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3103 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3104 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3105 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3106 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3107 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3108 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3109 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
3110 var-tracking.c, vtable-verify.c, web.c: Adjust.
3112 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3114 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
3115 * configure: Regenerate.
3117 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3119 * passes.c (dump_one_pass): don't check pass->has_gate.
3120 (execute_ipa_summary_passes): Likewise.
3121 (execute_one_pass): Likewise.
3122 (ipa_write_summaries_2): Likewise.
3123 (ipa_write_optimization_summaries_1): Likewise.
3124 (ipa_read_optimization_summaries_1): Likewise.
3125 (execute_ipa_stmt_fixups): Likewise.
3126 * tree-pass.h (pass_data::has_gate): Remove.
3127 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3128 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3129 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3130 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3131 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3132 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
3133 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
3134 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
3135 gimple-low.c, gimple-ssa-isolate-paths.c,
3136 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
3137 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3138 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3139 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
3140 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
3141 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
3142 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
3143 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3144 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
3145 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3146 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3147 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3148 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3149 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3150 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3151 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3152 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3153 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3154 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3155 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3156 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3157 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3160 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3162 * pass_manager.h (pass_manager::register_dump_files_1): Remove
3164 * passes.c (pass_manager::register_dump_files_1): Merge into
3165 (pass_manager::register_dump_files): this, and remove its handling of
3166 properties since the pass always has the properties anyway.
3167 (pass_manager::pass_manager): Adjust.
3169 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3171 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
3172 * passes.c (pass_manager::register_dump_files_1): Remove dead code
3173 dealing with properties.
3174 (pass_manager::register_dump_files): Adjust.
3176 2014-03-20 Mark Wielaard <mjw@redhat.com>
3178 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
3179 then represent the bound as normal constant value.
3181 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3184 Forward port from 4.8 branch
3185 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
3187 * config/i386/bmiintrin.h (_blsi_u32): New.
3191 (_blsmsk_u32): Ditto.
3192 (_blsmsk_u64): Ditto.
3193 (_tzcnt_u32): Ditto.
3194 (_tzcnt_u64): Ditto.
3196 2014-04-17 Kito Cheng <kito@0xlab.org>
3198 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
3200 2014-04-17 Richard Biener <rguenther@suse.de>
3203 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
3204 boolean results for comparisons.
3206 2014-04-17 Richard Biener <rguenther@suse.de>
3208 PR tree-optimization/60836
3209 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
3210 initial PHI args to be gimple values.
3212 2014-04-17 Richard Biener <rguenther@suse.de>
3214 PR tree-optimization/60841
3215 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
3216 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
3217 of stmts to SLP build.
3218 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
3219 (vect_analyze_slp): Likewise.
3220 (vect_analyze_slp_instance): Likewise.
3221 (vect_build_slp_tree): Limit overall SLP tree growth.
3222 * tree-vectorizer.h (vect_analyze_data_refs,
3223 vect_analyze_slp): Adjust prototypes.
3225 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3227 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
3230 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3232 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
3233 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
3234 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
3235 for TARGET_SLOW_PSHUFB
3237 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3239 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
3240 * config/i386/i386.c (intel_cost): Ditto.
3242 2014-04-17 Joey Ye <joey.ye@arm.com>
3244 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
3246 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3248 * opts.c (common_handle_option): Disable -fipa-reference coorectly
3251 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3253 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
3254 (type_all_derivations_known_p): New predicate.
3255 (type_all_ctors_visible_p): New predicate.
3256 (type_possibly_instantiated_p): New predicate.
3257 (get_odr_type): Compute all_derivations_known.
3258 (dump_odr_type): Dump the flag.
3259 (maybe_record_type): Cleanup.
3260 (record_target_from_binfo): Add bases_to_consider array;
3261 record bases for types w/o instances and skip CXX destructor.
3262 (possible_polymorphic_call_targets_1): Add bases_to_consider
3263 and consider_construction parameters; check if type may have instance.
3264 (get_polymorphic_call_info): Set maybe_in_construction to true
3265 when we know nothing.
3266 (record_targets_from_bases): Skip CXX destructors; they are
3267 never called for types in construction.
3268 (possible_polymorphic_call_targets): Do not record target when
3269 type may not have instance.
3271 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3274 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
3275 external aliases alive, too.
3277 2014-04-16 Andrew Pinski <apinski@cavium.com>
3279 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
3282 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3284 * final.c (compute_alignments): Do not apply loop alignment to a block
3285 falling through to the exit.
3287 2014-04-16 Catherine Moore <clm@codesourcery.com>
3289 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
3290 Adjust constraints for microMIPS store patterns.
3292 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3294 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
3296 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3298 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
3299 (append_use): Run at -O0.
3300 (append_vdef): Likewise.
3301 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
3302 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
3304 2014-04-16 Jakub Jelinek <jakub@redhat.com>
3306 PR tree-optimization/60844
3307 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
3308 (propagate_op_to_single_use, remove_visited_stmt_chain,
3309 linearize_expr, repropagate_negates, reassociate_bb): Use it
3310 instead of gsi_remove.
3312 2014-04-16 Martin Jambor <mjambor@suse.cz>
3314 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
3315 ipa_transforms_to_apply.
3316 (cgraph_function_versioning): Assert that old_node has empty
3317 ipa_transforms_to_apply.
3318 * trans-mem.c (ipa_tm_create_version): Likewise.
3319 * tree-inline.c (tree_function_versioning): Do not duplicate
3320 ipa_transforms_to_apply.
3322 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3325 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
3327 Pass necessary as flags on 64-bit Solaris/x86.
3328 Use lowercase relocs for x86_64-*-*.
3329 * configure: Regenerate.
3331 2014-04-15 Jan Hubicka <jh@suse.cz>
3333 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
3334 (maybe_record_node, likely_target_p): Use it.
3336 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3339 Revert following patch
3341 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3344 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3345 software floating point or no floating point registers, do not
3346 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3347 in GPRs that occurs after we tested for GPRs that would never be
3350 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3351 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3352 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3353 specifically allow DDmode, since that does not use the SPE SIMD
3356 2014-03-21 Mark Wielaard <mjw@redhat.com>
3358 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
3359 as unsigned or int depending on type and value used.
3361 2014-04-15 Richard Biener <rguenther@suse.de>
3363 PR rtl-optimization/56965
3364 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
3365 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
3367 * alias.c (true_dependence_1): Do not call
3368 nonoverlapping_component_refs_p.
3369 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
3370 nonoverlapping_component_refs_p.
3371 (indirect_refs_may_alias_p): Likewise.
3373 2014-04-15 Teresa Johnson <tejohnson@google.com>
3375 * cfg.c (dump_bb_info): Fix flags check.
3376 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
3378 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3380 PR rtl-optimization/60663
3381 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
3384 2014-04-15 Richard Biener <rguenther@suse.de>
3386 * lto-streamer.h (LTO_major_version): Bump to 4.
3388 2014-04-15 Richard Biener <rguenther@suse.de>
3390 * common.opt (lto_partition_model): New enum.
3391 (flto-partition=): Merge separate options with a single with argument,
3392 add -flto-partition=one support.
3393 * flag-types.h (enum lto_partition_model): Declare.
3394 * opts.c (finish_options): Remove duplicate -flto-partition=
3396 * lto-wrapper.c (run_gcc): Adjust.
3398 2014-04-15 Richard Biener <rguenther@suse.de>
3400 * alias.c (ncr_compar): New function.
3401 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
3403 2014-04-15 Richard Biener <rguenther@suse.de>
3405 * alias.c (record_component_aliases): Do not walk BINFOs.
3407 2014-04-15 Richard Biener <rguenther@suse.de>
3409 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
3410 Add struct function argument and adjust.
3411 (find_func_aliases_for_call): Likewise.
3412 (find_func_aliases): Likewise.
3413 (find_func_clobbers): Likewise.
3414 (intra_create_variable_infos): Likewise.
3415 (compute_points_to_sets): Likewise.
3416 (ipa_pta_execute): Adjust. Do not push/pop cfun.
3418 2014-04-15 Richard Biener <rguenther@suse.de>
3420 * tree.c (iterative_hash_expr): Use enum tree_code_class
3421 to store TREE_CODE_CLASS.
3422 (tree_block): Likewise.
3423 (tree_set_block): Likewise.
3424 * tree.h (fold_build_pointer_plus_loc): Use
3425 convert_to_ptrofftype_loc.
3427 2014-04-15 Jakub Jelinek <jakub@redhat.com>
3430 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
3433 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
3435 * cfgloop.h (struct loop): Move force_vectorize down.
3436 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
3437 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
3438 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
3439 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
3440 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
3441 * tree-core.h (enum annot_expr_kind): Add new kind values.
3442 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
3443 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
3445 * tree.def (ANNOTATE_EXPR): Tweak comment.
3447 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3449 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
3452 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
3454 * tree.h (TYPE_IDENTIFIER): Declare.
3455 * tree.c (subrange_type_for_debug_p): Use it.
3456 * godump.c (go_format_type): Likewise.
3457 * dwarf2out.c (is_cxx_auto, modified_type_die,
3458 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
3459 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
3461 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3464 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
3466 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
3468 * config/i386/i386.c (examine_argument): Return bool. Return true if
3469 parameter should be passed in memory.
3470 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
3471 (construct_container): Update calls to examine_argument.
3472 (function_arg_advance_64): Ditto.
3473 (return_in_memory_32): Merge with ix86_return_in_memory.
3474 (return_in_memory_64): Ditto.
3475 (return_in_memory_ms_64): Ditto.
3477 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3479 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
3480 * coverage.c (coverage_compute_profile_id): Handle externally visible
3483 2014-04-14 Martin Jambor <mjambor@suse.cz>
3485 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
3486 DECL_DISREGARD_INLINE_LIMITS functions.
3488 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3491 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
3493 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3496 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
3497 optimize_insn_for_speed_p instead of
3498 optimize_function_for_speed_p.
3500 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
3502 * doc/invoke.texi (free): Document AArch64.
3504 2014-04-14 Richard Biener <rguenther@suse.de>
3506 PR tree-optimization/60042
3507 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
3508 (insert_into_preds_of_block): Do not prevent PHI insertion
3509 for REFERENCE exprs here ...
3510 (eliminate_dom_walker::before_dom_children): ... but prevent
3511 their use here under similar conditions when applied to the
3512 IL after PRE optimizations.
3514 2014-04-14 Richard Biener <rguenther@suse.de>
3516 * passes.def: Move early points-to after early SRA.
3518 2014-04-14 Richard Biener <rguenther@suse.de>
3520 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
3521 check for which sign-changes we allow when forwarding
3522 a converted value into a switch.
3524 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3526 * stor-layout.c (place_field): Finalize non-constant offset for the
3529 2014-04-14 Richard Biener <rguenther@suse.de>
3531 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
3533 (expand_switch_using_bit_tests_p): Likewise.
3534 (process_switch): Compute and pass on speed_p based on the
3536 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
3537 optimize_bb_for_speed_p.
3539 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3541 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
3542 * function.h (struct function): Rename has_force_vect_loops into
3543 has_force_vectorize_loops.
3544 * lto-streamer-in.c (input_cfg): Adjust for renaming.
3545 (input_struct_function_base): Likewise.
3546 * lto-streamer-out.c (output_cfg): Likewise.
3547 (output_struct_function_base): Likewise.
3548 * omp-low.c (expand_omp_simd): Likewise.
3549 * tree-cfg.c (move_sese_region_to_fn): Likewise.
3550 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
3551 (version_loop_for_if_conversion): Likewise.
3552 (tree_if_conversion): Likewise.
3553 (main_tree_if_conversion): Likewise.
3554 (gate_tree_if_conversion): Likewise.
3555 * tree-inline.c (copy_loops): Likewise.
3556 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
3557 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
3558 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
3559 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
3560 * tree-vectorizer.c (vectorize_loops): Likewise.
3561 * tree-vectorizer.h (unlimited_cost_model): Likewise.
3563 2014-04-14 Richard Biener <rguenther@suse.de>
3566 * lto-streamer-out.c (wrap_refs): New function.
3567 (lto_output): Wrap symbol references in global initializes in
3568 type-preserving MEM_REFs.
3570 2014-04-14 Christian Bruel <christian.bruel@st.com>
3572 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
3574 2014-04-14 Christian Bruel <christian.bruel@st.com>
3576 * config/sh/sh.md (setmemqi): New expand pattern.
3577 * config/sh/sh.h (CLEAR_RATIO): Define.
3578 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
3579 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
3581 2014-04-14 Richard Biener <rguenther@suse.de>
3584 * fold-const.c (negate_expr_p): Don't negate directional rounding
3586 (fold_negate_expr): Likewise.
3588 2014-04-14 Richard Biener <rguenther@suse.de>
3590 PR tree-optimization/59817
3591 PR tree-optimization/60453
3592 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
3593 recursion to catch all CHRECs in the scalar evolution and restrict
3594 the predicate for the remains appropriately.
3596 2014-04-12 Catherine Moore <clm@codesourcery.com>
3598 * config/mips/constraints.md: Add new register constraint "kb".
3599 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
3600 (*movhi_internal): Likewise.
3601 (*movqi_internal): Likewise.
3602 * config/mips/mips.h (M16_STORE_REGS): New register class.
3603 (REG_CLASS_NAMES): Add M16_STORE_REGS.
3604 (REG_CLASS_CONTENTS): Likewise.
3605 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
3607 2014-04-11 Tobias Burnus <burnus@net-b.de>
3610 * doc/invoke.texi (-Wformat-signedness): Document it.
3611 (Wformat=2): Mention that this enables -Wformat-signedness.
3613 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3615 * common/config/epiphany/epiphany-common.c
3616 (epiphany_option_optimization_table): Enable section anchors by
3617 default at -O1 or higher.
3618 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
3619 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
3620 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
3621 carries no extra cost.
3622 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
3623 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
3624 * config/epiphany/predicates.md (memclob_operand): New predicate.
3625 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
3626 Use memclob_operand predicate and X constraint for operand 3.
3628 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3630 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
3631 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
3634 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3636 PR rtl-optimization/60651
3637 * mode-switching.c (optimize_mode_switching): Make sure to emit
3638 sets of a lower numbered entity before sets of a higher numbered
3639 entity to a mode of the same or lower priority.
3640 When creating a seginfo for a basic block that starts with a code
3641 label, move the insertion point past the code label.
3642 (new_seginfo): Document and enforce requirement that
3643 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
3644 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
3645 * doc/tm.texi: Regenerate.
3647 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
3650 * config/arc/arc.c (arc_save_restore): Fix assert typo.
3652 2013-04-11 Jakub Jelinek <jakub@redhat.com>
3654 * BASE-VER: Set to 4.10.0.
3656 2014-04-11 Tobias Burnus <burnus@net-b.de>
3659 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
3660 * doc/gcc.texi (Service): Update description in the @menu
3661 * doc/invoke.texi (Option Summary): Remove misplaced and
3664 2014-04-11 Steve Ellcey <sellcey@mips.com>
3665 Jakub Jelinek <jakub@redhat.com>
3668 * expr.c (convert_move): Use emit_store_flag_force instead of
3669 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
3672 2014-04-11 Richard Biener <rguenther@suse.de>
3675 * varasm.c (assemble_alias): Avoid endless error reporting
3676 recursion by setting TREE_ASM_WRITTEN.
3678 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3680 * config/s390/s390.md: Add a splitter for NOT rtx.
3682 2014-04-11 Jakub Jelinek <jakub@redhat.com>
3684 PR rtl-optimization/60663
3685 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
3687 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
3688 Jakub Jelinek <jakub@redhat.com>
3691 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
3692 flag from decl_node to node.
3694 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3697 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
3698 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
3699 ameliorating the cases where it can be.
3701 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
3704 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
3706 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3707 (loadsync_<mode>): Change mode.
3708 (load_quadpti, store_quadpti): New.
3709 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3710 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3711 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
3713 2014-04-09 Cong Hou <congh@google.com>
3716 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
3719 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3721 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
3722 instead of vnor to exploit possible fusion opportunity in the
3724 (altivec_expand_vec_perm_const_le): Likewise.
3726 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
3728 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3729 (loadsync_<mode>): Change mode.
3730 (load_quadpti, store_quadpti): New.
3731 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3732 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3734 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
3737 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
3738 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
3739 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
3741 2014-04-08 Richard Biener <rguenther@suse.de>
3744 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
3745 a 64bit widest int print double-int similar to on HWI64 hosts.
3747 2014-04-08 Richard Biener <rguenther@suse.de>
3749 PR tree-optimization/60785
3750 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
3751 default defs properly.
3753 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
3755 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
3756 (Weffc++): Likewise.
3758 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
3760 * ipa-devirt.c (maybe_record_node): When node is not recorded,
3761 set completep to false rather than true.
3763 2014-04-07 Douglas B Rupp <rupp@adacore.com>
3766 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
3767 ARM_TARGET2_DWARF_FORMAT.
3769 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
3772 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
3773 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
3776 2014-04-07 Richard Biener <rguenther@suse.de>
3778 PR tree-optimization/60766
3779 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
3780 (may_eliminate_iv): Convert cand_value_at result to desired type.
3782 2014-04-07 Jason Merrill <jason@redhat.com>
3785 * common.opt (-fno-gnu-unique): Add.
3786 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
3788 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3790 * haifa-sched.c: Fix outdated function reference and minor
3791 grammar errors in introductory comment.
3793 2014-04-07 Richard Biener <rguenther@suse.de>
3796 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
3798 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
3800 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
3803 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
3804 size accounting for thunks.
3805 (pa_asm_output_mi_thunk): Use final_start_function() and
3806 final_end_function() to output function start and end directives.
3808 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3810 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
3811 device specific ISA/ feature information. Remove short_sp and
3812 errata_skip ds. Add avr_device_specific_features enum to have device
3814 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
3815 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
3816 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
3817 updated device specific info.
3818 * config/avr/avr-mcus.def: Merge device specific details to
3819 dev_attribute field.
3820 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
3822 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
3823 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
3824 assembler if RMW isa supported by current device.
3825 * config/avr/genmultilib.awk: Update as device info structure changed.
3826 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
3828 2014-04-04 Cong Hou <congh@google.com>
3830 PR tree-optimization/60656
3831 * tree-vect-stmts.c (supportable_widening_operation):
3832 Fix a bug that elements in a vector with vect_used_by_reduction
3833 property are incorrectly reordered when the operation on it is not
3834 consistant with the one in reduction operation.
3836 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
3838 PR rtl-optimization/60155
3839 * gcse.c (record_set_data): New function.
3840 (single_set_gcse): New function.
3841 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
3842 (hoist_code): Likewise.
3843 (get_pressure_class_and_nregs): Likewise.
3845 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
3847 * explow.c (probe_stack_range): Emit a final optimization blockage.
3849 2014-04-04 Anthony Green <green@moxielogic.com>
3851 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
3854 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
3857 * lto-cgraph.c (input_overwrite_node): Check that partitioning
3858 flags are set only during streaming.
3859 * ipa.c (process_references, walk_polymorphic_call_targets,
3860 symtab_remove_unreachable_nodes): Drop bodies of always inline
3861 after early inlining.
3862 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
3864 2014-04-04 Jakub Jelinek <jakub@redhat.com>
3865 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3868 * dwarf2out.c (const_ok_for_output_1): Reject expressions
3871 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3874 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
3876 (cortex_a53_fdivd): Likewise.
3878 2014-04-04 Martin Jambor <mjambor@suse.cz>
3881 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
3883 * cgraph.c (clone_of_p): Also return true if thunks match.
3884 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
3885 cgraph_function_or_thunk_node and an obsolete comment.
3886 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
3888 (build_function_decl_skip_args): Likewise.
3889 (set_new_clone_decl_and_node_flags): New function.
3890 (duplicate_thunk_for_node): Likewise.
3891 (redirect_edge_duplicating_thunks): Likewise.
3892 (cgraph_clone_node): New parameter args_to_skip, pass it to
3893 redirect_edge_duplicating_thunks which is called instead of
3894 cgraph_redirect_edge_callee.
3895 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
3896 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
3898 2014-04-04 Jeff Law <law@redhat.com>
3901 * config/arm/predicates.md (const_int_I_operand): New predicate.
3902 (const_int_M_operand): Similarly.
3903 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
3905 (insv_t2, extv_reg, extzv_t2): Likewise.
3906 (load_multiple_with_writeback): Similarly for const_int_I_operand.
3907 (pop_multiple_with_writeback_and_return): Likewise.
3908 (vfp_pop_multiple_with_writeback): Likewise
3910 2014-04-04 Richard Biener <rguenther@suse.de>
3913 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
3914 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
3916 * gimplify.h (gimple_add_tmp_var_fn): Declare.
3917 * gimplify.c (gimple_add_tmp_var_fn): New function.
3918 * gimple-expr.h (create_tmp_reg_fn): Declare.
3919 * gimple-expr.c (create_tmp_reg_fn): New function.
3920 * gimple-low.c (record_vars_into): Don't change cfun.
3921 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
3922 code generation without cfun.
3924 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
3927 * Makefile.in (install-driver): Fix shell scripting.
3929 2014-04-03 Cong Hou <congh@google.com>
3931 PR tree-optimization/60505
3932 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
3933 threshold of number of iterations below which no vectorization
3935 * tree-vect-loop.c (new_loop_vec_info):
3936 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
3937 * tree-vect-loop.c (vect_analyze_loop_operations):
3938 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
3939 * tree-vect-loop.c (vect_transform_loop):
3940 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
3941 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
3942 of iterations of the loop and see if we should build the epilogue.
3944 2014-04-03 Richard Biener <rguenther@suse.de>
3946 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
3947 (streamer_tree_cache_create): Adjust.
3948 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
3949 to allow optional nodes array.
3950 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
3951 (streamer_tree_cache_append): Likewise.
3952 (streamer_tree_cache_create): Create nodes array optionally
3953 as specified by parameter.
3954 * lto-streamer-out.c (create_output_block): Avoid maintaining
3955 the node array in the writer cache.
3956 (DFS_write_tree): Remove assertion.
3957 (produce_asm_for_decls): Free the out decl state hash table early.
3958 * lto-streamer-in.c (lto_data_in_create): Adjust for
3959 streamer_tree_cache_create prototype change.
3961 2014-04-03 Richard Biener <rguenther@suse.de>
3963 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
3964 set TREE_CHAIN to NULL_TREE.
3966 2014-04-03 Richard Biener <rguenther@suse.de>
3968 PR tree-optimization/60740
3969 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
3970 over all GIMPLE_COND operands.
3972 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
3974 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
3975 (Weffc++): Remove Scott's numbering, merge lists and reference
3978 2014-04-03 Nick Clifton <nickc@redhat.com>
3980 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
3983 2014-04-03 Martin Jambor <mjambor@suse.cz>
3985 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
3986 mention gcc_unreachable before failing.
3987 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
3990 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
3993 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
3994 inconsistent code and instead mark the context inconsistent.
3995 (possible_polymorphic_call_targets): For inconsistent contexts
3996 return empty complete list.
3998 2014-04-02 Anthony Green <green@moxielogic.com>
4000 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
4001 (extendqisi2, extendhisi2): Define.
4002 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
4003 (WCHAR_TYPE): Change to unsigned int.
4005 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4007 PR tree-optimization/60733
4008 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
4009 insertion point for PHI candidates to be the end of the feeding
4010 block for the PHI argument.
4012 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
4014 PR rtl-optimization/60650
4015 * lra-constraints.c (process_alt_operands): Decrease reject for
4016 earlyclobber matching.
4018 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4020 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
4022 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4024 * config/spu/spu.c (pad_bb): Do not crash when the last
4025 insn is CODE_FOR_blockage.
4027 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4029 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
4030 lies outside the target mode.
4032 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4035 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4036 software floating point or no floating point registers, do not
4037 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4038 in GPRs that occurs after we tested for GPRs that would never be
4041 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4042 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4043 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4044 specifically allow DDmode, since that does not use the SPE SIMD
4047 2014-04-02 Richard Biener <rguenther@suse.de>
4050 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4051 MODE_INTs. Properly use negv_optab.
4052 (expand_abs): Likewise.
4054 2014-04-02 Richard Biener <rguenther@suse.de>
4057 * Makefile.in (install-driver): Guard extra installs with special
4060 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4062 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4065 2014-04-01 Richard Henderson <rth@redhat.com>
4068 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
4069 alternative enabled before register allocation.
4071 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
4073 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
4074 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
4076 (nios2_large_got_address): Remove unneeded 'sym' parameter.
4077 (nios2_got_address): Update nios2_large_got_address call site.
4078 (nios2_delegitimize_address): New function.
4079 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
4080 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
4081 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
4083 2014-04-01 Martin Husemann <martin@duskware.de>
4085 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
4088 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
4090 PR rtl-optimization/60604
4091 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
4092 check from register_operand.
4093 (register_operand): Redefine in terms of general_operand.
4094 (nonmemory_operand): Use register_operand for the non-constant cases.
4096 2014-04-01 Richard Biener <rguenther@suse.de>
4098 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
4100 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
4102 * doc/invoke.texi (mapp-regs): Clarify.
4104 2014-03-31 Ulrich Drepper <drepper@gmail.com>
4106 * config/i386/avx512fintrin.h (__v32hi): Define type.
4107 (__v64qi): Likewise.
4108 (_mm512_set1_epi8): Define.
4109 (_mm512_set1_epi16): Define.
4110 (_mm512_set4_epi32): Define.
4111 (_mm512_set4_epi64): Define.
4112 (_mm512_set4_pd): Define.
4113 (_mm512_set4_ps): Define.
4114 (_mm512_setr4_epi64): Define.
4115 (_mm512_setr4_epi32): Define.
4116 (_mm512_setr4_pd): Define.
4117 (_mm512_setr4_ps): Define.
4118 (_mm512_setzero_epi32): Define.
4120 2014-03-31 Martin Jambor <mjambor@suse.cz>
4123 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
4124 callsite_arguments_match_p. Updated all callers. Also check types of
4125 corresponding formal parameters and actual arguments.
4126 (not_all_callers_have_enough_arguments_p) Renamed to
4127 some_callers_have_mismatched_arguments_p.
4129 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
4131 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
4133 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
4136 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
4139 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
4141 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
4142 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
4144 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
4145 Use FMAMODE_NOVF512 mode iterator.
4146 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
4147 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
4148 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
4149 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
4151 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
4152 Use VF_128_256 mode iterator.
4153 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
4156 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4158 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
4159 static chain if needed.
4161 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4164 * lra-constraints.c (index_part_to_reg): New.
4165 (process_address): Use it.
4167 2014-03-27 Jeff Law <law@redhat.com>
4168 Jakub Jelinek <jakub@redhat.com>
4171 * expr.c (do_tablejump): Use simplify_gen_binary rather than
4172 gen_rtx_{PLUS,MULT} to build up the address expression.
4174 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
4175 creating non-canonical RTL.
4177 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4180 * ipa-inline.c (want_inline_small_function_p): Short circuit large
4181 functions; reorganize to make cheap checks first.
4182 (inline_small_functions): Do not estimate growth when dumping;
4184 * ipa-inline.h (inline_summary): Add min_size.
4185 (growth_likely_positive): New function.
4186 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
4187 (set_cond_stmt_execution_predicate): Cleanup.
4188 (estimate_edge_size_and_time): Compute min_size.
4189 (estimate_calls_size_and_time): Likewise.
4190 (estimate_node_size_and_time): Likewise.
4191 (inline_update_overall_summary): Update min_size.
4192 (do_estimate_edge_time): Likewise.
4193 (do_estimate_edge_size): Update.
4194 (do_estimate_edge_hints): Update.
4195 (growth_likely_positive): New function.
4197 2014-03-28 Jakub Jelinek <jakub@redhat.com>
4200 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
4201 also if addr has VOIDmode.
4203 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4205 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
4206 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
4208 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
4209 instructions as well as AdvancedSIMD loads.
4211 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4213 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
4214 Use crypto_aese type.
4215 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
4216 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
4217 crypto_aese, crypto_aesmc. Move to types.md.
4218 * config/arm/types.md (crypto_aes): Split into crypto_aese,
4220 * config/arm/iterators.md (crypto_type): Likewise.
4222 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4224 * cgraph.c: Include expr.h and tree-dfa.h.
4225 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
4228 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4231 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
4232 regs from checking multi-reg pseudos.
4234 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4236 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
4238 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4240 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
4241 if it would clobber the stack pointer, even temporarily.
4243 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
4245 * mode-switching.c: Make small adjustments to the top comment.
4247 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4249 * config/rs6000/constraints.md (wD constraint): New constraint to
4250 match the constant integer to get the top DImode/DFmode out of a
4251 vector in a VSX register.
4253 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
4254 match the constant integer to get the top DImode/DFmode out of a
4255 vector in a VSX register.
4257 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
4260 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
4263 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
4264 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
4266 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
4267 Optimize vec_extract of 64-bit values, where the value being
4268 extracted is in the top word, where we can use scalar
4269 instructions. Add direct move and store support. Combine the big
4270 endian/little endian vector select load support into a single insn.
4271 (vsx_extract_<mode>_internal1): Likewise.
4272 (vsx_extract_<mode>_internal2): Likewise.
4273 (vsx_extract_<mode>_load): Likewise.
4274 (vsx_extract_<mode>_store): Likewise.
4275 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
4276 combined into vsx_extract_<mode>_load.
4277 (vsx_extract_<mode>_one_le): Likewise.
4279 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
4280 define the top 64-bit vector element.
4282 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
4285 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4286 Document vec_vbpermq builtin.
4289 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
4290 enable use of xxsldwi and xxpermdi builtin functions.
4291 (vec_xxpermdi): Likewise.
4293 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4294 Document use of vec_xxsldwi and vec_xxpermdi builtins.
4296 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
4298 PR rtl-optimization/60650
4299 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
4301 (find_spills_for): New.
4302 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
4303 Spill all pseudos on the second iteration.
4305 2014-03-27 Marek Polacek <polacek@redhat.com>
4308 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
4311 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4313 * config/s390/s390.c (s390_can_use_return_insn): Check for
4314 call-saved FPRs on 31 bit.
4316 2014-03-27 Jakub Jelinek <jakub@redhat.com>
4319 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
4320 if they need regimplification, just drop them instead of
4321 calling gimple_regimplify_operands on them.
4323 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
4326 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
4327 (aarch64_frame_pointer_required): Adjust logic.
4328 (aarch64_can_eliminate): Adjust logic.
4329 (aarch64_override_options_after_change): Adjust logic.
4331 2014-03-27 Dehao Chen <dehao@google.com>
4333 * ipa-inline.c (early_inliner): Update node's inline info.
4335 2014-03-26 Dehao Chen <dehao@google.com>
4337 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
4338 compiler inserted conditional jumps for NAN float check.
4340 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4342 * ubsan.h (ubsan_create_data): Change second argument's type
4343 to const location_t *.
4344 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
4346 (ubsan_create_data): Change second argument to const location_t *PLOC.
4347 Create Loc field whenever PLOC is non-NULL.
4348 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
4349 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
4353 * real.c (real_to_integer2): Change type of low to UHWI.
4355 2014-03-26 Tobias Burnus <burnus@net-b.de>
4357 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
4358 (CILK_SELF_SPECS): New define.
4359 (driver_self_specs): Use it.
4361 2014-03-26 Richard Biener <rguenther@suse.de>
4363 * tree-pretty-print.c (percent_K_format): Implement special
4364 case for LTO and its stripped down BLOCK tree.
4366 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4369 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
4371 * tree-vrp.c (simplify_internal_call_using_ranges): If only
4372 one range is range_int_cst_p, but not both, at least optimize
4373 addition/subtraction of 0 and multiplication by 0 or 1.
4374 * gimple-fold.c (gimple_fold_call): Fold
4375 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
4376 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
4377 INTEGER_CSTs, try to fold at least x * 0 and y - y.
4379 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
4381 PR rtl-optimization/60452
4382 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
4383 <case REG>: Return 1 for invalid offsets from the frame pointer.
4385 2014-03-26 Marek Polacek <polacek@redhat.com>
4388 * doc/extend.texi (C Extensions): Mention variable-length arrays in
4391 2014-03-26 Marek Polacek <polacek@redhat.com>
4394 * doc/extend.texi (Designated Inits): Describe what happens to omitted
4397 2014-03-26 Marek Polacek <polacek@redhat.com>
4400 * ira-color.c (update_conflict_hard_regno_costs): Perform the
4401 multiplication in unsigned type.
4403 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4405 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
4407 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4409 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
4411 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4414 * cif-code.def (UNREACHABLE) New code.
4415 * ipa-inline.c (inline_small_functions): Skip edges to
4416 __builtlin_unreachable.
4417 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
4418 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
4419 predicate to __bulitin_unreachable.
4420 (set_cond_stmt_execution_predicate): Fix issue when
4421 invert_tree_comparison returns ERROR_MARK.
4422 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
4423 propagate to inline clones.
4424 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
4426 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
4427 * cgraphclones.c (cgraph_clone_node): If call destination is already
4428 ureachable, do not redirect it back.
4429 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
4432 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4434 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
4435 Do not modify inline clones.
4437 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4439 * config/i386/i386.md (general_sext_operand): New mode attr.
4440 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
4441 don't generate (sign_extend (const_int)).
4442 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
4443 operands[2]. Use We constraint instead of <i> and
4444 <general_sext_operand> predicate instead of <general_operand>.
4445 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
4446 * config/i386/constraints.md (We): New constraint.
4447 * config/i386/predicates.md (x86_64_sext_operand,
4448 sext_operand): New predicates.
4450 2014-03-25 Martin Jambor <mjambor@suse.cz>
4453 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
4454 inconsistent devirtualizations to __builtin_unreachable.
4456 2014-03-25 Marek Polacek <polacek@redhat.com>
4459 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
4461 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
4463 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
4464 order of elements for big-endian.
4466 2014-03-25 Richard Biener <rguenther@suse.de>
4469 * gimplify-me.c (gimple_regimplify_operands): Update the
4472 2014-03-25 Martin Jambor <mjambor@suse.cz>
4475 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
4476 (lto_output_varpool_node): Likewise.
4477 (input_overwrite_node): Likewise.
4478 (input_varpool_node): Likewise.
4480 2014-03-25 Richard Biener <rguenther@suse.de>
4482 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
4483 (run_gcc): Likewise.
4485 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4487 * combine.c (simplify_compare_const): Add MODE argument.
4488 Handle mode_width 0 as very large mode_width.
4489 (try_combine, simplify_comparison): Adjust callers.
4491 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
4492 type to avoid signed integer overflow.
4493 * explow.c (plus_constant): Likewise.
4495 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
4497 * doc/generic.texi: Correct typos.
4499 2014-03-24 Tobias Burnus <burnus@net-b.de>
4501 * doc/invoke.texi (-flto): Expand section about
4502 using static libraries with LTO.
4504 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4506 PR rtl-optimization/60501
4507 * optabs.def (addptr3_optab): New optab.
4508 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
4509 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
4510 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
4512 * lra.c (emit_add3_insn): Use the addptr pattern if available.
4514 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
4516 2014-03-24 Ulrich Drepper <drepper@gmail.com>
4518 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
4521 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
4522 (_mm256_undefined_ps): Define.
4523 (_mm256_undefined_pd): Define.
4524 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
4525 (_mm_undefined_pd): Define.
4526 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
4527 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
4528 (_mm512_undefined_ps): Define.
4529 (_mm512_undefined_pd): Define.
4530 Use _mm*_undefined_*.
4531 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
4533 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
4535 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
4536 (lshr_simd): DI mode added.
4537 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
4538 (aarch64_ushr_simddi): Likewise.
4539 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
4540 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
4541 (vshrd_n_u64): Likewise.
4543 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4545 * Makefile.in (s-macro_list): Depend on cc1.
4547 2014-03-23 Teresa Johnson <tejohnson@google.com>
4549 * ipa-utils.c (ipa_print_order): Use specified dump file.
4551 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
4553 PR rtl-optimization/60601
4554 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
4556 * gcc.c (eval_spec_function): Initialize save_growing_value.
4558 2014-03-22 Jakub Jelinek <jakub@redhat.com>
4561 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
4562 code == MINUS_EXPR, never swap op0 with op1.
4564 * toplev.c (init_local_tick): Avoid signed integer multiplication
4566 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
4567 shift by first operand's bitsize.
4569 2014-03-21 Jakub Jelinek <jakub@redhat.com>
4572 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
4574 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
4575 TARGET_ISA_64BIT_P(x).
4577 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4579 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
4580 pattern for vector nor instead of subtract from splat(-1).
4581 (altivec_expand_vec_perm_const_le): Likewise.
4583 2014-03-21 Richard Henderson <rth@twiddle.net>
4586 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
4587 related insns after epilogue_completed.
4589 2014-03-21 Martin Jambor <mjambor@suse.cz>
4592 * cgraph.h (symtab_node): New flag body_removed.
4593 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
4594 when removing bodies.
4595 * symtab.c (dump_symtab_base): Dump body_removed flag.
4596 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
4597 had their bodies removed.
4599 2014-03-21 Martin Jambor <mjambor@suse.cz>
4602 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
4605 2014-03-21 Richard Biener <rguenther@suse.de>
4607 PR tree-optimization/60577
4608 * tree-core.h (struct tree_base): Document nothrow_flag use
4609 in VAR_DECL_NONALIASED.
4610 * tree.h (VAR_DECL_NONALIASED): New.
4611 (may_be_aliased): Adjust.
4612 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
4614 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4616 * expr.c (expand_expr_real_1): Remove outdated comment.
4618 2014-03-20 Jakub Jelinek <jakub@redhat.com>
4621 * ira.c (adjust_cleared_regs): Call copy_rtx on
4622 *reg_equiv[REGNO (loc)].src_p before passing it to
4623 simplify_replace_fn_rtx.
4626 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
4627 into CONST, put pic register as first operand of PLUS. Use
4628 gen_const_mem for both 32-bit and 64-bit PIC got loads.
4630 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4632 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
4634 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4636 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
4637 around for store forwarding issue in the FPU on the UT699.
4638 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
4639 loads and operations if -mfix-ut699 is specified.
4640 (divtf3_hq): Tweak attribute.
4641 (sqrttf2_hq): Likewise.
4643 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4645 * calls.c (store_one_arg): Remove incorrect const qualification on the
4646 type of the temporary.
4647 * cfgexpand.c (expand_return): Likewise.
4648 * expr.c (expand_constructor): Likewise.
4649 (expand_expr_real_1): Likewise.
4651 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4653 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
4656 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
4659 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
4661 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
4663 * config/arm/aarch-common-protos.h
4664 (alu_cost_table): Fix spelling of "extend".
4665 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
4667 2014-03-19 Richard Biener <rguenther@suse.de>
4670 * tree-core.h (tree_type_common): Re-order pointer members
4671 to reduce recursion depth during GC walks.
4673 2014-03-19 Marek Polacek <polacek@redhat.com>
4676 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
4677 before accessing it.
4679 2014-03-19 Richard Biener <rguenther@suse.de>
4682 * lto-streamer-in.c (input_function): In WPA stage do not drop
4685 2014-03-19 Jakub Jelinek <jakub@redhat.com>
4687 PR tree-optimization/60559
4688 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
4689 with build_zero_cst assignment.
4691 2014-03-18 Kai Tietz <ktietz@redhat.com>
4693 PR rtl-optimization/56356
4694 * sdbout.c (sdbout_parms): Verify that parms'
4695 incoming argument is valid.
4696 (sdbout_reg_parms): Likewise.
4698 2014-03-18 Richard Henderson <rth@redhat.com>
4701 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
4702 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
4703 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
4705 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
4707 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
4708 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
4709 Italicize plugin event names in description. Explain that
4710 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
4711 Remind that no GCC functions should be called after PLUGIN_FINISH.
4712 Explain what pragmas with expansion are.
4714 2014-03-18 Martin Liska <mliska@suse.cz>
4716 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
4717 gimple call statement is update.
4718 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
4719 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
4721 2014-03-18 Jakub Jelinek <jakub@redhat.com>
4724 * ubsan.c (ubsan_instrument_unreachable): Call
4725 initialize_sanitizer_builtins.
4726 (ubsan_pass): Likewise.
4729 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
4730 varpool_finalize_decl instead of rest_of_decl_compilation.
4732 2014-03-18 Richard Biener <rguenther@suse.de>
4734 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
4735 by using bitmap_and_compl instead of bitmap_and_compl_into.
4736 (df_rd_transfer_function): Likewise.
4738 2014-03-18 Richard Biener <rguenther@suse.de>
4740 * doc/lto.texi (fresolution): Fix typo.
4742 2014-03-18 Richard Biener <rguenther@suse.de>
4744 * doc/invoke.texi (flto): Update for changes in 4.9.
4746 2014-03-18 Richard Biener <rguenther@suse.de>
4748 * doc/loop.texi: Remove section on the removed lambda framework.
4749 Update loop docs with recent changes in preserving loop structure.
4751 2014-03-18 Richard Biener <rguenther@suse.de>
4753 * doc/lto.texi (-fresolution): Document.
4755 2014-03-18 Richard Biener <rguenther@suse.de>
4757 * doc/contrib.texi: Adjust my name.
4759 2014-03-18 Jakub Jelinek <jakub@redhat.com>
4762 * internal-fn.c: Include diagnostic-core.h.
4763 (expand_BUILTIN_EXPECT): New function.
4764 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
4765 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
4766 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
4767 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
4769 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
4770 Revert 3 argument __builtin_expect code.
4771 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
4772 * gimple-fold.c (gimple_fold_call): Likewise.
4773 * tree.h (fold_builtin_expect): New prototype.
4774 * builtins.c (build_builtin_expect_predicate): Add predictor
4775 argument, if non-NULL, create 3 argument __builtin_expect.
4776 (fold_builtin_expect): No longer static. Add ARG2 argument,
4777 pass it through to build_builtin_expect_predicate.
4778 (fold_builtin_2): Adjust caller.
4779 (fold_builtin_3): Handle BUILT_IN_EXPECT.
4780 * internal-fn.def (BUILTIN_EXPECT): New.
4782 2014-03-18 Tobias Burnus <burnus@net-b.de>
4785 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
4786 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
4787 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
4789 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
4792 * predict.c (combine_predictions_for_bb): Fix up formatting.
4793 (expr_expected_value_1, expr_expected_value): Add predictor argument,
4794 fill what it points to if non-NULL.
4795 (tree_predict_by_opcode): Adjust caller, use the predictor.
4796 * predict.def (PRED_COMPARE_AND_SWAP): Add.
4798 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
4800 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
4801 proper constant for the store mode.
4803 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
4805 * symtab.c (change_decl_assembler_name): Fix transparent alias
4808 2014-03-16 Renlin Li <Renlin.Li@arm.com>
4810 * config/aarch64/aarch64.c: Correct the comments about the
4811 aarch64 stack layout.
4813 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
4815 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
4816 check for GF_OMP_FOR_KIND_FOR.
4818 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
4820 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
4821 ymm and zmm register names.
4823 2014-03-17 Jakub Jelinek <jakub@redhat.com>
4826 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
4827 note creation for the 2010-08-31 changes.
4829 2014-03-17 Marek Polacek <polacek@redhat.com>
4832 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
4833 as -fno-tree-loop-vectorize.
4834 (expand_omp_simd): Likewise.
4836 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
4838 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
4839 (eligible_for_call_delay): New prototype.
4840 * config/sparc/sparc.c (tls_call_delay): Rename into...
4841 (eligible_for_call_delay): ...this. Return false if the instruction
4842 cannot be put in the delay slot of a branch.
4843 (eligible_for_restore_insn): Simplify.
4844 (eligible_for_return_delay): Return false if the instruction cannot be
4845 put in the delay slot of a branch and simplify.
4846 (eligible_for_sibcall_delay): Return false if the instruction cannot be
4847 put in the delay slot of a branch.
4848 * config/sparc/sparc.md (fix_ut699): New attribute.
4849 (tls_call_delay): Delete.
4850 (in_call_delay): Reimplement.
4851 (eligible_for_sibcall_delay): Rename into...
4852 (in_sibcall_delay): ...this.
4853 (eligible_for_return_delay): Rename into...
4854 (in_return_delay): ...this.
4855 (in_branch_delay): Reimplement.
4856 (in_uncond_branch_delay): Delete.
4857 (in_annul_branch_delay): Delete.
4859 2014-03-14 Richard Henderson <rth@redhat.com>
4862 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
4863 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
4864 (*floathi<X87MODEF>2_i387_with_temp): Remove.
4865 (floathi splitters): Remove.
4866 (float<SWI48x>xf2): New pattern.
4867 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
4868 code that tried to handle DImode for 32-bit, but which was excluded
4869 by the pattern's condition. Drop allocation of stack temporary.
4870 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
4871 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
4872 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
4873 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
4874 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
4875 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
4876 (*float<SWI48><MODEF>2_sse_interunit): Remove.
4877 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
4878 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
4879 (*float<SWI48x><X87MODEF>2_i387): Remove.
4880 (all float _with_temp splitters): Remove.
4881 (*float<SWI48x><MODEF>2_i387): New pattern.
4882 (*float<SWI48><MODEF>2_sse): New pattern.
4883 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
4884 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
4886 2014-03-14 Jakub Jelinek <jakub@redhat.com>
4887 Marek Polacek <polacek@redhat.com>
4890 * common.opt (dump_base_name_prefixed): New Variable.
4891 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
4892 if x_dump_base_name_prefixed is already set, set it at the end.
4894 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
4896 PR rtl-optimization/60508
4897 * lra-constraints.c (get_reload_reg): Add new parameter
4899 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
4900 Pass the new parameter values.
4902 2014-03-14 Richard Biener <rguenther@suse.de>
4904 * common.opt: Revert unintented changes from r205065.
4907 2014-03-14 Richard Biener <rguenther@suse.de>
4910 * cfghooks.c (split_block): Properly adjust all loops the
4911 block was a latch of.
4913 2014-03-14 Martin Jambor <mjambor@suse.cz>
4916 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
4919 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
4922 * config/avr/avr.c (avr_set_current_function): Pass function name
4923 through default_strip_name_encoding before sanity checking instead
4924 of skipping the first char of the assembler name.
4926 2014-03-13 Richard Henderson <rth@redhat.com>
4929 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
4930 (ix86_force_to_memory, ix86_free_from_memory): Remove.
4931 * config/i386/i386-protos.h: Likewise.
4932 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
4933 in the expander instead of a splitter.
4934 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
4935 any possibility of requiring a memory.
4936 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
4937 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
4938 (fp branch splitters): Update for ix86_split_fp_branch.
4939 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
4940 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
4941 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
4942 (*fop_<MODEF>_2_i387): Remove f/r alternative.
4943 (*fop_<MODEF>_3_i387): Likewise.
4944 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
4945 (splitters for the fop_* register patterns): Remove.
4946 (fscalexf4_i387): Rename from *fscalexf4_i387.
4947 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
4949 2014-03-13 Jakub Jelinek <jakub@redhat.com>
4951 PR tree-optimization/59779
4952 * tree-dfa.c (get_ref_base_and_extent): Use double_int
4953 type for bitsize and maxsize instead of HOST_WIDE_INT.
4955 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
4957 PR rtl-optimization/57320
4958 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
4959 the CFG after thread_prologue_and_epilogue_insns.
4961 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
4963 PR rtl-optimization/57189
4964 * lra-constraints.c (process_alt_operands): Disfavor spilling
4967 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
4969 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
4971 2014-03-13 Jakub Jelinek <jakub@redhat.com>
4973 PR tree-optimization/59025
4975 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
4976 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
4978 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
4981 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
4982 calls of avr_out_plus_1.
4984 2014-03-13 Bin Cheng <bin.cheng@arm.com>
4986 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
4987 BB's single pred and update the father loop's latch info later.
4989 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
4991 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
4995 (VEC_base): Likewise.
4996 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
4997 registers, we need to swap double words in little endian mode.
4999 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
5000 to be a container mode for 128-bit integer operations added in ISA
5001 2.07. Unlike TImode and PTImode, the preferred register set is
5002 the Altivec/VMX registers for the 128-bit operations.
5004 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
5006 (rs6000_split_128bit_ok_p): Likewise.
5008 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
5009 macros for creating ISA 2.07 normal and overloaded builtin
5010 functions with 3 arguments.
5011 (BU_P8V_OVERLOAD_3): Likewise.
5012 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
5013 for use as overloaded functions.
5014 (VPERM_1TI_UNS): Likewise.
5015 (VSEL_1TI): Likewise.
5016 (VSEL_1TI_UNS): Likewise.
5017 (ST_INTERNAL_1ti): Likewise.
5018 (LD_INTERNAL_1ti): Likewise.
5019 (XXSEL_1TI): Likewise.
5020 (XXSEL_1TI_UNS): Likewise.
5021 (VPERM_1TI): Likewise.
5022 (VPERM_1TI_UNS): Likewise.
5023 (XXPERMDI_1TI): Likewise.
5024 (SET_1TI): Likewise.
5025 (LXVD2X_V1TI): Likewise.
5026 (STXVD2X_V1TI): Likewise.
5027 (VEC_INIT_V1TI): Likewise.
5028 (VEC_SET_V1TI): Likewise.
5029 (VEC_EXT_V1TI): Likewise.
5030 (EQV_V1TI): Likewise.
5031 (NAND_V1TI): Likewise.
5032 (ORC_V1TI): Likewise.
5033 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
5034 added in ISA 2.07. Add both normal 'altivec' builtins, and the
5036 (VADDUQM): Likewise.
5037 (VSUBCUQ): Likewise.
5038 (VADDEUQM): Likewise.
5039 (VADDECUQ): Likewise.
5040 (VSUBEUQM): Likewise.
5041 (VSUBECUQ): Likewise.
5043 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5044 __int128_t and __uint128_t types.
5045 (__uint128_type): Likewise.
5046 (altivec_categorize_keyword): Add support for vector __int128_t,
5047 vector __uint128_t, vector __int128, and vector unsigned __int128
5048 as a container type for TImode operations that need to be done in
5049 VSX/Altivec registers.
5050 (rs6000_macro_to_expand): Likewise.
5051 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5052 to support 128-bit integer instructions vaddcuq, vadduqm,
5053 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5054 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5056 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5057 for V1TImode, and set up preferences to use VSX/Altivec registers.
5058 Setup VSX reload handlers.
5059 (rs6000_debug_reg_global): Likewise.
5060 (rs6000_init_hard_regno_mode_ok): Likewise.
5061 (rs6000_preferred_simd_mode): Likewise.
5062 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5063 (easy_altivec_constant): Likewise.
5064 (output_vec_const_move): Likewise.
5065 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5067 (rs6000_expand_vector_extract): Likewise.
5068 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
5070 (rs6000_const_vec): Add support for V1TImode.
5071 (rs6000_emit_le_vsx_load): Swap double words when loading or
5072 storing TImode/V1TImode.
5073 (rs6000_emit_le_vsx_store): Likewise.
5074 (rs6000_emit_le_vsx_move): Likewise.
5075 (rs6000_emit_move): Add support for V1TImode.
5076 (altivec_expand_ld_builtin): Likewise.
5077 (altivec_expand_st_builtin): Likewise.
5078 (altivec_expand_vec_init_builtin): Likewise.
5079 (altivec_expand_builtin): Likewise.
5080 (rs6000_init_builtins): Add support for V1TImode type. Add
5081 support for ISA 2.07 128-bit integer builtins. Define type names
5082 for the VSX/Altivec vector types.
5083 (altivec_init_builtins): Add support for overloaded vector
5084 functions with V1TImode type.
5085 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
5086 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
5088 (rs6000_split_128bit_ok_p): Likewise.
5089 (rs6000_handle_altivec_attribute): Create V1TImode from vector
5090 __int128_t and vector __uint128_t.
5092 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
5093 and mode attributes.
5100 (VS_scalar): Likewise.
5101 (VS_double): Likewise.
5102 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
5104 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
5105 we support the ISA 2.07 128-bit integer arithmetic instructions.
5106 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
5107 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
5108 and TImode types for use with the builtin functions.
5109 (V1TI_type_node): Likewise.
5110 (unsigned_V1TI_type_node): Likewise.
5111 (intTI_type_internal_node): Likewise.
5112 (uintTI_type_internal_node): Likewise.
5114 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
5115 128-bit builtin functions.
5116 (UNSPEC_VADDEUQM): Likewise.
5117 (UNSPEC_VADDECUQ): Likewise.
5118 (UNSPEC_VSUBCUQ): Likewise.
5119 (UNSPEC_VSUBEUQM): Likewise.
5120 (UNSPEC_VSUBECUQ): Likewise.
5121 (VM): Add V1TImode to vector mode iterators.
5123 (VI_unit): Likewise.
5124 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
5125 (altivec_vaddcuq): Likewise.
5126 (altivec_vsubuqm): Likewise.
5127 (altivec_vsubcuq): Likewise.
5128 (altivec_vaddeuqm): Likewise.
5129 (altivec_vaddecuq): Likewise.
5130 (altivec_vsubeuqm): Likewise.
5131 (altivec_vsubecuq): Likewise.
5133 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
5135 (BOOL_128): Likewise.
5136 (BOOL_REGS_OUTPUT): Likewise.
5137 (BOOL_REGS_OP1): Likewise.
5138 (BOOL_REGS_OP2): Likewise.
5139 (BOOL_REGS_UNARY): Likewise.
5140 (BOOL_REGS_AND_CR0): Likewise.
5142 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
5143 128-bit integer builtin support.
5144 (vec_vadduqm): Likewise.
5145 (vec_vaddecuq): Likewise.
5146 (vec_vaddeuqm): Likewise.
5147 (vec_vsubecuq): Likewise.
5148 (vec_vsubeuqm): Likewise.
5149 (vec_vsubcuq): Likewise.
5150 (vec_vsubuqm): Likewise.
5152 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5153 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
5154 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
5155 128-bit integer add/subtract to ISA 2.07.
5157 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
5159 * config/arc/arc.c (arc_predicate_delay_insns):
5160 Fix third argument passed to conditionalize_nonjump.
5162 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
5164 * config/aarch64/aarch64-builtins.c
5165 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
5166 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
5167 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
5168 instead of __builtin_lfloor.
5169 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
5171 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5173 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
5174 (tree_ssa_ifcombine_bb_1): New function.
5175 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
5176 is an empty forwarder block to then_bb or vice versa and then_bb
5177 and else_bb are effectively swapped.
5179 2014-03-12 Christian Bruel <christian.bruel@st.com>
5182 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
5183 REG_CFA_DEF_CFA note.
5184 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
5185 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
5187 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
5189 PR tree-optimization/60454
5190 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
5192 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5194 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
5195 Do not define target_cpu_default2 to generic.
5196 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
5197 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
5198 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
5200 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5201 Marc Glisse <marc.glisse@inria.fr>
5203 PR tree-optimization/60502
5204 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
5205 instead of build_low_bits_mask.
5207 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5210 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
5211 if there are multiple uses, but op doesn't live on E edge.
5212 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
5213 clobber stmts before __builtin_unreachable.
5215 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
5217 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
5218 hard_frame_pointer_rtx.
5219 * cse.c (cse_insn): Remove volatile check.
5220 * cselib.c (cselib_process_insn): Likewise.
5221 * dse.c (scan_insn): Likewise.
5223 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5225 * config/arc/arc.c (conditionalize_nonjump): New function,
5227 (arc_ifcvt): ... this.
5228 (arc_predicate_delay_insns): Use it.
5230 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5232 * config/arc/predicates.md (extend_operand): During/after reload,
5233 allow const_int_operand.
5234 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
5235 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
5236 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
5238 (umulsi3_highpart_i): Likewise.
5240 2014-03-11 Richard Biener <rguenther@suse.de>
5242 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
5243 Add asserts to guard possible wrong-code bugs.
5245 2014-03-11 Richard Biener <rguenther@suse.de>
5247 PR tree-optimization/60429
5248 PR tree-optimization/60485
5249 * tree-ssa-structalias.c (set_union_with_increment): Properly
5250 take into account all fields that overlap the shifted vars.
5251 (do_sd_constraint): Likewise.
5252 (do_ds_constraint): Likewise.
5253 (get_constraint_for_ptr_offset): Likewise.
5255 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
5257 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
5258 (nios2_compute_frame_layout):
5259 Add calculation of cfun->machine->fp_save_offset.
5260 (nios2_expand_prologue): Correct setting of frame pointer register
5262 (nios2_expand_epilogue): Update recovery of stack pointer from
5263 frame pointer accordingly.
5264 (nios2_initial_elimination_offset): Update calculation of offset
5265 for eliminating to HARD_FRAME_POINTER_REGNUM.
5267 2014-03-10 Jakub Jelinek <jakub@redhat.com>
5270 * ipa.c (symtab_remove_unreachable_nodes): Don't call
5271 cgraph_get_create_node on VAR_DECLs.
5273 2014-03-10 Richard Biener <rguenther@suse.de>
5276 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
5278 2014-03-08 Douglas B Rupp <rupp@gnat.com>
5280 * config/vms/vms.opt (vms_float_format): New variable.
5282 2014-03-08 Tobias Burnus <burnus@net-b.de>
5284 * doc/invoke.texi (-fcilkplus): Update implementation status.
5286 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
5287 Richard Biener <rguenther@suse.de>
5289 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
5290 consistently accross all TUs.
5291 (run_gcc): Enable -fshort-double automatically at link at link-time
5292 and disallow override.
5294 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
5297 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
5298 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
5299 if they can't be used.
5301 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5303 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
5304 for Solaris 11/x86 ld.
5305 * configure: Regenerate.
5307 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5309 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
5310 (LIB_TLS_SPEC): Save as ld_tls_libs.
5311 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
5312 (HAVE_AS_IX86_TLSLDM): New test.
5313 * configure, config.in: Regenerate.
5314 * config/i386/i386.c (legitimize_tls_address): Fall back to
5315 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
5316 cannot support TLS_MODEL_LOCAL_DYNAMIC.
5317 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
5318 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
5320 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
5322 * common.opt (fira-loop-pressure): Mark as optimization.
5324 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
5326 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
5327 an OpenMP mappable type.
5329 2014-03-06 Matthias Klose <doko@ubuntu.com>
5331 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
5332 MULTILIB_OSDIRNAMES is not defined.
5334 2014-03-06 Jakub Jelinek <jakub@redhat.com>
5335 Meador Inge <meadori@codesourcery.com>
5338 * config/arm/arm.c (arm_tls_symbol_p): Remove.
5339 (arm_legitimize_address): Call legitimize_tls_address for any
5340 arm_tls_referenced_p expression, handle constant addend. Call it
5341 before testing for !TARGET_ARM.
5342 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
5344 2014-03-06 Richard Biener <rguenther@suse.de>
5350 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5352 * tree-streamer.c (record_common_node): Assert we don't record
5353 nodes with type double.
5354 (preload_common_node): Skip type double, complex double and double
5355 pointer since it is now frontend dependent due to fshort-double option.
5357 2014-03-06 Richard Biener <rguenther@suse.de>
5359 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
5360 or -fno-lto is specified and the linker has full plugin support.
5361 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
5362 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
5363 * lto-wrapper.c (merge_and_complain): Merge compile-time
5364 optimization levels.
5365 (run_gcc): And pass it through to the link options.
5367 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
5371 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5373 * cselib.c (remove_useless_values): Skip to avoid quadratic
5374 behavior if the condition moved from...
5375 (cselib_process_insn): ... here holds.
5377 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5380 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
5381 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
5384 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
5385 (TM_H): Add x86-tune.def.
5387 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5389 * config/aarch64/aarch64.c (generic_tunings):
5390 Use cortexa57_extra_costs.
5392 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5395 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
5396 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
5397 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
5400 2014-03-04 Heiher <r@hev.cc>
5402 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
5403 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
5405 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
5407 * config/i386/predicates.md (const2356_operand): Change to ...
5408 (const2367_operand): ... this.
5409 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
5411 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5412 (*avx512pf_scatterpf<mode>sf): Ditto.
5413 (avx512pf_scatterpf<mode>df): Ditto.
5414 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5415 (*avx512pf_scatterpf<mode>df): Ditto.
5416 * config/i386/i386.c (ix86_expand_builtin): Update
5417 incorrect hint operand error message.
5419 2014-03-04 Richard Biener <rguenther@suse.de>
5421 * lto-section-in.c (lto_get_section_data): Fix const cast.
5423 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5425 * tree-streamer.c (record_common_node): Assert we don't record
5426 nodes with type double.
5427 (preload_common_node): Skip type double, complex double and double
5428 pointer since it is now frontend dependent due to fshort-double option.
5430 2014-03-04 Richard Biener <rguenther@suse.de>
5433 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
5434 (lto_input_toplevel_asms): Likewise.
5435 * lto-section-in.c (lto_get_section_data): Instead do it here
5438 2014-03-04 Richard Biener <rguenther@suse.de>
5440 PR tree-optimization/60382
5441 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
5442 dead PHIs a reduction.
5444 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
5446 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
5448 (_mm_prefetch): Move out of GCC target("sse") pragma.
5449 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
5450 GCC target("prfchw") pragma.
5451 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
5453 * config/i386/i386.c (ix86_option_override_internal): Enable
5454 -mprfchw with -mprefetchwt1.
5456 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5458 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
5461 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5463 * opts.h (CL_PCH_IGNORE): Define.
5464 * targhooks.c (option_affects_pch_p):
5465 Return false for options that have CL_PCH_IGNORE set.
5466 * opt-functions.awk: Process PchIgnore.
5467 * doc/options.texi: Document PchIgnore.
5469 * config/arc/arc.opt (misize): Add PchIgnore property.
5471 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5473 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
5474 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
5475 constraint on constants to permit them being loaded into
5476 GENERAL_REGS or BASE_REGS.
5478 2014-03-03 Nick Clifton <nickc@redhat.com>
5480 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
5481 anti-cacnonical alternatives.
5482 (negandhi3_real): New pattern.
5483 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
5485 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5487 * config/avr/avr-mcus.def: Remove atxmega16x1.
5488 * config/avr/avr-tables.opt: Regenerate.
5489 * config/avr/t-multilib: Regenerate.
5490 * doc/avr-mmcu.texi: Regenerate.
5492 2014-03-03 Tobias Grosser <tobias@grosser.es>
5493 Mircea Namolaru <mircea.namolaru@inria.fr>
5495 PR tree-optimization/58028
5496 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
5499 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5501 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
5502 not handled by recognizers.
5504 2014-03-03 Jakub Jelinek <jakub@redhat.com>
5507 * function.c (expand_function_end): Don't emit
5508 clobber_return_register sequence if clobber_after is a BARRIER.
5509 * cfgexpand.c (construct_exit_block): Append instructions before
5510 return_label to prev_bb.
5512 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5514 * config/rs6000/constraints.md: Document reserved use of "wc".
5516 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5519 * ipa.c (function_and_variable_visibility): When dissolving comdat
5520 group, also set all symbols to local.
5522 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5527 2013-12-14 Jan Hubicka <jh@suse.cz>
5529 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
5531 2014-03-02 Jon Beniston <jon@beniston.com>
5537 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
5538 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
5539 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
5540 (simple_return, *simple_return): New patterns
5541 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
5542 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
5544 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
5546 * dwarf2out.c (gen_subprogram_die): Tidy.
5548 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
5551 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
5552 (*mov_t_msb_neg_negc): ... this new insn.
5554 2014-02-28 Jason Merrill <jason@redhat.com>
5557 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
5560 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
5563 * dwarf2out.c (decltype_auto_die): New static.
5564 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
5565 (gen_type_die_with_usage): Handle 'decltype(auto)'.
5566 (is_cxx_auto): Likewise.
5568 2014-02-28 Ian Bolton <ian.bolton@arm.com>
5570 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
5571 we are not using general regs only.
5573 2014-02-28 Richard Biener <rguenther@suse.de>
5576 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
5577 previous fix and only allow to remove trivial pre-headers
5578 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
5579 (remove_forwarder_block): Properly update the latch of a loop.
5581 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5584 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
5585 (cselib_preserved_hash_table): New.
5586 (preserve_constants_and_equivs): Move preserved vals to it.
5587 (cselib_find_slot): Look it up first.
5588 (cselib_init): Initialize it.
5589 (cselib_finish): Release it.
5590 (dump_cselib_table): Dump it.
5592 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5595 * cselib.c (remove_useless_values): Skip to avoid quadratic
5596 behavior if the condition moved from...
5597 (cselib_process_insn): ... here holds.
5599 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5602 * var-tracking.c (vt_initialize): Apply the same condition to
5603 preserve the CFA base value.
5605 2014-02-28 Joey Ye <joey.ye@arm.com>
5608 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
5609 if reload in progress or completed.
5611 2014-02-28 Tobias Burnus <burnus@net-b.de>
5614 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
5617 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
5619 * doc/tm.texi.in (Condition Code Status): Update documention for
5620 relative locations of cc0-setter and cc0-user.
5622 2014-02-27 Jeff Law <law@redhat.com>
5624 PR rtl-optimization/52714
5625 * combine.c (try_combine): When splitting an unrecognized PARALLEL
5626 into two independent simple sets, if I3 is a jump, ensure the
5627 pattern we place into I3 is a (set (pc) ...).
5629 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
5630 Jeff Law <law@redhat.com>
5632 PR rtl-optimization/49847
5633 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
5634 are in different blocks.
5635 * doc/tm.texi (Condition Code Status): Update documention for
5636 relative locations of cc0-setter and cc0-user.
5638 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
5641 * lra.c (lra_emit_add): Check SUBREG too.
5643 2014-02-27 Andreas Schwab <schwab@suse.de>
5645 * config/m68k/m68k.c (m68k_option_override): Disable
5646 -flive-range-shrinkage for classic m68k.
5647 (m68k_override_options_after_change): Likewise.
5649 2014-02-27 Marek Polacek <polacek@redhat.com>
5652 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
5653 -Wmaybe-uninitialized.
5655 2014-02-27 Alan Modra <amodra@gmail.com>
5658 * reload1.c (emit_input_reload_insns): When reload_override_in,
5659 set old to rl->in_reg when rl->in_reg is a subreg.
5661 2014-02-26 Richard Biener <rguenther@suse.de>
5664 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
5666 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
5668 * common/config/i386/predicates.md (const1256_operand): Remove.
5669 (const2356_operand): New.
5670 (const_1_to_2_operand): Remove.
5671 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5672 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5673 (*avx512pf_gatherpf<mode>sf): Ditto.
5674 (avx512pf_gatherpf<mode>df): Ditto.
5675 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5676 (*avx512pf_gatherpf<mode>df): Ditto.
5677 (avx512pf_scatterpf<mode>sf): Ditto.
5678 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5679 (*avx512pf_scatterpf<mode>sf): Ditto.
5680 (avx512pf_scatterpf<mode>df): Ditto.
5681 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5682 (*avx512pf_scatterpf<mode>df): Ditto.
5683 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
5685 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
5687 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
5688 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
5689 (_mm512_mask_testn_epi64_mask): Move to ...
5690 * config/i386/avx512cdintrin.h: Here.
5691 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
5692 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
5693 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
5694 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
5695 TARGET_AVX512F from TARGET_AVX512CD.
5697 2014-02-26 Richard Biener <rguenther@suse.de>
5700 * ipa.c (walk_polymorphic_call_targets): Properly guard
5701 call to inline_update_overall_summary.
5703 2014-02-26 Bin Cheng <bin.cheng@arm.com>
5706 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
5707 and latches only if requested. Fix latch if it is removed.
5708 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
5709 LOOPS_HAVE_PREHEADERS.
5711 2014-02-25 Andrew Pinski <apinski@cavium.com>
5713 * builtins.c (expand_builtin_thread_pointer): Create a new target
5714 when the target is NULL.
5716 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
5718 PR rtl-optimization/60317
5719 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
5720 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
5721 * lra-assigns.c: Include params.h.
5722 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
5723 other reload pseudos considerations.
5725 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5727 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
5728 to use canonical form for nor<mode>3.
5730 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5733 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
5736 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
5738 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
5739 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
5740 (ix86_handle_option): Handle OPT_mprefetchwt1.
5741 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
5742 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
5744 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
5745 OPTION_MASK_ISA_PREFETCHWT1.
5746 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
5747 (PTA_PREFETCHWT1): New.
5748 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
5749 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
5750 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
5751 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
5752 (*prefetch_avx512pf_<mode>_: Change into ...
5753 (*prefetch_prefetchwt1_<mode>: This.
5754 * config/i386/i386.opt (mprefetchwt1): New.
5755 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
5756 (_mm_prefetch): Handle intent to write.
5757 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
5759 2014-02-25 Richard Biener <rguenther@suse.de>
5762 * emit-rtl.c (mem_attrs_htab): Remove.
5763 (mem_attrs_htab_hash): Likewise.
5764 (mem_attrs_htab_eq): Likewise.
5765 (set_mem_attrs): Always allocate new mem-attrs when something changed.
5766 (init_emit_once): Do not allocate mem_attrs_htab.
5768 2014-02-25 Richard Biener <rguenther@suse.de>
5771 * lto-opts.c (lto_write_options): Output non-explicit conservative
5772 -fwrapv, -fno-trapv and -fno-strict-overflow.
5773 * lto-wrapper.c (merge_and_complain): Handle merging those options.
5774 (run_gcc): And pass them through.
5776 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
5778 * sel-sched.c (calculate_new_fences): New parameter ptime.
5779 Calculate it as a maximum over all fence cycles.
5780 (sel_sched_region_2): Adjust the call to calculate_new_fences.
5781 Print the final schedule timing when sched_verbose.
5783 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
5785 PR rtl-optimization/60292
5786 * sel-sched.c (fill_vec_av_set): Do not reset target availability
5787 bit fot the fence instruction.
5789 2014-02-24 Alangi Derick <alangiderick@gmail.com>
5791 * calls.h: Fix typo in comment.
5793 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
5795 * config/pa/pa.c (pa_output_move_double): Don't valididate when
5796 adjusting offsetable addresses.
5798 2014-02-24 Guozhi Wei <carrot@google.com>
5800 * sparseset.h (sparseset_pop): Fix the wrong index.
5802 2014-02-24 Walter Lee <walt@tilera.com>
5804 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
5805 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
5807 * common/config/tilegx/tilegx-common.c
5808 (TARGET_DEFAULT_TARGET_FLAGS): Define.
5809 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
5811 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
5812 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
5813 (tilegx_gimplify_va_arg_expr): Handle big endian.
5814 (tilegx_expand_unaligned_load): Ditto.
5815 (tilegx_expand_unaligned_store): Ditto.
5816 (TARGET_RETURN_IN_MSB): New.
5817 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
5818 (TARGET_ENDIAN_DEFAULT): New.
5819 (TARGET_BIG_ENDIAN): Handle big endian.
5820 (BYTES_BIG_ENDIAN): Ditto.
5821 (WORDS_BIG_ENDIAN): Ditto.
5822 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
5825 * config/tilegx/tilegx.md (extv): Handle big endian.
5827 (insn_st<n>): Ditto.
5828 (insn_st<n>_add<bitsuffix>): Ditto.
5829 (insn_stnt<n>): Ditto.
5830 (insn_stnt<n>_add<bitsuffix>):Ditto.
5831 (vec_interleave_highv8qi): Handle big endian.
5832 (vec_interleave_highv8qi_be): New.
5833 (vec_interleave_highv8qi_le): New.
5834 (insn_v1int_h): Handle big endian.
5835 (vec_interleave_lowv8qi): Handle big endian.
5836 (vec_interleave_lowv8qi_be): New.
5837 (vec_interleave_lowv8qi_le): New.
5838 (insn_v1int_l): Handle big endian.
5839 (vec_interleave_highv4hi): Handle big endian.
5840 (vec_interleave_highv4hi_be): New.
5841 (vec_interleave_highv4hi_le): New.
5842 (insn_v2int_h): Handle big endian.
5843 (vec_interleave_lowv4hi): Handle big endian.
5844 (vec_interleave_lowv4hi_be): New.
5845 (vec_interleave_lowv4hi_le): New.
5846 (insn_v2int_l): Handle big endian.
5847 (vec_interleave_highv2si): Handle big endian.
5848 (vec_interleave_highv2si_be): New.
5849 (vec_interleave_highv2si_le): New.
5850 (insn_v4int_h): Handle big endian.
5851 (vec_interleave_lowv2si): Handle big endian.
5852 (vec_interleave_lowv2si_be): New.
5853 (vec_interleave_lowv2si_le): New.
5854 (insn_v4int_l): Handle big endian.
5855 * config/tilegx/tilegx.opt (mbig-endian): New option.
5856 (mlittle-endian): New option.
5857 * doc/install.texi: Document tilegxbe-linux.
5858 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
5860 2014-02-24 Martin Jambor <mjambor@suse.cz>
5863 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
5864 there are no parameter descriptors.
5866 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
5868 PR rtl-optimization/60268
5869 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
5870 initialization to ...
5871 (sched_rgn_init): ... here.
5872 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
5874 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5876 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
5879 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
5881 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
5884 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5886 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
5887 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
5889 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5891 * config/microblaze/predicates.md: Add cmp_op predicate.
5892 * config/microblaze/microblaze.md: Add branch_compare instruction
5893 which uses cmp_op predicate and emits cmp insn before branch.
5894 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
5895 to microblaze_expand_conditional_branch and consolidate logic.
5896 (microblaze_expand_conditional_branch): emit branch_compare
5897 insn instead of handling cmp op separate from branch insn.
5899 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5901 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
5904 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5906 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
5907 define_insn with define_expand and new define_insn
5908 *altivec_lve<VI_char>x_internal.
5909 (altivec_stve<VI_char>x): Replace define_insn with define_expand
5910 and new define_insn *altivec_stve<VI_char>x_internal.
5911 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
5913 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
5915 (altivec_expand_stvex_be): New function.
5917 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
5919 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
5920 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
5921 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
5922 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
5924 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
5927 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
5928 instead of emit_move_insn.
5930 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5932 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
5934 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
5935 gen_altivec_vsumsws.
5937 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5939 * config/rs6000/altivec.md (altivec_lvxl): Rename as
5940 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
5941 (altivec_lvxl_<mode>): New define_expand incorporating
5942 -maltivec=be semantics where needed.
5943 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
5944 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
5945 semantics where needed.
5946 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
5947 (altivec_stvx_<mode>): New define_expand incorporating
5948 -maltivec=be semantics where needed.
5949 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
5950 VM2 iterator instead of V4SI.
5951 (altivec_stvxl_<mode>): New define_expand incorporating
5952 -maltivec=be semantics where needed.
5953 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
5954 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
5955 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
5956 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
5957 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
5958 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
5959 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
5960 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
5961 ALTIVEC_BUILTIN_STVXL.
5962 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
5963 (altivec_expand_stvx_be): Likewise.
5964 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
5965 (altivec_expand_lvx_be): Likewise.
5966 (altivec_expand_stvx_be): Likewise.
5967 (altivec_expand_builtin): Add cases for
5968 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
5969 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
5970 (altivec_init_builtins): Add definitions for
5971 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
5972 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
5974 2014-02-21 Catherine Moore <clm@codesourcery.com>
5976 * doc/invoke.texi (mvirt, mno-virt): Document.
5977 * config/mips/mips.opt (mvirt): New option.
5978 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
5980 2014-02-21 Richard Biener <rguenther@suse.de>
5982 PR tree-optimization/60276
5983 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
5984 (STMT_VINFO_MIN_NEG_DIST): New macro.
5985 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
5986 STMT_VINFO_MIN_NEG_DIST.
5987 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
5988 made for negative dependence distances still hold.
5990 2014-02-21 Richard Biener <rguenther@suse.de>
5993 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
5994 DECL_INITIAL for globals not in the current function context.
5996 2014-02-21 Jakub Jelinek <jakub@redhat.com>
5998 PR tree-optimization/56490
5999 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
6000 * tree-ssa-uninit.c: Include params.h.
6001 (compute_control_dep_chain): Add num_calls argument, return false
6002 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
6003 num_calls to recursive call.
6004 (find_predicates): Change dep_chain into normal array,
6005 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
6006 variable and adjust compute_control_dep_chain caller.
6007 (find_def_preds): Likewise.
6009 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
6011 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
6012 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
6014 2014-02-21 Nick Clifton <nickc@redhat.com>
6016 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
6017 (pushhi1): Likewise.
6018 (popqi1): Add mode to pre_dec.
6021 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6023 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
6024 mode for mask of V8SFmode permutation.
6026 2014-02-20 Richard Henderson <rth@redhat.com>
6029 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
6030 a new pseudo for OLDVAL.
6032 2014-02-20 Jakub Jelinek <jakub@redhat.com>
6035 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
6036 gen_reg_rtx if d->testing_p.
6037 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
6038 if d->testing_p and we will certainly return true.
6039 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6042 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6044 * emit-rtl.c (gen_reg_rtx): Assert that
6045 crtl->emit.regno_pointer_align_length is non-zero.
6047 2014-02-20 Richard Henderson <rth@redhat.com>
6050 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6051 on failure the store back into EXPECT.
6053 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6054 Sandra Loosemore <sandra@codesourcery.com>
6056 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6057 * config/nios2/nios2.c (nios2_function_profiler): Add
6058 -fPIC (flag_pic == 2) support.
6059 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6060 (nios2_large_offset_p): New function.
6061 (nios2_unspec_reloc_p): Move up position, update to use
6062 nios2_large_offset_p.
6063 (nios2_unspec_address): Remove function.
6064 (nios2_unspec_offset): New function.
6065 (nios2_large_got_address): New function.
6066 (nios2_got_address): Add large offset support.
6067 (nios2_legitimize_tls_address): Update usage of removed and new
6069 (nios2_symbol_binds_local_p): New function.
6070 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
6071 (nios2_legitimize_address): Update to use nios2_large_offset_p.
6072 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
6073 (nios2_print_operand): Merge H/L processing, add hiadj/lo
6074 processing for (const (unspec ...)).
6075 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
6077 2014-02-20 Richard Biener <rguenther@suse.de>
6079 * tree-cfg.c (replace_uses_by): Mark altered BBs before
6080 doing the substitution.
6081 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
6083 2014-02-20 Martin Jambor <mjambor@suse.cz>
6086 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
6087 info when checking whether lattices are bottom.
6089 2014-02-20 Richard Biener <rguenther@suse.de>
6092 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
6095 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
6098 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
6099 parameter specifying the scaling.
6100 (inline_call): Update.
6101 (want_inline_recursively): Guard division by zero.
6102 (recursive_inlining): Update.
6103 * ipa-inline.h (clone_inlined_nodes): Update.
6105 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6108 * config/i386/i386.c (classify_argument): Pass structures of size
6109 64 bytes or less in register.
6111 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6112 Kirill Yukhin <kirill.yukhin@intel.com>
6114 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
6115 (_mm_rcp28_round_ss): Ditto.
6116 (_mm_rsqrt28_round_sd): Ditto.
6117 (_mm_rsqrt28_round_ss): Ditto.
6118 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
6119 (_mm_rcp14_round_ss): Ditto.
6120 (_mm_rsqrt14_round_sd): Ditto.
6121 (_mm_rsqrt14_round_ss): Ditto.
6122 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
6123 the first input operand, get rid of match_dup.
6124 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
6126 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
6128 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
6129 operand as the first input operand, set type attribute.
6130 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
6132 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
6133 operand as the first input operand, set type attribute.
6135 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6137 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
6140 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
6143 * config/i386/i386.c (construct_container): Remove TFmode check
6144 for X86_64_INTEGER_CLASS.
6146 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
6149 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
6150 only when -Wpsabi is enabled.
6152 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
6155 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
6156 passing arrays in registers are the same as for structs, so remove the
6157 special case for them.
6159 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
6161 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
6162 destination type, extract only the valid bits if the source type is not
6163 integral and has a different mode.
6165 2014-02-19 Richard Biener <rguenther@suse.de>
6168 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
6171 2014-02-19 Richard Biener <rguenther@suse.de>
6174 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
6175 (ipa_modify_call_arguments): Emit an argument load explicitely and
6176 preserve virtual SSA form there and for the replacement call.
6177 Do not update SSA form nor free dominance info.
6179 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6181 * ipa.c (function_and_variable_visibility): Also clear WEAK
6182 flag when disolving COMDAT_GROUP.
6184 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6186 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
6187 * ipa-prop.c (ipa_set_jf_known_type): Return early when
6189 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
6190 do more sanity checks.
6191 (detect_type_change): Return true when giving up early.
6192 (compute_complex_assign_jump_func): Fix type parameter of
6193 ipa_set_ancestor_jf.
6194 (compute_complex_ancestor_jump_func): Likewise.
6195 (update_jump_functions_after_inlining): Fix updating of
6197 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
6199 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6201 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
6202 inline clones when edge disappears.
6204 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6207 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
6208 Split 64-bit moves into 2 patterns. Do not allow the use of
6209 direct move for TDmode in little endian, since the decimal value
6210 has little endian bytes within a word, but the 64-bit pieces are
6211 ordered in a big endian fashion, and normal subreg's of TDmode are
6213 (mov<mode>_64bit_dm): Likewise.
6214 (movtd_64bit_nodm): Likewise.
6216 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6218 PR tree-optimization/60174
6219 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
6220 statement of an SSA_NAME that occurs in an abnormal PHI node.
6222 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6225 * final.c (SEEN_BB): Remove.
6226 (SEEN_NOTE, SEEN_EMITTED): Renumber.
6227 (final_scan_insn): Don't force_source_line on second
6228 NOTE_INSN_BASIC_BLOCK.
6230 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
6233 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
6234 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
6235 (type_natural_mode): Warn ABI change when %zmm register is not
6236 available for AVX512F vector value passing.
6238 2014-02-18 Kai Tietz <ktietz@redhat.com>
6241 * config/i386/i386.c (ix86_expand_prologue): Use value in
6242 rax register as displacement when restoring %r10 or %rax.
6243 Fix wrong offset when restoring both registers.
6245 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6247 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
6248 assertion with conditional return.
6250 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6251 Uros Bizjak <ubizjak@gmail.com>
6254 * config/i386/driver-i386.c (host_detect_local_cpu): If
6255 YMM state is not saved by the OS, also clear has_f16c. Move
6256 CPUID 0x80000001 handling before YMM state saving checking.
6258 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
6260 PR rtl-optimization/58960
6261 * haifa-sched.c (alloc_global_sched_pressure_data): New,
6262 factored out from ...
6263 (sched_init): ... here.
6264 (free_global_sched_pressure_data): New, factored out from ...
6265 (sched_finish): ... here.
6266 * sched-int.h (free_global_sched_pressure_data): Declare.
6267 * sched-rgn.c (nr_regions_initial): New static global.
6268 (haifa_find_rgns): Initialize it.
6269 (schedule_region): Disable sched-pressure for the newly
6272 2014-02-17 Richard Biener <rguenther@suse.de>
6274 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
6275 release SSA defs of pattern stmts.
6277 2014-02-17 Richard Biener <rguenther@suse.de>
6279 * tree-inline.c (expand_call_inline): Release the virtual
6280 operand defined by the call we are about to inline.
6282 2014-02-17 Richard Biener <rguenther@suse.de>
6284 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
6286 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
6287 Ilya Tocar <ilya.tocar@intel.com>
6289 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
6290 arguments order in builtin.
6291 (_mm512_permutexvar_epi64): Ditto.
6292 (_mm512_mask_permutexvar_epi64): Ditto
6293 (_mm512_maskz_permutexvar_epi32): Ditto
6294 (_mm512_permutexvar_epi32): Ditto
6295 (_mm512_mask_permutexvar_epi32): Ditto
6297 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6299 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
6300 (p8_vmrgow): Likewise.
6302 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6304 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
6307 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
6310 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
6311 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
6312 into 64-bit and 32-bit moves. On 64-bit moves, add support for
6313 using direct move instructions on ISA 2.07. Also adjust
6314 instruction length for 64-bit.
6315 (mov<mode>_64bit, TFmode/TDmode): Likewise.
6316 (mov<mode>_32bit, TFmode/TDmode): Likewise.
6318 2014-02-15 Alan Modra <amodra@gmail.com>
6322 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
6323 find_replacement on parts of insn rtl that might be reloaded.
6325 2014-02-15 Richard Biener <rguenther@suse.de>
6327 PR tree-optimization/60183
6328 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
6329 (tree_ssa_phiprop): Calculate and free post-dominators.
6331 2014-02-14 Jeff Law <law@redhat.com>
6333 PR rtl-optimization/60131
6334 * ree.c (get_extended_src_reg): New function.
6335 (combine_reaching_defs): Use it rather than assuming location of REG.
6336 (find_and_remove_re): Verify first operand of extension is
6337 a REG before adding the insns to the copy list.
6339 2014-02-14 Roland McGrath <mcgrathr@google.com>
6341 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
6342 * configure: Regenerated.
6343 * config.in: Regenerated.
6344 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
6345 instead of ASM_SHORT.
6347 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
6348 Richard Earnshaw <rearnsha@arm.com>
6350 PR rtl-optimization/59535
6351 * lra-constraints.c (process_alt_operands): Encourage alternative
6352 when unassigned pseudo class is superset of the alternative class.
6353 (inherit_reload_reg): Don't inherit when optimizing for code size.
6354 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
6355 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
6356 modes not less than 4 for Thumb1.
6358 2014-02-14 Kyle McMartin <kyle@redhat.com>
6361 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
6363 2014-02-14 Richard Biener <rguenther@suse.de>
6365 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
6366 (get_frame_arg): Drop the assert with langhook types_compatible_p.
6367 Do not strip INDIRECT_REFs.
6369 2014-02-14 Richard Biener <rguenther@suse.de>
6372 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
6373 DECL_FUNCTION_SPECIFIC_TARGET.
6374 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
6375 * tree-streamer-out.c (pack_ts_target_option): Remove.
6376 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
6377 (write_ts_function_decl_tree_pointers): Do not stream
6378 DECL_FUNCTION_SPECIFIC_TARGET.
6379 * tree-streamer-in.c (unpack_ts_target_option): Remove.
6380 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
6381 (lto_input_ts_function_decl_tree_pointers): Do not stream
6382 DECL_FUNCTION_SPECIFIC_TARGET.
6384 2014-02-14 Jakub Jelinek <jakub@redhat.com>
6386 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
6387 (get_initial_def_for_induction, vectorizable_induction): Ignore
6388 debug stmts when looking for exit_phi.
6389 (vectorizable_live_operation): Fix up condition.
6391 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6393 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
6394 nreverse() because it changes the content of original tree list.
6396 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6398 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
6399 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
6401 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6403 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
6404 GNU coding standards.
6406 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6409 * dwarf2out.c (gen_subprogram_die): Don't call
6410 add_calling_convention_attribute if subr_die is old_die.
6412 2014-02-13 Sharad Singhai <singhai@google.com>
6414 * doc/optinfo.texi: Fix order of nodes.
6416 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
6418 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
6419 operands[2], not operands[3].
6421 2014-02-13 Richard Biener <rguenther@suse.de>
6424 * doc/install.texi (ISL): Update recommended version to 0.12.2,
6425 mention the possibility of an in-tree build.
6426 (CLooG): Update recommended version to 0.18.1, mention the
6427 possibility of an in-tree build and clarify that the ISL
6428 bundled with CLooG does not work.
6430 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6433 * expr.c (compress_float_constant): If x is a hard register,
6434 extend into a pseudo and then move to x.
6436 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
6438 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
6439 caused by bad second argument to warning_at() with -mhotpatch and
6440 nested functions (e.g. with gfortran).
6442 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
6444 * opts.c (option_name): Remove "enabled by default" rider.
6446 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
6448 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
6450 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
6451 Uros Bizjak <ubizjak@gmail.com>
6454 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
6455 * configure: Regenerated.
6457 2014-02-12 Richard Biener <rguenther@suse.de>
6459 * vec.c (vec_prefix::calculate_allocation): Move as
6460 inline variant to vec.h.
6461 (vec_prefix::calculate_allocation_1): New out-of-line version.
6462 * vec.h (vec_prefix::calculate_allocation_1): Declare.
6463 (vec_prefix::m_has_auto_buf): Rename to ...
6464 (vec_prefix::m_using_auto_storage): ... this.
6465 (vec_prefix::calculate_allocation): Inline the easy cases
6466 and dispatch to calculate_allocation_1 which doesn't need the
6468 (va_heap::reserve): Use gcc_checking_assert.
6469 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
6470 m_using_auto_storage.
6471 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
6473 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
6474 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
6475 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
6477 2014-02-12 Richard Biener <rguenther@suse.de>
6479 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
6480 when we found a dependence.
6482 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
6484 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
6486 (maybe_fold_stmt): ... into this new function.
6487 * omp-low.c (lower_omp): Update comment.
6489 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
6492 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
6495 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
6497 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
6498 identifiers in comments.
6499 (cortexa53_extra_costs): Likewise.
6500 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
6501 (cortexa7_extra_costs): Likewise.
6502 (cortexa12_extra_costs): Likewise.
6503 (cortexa15_extra_costs): Likewise.
6504 (v7m_extra_costs): Likewise.
6506 2014-02-12 Richard Biener <rguenther@suse.de>
6509 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
6510 of posix_memalign being successful.
6511 (lower_stmt): Restrict lowering of posix_memalign to when
6512 -ftree-bit-ccp is enabled.
6514 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6516 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
6518 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
6520 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
6522 PR rtl-optimization/60116
6523 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
6524 other_insn once the combination has been validated.
6526 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
6529 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
6531 * ipa-devirt.c: Include demangle.h
6532 (odr_violation_reported): New static variable.
6533 (add_type_duplicate): Update odr_violations.
6534 (maybe_record_node): Add completep parameter; update it.
6535 (record_target_from_binfo): Add COMPLETEP parameter;
6536 update it as needed.
6537 (possible_polymorphic_call_targets_1): Likewise.
6538 (struct polymorphic_call_target_d): Add nonconstruction_targets;
6539 rename FINAL to COMPLETE.
6540 (record_targets_from_bases): Sanity check we found the binfo;
6541 fix COMPLETEP updating.
6542 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
6543 parameter, fix computing of COMPLETEP.
6544 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
6545 at LTO time do demangling.
6546 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
6547 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
6549 (gimple_get_virt_method_for_binfo): Likewise.
6550 * gimple-fold.h (gimple_get_virt_method_for_binfo,
6551 gimple_get_virt_method_for_vtable): Update prototypes.
6553 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
6556 * genautomata.c (add_presence_absence): Fix typo with
6557 {final_}presence_list.
6559 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
6562 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
6563 for VSX/Altivec vectors that land in GPR registers.
6565 2014-02-11 Richard Henderson <rth@redhat.com>
6566 Jakub Jelinek <jakub@redhat.com>
6569 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
6570 around drhs if type conversion to lacc->type is not useless.
6572 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6574 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
6576 (cortex-a57.cortex-a53): Likewise.
6577 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
6579 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6581 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
6584 2014-02-11 Renlin Li <Renlin.Li@arm.com>
6586 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
6587 add_options_for_arm_vfp3.
6589 2014-02-11 Jeff Law <law@redhat.com>
6592 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
6593 object with an undesirable mode.
6595 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6598 * config/i386/sol2-9.h: New file.
6599 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
6600 *-*-solaris2.9*): Use it.
6602 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6604 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
6605 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
6607 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6609 * config/microblaze/microblaze.c: Extend mcpu version format
6611 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
6613 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
6615 2014-02-10 Richard Henderson <rth@redhat.com>
6618 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
6619 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
6620 ms-abi vs -mno-accumulate-outgoing-args.
6621 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
6622 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
6625 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
6628 * cfgexpand.c (expand_asm_operands): Attach source location to
6629 ASM_INPUT rtx objects.
6630 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
6632 2014-02-10 Nick Clifton <nickc@redhat.com>
6634 * config/mn10300/mn10300.c (popcount): New function.
6635 (mn10300_expand_prologue): Include saved registers in stack usage
6638 2014-02-10 Jeff Law <law@redhat.com>
6641 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
6642 when changing the SET_DEST of a prior insn to avoid an input reload.
6644 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6646 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
6647 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
6648 -mcall-openbsd, or -mcall-linux.
6649 (CC1_ENDIAN_BIG_SPEC): Remove.
6650 (CC1_ENDIAN_LITTLE_SPEC): Remove.
6651 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
6652 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
6653 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
6654 and %cc1_endian_default.
6655 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
6657 2014-02-10 Richard Biener <rguenther@suse.de>
6659 PR tree-optimization/60115
6660 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
6661 MEM_REF handling. Properly verify that the accesses are not
6662 out of the objects bound.
6664 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6666 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
6669 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
6671 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
6672 proper constants and fix formatting.
6673 (possible_polymorphic_call_targets): Fix formatting.
6675 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
6676 Ilya Tocar <ilya.tocar@intel.com>
6678 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
6679 (_mm512_loadu_epi32): Renamed into...
6680 (_mm512_loadu_si512): This.
6681 (_mm512_storeu_epi32): Renamed into...
6682 (_mm512_storeu_si512): This.
6683 (_mm512_maskz_ceil_ps): Removed.
6684 (_mm512_maskz_ceil_pd): Ditto.
6685 (_mm512_maskz_floor_ps): Ditto.
6686 (_mm512_maskz_floor_pd): Ditto.
6687 (_mm512_floor_round_ps): Ditto.
6688 (_mm512_floor_round_pd): Ditto.
6689 (_mm512_ceil_round_ps): Ditto.
6690 (_mm512_ceil_round_pd): Ditto.
6691 (_mm512_mask_floor_round_ps): Ditto.
6692 (_mm512_mask_floor_round_pd): Ditto.
6693 (_mm512_mask_ceil_round_ps): Ditto.
6694 (_mm512_mask_ceil_round_pd): Ditto.
6695 (_mm512_maskz_floor_round_ps): Ditto.
6696 (_mm512_maskz_floor_round_pd): Ditto.
6697 (_mm512_maskz_ceil_round_ps): Ditto.
6698 (_mm512_maskz_ceil_round_pd): Ditto.
6699 (_mm512_expand_pd): Ditto.
6700 (_mm512_expand_ps): Ditto.
6701 * config/i386/i386.c (ix86_builtins): Remove
6702 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
6703 (bdesc_args): Ditto.
6704 * config/i386/predicates.md (const1256_operand): New.
6705 (const_1_to_2_operand): Ditto.
6706 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6707 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6708 (*avx512pf_gatherpf<mode>sf): Ditto.
6709 (avx512pf_gatherpf<mode>df): Ditto.
6710 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6711 (*avx512pf_gatherpf<mode>df): Ditto.
6712 (avx512pf_scatterpf<mode>sf): Ditto.
6713 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6714 (*avx512pf_scatterpf<mode>sf): Ditto.
6715 (avx512pf_scatterpf<mode>df): Ditto.
6716 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6717 (*avx512pf_scatterpf<mode>df): Ditto.
6718 (avx512f_expand<mode>): Removed.
6719 (<shift_insn><mode>3<mask_name>): Change predicate type.
6721 2014-02-08 Jakub Jelinek <jakub@redhat.com>
6723 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
6724 not at the end of datarefs vector use ordered_remove to avoid
6725 reordering datarefs vector.
6728 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
6729 mark local addressable non-static vars as GOVD_PRIVATE
6730 instead of GOVD_LOCAL.
6731 * omp-low.c (lower_omp_for): Move gimple_bind_vars
6732 and BLOCK_VARS of gimple_bind_block to new_stmt rather
6736 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
6737 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
6738 assume_aligned or alloc_align attributes.
6739 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
6740 arguments. Handle also assume_aligned and alloc_align attributes.
6741 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
6742 calls to functions with assume_aligned or alloc_align attributes.
6743 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
6745 2014-02-08 Terry Guo <terry.guo@arm.com>
6747 * doc/invoke.texi: Document ARM -march=armv7e-m.
6749 2014-02-08 Jakub Jelinek <jakub@redhat.com>
6751 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
6752 flag on __cilkrts_rethrow builtin.
6755 * ipa-cp.c (determine_versionability): Fail at -O0
6756 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
6757 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
6760 2014-02-04 Jakub Jelinek <jakub@redhat.com>
6763 * tree-inline.c (copy_forbidden): Fail for
6764 __attribute__((optimize (0))) functions.
6766 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
6768 * varpool.c: Include pointer-set.h.
6769 (varpool_remove_unreferenced_decls): Variables in other partitions
6770 will not be output; be however careful to not lose information
6773 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
6775 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
6776 lookup in the vtable constructor.
6778 2014-02-07 Jeff Law <law@redhat.com>
6781 * config/m68k/m68k.md (ashldi_extsi): Turn into a
6782 define_insn_and_split.
6784 * ipa-inline.c (inline_small_functions): Fix typos.
6786 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6788 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
6789 (s390_can_use_return_insn): Declare.
6790 * config/s390/s390.h (EPILOGUE_USES): Define.
6791 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
6793 (s390_chunkify_start): Handle return JUMP_LABELs.
6794 (s390_early_mach): Emit a main_pool instruction on the entry edge.
6795 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
6796 (s390_can_use_return_insn): New functions.
6797 (s390_fix_long_loop_prediction): Handle conditional returns.
6798 (TARGET_SET_UP_BY_PROLOGUE): Define.
6799 * config/s390/s390.md (ANY_RETURN): New code iterator.
6800 (*creturn, *csimple_return, return, simple_return): New patterns.
6802 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6804 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
6805 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
6806 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
6807 REG_CFA_RESTORE list when deciding not to restore a register.
6809 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6811 * config/s390/s390.c: Include tree-pass.h and context.h.
6812 (s390_early_mach): New function, split out from...
6813 (s390_emit_prologue): ...here.
6814 (pass_data_s390_early_mach): New pass structure.
6815 (pass_s390_early_mach): New class.
6816 (s390_option_override): Create and register early_mach pass.
6817 Move to end of file.
6819 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6821 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
6822 to match for the exit block.
6824 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6826 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
6827 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
6828 Reject misaligned operands.
6830 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6832 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
6834 2014-02-07 Richard Biener <rguenther@suse.de>
6837 * gimple-low.c (lower_builtin_posix_memalign): New function.
6838 (lower_stmt): Call it to lower posix_memalign in a way
6839 to make alignment info accessible.
6841 2014-02-07 Jakub Jelinek <jakub@redhat.com>
6844 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
6845 __builtin_setjmp_receiver.
6847 2014-02-07 Richard Biener <rguenther@suse.de>
6850 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
6851 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
6852 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
6853 Handle BUILT_IN_POSIX_MEMALIGN.
6854 (find_func_clobbers): Likewise.
6855 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
6856 (call_may_clobber_ref_p_1): Likewise.
6858 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6861 * ipa-devirt.c (record_target_from_binfo): Remove overactive
6864 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6867 * lto-cgraph.c (lto_output_node): Use
6868 symtab_get_symbol_partitioning_class.
6869 (lto_output_varpool_node): likewise.
6870 (symtab_get_symbol_partitioning_class): Move here from
6872 * cgraph.h (symbol_partitioning_class): Likewise.
6873 (symtab_get_symbol_partitioning_class): Declare.
6875 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6877 * ggc.h (ggc_internal_cleared_alloc): New macro.
6878 * vec.h (vec_safe_copy): Handle memory stats.
6879 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
6880 * target-globals.c (save_target_globals): Likewise.
6882 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6885 * expr.c (emit_move_resolve_push): Export; be bit more selective
6886 on when to clear alias set.
6887 * expr.h (emit_move_resolve_push): Declare.
6888 * function.h (struct function): Add tail_call_marked.
6889 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
6890 * config/i386/i386-protos.h (ix86_expand_push): Remove.
6891 * config/i386/i386.md (TImode move expander): De not call
6893 (FP push expanders): Preserve memory attributes.
6894 * config/i386/sse.md (push<mode>1): Remove.
6895 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
6896 (ix86_expand_push): Remove.
6897 * config/i386/mmx.md (push<mode>1): Remove.
6899 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6901 PR rtl-optimization/60030
6902 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
6903 lopart with paradoxical subreg before shifting it up by hprec.
6905 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6907 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
6908 Remove extra newline at end of file.
6909 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
6910 (arm_issue_rate): Handle cortexa57.
6911 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
6912 (cortex-a57.cortex-a53): Likewise.
6914 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6917 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
6918 don't record in REG_FRAME_RELATED_EXPR registers not set in that
6920 (arm_expand_prologue): Adjust all callers.
6921 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
6922 info, registers also at the lowest numbered registers side. Use
6923 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
6927 * var-tracking.c (adjust_mems): Before adding a SET to
6928 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
6930 2014-02-06 Alan Modra <amodra@gmail.com>
6933 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
6934 change SDmode to DDmode when lra_in_progress.
6936 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6939 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
6940 free_data_ref on the dr first, and before goto again also set dr
6941 to the next dr. For simd_lane_access, free old datarefs[i] before
6942 overwriting it. For get_vectype_for_scalar_type failure, don't
6943 free_data_ref if simd_lane_access.
6945 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
6948 * tree.h (opts_for_fn): New inline function.
6949 (opt_for_fn): Define.
6950 * config/i386/i386.c (ix86_function_regparm): Use
6951 opt_for_fn (decl, optimize) instead of optimize.
6953 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
6955 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
6956 for SYMBOL_REF in large memory model.
6958 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6960 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
6962 (cortex-a57): Likewise.
6963 (cortex-a57.cortex-a53): Likewise.
6965 2014-02-06 Yury Gribov <y.gribov@samsung.com>
6966 Kugan Vivekanandarajah <kuganv@linaro.org>
6968 * config/arm/arm.c (arm_vector_alignment_reachable): Check
6970 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
6972 2014-02-06 Richard Biener <rguenther@suse.de>
6974 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
6975 set_loop_copy and initialize_original_copy_tables.
6977 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
6979 * config/aarch64/aarch64-simd.md
6980 (aarch64_ashr_simddi): Change QI to SI.
6982 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
6983 Jakub Jelinek <jakub@redhat.com>
6986 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
6989 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6991 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
6992 CODE_FOR_altivec_vpku[hw]um to
6993 CODE_FOR_altivec_vpku[hw]um_direct.
6994 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
6995 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
6996 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
6997 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
6999 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7001 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
7002 generation for -maltivec=be.
7003 (altivec_vsumsws): Simplify redundant test.
7005 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7007 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
7008 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
7009 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
7010 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
7011 gen_altivec_vpkuwum.
7012 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
7014 (altivec_vpks<VI_char>ss): Likewise.
7015 (altivec_vpks<VI_char>us): Likewise.
7016 (altivec_vpku<VI_char>us): Likewise.
7017 (altivec_vpku<VI_char>um): Likewise.
7018 (altivec_vpku<VI_char>um_direct): New (copy of
7019 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
7021 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
7022 target is little endian and -maltivec=be is not specified.
7023 (*altivec_vupkhs<VU_char>_direct): New (copy of
7024 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
7025 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
7026 target is little endian and -maltivec=be is not specified.
7027 (*altivec_vupkls<VU_char>_direct): New (copy of
7028 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
7029 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
7030 little endian and -maltivec=be is not specified.
7031 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
7032 little endian and -maltivec=be is not specified.
7034 2014-02-05 Richard Henderson <rth@redhat.com>
7037 * combine-stack-adj.c: Revert r206943.
7038 * sched-int.h (struct deps_desc): Add last_args_size.
7039 * sched-deps.c (init_deps): Initialize it.
7040 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7041 contain REG_ARGS_SIZE notes.
7043 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7045 * lto-cgraph.c (asm_nodes_output): Make global.
7046 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7047 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7048 (driver_handle_option): Handle OPT_fwpa.
7050 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7053 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7054 a comment typo and formatting issue. If odr_hash hasn't been
7055 created, return vNULL and set *completep to false.
7058 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7059 bb with no successors.
7061 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7064 * doc/invoke.texi (-march): Clarify documentation for ARM.
7068 2014-02-05 Richard Biener <rguenther@suse.de>
7070 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
7071 when not vectorizing because of too many alias checks.
7072 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
7073 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
7075 2014-02-05 Nick Clifton <nickc@redhat.com>
7077 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
7078 accept extended registers in any mode when compiling for the MN10300.
7080 2014-02-05 Yury Gribov <y.gribov@samsung.com>
7082 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
7083 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
7084 sanitization attributes.
7085 (can_inline_edge_p): Likewise.
7086 (sanitize_attrs_match_for_inline_p): New function.
7088 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7090 * ipa-prop.c (detect_type_change): Shor circuit testing of
7091 type changes on THIS pointer.
7093 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
7096 * config/pa/pa.c (legitimize_tls_address): Return original address
7097 if not passed a SYMBOL_REF rtx.
7098 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
7100 (pa_emit_move_sequence): Simplify TLS source operands.
7101 (pa_legitimate_constant_p): Reject all TLS constants.
7102 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
7103 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
7105 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7107 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
7108 groups when we know they are controlled by LTO.
7109 * varasm.c (default_binds_local_p_1): If object is in other partition,
7110 it will be resolved locally.
7112 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7114 * config/host-linux.c (linux_gt_pch_use_address): Don't
7115 use SSIZE_MAX because it is not always defined.
7117 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
7120 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
7121 threshold for pseudo splitting.
7122 (update_ebb_live_info): Process call argument hard registers and
7123 hard registers from insn definition too.
7124 (max_small_class_regs_num): New constant.
7125 (inherit_in_ebb): Update live hard regs through EBBs. Update
7126 reloads_num only for small register classes. Don't split for
7129 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
7132 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
7135 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7137 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
7140 2014-02-04 Marek Polacek <polacek@redhat.com>
7142 * gdbinit.in (pel): Define.
7144 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7146 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
7149 2014-02-04 Richard Biener <rguenther@suse.de>
7152 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
7153 in function context local.
7154 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
7155 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
7156 similar to LTO_imported_decl_ref.
7158 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7160 PR tree-optimization/60002
7161 * cgraphclones.c (build_function_decl_skip_args): Clear
7164 PR tree-optimization/60023
7165 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
7166 false to gsi_replace.
7167 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
7168 has been in some EH region and vec_stmt could throw, add
7169 vec_stmt into the same EH region.
7170 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
7171 has no lhs, ignore it.
7172 * internal-fn.c (expand_MASK_LOAD): Likewise.
7175 * tree-inline.c (copy_forbidden): Fail for
7176 __attribute__((optimize (0))) functions.
7179 * omp-low.c (simd_clone_struct_copy): If from->inbranch
7180 is set, copy one less argument.
7181 (expand_simd_clones): Don't subtract clone_info->inbranch
7182 from simd_clone_struct_alloc argument.
7184 PR rtl-optimization/57915
7185 * recog.c (simplify_while_replacing): If all unary/binary/relational
7186 operation arguments are constant, attempt to simplify those.
7189 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
7190 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
7192 2014-02-04 Richard Biener <rguenther@suse.de>
7194 PR tree-optimization/60012
7195 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
7196 TBAA disambiguation to all DDRs.
7198 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7201 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
7202 (LINK_SPEC): Use it for -shared, -shared-libgcc.
7204 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7207 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
7209 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7211 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
7212 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
7214 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7217 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
7218 to figure out targets of polymorphic calls with known decl.
7219 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7220 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
7221 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
7222 (get_polymorphic_call_info): ... here.
7223 (get_polymorphic_call_info_from_invariant): New function.
7225 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7227 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
7228 lookup via vtable pointer; check for type consistency
7229 and turn inconsitent facts into UNREACHABLE.
7230 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7231 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
7232 type inconsistent querries; return UNREACHABLE instead.
7234 2014-02-03 Richard Henderson <rth@twiddle.net>
7237 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
7238 already processed this node.
7239 (normalize_one_pred_1): Pass along mark_set.
7240 (normalize_one_pred): Create and destroy a pointer_set_t.
7241 (normalize_one_pred_chain): Likewise.
7243 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
7245 PR gcov-profile/58602
7246 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
7248 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7251 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
7252 -fno-devirtualize; try to devirtualize by the knowledge of
7253 virtual table pointer given by aggregate propagation.
7254 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7255 (ipa_print_node_jump_functions): Dump also offset that
7256 is relevant for polymorphic calls.
7257 (determine_known_aggregate_parts): Add arg_type parameter; use it
7258 instead of determining the type from pointer type.
7259 (ipa_compute_jump_functions_for_edge): Update call of
7260 determine_known_aggregate_parts.
7261 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
7262 (gimple_get_virt_method_for_binfo): ... here; simplify using
7263 vtable_pointer_value_to_vtable.
7264 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
7265 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
7266 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
7267 (vtable_pointer_value_to_vtable): Break out from ...; handle also
7269 (vtable_pointer_value_to_binfo): ... here.
7270 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
7272 2014-02-03 Teresa Johnson <tejohnson@google.com>
7274 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
7275 redef of outer loop index variable.
7277 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
7281 * doc/extend.texi (Function Attributes): Typo.
7283 2014-02-03 Cong Hou <congh@google.com>
7285 PR tree-optimization/60000
7286 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
7287 if the vectorized statement is a store. A store statement can only
7288 appear at the end of pattern statements.
7290 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7292 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
7293 (ix86_option_override_internal): Default long double to 64-bit for
7294 32-bit Bionic and to 128-bit for 64-bit Bionic.
7296 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
7297 TARGET_LONG_DOUBLE_128 is true.
7298 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
7300 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
7301 (mlong-double-64): Negate -mlong-double-128.
7302 (mlong-double-128): New option.
7304 * config/i386/i386-c.c (ix86_target_macros): Define
7305 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
7307 * doc/invoke.texi: Document -mlong-double-128.
7309 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7311 PR rtl-optimization/60024
7312 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
7314 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
7316 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
7318 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7320 PR rtl-optimization/57662
7321 * sel-sched.c (code_motion_path_driver): Do not mark already not
7322 existing blocks in the visiting bitmap.
7324 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7326 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
7327 on the insn being emitted.
7329 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
7330 Will Deacon <will.deacon@arm.com>
7332 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
7334 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7336 * config/arm/arm-tables.opt: Regenerate.
7338 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7340 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
7341 for vector types other than V16QImode.
7342 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
7343 define_expand, and call altivec_expand_vec_perm_le when producing
7344 code with little endian element order.
7345 (*altivec_vperm_<mode>_internal): New insn having previous
7346 behavior of altivec_vperm_<mode>.
7347 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
7348 altivec_expand_vec_perm_le when producing code with little endian
7350 (*altivec_vperm_<mode>_uns_internal): New insn having previous
7351 behavior of altivec_vperm_<mode>_uns.
7353 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7355 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
7356 (altivec_vsumsws): Add handling for -maltivec=be with a little
7358 (altivec_vsumsws_direct): New.
7359 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
7360 gen_altivec_vsumsws.
7362 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7364 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
7365 vtable_pointer_value_to_binfo): New functions.
7366 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
7367 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
7369 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
7371 * config/nios2/nios2.md (load_got_register): Initialize GOT
7372 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
7373 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
7375 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7377 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
7378 preserverd by passthrough, do not propagate the type.
7380 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7382 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
7383 (mips_atomic_assign_expand_fenv): New function.
7384 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
7386 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7388 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
7389 (__builtin_mips_set_fcsr): Likewise.
7390 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
7391 MIPS_USI_FTYPE_VOID.
7392 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
7393 (mips16_expand_set_fcsr): Likewise.
7394 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
7395 (mips16_set_fcsr_stub): Likewise.
7396 (mips16_get_fcsr_one_only_stub): New class.
7397 (mips16_set_fcsr_one_only_stub): Likewise.
7398 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
7399 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
7400 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
7401 (hard_float): New availability predicate.
7402 (mips_builtins): Add get_fcsr and set_fcsr.
7403 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
7404 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
7405 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
7406 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
7407 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
7410 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7412 * config/mips/mips.c (mips_one_only_stub): New class.
7413 (mips_need_mips16_rdhwr_p): Replace with...
7414 (mips16_rdhwr_stub): ...this new variable.
7415 (mips16_stub_call_address): New function.
7416 (mips16_rdhwr_one_only_stub): New class.
7417 (mips_expand_thread_pointer): Use mips16_stub_call_address.
7418 (mips_output_mips16_rdhwr): Delete.
7419 (mips_finish_stub): New function.
7420 (mips_code_end): Use it to handle rdhwr stubs.
7422 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
7425 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
7426 when calculating size of integer atomic types.
7428 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
7430 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
7432 2014-02-01 Jakub Jelinek <jakub@redhat.com>
7434 PR tree-optimization/60003
7435 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
7436 * profile.c (branch_prob): Use gimple_call_builtin_p
7437 to check for BUILT_IN_SETJMP_RECEIVER.
7438 * tree-inline.c (copy_bb): Call notice_special_calls.
7440 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
7443 * lra-constraints.c (process_alt_operands): Update reload_sum only
7446 2014-01-31 Richard Henderson <rth@redhat.com>
7449 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
7450 until after else_eh is processed.
7452 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7454 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
7455 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
7456 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
7457 in smmintrin.h, remove them.
7458 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
7459 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
7460 * config/i386/i386.md (ROUND_SAE): Fix value.
7461 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
7462 (const48_operand): New.
7463 * config/i386/subst.md (round), (round_expand): Use
7464 const_4_or_8_to_11_operand.
7465 (round_saeonly), (round_saeonly_expand): Use const48_operand.
7467 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7469 * config/i386/constraints.md (Yk): Swap meaning with k.
7470 * config/i386/i386.md (movhi_internal): Change Yk to k.
7471 (movqi_internal): Ditto.
7472 (*k<logic><mode>): Ditto.
7475 (kandn<mode>): Ditto.
7476 (*<code>hi_1): Ditto.
7477 (*<code>qi_1): Ditto.
7478 (kxnor<mode>): Ditto.
7479 (kortestzhi): Ditto.
7480 (kortestchi): Ditto.
7482 (*one_cmplhi2_1): Ditto.
7483 (*one_cmplqi2_1): Ditto.
7484 * config/i386/sse.md (): Change k to Yk.
7485 (avx512f_load<mode>_mask): Ditto.
7486 (avx512f_blendm<mode>): Ditto.
7487 (avx512f_store<mode>_mask): Ditto.
7488 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
7489 (avx512f_storedqu<mode>_mask): Ditto.
7490 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
7492 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
7493 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
7494 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
7495 (avx512f_maskcmp<mode>3): Ditto.
7496 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
7497 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
7498 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
7499 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
7500 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
7501 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
7502 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
7503 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
7504 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
7505 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
7506 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
7507 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
7508 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
7509 (vec_extract_lo_<mode>_maskm): Ditto.
7510 (vec_extract_hi_<mode>_maskm): Ditto.
7511 (avx512f_vternlog<mode>_mask): Ditto.
7512 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
7513 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
7514 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
7515 (avx512f_<code>v8div16qi2_mask): Ditto.
7516 (avx512f_<code>v8div16qi2_mask_store): Ditto.
7517 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
7518 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
7519 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
7520 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
7521 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7522 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7523 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7524 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7525 (avx512cd_maskb_vec_dupv8di): Ditto.
7526 (avx512cd_maskw_vec_dupv16si): Ditto.
7527 (avx512f_vpermi2var<mode>3_maskz): Ditto.
7528 (avx512f_vpermi2var<mode>3_mask): Ditto.
7529 (avx512f_vpermi2var<mode>3_mask): Ditto.
7530 (avx512f_vpermt2var<mode>3_maskz): Ditto.
7531 (*avx512f_gathersi<mode>): Ditto.
7532 (*avx512f_gathersi<mode>_2): Ditto.
7533 (*avx512f_gatherdi<mode>): Ditto.
7534 (*avx512f_gatherdi<mode>_2): Ditto.
7535 (*avx512f_scattersi<mode>): Ditto.
7536 (*avx512f_scatterdi<mode>): Ditto.
7537 (avx512f_compress<mode>_mask): Ditto.
7538 (avx512f_compressstore<mode>_mask): Ditto.
7539 (avx512f_expand<mode>_mask): Ditto.
7540 * config/i386/subst.md (mask): Change k to Yk.
7541 (mask_scalar_merge): Ditto.
7544 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
7546 * doc/extend.texi (Vector Extensions): Document ?: in C++.
7548 2014-01-31 Richard Biener <rguenther@suse.de>
7551 * builtins.c (fold_builtin_memory_op): Make sure to not
7552 use a floating-point mode or a boolean or enumeral type for
7555 2014-01-30 DJ Delorie <dj@redhat.com>
7557 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
7558 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
7559 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
7560 whenever main() has an epilogue.
7562 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7564 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
7565 unused variable "field".
7566 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
7567 (vsx_mergeh_<mode>): Likewise.
7568 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
7569 (altivec_vmrghh): Likewise.
7570 (altivec_vmrghw): Likewise.
7571 (altivec_vmrglb): Likewise.
7572 (altivec_vmrglh): Likewise.
7573 (altivec_vmrglw): Likewise.
7574 (altivec_vspltb): Add missing uses.
7575 (altivec_vsplth): Likewise.
7576 (altivec_vspltw): Likewise.
7577 (altivec_vspltsf): Likewise.
7579 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7582 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
7583 frame related instructions.
7585 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
7587 PR rtl-optimization/59959
7588 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
7589 any reload of register whose subreg is invalid.
7591 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7593 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
7594 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
7595 Add missing return type - void.
7597 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7599 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
7600 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
7601 remove element index adjustment for endian (now handled in vsx.md
7603 (altivec_expand_vec_perm_const): Use
7604 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
7605 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
7606 (vsx_xxspltw_<mode>): Adjust element index for little endian.
7607 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
7608 define_expand and a new define_insn *altivec_vspltb_internal;
7609 adjust for -maltivec=be on a little endian target.
7610 (altivec_vspltb_direct): New.
7611 (altivec_vsplth): Divide into a define_expand and a new
7612 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
7613 little endian target.
7614 (altivec_vsplth_direct): New.
7615 (altivec_vspltw): Divide into a define_expand and a new
7616 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
7617 little endian target.
7618 (altivec_vspltw_direct): New.
7619 (altivec_vspltsf): Divide into a define_expand and a new
7620 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
7621 a little endian target.
7623 2014-01-30 Richard Biener <rguenther@suse.de>
7625 PR tree-optimization/59993
7626 * tree-ssa-forwprop.c (associate_pointerplus): Check we
7627 can propagate form the earlier stmt and avoid the transform
7628 when the intermediate result is needed.
7630 2014-01-30 Alangi Derick <alangiderick@gmail.com>
7632 * README.Portability: Fix typo.
7634 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
7636 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
7637 comparison_operator with ordered_comparison_operator.
7639 2014-01-30 Nick Clifton <nickc@redhat.com>
7641 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
7642 Rename to mn10300_store_multiple_regs.
7643 * config/mn10300/mn10300.c: Likewise.
7644 * config/mn10300/mn10300.md (store_movm): Fix typo: call
7645 store_multiple_regs.
7646 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
7647 Call mn10300_store_multiple_regs.
7649 2014-01-30 Nick Clifton <nickc@redhat.com>
7650 DJ Delorie <dj@redhat.com>
7652 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
7653 %fp 2 to keep registers after it properly word-aligned.
7654 (rl78_alloc_physical_registers_umul): Handle the case where both
7655 input operands are the same.
7657 2014-01-30 Richard Biener <rguenther@suse.de>
7659 PR tree-optimization/59903
7660 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
7663 2014-01-30 Jason Merrill <jason@redhat.com>
7666 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
7669 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
7671 2014-01-30 Richard Biener <rguenther@suse.de>
7673 PR tree-optimization/59951
7674 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
7676 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
7679 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
7680 SFmode to DFmode case.
7682 2014-01-29 DJ Delorie <dj@redhat.com>
7684 * config/msp430/msp430.opt (-minrt): New.
7685 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
7687 (ENDFILE_SPEC): Likewise.
7689 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
7691 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
7692 (estimate_function_body_sizes): Use it.
7694 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
7697 * dwarf2out.c (is_cxx_auto): New.
7698 (is_base_type): Use it.
7699 (gen_type_die_with_usage): Likewise.
7701 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7703 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
7704 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
7705 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
7706 -maltivec=be with LE targets.
7707 (vsx_mergeh_<mode>): Likewise.
7708 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
7709 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
7710 (altivec_vmrghb): Replace with define_expand and new
7711 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
7712 (altivec_vmrghb_direct): New define_insn.
7713 (altivec_vmrghh): Replace with define_expand and new
7714 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
7715 (altivec_vmrghh_direct): New define_insn.
7716 (altivec_vmrghw): Replace with define_expand and new
7717 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
7718 (altivec_vmrghw_direct): New define_insn.
7719 (*altivec_vmrghsf): Adjust for endianness.
7720 (altivec_vmrglb): Replace with define_expand and new
7721 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
7722 (altivec_vmrglb_direct): New define_insn.
7723 (altivec_vmrglh): Replace with define_expand and new
7724 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
7725 (altivec_vmrglh_direct): New define_insn.
7726 (altivec_vmrglw): Replace with define_expand and new
7727 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
7728 (altivec_vmrglw_direct): New define_insn.
7729 (*altivec_vmrglsf): Adjust for endianness.
7730 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
7731 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
7732 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
7733 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
7734 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
7735 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
7736 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
7737 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
7739 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
7741 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
7742 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
7745 2014-01-29 Richard Biener <rguenther@suse.de>
7747 PR tree-optimization/58742
7748 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
7749 associate_pointerplus_align.
7750 (associate_pointerplus_diff): New function.
7751 (associate_pointerplus): Likewise. Call associate_pointerplus_align
7752 and associate_pointerplus_diff.
7754 2014-01-29 Richard Biener <rguenther@suse.de>
7756 * lto-streamer.h (LTO_major_version): Bump to 3.
7757 (LTO_minor_version): Reset to 0.
7759 2014-01-29 Renlin Li <Renlin.Li@arm.com>
7761 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
7762 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
7763 (arm_file_start): Generate correct asm header for armv7ve.
7764 * config/arm/bpabi.h: Add multilib support for armv7ve.
7765 * config/arm/driver-arm.c: Change the architectures of cortex-a7
7766 and cortex-a15 to armv7ve.
7767 * config/arm/t-aprofile: Add multilib support for armv7ve.
7768 * doc/invoke.texi: Document -march=armv7ve.
7770 2014-01-29 Richard Biener <rguenther@suse.de>
7772 PR tree-optimization/58742
7773 * tree-ssa-forwprop.c (associate_plusminus): Return true
7774 if we changed sth, defer EH cleanup to ...
7775 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
7776 (simplify_mult): New function.
7778 2014-01-29 Jakub Jelinek <jakub@redhat.com>
7781 PR tree-optimization/59920
7782 * tree.c (build_common_builtin_nodes): Remove
7783 __builtin_setjmp_dispatcher initialization.
7784 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
7785 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
7786 instead of gsi_after_labels + manually skipping debug stmts.
7787 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
7788 ignore bbs with IFN_ABNORMAL_DISPATCHER.
7789 * tree-inline.c (copy_edges_for_bb): Remove
7790 can_make_abnormal_goto argument, instead add abnormal_goto_dest
7791 argument. Ignore computed_goto_p stmts. Don't call
7792 make_abnormal_goto_edges. If a call might need abnormal edges
7793 for non-local gotos, see if it already has an edge to
7794 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
7795 with true argument, don't do anything then, otherwise add
7796 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
7797 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
7799 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
7800 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
7801 (lower_stmt): Don't set data->calls_builtin_setjmp.
7802 (lower_builtin_setjmp): Adjust comment.
7803 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
7804 * tree-cfg.c (found_computed_goto): Remove.
7805 (factor_computed_gotos): Remove.
7806 (make_goto_expr_edges): Return bool, true for computed gotos.
7807 Don't call make_abnormal_goto_edges.
7808 (build_gimple_cfg): Don't set found_computed_goto, don't call
7809 factor_computed_gotos.
7810 (computed_goto_p): No longer static.
7811 (make_blocks): Don't set found_computed_goto.
7812 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
7813 (make_edges): If make_goto_expr_edges returns true, push bb
7814 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
7815 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
7816 vector. Record mapping between bbs and OpenMP regions if there
7817 are any, adjust make_gimple_omp_edges caller. Call
7818 handle_abnormal_edges.
7819 (make_abnormal_goto_edges): Remove.
7820 * tree-cfg.h (make_abnormal_goto_edges): Remove.
7821 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
7822 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
7823 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
7824 * internal-fn.def (ABNORMAL_DISPATCHER): New.
7825 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
7826 filling *region also set *region_idx to (*region)->entry->index.
7829 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
7830 For REGs set ORIGINAL_REGNO.
7832 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
7834 * doc/md.texi: Mention that a target shouldn't implement
7835 vec_widen_(s|u)mul_even/odd pair if it is less efficient
7838 2014-01-29 Jakub Jelinek <jakub@redhat.com>
7840 PR tree-optimization/59594
7841 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
7842 a copy of the datarefs vector rather than the vector itself.
7844 2014-01-28 Jason Merrill <jason@redhat.com>
7847 * dwarf2out.c (auto_die): New static.
7848 (gen_type_die_with_usage): Handle C++1y 'auto'.
7849 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
7852 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
7855 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
7856 (SPEC_X32): Likewise.
7857 (SPEC_64): Likewise.
7858 * config/i386/i386.c (ix86_option_override_internal): Turn off
7859 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
7861 (x86_file_start): Output .code16gcc for TARGET_16BIT.
7862 * config/i386/i386.h (TARGET_16BIT): New macro.
7863 (TARGET_16BIT_P): Likewise.
7864 * config/i386/i386.opt: Add m16.
7865 * doc/invoke.texi: Document -m16.
7867 2014-01-28 Jakub Jelinek <jakub@redhat.com>
7869 PR preprocessor/59935
7870 * input.c (location_get_source_line): Bail out on when line number
7871 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
7873 2014-01-28 Richard Biener <rguenther@suse.de>
7875 PR tree-optimization/58742
7876 * tree-ssa-forwprop.c (associate_plusminus): Handle
7877 pointer subtraction of the form (T)(P + A) - (T)P.
7879 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7881 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
7884 2014-01-28 Richard Biener <rguenther@suse.de>
7887 2014-01-28 Richard Biener <rguenther@suse.de>
7889 PR rtl-optimization/45364
7890 PR rtl-optimization/59890
7891 * var-tracking.c (local_get_addr_clear_given_value): Handle
7892 already cleared slot.
7893 (val_reset): Handle not allocated local_get_addr_cache.
7894 (vt_find_locations): Use post-order on the inverted CFG.
7896 2014-01-28 Richard Biener <rguenther@suse.de>
7898 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
7900 2014-01-28 Richard Biener <rguenther@suse.de>
7902 PR rtl-optimization/45364
7903 PR rtl-optimization/59890
7904 * var-tracking.c (local_get_addr_clear_given_value): Handle
7905 already cleared slot.
7906 (val_reset): Handle not allocated local_get_addr_cache.
7907 (vt_find_locations): Use post-order on the inverted CFG.
7909 2014-01-28 Alan Modra <amodra@gmail.com>
7911 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
7912 * configure.ac <recursive call for build != host>: Define
7913 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
7914 and LD_FOR_BUILD too.
7915 * configure: Regenerate.
7917 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
7919 * config/i386/i386.c (get_builtin_code_for_version): Separate
7920 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
7921 Broadwell from Haswell.
7923 2014-01-27 Steve Ellcey <sellcey@mips.com>
7925 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
7926 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
7927 * config/mips/mips.c (mips_option_override): Change setting
7929 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
7930 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
7931 Change from Mask to Var.
7933 2014-01-27 Jeff Law <law@redhat.com>
7935 * ipa-inline.c (inline_small_functions): Fix typo.
7937 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
7939 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
7940 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
7941 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
7942 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
7943 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
7944 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
7945 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
7946 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
7947 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
7948 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
7949 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
7950 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
7951 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
7952 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
7953 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
7954 (_mm512_storeu_epi64): Ditto.
7955 (_mm512_cmpge_epi32_mask): Ditto.
7956 (_mm512_cmpge_epu32_mask): Ditto.
7957 (_mm512_cmpge_epi64_mask): Ditto.
7958 (_mm512_cmpge_epu64_mask): Ditto.
7959 (_mm512_cmple_epi32_mask): Ditto.
7960 (_mm512_cmple_epu32_mask): Ditto.
7961 (_mm512_cmple_epi64_mask): Ditto.
7962 (_mm512_cmple_epu64_mask): Ditto.
7963 (_mm512_cmplt_epi32_mask): Ditto.
7964 (_mm512_cmplt_epu32_mask): Ditto.
7965 (_mm512_cmplt_epi64_mask): Ditto.
7966 (_mm512_cmplt_epu64_mask): Ditto.
7967 (_mm512_cmpneq_epi32_mask): Ditto.
7968 (_mm512_cmpneq_epu32_mask): Ditto.
7969 (_mm512_cmpneq_epi64_mask): Ditto.
7970 (_mm512_cmpneq_epu64_mask): Ditto.
7971 (_mm512_expand_pd): Ditto.
7972 (_mm512_expand_ps): Ditto.
7973 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
7974 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
7975 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
7976 * config/i386/i386.c (ix86_builtins): Add
7977 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
7978 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
7979 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
7980 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
7981 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
7982 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
7983 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
7984 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
7985 IX86_BUILTIN_PMOVUSQW512_MEM.
7986 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
7987 __builtin_ia32_pmovsqd512mem_mask,
7988 __builtin_ia32_pmovqd512mem_mask,
7989 __builtin_ia32_pmovusqw512mem_mask,
7990 __builtin_ia32_pmovsqw512mem_mask,
7991 __builtin_ia32_pmovqw512mem_mask,
7992 __builtin_ia32_pmovusdw512mem_mask,
7993 __builtin_ia32_pmovsdw512mem_mask,
7994 __builtin_ia32_pmovdw512mem_mask,
7995 __builtin_ia32_pmovqb512mem_mask,
7996 __builtin_ia32_pmovusqb512mem_mask,
7997 __builtin_ia32_pmovsqb512mem_mask,
7998 __builtin_ia32_pmovusdb512mem_mask,
7999 __builtin_ia32_pmovsdb512mem_mask,
8000 __builtin_ia32_pmovdb512mem_mask.
8001 (bdesc_args): Add __builtin_ia32_expanddf512,
8002 __builtin_ia32_expandsf512.
8003 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
8004 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
8005 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
8006 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
8007 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
8008 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
8009 (avx512f_<code>v8div16qi2_mask_store): This.
8010 (avx512f_expand<mode>): New.
8012 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
8014 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
8016 (_mm512_mask_prefetch_i64gather_pd): Ditto.
8017 (_mm512_prefetch_i32scatter_pd): Ditto.
8018 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
8019 (_mm512_prefetch_i64scatter_pd): Ditto.
8020 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
8021 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
8022 (_mm512_mask_prefetch_i64gather_ps): Ditto.
8023 (_mm512_prefetch_i32scatter_ps): Ditto.
8024 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
8025 (_mm512_prefetch_i64scatter_ps): Ditto.
8026 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
8027 * config/i386/i386-builtin-types.def: Define
8028 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
8029 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
8030 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
8031 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
8032 IX86_BUILTIN_SCATTERPFQPD.
8033 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
8034 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
8035 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
8036 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
8037 __builtin_ia32_scatterpfqps.
8038 (ix86_expand_builtin): Expand new built-ins.
8039 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8040 fix memory access data type.
8041 (*avx512pf_gatherpf<mode>_mask): Ditto.
8042 (*avx512pf_gatherpf<mode>): Ditto.
8043 (avx512pf_scatterpf<mode>): Ditto.
8044 (*avx512pf_scatterpf<mode>_mask): Ditto.
8045 (*avx512pf_scatterpf<mode>): Ditto.
8046 (GATHER_SCATTER_SF_MEM_MODE): New.
8047 (avx512pf_gatherpf<mode>df): Ditto.
8048 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8049 (*avx512pf_scatterpf<mode>df): Ditto.
8051 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8054 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8055 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8058 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8060 * common/config/arm/arm-common.c
8061 (arm_rewrite_mcpu): Handle multiple names.
8063 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8065 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8067 * gimple-builder.h (create_gimple_tmp): Delete.
8069 2014-01-27 Christian Bruel <christian.bruel@st.com>
8071 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
8074 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
8076 * config/pa/pa.md (call): Generate indirect long calls to non-local
8077 functions when outputing 32-bit code.
8078 (call_value): Likewise except for special call to buggy powf function.
8080 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
8081 portable runtime and PIC indirect calls.
8082 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
8083 and PIC call sequences. Use ldo instead of blr to set return register
8084 in PIC call sequence.
8086 2014-01-25 Walter Lee <walt@tilera.com>
8088 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
8089 avoid clobbering a live register.
8091 2014-01-25 Walter Lee <walt@tilera.com>
8093 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
8094 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
8095 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
8096 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
8098 2014-01-25 Walter Lee <walt@tilera.com>
8100 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
8101 arguments on even registers.
8102 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
8104 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
8105 (BIGGEST_ALIGNMENT): Ditto.
8106 (BIGGEST_FIELD_ALIGNMENT): Ditto.
8108 2014-01-25 Walter Lee <walt@tilera.com>
8110 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
8111 insns before bundling.
8112 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8114 2014-01-25 Walter Lee <walt@tilera.com>
8116 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
8117 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
8118 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
8120 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8122 * config/mips/constraints.md (kl): Delete.
8123 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
8124 define expands, using...
8125 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
8126 instructions for MIPS16.
8127 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
8128 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
8130 2014-01-25 Walter Lee <walt@tilera.com>
8132 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
8136 2014-01-25 Walter Lee <walt@tilera.com>
8138 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
8139 (TARGET_EXPAND_TO_RTL_HOOK): Define.
8141 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8143 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
8146 2014-01-25 Jakub Jelinek <jakub@redhat.com>
8148 * print-rtl.c (in_call_function_usage): New var.
8149 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
8150 EXPR_LIST mode as mode and not as reg note name.
8153 * cfgloopmanip.c (copy_loop_info): If
8154 loop->warned_aggressive_loop_optimizations, make sure
8155 the flag is set in target loop too.
8157 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
8159 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
8161 * builtins.def: Likewise.
8162 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
8163 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
8164 * ira.c (ira_setup_eliminable_regset): Likewise.
8165 * omp-low.c (gate_expand_omp): Likewise.
8166 (execute_lower_omp): Likewise.
8167 (diagnose_sb_0): Likewise.
8168 (gate_diagnose_omp_blocks): Likewise.
8169 (simd_clone_clauses_extract): Likewise.
8172 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8174 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
8175 correction for little endian...
8176 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
8179 2014-01-24 Jeff Law <law@redhat.com>
8181 PR tree-optimization/59919
8182 * tree-vrp.c (find_assert_locations_1): Do not register asserts
8183 for non-returning calls.
8185 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
8187 * common/config/aarch64/aarch64-common.c
8188 (aarch64_rewrite_mcpu): Handle multiple names.
8189 * config/aarch64/aarch64.h
8190 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8192 2014-01-24 Dodji Seketeli <dodji@redhat.com>
8194 * input.c (add_file_to_cache_tab): Handle the case where fopen
8197 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
8200 * config/i386/i386.md (pushsf splitter): Get stack adjustment
8201 from push operand if code of push isn't PRE_DEC.
8203 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8206 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
8207 -mquad-memory-atomic. Update -mquad-memory documentation to say
8208 it is only used for non-atomic loads/stores.
8210 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
8211 -mquad-memory or -mquad-memory-atomic switches.
8213 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
8214 -mquad-memory-atomic to ISA 2.07 support.
8216 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
8217 to separate support of normal quad word memory operations (ldq, stq)
8218 from the atomic quad word memory operations.
8220 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
8221 support to separate non-atomic quad word operations from atomic
8222 quad word operations. Disable non-atomic quad word operations in
8223 little endian mode so that we don't have to swap words after the
8224 load and before the store.
8225 (quad_load_store_p): Add comment about atomic quad word support.
8226 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
8227 options printed with -mdebug=reg.
8229 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
8230 -mquad-memory-atomic as the test for whether we have quad word
8231 atomic instructions.
8232 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
8233 or -mp8-vector are used, allow byte/half-word atomic operations.
8235 * config/rs6000/sync.md (load_lockedti): Insure that the address
8236 is a proper indexed or indirect address for the lqarx instruction.
8237 On little endian systems, swap the hi/lo registers after the lqarx
8239 (load_lockedpti): Use indexed_or_indirect_operand predicate to
8240 insure the address is valid for the lqarx instruction.
8241 (store_conditionalti): Insure that the address is a proper indexed
8242 or indirect address for the stqcrx. instruction. On little endian
8243 systems, swap the hi/lo registers before doing the stqcrx.
8245 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
8246 insure the address is valid for the stqcrx. instruction.
8248 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8249 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
8250 type of quad memory support is available.
8252 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
8255 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
8256 there is a danger of looping.
8258 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
8260 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8261 force flag_ira_loop_pressure if set via command line.
8263 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8265 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
8266 (ashr_simd): New builtin handling DI mode.
8267 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
8268 (aarch64_sshr_simddi): New match pattern.
8269 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
8270 (vshrd_n_s64): Likewise.
8271 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
8273 2014-01-23 Nick Clifton <nickc@redhat.com>
8275 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
8276 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
8277 favour of mcu specific scripts.
8278 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
8281 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8282 Alex Velenko <Alex.Velenko@arm.com>
8284 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
8285 (vaddv_s16): Likewise.
8286 (vaddv_s32): Likewise.
8287 (vaddv_u8): Likewise.
8288 (vaddv_u16): Likewise.
8289 (vaddv_u32): Likewise.
8290 (vaddvq_s8): Likewise.
8291 (vaddvq_s16): Likewise.
8292 (vaddvq_s32): Likewise.
8293 (vaddvq_s64): Likewise.
8294 (vaddvq_u8): Likewise.
8295 (vaddvq_u16): Likewise.
8296 (vaddvq_u32): Likewise.
8297 (vaddvq_u64): Likewise.
8298 (vaddv_f32): Likewise.
8299 (vaddvq_f32): Likewise.
8300 (vaddvq_f64): Likewise.
8301 (vmaxv_f32): Likewise.
8302 (vmaxv_s8): Likewise.
8303 (vmaxv_s16): Likewise.
8304 (vmaxv_s32): Likewise.
8305 (vmaxv_u8): Likewise.
8306 (vmaxv_u16): Likewise.
8307 (vmaxv_u32): Likewise.
8308 (vmaxvq_f32): Likewise.
8309 (vmaxvq_f64): Likewise.
8310 (vmaxvq_s8): Likewise.
8311 (vmaxvq_s16): Likewise.
8312 (vmaxvq_s32): Likewise.
8313 (vmaxvq_u8): Likewise.
8314 (vmaxvq_u16): Likewise.
8315 (vmaxvq_u32): Likewise.
8316 (vmaxnmv_f32): Likewise.
8317 (vmaxnmvq_f32): Likewise.
8318 (vmaxnmvq_f64): Likewise.
8319 (vminv_f32): Likewise.
8320 (vminv_s8): Likewise.
8321 (vminv_s16): Likewise.
8322 (vminv_s32): Likewise.
8323 (vminv_u8): Likewise.
8324 (vminv_u16): Likewise.
8325 (vminv_u32): Likewise.
8326 (vminvq_f32): Likewise.
8327 (vminvq_f64): Likewise.
8328 (vminvq_s8): Likewise.
8329 (vminvq_s16): Likewise.
8330 (vminvq_s32): Likewise.
8331 (vminvq_u8): Likewise.
8332 (vminvq_u16): Likewise.
8333 (vminvq_u32): Likewise.
8334 (vminnmv_f32): Likewise.
8335 (vminnmvq_f32): Likewise.
8336 (vminnmvq_f64): Likewise.
8338 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8340 * config/aarch64/aarch64-simd.md
8341 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
8342 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
8343 (*aarch64_mul3_elt<mode>): Likewise.
8344 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
8345 (*aarch64_mul3_elt_to_64v2df): Likewise.
8346 (*aarch64_mla_elt<mode>): Likewise.
8347 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8348 (*aarch64_mls_elt<mode>): Likewise.
8349 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8350 (*aarch64_fma4_elt<mode>): Likewise.
8351 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8352 (*aarch64_fma4_elt_to_64v2df): Likewise.
8353 (*aarch64_fnma4_elt<mode>): Likewise.
8354 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8355 (*aarch64_fnma4_elt_to_64v2df): Likewise.
8356 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
8357 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
8358 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8359 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8360 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
8361 (aarch64_sqdmull_lane<mode>_internal): Likewise.
8362 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
8364 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
8366 * config/aarch64/aarch64-simd.md
8367 (aarch64_be_checked_get_lane<mode>): New define_expand.
8368 * config/aarch64/aarch64-simd-builtins.def
8369 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
8370 New builtin definition.
8371 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
8372 Use new safe be builtin.
8374 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8376 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
8378 (aarch64_be_st1<mode>): Likewise.
8379 (aarch_ld1<VALL:mode>): Define_expand modified.
8380 (aarch_st1<VALL:mode>): Likewise.
8381 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
8382 (UNSPEC_ST1): Likewise.
8384 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
8386 * config/microblaze/microblaze.md: Add trap insn and attribute
8388 2014-01-23 Dodji Seketeli <dodji@redhat.com>
8390 PR preprocessor/58580
8391 * input.h (location_get_source_line): Take an additional line_size
8393 (void diagnostics_file_cache_fini): Declare new function.
8394 * input.c (struct fcache): New type.
8395 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
8396 New static constants.
8397 (diagnostic_file_cache_init, total_lines_num)
8398 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
8399 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
8400 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
8401 (get_next_line, read_next_line, goto_next_line, read_line_num):
8402 New static function definitions.
8403 (diagnostic_file_cache_fini): New function.
8404 (location_get_source_line): Take an additional output line_len
8405 parameter. Re-write using lookup_or_add_file_to_cache_tab and
8407 * diagnostic.c (diagnostic_finish): Call
8408 diagnostic_file_cache_fini.
8409 (adjust_line): Take an additional input parameter for the length
8410 of the line, rather than calculating it with strlen.
8411 (diagnostic_show_locus): Adjust the use of
8412 location_get_source_line and adjust_line with respect to their new
8413 signature. While displaying a line now, do not stop at the first
8414 null byte. Rather, display the zero byte as a space and keep
8415 going until we reach the size of the line.
8416 * Makefile.in: Add vec.o to OBJS-libcommon
8418 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8419 Ilya Tocar <ilya.tocar@intel.com>
8421 * config/i386/avx512fintrin.h (_mm512_kmov): New.
8422 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
8423 (__builtin_ia32_kmov16): Ditto.
8424 * config/i386/i386.md (UNSPEC_KMOV): New.
8427 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8429 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
8430 (_mm512_storeu_si512): Ditto.
8432 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
8435 * rtl.h (get_referenced_operands): Declare.
8436 * recog.c (get_referenced_operands): New function.
8437 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
8438 operands have been referenced when recording LO_SUM references.
8440 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
8442 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
8444 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8446 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
8447 Enable for generic and recent AMD targets.
8449 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8451 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
8452 ARG_SIZE note when adjustment was eliminated.
8454 2014-01-22 Jeff Law <law@redhat.com>
8456 PR tree-optimization/59597
8457 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
8458 in file. Accept new argument REGISTERING and use it to modify
8459 dump output appropriately.
8460 (register_jump_thread): Corresponding changes.
8461 (mark_threaded_blocks): Reinstate code to cancel unprofitable
8462 thread paths involving joiner blocks. Add code to dump cancelled
8463 jump threading paths.
8465 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
8467 PR rtl-optimization/59477
8468 * lra-constraints.c (inherit_in_ebb): Process call for living hard
8469 regs. Update reloads_num and potential_reload_hard_regs for all insns.
8471 2014-01-22 Tom Tromey <tromey@redhat.com>
8473 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
8475 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
8477 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8479 PR rtl-optimization/59896
8480 * lra-constraints.c (process_alt_operands): Check unused note for
8481 matched operands of insn with no output reloads.
8483 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
8485 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
8486 (mips_move_from_gpr_cost): Likewise.
8488 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8490 PR rtl-optimization/59858
8491 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
8492 ira_class_hard_regs_num.
8493 (process_alt_operands): Increase reject for dying matched operand.
8495 2014-01-21 Jakub Jelinek <jakub@redhat.com>
8498 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
8499 smaller than size, perform several stores or loads and stores
8500 at dst + count - size to store or copy all of size bytes, rather
8501 than just last modesize bytes.
8503 2014-01-20 DJ Delorie <dj@redhat.com>
8505 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
8506 that CLOBBERs are REGs before propogating their values.
8508 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
8511 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
8512 (cgraph_inline_failed_type): New function.
8513 * cgraph.h (DEFCIFCODE): Add type.
8514 (cgraph_inline_failed_type_t): New enum.
8515 (cgraph_inline_failed_type): New prototype.
8516 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
8517 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
8518 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
8519 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
8520 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
8521 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
8522 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
8523 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
8524 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
8525 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
8526 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
8527 OPTIMIZATION_MISMATCH.
8528 * tree-inline.c (expand_call_inline): Emit errors during
8529 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
8531 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8534 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
8535 mode attribute in insn output.
8537 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
8539 * output.h (output_constant): Delete.
8540 * varasm.c (output_constant): Make private.
8542 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
8544 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
8546 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8549 * tree.h (fold_builtin_strcat): New prototype.
8550 * builtins.c (fold_builtin_strcat): No longer static. Add len
8551 argument, if non-NULL, don't call c_strlen. Optimize
8552 directly into __builtin_memcpy instead of __builtin_strcpy.
8553 (fold_builtin_2): Adjust fold_builtin_strcat caller.
8554 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
8556 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8558 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8559 for SImode_address_operand operands, having only a REG argument.
8561 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
8563 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
8564 loader name using mbig-endian.
8565 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
8567 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8569 * doc/invoke.texi (-march): Clarify documentation for AArch64.
8573 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
8575 * config/aarch64/aarch64-protos.h
8576 (aarch64_cannot_change_mode_class_ptr): Declare.
8577 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
8578 aarch64_cannot_change_mode_class_ptr): New.
8579 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
8580 backend hook aarch64_cannot_change_mode_class.
8582 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8584 * common/config/aarch64/aarch64-common.c
8585 (aarch64_handle_option): Don't handle any option order logic here.
8586 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
8587 selected_cpu, warn on architecture version mismatch.
8588 (aarch64_override_options): Fix parsing order for option strings.
8590 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8591 Iain Sandoe <iain@codesourcery.com>
8594 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
8595 warning. Amend comment to reflect current functionality.
8597 2014-01-20 Richard Biener <rguenther@suse.de>
8600 * builtins.c (fold_builtin_strcat): Remove case better handled
8601 by tree-ssa-strlen.c.
8603 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
8605 * config/aarch64/aarch64.opt
8606 (mcpu, march, mtune): Make case-insensitive.
8608 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8611 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8612 if operands[1] is a REG or ZERO_EXTEND of a REG.
8614 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
8616 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
8618 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
8620 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
8621 long non-pic millicode calls.
8623 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8625 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
8627 2014-01-19 Kito Cheng <kito@0xlab.org>
8629 * builtins.c (expand_movstr): Check movstr expand done or fail.
8631 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8632 H.J. Lu <hongjiu.lu@intel.com>
8635 * config/i386/i386.md (*lea<mode>): Zero-extend return register
8636 to DImode for zero-extended addresses.
8638 2014-01-19 Jakub Jelinek <jakub@redhat.com>
8640 PR rtl-optimization/57763
8641 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
8642 on the new indirect jump_insn and increment LABEL_NUSES (label).
8644 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
8648 * config.gcc (x86_archs): New variable.
8649 (x86_64_archs): Likewise.
8650 (x86_cpus): Likewise.
8651 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
8652 --with-arch/--with-cpu= options.
8653 Support --with-arch=/--with-cpu={nehalem,westmere,
8654 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
8656 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8658 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
8659 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
8661 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8663 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
8665 2014-01-18 Jakub Jelinek <jakub@redhat.com>
8668 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
8669 clear cpp_get_options (parse_in)->warn_unused_macros for
8670 ix86_target_macros_internal with cpp_define.
8672 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
8674 * jump.c (delete_related_insns): Keep (use (insn))s.
8675 * reorg.c (redundant_insn): Check for barriers too.
8677 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8679 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
8681 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
8683 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
8684 call to $$dyncall when TARGET_LONG_CALLS is true.
8686 2014-01-17 Jeff Law <law@redhat.com>
8688 * ree.c (combine_set_extension): Temporarily disable test for
8689 changing number of hard registers.
8691 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
8694 * ipa-inline-analysis.c (inline_free_summary):
8695 Do not free summary of aliases.
8697 2014-01-17 Jakub Jelinek <jakub@redhat.com>
8700 * gimplify.c (gimplify_expr): Use create_tmp_var
8701 instead of create_tmp_var_raw. If cond doesn't have
8702 integral type, don't add the IFN_ANNOTATE builtin at all.
8704 2014-01-17 Martin Jambor <mjambor@suse.cz>
8707 * ipa-cp.c (prev_edge_clone): New variable.
8708 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
8709 Also resize prev_edge_clone vector.
8710 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
8711 (ipcp_edge_removal_hook): New function.
8712 (ipcp_driver): Register ipcp_edge_removal_hook.
8714 2014-01-17 Andrew Pinski <apinski@cavium.com>
8715 Steve Ellcey <sellcey@mips.com>
8718 * config/mips/mips.c (mips_print_operand): Check operand mode instead
8721 2014-01-17 Jeff Law <law@redhat.com>
8724 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
8725 so that pass_ccp runs first.
8727 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8729 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
8730 (ix86_adjust_cost): Use !TARGET_XXX.
8731 (do_reorder_for_imul): Likewise.
8732 (swap_top_of_ready_list): Likewise.
8733 (ix86_sched_reorder): Likewise.
8735 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8737 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
8738 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
8739 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
8740 (intel_memset): New. Duplicate slm_memset.
8741 (intel_cost): New. Duplicate slm_cost.
8742 (m_INTEL): New macro.
8743 (processor_target_table): Add "intel".
8744 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
8745 with PROCESSOR_INTEL for "intel".
8746 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
8747 PROCESSOR_SILVERMONT.
8748 (ix86_issue_rate): Likewise.
8749 (ix86_adjust_cost): Likewise.
8750 (ia32_multipass_dfa_lookahead): Likewise.
8751 (swap_top_of_ready_list): Likewise.
8752 (ix86_sched_reorder): Likewise.
8753 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
8754 instead of TARGET_OPT_AGU.
8755 * config/i386/i386.h (TARGET_INTEL): New.
8756 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
8757 (processor_type): Add PROCESSOR_INTEL.
8758 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
8759 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
8761 2014-01-17 Marek Polacek <polacek@redhat.com>
8764 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
8767 2014-01-17 Richard Biener <rguenther@suse.de>
8769 PR tree-optimization/46590
8770 * opts.c (default_options_table): Add entries for
8771 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
8772 all enabled at -O1 but not for -Og.
8773 * common.opt (fbranch-count-reg): Remove Init(1).
8774 (fmove-loop-invariants): Likewise.
8775 (ftree-pta): Likewise.
8777 2014-01-17 Jakub Jelinek <jakub@redhat.com>
8779 * config/i386/i386.c (ix86_data_alignment): For compatibility with
8780 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
8781 decls to at least the GCC 4.8 used alignments.
8784 * tree-nested.c (convert_nonlocal_reference_stmt,
8785 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
8786 of GIMPLE_BIND stmts, adjust associated decls.
8788 2014-01-17 Richard Biener <rguenther@suse.de>
8790 PR tree-optimization/46590
8791 * vec.h (vec<>::bseach): New member function implementing
8792 binary search according to C89 bsearch.
8793 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
8794 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
8795 bitmap pointer again. Make accesses_in_loop a flat array.
8796 (mem_ref_obstack): New global.
8797 (outermost_indep_loop): Adjust for mem_ref->stored changes.
8798 (mark_ref_stored): Likewise.
8799 (ref_indep_loop_p_2): Likewise.
8800 (set_ref_stored_in_loop): New helper function.
8801 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
8802 (memref_free): Adjust.
8803 (record_mem_ref_loc): Simplify.
8804 (gather_mem_refs_stmt): Adjust.
8805 (sort_locs_in_loop_postorder_cmp): New function.
8806 (analyze_memory_references): Sort accesses_in_loop after
8807 loop postorder number.
8808 (find_ref_loc_in_loop_cmp): New function.
8809 (for_all_locs_in_loop): Find relevant cluster of locs in
8810 accesses_in_loop and iterate without recursion.
8811 (execute_sm): Avoid uninit warning.
8812 (struct ref_always_accessed): Simplify.
8813 (ref_always_accessed::operator ()): Likewise.
8814 (ref_always_accessed_p): Likewise.
8815 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
8816 loop postorder numbers here.
8817 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
8820 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
8823 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
8824 on decls for which assemble_alias has been called.
8826 2014-01-17 Nick Clifton <nickc@redhat.com>
8828 * config/msp430/msp430.opt: (mcpu): New option.
8829 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
8830 (msp430_option_override): Parse target_cpu. If the MCU name
8831 matches a generic string, clear target_mcu.
8832 (msp430_attr): Allow numeric interrupt values up to 63.
8833 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
8834 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
8836 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
8838 * config/msp430/msp430.md (popm): Use %J rather than %I.
8839 (addsi3): Use msp430_nonimmediate_operand for operand 2.
8840 (addhi_cy_i): Use immediate_operand for operand 2.
8841 * doc/invoke.texi: Document -mcpu option.
8843 2014-01-17 Richard Biener <rguenther@suse.de>
8845 PR rtl-optimization/38518
8846 * df.h (df_analyze_loop): Declare.
8847 * df-core.c: Include cfgloop.h.
8848 (df_analyze_1): Split out main part of df_analyze.
8849 (df_analyze): Adjust.
8850 (loop_inverted_post_order_compute): New function.
8851 (loop_post_order_compute): Likewise.
8852 (df_analyze_loop): New function avoiding whole-function
8854 * loop-invariant.c (find_defs): Use df_analyze_loop.
8855 (find_invariants): Adjust.
8856 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
8858 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
8860 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
8861 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
8863 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
8865 * ipa-ref.c (ipa_remove_stmt_references): Fix references
8866 traversal when removing references.
8868 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
8871 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
8873 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
8876 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
8877 pushing a reload for an autoinc when we had previously reloaded an
8878 inner part of the address.
8880 2014-01-16 Jakub Jelinek <jakub@redhat.com>
8882 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
8884 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
8885 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
8886 when not giving up or versioning for alias only because of
8888 (vect_analyze_data_ref_dependences): Set to true.
8889 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
8891 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
8892 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
8896 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
8899 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
8900 operand 0 predicate for gathers, use a new pseudo as subtarget.
8902 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
8905 * lra-constraints.c (process_alt_operands): Add printing debug info.
8906 Check absence of input/output reloads for matched operands too.
8908 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
8910 PR rtl-optimization/59835
8911 * ira.c (ira_init_register_move_cost): Increase cost for
8914 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
8916 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
8918 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
8921 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
8922 non-register objects. Use gen_(high/low)part more consistently.
8925 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8928 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
8929 endian support, remove tests for WORDS_BIG_ENDIAN.
8930 (p8_mfvsrd_3_<mode>): Likewise.
8931 (reload_gpr_from_vsx<mode>): Likewise.
8932 (reload_gpr_from_vsxsf): Likewise.
8933 (p8_mfvsrd_4_disf): Likewise.
8935 2014-01-16 Richard Biener <rguenther@suse.de>
8937 PR rtl-optimization/46590
8938 * lcm.c (compute_antinout_edge): Use postorder iteration.
8939 (compute_laterin): Use inverted postorder iteration.
8941 2014-01-16 Nick Clifton <nickc@redhat.com>
8944 * varasm.c (output_constant): Return the number of bytes actually
8946 (output_constructor_array_range): Update the field size with the
8947 number of bytes emitted by output_constant.
8948 (output_constructor_regular_field): Likewise. Also do not
8949 complain if the total number of bytes emitted is now greater
8950 than the expected fieldpos.
8951 * output.h (output_constant): Update prototype and descriptive comment.
8953 2014-01-16 Marek Polacek <polacek@redhat.com>
8956 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
8957 it is error_mark_node.
8959 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
8961 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
8962 VALID_AVX256_REG_OR_OI_MODE.
8964 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
8966 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
8967 current procedure should be profiled.
8969 2014-01-15 Andrew Pinski <apinski@cavium.com>
8971 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
8972 of moving from/to the STACK_REG register class.
8974 2014-01-15 Richard Henderson <rth@redhat.com>
8977 * reginfo.c (global_regs_decl): Globalize.
8978 * rtl.h (global_regs_decl): Declare.
8979 * ira.c (do_reload): Diagnose frame_pointer_needed and it
8980 reserved via global_regs.
8982 2014-01-15 Teresa Johnson <tejohnson@google.com>
8984 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
8986 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8988 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
8989 and vmulosh rather than call gen_vec_widen_smult_*.
8990 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
8991 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
8992 (vec_widen_smult_even_v16qi): Likewise.
8993 (vec_widen_umult_even_v8hi): Likewise.
8994 (vec_widen_smult_even_v8hi): Likewise.
8995 (vec_widen_umult_odd_v16qi): Likewise.
8996 (vec_widen_smult_odd_v16qi): Likewise.
8997 (vec_widen_umult_odd_v8hi): Likewise.
8998 (vec_widen_smult_odd_v8hi): Likewise.
8999 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
9000 vmuloub rather than call gen_vec_widen_umult_*.
9001 (vec_widen_umult_lo_v16qi): Likewise.
9002 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
9003 vmulosb rather than call gen_vec_widen_smult_*.
9004 (vec_widen_smult_lo_v16qi): Likewise.
9005 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
9006 rather than call gen_vec_widen_umult_*.
9007 (vec_widen_umult_lo_v8hi): Likewise.
9008 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
9009 rather than call gen_vec_widen_smult_*.
9010 (vec_widen_smult_lo_v8hi): Likewise.
9012 2014-01-15 Jeff Law <law@redhat.com>
9014 PR tree-optimization/59747
9015 * ree.c (find_and_remove_re): Properly handle case where a second
9016 eliminated extension requires widening a copy created for elimination
9017 of a prior extension.
9018 (combine_set_extension): Ensure that the number of hard regs needed
9019 for a destination register does not change when we widen it.
9021 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
9023 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
9024 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
9025 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
9026 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
9027 (avr-*-rtems*): Likewise.
9028 (bfin*-rtems*): Likewise.
9029 (moxie-*-rtems*): Likewise.
9030 (h8300-*-rtems*): Likewise.
9031 (i[34567]86-*-rtems*): Likewise.
9032 (lm32-*-rtems*): Likewise.
9033 (m32r-*-rtems*): Likewise.
9034 (m68k-*-rtems*): Likewise.
9035 (microblaze*-*-rtems*): Likewise.
9036 (mips*-*-rtems*): Likewise.
9037 (powerpc-*-rtems*): Likewise.
9038 (sh-*-rtems*): Likewise.
9039 (sparc-*-rtems*): Likewise.
9040 (sparc64-*-rtems*): Likewise.
9041 (v850-*-rtems*): Likewise.
9042 (m32c-*-rtems*): Likewise.
9044 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9046 PR rtl-optimization/59511
9047 * ira.c (ira_init_register_move_cost): Use memory costs for some
9048 cases of register move cost calculations.
9049 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9050 instead of BB frequency.
9051 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9052 * lra-assigns.c (find_hard_regno_for): Ditto.
9054 2014-01-15 Richard Biener <rguenther@suse.de>
9056 PR tree-optimization/59822
9057 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9058 (vectorizable_load): Use it to hoist defs of uses of invariant
9059 loads out of the loop.
9061 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9062 Kugan Vivekanandarajah <kuganv@linaro.org>
9065 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9068 2014-01-15 Richard Biener <rguenther@suse.de>
9070 PR rtl-optimization/59802
9071 * lcm.c (compute_available): Use inverted postorder to seed
9072 the initial worklist.
9074 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9077 * config/s390/s390.c (s390_preferred_reload_class): Don't return
9078 ADDR_REGS for invalid symrefs in non-PIC code.
9080 2014-01-15 Jakub Jelinek <jakub@redhat.com>
9083 * builtins.c (determine_block_size): Initialize *probable_max_size
9084 even if len_rtx is CONST_INT.
9086 2014-01-14 Andrew Pinski <apinski@cavium.com>
9088 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
9089 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
9090 (cortexa53_tunings): Likewise.
9091 (aarch64_sched_issue_rate): New function.
9092 (TARGET_SCHED_ISSUE_RATE): Define.
9094 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9096 * ira-costs.c (find_costs_and_classes): Add missed
9097 ira_init_register_move_cost_if_necessary.
9099 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9102 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
9104 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
9107 * config/i386/i386.c (type_natural_mode): Add a bool parameter
9108 to indicate if type is used for function return value. Warn ABI
9109 change if the vector mode isn't available for function return value.
9110 (ix86_function_arg_advance): Pass false to type_natural_mode.
9111 (ix86_function_arg): Likewise.
9112 (ix86_gimplify_va_arg): Likewise.
9113 (function_arg_32): Don't warn ABI change.
9114 (ix86_function_value): Pass true to type_natural_mode.
9115 (ix86_return_in_memory): Likewise.
9116 (ix86_struct_value_rtx): Removed.
9117 (TARGET_STRUCT_VALUE_RTX): Likewise.
9119 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9121 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
9122 converting a conditional jump into a conditional return.
9124 2014-01-14 Richard Biener <rguenther@suse.de>
9126 PR tree-optimization/58921
9127 PR tree-optimization/59006
9128 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
9129 hoisting invariant stmts.
9130 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
9131 invariant loads on the preheader edge if possible.
9133 2014-01-14 Joey Ye <joey.ye@arm.com>
9135 * doc/plugin.texi (Building GCC plugins): Update to C++.
9137 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
9139 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
9140 (_mm_rcp28_round_ss): Ditto.
9141 (_mm_rsqrt28_round_sd): Ditto.
9142 (_mm_rsqrt28_round_ss): Ditto.
9143 (_mm_rcp28_sd): Ditto.
9144 (_mm_rcp28_ss): Ditto.
9145 (_mm_rsqrt28_sd): Ditto.
9146 (_mm_rsqrt28_ss): Ditto.
9147 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
9148 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
9149 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
9150 (IX86_BUILTIN_RCP28SD): Ditto.
9151 (IX86_BUILTIN_RCP28SS): Ditto.
9152 (IX86_BUILTIN_RSQRT28SD): Ditto.
9153 (IX86_BUILTIN_RSQRT28SS): Ditto.
9154 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
9155 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
9156 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
9157 (ix86_expand_special_args_builtin): Expand new FTYPE.
9158 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
9159 (srcp14<mode>): Make insn unary.
9160 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
9161 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
9162 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
9163 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
9164 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
9165 Fix rounding: make it SAE only.
9166 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9168 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9170 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
9171 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
9172 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
9173 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
9174 (round_saeonly_mask_scalar_operand4): Ditto.
9175 (round_saeonly_mask_scalar_op3): Ditto.
9176 (round_saeonly_mask_scalar_op4): Ditto.
9178 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9180 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9181 Implement -maltivec=be for vec_insert and vec_extract.
9183 2014-01-10 DJ Delorie <dj@redhat.com>
9185 * config/msp430/msp430.md (call_internal): Don't allow memory
9186 references with SP as the base register.
9187 (call_value_internal): Likewise.
9188 * config/msp430/constraints.md (Yc): New. For memory references
9189 that don't use SP as a base register.
9191 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
9192 "an integer without a # prefix"
9193 * config/msp430/msp430.md (epilogue_helper): Use it.
9195 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9198 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
9199 AVX512F gather builtins.
9200 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
9201 on gather decls with INTEGER_TYPE masktype.
9202 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
9203 directly into the builtin rather than hoisting it before loop.
9205 PR tree-optimization/59387
9206 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
9207 (scev_const_prop): If folded_casts and type has undefined overflow,
9208 use force_gimple_operand instead of force_gimple_operand_gsi and
9209 for each added stmt if it is assign with
9210 arith_code_with_undefined_signed_overflow, call
9211 rewrite_to_defined_overflow.
9212 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
9213 gimple-fold.h instead.
9214 (arith_code_with_undefined_signed_overflow,
9215 rewrite_to_defined_overflow): Moved to ...
9216 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
9217 rewrite_to_defined_overflow): ... here. No longer static.
9218 Include gimplify-me.h.
9219 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
9220 rewrite_to_defined_overflow): New prototypes.
9222 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9224 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
9226 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
9228 * builtins.c (get_object_alignment_2): Minor tweak.
9229 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
9231 2014-01-13 Christian Bruel <christian.bruel@st.com>
9233 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
9234 optimized non constant lengths.
9236 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9239 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
9240 load as __atomic_load_N if possible.
9242 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9244 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
9246 (rs6000_expand_builtin): Adjust call.
9248 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9251 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
9252 * config/rs6000/rs6000.c: Include target-globals.h.
9253 (rs6000_set_current_function): Instead of doing target_reinit
9254 unconditionally, use save_target_globals_default_opts and
9255 restore_target_globals.
9257 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
9259 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
9260 (rs6000_expand_builtin): Handle mffs and mtfsf.
9261 (rs6000_init_builtins): Define mffs and mtfsf.
9262 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
9263 (rs6000_mffs): New pattern.
9264 (rs6000_mtfsf): New pattern.
9266 2014-01-11 Bin Cheng <bin.cheng@arm.com>
9268 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
9269 Start narrowing with START. Apply candidate-use pair
9270 and check overall cost in narrowing.
9271 (iv_ca_prune): Pass new argument.
9273 2014-01-10 Jeff Law <law@redhat.com>
9276 * ree.c (combine_reaching_defs): Ensure the defining statement
9277 occurs before the extension when optimizing extensions with
9278 different source and destination hard registers.
9280 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9283 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
9284 vtables into the type inheritance graph.
9286 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9288 PR rtl-optimization/59754
9289 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
9290 modes in the REGNO != REGNO case.
9292 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9294 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
9296 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9298 PR tree-optimization/59745
9299 * tree-predcom.c (tree_predictive_commoning_loop): Call
9300 free_affine_expand_cache if giving up because components is NULL.
9302 * target-globals.c (save_target_globals): Allocate < 4KB structs using
9303 GC in payload of target_globals struct instead of allocating them on
9304 the heap and the larger structs separately using GC.
9305 * target-globals.h (struct target_globals): Make regs, hard_regs,
9306 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
9307 of GTY((skip)) and change type to void *.
9308 (reset_target_globals): Cast loads from those fields to corresponding
9311 2014-01-10 Steve Ellcey <sellcey@mips.com>
9314 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
9315 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
9316 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
9318 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
9321 * aarch64-modes.def (CC_Zmode): New flags mode.
9322 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
9323 represents an equality.
9324 (aarch64_get_condition_code): Handle CC_Zmode.
9325 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
9327 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9329 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
9330 extraction in good case.
9332 2014-01-10 Richard Biener <rguenther@suse.de>
9334 PR tree-optimization/59374
9335 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
9336 checking after SLP discovery. Mark stmts not participating
9337 in any SLP instance properly.
9339 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9341 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
9342 when handling a SET rtx.
9344 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9346 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
9347 (cortex-a57): Likewise.
9348 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
9350 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9352 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
9353 non-iwmmxt builtins.
9355 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9359 * ipa-devirt.c record_target_from_binfo): Take as argument
9360 stack of binfos and lookup matching one for virtual inheritance.
9361 (possible_polymorphic_call_targets_1): Update.
9363 2014-01-10 Huacai Chen <chenhc@lemote.com>
9365 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
9366 kernel strings for Loongson-2E/2F/3A.
9368 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9371 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
9372 is_gimple_call before calling gimple_call_internal_p.
9374 2014-01-09 Steve Ellcey <sellcey@mips.com>
9376 * Makefile.in (TREE_FLOW_H): Remove.
9377 (TREE_SSA_H): Add file names from tree-flow.h.
9378 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
9379 * tree.h: Remove tree-flow.h reference.
9380 * hash-table.h: Remove tree-flow.h reference.
9381 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
9382 reference with tree-ssa-loop.h.
9384 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9386 * doc/invoke.texi: Add -maltivec={be,le} options, and document
9387 default element-order behavior for -maltivec.
9388 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
9389 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
9390 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
9391 when targeting big endian, at least for now.
9392 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
9394 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9397 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
9398 var satisfies use_register_for_decl, just take into account type
9399 alignment, rather than decl alignment.
9401 PR tree-optimization/59622
9402 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
9403 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
9404 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
9405 Don't devirtualize for inplace at all. For targets.length () == 1,
9406 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
9408 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9410 * config/i386/i386.md (cpu): Remove the unused btver1.
9412 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9414 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
9416 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9419 * tree-core.h (struct target_globals): New forward declaration.
9420 (struct tree_target_option): Add globals field.
9421 * tree.h (TREE_TARGET_GLOBALS): Define.
9422 (prepare_target_option_nodes_for_pch): New prototype.
9423 * target-globals.h (struct target_globals): Define even if
9425 * tree.c (prepare_target_option_node_for_pch,
9426 prepare_target_option_nodes_for_pch): New functions.
9427 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
9428 * config/i386/i386.c: Include target-globals.h.
9429 (ix86_set_current_function): Instead of doing target_reinit
9430 unconditionally, use save_target_globals_default_opts and
9431 restore_target_globals.
9433 2014-01-09 Richard Biener <rguenther@suse.de>
9435 PR tree-optimization/59715
9436 * tree-cfg.h (split_critical_edges): Declare.
9437 * tree-cfg.c (split_critical_edges): Export.
9438 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
9440 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
9442 * cfgexpand.c (expand_stack_vars): Optionally disable
9443 asan stack protection.
9444 (expand_used_vars): Likewise.
9445 (partition_stack_vars): Likewise.
9446 * asan.c (asan_emit_stack_protection): Optionally disable
9447 after return stack usage.
9448 (instrument_derefs): Optionally disable memory access instrumentation.
9449 (instrument_builtin_call): Likewise.
9450 (instrument_strlen_call): Likewise.
9451 (asan_protect_global): Optionally disable global variables protection.
9452 * doc/invoke.texi: Added doc for new options.
9453 * params.def: Added new options.
9454 * params.h: Likewise.
9456 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9458 PR rtl-optimization/59724
9459 * ifcvt.c (cond_exec_process_if_block): Don't call
9460 flow_find_head_matching_sequence with 0 longest_match.
9461 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
9462 non-active insns if !stop_after.
9463 (try_head_merge_bb): Revert 2014-01-07 changes.
9465 2014-01-08 Jeff Law <law@redhat.com>
9467 * ree.c (get_sub_rtx): New function, extracted from...
9468 (merge_def_and_ext): Here.
9469 (combine_reaching_defs): Use get_sub_rtx.
9471 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
9473 * cgraph.h (varpool_variable_node): Do not choke on null node.
9475 2014-01-08 Catherine Moore <clm@codesourcery.com>
9477 * config/mips/mips.md (simple_return): Attempt to use JRC
9479 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
9481 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9483 PR rtl-optimization/59137
9484 * reorg.c (steal_delay_list_from_target): Call update_block for
9486 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
9488 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9490 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
9491 two duplicate entries.
9493 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9496 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
9498 * config/mips/mips.c (mips_truncated_op_cost): New function.
9499 (mips_rtx_costs): Adjust test for BADDU.
9500 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
9502 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
9504 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
9505 (*baddu_si): ...this new pattern.
9507 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9510 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
9512 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9515 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
9517 (expand_expr, expand_normal): Adjust.
9518 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
9519 inner_reference_p. Use inner_reference_p to expand inner references.
9520 (store_expr): Adjust.
9521 * cfgexpand.c (expand_call_stmt): Adjust.
9523 2014-01-08 Rong Xu <xur@google.com>
9525 * gcov-io.c (gcov_var): Move from gcov-io.h.
9526 (gcov_position): Ditto.
9527 (gcov_is_error): Ditto.
9528 (gcov_rewrite): Ditto.
9529 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
9530 only part to libgcc/libgcov.h.
9532 2014-01-08 Marek Polacek <polacek@redhat.com>
9535 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
9537 2014-01-08 Marek Polacek <polacek@redhat.com>
9540 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
9542 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9544 PR rtl-optimization/59649
9545 * stor-layout.c (get_mode_bounds): For BImode return
9546 0 and STORE_FLAG_VALUE.
9548 2014-01-08 Richard Biener <rguenther@suse.de>
9551 * gimple.h (is_gimple_builtin_call): Remove.
9552 (gimple_builtin_call_types_compatible_p): New.
9553 (gimple_call_builtin_p): New overload.
9554 * gimple.c (is_gimple_builtin_call): Remove.
9555 (validate_call): Rename to ...
9556 (gimple_builtin_call_types_compatible_p): ... this and export. Also
9558 (validate_type): New static function.
9559 (gimple_call_builtin_p): New overload and adjust.
9560 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
9561 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
9562 (gimple_fold_stmt_to_constant_1): Likewise.
9563 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
9565 2014-01-08 Richard Biener <rguenther@suse.de>
9568 * gimplify.c (gimplify_expr): Gimplify register-register type
9569 VIEW_CONVERT_EXPRs to separate stmts.
9571 2014-01-07 Jeff Law <law@redhat.com>
9574 * ree.c (combine_set_extension): Handle case where source
9575 and destination registers in an extension insn are different.
9576 (combine_reaching_defs): Allow source and destination registers
9577 in extension to be different under limited circumstances.
9578 (add_removable_extension): Remove restriction that the
9579 source and destination registers in the extension are the same.
9580 (find_and_remove_re): Emit a copy from the extension's
9581 destination to its source after the defining insn if
9582 the source and destination registers are different.
9585 * ifcvt.c (merge_if_block): If we are merging a block with more than
9586 one successor with a block with no successors, remove any BARRIER
9587 after the second block.
9589 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
9591 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
9593 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
9596 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
9597 for 14-bit register offsets when INT14_OK_STRICT is false.
9599 2014-01-07 Roland Stigge <stigge@antcom.de>
9600 Michael Meissner <meissner@linux.vnet.ibm.com>
9603 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
9604 Only check TFmode for SPE constants. Don't check TImode or TDmode.
9606 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
9608 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
9611 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
9613 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
9614 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
9615 rtx is const0_rtx or not.
9617 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
9620 * target-globals.c (save_target_globals): Remove this_fn_optab
9622 * toplev.c: Include optabs.h.
9623 (target_reinit): Temporarily restore the global options if another
9624 set of options are in force.
9626 2014-01-07 Jakub Jelinek <jakub@redhat.com>
9628 PR rtl-optimization/58668
9629 * cfgcleanup.c (flow_find_cross_jump): Don't count
9630 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
9631 to determine what is counted.
9632 (flow_find_head_matching_sequence): Use active_insn_p to determine
9634 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
9636 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
9637 determine what is counted.
9639 PR tree-optimization/59643
9640 * tree-predcom.c (split_data_refs_to_components): If one dr is
9641 read and one write, determine_offset fails and the write isn't
9642 in the bad component, just put the read into the bad component.
9644 2014-01-07 Mike Stump <mikestump@comcast.net>
9645 Jakub Jelinek <jakub@redhat.com>
9648 * tree-core.h (struct tree_optimization_option): Change optabs
9649 type from unsigned char * to void *.
9650 * optabs.c (init_tree_optimization_optabs): Adjust
9651 TREE_OPTIMIZATION_OPTABS initialization.
9653 2014-01-06 Jakub Jelinek <jakub@redhat.com>
9656 * config/i386/i386.h (struct machine_function): Add
9657 no_drap_save_restore field.
9658 * config/i386/i386.c (ix86_save_reg): Use
9659 !cfun->machine->no_drap_save_restore instead of
9660 crtl->stack_realign_needed.
9661 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
9662 this function clears frame_pointer_needed. Set
9663 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
9664 and DRAP reg is needed.
9666 2014-01-06 Marek Polacek <polacek@redhat.com>
9669 * doc/implement-c.texi: Mention that other integer types are
9670 permitted as bit-field types in strictly conforming mode.
9672 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
9674 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
9677 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
9679 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
9681 2014-01-06 Martin Jambor <mjambor@suse.cz>
9684 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
9686 * ipa-prop.c (ipa_print_node_params): Fix indentation.
9688 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
9692 * var-tracking.c (add_stores): Preserve the value of the source even if
9693 we don't record the store.
9695 2014-01-06 Terry Guo <terry.guo@arm.com>
9697 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
9699 2014-01-05 Iain Sandoe <iain@codesourcery.com>
9702 * config/darwin.c (darwin_function_section): Adjust return values to
9703 correspond to optimisation changes made in r206070.
9705 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
9707 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
9708 from prefetch_block tune setting.
9709 (nocona_cost): Correct size of prefetch block to 64.
9711 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
9713 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
9714 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
9715 used to save the static chain register in the computation of the offset
9716 from which the FP registers need to be restored.
9718 2014-01-04 Jakub Jelinek <jakub@redhat.com>
9720 PR tree-optimization/59519
9721 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
9722 ICE if get_current_def (current_new_name) is already non-NULL, as long
9723 as it is a phi result of some other phi in *new_exit_bb that has
9726 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
9727 or vmovdqu* for misaligned_operand.
9728 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
9729 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
9730 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
9731 aligned_mem for AVX512F masked aligned load and store builtins and for
9734 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
9736 PR tree-optimization/59651
9737 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
9738 Address range for negative step should be added by TYPE_SIZE_UNIT.
9740 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
9742 * config/m68k/m68k.c (handle_move_double): Handle pushes with
9743 overlapping registers also for registers other than the stack pointer.
9745 2014-01-03 Marek Polacek <polacek@redhat.com>
9748 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
9751 2014-01-03 Jakub Jelinek <jakub@redhat.com>
9754 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
9757 * config/i386/i386.md (MODE_SIZE): New mode attribute.
9758 (push splitter): Use <P:MODE_SIZE> instead of
9759 GET_MODE_SIZE (<P:MODE>mode).
9760 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
9761 (mov -1, reg peephole2): Likewise.
9762 * config/i386/sse.md (*mov<mode>_internal,
9763 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
9764 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
9765 *<code><mode>3, *andnot<mode>3<mask_name>,
9766 <mask_codefor><code><mode>3<mask_name>): Likewise.
9767 * config/i386/subst.md (mask_mode512bit_condition,
9768 sd_mask_mode512bit_condition): Likewise.
9770 2014-01-02 Xinliang David Li <davidxl@google.com>
9772 PR tree-optimization/59303
9773 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
9774 (dump_predicates): Better output format.
9775 (pred_equal_p): New function.
9776 (is_neq_relop_p): Ditto.
9777 (is_neq_zero_form_p): Ditto.
9778 (pred_expr_equal_p): Ditto.
9779 (pred_neg_p): Ditto.
9780 (simplify_pred): Ditto.
9781 (simplify_preds_2): Ditto.
9782 (simplify_preds_3): Ditto.
9783 (simplify_preds_4): Ditto.
9784 (simplify_preds): Ditto.
9786 (push_to_worklist): Ditto.
9787 (get_pred_info_from_cmp): Ditto.
9788 (is_degenerated_phi): Ditto.
9789 (normalize_one_pred_1): Ditto.
9790 (normalize_one_pred): Ditto.
9791 (normalize_one_pred_chain): Ditto.
9792 (normalize_preds): Ditto.
9793 (normalize_cond_1): Remove function.
9794 (normalize_cond): Ditto.
9795 (is_gcond_subset_of): Ditto.
9796 (is_subset_of_any): Ditto.
9797 (is_or_set_subset_of): Ditto.
9798 (is_and_set_subset_of): Ditto.
9799 (is_norm_cond_subset_of): Ditto.
9800 (pred_chain_length_cmp): Ditto.
9801 (convert_control_dep_chain_into_preds): Type change.
9802 (find_predicates): Ditto.
9803 (find_def_preds): Ditto.
9804 (destroy_predicates_vecs): Ditto.
9805 (find_matching_predicates_in_rest_chains): Ditto.
9806 (use_pred_not_overlap_with_undef_path_pred): Ditto.
9807 (is_pred_expr_subset): Ditto.
9808 (is_pred_chain_subset_of): Ditto.
9809 (is_included_in): Ditto.
9810 (is_superset_of): Ditto.
9812 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
9814 Update copyright years.
9816 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
9818 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
9819 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
9820 config/arc/arc.md, config/arc/arc.opt,
9821 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
9822 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
9823 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
9824 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
9825 config/linux-protos.h, config/linux.c, config/winnt-c.c,
9826 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
9827 vtable-verify.c, vtable-verify.h: Use the standard form for the
9830 2014-01-02 Tobias Burnus <burnus@net-b.de>
9832 * gcc.c (process_command): Update copyright notice dates.
9833 * gcov-dump.c: Ditto.
9835 * doc/cpp.texi: Bump @copying's copyright year.
9836 * doc/cppinternals.texi: Ditto.
9837 * doc/gcc.texi: Ditto.
9838 * doc/gccint.texi: Ditto.
9839 * doc/gcov.texi: Ditto.
9840 * doc/install.texi: Ditto.
9841 * doc/invoke.texi: Ditto.
9843 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9845 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
9847 2014-01-01 Jakub Jelinek <jakub@redhat.com>
9849 * config/i386/sse.md (*mov<mode>_internal): Guard
9850 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
9852 PR rtl-optimization/59647
9853 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
9854 new_rtx into UNSIGNED_FLOAT rtxes.
9856 Copyright (C) 2014 Free Software Foundation, Inc.
9858 Copying and distribution of this file, with or without modification,
9859 are permitted in any medium without royalty provided the copyright
9860 notice and this notice are preserved.