1 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
3 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
4 of using optimize_size.
6 2014-05-06 Mike Stump <mikestump@comcast.net>
8 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
10 2014-05-06 Joseph Myers <joseph@codesourcery.com>
12 * config/i386/sse.md (*mov<mode>_internal)
13 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
14 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
15 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
16 (*<code><mode>3, *andnot<mode>3<mask_name>)
17 (<mask_codefor><code><mode>3<mask_name>): Only consider
18 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
20 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
23 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
25 * lra-constraints.c (valid_address_p): Move earlier in file.
26 Add a constraint argument to the address_info version.
27 (satisfies_memory_constraint_p): New function.
28 (satisfies_address_constraint_p): Likewise.
29 (process_alt_operands, curr_insn_transform): Use them.
30 (process_address): Pass the constraint to valid_address_p when
31 checking address operands.
33 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
35 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
36 to their respective blocks. Fix inadvertent use of "node".
38 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
40 * emit-rtl.c (init_derived_machine_modes): New functionm, split
42 (init_emit_once): ...here.
43 * rtl.h (init_derived_machine_modes): Declare.
44 * toplev.c (do_compile): Call it even if no_backend.
46 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
47 Mike Stump <mikestump@comcast.net>
48 Richard Sandiford <rdsandiford@googlemail.com>
49 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
51 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
52 (rtx_equal_for_memref_p): Update comment.
53 (adjust_offset_for_component_ref): Use wide-int interfaces.
54 * builtins.c (get_object_alignment_2): Likewise.
55 (c_readstr): Likewise.
56 (target_char_cast): Add comment.
57 (determine_block_size): Use wide-int interfaces.
58 (expand_builtin_signbit): Likewise.
59 (fold_builtin_int_roundingfn): Likewise.
60 (fold_builtin_bitop): Likewise.
61 (fold_builtin_bswap): Likewise.
62 (fold_builtin_logarithm): Use signop.
63 (fold_builtin_pow): Likewise.
64 (fold_builtin_memory_op): Use wide-int interfaces.
65 (fold_builtin_object_size): Likewise.
66 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
67 nb_iterations_estimate.
68 (record_niter_bound): Use wide-int interfaces.
69 (get_estimated_loop_iterations_int): Likewise.
70 (get_estimated_loop_iterations): Likewise.
71 (get_max_loop_iterations): Likewise.
72 * cfgloop.h: Include wide-int.h.
73 (struct nb_iter_bound): Change bound to widest_int.
74 (struct loop): Change nb_iterations_upper_bound and
75 nb_iterations_estimate to widest_int.
76 (record_niter_bound): Switch to use widest_int.
77 (get_estimated_loop_iterations): Likewise.
78 (get_max_loop_iterations): Likewise.
79 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
81 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
82 * combine.c (try_combine): Likewise.
83 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
84 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
86 (aarch64_float_const_representable_p): Likewise.
87 * config/arc/arc.c: Include wide-int.h.
88 (arc_can_use_doloop_p): Use wide-int interfaces.
89 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
90 (vfp3_const_double_index): Likewise.
91 * config/avr/avr.c (avr_out_round): Likewise.
92 (avr_fold_builtin): Likewise.
93 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
94 (bfin_can_use_doloop_p): Likewise.
95 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
96 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
97 * config/i386/i386.c: Include wide-int.h.
98 (ix86_data_alignment): Use wide-int interfaces.
99 (ix86_local_alignment): Likewise.
100 (ix86_emit_swsqrtsf): Update real_from_integer.
101 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
102 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
103 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
104 (zero_constant): Likewise.
105 (input_operand): Likewise.
106 (splat_input_operand): Likewise.
107 (non_logical_cint_operand): Change const_double to const_wide_int.
108 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
109 (easy_altivec_constant): Remove comment.
110 (paired_expand_vector_init): Use CONSTANT_P.
111 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
112 (rs6000_emit_move): Update checks.
113 (rs6000_aggregate_candidate): Use wide-int interfaces.
114 (rs6000_expand_ternop_builtin): Likewise.
115 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
116 (rs6000_assemble_integer): Likewise.
117 (rs6000_hash_constant): Likewise.
118 (output_toc): Likewise.
119 (rs6000_rtx_costs): Likewise.
120 (rs6000_emit_swrsqrt); Update call to real_from_integer.
121 * config/rs6000/rs6000-c.c: Include wide-int.h.
122 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
123 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
124 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
125 Handle CONST_WIDE_INT.
126 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
127 Use tree_fits_uhwi_p.
128 * config/sparc/sparc.c: Include wide-int.h.
129 (sparc_fold_builtin): Use wide-int interfaces.
130 * config/vax/vax.c: Include wide-int.h.
131 (vax_float_literal): Use real_from_integer.
132 * coretypes.h (struct hwivec_def): New.
135 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
136 (equiv_constant): Handle CONST_WIDE_INT.
137 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
138 (cselib_hash_rtx): Handle CONST_WIDE_INT.
139 * dbxout.c (stabstr_U): Use wide-int interfaces.
140 (dbxout_type): Update to use cst_fits_shwi_p.
141 * defaults.h (LOG2_BITS_PER_UNIT): Define.
142 (TARGET_SUPPORTS_WIDE_INT): Add default.
143 * dfp.c: Include wide-int.h.
144 (decimal_real_to_integer2): Use wide-int interfaces and rename to
145 decimal_real_to_integer.
146 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
147 decimal_real_to_integer.
148 * doc/generic.texi (Constant expressions): Update for wide_int.
149 * doc/rtl.texi (const_double): Likewise.
150 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
151 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
152 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
153 (REAL_VALUE_FROM_INT): Remove.
154 (TARGET_SUPPORTS_WIDE_INT): New.
155 * doc/tm.texi: Regenerate.
156 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
157 * double-int.h: Include wide-int.h.
158 (struct wi::int_traits): New.
159 * dwarf2out.c (get_full_len): New.
160 (dw_val_equal_p): Add case dw_val_class_wide_int.
161 (size_of_loc_descr): Likewise.
162 (output_loc_operands): Likewise.
163 (insert_double): Remove.
164 (insert_wide_int): New.
166 (print_die): Add case dw_val_class_wide_int.
167 (attr_checksum): Likewise.
168 (attr_checksum_ordered): Likewise.
169 (same_dw_val_p): Likewise.
170 (size_of_die): Likewise.
171 (value_format): Likewise.
172 (output_die): Likewise.
173 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
175 (clz_loc_descriptor): Use wide-int interfaces.
176 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
177 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
178 (round_up_to_align): Use wide-int interfaces.
179 (field_byte_offset): Likewise.
180 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
181 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
182 CONST_DOUBLE handling. Use wide-int interfaces.
183 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
184 (gen_enumeration_type_die): Use add_AT_wide.
185 (hash_loc_operands): Add case dw_val_class_wide_int.
186 (compare_loc_operands): Likewise.
187 * dwarf2out.h: Include wide-int.h.
189 (enum dw_val_class): Add dw_val_class_wide_int.
190 (struct dw_val_struct): Add val_wide.
191 * emit-rtl.c (const_wide_int_htab): New.
192 (const_wide_int_htab_hash): New.
193 (const_wide_int_htab_eq): New.
194 (lookup_const_wide_int): New.
195 (const_double_htab_hash): Use wide-int interfaces.
196 (const_double_htab_eq): Likewise.
197 (rtx_to_double_int): Conditionally compile for wide-int.
198 (immed_double_int_const): Rename to immed_wide_int_const and
200 (immed_double_const): Conditionally compile for wide-int.
201 (init_emit_once): Use wide-int interfaces.
202 * explow.c (plus_constant): Likewise.
203 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
204 (lshift_value): Use wide-int interfaces.
205 (expand_mult): Likewise.
206 (choose_multiplier): Likewise.
207 (expand_smod_pow2): Likewise.
208 (make_tree): Likewise.
209 * expr.c (convert_modes): Consolidate handling of constants.
210 Use wide-int interfaces.
211 (emit_group_load_1): Add note.
212 (store_expr): Update comment.
213 (get_inner_reference): Use wide-int interfaces.
214 (expand_constructor): Update comment.
215 (expand_expr_real_2): Use wide-int interfaces.
216 (expand_expr_real_1): Likewise.
217 (reduce_to_bit_field_precision): Likewise.
218 (const_vector_from_tree): Likewise.
219 * final.c: Include wide-int-print.h.
220 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
221 * fixed-value.c: Include wide-int.h.
222 (fixed_from_string): Use wide-int interfaces.
223 (fixed_to_decimal): Likewise.
224 (fixed_convert_from_real): Likewise.
225 (real_convert_from_fixed): Likewise.
226 * fold-const.h (mem_ref_offset): Return an offset_int.
227 (div_if_zero_remainder): Remove code parameter.
228 * fold-const.c (div_if_zero_remainder): Remove code parameter.
229 Use wide-int interfaces.
230 (may_negate_without_overflow_p): Use wide-int interfaces.
231 (negate_expr_p): Likewise.
232 (fold_negate_expr): Likewise.
233 (int_const_binop_1): Likewise.
234 (const_binop): Likewise.
235 (fold_convert_const_int_from_int): Likewise.
236 (fold_convert_const_int_from_real): Likewise.
237 (fold_convert_const_int_from_fixed): Likewise.
238 (fold_convert_const_fixed_from_int): Likewise.
239 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
240 (sign_bit_p): Use wide-int interfaces.
241 (make_range_step): Likewise.
242 (build_range_check): Likewise. Pass an integer of the correct type
243 instead of using integer_one_node.
244 (range_predecessor): Pass an integer of the correct type instead
245 of using integer_one_node.
246 (range_successor): Likewise.
247 (merge_ranges): Likewise.
248 (unextend): Use wide-int interfaces.
249 (extract_muldiv_1): Likewise.
250 (fold_div_compare): Likewise.
251 (fold_single_bit_test): Likewise.
252 (fold_sign_changed_comparison): Likewise.
253 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
254 (fold_plusminus_mult_expr): Use wide-int interfaces.
255 (native_encode_int): Likewise.
256 (native_interpret_int): Likewise.
257 (fold_unary_loc): Likewise.
258 (pointer_may_wrap_p): Likewise.
259 (size_low_cst): Likewise.
260 (mask_with_tz): Likewise.
261 (fold_binary_loc): Likewise.
262 (fold_ternary_loc): Likewise.
263 (multiple_of_p): Likewise.
264 (tree_call_nonnegative_warnv_p): Update calls to
265 tree_int_cst_min_precision and real_from_integer.
266 (fold_negate_const): Use wide-int interfaces.
267 (fold_abs_const): Likewise.
268 (fold_relational_const): Use tree_int_cst_lt.
269 (round_up_loc): Use wide-int interfaces.
270 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
271 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
272 * gengtype.c: Remove include of double-int.h.
273 (do_typedef): Use wide-int interfaces.
274 (open_base_files): Add wide-int.h.
275 (main): Add offset_int and widest_int typedefs.
276 * gengtype-lex.l: Handle "^".
277 (CXX_KEYWORD): Add "static".
278 * gengtype-parse.c (require3): New.
279 (require_template_declaration): Handle constant template arguments
280 and nested templates.
281 * gengtype-state.c: Don't include "double-int.h".
282 * genpreds.c (write_one_predicate_function): Update comment.
283 (write_tm_constrs_h): Add check for hval and lval use in
285 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
286 (add_to_sequence): Likewise.
287 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
288 and const_double_operand.
289 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
291 * gimple-fold.c (get_base_constructor): Likewise.
292 (fold_array_ctor_reference): Likewise.
293 (fold_nonarray_ctor_reference): Likewise.
294 (fold_const_aggregate_ref_1): Likewise.
295 (gimple_val_nonnegative_real_p): Likewise.
296 (gimple_fold_indirect_ref): Likewise.
297 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
298 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
299 (struct slsr_cand_d): Change index to be widest_int.
300 (struct incr_info_d): Change incr to be widest_int.
301 (alloc_cand_and_find_basis): Use wide-int interfaces.
302 (slsr_process_phi): Likewise.
303 (backtrace_base_for_ref): Likewise. Return a widest_int.
304 (restructure_reference): Take a widest_int instead of a double_int.
305 (slsr_process_ref): Use wide-int interfaces.
306 (create_mul_ssa_cand): Likewise.
307 (create_mul_imm_cand): Likewise.
308 (create_add_ssa_cand): Likewise.
309 (create_add_imm_cand): Take a widest_int instead of a double_int.
310 (slsr_process_add): Use wide-int interfaces.
311 (slsr_process_cast): Likewise.
312 (slsr_process_copy): Likewise.
313 (dump_candidate): Likewise.
314 (dump_incr_vec): Likewise.
315 (replace_ref): Likewise.
316 (cand_increment): Likewise. Return a widest_int.
317 (cand_abs_increment): Likewise.
318 (replace_mult_candidate): Take a widest_int instead of a double_int.
319 (replace_unconditional_candidate): Use wide-int interfaces.
320 (incr_vec_index): Take a widest_int instead of a double_int.
321 (create_add_on_incoming_edge): Likewise.
322 (create_phi_basis): Use wide-int interfaces.
323 (replace_conditional_candidate): Likewise.
324 (record_increment): Take a widest_int instead of a double_int.
325 (record_phi_increments): Use wide-int interfaces.
326 (phi_incr_cost): Take a widest_int instead of a double_int.
327 (lowest_cost_path): Likewise.
328 (total_savings): Likewise.
329 (analyze_increments): Use wide-int interfaces.
330 (ncd_with_phi): Take a widest_int instead of a double_int.
331 (ncd_of_cand_and_phis): Likewise.
332 (nearest_common_dominator_for_cands): Likewise.
333 (insert_initializers): Use wide-int interfaces.
334 (all_phi_incrs_profitable): Likewise.
335 (replace_one_candidate): Likewise.
336 (replace_profitable_candidates): Likewise.
337 * godump.c: Include wide-int-print.h.
338 (go_output_typedef): Use wide-int interfaces.
339 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
340 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
341 (build_loop_iteration_domains): Likewise.
342 * hooks.h: Include wide-int.h rather than double-int.h.
343 (hook_bool_dint_dint_uint_bool_true): Delete.
344 (hook_bool_wint_wint_uint_bool_true): Declare.
345 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
346 (hook_bool_wint_wint_uint_bool_true): New.
347 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
349 (ubsan_expand_si_overflow_mul_check): Likewise.
350 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
351 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
352 (get_ancestor_addr_info): Likewise.
353 (ipa_modify_call_arguments): Likewise.
354 * loop-doloop.c (doloop_modify): Likewise.
355 (doloop_optimize): Likewise.
356 * loop-iv.c (iv_number_of_iterations): Likewise.
357 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
358 (unroll_loop_constant_iterations): Likewise.
359 (decide_unroll_runtime_iterations): Likewise.
360 (unroll_loop_runtime_iterations): Likewise.
361 (decide_peel_simple): Likewise.
362 (decide_unroll_stupid): Likewise.
363 * lto-streamer-in.c (streamer_read_wi): Add.
364 (input_cfg): Use wide-int interfaces.
365 (lto_input_tree_1): Likewise.
366 * lto-streamer-out.c (streamer_write_wi): Add.
367 (hash_tree): Use wide-int interfaces.
368 (output_cfg): Likewise.
369 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
370 (GTFILES): Add wide-int.h and signop.h.
371 (TAGS): Look for .cc files too.
372 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
373 * optabs.c (expand_subword_shift): Likewise.
374 (expand_doubleword_shift): Likewise.
375 (expand_absneg_bit): Likewise.
376 (expand_copysign_absneg): Likewise.
377 (expand_copysign_bit): Likewise.
378 * postreload.c (reload_cse_simplify_set): Likewise.
379 * predict.c (predict_iv_comparison): Likewise.
380 * pretty-print.h: Include wide-int-print.h.
382 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
383 * print-tree.c: Include wide-int-print.h.
384 (print_node_brief): Use wide-int interfaces.
385 (print_node): Likewise.
386 * read-rtl.c (validate_const_wide_int): New.
387 (read_rtx_code): Add CONST_WIDE_INT case.
388 * real.c: Include wide-int.h.
389 (real_to_integer2): Delete.
390 (real_to_integer): New function, returning a wide_int.
391 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
392 (ten_to_ptwo): Update call to real_from_integer.
393 (real_digit): Likewise.
394 * real.h: Include signop.h, wide-int.h and insn-modes.h.
395 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
396 (REAL_VALUE_TO_INT): Delete.
397 (real_to_integer): Declare a wide-int form.
398 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
399 * recog.c (const_int_operand): Improve comment.
400 (const_scalar_int_operand): New.
401 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
402 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
403 (split_double): Likewise.
404 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
405 (rtx_size): Likewise.
406 (rtx_alloc_stat_v): New.
407 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
408 (cwi_output_hex): New.
409 (iterative_hash_rtx): Handle CONST_WIDE_INT.
410 (cwi_check_failed_bounds): New.
411 * rtl.def (CONST_WIDE_INT): New.
412 * rtl.h: Include <utility> and wide-int.h.
413 (struct hwivec_def): New.
414 (CWI_GET_NUM_ELEM): New.
415 (CWI_PUT_NUM_ELEM): New.
416 (struct rtx_def): Add num_elem and hwiv.
417 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
418 (CASE_CONST_UNIQUE): Likewise.
419 (CASE_CONST_ANY): Likewise.
420 (CONST_SCALAR_INT_P): Likewise.
421 (CONST_WIDE_INT_P): New.
424 (cwi_check_failed_bounds): New.
427 (CONST_WIDE_INT_VEC) New.
428 (CONST_WIDE_INT_NUNITS) New.
429 (CONST_WIDE_INT_ELT) New.
430 (rtx_mode_t): New type.
431 (wi::int_traits <rtx_mode_t>): New.
433 (wi::min_value): New.
434 (wi::max_value): New.
436 (const_wide_int_alloc): New.
437 (immed_wide_int_const): New.
438 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
439 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
440 * signop.h: New file.
441 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
442 (simplify_const_unary_operation): Use wide-int interfaces.
443 (simplify_binary_operation_1): Likewise.
444 (simplify_const_binary_operation): Likewise.
445 (simplify_const_relational_operation): Likewise.
446 (simplify_immed_subreg): Likewise.
447 * stmt.c (expand_case): Likewise.
448 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
449 signop rather than a bool.
450 * stor-layout.c (layout_type): Use wide-int interfaces.
451 (initialize_sizetypes): Update calls to
452 set_min_and_max_values_for_integral_type.
453 (set_min_and_max_values_for_integral_type): Take a signop rather
454 than a bool. Use wide-int interfaces.
455 (fixup_signed_type): Update accordingly. Remove
456 HOST_BITS_PER_DOUBLE_INT limit.
457 (fixup_unsigned_type): Likewise.
458 * system.h (STATIC_CONSTANT_P): New.
459 (STATIC_ASSERT): New.
460 * target.def (can_use_doloop_p): Take widest_ints rather than
462 * target.h: Include wide-int.h rather than double-int.h.
463 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
465 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
466 rather than INT_CST_LT_UNSIGNED.
467 (can_use_doloop_if_innermost): Take widest_ints rather than
469 * tree-affine.c: Include wide-int-print.h.
470 (double_int_ext_for_comb): Delete.
471 (wide_int_ext_for_comb): New.
472 (aff_combination_zero): Use wide-int interfaces.
473 (aff_combination_const): Take a widest_int instead of a double_int.
474 (aff_combination_elt): Use wide-int interfaces.
475 (aff_combination_scale): Take a widest_int instead of a double_int.
476 (aff_combination_add_elt): Likewise.
477 (aff_combination_add_cst): Likewise.
478 (aff_combination_add): Use wide-int interfaces.
479 (aff_combination_convert): Likewise.
480 (tree_to_aff_combination): Likewise.
481 (add_elt_to_tree): Take a widest_int instead of a double_int.
482 (aff_combination_to_tree): Use wide-int interfaces.
483 (aff_combination_remove_elt): Likewise.
484 (aff_combination_add_product): Take a widest_int instead of
486 (aff_combination_mult): Use wide-int interfaces.
487 (aff_combination_expand): Likewise.
488 (double_int_constant_multiple_p): Delete.
489 (wide_int_constant_multiple_p): New.
490 (aff_combination_constant_multiple_p): Take a widest_int pointer
491 instead of a double_int pointer.
492 (print_aff): Use wide-int interfaces.
493 (get_inner_reference_aff): Take a widest_int pointer
494 instead of a double_int pointer.
495 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
496 * tree-affine.h: Include wide-int.h.
497 (struct aff_comb_elt): Change type of coef to widest_int.
498 (struct affine_tree_combination): Change type of offset to widest_int.
499 (double_int_ext_for_comb): Delete.
500 (wide_int_ext_for_comb): New.
501 (aff_combination_const): Use widest_int instead of double_int.
502 (aff_combination_scale): Likewise.
503 (aff_combination_add_elt): Likewise.
504 (aff_combination_constant_multiple_p): Likewise.
505 (get_inner_reference_aff): Likewise.
506 (aff_comb_cannot_overlap_p): Likewise.
507 (aff_combination_zero_p): Use wide-int interfaces.
508 * tree.c: Include tree.h.
509 (init_ttree): Use make_int_cst.
510 (tree_code_size): Removed code for INTEGER_CST case.
511 (tree_size): Add INTEGER_CST case.
512 (make_node_stat): Update comment.
513 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
514 (build_int_cst_type): Use wide-int interfaces.
515 (double_int_to_tree): Likewise.
516 (double_int_fits_to_tree_p): Delete.
517 (force_fit_type_double): Delete.
518 (force_fit_type): New.
519 (int_cst_hash_hash): Use wide-int interfaces.
520 (int_cst_hash_eq): Likewise.
521 (build_int_cst_wide): Delete.
522 (wide_int_to_tree): New.
523 (cache_integer_cst): Use wide-int interfaces.
524 (build_low_bits_mask): Likewise.
525 (cst_and_fits_in_hwi): Likewise.
526 (real_value_from_int_cst): Likewise.
527 (make_int_cst_stat): New.
528 (integer_zerop): Use wide_int interfaces.
529 (integer_onep): Likewise.
530 (integer_all_onesp): Likewise.
531 (integer_pow2p): Likewise.
532 (integer_nonzerop): Likewise.
533 (tree_log2): Likewise.
534 (tree_floor_log2): Likewise.
535 (tree_ctz): Likewise.
536 (int_size_in_bytes): Likewise.
537 (mem_ref_offset): Return an offset_int rather than a double_int.
538 (build_type_attribute_qual_variant): Use wide_int interfaces.
539 (type_hash_eq): Likewise
540 (tree_int_cst_equal): Likewise.
541 (tree_int_cst_lt): Delete.
542 (tree_int_cst_compare): Likewise.
543 (tree_fits_shwi_p): Use wide_int interfaces.
544 (tree_fits_uhwi_p): Likewise.
545 (tree_int_cst_sign_bit): Likewise.
546 (tree_int_cst_sgn): Likewise.
547 (tree_int_cst_min_precision): Take a signop rather than a bool.
548 (simple_cst_equal): Use wide_int interfaces.
549 (compare_tree_int): Likewise.
550 (iterative_hash_expr): Likewise.
551 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
553 (get_type_static_bounds): Use wide_int interfaces.
554 (tree_int_cst_elt_check_failed): New.
555 (build_common_tree_nodes): Reordered to set prec before filling in
557 (int_cst_value): Check cst_and_fits_in_hwi.
558 (widest_int_cst_value): Use wide_int interfaces.
559 (upper_bound_in_type): Likewise.
560 (lower_bound_in_type): Likewise.
561 (num_ending_zeros): Likewise.
562 (drop_tree_overflow): Likewise.
563 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
564 (gen_conditions_for_pow_cst_base): Likewise.
565 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
566 (group_case_labels_stmt): Use wide-int interfaces.
567 (verify_gimple_assign_binary): Likewise.
568 (print_loop): Likewise.
569 * tree-chrec.c (tree_fold_binomial): Likewise.
570 * tree-core.h (struct tree_base): Add int_length.
571 (struct tree_int_cst): Change rep of value.
572 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
573 (dr_may_alias_p): Likewise.
574 (max_stmt_executions_tree): Likewise.
575 * tree.def (INTEGER_CST): Update comment.
576 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
577 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
578 * tree-dump.c: Include wide-int.h and wide-int-print.h.
579 (dequeue_and_dump): Use wide-int interfaces.
580 * tree.h: Include wide-int.h.
581 (NULL_TREE): Moved to earlier loc in file.
582 (TREE_INT_CST_ELT_CHECK): New.
583 (tree_int_cst_elt_check_failed): New.
585 (TREE_INT_CST): Delete.
586 (TREE_INT_CST_LOW): Use wide-int interfaces.
587 (TREE_INT_CST_HIGH): Delete.
588 (TREE_INT_CST_NUNITS): New.
589 (TREE_INT_CST_EXT_NUNITS): Likewise.
590 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
591 (TREE_INT_CST_ELT): Likewise.
592 (INT_CST_LT): Delete.
593 (tree_int_cst_elt_check): New (two forms).
594 (type_code_size): Update comment.
595 (make_int_cst_stat, make_int_cst): New.
596 (tree_to_double_int): Delete.
597 (double_int_fits_to_tree_p): Delete.
598 (force_fit_type_double): Delete.
599 (build_int_cstu): Replace with out-of-line function.
600 (build_int_cst_wide): Delete.
601 (tree_int_cst_lt): Define inline.
602 (tree_int_cst_le): New.
603 (tree_int_cst_compare): Define inline.
604 (tree_int_cst_min_precision): Take a signop rather than a bool.
605 (wi::int_traits <const_tree>): New.
606 (wi::int_traits <tree>): New.
607 (wi::extended_tree): New.
608 (wi::int_traits <wi::extended_tree>): New.
609 (wi::to_widest): New.
610 (wi::to_offset): New.
611 (wi::fits_to_tree_p): New.
612 (wi::min_value): New.
613 (wi::max_value): New.
614 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
615 (copy_tree_body_r): Likewise.
616 * tree-object-size.c (compute_object_offset): Likewise.
617 (addr_object_size): Likewise.
618 * tree-predcom.c: Include wide-int-print.h.
619 (struct dref_d): Change type of offset to widest_int.
620 (dump_dref): Call wide-int printer.
621 (aff_combination_dr_offset): Use wide-int interfaces.
622 (determine_offset): Take a widest_int pointer rather than a
624 (split_data_refs_to_components): Use wide-int interfaces.
625 (suitable_component_p): Likewise.
626 (order_drefs): Likewise.
627 (add_ref_to_chain): Likewise.
628 (valid_initializer_p): Likewise.
629 (determine_roots_comp): Likewise.
630 * tree-pretty-print.c: Include wide-int-print.h.
631 (dump_generic_node): Use wide-int interfaces.
632 * tree-sra.c (sra_ipa_modify_expr): Likewise.
633 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
634 (move_fixed_address_to_symbol): Likewise.
635 (move_hint_to_base): Likewise.
636 (move_pointer_to_base): Likewise.
637 (move_variant_to_index): Likewise.
638 (most_expensive_mult_to_index): Likewise.
639 (addr_to_parts): Likewise.
640 (copy_ref_info): Likewise.
641 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
642 (indirect_refs_may_alias_p): Likewise.
643 (stmt_kills_ref_p_1): Likewise.
644 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
645 * tree-ssa-ccp.c: Update comment at top of file. Include
647 (struct prop_value_d): Change type of mask to widest_int.
648 (extend_mask): New function.
649 (dump_lattice_value): Use wide-int interfaces.
650 (get_default_value): Likewise.
651 (set_constant_value): Likewise.
652 (set_value_varying): Likewise.
653 (valid_lattice_transition): Likewise.
654 (set_lattice_value): Likewise.
655 (value_to_double_int): Delete.
656 (value_to_wide_int): New.
657 (get_value_from_alignment): Use wide-int interfaces.
658 (get_value_for_expr): Likewise.
659 (do_dbg_cnt): Likewise.
660 (ccp_finalize): Likewise.
661 (ccp_lattice_meet): Likewise.
662 (bit_value_unop_1): Use widest_ints rather than double_ints.
663 (bit_value_binop_1): Likewise.
664 (bit_value_unop): Use wide-int interfaces.
665 (bit_value_binop): Likewise.
666 (bit_value_assume_aligned): Likewise.
667 (evaluate_stmt): Likewise.
668 (ccp_fold_stmt): Likewise.
669 (visit_cond_stmt): Likewise.
670 (ccp_visit_stmt): Likewise.
671 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
672 (constant_pointer_difference): Likewise.
673 (associate_pointerplus): Likewise.
674 (combine_conversions): Likewise.
675 * tree-ssa-loop.h: Include wide-int.h.
676 (struct tree_niter_desc): Change type of max to widest_int.
677 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
678 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
679 (remove_redundant_iv_tests): Likewise.
680 (canonicalize_loop_induction_variables): Likewise.
681 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
682 (constant_multiple_of): Take a widest_int pointer instead of
683 a double_int pointer.
684 (get_computation_aff): Use wide-int interfaces.
685 (ptr_difference_cost): Likewise.
686 (difference_cost): Likewise.
687 (get_loop_invariant_expr_id): Likewise.
688 (get_computation_cost_at): Likewise.
689 (iv_elimination_compare_lt): Likewise.
690 (may_eliminate_iv): Likewise.
691 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
692 instead of double_int.
693 (max_loop_iterations): Likewise.
694 (max_stmt_executions): Likewise.
695 (estimated_stmt_executions): Likewise.
696 * tree-ssa-loop-niter.c: Include wide-int-print.h.
697 (split_to_var_and_offset): Use wide-int interfaces.
698 (determine_value_range): Likewise.
699 (bound_difference_of_offsetted_base): Likewise.
700 (bounds_add): Take a widest_int instead of a double_int.
701 (number_of_iterations_ne_max): Use wide-int interfaces.
702 (number_of_iterations_ne): Likewise.
703 (number_of_iterations_lt_to_ne): Likewise.
704 (assert_loop_rolls_lt): Likewise.
705 (number_of_iterations_lt): Likewise.
706 (number_of_iterations_le): Likewise.
707 (number_of_iterations_cond): Likewise.
708 (number_of_iterations_exit): Likewise.
709 (finite_loop_p): Likewise.
710 (derive_constant_upper_bound_assign): Likewise.
711 (derive_constant_upper_bound): Return a widest_int.
712 (derive_constant_upper_bound_ops): Likewise.
713 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
714 (record_estimate): Take a widest_int rather than a double_int.
715 (record_nonwrapping_iv): Use wide-int interfaces.
716 (double_int_cmp): Delete.
718 (bound_index): Take a widest_int rather than a double_int.
719 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
720 (maybe_lower_iteration_bound): Likewise.
721 (estimate_numbers_of_iterations_loop): Likewise.
722 (estimated_loop_iterations): Take a widest_int pointer than than
723 a double_int pointer.
724 (estimated_loop_iterations_int): Use wide-int interfaces.
725 (max_loop_iterations): Take a widest_int pointer than than
726 a double_int pointer.
727 (max_loop_iterations_int): Use wide-int interfaces.
728 (max_stmt_executions): Take a widest_int pointer than than
729 a double_int pointer.
730 (estimated_stmt_executions): Likewise.
731 (n_of_executions_at_most): Use wide-int interfaces.
732 (scev_probably_wraps_p): Likewise.
733 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
735 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
737 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
738 double_ints. Adjust for trailing_wide_ints <3> representation.
739 (set_nonzero_bits): Likewise.
740 (get_range_info): Return wide_ints rather than double_ints.
741 Adjust for trailing_wide_ints <3> representation.
742 (get_nonzero_bits): Likewise.
743 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
745 * tree-ssanames.h (struct range_info_def): Replace min, max and
746 nonzero_bits with a trailing_wide_ints <3>.
747 (set_range_info): Use wide_int_refs rather than double_ints.
748 (set_nonzero_bits): Likewise.
749 (get_range_info): Return wide_ints rather than double_ints.
750 (get_nonzero_bits): Likewise.
751 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
752 * tree-ssa-pre.c (phi_translate_1): Likewise.
753 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
754 (acceptable_pow_call): Likewise.
755 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
757 (vn_reference_fold_indirect): Likewise.
758 (vn_reference_maybe_forwprop_address): Likewise.
759 (valueize_refs_1): Likewise.
760 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
761 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
762 tree_int_cst_lt and tree_int_cst_le.
763 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
765 (streamer_alloc_tree): Likewise.
766 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
767 (streamer_write_tree_header): Likewise.
768 (streamer_write_integer_cst): Likewise.
769 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
770 (build_constructors): Likewise.
771 (array_value_type): Likewise.
772 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
773 (vect_check_gather): Likewise.
774 * tree-vect-generic.c (build_replicated_const): Likewise.
775 (expand_vector_divmod): Likewise.
776 * tree-vect-loop.c (vect_transform_loop): Likewise.
777 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
778 (vect_do_peeling_for_alignment): Likewise.
779 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
780 * tree-vrp.c: Include wide-int.h.
781 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
782 (extract_range_from_assert): Use wide-int interfaces.
783 (vrp_int_const_binop): Likewise.
784 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
786 (ranges_from_anti_range): Use wide-int interfaces.
787 (quad_int_cmp): Delete.
788 (quad_int_pair_sort): Likewise.
789 (extract_range_from_binary_expr_1): Use wide-int interfaces.
790 (extract_range_from_unary_expr_1): Likewise.
791 (adjust_range_with_scev): Likewise.
792 (masked_increment): Take and return wide_ints rather than double_ints.
793 (register_edge_assert_for_2): Use wide-int interfaces.
794 (check_array_ref): Likewise.
795 (search_for_addr_array): Likewise.
796 (maybe_set_nonzero_bits): Likewise.
797 (union_ranges): Pass an integer of the correct type instead of
798 using integer_one_node.
799 (intersect_ranges): Likewise.
800 (simplify_truth_ops_using_ranges): Likewise.
801 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
802 (range_fits_type_p): Likewise.
803 (simplify_cond_using_ranges): Likewise. Take a signop rather than
805 (simplify_conversion_using_ranges): Use wide-int interfaces.
806 (simplify_float_conversion_using_ranges): Likewise.
807 (vrp_finalize): Likewise.
808 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
809 (gimple_stringops_transform): Likewise.
810 * varasm.c (decode_addr_const): Likewise.
811 (const_hash_1): Likewise.
812 (const_rtx_hash_1): Likewise
813 (output_constant): Likewise.
814 (array_size_for_constructor): Likewise.
815 (output_constructor_regular_field): Likewise.
816 (output_constructor_bitfield): Likewise.
817 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
818 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
820 * gencheck.c: Define BITS_PER_UNIT.
823 * wide-int-print.cc: New.
824 * wide-int-print.h: New.
826 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
828 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
830 2014-05-06 Richard Biener <rguenther@suse.de>
832 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
833 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
834 (TODO_verify_all): Adjust.
835 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
836 TODO_verify_stmts and TODO_verify_rtl_sharing.
837 * bb-reorder.c: Likewise.
838 * cfgexpand.c: Likewise.
841 * function.c: Likewise.
842 * fwprop.c: Likewise.
844 * gimple-ssa-isolate-paths.c: Likewise.
845 * gimple-ssa-strength-reduction.c: Likewise.
846 * ipa-split.c: Likewise.
847 * loop-init.c: Likewise.
848 * loop-unroll.c: Likewise.
849 * lower-subreg.c: Likewise.
850 * modulo-sched.c: Likewise.
851 * postreload-gcse.c: Likewise.
852 * predict.c: Likewise.
854 * sched-rgn.c: Likewise.
855 * store-motion.c: Likewise.
856 * tracer.c: Likewise.
857 * trans-mem.c: Likewise.
858 * tree-call-cdce.c: Likewise.
859 * tree-cfg.c: Likewise.
860 * tree-cfgcleanup.c: Likewise.
861 * tree-complex.c: Likewise.
862 * tree-eh.c: Likewise.
863 * tree-emutls.c: Likewise.
864 * tree-if-conv.c: Likewise.
865 * tree-into-ssa.c: Likewise.
866 * tree-loop-distribution.c: Likewise.
867 * tree-object-size.c: Likewise.
868 * tree-parloops.c: Likewise.
869 * tree-pass.h: Likewise.
870 * tree-sra.c: Likewise.
871 * tree-ssa-ccp.c: Likewise.
872 * tree-ssa-copy.c: Likewise.
873 * tree-ssa-copyrename.c: Likewise.
874 * tree-ssa-dce.c: Likewise.
875 * tree-ssa-dom.c: Likewise.
876 * tree-ssa-dse.c: Likewise.
877 * tree-ssa-forwprop.c: Likewise.
878 * tree-ssa-ifcombine.c: Likewise.
879 * tree-ssa-loop-ch.c: Likewise.
880 * tree-ssa-loop-ivcanon.c: Likewise.
881 * tree-ssa-loop.c: Likewise.
882 * tree-ssa-math-opts.c: Likewise.
883 * tree-ssa-phiopt.c: Likewise.
884 * tree-ssa-phiprop.c: Likewise.
885 * tree-ssa-pre.c: Likewise.
886 * tree-ssa-reassoc.c: Likewise.
887 * tree-ssa-sink.c: Likewise.
888 * tree-ssa-strlen.c: Likewise.
889 * tree-ssa-tail-merge.c: Likewise.
890 * tree-ssa-uncprop.c: Likewise.
891 * tree-switch-conversion.c: Likewise.
892 * tree-tailcall.c: Likewise.
893 * tree-vect-generic.c: Likewise.
894 * tree-vectorizer.c: Likewise.
895 * tree-vrp.c: Likewise.
897 * var-tracking.c: Likewise.
898 * bt-load.c: Likewise.
899 * cfgcleanup.c: Likewise.
900 * combine-stack-adj.c: Likewise.
901 * combine.c: Likewise.
902 * compare-elim.c: Likewise.
903 * config/epiphany/resolve-sw-modes.c: Likewise.
904 * config/i386/i386.c: Likewise.
905 * config/mips/mips.c: Likewise.
906 * config/s390/s390.c: Likewise.
907 * config/sh/sh_treg_combine.cc: Likewise.
908 * config/sparc/sparc.c: Likewise.
913 * mode-switching.c: Likewise.
914 * passes.c: Likewise.
915 * postreload.c: Likewise.
917 * reg-stack.c: Likewise.
918 * regcprop.c: Likewise.
919 * regrename.c: Likewise.
922 2014-05-06 Richard Biener <rguenther@suse.de>
925 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
926 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
928 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
931 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
933 2014-05-05 Radovan Obradovic <robradovic@mips.com>
934 Tom de Vries <tom@codesourcery.com>
936 * target.def (call_fusage_contains_non_callee_clobbers): New DEFHOOKPOD.
937 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
939 (@node Miscellaneous Register Hooks): New node.
940 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
941 * doc/tm.texi: Regenerate.
943 2014-05-05 Marek Polacek <polacek@redhat.com>
946 * opts.c (common_handle_option): Call error_at instead of warning_at.
948 2014-05-05 Richard Biener <rguenther@suse.de>
950 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
951 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
952 under the TODO_verify_il umbrella.
954 2014-05-05 Richard Biener <rguenther@suse.de>
956 * passes.c (execute_function_todo): Move TODO_verify_flow under
957 the TODO_verify_ul umbrella.
959 2014-05-05 Richard Biener <rguenther@suse.de>
962 * fold-const.c (fold_binary_loc): Consistently avoid
963 canonicalizing X & CST away from a CST that is the mask
966 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
968 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
969 int argument to enum machine_mode.
970 (picochip_class_max_nregs): Ditto.
971 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
972 (picochip_class_max_nregs): Ditto.
974 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
976 * target.def: Add new target hook.
977 * doc/tm.texi: Regenerate.
978 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
979 * targhooks.c (default_keep_leaf_when_profiled): New function.
981 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
982 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
984 2014-05-05 Bin Cheng <bin.cheng@arm.com>
986 PR tree-optimization/60363
987 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
988 (copy_phi_args): New parameters. Call get_value_locus_in_path.
989 (update_destination_phis): New parameter.
990 (create_edge_and_update_destination_phis): Ditto.
991 (ssa_fix_duplicate_block_edges): Pass new arguments.
992 (thread_single_edge): Ditto.
994 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
996 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
997 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
998 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
999 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
1000 Use RS6000_BTM_HARD_FLOAT.
1001 (BU_MISC_2): Likewise.
1002 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
1003 RS6000_BTM_HARD_FLOAT.
1004 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
1006 (rs6000_invalid_builtin): Add hard floating builtin support.
1007 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
1008 hard float builtins.
1009 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
1011 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1013 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
1014 Add missing function* argument.
1016 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1018 * lra-constraints.c (valid_address_p): Move earlier in file.
1019 Add a constraint argument to the address_info version.
1020 (satisfies_memory_constraint_p): New function.
1021 (satisfies_address_constraint_p): Likewise.
1022 (process_alt_operands, curr_insn_transform): Use them.
1023 (process_address): Pass the constraint to valid_address_p when
1024 checking address operands.
1026 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1028 * config/mips/mips.c (mips_isa_rev): New variable.
1029 (mips_set_architecture): Set it.
1030 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1032 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1033 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1034 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1035 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1036 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
1037 conditions in terms of mips_isa_rev.
1038 (mips_isa_rev): Declare.
1040 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1042 * config/sh/sh-mem.cc: Use tabs instead of spaces.
1043 (prob_unlikely, prob_likely): Make variables const.
1045 2014-05-03 Denis Chertykov <chertykov@gmail.com>
1047 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
1049 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1051 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
1053 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1055 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
1056 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
1057 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
1059 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
1060 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
1062 Replace usage of ROUND_REG with sh_round_reg.
1063 Use CEIL instead of ROUND_ADVANCE.
1065 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1068 * config/sh/sh.c: Include stdlib headers before everything else.
1070 2014-05-02 Jakub Jelinek <jakub@redhat.com>
1072 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
1073 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
1074 (gimplify_adjust_omp_clauses): Simd region is never
1075 directly nested in combined parallel. Instead, for linear
1076 with copyin/copyout, if in combined for simd loop, make decl
1077 firstprivate/lastprivate on OMP_FOR.
1078 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
1079 expand_omp_for_static_chunk): When setting endvar, also set
1080 fd->loop.v to the same value.
1082 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1084 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
1086 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
1088 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
1091 2014-05-02 Marek Polacek <polacek@redhat.com>
1093 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
1095 2014-05-02 Kito Cheng <kito@0xlab.org>
1097 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
1098 to a C expression marco.
1099 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
1100 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
1101 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
1102 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
1103 HONOR_REG_ALLOC_ORDER.
1104 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
1106 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1108 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
1110 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1112 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
1114 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
1116 * tree-if-conv.c (is_cond_scalar_reduction): New function.
1117 (convert_scalar_cond_reduction): Likewise.
1118 (predicate_scalar_phi): Add recognition and transformation
1119 of simple conditioanl reduction to be vectorizable.
1121 2014-05-01 Marek Polacek <polacek@redhat.com>
1124 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
1126 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
1128 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
1129 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
1130 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
1131 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
1132 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
1133 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
1134 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
1135 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
1137 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
1139 * config/arc/arc.opt (mlra): Move comment above option name
1140 to avoid mis-parsing as language options.
1142 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1144 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
1145 * config/sol2.h: ... here.
1146 * config/sol2-10.h: Remove.
1148 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
1149 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
1150 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
1151 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
1152 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
1153 * config/sol2.h: ... here.
1154 (SECTION_NAME_FORMAT): Don't redefine.
1155 (STARTFILE_ARCH32_SPEC): Rename to ...
1156 (STARTFILE_ARCH_SPEC): ... this.
1157 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
1158 * config/sparc/sol2.h: ... here.
1159 (SECTION_NAME_FORMAT): Don't undef.
1160 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
1161 (SUBTARGET_EXTRA_SPECS): Remove.
1162 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
1164 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
1165 (MD_STARTFILE_PREFIX): Remove.
1166 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
1167 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
1168 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
1169 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
1170 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
1171 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
1172 * config/i386/sol2.h: ... here.
1173 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
1174 * config/i386/sol2-bi.h: Remove.
1175 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
1176 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
1178 * config/i386/t-sol2-64: Rename to ...
1179 * config/i386/t-sol2: ... this.
1180 * config/sparc/t-sol2-64: Rename to ...
1181 * config/sparc/t-sol2: ... this.
1183 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
1184 sol2_tm_file_head, sol2_tm_file_tail.
1185 Include ${cpu_type}/sol2.h before sol2.h.
1187 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
1188 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
1189 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
1190 Reflect i386/t-sol2-64 renaming.
1191 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
1192 Reflect sparc/t-sol2-64 renaming.
1194 2014-04-30 Richard Biener <rguenther@suse.de>
1196 * passes.c (execute_function_todo): Move TODO_verify_stmts
1197 and TODO_verify_ssa under the TODO_verify_il umbrella.
1198 * tree-ssa.h (verify_ssa): Adjust prototype.
1199 * tree-ssa.c (verify_ssa): Add parameter to tell whether
1200 we should verify SSA operands.
1201 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
1202 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
1203 whether we should verify whether not throwing stmts have EH info.
1204 * graphite-scop-detection.c (create_sese_edges): Adjust.
1205 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
1206 * tree-eh.c (lower_try_finally_switch): Do not add the
1207 default case label twice.
1209 2014-04-30 Marek Polacek <polacek@redhat.com>
1211 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
1212 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
1213 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
1214 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
1216 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
1218 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
1219 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
1220 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
1221 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
1222 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
1223 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
1224 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
1225 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
1227 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1229 * tree-cfg.c (dump_function_to_file): Dump the return type of
1230 functions, in a line to itself before the function body, mimicking
1231 the layout of a C function.
1233 2014-04-29 Jakub Jelinek <jakub@redhat.com>
1235 PR tree-optimization/60971
1236 * tree-tailcall.c (process_assignment): Reject conversions which
1239 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
1241 * calls.c (initialize_argument_information): Always treat
1242 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
1243 (expand_call): Likewise.
1244 (emit_library_call_calue_1): Likewise.
1245 * expr.c (PUSH_ARGS_REVERSED): Do not define.
1246 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
1249 2014-04-29 Nick Clifton <nickc@redhat.com>
1251 * config/msp430/msp430.md (umulsidi): Fix typo.
1252 (mulhisi3): Enable even inside interrupt handlers.
1253 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
1254 bigger return address pushed in large mode.
1256 2014-04-29 Nick Clifton <nickc@redhat.com>
1258 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
1259 (arc_init_reg_tables): Use a machine_mode enum to iterate over
1261 * config/m32r/m32r.c (init_reg_tables): Likewise.
1262 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
1263 enum to hold the modes.
1265 2014-04-29 Richard Biener <rguenther@suse.de>
1267 * dominance.c (free_dominance_info): Add overload with
1269 (dom_info_state): Likewise.
1270 (dom_info_available_p): Likewise.
1271 * basic-block.h (free_dominance_info, dom_info_state,
1272 dom_info_available_p): Declare overloads.
1273 * passes.c (execute_function_todo): Verify that verifiers
1274 don't change dominator info state. Drop dominator info
1275 for IPA pass invocations.
1276 * cgraph.c (release_function_body): Restore asserts that
1277 dominator information is released.
1279 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
1281 * doc/invoke.texi: Fix typo.
1282 * tree-vrp.c: Fix typos.
1283 * gimple.c (infer_nonnull_range): Reorder operands of an &&
1286 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1288 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
1290 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1292 * config/aarch64/aarch64-builtins.c
1293 (aarch64_types_storestruct_lane_qualifiers): New.
1294 (TYPES_STORESTRUCT_LANE): Likewise.
1295 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
1296 (st3_lane): Likewise.
1297 (st4_lane): Likewise.
1298 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
1299 (vec_store_lanesci_lane<mode>): Likewise.
1300 (vec_store_lanesxi_lane<mode>): Likewise.
1301 (aarch64_st2_lane<VQ:mode>): Likewise.
1302 (aarch64_st3_lane<VQ:mode>): Likewise.
1303 (aarch64_st4_lane<VQ:mode>): Likewise.
1304 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
1305 * config/aarch64/arm_neon.h
1306 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
1307 use new macro arguments.
1308 (__ST3_LANE_FUNC): Likewise.
1309 (__ST4_LANE_FUNC): Likewise.
1310 * config/aarch64/iterators.md (V_TWO_ELEM): New.
1311 (V_THREE_ELEM): Likewise.
1312 (V_FOUR_ELEM): Likewise.
1314 2014-04-28 David Malcolm <dmalcolm@redhat.com>
1316 * doc/gimple.texi: Replace the description of the now-defunct
1317 union gimple_statement_d with a diagram showing the
1318 gimple_statement_base class hierarchy and its relationships to
1319 the GSS_ and GIMPLE_ enums.
1321 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1323 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
1324 * config/aarch64/aarch64.c
1325 (aarch64_cannot_change_mode_class): Weaken conditions.
1326 (aarch64_modes_tieable_p): New.
1327 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
1329 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
1331 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
1332 (loadsync_<mode>): Change mode.
1333 (load_quadpti, store_quadpti): New.
1334 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
1335 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
1337 2014-04-28 Martin Jambor <mjambor@suse.cz>
1339 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
1340 same alias type as the original statement.
1341 (subreplacement_assignment_data): New type.
1342 (handle_unscalarized_data_in_subtree): New type of parameter,
1343 generate new memory accesses with same alias type as the original
1345 (load_assign_lhs_subreplacements): Likewise.
1346 (sra_modify_constructor_assign): Generate new memory accesses with
1347 same alias type as the original statement.
1349 2014-04-28 Richard Biener <rguenther@suse.de>
1351 * tree-pass.h (TODO_verify_il): Define.
1352 (TODO_verify_all): Complete properly.
1353 * passes.c (execute_function_todo): Move existing loop-closed
1354 SSA verification under TODO_verify_il.
1355 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
1356 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
1357 Fix tree sharing issue.
1359 2014-04-28 Richard Biener <rguenther@suse.de>
1362 * builtins.def (DEF_C11_BUILTIN): Add.
1363 (BUILT_IN_ALIGNED_ALLOC): Likewise.
1364 * coretypes.h (enum function_class): Add function_c11_misc.
1365 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1366 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
1367 (call_may_clobber_ref_p_1): Likewise.
1368 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
1369 (mark_all_reaching_defs_necessary_1): Likewise.
1370 (propagate_necessity): Likewise.
1371 (eliminate_unnecessary_stmts): Likewise.
1372 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
1374 2014-04-28 Richard Biener <rguenther@suse.de>
1376 * tree-vrp.c (vrp_var_may_overflow): Remove.
1377 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
1378 with overflow immediately bump to one before that value and
1379 let iteration figure out overflow status.
1381 2014-04-28 Richard Biener <rguenther@suse.de>
1383 * configure.ac: Do valgrind header checks unconditionally.
1384 Add --enable-valgrind-annotations.
1385 * system.h: Guard valgrind header inclusion with
1386 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
1387 * alloc-pool.c (pool_alloc, pool_free): Use
1388 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
1389 to guard possibly dead code.
1390 * config.in: Regenerated.
1391 * configure: Likewise.
1393 2014-04-28 Jeff Law <law@redhat.com>
1395 PR tree-optimization/60902
1396 * tree-ssa-threadedge.c
1397 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
1398 over real defs when invalidating outputs from statements that do not
1399 produce useful outputs for threading.
1401 2014-04-28 Richard Biener <rguenther@suse.de>
1403 PR tree-optimization/60979
1404 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
1405 SCOPs that end in a block with a successor with abnormal
1408 2014-04-28 Richard Biener <rguenther@suse.de>
1410 * tree-pass.h (execute_pass_list): Adjust prototype.
1411 * passes.c (pass_manager::execute_early_local_passes):
1413 (do_per_function): Change callback signature, push all actual
1414 work to the callbals.
1415 (do_per_function_toporder): Likewise.
1416 (execute_function_dump): Adjust.
1417 (execute_function_todo): Likewise.
1418 (clear_last_verified): Likewise.
1419 (verify_curr_properties): Likewise.
1420 (update_properties_after_pass): Likewise.
1421 (execute_pass_list_1): Split out from ...
1422 (execute_pass_list): ... here. Adjust.
1423 (execute_ipa_pass_list): Likewise.
1424 * cgraphunit.c (cgraph_add_new_function): Adjust.
1425 (analyze_function): Likewise.
1426 (expand_function): Likewise.
1427 * cgraph.c (release_function_body): Free dominance info
1428 here instead of asserting it was magically freed elsewhere.
1430 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
1432 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
1433 * configure: Regenerate.
1434 * config/sparc/sparc.opt (muser-mode): New option.
1435 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
1437 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
1438 * doc/invoke.texi (SPARC options): Document -muser-mode.
1440 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
1442 * cselib.c (find_slot_memmode): Delete.
1443 (cselib_hasher): Change compare_type to a struct.
1444 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
1446 (preserve_constants_and_equivs): Adjust for new compare_type.
1447 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
1448 (wrap_constant): Delete.
1449 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
1451 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
1453 * doc/install.texi (Building with profile feedback): Remove
1456 2014-04-26 Tom de Vries <tom@codesourcery.com>
1458 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
1461 2014-04-25 Cary Coutant <ccoutant@google.com>
1464 * dwarf2out.c (should_move_die_to_comdat): A type definition
1465 can contain a subprogram definition, but don't move it to a
1467 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
1468 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
1470 (clone_tree_hash): Rename to...
1471 (clone_tree_partial): ...this; change callers. Copy
1472 DW_TAG_subprogram DIEs as declarations.
1473 (copy_decls_walk): Don't copy children of a declaration into a
1476 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
1479 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
1482 2014-04-25 Jiong Wang <jiong.wang@arm.com>
1484 * config/arm/predicates.md (call_insn_operand): Add long_call check.
1485 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
1487 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
1490 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1492 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
1495 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1497 PR tree-optimization/60930
1498 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
1499 creating a multiply candidate by folding two constant
1500 multiplicands when the result overflows.
1502 2014-04-25 Jakub Jelinek <jakub@redhat.com>
1504 PR tree-optimization/60960
1505 * tree-vect-generic.c (expand_vector_operation): Only call
1506 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
1508 2014-04-25 Tom de Vries <tom@codesourcery.com>
1510 * expr.c (clobber_reg_mode): New function.
1511 * expr.h (clobber_reg): New function.
1513 2014-04-25 Tom de Vries <tom@codesourcery.com>
1515 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
1518 2014-04-25 Radovan Obradovic <robradovic@mips.com>
1519 Tom de Vries <tom@codesourcery.com>
1521 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
1523 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
1524 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
1525 new argument to find_all_hard_reg_sets call.
1527 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1529 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
1530 Use HOST_WIDE_INT_C for mask literal.
1531 (aarch_rev16_shleft_mask_imm_p): Likewise.
1533 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
1536 * config/sparc/sparc.md (ashlsi3_extend): Delete.
1538 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
1540 PR preprocessor/56540
1541 * config/i386/i386-c.c (ix86_target_macros): Define
1542 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
1544 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1546 * configure.ac (tga_func): Remove.
1547 (LIB_TLS_SPEC): Remove.
1548 * configure: Regenerate.
1549 * config.in: Regenerate.
1550 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
1552 2014-04-25 Richard Biener <rguenther@suse.de>
1555 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
1556 call stmt use/clobber sets during stmt walk instead of
1557 walking the possibly incomplete set of caller edges.
1559 2014-04-25 Richard Biener <rguenther@suse.de>
1562 * passes.c (apply_ipa_transforms): Inline into only caller ...
1563 (execute_one_pass): ... here. Properly bring in function
1564 bodies for nodes we want to apply IPA transforms to.
1566 2014-04-24 Cong Hou <congh@google.com>
1568 PR tree-optimization/60896
1569 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
1570 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
1571 (vect_mark_pattern_stmts): Set the def type of all statements in
1572 PATTERN_DEF_SEQ as vect_internal_def.
1574 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
1576 * doc/extend.texi (PowerPC Built-in Functions): Document new
1577 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
1579 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
1581 * config/rs6000/predicates.md (const_0_to_3_operand): New
1582 predicate to match 0..3 integer constants.
1584 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
1585 to support adding miscellaneous builtin functions.
1586 (BU_DFP_MISC_2): Likewise.
1587 (BU_P7_MISC_1): Likewise.
1588 (BU_P7_MISC_2): Likewise.
1589 (BU_P8V_MISC_3): Likewise.
1590 (BU_MISC_1): Likewise.
1591 (BU_MISC_2): Likewise.
1592 (DIVWE): Add extended divide builtin functions.
1595 (DIVWEUO): Likewise.
1599 (DIVDEUO): Likewise.
1600 (DXEX): Add decimal floating-point builtin functions.
1603 (DDEDPDQ): Likewise.
1605 (DENBCDQ): Likewise.
1612 (CDTBCD): Add new BCD builtin functions.
1616 (BCDADD_LT): Likewise.
1617 (BCDADD_EQ): Likewise.
1618 (BCDADD_GT): Likewise.
1619 (BCDADD_OV): Likewise.
1621 (BCDSUB_LT): Likewise.
1622 (BCDSUB_EQ): Likewise.
1623 (BCDSUB_GT): Likewise.
1624 (BCDSUB_OV): Likewise.
1625 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
1626 (UNPACK_TD): Likewise.
1627 (PACK_TF): Likewise.
1628 (UNPACK_TF): Likewise.
1629 (UNPACK_TF_0): Likewise.
1630 (UNPACK_TF_1): Likewise.
1631 (PACK_V1TI): Likewise.
1632 (UNPACK_V1TI): Likewise.
1634 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
1635 support for decimal floating point builtin functions.
1636 (rs6000_expand_ternop_builtin): Add checks for the new builtin
1637 functions that take constant arguments.
1638 (rs6000_invalid_builtin): Add decimal floating point builtin
1640 (rs6000_init_builtins): Setup long double, _Decimal64, and
1641 _Decimal128 types for new builtin functions.
1642 (builtin_function_type): Set the unsigned flags appropriately for
1643 the new builtin functions.
1644 (rs6000_opt_masks): Add support for decimal floating point builtin
1647 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
1648 floating point builtin functions.
1649 (RS6000_BTM_COMMON): Likewise.
1650 (RS6000_BTI_long_double): Likewise.
1651 (RS6000_BTI_dfloat64): Likewise.
1652 (RS6000_BTI_dfloat128): Likewise.
1653 (long_double_type_internal_node): Likewise.
1654 (dfloat64_type_internal_node): Likewise.
1655 (dfloat128_type_internal_node): Likewise.
1657 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
1658 2.07 bcd arithmetic instructions.
1659 (UNSPEC_BCDSUB): Likewise.
1660 (UNSPEC_BCD_OVERFLOW): Likewise.
1661 (UNSPEC_BCD_ADD_SUB): Likewise.
1662 (bcd_add_sub): Likewise.
1663 (BCD_TEST): Likewise.
1664 (bcd<bcd_add_sub>): Likewise.
1665 (bcd<bcd_add_sub>_test): Likewise.
1666 (bcd<bcd_add_sub>_test2): Likewise.
1667 (bcd<bcd_add_sub>_<code>): Likewise.
1668 (peephole2 for combined bcd ops): Likewise.
1670 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
1671 decimal floating point builtin functions.
1672 (UNSPEC_DENBCD): Likewise.
1673 (UNSPEC_DXEX): Likewise.
1674 (UNSPEC_DIEX): Likewise.
1675 (UNSPEC_DSCLI): Likewise.
1676 (UNSPEC_DSCRI): Likewise.
1677 (D64_D128): Likewise.
1678 (dfp_suffix): Likewise.
1679 (dfp_ddedpd_<mode>): Likewise.
1680 (dfp_denbcd_<mode>): Likewise.
1681 (dfp_dxex_<mode>): Likewise.
1682 (dfp_diex_<mode>): Likewise.
1683 (dfp_dscli_<mode>): Likewise.
1684 (dfp_dscri_<mode>): Likewise.
1686 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
1688 (UNSPEC_CDTBCD): Likewise.
1689 (UNSPEC_CBCDTD): Likewise.
1690 (UNSPEC_DIVE): Add support for new extended divide builtin
1692 (UNSPEC_DIVEO): Likewise.
1693 (UNSPEC_DIVEU): Likewise.
1694 (UNSPEC_DIVEUO): Likewise.
1695 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
1696 pack/unpack 128-bit types.
1697 (UNSPEC_PACK_128BIT): Likewise.
1698 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
1699 (udiv<mode>3): Use idiv_ldiv mode attribute.
1700 (div<mode>3): Likewise.
1701 (addg6s): Add new BCD builtin functions.
1704 (UNSPEC_DIV_EXTEND): Add support for new extended divide
1706 (div_extend): Likewise.
1707 (div<div_extend>_<mode>"): Likewise.
1708 (FP128_64): Add support for new builtin functions to pack/unpack
1710 (unpack<mode>): Likewise.
1711 (unpacktf_0): Likewise.
1712 (unpacktf_1): Likewise.
1713 (unpack<mode>_dm): Likewise.
1714 (unpack<mode>_nodm): Likewise.
1715 (pack<mode>): Likewise.
1716 (unpackv1ti): Likewise.
1717 (packv1ti): Likewise.
1719 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
1721 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
1724 2014-04-24 Jakub Jelinek <jakub@redhat.com>
1726 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
1727 * gimplify.c (omp_is_private): Change last argument's type to int.
1728 Only diagnose lastprivate if the simd argument is 1, only diagnose
1729 linear if the simd argument is 2.
1730 (gimplify_omp_for): Adjust omp_is_private callers. When adding
1731 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
1732 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
1733 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
1734 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
1735 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
1736 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
1737 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
1738 * tree-nested.c (convert_nonlocal_omp_clauses,
1739 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
1741 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
1744 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
1747 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
1749 * flag-types.h (enum ivar_visibility): Add.
1751 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
1753 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
1754 function * argument.
1756 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
1758 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
1760 2014-04-24 Radovan Obradovic <robradovic@mips.com>
1761 Tom de Vries <tom@codesourcery.com>
1763 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
1764 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
1766 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
1767 * emit-rtl.c (try_split): Same.
1769 2014-04-24 Radovan Obradovic <robradovic@mips.com>
1770 Tom de Vries <tom@codesourcery.com>
1772 * common.opt (fuse-caller-save): New option.
1774 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
1776 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
1779 2014-04-24 Richard Biener <rguenther@suse.de>
1781 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
1782 during TER and instead use the sepops interface for expanding
1783 non-GIMPLE_SINGLE_RHS.
1785 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1787 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
1788 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
1790 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1792 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
1793 assembler 64-bit option.
1794 * configure: Regenerate.
1796 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1798 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
1799 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
1800 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
1801 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
1802 (TARGET_CRYPTO): Take TARGET_SIMD into account.
1804 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1806 * config/aarch64/aarch64-builtins.c
1807 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
1808 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
1809 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
1810 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
1812 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
1813 (Vrevsuff): New mode attribute.
1815 2014-04-24 Terry Guo <terry.guo@arm.com>
1817 * config/arm/arm.h (machine_function): Define variable
1818 after_arm_reorg here.
1819 * config/arm/arm.c (after_arm_reorg): Remove the definition.
1820 (arm_split_constant): Update the way to access variable
1823 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
1825 2014-04-23 Tom de Vries <tom@codesourcery.com>
1827 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
1829 2014-04-23 David Malcolm <dmalcolm@redhat.com>
1831 * is-a.h: Update comments to reflect the following changes to the
1832 "pointerness" of the API, making the template parameter match the
1833 return type, allowing use of is-a.h with typedefs of pointers.
1834 (is_a_helper::cast): Return a T rather then a pointer to a T, so
1835 that the return type matches the parameter to the is_a_helper.
1837 (dyn_cast): Likewise.
1839 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
1840 pointer from the is-a.h API.
1842 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
1843 (is_a_helper <cgraph_node *>::test): ...this, matching change to
1845 (is_a_helper <varpool_node>::test): Likewise, convert to...
1846 (is_a_helper <varpool_node *>::test): ...this.
1848 (varpool_first_variable): Update for removal of implicit pointer
1849 from the is-a.h API.
1850 (varpool_next_variable): Likewise.
1851 (varpool_first_static_initializer): Likewise.
1852 (varpool_next_static_initializer): Likewise.
1853 (varpool_first_defined_variable): Likewise.
1854 (varpool_next_defined_variable): Likewise.
1855 (cgraph_first_defined_function): Likewise.
1856 (cgraph_next_defined_function): Likewise.
1857 (cgraph_first_function): Likewise.
1858 (cgraph_next_function): Likewise.
1859 (cgraph_first_function_with_gimple_body): Likewise.
1860 (cgraph_next_function_with_gimple_body): Likewise.
1861 (cgraph_alias_target): Likewise.
1862 (varpool_alias_target): Likewise.
1863 (cgraph_function_or_thunk_node): Likewise.
1864 (varpool_variable_node): Likewise.
1865 (symtab_real_symbol_p): Likewise.
1866 * cgraphunit.c (referred_to_p): Likewise.
1867 (analyze_functions): Likewise.
1868 (handle_alias_pairs): Likewise.
1869 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1870 * gimple-ssa.h (gimple_vuse_op): Likewise.
1871 (gimple_vdef_op): Likewise.
1872 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
1873 * gimple.c (gimple_build_asm_1): Likewise.
1874 (gimple_build_try): Likewise.
1875 (gimple_build_resx): Likewise.
1876 (gimple_build_eh_dispatch): Likewise.
1877 (gimple_build_omp_for): Likewise.
1878 (gimple_omp_for_set_clauses): Likewise.
1880 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
1881 (is_a_helper <gimple_statement_asm *>::test): ...this.
1882 (is_a_helper <gimple_statement_bind>::test): Convert to...
1883 (is_a_helper <gimple_statement_bind *>::test): ...this.
1884 (is_a_helper <gimple_statement_call>::test): Convert to...
1885 (is_a_helper <gimple_statement_call *>::test): ...this.
1886 (is_a_helper <gimple_statement_catch>::test): Convert to...
1887 (is_a_helper <gimple_statement_catch *>::test): ...this.
1888 (is_a_helper <gimple_statement_resx>::test): Convert to...
1889 (is_a_helper <gimple_statement_resx *>::test): ...this.
1890 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
1891 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
1892 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
1893 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
1894 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
1895 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
1896 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
1897 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
1898 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
1899 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
1900 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
1901 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
1902 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
1903 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
1904 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
1905 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
1906 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
1907 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
1908 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
1909 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
1910 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
1911 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
1912 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
1913 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
1914 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
1915 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
1916 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
1917 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
1918 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
1919 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
1920 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
1921 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
1922 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
1923 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
1924 (is_a_helper <gimple_statement_phi>::test): Convert to...
1925 (is_a_helper <gimple_statement_phi *>::test): ...this.
1926 (is_a_helper <gimple_statement_transaction>::test): Convert to...
1927 (is_a_helper <gimple_statement_transaction *>::test): ...this.
1928 (is_a_helper <gimple_statement_try>::test): Convert to...
1929 (is_a_helper <gimple_statement_try *>::test): ...this.
1930 (is_a_helper <gimple_statement_wce>::test): Convert to...
1931 (is_a_helper <gimple_statement_wce *>::test): ...this.
1932 (is_a_helper <const gimple_statement_asm>::test): Convert to...
1933 (is_a_helper <const gimple_statement_asm *>::test): ...this.
1934 (is_a_helper <const gimple_statement_bind>::test): Convert to...
1935 (is_a_helper <const gimple_statement_bind *>::test): ...this.
1936 (is_a_helper <const gimple_statement_call>::test): Convert to...
1937 (is_a_helper <const gimple_statement_call *>::test): ...this.
1938 (is_a_helper <const gimple_statement_catch>::test): Convert to...
1939 (is_a_helper <const gimple_statement_catch *>::test): ...this.
1940 (is_a_helper <const gimple_statement_resx>::test): Convert to...
1941 (is_a_helper <const gimple_statement_resx *>::test): ...this.
1942 (is_a_helper <const gimple_statement_eh_dispatch>::test):
1944 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
1946 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
1948 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
1949 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
1951 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
1953 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
1955 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
1957 (is_a_helper <const gimple_statement_omp_return>::test): Convert
1959 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
1960 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
1962 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
1963 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
1965 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
1966 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
1967 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
1968 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
1970 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
1971 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
1973 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
1974 (is_a_helper <const gimple_statement_omp_target>::test): Convert
1976 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
1977 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
1979 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
1980 (is_a_helper <const gimple_statement_omp_single>::test): Convert
1982 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
1983 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
1985 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
1986 (is_a_helper <const gimple_statement_omp_task>::test): Convert
1988 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
1989 (is_a_helper <const gimple_statement_phi>::test): Convert to...
1990 (is_a_helper <const gimple_statement_phi *>::test): ...this.
1991 (is_a_helper <const gimple_statement_transaction>::test): Convert
1993 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
1994 (is_a_helper <const gimple_statement_with_ops>::test): Convert
1996 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
1997 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
1998 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
1999 (is_a_helper <const gimple_statement_with_memory_ops>::test):
2001 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
2003 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
2005 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
2007 (gimple_use_ops): Update for removal of implicit pointer from the
2009 (gimple_set_use_ops): Likewise.
2010 (gimple_vuse): Likewise.
2011 (gimple_vdef): Likewise.
2012 (gimple_vuse_ptr): Likewise.
2013 (gimple_vdef_ptr): Likewise.
2014 (gimple_set_vuse): Likewise.
2015 (gimple_set_vdef): Likewise.
2016 (gimple_omp_return_set_lhs): Likewise.
2017 (gimple_omp_return_lhs): Likewise.
2018 (gimple_omp_return_lhs_ptr): Likewise.
2019 (gimple_call_fntype): Likewise.
2020 (gimple_call_set_fntype): Likewise.
2021 (gimple_call_set_internal_fn): Likewise.
2022 (gimple_call_use_set): Likewise.
2023 (gimple_call_clobber_set): Likewise.
2024 (gimple_bind_vars): Likewise.
2025 (gimple_bind_set_vars): Likewise.
2026 (gimple_bind_body_ptr): Likewise.
2027 (gimple_bind_set_body): Likewise.
2028 (gimple_bind_add_stmt): Likewise.
2029 (gimple_bind_block): Likewise.
2030 (gimple_bind_set_block): Likewise.
2031 (gimple_asm_ninputs): Likewise.
2032 (gimple_asm_noutputs): Likewise.
2033 (gimple_asm_nclobbers): Likewise.
2034 (gimple_asm_nlabels): Likewise.
2035 (gimple_asm_input_op): Likewise.
2036 (gimple_asm_input_op_ptr): Likewise.
2037 (gimple_asm_output_op): Likewise.
2038 (gimple_asm_output_op_ptr): Likewise.
2039 (gimple_asm_set_output_op): Likewise.
2040 (gimple_asm_clobber_op): Likewise.
2041 (gimple_asm_set_clobber_op): Likewise.
2042 (gimple_asm_label_op): Likewise.
2043 (gimple_asm_set_label_op): Likewise.
2044 (gimple_asm_string): Likewise.
2045 (gimple_catch_types): Likewise.
2046 (gimple_catch_types_ptr): Likewise.
2047 (gimple_catch_handler_ptr): Likewise.
2048 (gimple_catch_set_types): Likewise.
2049 (gimple_catch_set_handler): Likewise.
2050 (gimple_eh_filter_types): Likewise.
2051 (gimple_eh_filter_types_ptr): Likewise.
2052 (gimple_eh_filter_failure_ptr): Likewise.
2053 (gimple_eh_filter_set_types): Likewise.
2054 (gimple_eh_filter_set_failure): Likewise.
2055 (gimple_eh_must_not_throw_fndecl): Likewise.
2056 (gimple_eh_must_not_throw_set_fndecl): Likewise.
2057 (gimple_eh_else_n_body_ptr): Likewise.
2058 (gimple_eh_else_e_body_ptr): Likewise.
2059 (gimple_eh_else_set_n_body): Likewise.
2060 (gimple_eh_else_set_e_body): Likewise.
2061 (gimple_try_eval_ptr): Likewise.
2062 (gimple_try_cleanup_ptr): Likewise.
2063 (gimple_try_set_eval): Likewise.
2064 (gimple_try_set_cleanup): Likewise.
2065 (gimple_wce_cleanup_ptr): Likewise.
2066 (gimple_wce_set_cleanup): Likewise.
2067 (gimple_phi_capacity): Likewise.
2068 (gimple_phi_num_args): Likewise.
2069 (gimple_phi_result): Likewise.
2070 (gimple_phi_result_ptr): Likewise.
2071 (gimple_phi_set_result): Likewise.
2072 (gimple_phi_arg): Likewise.
2073 (gimple_phi_set_arg): Likewise.
2074 (gimple_resx_region): Likewise.
2075 (gimple_resx_set_region): Likewise.
2076 (gimple_eh_dispatch_region): Likewise.
2077 (gimple_eh_dispatch_set_region): Likewise.
2078 (gimple_omp_critical_name): Likewise.
2079 (gimple_omp_critical_name_ptr): Likewise.
2080 (gimple_omp_critical_set_name): Likewise.
2081 (gimple_omp_for_clauses): Likewise.
2082 (gimple_omp_for_clauses_ptr): Likewise.
2083 (gimple_omp_for_set_clauses): Likewise.
2084 (gimple_omp_for_collapse): Likewise.
2085 (gimple_omp_for_index): Likewise.
2086 (gimple_omp_for_index_ptr): Likewise.
2087 (gimple_omp_for_set_index): Likewise.
2088 (gimple_omp_for_initial): Likewise.
2089 (gimple_omp_for_initial_ptr): Likewise.
2090 (gimple_omp_for_set_initial): Likewise.
2091 (gimple_omp_for_final): Likewise.
2092 (gimple_omp_for_final_ptr): Likewise.
2093 (gimple_omp_for_set_final): Likewise.
2094 (gimple_omp_for_incr): Likewise.
2095 (gimple_omp_for_incr_ptr): Likewise.
2096 (gimple_omp_for_set_incr): Likewise.
2097 (gimple_omp_for_pre_body_ptr): Likewise.
2098 (gimple_omp_for_set_pre_body): Likewise.
2099 (gimple_omp_parallel_clauses): Likewise.
2100 (gimple_omp_parallel_clauses_ptr): Likewise.
2101 (gimple_omp_parallel_set_clauses): Likewise.
2102 (gimple_omp_parallel_child_fn): Likewise.
2103 (gimple_omp_parallel_child_fn_ptr): Likewise.
2104 (gimple_omp_parallel_set_child_fn): Likewise.
2105 (gimple_omp_parallel_data_arg): Likewise.
2106 (gimple_omp_parallel_data_arg_ptr): Likewise.
2107 (gimple_omp_parallel_set_data_arg): Likewise.
2108 (gimple_omp_task_clauses): Likewise.
2109 (gimple_omp_task_clauses_ptr): Likewise.
2110 (gimple_omp_task_set_clauses): Likewise.
2111 (gimple_omp_task_child_fn): Likewise.
2112 (gimple_omp_task_child_fn_ptr): Likewise.
2113 (gimple_omp_task_set_child_fn): Likewise.
2114 (gimple_omp_task_data_arg): Likewise.
2115 (gimple_omp_task_data_arg_ptr): Likewise.
2116 (gimple_omp_task_set_data_arg): Likewise.
2117 (gimple_omp_taskreg_clauses): Likewise.
2118 (gimple_omp_taskreg_clauses_ptr): Likewise.
2119 (gimple_omp_taskreg_set_clauses): Likewise.
2120 (gimple_omp_taskreg_child_fn): Likewise.
2121 (gimple_omp_taskreg_child_fn_ptr): Likewise.
2122 (gimple_omp_taskreg_set_child_fn): Likewise.
2123 (gimple_omp_taskreg_data_arg): Likewise.
2124 (gimple_omp_taskreg_data_arg_ptr): Likewise.
2125 (gimple_omp_taskreg_set_data_arg): Likewise.
2126 (gimple_omp_task_copy_fn): Likewise.
2127 (gimple_omp_task_copy_fn_ptr): Likewise.
2128 (gimple_omp_task_set_copy_fn): Likewise.
2129 (gimple_omp_task_arg_size): Likewise.
2130 (gimple_omp_task_arg_size_ptr): Likewise.
2131 (gimple_omp_task_set_arg_size): Likewise.
2132 (gimple_omp_task_arg_align): Likewise.
2133 (gimple_omp_task_arg_align_ptr): Likewise.
2134 (gimple_omp_task_set_arg_align): Likewise.
2135 (gimple_omp_single_clauses): Likewise.
2136 (gimple_omp_single_clauses_ptr): Likewise.
2137 (gimple_omp_single_set_clauses): Likewise.
2138 (gimple_omp_target_clauses): Likewise.
2139 (gimple_omp_target_clauses_ptr): Likewise.
2140 (gimple_omp_target_set_clauses): Likewise.
2141 (gimple_omp_target_child_fn): Likewise.
2142 (gimple_omp_target_child_fn_ptr): Likewise.
2143 (gimple_omp_target_set_child_fn): Likewise.
2144 (gimple_omp_target_data_arg): Likewise.
2145 (gimple_omp_target_data_arg_ptr): Likewise.
2146 (gimple_omp_target_set_data_arg): Likewise.
2147 (gimple_omp_teams_clauses): Likewise.
2148 (gimple_omp_teams_clauses_ptr): Likewise.
2149 (gimple_omp_teams_set_clauses): Likewise.
2150 (gimple_omp_sections_clauses): Likewise.
2151 (gimple_omp_sections_clauses_ptr): Likewise.
2152 (gimple_omp_sections_set_clauses): Likewise.
2153 (gimple_omp_sections_control): Likewise.
2154 (gimple_omp_sections_control_ptr): Likewise.
2155 (gimple_omp_sections_set_control): Likewise.
2156 (gimple_omp_for_set_cond): Likewise.
2157 (gimple_omp_for_cond): Likewise.
2158 (gimple_omp_atomic_store_set_val): Likewise.
2159 (gimple_omp_atomic_store_val): Likewise.
2160 (gimple_omp_atomic_store_val_ptr): Likewise.
2161 (gimple_omp_atomic_load_set_lhs): Likewise.
2162 (gimple_omp_atomic_load_lhs): Likewise.
2163 (gimple_omp_atomic_load_lhs_ptr): Likewise.
2164 (gimple_omp_atomic_load_set_rhs): Likewise.
2165 (gimple_omp_atomic_load_rhs): Likewise.
2166 (gimple_omp_atomic_load_rhs_ptr): Likewise.
2167 (gimple_omp_continue_control_def): Likewise.
2168 (gimple_omp_continue_control_def_ptr): Likewise.
2169 (gimple_omp_continue_set_control_def): Likewise.
2170 (gimple_omp_continue_control_use): Likewise.
2171 (gimple_omp_continue_control_use_ptr): Likewise.
2172 (gimple_omp_continue_set_control_use): Likewise.
2173 (gimple_transaction_body_ptr): Likewise.
2174 (gimple_transaction_label): Likewise.
2175 (gimple_transaction_label_ptr): Likewise.
2176 (gimple_transaction_set_body): Likewise.
2177 (gimple_transaction_set_label): Likewise.
2179 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
2180 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2181 * ipa-ref.c (ipa_record_reference): Likewise.
2182 * ipa-reference.c (analyze_function): Likewise.
2183 (ipa_reference_write_optimization_summary): Likewise.
2184 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
2185 (address_taken_from_non_vtable_p): Likewise.
2186 (comdat_can_be_unshared_p_1): Likewise.
2187 * lto-cgraph.c (lto_output_ref): Likewise.
2188 (add_references): Likewise.
2189 (compute_ltrans_boundary): Likewise.
2190 (output_symtab): Likewise.
2191 (input_ref): Likewise.
2192 (input_cgraph_1): Likewise.
2193 (output_cgraph_opt_summary): Likewise.
2194 * lto-streamer-out.c (lto_output): Likewise.
2195 (output_symbol_p): Likewise.
2196 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
2197 (lsei_start_function_in_partition): Likewise.
2198 (lsei_next_variable_in_partition): Likewise.
2199 (lsei_start_variable_in_partition): Likewise.
2200 * symtab.c (insert_to_assembler_name_hash): Likewise.
2201 (unlink_from_assembler_name_hash): Likewise.
2202 (symtab_unregister_node): Likewise.
2203 (symtab_remove_node): Likewise.
2204 (dump_symtab_node): Likewise.
2205 (verify_symtab_base): Likewise.
2206 (verify_symtab_node): Likewise.
2207 (symtab_make_decl_local): Likewise.
2208 (symtab_alias_ultimate_target): Likewise.
2209 (symtab_resolve_alias): Likewise.
2210 (symtab_get_symbol_partitioning_class): Likewise.
2211 * tree-phinodes.c (allocate_phi_node): Likewise.
2212 (reserve_phi_args_for_new_edge): Likewise.
2213 (remove_phi_args): Likewise.
2214 * varpool.c (varpool_node_for_asm): Likewise.
2215 (varpool_remove_unreferenced_decls): Likewise.
2217 2014-04-23 Jeff Law <law@redhat.com>
2219 PR tree-optimization/60902
2220 * tree-ssa-threadedge.c
2221 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
2222 invalidate outputs from statements that do not produce useful
2223 outputs for threading.
2225 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2227 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
2228 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
2229 machine descriptions for Stack Smashing Protector.
2231 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
2233 * aarch64.md (<optab>_rol<mode>3): New pattern.
2234 (<optab>_rolsi3_uxtw): Likewise.
2235 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
2237 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
2239 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
2240 (arm_cortex_a12_tune): Likewise.
2242 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2244 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
2246 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2248 * config/arm/arm.md (arm_rev16si2): New pattern.
2249 (arm_rev16si2_alt): Likewise.
2250 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
2252 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2254 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
2255 (rev16<mode>2_alt): Likewise.
2256 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
2257 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
2258 (aarch_rev16_shleft_mask_imm_p): Likewise.
2259 (aarch_rev16_p_1): Likewise.
2260 (aarch_rev16_p): Likewise.
2261 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
2262 (aarch_rev16_shright_mask_imm_p): Likewise.
2263 (aarch_rev16_shleft_mask_imm_p): Likewise.
2265 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2267 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
2268 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
2270 (cortex_a53_extra_costs): Likewise.
2271 (cortex_a57_extra_costs): Likewise.
2272 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
2273 (cortexa7_extra_costs): Likewise.
2274 (cortexa8_extra_costs): Likewise.
2275 (cortexa12_extra_costs): Likewise.
2276 (cortexa15_extra_costs): Likewise.
2277 (v7m_extra_costs): Likewise.
2278 (arm_new_rtx_costs): Handle BSWAP.
2280 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2282 * config/arm/arm.c (cortexa8_extra_costs): New table.
2283 (arm_cortex_a8_tune): New tuning struct.
2284 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
2286 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2288 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
2290 2014-04-23 Richard Biener <rguenther@suse.de>
2292 * Makefile.in (OBJS): Remove loop-unswitch.o.
2293 * loop-unswitch.c: Delete.
2294 * tree-pass.h (make_pass_rtl_unswitch): Remove.
2295 * passes.def (pass_rtl_unswitch): Likewise.
2296 * loop-init.c (gate_rtl_unswitch): Likewise.
2297 (rtl_unswitch): Likewise.
2298 (pass_data_rtl_unswitch): Likewise.
2299 (pass_rtl_unswitch): Likewise.
2300 (make_pass_rtl_unswitch): Likewise.
2301 * rtl.h (reversed_condition): Likewise.
2302 (compare_and_jump_seq): Likewise.
2303 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
2305 * loop-unroll.c (compare_and_jump_seq): Likewise.
2307 2014-04-23 Richard Biener <rguenther@suse.de>
2309 PR tree-optimization/60903
2310 * tree-ssa-loop-im.c (analyze_memory_references): Remove
2311 commented code block.
2312 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
2313 loop flags to newly created BBs and edges.
2315 2014-04-23 Nick Clifton <nickc@redhat.com>
2317 * config/msp430/msp430.c (msp430_handle_option): Move function
2319 (msp430_option_override): Simplify mcu and mcpu option handling.
2320 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
2321 support for -mhwmult command line option.
2322 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
2323 -mhwmult command line option.
2324 (msp430_hwmult_enabled): Delete.
2325 (msp43o_output_labelref): Add support for -mhwmult command line
2327 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
2328 (umulsidi3): Likewise.
2329 * config/msp430/msp430.opt (mmcu): Add Report attribute.
2330 (mcpu, mlarge, msmall): Likewise.
2331 (mhwmult): New option.
2332 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
2334 (msp430_is_f5_mcu): Remove prototype.
2335 (msp430_use_f5_series_hwmult): Add prototype.
2336 * config/msp430/msp430-opts.h: New file.
2337 * common/config/msp430: New directory.
2338 * common/config/msp430/msp430-common.c: New file.
2339 * config.gcc (msp430): Remove target_has_targetm_common.
2340 * doc/invoke.texi: Document -mhwmult command line option.
2342 2014-04-23 Nick Clifton <nickc@redhat.com>
2344 * config/i386/cygwin.h (ENDFILE_SPEC): Include
2345 default-manifest.o if it can be found in the search path.
2346 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
2348 2014-04-23 Terry Guo <terry.guo@arm.com>
2350 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
2352 2014-04-23 Richard Biener <rguenther@suse.de>
2355 * tree-inline.c (declare_return_variable): Use mark_addressable.
2357 2014-04-23 Richard Biener <rguenther@suse.de>
2360 * loop-init.c (loop_optimizer_init): Make sure to apply
2361 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
2363 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2366 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
2368 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
2369 if flag_sanitize_undefined_trap_on_error.
2370 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
2371 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
2372 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
2373 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
2374 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
2375 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
2376 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
2377 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
2378 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
2379 * ubsan.c (ubsan_instrument_unreachable): Return
2380 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
2381 (ubsan_expand_null_ifn): Emit __builtin_trap ()
2382 if flag_sanitize_undefined_trap_on_error and
2383 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
2384 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
2385 instrument_bool_enum_load): Emit __builtin_trap () if
2386 flag_sanitize_undefined_trap_on_error and
2387 __builtin_handle_*_abort () if !flag_sanitize_recover.
2388 * doc/invoke.texi (-fsanitize-recover,
2389 -fsanitize-undefined-trap-on-error): Document.
2391 2014-04-22 Christian Bruel <christian.bruel@st.com>
2393 * config/sh/sh.md (mov<mode>): Replace movQIHI.
2394 Force immediates to SImode.
2396 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
2398 * config/nios2/nios2.md (UNSPEC_ROUND): New.
2400 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
2401 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
2402 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
2403 (nios2_fpu_insn): Add entry for round.
2404 (N2FPU_NO_ERRNO_P): Define.
2405 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
2407 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
2409 2014-04-22 Richard Henderson <rth@redhat.com>
2411 * config/aarch64/aarch64 (addti3, subti3): New expanders.
2412 (add<GPI>3_compare0): Remove leading * from name.
2413 (add<GPI>3_carryin): Likewise.
2414 (sub<GPI>3_compare0): Likewise.
2415 (sub<GPI>3_carryin): Likewise.
2416 (<su_optab>mulditi3): New expander.
2417 (multi3): New expander.
2418 (madd<GPI>): Remove leading * from name.
2420 2014-04-22 Martin Jambor <mjambor@suse.cz>
2422 * cgraphclones.c (cgraph_function_versioning): Copy
2423 ipa_transforms_to_apply instead of asserting it is empty.
2425 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
2428 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
2429 on count_exp to get mode.
2431 2014-04-22 Andrew Pinski <apinski@cavium.com>
2433 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
2434 Handle TLS for ILP32.
2435 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
2436 (tlsie_small_<mode>): this and handle PTR.
2437 (tlsie_small_sidi): New pattern.
2438 (tlsle_small): Change to an expand to handle ILP32.
2439 (tlsle_small_<mode>): New pattern.
2440 (tlsdesc_small): Rename to ...
2441 (tlsdesc_small_<mode>): this and handle PTR.
2443 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2445 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
2447 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2449 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2450 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
2451 (aarch64_types_signed_poly_qualifiers): Likewise.
2452 (aarch64_types_unsigned_signed_qualifiers): Likewise.
2453 (aarch64_types_poly_signed_qualifiers): Likewise.
2454 (TYPES_REINTERP_SS): Type macro added.
2455 (TYPES_REINTERP_SU): Likewise.
2456 (TYPES_REINTERP_SP): Likewise.
2457 (TYPES_REINTERP_US): Likewise.
2458 (TYPES_REINTERP_PS): Likewise.
2459 (aarch64_fold_builtin): New expression folding added.
2460 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
2461 Declarations removed.
2462 (REINTERP_SS): Declarations added.
2463 (REINTERP_US): Likewise.
2464 (REINTERP_PS): Likewise.
2465 (REINTERP_SU): Likewise.
2466 (REINTERP_SP): Likewise.
2467 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
2468 (vreinterpretq_p8_f64): Likewise.
2469 (vreinterpret_p16_f64): Likewise.
2470 (vreinterpretq_p16_f64): Likewise.
2471 (vreinterpret_f32_f64): Likewise.
2472 (vreinterpretq_f32_f64): Likewise.
2473 (vreinterpret_f64_f32): Likewise.
2474 (vreinterpret_f64_p8): Likewise.
2475 (vreinterpret_f64_p16): Likewise.
2476 (vreinterpret_f64_s8): Likewise.
2477 (vreinterpret_f64_s16): Likewise.
2478 (vreinterpret_f64_s32): Likewise.
2479 (vreinterpret_f64_s64): Likewise.
2480 (vreinterpret_f64_u8): Likewise.
2481 (vreinterpret_f64_u16): Likewise.
2482 (vreinterpret_f64_u32): Likewise.
2483 (vreinterpret_f64_u64): Likewise.
2484 (vreinterpretq_f64_f32): Likewise.
2485 (vreinterpretq_f64_p8): Likewise.
2486 (vreinterpretq_f64_p16): Likewise.
2487 (vreinterpretq_f64_s8): Likewise.
2488 (vreinterpretq_f64_s16): Likewise.
2489 (vreinterpretq_f64_s32): Likewise.
2490 (vreinterpretq_f64_s64): Likewise.
2491 (vreinterpretq_f64_u8): Likewise.
2492 (vreinterpretq_f64_u16): Likewise.
2493 (vreinterpretq_f64_u32): Likewise.
2494 (vreinterpretq_f64_u64): Likewise.
2495 (vreinterpret_s64_f64): Likewise.
2496 (vreinterpretq_s64_f64): Likewise.
2497 (vreinterpret_u64_f64): Likewise.
2498 (vreinterpretq_u64_f64): Likewise.
2499 (vreinterpret_s8_f64): Likewise.
2500 (vreinterpretq_s8_f64): Likewise.
2501 (vreinterpret_s16_f64): Likewise.
2502 (vreinterpretq_s16_f64): Likewise.
2503 (vreinterpret_s32_f64): Likewise.
2504 (vreinterpretq_s32_f64): Likewise.
2505 (vreinterpret_u8_f64): Likewise.
2506 (vreinterpretq_u8_f64): Likewise.
2507 (vreinterpret_u16_f64): Likewise.
2508 (vreinterpretq_u16_f64): Likewise.
2509 (vreinterpret_u32_f64): Likewise.
2510 (vreinterpretq_u32_f64): Likewise.
2512 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2514 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2515 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
2516 (vreinterpret_p8_s8): Likewise.
2517 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
2518 (vreinterpret_p8_s16): Likewise.
2519 (vreinterpret_p8_s32): Likewise.
2520 (vreinterpret_p8_s64): Likewise.
2521 (vreinterpret_p8_f32): Likewise.
2522 (vreinterpret_p8_u8): Likewise.
2523 (vreinterpret_p8_u16): Likewise.
2524 (vreinterpret_p8_u32): Likewise.
2525 (vreinterpret_p8_u64): Likewise.
2526 (vreinterpret_p8_p16): Likewise.
2527 (vreinterpretq_p8_s8): Likewise.
2528 (vreinterpretq_p8_s16): Likewise.
2529 (vreinterpretq_p8_s32): Likewise.
2530 (vreinterpretq_p8_s64): Likewise.
2531 (vreinterpretq_p8_f32): Likewise.
2532 (vreinterpretq_p8_u8): Likewise.
2533 (vreinterpretq_p8_u16): Likewise.
2534 (vreinterpretq_p8_u32): Likewise.
2535 (vreinterpretq_p8_u64): Likewise.
2536 (vreinterpretq_p8_p16): Likewise.
2537 (vreinterpret_p16_s8): Likewise.
2538 (vreinterpret_p16_s16): Likewise.
2539 (vreinterpret_p16_s32): Likewise.
2540 (vreinterpret_p16_s64): Likewise.
2541 (vreinterpret_p16_f32): Likewise.
2542 (vreinterpret_p16_u8): Likewise.
2543 (vreinterpret_p16_u16): Likewise.
2544 (vreinterpret_p16_u32): Likewise.
2545 (vreinterpret_p16_u64): Likewise.
2546 (vreinterpret_p16_p8): Likewise.
2547 (vreinterpretq_p16_s8): Likewise.
2548 (vreinterpretq_p16_s16): Likewise.
2549 (vreinterpretq_p16_s32): Likewise.
2550 (vreinterpretq_p16_s64): Likewise.
2551 (vreinterpretq_p16_f32): Likewise.
2552 (vreinterpretq_p16_u8): Likewise.
2553 (vreinterpretq_p16_u16): Likewise.
2554 (vreinterpretq_p16_u32): Likewise.
2555 (vreinterpretq_p16_u64): Likewise.
2556 (vreinterpretq_p16_p8): Likewise.
2557 (vreinterpret_f32_s8): Likewise.
2558 (vreinterpret_f32_s16): Likewise.
2559 (vreinterpret_f32_s32): Likewise.
2560 (vreinterpret_f32_s64): Likewise.
2561 (vreinterpret_f32_u8): Likewise.
2562 (vreinterpret_f32_u16): Likewise.
2563 (vreinterpret_f32_u32): Likewise.
2564 (vreinterpret_f32_u64): Likewise.
2565 (vreinterpret_f32_p8): Likewise.
2566 (vreinterpret_f32_p16): Likewise.
2567 (vreinterpretq_f32_s8): Likewise.
2568 (vreinterpretq_f32_s16): Likewise.
2569 (vreinterpretq_f32_s32): Likewise.
2570 (vreinterpretq_f32_s64): Likewise.
2571 (vreinterpretq_f32_u8): Likewise.
2572 (vreinterpretq_f32_u16): Likewise.
2573 (vreinterpretq_f32_u32): Likewise.
2574 (vreinterpretq_f32_u64): Likewise.
2575 (vreinterpretq_f32_p8): Likewise.
2576 (vreinterpretq_f32_p16): Likewise.
2577 (vreinterpret_s64_s8): Likewise.
2578 (vreinterpret_s64_s16): Likewise.
2579 (vreinterpret_s64_s32): Likewise.
2580 (vreinterpret_s64_f32): Likewise.
2581 (vreinterpret_s64_u8): Likewise.
2582 (vreinterpret_s64_u16): Likewise.
2583 (vreinterpret_s64_u32): Likewise.
2584 (vreinterpret_s64_u64): Likewise.
2585 (vreinterpret_s64_p8): Likewise.
2586 (vreinterpret_s64_p16): Likewise.
2587 (vreinterpretq_s64_s8): Likewise.
2588 (vreinterpretq_s64_s16): Likewise.
2589 (vreinterpretq_s64_s32): Likewise.
2590 (vreinterpretq_s64_f32): Likewise.
2591 (vreinterpretq_s64_u8): Likewise.
2592 (vreinterpretq_s64_u16): Likewise.
2593 (vreinterpretq_s64_u32): Likewise.
2594 (vreinterpretq_s64_u64): Likewise.
2595 (vreinterpretq_s64_p8): Likewise.
2596 (vreinterpretq_s64_p16): Likewise.
2597 (vreinterpret_u64_s8): Likewise.
2598 (vreinterpret_u64_s16): Likewise.
2599 (vreinterpret_u64_s32): Likewise.
2600 (vreinterpret_u64_s64): Likewise.
2601 (vreinterpret_u64_f32): Likewise.
2602 (vreinterpret_u64_u8): Likewise.
2603 (vreinterpret_u64_u16): Likewise.
2604 (vreinterpret_u64_u32): Likewise.
2605 (vreinterpret_u64_p8): Likewise.
2606 (vreinterpret_u64_p16): Likewise.
2607 (vreinterpretq_u64_s8): Likewise.
2608 (vreinterpretq_u64_s16): Likewise.
2609 (vreinterpretq_u64_s32): Likewise.
2610 (vreinterpretq_u64_s64): Likewise.
2611 (vreinterpretq_u64_f32): Likewise.
2612 (vreinterpretq_u64_u8): Likewise.
2613 (vreinterpretq_u64_u16): Likewise.
2614 (vreinterpretq_u64_u32): Likewise.
2615 (vreinterpretq_u64_p8): Likewise.
2616 (vreinterpretq_u64_p16): Likewise.
2617 (vreinterpret_s8_s16): Likewise.
2618 (vreinterpret_s8_s32): Likewise.
2619 (vreinterpret_s8_s64): Likewise.
2620 (vreinterpret_s8_f32): Likewise.
2621 (vreinterpret_s8_u8): Likewise.
2622 (vreinterpret_s8_u16): Likewise.
2623 (vreinterpret_s8_u32): Likewise.
2624 (vreinterpret_s8_u64): Likewise.
2625 (vreinterpret_s8_p8): Likewise.
2626 (vreinterpret_s8_p16): Likewise.
2627 (vreinterpretq_s8_s16): Likewise.
2628 (vreinterpretq_s8_s32): Likewise.
2629 (vreinterpretq_s8_s64): Likewise.
2630 (vreinterpretq_s8_f32): Likewise.
2631 (vreinterpretq_s8_u8): Likewise.
2632 (vreinterpretq_s8_u16): Likewise.
2633 (vreinterpretq_s8_u32): Likewise.
2634 (vreinterpretq_s8_u64): Likewise.
2635 (vreinterpretq_s8_p8): Likewise.
2636 (vreinterpretq_s8_p16): Likewise.
2637 (vreinterpret_s16_s8): Likewise.
2638 (vreinterpret_s16_s32): Likewise.
2639 (vreinterpret_s16_s64): Likewise.
2640 (vreinterpret_s16_f32): Likewise.
2641 (vreinterpret_s16_u8): Likewise.
2642 (vreinterpret_s16_u16): Likewise.
2643 (vreinterpret_s16_u32): Likewise.
2644 (vreinterpret_s16_u64): Likewise.
2645 (vreinterpret_s16_p8): Likewise.
2646 (vreinterpret_s16_p16): Likewise.
2647 (vreinterpretq_s16_s8): Likewise.
2648 (vreinterpretq_s16_s32): Likewise.
2649 (vreinterpretq_s16_s64): Likewise.
2650 (vreinterpretq_s16_f32): Likewise.
2651 (vreinterpretq_s16_u8): Likewise.
2652 (vreinterpretq_s16_u16): Likewise.
2653 (vreinterpretq_s16_u32): Likewise.
2654 (vreinterpretq_s16_u64): Likewise.
2655 (vreinterpretq_s16_p8): Likewise.
2656 (vreinterpretq_s16_p16): Likewise.
2657 (vreinterpret_s32_s8): Likewise.
2658 (vreinterpret_s32_s16): Likewise.
2659 (vreinterpret_s32_s64): Likewise.
2660 (vreinterpret_s32_f32): Likewise.
2661 (vreinterpret_s32_u8): Likewise.
2662 (vreinterpret_s32_u16): Likewise.
2663 (vreinterpret_s32_u32): Likewise.
2664 (vreinterpret_s32_u64): Likewise.
2665 (vreinterpret_s32_p8): Likewise.
2666 (vreinterpret_s32_p16): Likewise.
2667 (vreinterpretq_s32_s8): Likewise.
2668 (vreinterpretq_s32_s16): Likewise.
2669 (vreinterpretq_s32_s64): Likewise.
2670 (vreinterpretq_s32_f32): Likewise.
2671 (vreinterpretq_s32_u8): Likewise.
2672 (vreinterpretq_s32_u16): Likewise.
2673 (vreinterpretq_s32_u32): Likewise.
2674 (vreinterpretq_s32_u64): Likewise.
2675 (vreinterpretq_s32_p8): Likewise.
2676 (vreinterpretq_s32_p16): Likewise.
2677 (vreinterpret_u8_s8): Likewise.
2678 (vreinterpret_u8_s16): Likewise.
2679 (vreinterpret_u8_s32): Likewise.
2680 (vreinterpret_u8_s64): Likewise.
2681 (vreinterpret_u8_f32): Likewise.
2682 (vreinterpret_u8_u16): Likewise.
2683 (vreinterpret_u8_u32): Likewise.
2684 (vreinterpret_u8_u64): Likewise.
2685 (vreinterpret_u8_p8): Likewise.
2686 (vreinterpret_u8_p16): Likewise.
2687 (vreinterpretq_u8_s8): Likewise.
2688 (vreinterpretq_u8_s16): Likewise.
2689 (vreinterpretq_u8_s32): Likewise.
2690 (vreinterpretq_u8_s64): Likewise.
2691 (vreinterpretq_u8_f32): Likewise.
2692 (vreinterpretq_u8_u16): Likewise.
2693 (vreinterpretq_u8_u32): Likewise.
2694 (vreinterpretq_u8_u64): Likewise.
2695 (vreinterpretq_u8_p8): Likewise.
2696 (vreinterpretq_u8_p16): Likewise.
2697 (vreinterpret_u16_s8): Likewise.
2698 (vreinterpret_u16_s16): Likewise.
2699 (vreinterpret_u16_s32): Likewise.
2700 (vreinterpret_u16_s64): Likewise.
2701 (vreinterpret_u16_f32): Likewise.
2702 (vreinterpret_u16_u8): Likewise.
2703 (vreinterpret_u16_u32): Likewise.
2704 (vreinterpret_u16_u64): Likewise.
2705 (vreinterpret_u16_p8): Likewise.
2706 (vreinterpret_u16_p16): Likewise.
2707 (vreinterpretq_u16_s8): Likewise.
2708 (vreinterpretq_u16_s16): Likewise.
2709 (vreinterpretq_u16_s32): Likewise.
2710 (vreinterpretq_u16_s64): Likewise.
2711 (vreinterpretq_u16_f32): Likewise.
2712 (vreinterpretq_u16_u8): Likewise.
2713 (vreinterpretq_u16_u32): Likewise.
2714 (vreinterpretq_u16_u64): Likewise.
2715 (vreinterpretq_u16_p8): Likewise.
2716 (vreinterpretq_u16_p16): Likewise.
2717 (vreinterpret_u32_s8): Likewise.
2718 (vreinterpret_u32_s16): Likewise.
2719 (vreinterpret_u32_s32): Likewise.
2720 (vreinterpret_u32_s64): Likewise.
2721 (vreinterpret_u32_f32): Likewise.
2722 (vreinterpret_u32_u8): Likewise.
2723 (vreinterpret_u32_u16): Likewise.
2724 (vreinterpret_u32_u64): Likewise.
2725 (vreinterpret_u32_p8): Likewise.
2726 (vreinterpret_u32_p16): Likewise.
2727 (vreinterpretq_u32_s8): Likewise.
2728 (vreinterpretq_u32_s16): Likewise.
2729 (vreinterpretq_u32_s32): Likewise.
2730 (vreinterpretq_u32_s64): Likewise.
2731 (vreinterpretq_u32_f32): Likewise.
2732 (vreinterpretq_u32_u8): Likewise.
2733 (vreinterpretq_u32_u16): Likewise.
2734 (vreinterpretq_u32_u64): Likewise.
2735 (vreinterpretq_u32_p8): Likewise.
2736 (vreinterpretq_u32_p16): Likewise.
2738 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2740 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
2742 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
2745 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
2746 (vqnegd_s64): Likewise.
2747 (vqabs_s64): Likewise.
2748 (vqabsd_s64): Likewise.
2750 2014-04-22 Richard Henderson <rth@redhat.com>
2752 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
2753 computation to the top of the loop.
2755 2014-04-22 Renlin <renlin.li@arm.com>
2756 Jiong Wang <jiong.wang@arm.com>
2758 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
2759 * config/aarch64/aarch64.c (aarch64_layout_frame)
2760 (aarch64_initial_elimination_offset): Likewise.
2762 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
2764 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
2767 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
2769 * machmode.h (bitwise_mode_for_mode): Declare.
2770 * stor-layout.h (bitwise_type_for_mode): Likewise.
2771 * stor-layout.c (bitwise_mode_for_mode): New function.
2772 (bitwise_type_for_mode): Likewise.
2773 * builtins.c (fold_builtin_memory_op): Use it instead of
2774 int_mode_for_mode and build_nonstandard_integer_type.
2776 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2778 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
2779 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
2780 (*-*-solaris2*): Simplify.
2781 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
2782 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
2783 *-*-solaris2.9* handling.
2785 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
2787 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
2788 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
2790 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
2791 * configure: Regenerate.
2793 * config/i386/sol2-9.h: Remove.
2795 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
2796 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
2797 Remove Solaris 9 references.
2799 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
2801 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
2802 (floatuns<GPI:mode><GPF:mode>2): Remove.
2803 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
2804 and floatuns conversions.
2805 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
2806 and floatuns conversions.
2807 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
2808 (w1,w2): New mode attributes for inequal width conversions.
2810 2014-04-22 Renlin Li <Renlin.Li@arm.com>
2812 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
2813 the output asm format.
2815 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
2817 * config/aarch64/aarch64-simd.md
2818 (aarch64_cm<optab>di): Always split.
2819 (*aarch64_cm<optab>di): New.
2820 (aarch64_cmtstdi): Always split.
2821 (*aarch64_cmtstdi): New.
2823 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2825 PR tree-optimization/60823
2826 * omp-low.c (ipa_simd_modify_function_body): Go through
2827 all SSA_NAMEs and for those refering to vector arguments
2828 which are going to be replaced adjust SSA_NAME_VAR and,
2829 if it is a default definition, change it into a non-default
2830 definition assigned at the beginning of function from new_decl.
2831 (ipa_simd_modify_stmt_ops): Rewritten.
2832 * tree-dfa.c (set_ssa_default_def): When removing default def,
2833 check for NULL loc instead of NULL *loc.
2835 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2837 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
2838 restrictions on core registers for DImode values in Thumb2.
2840 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2842 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
2843 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
2845 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2847 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
2848 (*iordi_notzesidi_di): Likewise.
2849 (*iordi_notsesidi_di): Likewise.
2851 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2853 * config/arm/arm-protos.h (tune_params): New struct members.
2854 * config/arm/arm.c: Initialise tune_params per processor.
2855 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
2856 for speed, based on new tune_params.
2858 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2860 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
2862 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
2864 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
2866 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
2867 * config/aarch64/arm_neon.h (vrnd_f64): Added.
2868 (vrnda_f64): Likewise.
2869 (vrndi_f64): Likewise.
2870 (vrndm_f64): Likewise.
2871 (vrndn_f64): Likewise.
2872 (vrndp_f64): Likewise.
2873 (vrndx_f64): Likewise.
2875 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2877 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
2878 GET_MODE_SIZE argument is enum machine_mode.
2880 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2883 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
2884 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
2886 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
2889 * asan.c (asan_emit_stack_protection): Force the base to align to
2890 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
2891 appropriate bits if STRICT_ALIGNMENT.
2892 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
2894 (expand_used_vars): Leave a space in the stack frame for alignment
2895 if STRICT_ALIGNMENT.
2897 2014-04-21 David Malcolm <dmalcolm@redhat.com>
2899 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
2901 (gimple_store_p): Likewise.
2902 (gimple_assign_load_p): Likewise.
2903 (gimple_assign_cast_p): Likewise.
2904 (gimple_clobber_p): Likewise.
2906 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
2907 rather than a gimple.
2908 (gimple_assign_cast_p): Likewise.
2910 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
2913 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
2914 If mode is DDmode and TARGET_E500_DOUBLE allow move.
2916 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
2917 more debug information for E500 if -mdebug=reg.
2919 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
2922 * config/i386/i386.c (ix86_expand_builtin)
2923 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
2924 register for target RTX.
2925 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
2927 2014-04-18 Cong Hou <congh@google.com>
2929 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
2930 the widen-mult pattern by handling two operands with different sizes,
2931 and operands whose size is smaller than half of the result type.
2933 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
2935 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
2936 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
2937 (do_estimate_edge_time): Compute it.
2938 * ipa-inline.c (want_inline_small_function_p): Bypass
2939 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
2941 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
2943 * ipa-inline.c (spec_rem): New static variable.
2944 (dump_overall_stats): New function.
2945 (dump_inline_stats): New function.
2947 2014-04-18 Richard Henderson <rth@redhat.com>
2949 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
2950 to GET_MODE_SIZE, not a reg_class_t.
2952 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2954 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
2955 (vsx_xxmrglw_<mode>): Likewise.
2957 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2960 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
2961 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
2962 (rs6000_init_hard_regno_mode_ok): Likewise.
2964 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
2966 * ipa-inline.c (inline_small_functions): Account only non-cold
2968 * doc/invoke.texi (inline-unit-growth): Update documentation.
2970 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
2972 * config/rs6000/rs6000.md (addti3, subti3): New.
2974 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
2977 * config/i386/i386.c (ix86_expand_clear): Remove outdated
2978 comment. Check optimize_insn_for_size_p instead of
2979 optimize_insn_for_speed_p.
2981 2014-04-17 Martin Jambor <mjambor@suse.cz>
2983 * gimple-iterator.c (gsi_start_edge): New function.
2984 * gimple-iterator.h (gsi_start_edge): Declare.
2985 * tree-sra.c (single_non_eh_succ): New function.
2986 (disqualify_ops_if_throwing_stmt): Renamed to
2987 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
2988 having one non-EH successor BB.
2989 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
2990 generate loads into replacements.
2991 (sra_modify_assign): Likewise and and also use the simple path for
2993 (sra_modify_function_body): Commit statements on edges.
2995 2014-04-17 Richard Biener <rguenther@suse.de>
2998 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
2999 comparison results and add clarifying comment.
3001 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3003 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
3004 (blank_mode): Initialize it.
3005 (emit_mode_size_inline, emit_mode_nunits_inline,
3006 emit_mode_inner_inline): New functions.
3007 (emit_insn_modes_h): Call them and surround their output with
3008 #if GCC_VERSION >= 4001 ... #endif.
3009 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
3010 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
3011 mode_* arrays if the argument is __builtin_constant_p.
3012 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
3013 is enum machine_mode.
3015 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3017 * passes.c (opt_pass::execute): Adjust.
3018 (pass_manager::execute_pass_mode_switching): Likewise.
3019 (early_local_passes::execute): Likewise.
3020 (execute_one_pass): Pass cfun to the pass's execute method.
3021 * tree-pass.h (opt_pass::execute): Add function * argument.
3022 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3023 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3024 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3025 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3026 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3027 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
3028 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
3029 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3030 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3031 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3032 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3033 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3034 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3035 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3036 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3037 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3038 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3039 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3040 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3041 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3042 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3043 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3044 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3045 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3046 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3047 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3048 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3049 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3050 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3053 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3055 * passes.c (opt_pass::gate): Take function * argument.
3056 (gate_all_early_local_passes): Merge into
3057 (early_local_passes::gate): this.
3058 (gate_all_early_optimizations): Merge into
3059 (all_early_optimizations::gate): this.
3060 (gate_all_optimizations): Mege into
3061 (all_optimizations::gate): this.
3062 (gate_all_optimizations_g): Merge into
3063 (all_optimizations_g::gate): this.
3064 (gate_rest_of_compilation): Mege into
3065 (rest_of_compilation::gate): this.
3066 (gate_postreload): Merge into
3067 (postreload::gate): this.
3068 (dump_one_pass): Pass cfun to the pass's gate method.
3069 (execute_ipa_summary_passes): Likewise.
3070 (execute_one_pass): Likewise.
3071 (ipa_write_summaries_2): Likewise.
3072 (ipa_write_optimization_summaries_1): Likewise.
3073 (ipa_read_summaries_1): Likewise.
3074 (ipa_read_optimization_summaries_1): Likewise.
3075 (execute_ipa_stmt_fixups): Likewise.
3076 * tree-pass.h (opt_pass::gate): Add function * argument.
3077 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
3078 combine-stack-adj.c, combine.c, compare-elim.c,
3079 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3080 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
3081 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
3082 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
3083 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3084 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3085 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3086 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
3087 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
3088 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
3089 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3090 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3091 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
3092 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3093 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3094 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3095 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3096 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3097 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3098 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3099 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3100 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3101 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3102 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
3103 var-tracking.c, vtable-verify.c, web.c: Adjust.
3105 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3107 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
3108 * configure: Regenerate.
3110 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3112 * passes.c (dump_one_pass): don't check pass->has_gate.
3113 (execute_ipa_summary_passes): Likewise.
3114 (execute_one_pass): Likewise.
3115 (ipa_write_summaries_2): Likewise.
3116 (ipa_write_optimization_summaries_1): Likewise.
3117 (ipa_read_optimization_summaries_1): Likewise.
3118 (execute_ipa_stmt_fixups): Likewise.
3119 * tree-pass.h (pass_data::has_gate): Remove.
3120 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3121 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3122 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3123 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3124 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3125 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
3126 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
3127 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
3128 gimple-low.c, gimple-ssa-isolate-paths.c,
3129 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
3130 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3131 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3132 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
3133 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
3134 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
3135 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
3136 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3137 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
3138 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3139 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3140 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3141 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3142 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3143 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3144 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3145 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3146 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3147 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3148 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3149 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3150 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3153 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3155 * pass_manager.h (pass_manager::register_dump_files_1): Remove
3157 * passes.c (pass_manager::register_dump_files_1): Merge into
3158 (pass_manager::register_dump_files): this, and remove its handling of
3159 properties since the pass always has the properties anyway.
3160 (pass_manager::pass_manager): Adjust.
3162 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3164 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
3165 * passes.c (pass_manager::register_dump_files_1): Remove dead code
3166 dealing with properties.
3167 (pass_manager::register_dump_files): Adjust.
3169 2014-03-20 Mark Wielaard <mjw@redhat.com>
3171 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
3172 then represent the bound as normal constant value.
3174 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3177 Forward port from 4.8 branch
3178 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
3180 * config/i386/bmiintrin.h (_blsi_u32): New.
3184 (_blsmsk_u32): Ditto.
3185 (_blsmsk_u64): Ditto.
3186 (_tzcnt_u32): Ditto.
3187 (_tzcnt_u64): Ditto.
3189 2014-04-17 Kito Cheng <kito@0xlab.org>
3191 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
3193 2014-04-17 Richard Biener <rguenther@suse.de>
3196 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
3197 boolean results for comparisons.
3199 2014-04-17 Richard Biener <rguenther@suse.de>
3201 PR tree-optimization/60836
3202 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
3203 initial PHI args to be gimple values.
3205 2014-04-17 Richard Biener <rguenther@suse.de>
3207 PR tree-optimization/60841
3208 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
3209 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
3210 of stmts to SLP build.
3211 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
3212 (vect_analyze_slp): Likewise.
3213 (vect_analyze_slp_instance): Likewise.
3214 (vect_build_slp_tree): Limit overall SLP tree growth.
3215 * tree-vectorizer.h (vect_analyze_data_refs,
3216 vect_analyze_slp): Adjust prototypes.
3218 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3220 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
3223 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3225 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
3226 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
3227 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
3228 for TARGET_SLOW_PSHUFB
3230 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3232 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
3233 * config/i386/i386.c (intel_cost): Ditto.
3235 2014-04-17 Joey Ye <joey.ye@arm.com>
3237 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
3239 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3241 * opts.c (common_handle_option): Disable -fipa-reference coorectly
3244 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3246 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
3247 (type_all_derivations_known_p): New predicate.
3248 (type_all_ctors_visible_p): New predicate.
3249 (type_possibly_instantiated_p): New predicate.
3250 (get_odr_type): Compute all_derivations_known.
3251 (dump_odr_type): Dump the flag.
3252 (maybe_record_type): Cleanup.
3253 (record_target_from_binfo): Add bases_to_consider array;
3254 record bases for types w/o instances and skip CXX destructor.
3255 (possible_polymorphic_call_targets_1): Add bases_to_consider
3256 and consider_construction parameters; check if type may have instance.
3257 (get_polymorphic_call_info): Set maybe_in_construction to true
3258 when we know nothing.
3259 (record_targets_from_bases): Skip CXX destructors; they are
3260 never called for types in construction.
3261 (possible_polymorphic_call_targets): Do not record target when
3262 type may not have instance.
3264 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3267 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
3268 external aliases alive, too.
3270 2014-04-16 Andrew Pinski <apinski@cavium.com>
3272 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
3275 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3277 * final.c (compute_alignments): Do not apply loop alignment to a block
3278 falling through to the exit.
3280 2014-04-16 Catherine Moore <clm@codesourcery.com>
3282 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
3283 Adjust constraints for microMIPS store patterns.
3285 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3287 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
3289 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3291 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
3292 (append_use): Run at -O0.
3293 (append_vdef): Likewise.
3294 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
3295 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
3297 2014-04-16 Jakub Jelinek <jakub@redhat.com>
3299 PR tree-optimization/60844
3300 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
3301 (propagate_op_to_single_use, remove_visited_stmt_chain,
3302 linearize_expr, repropagate_negates, reassociate_bb): Use it
3303 instead of gsi_remove.
3305 2014-04-16 Martin Jambor <mjambor@suse.cz>
3307 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
3308 ipa_transforms_to_apply.
3309 (cgraph_function_versioning): Assert that old_node has empty
3310 ipa_transforms_to_apply.
3311 * trans-mem.c (ipa_tm_create_version): Likewise.
3312 * tree-inline.c (tree_function_versioning): Do not duplicate
3313 ipa_transforms_to_apply.
3315 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3318 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
3320 Pass necessary as flags on 64-bit Solaris/x86.
3321 Use lowercase relocs for x86_64-*-*.
3322 * configure: Regenerate.
3324 2014-04-15 Jan Hubicka <jh@suse.cz>
3326 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
3327 (maybe_record_node, likely_target_p): Use it.
3329 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3332 Revert following patch
3334 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3337 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3338 software floating point or no floating point registers, do not
3339 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3340 in GPRs that occurs after we tested for GPRs that would never be
3343 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3344 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3345 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3346 specifically allow DDmode, since that does not use the SPE SIMD
3349 2014-03-21 Mark Wielaard <mjw@redhat.com>
3351 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
3352 as unsigned or int depending on type and value used.
3354 2014-04-15 Richard Biener <rguenther@suse.de>
3356 PR rtl-optimization/56965
3357 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
3358 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
3360 * alias.c (true_dependence_1): Do not call
3361 nonoverlapping_component_refs_p.
3362 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
3363 nonoverlapping_component_refs_p.
3364 (indirect_refs_may_alias_p): Likewise.
3366 2014-04-15 Teresa Johnson <tejohnson@google.com>
3368 * cfg.c (dump_bb_info): Fix flags check.
3369 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
3371 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3373 PR rtl-optimization/60663
3374 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
3377 2014-04-15 Richard Biener <rguenther@suse.de>
3379 * lto-streamer.h (LTO_major_version): Bump to 4.
3381 2014-04-15 Richard Biener <rguenther@suse.de>
3383 * common.opt (lto_partition_model): New enum.
3384 (flto-partition=): Merge separate options with a single with argument,
3385 add -flto-partition=one support.
3386 * flag-types.h (enum lto_partition_model): Declare.
3387 * opts.c (finish_options): Remove duplicate -flto-partition=
3389 * lto-wrapper.c (run_gcc): Adjust.
3391 2014-04-15 Richard Biener <rguenther@suse.de>
3393 * alias.c (ncr_compar): New function.
3394 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
3396 2014-04-15 Richard Biener <rguenther@suse.de>
3398 * alias.c (record_component_aliases): Do not walk BINFOs.
3400 2014-04-15 Richard Biener <rguenther@suse.de>
3402 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
3403 Add struct function argument and adjust.
3404 (find_func_aliases_for_call): Likewise.
3405 (find_func_aliases): Likewise.
3406 (find_func_clobbers): Likewise.
3407 (intra_create_variable_infos): Likewise.
3408 (compute_points_to_sets): Likewise.
3409 (ipa_pta_execute): Adjust. Do not push/pop cfun.
3411 2014-04-15 Richard Biener <rguenther@suse.de>
3413 * tree.c (iterative_hash_expr): Use enum tree_code_class
3414 to store TREE_CODE_CLASS.
3415 (tree_block): Likewise.
3416 (tree_set_block): Likewise.
3417 * tree.h (fold_build_pointer_plus_loc): Use
3418 convert_to_ptrofftype_loc.
3420 2014-04-15 Jakub Jelinek <jakub@redhat.com>
3423 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
3426 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
3428 * cfgloop.h (struct loop): Move force_vectorize down.
3429 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
3430 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
3431 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
3432 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
3433 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
3434 * tree-core.h (enum annot_expr_kind): Add new kind values.
3435 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
3436 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
3438 * tree.def (ANNOTATE_EXPR): Tweak comment.
3440 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3442 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
3445 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
3447 * tree.h (TYPE_IDENTIFIER): Declare.
3448 * tree.c (subrange_type_for_debug_p): Use it.
3449 * godump.c (go_format_type): Likewise.
3450 * dwarf2out.c (is_cxx_auto, modified_type_die,
3451 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
3452 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
3454 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3457 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
3459 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
3461 * config/i386/i386.c (examine_argument): Return bool. Return true if
3462 parameter should be passed in memory.
3463 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
3464 (construct_container): Update calls to examine_argument.
3465 (function_arg_advance_64): Ditto.
3466 (return_in_memory_32): Merge with ix86_return_in_memory.
3467 (return_in_memory_64): Ditto.
3468 (return_in_memory_ms_64): Ditto.
3470 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3472 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
3473 * coverage.c (coverage_compute_profile_id): Handle externally visible
3476 2014-04-14 Martin Jambor <mjambor@suse.cz>
3478 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
3479 DECL_DISREGARD_INLINE_LIMITS functions.
3481 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3484 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
3486 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3489 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
3490 optimize_insn_for_speed_p instead of
3491 optimize_function_for_speed_p.
3493 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
3495 * doc/invoke.texi (free): Document AArch64.
3497 2014-04-14 Richard Biener <rguenther@suse.de>
3499 PR tree-optimization/60042
3500 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
3501 (insert_into_preds_of_block): Do not prevent PHI insertion
3502 for REFERENCE exprs here ...
3503 (eliminate_dom_walker::before_dom_children): ... but prevent
3504 their use here under similar conditions when applied to the
3505 IL after PRE optimizations.
3507 2014-04-14 Richard Biener <rguenther@suse.de>
3509 * passes.def: Move early points-to after early SRA.
3511 2014-04-14 Richard Biener <rguenther@suse.de>
3513 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
3514 check for which sign-changes we allow when forwarding
3515 a converted value into a switch.
3517 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3519 * stor-layout.c (place_field): Finalize non-constant offset for the
3522 2014-04-14 Richard Biener <rguenther@suse.de>
3524 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
3526 (expand_switch_using_bit_tests_p): Likewise.
3527 (process_switch): Compute and pass on speed_p based on the
3529 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
3530 optimize_bb_for_speed_p.
3532 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3534 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
3535 * function.h (struct function): Rename has_force_vect_loops into
3536 has_force_vectorize_loops.
3537 * lto-streamer-in.c (input_cfg): Adjust for renaming.
3538 (input_struct_function_base): Likewise.
3539 * lto-streamer-out.c (output_cfg): Likewise.
3540 (output_struct_function_base): Likewise.
3541 * omp-low.c (expand_omp_simd): Likewise.
3542 * tree-cfg.c (move_sese_region_to_fn): Likewise.
3543 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
3544 (version_loop_for_if_conversion): Likewise.
3545 (tree_if_conversion): Likewise.
3546 (main_tree_if_conversion): Likewise.
3547 (gate_tree_if_conversion): Likewise.
3548 * tree-inline.c (copy_loops): Likewise.
3549 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
3550 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
3551 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
3552 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
3553 * tree-vectorizer.c (vectorize_loops): Likewise.
3554 * tree-vectorizer.h (unlimited_cost_model): Likewise.
3556 2014-04-14 Richard Biener <rguenther@suse.de>
3559 * lto-streamer-out.c (wrap_refs): New function.
3560 (lto_output): Wrap symbol references in global initializes in
3561 type-preserving MEM_REFs.
3563 2014-04-14 Christian Bruel <christian.bruel@st.com>
3565 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
3567 2014-04-14 Christian Bruel <christian.bruel@st.com>
3569 * config/sh/sh.md (setmemqi): New expand pattern.
3570 * config/sh/sh.h (CLEAR_RATIO): Define.
3571 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
3572 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
3574 2014-04-14 Richard Biener <rguenther@suse.de>
3577 * fold-const.c (negate_expr_p): Don't negate directional rounding
3579 (fold_negate_expr): Likewise.
3581 2014-04-14 Richard Biener <rguenther@suse.de>
3583 PR tree-optimization/59817
3584 PR tree-optimization/60453
3585 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
3586 recursion to catch all CHRECs in the scalar evolution and restrict
3587 the predicate for the remains appropriately.
3589 2014-04-12 Catherine Moore <clm@codesourcery.com>
3591 * config/mips/constraints.md: Add new register constraint "kb".
3592 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
3593 (*movhi_internal): Likewise.
3594 (*movqi_internal): Likewise.
3595 * config/mips/mips.h (M16_STORE_REGS): New register class.
3596 (REG_CLASS_NAMES): Add M16_STORE_REGS.
3597 (REG_CLASS_CONTENTS): Likewise.
3598 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
3600 2014-04-11 Tobias Burnus <burnus@net-b.de>
3603 * doc/invoke.texi (-Wformat-signedness): Document it.
3604 (Wformat=2): Mention that this enables -Wformat-signedness.
3606 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3608 * common/config/epiphany/epiphany-common.c
3609 (epiphany_option_optimization_table): Enable section anchors by
3610 default at -O1 or higher.
3611 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
3612 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
3613 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
3614 carries no extra cost.
3615 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
3616 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
3617 * config/epiphany/predicates.md (memclob_operand): New predicate.
3618 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
3619 Use memclob_operand predicate and X constraint for operand 3.
3621 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3623 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
3624 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
3627 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3629 PR rtl-optimization/60651
3630 * mode-switching.c (optimize_mode_switching): Make sure to emit
3631 sets of a lower numbered entity before sets of a higher numbered
3632 entity to a mode of the same or lower priority.
3633 When creating a seginfo for a basic block that starts with a code
3634 label, move the insertion point past the code label.
3635 (new_seginfo): Document and enforce requirement that
3636 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
3637 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
3638 * doc/tm.texi: Regenerate.
3640 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
3643 * config/arc/arc.c (arc_save_restore): Fix assert typo.
3645 2013-04-11 Jakub Jelinek <jakub@redhat.com>
3647 * BASE-VER: Set to 4.10.0.
3649 2014-04-11 Tobias Burnus <burnus@net-b.de>
3652 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
3653 * doc/gcc.texi (Service): Update description in the @menu
3654 * doc/invoke.texi (Option Summary): Remove misplaced and
3657 2014-04-11 Steve Ellcey <sellcey@mips.com>
3658 Jakub Jelinek <jakub@redhat.com>
3661 * expr.c (convert_move): Use emit_store_flag_force instead of
3662 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
3665 2014-04-11 Richard Biener <rguenther@suse.de>
3668 * varasm.c (assemble_alias): Avoid endless error reporting
3669 recursion by setting TREE_ASM_WRITTEN.
3671 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3673 * config/s390/s390.md: Add a splitter for NOT rtx.
3675 2014-04-11 Jakub Jelinek <jakub@redhat.com>
3677 PR rtl-optimization/60663
3678 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
3680 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
3681 Jakub Jelinek <jakub@redhat.com>
3684 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
3685 flag from decl_node to node.
3687 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3690 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
3691 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
3692 ameliorating the cases where it can be.
3694 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
3697 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
3699 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3700 (loadsync_<mode>): Change mode.
3701 (load_quadpti, store_quadpti): New.
3702 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3703 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3704 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
3706 2014-04-09 Cong Hou <congh@google.com>
3709 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
3712 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3714 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
3715 instead of vnor to exploit possible fusion opportunity in the
3717 (altivec_expand_vec_perm_const_le): Likewise.
3719 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
3721 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3722 (loadsync_<mode>): Change mode.
3723 (load_quadpti, store_quadpti): New.
3724 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3725 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3727 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
3730 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
3731 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
3732 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
3734 2014-04-08 Richard Biener <rguenther@suse.de>
3737 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
3738 a 64bit widest int print double-int similar to on HWI64 hosts.
3740 2014-04-08 Richard Biener <rguenther@suse.de>
3742 PR tree-optimization/60785
3743 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
3744 default defs properly.
3746 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
3748 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
3749 (Weffc++): Likewise.
3751 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
3753 * ipa-devirt.c (maybe_record_node): When node is not recorded,
3754 set completep to false rather than true.
3756 2014-04-07 Douglas B Rupp <rupp@adacore.com>
3759 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
3760 ARM_TARGET2_DWARF_FORMAT.
3762 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
3765 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
3766 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
3769 2014-04-07 Richard Biener <rguenther@suse.de>
3771 PR tree-optimization/60766
3772 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
3773 (may_eliminate_iv): Convert cand_value_at result to desired type.
3775 2014-04-07 Jason Merrill <jason@redhat.com>
3778 * common.opt (-fno-gnu-unique): Add.
3779 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
3781 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3783 * haifa-sched.c: Fix outdated function reference and minor
3784 grammar errors in introductory comment.
3786 2014-04-07 Richard Biener <rguenther@suse.de>
3789 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
3791 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
3793 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
3796 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
3797 size accounting for thunks.
3798 (pa_asm_output_mi_thunk): Use final_start_function() and
3799 final_end_function() to output function start and end directives.
3801 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3803 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
3804 device specific ISA/ feature information. Remove short_sp and
3805 errata_skip ds. Add avr_device_specific_features enum to have device
3807 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
3808 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
3809 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
3810 updated device specific info.
3811 * config/avr/avr-mcus.def: Merge device specific details to
3812 dev_attribute field.
3813 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
3815 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
3816 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
3817 assembler if RMW isa supported by current device.
3818 * config/avr/genmultilib.awk: Update as device info structure changed.
3819 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
3821 2014-04-04 Cong Hou <congh@google.com>
3823 PR tree-optimization/60656
3824 * tree-vect-stmts.c (supportable_widening_operation):
3825 Fix a bug that elements in a vector with vect_used_by_reduction
3826 property are incorrectly reordered when the operation on it is not
3827 consistant with the one in reduction operation.
3829 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
3831 PR rtl-optimization/60155
3832 * gcse.c (record_set_data): New function.
3833 (single_set_gcse): New function.
3834 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
3835 (hoist_code): Likewise.
3836 (get_pressure_class_and_nregs): Likewise.
3838 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
3840 * explow.c (probe_stack_range): Emit a final optimization blockage.
3842 2014-04-04 Anthony Green <green@moxielogic.com>
3844 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
3847 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
3850 * lto-cgraph.c (input_overwrite_node): Check that partitioning
3851 flags are set only during streaming.
3852 * ipa.c (process_references, walk_polymorphic_call_targets,
3853 symtab_remove_unreachable_nodes): Drop bodies of always inline
3854 after early inlining.
3855 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
3857 2014-04-04 Jakub Jelinek <jakub@redhat.com>
3858 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3861 * dwarf2out.c (const_ok_for_output_1): Reject expressions
3864 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3867 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
3869 (cortex_a53_fdivd): Likewise.
3871 2014-04-04 Martin Jambor <mjambor@suse.cz>
3874 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
3876 * cgraph.c (clone_of_p): Also return true if thunks match.
3877 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
3878 cgraph_function_or_thunk_node and an obsolete comment.
3879 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
3881 (build_function_decl_skip_args): Likewise.
3882 (set_new_clone_decl_and_node_flags): New function.
3883 (duplicate_thunk_for_node): Likewise.
3884 (redirect_edge_duplicating_thunks): Likewise.
3885 (cgraph_clone_node): New parameter args_to_skip, pass it to
3886 redirect_edge_duplicating_thunks which is called instead of
3887 cgraph_redirect_edge_callee.
3888 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
3889 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
3891 2014-04-04 Jeff Law <law@redhat.com>
3894 * config/arm/predicates.md (const_int_I_operand): New predicate.
3895 (const_int_M_operand): Similarly.
3896 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
3898 (insv_t2, extv_reg, extzv_t2): Likewise.
3899 (load_multiple_with_writeback): Similarly for const_int_I_operand.
3900 (pop_multiple_with_writeback_and_return): Likewise.
3901 (vfp_pop_multiple_with_writeback): Likewise
3903 2014-04-04 Richard Biener <rguenther@suse.de>
3906 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
3907 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
3909 * gimplify.h (gimple_add_tmp_var_fn): Declare.
3910 * gimplify.c (gimple_add_tmp_var_fn): New function.
3911 * gimple-expr.h (create_tmp_reg_fn): Declare.
3912 * gimple-expr.c (create_tmp_reg_fn): New function.
3913 * gimple-low.c (record_vars_into): Don't change cfun.
3914 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
3915 code generation without cfun.
3917 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
3920 * Makefile.in (install-driver): Fix shell scripting.
3922 2014-04-03 Cong Hou <congh@google.com>
3924 PR tree-optimization/60505
3925 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
3926 threshold of number of iterations below which no vectorization
3928 * tree-vect-loop.c (new_loop_vec_info):
3929 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
3930 * tree-vect-loop.c (vect_analyze_loop_operations):
3931 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
3932 * tree-vect-loop.c (vect_transform_loop):
3933 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
3934 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
3935 of iterations of the loop and see if we should build the epilogue.
3937 2014-04-03 Richard Biener <rguenther@suse.de>
3939 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
3940 (streamer_tree_cache_create): Adjust.
3941 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
3942 to allow optional nodes array.
3943 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
3944 (streamer_tree_cache_append): Likewise.
3945 (streamer_tree_cache_create): Create nodes array optionally
3946 as specified by parameter.
3947 * lto-streamer-out.c (create_output_block): Avoid maintaining
3948 the node array in the writer cache.
3949 (DFS_write_tree): Remove assertion.
3950 (produce_asm_for_decls): Free the out decl state hash table early.
3951 * lto-streamer-in.c (lto_data_in_create): Adjust for
3952 streamer_tree_cache_create prototype change.
3954 2014-04-03 Richard Biener <rguenther@suse.de>
3956 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
3957 set TREE_CHAIN to NULL_TREE.
3959 2014-04-03 Richard Biener <rguenther@suse.de>
3961 PR tree-optimization/60740
3962 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
3963 over all GIMPLE_COND operands.
3965 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
3967 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
3968 (Weffc++): Remove Scott's numbering, merge lists and reference
3971 2014-04-03 Nick Clifton <nickc@redhat.com>
3973 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
3976 2014-04-03 Martin Jambor <mjambor@suse.cz>
3978 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
3979 mention gcc_unreachable before failing.
3980 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
3983 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
3986 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
3987 inconsistent code and instead mark the context inconsistent.
3988 (possible_polymorphic_call_targets): For inconsistent contexts
3989 return empty complete list.
3991 2014-04-02 Anthony Green <green@moxielogic.com>
3993 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
3994 (extendqisi2, extendhisi2): Define.
3995 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
3996 (WCHAR_TYPE): Change to unsigned int.
3998 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4000 PR tree-optimization/60733
4001 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
4002 insertion point for PHI candidates to be the end of the feeding
4003 block for the PHI argument.
4005 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
4007 PR rtl-optimization/60650
4008 * lra-constraints.c (process_alt_operands): Decrease reject for
4009 earlyclobber matching.
4011 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4013 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
4015 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4017 * config/spu/spu.c (pad_bb): Do not crash when the last
4018 insn is CODE_FOR_blockage.
4020 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4022 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
4023 lies outside the target mode.
4025 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4028 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4029 software floating point or no floating point registers, do not
4030 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4031 in GPRs that occurs after we tested for GPRs that would never be
4034 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4035 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4036 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4037 specifically allow DDmode, since that does not use the SPE SIMD
4040 2014-04-02 Richard Biener <rguenther@suse.de>
4043 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4044 MODE_INTs. Properly use negv_optab.
4045 (expand_abs): Likewise.
4047 2014-04-02 Richard Biener <rguenther@suse.de>
4050 * Makefile.in (install-driver): Guard extra installs with special
4053 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4055 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4058 2014-04-01 Richard Henderson <rth@redhat.com>
4061 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
4062 alternative enabled before register allocation.
4064 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
4066 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
4067 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
4069 (nios2_large_got_address): Remove unneeded 'sym' parameter.
4070 (nios2_got_address): Update nios2_large_got_address call site.
4071 (nios2_delegitimize_address): New function.
4072 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
4073 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
4074 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
4076 2014-04-01 Martin Husemann <martin@duskware.de>
4078 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
4081 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
4083 PR rtl-optimization/60604
4084 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
4085 check from register_operand.
4086 (register_operand): Redefine in terms of general_operand.
4087 (nonmemory_operand): Use register_operand for the non-constant cases.
4089 2014-04-01 Richard Biener <rguenther@suse.de>
4091 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
4093 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
4095 * doc/invoke.texi (mapp-regs): Clarify.
4097 2014-03-31 Ulrich Drepper <drepper@gmail.com>
4099 * config/i386/avx512fintrin.h (__v32hi): Define type.
4100 (__v64qi): Likewise.
4101 (_mm512_set1_epi8): Define.
4102 (_mm512_set1_epi16): Define.
4103 (_mm512_set4_epi32): Define.
4104 (_mm512_set4_epi64): Define.
4105 (_mm512_set4_pd): Define.
4106 (_mm512_set4_ps): Define.
4107 (_mm512_setr4_epi64): Define.
4108 (_mm512_setr4_epi32): Define.
4109 (_mm512_setr4_pd): Define.
4110 (_mm512_setr4_ps): Define.
4111 (_mm512_setzero_epi32): Define.
4113 2014-03-31 Martin Jambor <mjambor@suse.cz>
4116 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
4117 callsite_arguments_match_p. Updated all callers. Also check types of
4118 corresponding formal parameters and actual arguments.
4119 (not_all_callers_have_enough_arguments_p) Renamed to
4120 some_callers_have_mismatched_arguments_p.
4122 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
4124 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
4126 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
4129 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
4132 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
4134 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
4135 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
4137 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
4138 Use FMAMODE_NOVF512 mode iterator.
4139 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
4140 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
4141 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
4142 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
4144 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
4145 Use VF_128_256 mode iterator.
4146 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
4149 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4151 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
4152 static chain if needed.
4154 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4157 * lra-constraints.c (index_part_to_reg): New.
4158 (process_address): Use it.
4160 2014-03-27 Jeff Law <law@redhat.com>
4161 Jakub Jelinek <jakub@redhat.com>
4164 * expr.c (do_tablejump): Use simplify_gen_binary rather than
4165 gen_rtx_{PLUS,MULT} to build up the address expression.
4167 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
4168 creating non-canonical RTL.
4170 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4173 * ipa-inline.c (want_inline_small_function_p): Short circuit large
4174 functions; reorganize to make cheap checks first.
4175 (inline_small_functions): Do not estimate growth when dumping;
4177 * ipa-inline.h (inline_summary): Add min_size.
4178 (growth_likely_positive): New function.
4179 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
4180 (set_cond_stmt_execution_predicate): Cleanup.
4181 (estimate_edge_size_and_time): Compute min_size.
4182 (estimate_calls_size_and_time): Likewise.
4183 (estimate_node_size_and_time): Likewise.
4184 (inline_update_overall_summary): Update min_size.
4185 (do_estimate_edge_time): Likewise.
4186 (do_estimate_edge_size): Update.
4187 (do_estimate_edge_hints): Update.
4188 (growth_likely_positive): New function.
4190 2014-03-28 Jakub Jelinek <jakub@redhat.com>
4193 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
4194 also if addr has VOIDmode.
4196 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4198 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
4199 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
4201 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
4202 instructions as well as AdvancedSIMD loads.
4204 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4206 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
4207 Use crypto_aese type.
4208 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
4209 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
4210 crypto_aese, crypto_aesmc. Move to types.md.
4211 * config/arm/types.md (crypto_aes): Split into crypto_aese,
4213 * config/arm/iterators.md (crypto_type): Likewise.
4215 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4217 * cgraph.c: Include expr.h and tree-dfa.h.
4218 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
4221 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4224 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
4225 regs from checking multi-reg pseudos.
4227 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4229 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
4231 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4233 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
4234 if it would clobber the stack pointer, even temporarily.
4236 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
4238 * mode-switching.c: Make small adjustments to the top comment.
4240 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4242 * config/rs6000/constraints.md (wD constraint): New constraint to
4243 match the constant integer to get the top DImode/DFmode out of a
4244 vector in a VSX register.
4246 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
4247 match the constant integer to get the top DImode/DFmode out of a
4248 vector in a VSX register.
4250 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
4253 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
4256 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
4257 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
4259 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
4260 Optimize vec_extract of 64-bit values, where the value being
4261 extracted is in the top word, where we can use scalar
4262 instructions. Add direct move and store support. Combine the big
4263 endian/little endian vector select load support into a single insn.
4264 (vsx_extract_<mode>_internal1): Likewise.
4265 (vsx_extract_<mode>_internal2): Likewise.
4266 (vsx_extract_<mode>_load): Likewise.
4267 (vsx_extract_<mode>_store): Likewise.
4268 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
4269 combined into vsx_extract_<mode>_load.
4270 (vsx_extract_<mode>_one_le): Likewise.
4272 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
4273 define the top 64-bit vector element.
4275 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
4278 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4279 Document vec_vbpermq builtin.
4282 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
4283 enable use of xxsldwi and xxpermdi builtin functions.
4284 (vec_xxpermdi): Likewise.
4286 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4287 Document use of vec_xxsldwi and vec_xxpermdi builtins.
4289 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
4291 PR rtl-optimization/60650
4292 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
4294 (find_spills_for): New.
4295 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
4296 Spill all pseudos on the second iteration.
4298 2014-03-27 Marek Polacek <polacek@redhat.com>
4301 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
4304 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4306 * config/s390/s390.c (s390_can_use_return_insn): Check for
4307 call-saved FPRs on 31 bit.
4309 2014-03-27 Jakub Jelinek <jakub@redhat.com>
4312 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
4313 if they need regimplification, just drop them instead of
4314 calling gimple_regimplify_operands on them.
4316 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
4319 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
4320 (aarch64_frame_pointer_required): Adjust logic.
4321 (aarch64_can_eliminate): Adjust logic.
4322 (aarch64_override_options_after_change): Adjust logic.
4324 2014-03-27 Dehao Chen <dehao@google.com>
4326 * ipa-inline.c (early_inliner): Update node's inline info.
4328 2014-03-26 Dehao Chen <dehao@google.com>
4330 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
4331 compiler inserted conditional jumps for NAN float check.
4333 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4335 * ubsan.h (ubsan_create_data): Change second argument's type
4336 to const location_t *.
4337 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
4339 (ubsan_create_data): Change second argument to const location_t *PLOC.
4340 Create Loc field whenever PLOC is non-NULL.
4341 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
4342 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
4346 * real.c (real_to_integer2): Change type of low to UHWI.
4348 2014-03-26 Tobias Burnus <burnus@net-b.de>
4350 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
4351 (CILK_SELF_SPECS): New define.
4352 (driver_self_specs): Use it.
4354 2014-03-26 Richard Biener <rguenther@suse.de>
4356 * tree-pretty-print.c (percent_K_format): Implement special
4357 case for LTO and its stripped down BLOCK tree.
4359 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4362 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
4364 * tree-vrp.c (simplify_internal_call_using_ranges): If only
4365 one range is range_int_cst_p, but not both, at least optimize
4366 addition/subtraction of 0 and multiplication by 0 or 1.
4367 * gimple-fold.c (gimple_fold_call): Fold
4368 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
4369 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
4370 INTEGER_CSTs, try to fold at least x * 0 and y - y.
4372 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
4374 PR rtl-optimization/60452
4375 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
4376 <case REG>: Return 1 for invalid offsets from the frame pointer.
4378 2014-03-26 Marek Polacek <polacek@redhat.com>
4381 * doc/extend.texi (C Extensions): Mention variable-length arrays in
4384 2014-03-26 Marek Polacek <polacek@redhat.com>
4387 * doc/extend.texi (Designated Inits): Describe what happens to omitted
4390 2014-03-26 Marek Polacek <polacek@redhat.com>
4393 * ira-color.c (update_conflict_hard_regno_costs): Perform the
4394 multiplication in unsigned type.
4396 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4398 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
4400 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4402 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
4404 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4407 * cif-code.def (UNREACHABLE) New code.
4408 * ipa-inline.c (inline_small_functions): Skip edges to
4409 __builtlin_unreachable.
4410 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
4411 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
4412 predicate to __bulitin_unreachable.
4413 (set_cond_stmt_execution_predicate): Fix issue when
4414 invert_tree_comparison returns ERROR_MARK.
4415 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
4416 propagate to inline clones.
4417 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
4419 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
4420 * cgraphclones.c (cgraph_clone_node): If call destination is already
4421 ureachable, do not redirect it back.
4422 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
4425 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4427 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
4428 Do not modify inline clones.
4430 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4432 * config/i386/i386.md (general_sext_operand): New mode attr.
4433 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
4434 don't generate (sign_extend (const_int)).
4435 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
4436 operands[2]. Use We constraint instead of <i> and
4437 <general_sext_operand> predicate instead of <general_operand>.
4438 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
4439 * config/i386/constraints.md (We): New constraint.
4440 * config/i386/predicates.md (x86_64_sext_operand,
4441 sext_operand): New predicates.
4443 2014-03-25 Martin Jambor <mjambor@suse.cz>
4446 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
4447 inconsistent devirtualizations to __builtin_unreachable.
4449 2014-03-25 Marek Polacek <polacek@redhat.com>
4452 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
4454 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
4456 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
4457 order of elements for big-endian.
4459 2014-03-25 Richard Biener <rguenther@suse.de>
4462 * gimplify-me.c (gimple_regimplify_operands): Update the
4465 2014-03-25 Martin Jambor <mjambor@suse.cz>
4468 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
4469 (lto_output_varpool_node): Likewise.
4470 (input_overwrite_node): Likewise.
4471 (input_varpool_node): Likewise.
4473 2014-03-25 Richard Biener <rguenther@suse.de>
4475 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
4476 (run_gcc): Likewise.
4478 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4480 * combine.c (simplify_compare_const): Add MODE argument.
4481 Handle mode_width 0 as very large mode_width.
4482 (try_combine, simplify_comparison): Adjust callers.
4484 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
4485 type to avoid signed integer overflow.
4486 * explow.c (plus_constant): Likewise.
4488 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
4490 * doc/generic.texi: Correct typos.
4492 2014-03-24 Tobias Burnus <burnus@net-b.de>
4494 * doc/invoke.texi (-flto): Expand section about
4495 using static libraries with LTO.
4497 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4499 PR rtl-optimization/60501
4500 * optabs.def (addptr3_optab): New optab.
4501 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
4502 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
4503 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
4505 * lra.c (emit_add3_insn): Use the addptr pattern if available.
4507 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
4509 2014-03-24 Ulrich Drepper <drepper@gmail.com>
4511 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
4514 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
4515 (_mm256_undefined_ps): Define.
4516 (_mm256_undefined_pd): Define.
4517 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
4518 (_mm_undefined_pd): Define.
4519 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
4520 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
4521 (_mm512_undefined_ps): Define.
4522 (_mm512_undefined_pd): Define.
4523 Use _mm*_undefined_*.
4524 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
4526 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
4528 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
4529 (lshr_simd): DI mode added.
4530 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
4531 (aarch64_ushr_simddi): Likewise.
4532 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
4533 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
4534 (vshrd_n_u64): Likewise.
4536 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4538 * Makefile.in (s-macro_list): Depend on cc1.
4540 2014-03-23 Teresa Johnson <tejohnson@google.com>
4542 * ipa-utils.c (ipa_print_order): Use specified dump file.
4544 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
4546 PR rtl-optimization/60601
4547 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
4549 * gcc.c (eval_spec_function): Initialize save_growing_value.
4551 2014-03-22 Jakub Jelinek <jakub@redhat.com>
4554 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
4555 code == MINUS_EXPR, never swap op0 with op1.
4557 * toplev.c (init_local_tick): Avoid signed integer multiplication
4559 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
4560 shift by first operand's bitsize.
4562 2014-03-21 Jakub Jelinek <jakub@redhat.com>
4565 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
4567 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
4568 TARGET_ISA_64BIT_P(x).
4570 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4572 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
4573 pattern for vector nor instead of subtract from splat(-1).
4574 (altivec_expand_vec_perm_const_le): Likewise.
4576 2014-03-21 Richard Henderson <rth@twiddle.net>
4579 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
4580 related insns after epilogue_completed.
4582 2014-03-21 Martin Jambor <mjambor@suse.cz>
4585 * cgraph.h (symtab_node): New flag body_removed.
4586 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
4587 when removing bodies.
4588 * symtab.c (dump_symtab_base): Dump body_removed flag.
4589 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
4590 had their bodies removed.
4592 2014-03-21 Martin Jambor <mjambor@suse.cz>
4595 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
4598 2014-03-21 Richard Biener <rguenther@suse.de>
4600 PR tree-optimization/60577
4601 * tree-core.h (struct tree_base): Document nothrow_flag use
4602 in VAR_DECL_NONALIASED.
4603 * tree.h (VAR_DECL_NONALIASED): New.
4604 (may_be_aliased): Adjust.
4605 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
4607 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4609 * expr.c (expand_expr_real_1): Remove outdated comment.
4611 2014-03-20 Jakub Jelinek <jakub@redhat.com>
4614 * ira.c (adjust_cleared_regs): Call copy_rtx on
4615 *reg_equiv[REGNO (loc)].src_p before passing it to
4616 simplify_replace_fn_rtx.
4619 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
4620 into CONST, put pic register as first operand of PLUS. Use
4621 gen_const_mem for both 32-bit and 64-bit PIC got loads.
4623 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4625 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
4627 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4629 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
4630 around for store forwarding issue in the FPU on the UT699.
4631 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
4632 loads and operations if -mfix-ut699 is specified.
4633 (divtf3_hq): Tweak attribute.
4634 (sqrttf2_hq): Likewise.
4636 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4638 * calls.c (store_one_arg): Remove incorrect const qualification on the
4639 type of the temporary.
4640 * cfgexpand.c (expand_return): Likewise.
4641 * expr.c (expand_constructor): Likewise.
4642 (expand_expr_real_1): Likewise.
4644 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4646 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
4649 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
4652 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
4654 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
4656 * config/arm/aarch-common-protos.h
4657 (alu_cost_table): Fix spelling of "extend".
4658 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
4660 2014-03-19 Richard Biener <rguenther@suse.de>
4663 * tree-core.h (tree_type_common): Re-order pointer members
4664 to reduce recursion depth during GC walks.
4666 2014-03-19 Marek Polacek <polacek@redhat.com>
4669 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
4670 before accessing it.
4672 2014-03-19 Richard Biener <rguenther@suse.de>
4675 * lto-streamer-in.c (input_function): In WPA stage do not drop
4678 2014-03-19 Jakub Jelinek <jakub@redhat.com>
4680 PR tree-optimization/60559
4681 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
4682 with build_zero_cst assignment.
4684 2014-03-18 Kai Tietz <ktietz@redhat.com>
4686 PR rtl-optimization/56356
4687 * sdbout.c (sdbout_parms): Verify that parms'
4688 incoming argument is valid.
4689 (sdbout_reg_parms): Likewise.
4691 2014-03-18 Richard Henderson <rth@redhat.com>
4694 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
4695 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
4696 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
4698 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
4700 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
4701 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
4702 Italicize plugin event names in description. Explain that
4703 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
4704 Remind that no GCC functions should be called after PLUGIN_FINISH.
4705 Explain what pragmas with expansion are.
4707 2014-03-18 Martin Liska <mliska@suse.cz>
4709 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
4710 gimple call statement is update.
4711 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
4712 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
4714 2014-03-18 Jakub Jelinek <jakub@redhat.com>
4717 * ubsan.c (ubsan_instrument_unreachable): Call
4718 initialize_sanitizer_builtins.
4719 (ubsan_pass): Likewise.
4722 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
4723 varpool_finalize_decl instead of rest_of_decl_compilation.
4725 2014-03-18 Richard Biener <rguenther@suse.de>
4727 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
4728 by using bitmap_and_compl instead of bitmap_and_compl_into.
4729 (df_rd_transfer_function): Likewise.
4731 2014-03-18 Richard Biener <rguenther@suse.de>
4733 * doc/lto.texi (fresolution): Fix typo.
4735 2014-03-18 Richard Biener <rguenther@suse.de>
4737 * doc/invoke.texi (flto): Update for changes in 4.9.
4739 2014-03-18 Richard Biener <rguenther@suse.de>
4741 * doc/loop.texi: Remove section on the removed lambda framework.
4742 Update loop docs with recent changes in preserving loop structure.
4744 2014-03-18 Richard Biener <rguenther@suse.de>
4746 * doc/lto.texi (-fresolution): Document.
4748 2014-03-18 Richard Biener <rguenther@suse.de>
4750 * doc/contrib.texi: Adjust my name.
4752 2014-03-18 Jakub Jelinek <jakub@redhat.com>
4755 * internal-fn.c: Include diagnostic-core.h.
4756 (expand_BUILTIN_EXPECT): New function.
4757 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
4758 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
4759 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
4760 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
4762 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
4763 Revert 3 argument __builtin_expect code.
4764 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
4765 * gimple-fold.c (gimple_fold_call): Likewise.
4766 * tree.h (fold_builtin_expect): New prototype.
4767 * builtins.c (build_builtin_expect_predicate): Add predictor
4768 argument, if non-NULL, create 3 argument __builtin_expect.
4769 (fold_builtin_expect): No longer static. Add ARG2 argument,
4770 pass it through to build_builtin_expect_predicate.
4771 (fold_builtin_2): Adjust caller.
4772 (fold_builtin_3): Handle BUILT_IN_EXPECT.
4773 * internal-fn.def (BUILTIN_EXPECT): New.
4775 2014-03-18 Tobias Burnus <burnus@net-b.de>
4778 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
4779 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
4780 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
4782 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
4785 * predict.c (combine_predictions_for_bb): Fix up formatting.
4786 (expr_expected_value_1, expr_expected_value): Add predictor argument,
4787 fill what it points to if non-NULL.
4788 (tree_predict_by_opcode): Adjust caller, use the predictor.
4789 * predict.def (PRED_COMPARE_AND_SWAP): Add.
4791 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
4793 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
4794 proper constant for the store mode.
4796 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
4798 * symtab.c (change_decl_assembler_name): Fix transparent alias
4801 2014-03-16 Renlin Li <Renlin.Li@arm.com>
4803 * config/aarch64/aarch64.c: Correct the comments about the
4804 aarch64 stack layout.
4806 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
4808 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
4809 check for GF_OMP_FOR_KIND_FOR.
4811 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
4813 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
4814 ymm and zmm register names.
4816 2014-03-17 Jakub Jelinek <jakub@redhat.com>
4819 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
4820 note creation for the 2010-08-31 changes.
4822 2014-03-17 Marek Polacek <polacek@redhat.com>
4825 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
4826 as -fno-tree-loop-vectorize.
4827 (expand_omp_simd): Likewise.
4829 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
4831 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
4832 (eligible_for_call_delay): New prototype.
4833 * config/sparc/sparc.c (tls_call_delay): Rename into...
4834 (eligible_for_call_delay): ...this. Return false if the instruction
4835 cannot be put in the delay slot of a branch.
4836 (eligible_for_restore_insn): Simplify.
4837 (eligible_for_return_delay): Return false if the instruction cannot be
4838 put in the delay slot of a branch and simplify.
4839 (eligible_for_sibcall_delay): Return false if the instruction cannot be
4840 put in the delay slot of a branch.
4841 * config/sparc/sparc.md (fix_ut699): New attribute.
4842 (tls_call_delay): Delete.
4843 (in_call_delay): Reimplement.
4844 (eligible_for_sibcall_delay): Rename into...
4845 (in_sibcall_delay): ...this.
4846 (eligible_for_return_delay): Rename into...
4847 (in_return_delay): ...this.
4848 (in_branch_delay): Reimplement.
4849 (in_uncond_branch_delay): Delete.
4850 (in_annul_branch_delay): Delete.
4852 2014-03-14 Richard Henderson <rth@redhat.com>
4855 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
4856 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
4857 (*floathi<X87MODEF>2_i387_with_temp): Remove.
4858 (floathi splitters): Remove.
4859 (float<SWI48x>xf2): New pattern.
4860 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
4861 code that tried to handle DImode for 32-bit, but which was excluded
4862 by the pattern's condition. Drop allocation of stack temporary.
4863 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
4864 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
4865 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
4866 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
4867 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
4868 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
4869 (*float<SWI48><MODEF>2_sse_interunit): Remove.
4870 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
4871 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
4872 (*float<SWI48x><X87MODEF>2_i387): Remove.
4873 (all float _with_temp splitters): Remove.
4874 (*float<SWI48x><MODEF>2_i387): New pattern.
4875 (*float<SWI48><MODEF>2_sse): New pattern.
4876 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
4877 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
4879 2014-03-14 Jakub Jelinek <jakub@redhat.com>
4880 Marek Polacek <polacek@redhat.com>
4883 * common.opt (dump_base_name_prefixed): New Variable.
4884 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
4885 if x_dump_base_name_prefixed is already set, set it at the end.
4887 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
4889 PR rtl-optimization/60508
4890 * lra-constraints.c (get_reload_reg): Add new parameter
4892 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
4893 Pass the new parameter values.
4895 2014-03-14 Richard Biener <rguenther@suse.de>
4897 * common.opt: Revert unintented changes from r205065.
4900 2014-03-14 Richard Biener <rguenther@suse.de>
4903 * cfghooks.c (split_block): Properly adjust all loops the
4904 block was a latch of.
4906 2014-03-14 Martin Jambor <mjambor@suse.cz>
4909 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
4912 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
4915 * config/avr/avr.c (avr_set_current_function): Pass function name
4916 through default_strip_name_encoding before sanity checking instead
4917 of skipping the first char of the assembler name.
4919 2014-03-13 Richard Henderson <rth@redhat.com>
4922 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
4923 (ix86_force_to_memory, ix86_free_from_memory): Remove.
4924 * config/i386/i386-protos.h: Likewise.
4925 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
4926 in the expander instead of a splitter.
4927 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
4928 any possibility of requiring a memory.
4929 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
4930 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
4931 (fp branch splitters): Update for ix86_split_fp_branch.
4932 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
4933 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
4934 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
4935 (*fop_<MODEF>_2_i387): Remove f/r alternative.
4936 (*fop_<MODEF>_3_i387): Likewise.
4937 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
4938 (splitters for the fop_* register patterns): Remove.
4939 (fscalexf4_i387): Rename from *fscalexf4_i387.
4940 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
4942 2014-03-13 Jakub Jelinek <jakub@redhat.com>
4944 PR tree-optimization/59779
4945 * tree-dfa.c (get_ref_base_and_extent): Use double_int
4946 type for bitsize and maxsize instead of HOST_WIDE_INT.
4948 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
4950 PR rtl-optimization/57320
4951 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
4952 the CFG after thread_prologue_and_epilogue_insns.
4954 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
4956 PR rtl-optimization/57189
4957 * lra-constraints.c (process_alt_operands): Disfavor spilling
4960 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
4962 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
4964 2014-03-13 Jakub Jelinek <jakub@redhat.com>
4966 PR tree-optimization/59025
4968 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
4969 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
4971 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
4974 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
4975 calls of avr_out_plus_1.
4977 2014-03-13 Bin Cheng <bin.cheng@arm.com>
4979 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
4980 BB's single pred and update the father loop's latch info later.
4982 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
4984 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
4988 (VEC_base): Likewise.
4989 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
4990 registers, we need to swap double words in little endian mode.
4992 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
4993 to be a container mode for 128-bit integer operations added in ISA
4994 2.07. Unlike TImode and PTImode, the preferred register set is
4995 the Altivec/VMX registers for the 128-bit operations.
4997 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
4999 (rs6000_split_128bit_ok_p): Likewise.
5001 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
5002 macros for creating ISA 2.07 normal and overloaded builtin
5003 functions with 3 arguments.
5004 (BU_P8V_OVERLOAD_3): Likewise.
5005 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
5006 for use as overloaded functions.
5007 (VPERM_1TI_UNS): Likewise.
5008 (VSEL_1TI): Likewise.
5009 (VSEL_1TI_UNS): Likewise.
5010 (ST_INTERNAL_1ti): Likewise.
5011 (LD_INTERNAL_1ti): Likewise.
5012 (XXSEL_1TI): Likewise.
5013 (XXSEL_1TI_UNS): Likewise.
5014 (VPERM_1TI): Likewise.
5015 (VPERM_1TI_UNS): Likewise.
5016 (XXPERMDI_1TI): Likewise.
5017 (SET_1TI): Likewise.
5018 (LXVD2X_V1TI): Likewise.
5019 (STXVD2X_V1TI): Likewise.
5020 (VEC_INIT_V1TI): Likewise.
5021 (VEC_SET_V1TI): Likewise.
5022 (VEC_EXT_V1TI): Likewise.
5023 (EQV_V1TI): Likewise.
5024 (NAND_V1TI): Likewise.
5025 (ORC_V1TI): Likewise.
5026 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
5027 added in ISA 2.07. Add both normal 'altivec' builtins, and the
5029 (VADDUQM): Likewise.
5030 (VSUBCUQ): Likewise.
5031 (VADDEUQM): Likewise.
5032 (VADDECUQ): Likewise.
5033 (VSUBEUQM): Likewise.
5034 (VSUBECUQ): Likewise.
5036 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5037 __int128_t and __uint128_t types.
5038 (__uint128_type): Likewise.
5039 (altivec_categorize_keyword): Add support for vector __int128_t,
5040 vector __uint128_t, vector __int128, and vector unsigned __int128
5041 as a container type for TImode operations that need to be done in
5042 VSX/Altivec registers.
5043 (rs6000_macro_to_expand): Likewise.
5044 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5045 to support 128-bit integer instructions vaddcuq, vadduqm,
5046 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5047 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5049 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5050 for V1TImode, and set up preferences to use VSX/Altivec registers.
5051 Setup VSX reload handlers.
5052 (rs6000_debug_reg_global): Likewise.
5053 (rs6000_init_hard_regno_mode_ok): Likewise.
5054 (rs6000_preferred_simd_mode): Likewise.
5055 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5056 (easy_altivec_constant): Likewise.
5057 (output_vec_const_move): Likewise.
5058 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5060 (rs6000_expand_vector_extract): Likewise.
5061 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
5063 (rs6000_const_vec): Add support for V1TImode.
5064 (rs6000_emit_le_vsx_load): Swap double words when loading or
5065 storing TImode/V1TImode.
5066 (rs6000_emit_le_vsx_store): Likewise.
5067 (rs6000_emit_le_vsx_move): Likewise.
5068 (rs6000_emit_move): Add support for V1TImode.
5069 (altivec_expand_ld_builtin): Likewise.
5070 (altivec_expand_st_builtin): Likewise.
5071 (altivec_expand_vec_init_builtin): Likewise.
5072 (altivec_expand_builtin): Likewise.
5073 (rs6000_init_builtins): Add support for V1TImode type. Add
5074 support for ISA 2.07 128-bit integer builtins. Define type names
5075 for the VSX/Altivec vector types.
5076 (altivec_init_builtins): Add support for overloaded vector
5077 functions with V1TImode type.
5078 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
5079 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
5081 (rs6000_split_128bit_ok_p): Likewise.
5082 (rs6000_handle_altivec_attribute): Create V1TImode from vector
5083 __int128_t and vector __uint128_t.
5085 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
5086 and mode attributes.
5093 (VS_scalar): Likewise.
5094 (VS_double): Likewise.
5095 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
5097 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
5098 we support the ISA 2.07 128-bit integer arithmetic instructions.
5099 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
5100 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
5101 and TImode types for use with the builtin functions.
5102 (V1TI_type_node): Likewise.
5103 (unsigned_V1TI_type_node): Likewise.
5104 (intTI_type_internal_node): Likewise.
5105 (uintTI_type_internal_node): Likewise.
5107 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
5108 128-bit builtin functions.
5109 (UNSPEC_VADDEUQM): Likewise.
5110 (UNSPEC_VADDECUQ): Likewise.
5111 (UNSPEC_VSUBCUQ): Likewise.
5112 (UNSPEC_VSUBEUQM): Likewise.
5113 (UNSPEC_VSUBECUQ): Likewise.
5114 (VM): Add V1TImode to vector mode iterators.
5116 (VI_unit): Likewise.
5117 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
5118 (altivec_vaddcuq): Likewise.
5119 (altivec_vsubuqm): Likewise.
5120 (altivec_vsubcuq): Likewise.
5121 (altivec_vaddeuqm): Likewise.
5122 (altivec_vaddecuq): Likewise.
5123 (altivec_vsubeuqm): Likewise.
5124 (altivec_vsubecuq): Likewise.
5126 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
5128 (BOOL_128): Likewise.
5129 (BOOL_REGS_OUTPUT): Likewise.
5130 (BOOL_REGS_OP1): Likewise.
5131 (BOOL_REGS_OP2): Likewise.
5132 (BOOL_REGS_UNARY): Likewise.
5133 (BOOL_REGS_AND_CR0): Likewise.
5135 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
5136 128-bit integer builtin support.
5137 (vec_vadduqm): Likewise.
5138 (vec_vaddecuq): Likewise.
5139 (vec_vaddeuqm): Likewise.
5140 (vec_vsubecuq): Likewise.
5141 (vec_vsubeuqm): Likewise.
5142 (vec_vsubcuq): Likewise.
5143 (vec_vsubuqm): Likewise.
5145 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5146 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
5147 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
5148 128-bit integer add/subtract to ISA 2.07.
5150 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
5152 * config/arc/arc.c (arc_predicate_delay_insns):
5153 Fix third argument passed to conditionalize_nonjump.
5155 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
5157 * config/aarch64/aarch64-builtins.c
5158 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
5159 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
5160 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
5161 instead of __builtin_lfloor.
5162 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
5164 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5166 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
5167 (tree_ssa_ifcombine_bb_1): New function.
5168 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
5169 is an empty forwarder block to then_bb or vice versa and then_bb
5170 and else_bb are effectively swapped.
5172 2014-03-12 Christian Bruel <christian.bruel@st.com>
5175 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
5176 REG_CFA_DEF_CFA note.
5177 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
5178 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
5180 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
5182 PR tree-optimization/60454
5183 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
5185 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5187 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
5188 Do not define target_cpu_default2 to generic.
5189 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
5190 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
5191 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
5193 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5194 Marc Glisse <marc.glisse@inria.fr>
5196 PR tree-optimization/60502
5197 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
5198 instead of build_low_bits_mask.
5200 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5203 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
5204 if there are multiple uses, but op doesn't live on E edge.
5205 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
5206 clobber stmts before __builtin_unreachable.
5208 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
5210 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
5211 hard_frame_pointer_rtx.
5212 * cse.c (cse_insn): Remove volatile check.
5213 * cselib.c (cselib_process_insn): Likewise.
5214 * dse.c (scan_insn): Likewise.
5216 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5218 * config/arc/arc.c (conditionalize_nonjump): New function,
5220 (arc_ifcvt): ... this.
5221 (arc_predicate_delay_insns): Use it.
5223 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5225 * config/arc/predicates.md (extend_operand): During/after reload,
5226 allow const_int_operand.
5227 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
5228 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
5229 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
5231 (umulsi3_highpart_i): Likewise.
5233 2014-03-11 Richard Biener <rguenther@suse.de>
5235 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
5236 Add asserts to guard possible wrong-code bugs.
5238 2014-03-11 Richard Biener <rguenther@suse.de>
5240 PR tree-optimization/60429
5241 PR tree-optimization/60485
5242 * tree-ssa-structalias.c (set_union_with_increment): Properly
5243 take into account all fields that overlap the shifted vars.
5244 (do_sd_constraint): Likewise.
5245 (do_ds_constraint): Likewise.
5246 (get_constraint_for_ptr_offset): Likewise.
5248 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
5250 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
5251 (nios2_compute_frame_layout):
5252 Add calculation of cfun->machine->fp_save_offset.
5253 (nios2_expand_prologue): Correct setting of frame pointer register
5255 (nios2_expand_epilogue): Update recovery of stack pointer from
5256 frame pointer accordingly.
5257 (nios2_initial_elimination_offset): Update calculation of offset
5258 for eliminating to HARD_FRAME_POINTER_REGNUM.
5260 2014-03-10 Jakub Jelinek <jakub@redhat.com>
5263 * ipa.c (symtab_remove_unreachable_nodes): Don't call
5264 cgraph_get_create_node on VAR_DECLs.
5266 2014-03-10 Richard Biener <rguenther@suse.de>
5269 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
5271 2014-03-08 Douglas B Rupp <rupp@gnat.com>
5273 * config/vms/vms.opt (vms_float_format): New variable.
5275 2014-03-08 Tobias Burnus <burnus@net-b.de>
5277 * doc/invoke.texi (-fcilkplus): Update implementation status.
5279 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
5280 Richard Biener <rguenther@suse.de>
5282 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
5283 consistently accross all TUs.
5284 (run_gcc): Enable -fshort-double automatically at link at link-time
5285 and disallow override.
5287 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
5290 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
5291 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
5292 if they can't be used.
5294 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5296 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
5297 for Solaris 11/x86 ld.
5298 * configure: Regenerate.
5300 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5302 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
5303 (LIB_TLS_SPEC): Save as ld_tls_libs.
5304 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
5305 (HAVE_AS_IX86_TLSLDM): New test.
5306 * configure, config.in: Regenerate.
5307 * config/i386/i386.c (legitimize_tls_address): Fall back to
5308 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
5309 cannot support TLS_MODEL_LOCAL_DYNAMIC.
5310 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
5311 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
5313 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
5315 * common.opt (fira-loop-pressure): Mark as optimization.
5317 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
5319 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
5320 an OpenMP mappable type.
5322 2014-03-06 Matthias Klose <doko@ubuntu.com>
5324 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
5325 MULTILIB_OSDIRNAMES is not defined.
5327 2014-03-06 Jakub Jelinek <jakub@redhat.com>
5328 Meador Inge <meadori@codesourcery.com>
5331 * config/arm/arm.c (arm_tls_symbol_p): Remove.
5332 (arm_legitimize_address): Call legitimize_tls_address for any
5333 arm_tls_referenced_p expression, handle constant addend. Call it
5334 before testing for !TARGET_ARM.
5335 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
5337 2014-03-06 Richard Biener <rguenther@suse.de>
5343 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5345 * tree-streamer.c (record_common_node): Assert we don't record
5346 nodes with type double.
5347 (preload_common_node): Skip type double, complex double and double
5348 pointer since it is now frontend dependent due to fshort-double option.
5350 2014-03-06 Richard Biener <rguenther@suse.de>
5352 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
5353 or -fno-lto is specified and the linker has full plugin support.
5354 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
5355 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
5356 * lto-wrapper.c (merge_and_complain): Merge compile-time
5357 optimization levels.
5358 (run_gcc): And pass it through to the link options.
5360 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
5364 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5366 * cselib.c (remove_useless_values): Skip to avoid quadratic
5367 behavior if the condition moved from...
5368 (cselib_process_insn): ... here holds.
5370 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5373 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
5374 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
5377 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
5378 (TM_H): Add x86-tune.def.
5380 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5382 * config/aarch64/aarch64.c (generic_tunings):
5383 Use cortexa57_extra_costs.
5385 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5388 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
5389 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
5390 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
5393 2014-03-04 Heiher <r@hev.cc>
5395 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
5396 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
5398 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
5400 * config/i386/predicates.md (const2356_operand): Change to ...
5401 (const2367_operand): ... this.
5402 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
5404 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5405 (*avx512pf_scatterpf<mode>sf): Ditto.
5406 (avx512pf_scatterpf<mode>df): Ditto.
5407 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5408 (*avx512pf_scatterpf<mode>df): Ditto.
5409 * config/i386/i386.c (ix86_expand_builtin): Update
5410 incorrect hint operand error message.
5412 2014-03-04 Richard Biener <rguenther@suse.de>
5414 * lto-section-in.c (lto_get_section_data): Fix const cast.
5416 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5418 * tree-streamer.c (record_common_node): Assert we don't record
5419 nodes with type double.
5420 (preload_common_node): Skip type double, complex double and double
5421 pointer since it is now frontend dependent due to fshort-double option.
5423 2014-03-04 Richard Biener <rguenther@suse.de>
5426 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
5427 (lto_input_toplevel_asms): Likewise.
5428 * lto-section-in.c (lto_get_section_data): Instead do it here
5431 2014-03-04 Richard Biener <rguenther@suse.de>
5433 PR tree-optimization/60382
5434 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
5435 dead PHIs a reduction.
5437 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
5439 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
5441 (_mm_prefetch): Move out of GCC target("sse") pragma.
5442 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
5443 GCC target("prfchw") pragma.
5444 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
5446 * config/i386/i386.c (ix86_option_override_internal): Enable
5447 -mprfchw with -mprefetchwt1.
5449 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5451 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
5454 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5456 * opts.h (CL_PCH_IGNORE): Define.
5457 * targhooks.c (option_affects_pch_p):
5458 Return false for options that have CL_PCH_IGNORE set.
5459 * opt-functions.awk: Process PchIgnore.
5460 * doc/options.texi: Document PchIgnore.
5462 * config/arc/arc.opt (misize): Add PchIgnore property.
5464 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5466 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
5467 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
5468 constraint on constants to permit them being loaded into
5469 GENERAL_REGS or BASE_REGS.
5471 2014-03-03 Nick Clifton <nickc@redhat.com>
5473 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
5474 anti-cacnonical alternatives.
5475 (negandhi3_real): New pattern.
5476 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
5478 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5480 * config/avr/avr-mcus.def: Remove atxmega16x1.
5481 * config/avr/avr-tables.opt: Regenerate.
5482 * config/avr/t-multilib: Regenerate.
5483 * doc/avr-mmcu.texi: Regenerate.
5485 2014-03-03 Tobias Grosser <tobias@grosser.es>
5486 Mircea Namolaru <mircea.namolaru@inria.fr>
5488 PR tree-optimization/58028
5489 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
5492 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5494 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
5495 not handled by recognizers.
5497 2014-03-03 Jakub Jelinek <jakub@redhat.com>
5500 * function.c (expand_function_end): Don't emit
5501 clobber_return_register sequence if clobber_after is a BARRIER.
5502 * cfgexpand.c (construct_exit_block): Append instructions before
5503 return_label to prev_bb.
5505 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5507 * config/rs6000/constraints.md: Document reserved use of "wc".
5509 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5512 * ipa.c (function_and_variable_visibility): When dissolving comdat
5513 group, also set all symbols to local.
5515 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5520 2013-12-14 Jan Hubicka <jh@suse.cz>
5522 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
5524 2014-03-02 Jon Beniston <jon@beniston.com>
5530 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
5531 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
5532 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
5533 (simple_return, *simple_return): New patterns
5534 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
5535 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
5537 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
5539 * dwarf2out.c (gen_subprogram_die): Tidy.
5541 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
5544 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
5545 (*mov_t_msb_neg_negc): ... this new insn.
5547 2014-02-28 Jason Merrill <jason@redhat.com>
5550 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
5553 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
5556 * dwarf2out.c (decltype_auto_die): New static.
5557 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
5558 (gen_type_die_with_usage): Handle 'decltype(auto)'.
5559 (is_cxx_auto): Likewise.
5561 2014-02-28 Ian Bolton <ian.bolton@arm.com>
5563 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
5564 we are not using general regs only.
5566 2014-02-28 Richard Biener <rguenther@suse.de>
5569 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
5570 previous fix and only allow to remove trivial pre-headers
5571 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
5572 (remove_forwarder_block): Properly update the latch of a loop.
5574 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5577 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
5578 (cselib_preserved_hash_table): New.
5579 (preserve_constants_and_equivs): Move preserved vals to it.
5580 (cselib_find_slot): Look it up first.
5581 (cselib_init): Initialize it.
5582 (cselib_finish): Release it.
5583 (dump_cselib_table): Dump it.
5585 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5588 * cselib.c (remove_useless_values): Skip to avoid quadratic
5589 behavior if the condition moved from...
5590 (cselib_process_insn): ... here holds.
5592 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5595 * var-tracking.c (vt_initialize): Apply the same condition to
5596 preserve the CFA base value.
5598 2014-02-28 Joey Ye <joey.ye@arm.com>
5601 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
5602 if reload in progress or completed.
5604 2014-02-28 Tobias Burnus <burnus@net-b.de>
5607 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
5610 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
5612 * doc/tm.texi.in (Condition Code Status): Update documention for
5613 relative locations of cc0-setter and cc0-user.
5615 2014-02-27 Jeff Law <law@redhat.com>
5617 PR rtl-optimization/52714
5618 * combine.c (try_combine): When splitting an unrecognized PARALLEL
5619 into two independent simple sets, if I3 is a jump, ensure the
5620 pattern we place into I3 is a (set (pc) ...).
5622 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
5623 Jeff Law <law@redhat.com>
5625 PR rtl-optimization/49847
5626 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
5627 are in different blocks.
5628 * doc/tm.texi (Condition Code Status): Update documention for
5629 relative locations of cc0-setter and cc0-user.
5631 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
5634 * lra.c (lra_emit_add): Check SUBREG too.
5636 2014-02-27 Andreas Schwab <schwab@suse.de>
5638 * config/m68k/m68k.c (m68k_option_override): Disable
5639 -flive-range-shrinkage for classic m68k.
5640 (m68k_override_options_after_change): Likewise.
5642 2014-02-27 Marek Polacek <polacek@redhat.com>
5645 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
5646 -Wmaybe-uninitialized.
5648 2014-02-27 Alan Modra <amodra@gmail.com>
5651 * reload1.c (emit_input_reload_insns): When reload_override_in,
5652 set old to rl->in_reg when rl->in_reg is a subreg.
5654 2014-02-26 Richard Biener <rguenther@suse.de>
5657 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
5659 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
5661 * common/config/i386/predicates.md (const1256_operand): Remove.
5662 (const2356_operand): New.
5663 (const_1_to_2_operand): Remove.
5664 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5665 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5666 (*avx512pf_gatherpf<mode>sf): Ditto.
5667 (avx512pf_gatherpf<mode>df): Ditto.
5668 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5669 (*avx512pf_gatherpf<mode>df): Ditto.
5670 (avx512pf_scatterpf<mode>sf): Ditto.
5671 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5672 (*avx512pf_scatterpf<mode>sf): Ditto.
5673 (avx512pf_scatterpf<mode>df): Ditto.
5674 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5675 (*avx512pf_scatterpf<mode>df): Ditto.
5676 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
5678 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
5680 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
5681 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
5682 (_mm512_mask_testn_epi64_mask): Move to ...
5683 * config/i386/avx512cdintrin.h: Here.
5684 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
5685 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
5686 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
5687 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
5688 TARGET_AVX512F from TARGET_AVX512CD.
5690 2014-02-26 Richard Biener <rguenther@suse.de>
5693 * ipa.c (walk_polymorphic_call_targets): Properly guard
5694 call to inline_update_overall_summary.
5696 2014-02-26 Bin Cheng <bin.cheng@arm.com>
5699 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
5700 and latches only if requested. Fix latch if it is removed.
5701 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
5702 LOOPS_HAVE_PREHEADERS.
5704 2014-02-25 Andrew Pinski <apinski@cavium.com>
5706 * builtins.c (expand_builtin_thread_pointer): Create a new target
5707 when the target is NULL.
5709 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
5711 PR rtl-optimization/60317
5712 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
5713 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
5714 * lra-assigns.c: Include params.h.
5715 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
5716 other reload pseudos considerations.
5718 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5720 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
5721 to use canonical form for nor<mode>3.
5723 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5726 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
5729 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
5731 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
5732 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
5733 (ix86_handle_option): Handle OPT_mprefetchwt1.
5734 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
5735 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
5737 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
5738 OPTION_MASK_ISA_PREFETCHWT1.
5739 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
5740 (PTA_PREFETCHWT1): New.
5741 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
5742 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
5743 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
5744 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
5745 (*prefetch_avx512pf_<mode>_: Change into ...
5746 (*prefetch_prefetchwt1_<mode>: This.
5747 * config/i386/i386.opt (mprefetchwt1): New.
5748 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
5749 (_mm_prefetch): Handle intent to write.
5750 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
5752 2014-02-25 Richard Biener <rguenther@suse.de>
5755 * emit-rtl.c (mem_attrs_htab): Remove.
5756 (mem_attrs_htab_hash): Likewise.
5757 (mem_attrs_htab_eq): Likewise.
5758 (set_mem_attrs): Always allocate new mem-attrs when something changed.
5759 (init_emit_once): Do not allocate mem_attrs_htab.
5761 2014-02-25 Richard Biener <rguenther@suse.de>
5764 * lto-opts.c (lto_write_options): Output non-explicit conservative
5765 -fwrapv, -fno-trapv and -fno-strict-overflow.
5766 * lto-wrapper.c (merge_and_complain): Handle merging those options.
5767 (run_gcc): And pass them through.
5769 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
5771 * sel-sched.c (calculate_new_fences): New parameter ptime.
5772 Calculate it as a maximum over all fence cycles.
5773 (sel_sched_region_2): Adjust the call to calculate_new_fences.
5774 Print the final schedule timing when sched_verbose.
5776 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
5778 PR rtl-optimization/60292
5779 * sel-sched.c (fill_vec_av_set): Do not reset target availability
5780 bit fot the fence instruction.
5782 2014-02-24 Alangi Derick <alangiderick@gmail.com>
5784 * calls.h: Fix typo in comment.
5786 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
5788 * config/pa/pa.c (pa_output_move_double): Don't valididate when
5789 adjusting offsetable addresses.
5791 2014-02-24 Guozhi Wei <carrot@google.com>
5793 * sparseset.h (sparseset_pop): Fix the wrong index.
5795 2014-02-24 Walter Lee <walt@tilera.com>
5797 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
5798 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
5800 * common/config/tilegx/tilegx-common.c
5801 (TARGET_DEFAULT_TARGET_FLAGS): Define.
5802 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
5804 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
5805 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
5806 (tilegx_gimplify_va_arg_expr): Handle big endian.
5807 (tilegx_expand_unaligned_load): Ditto.
5808 (tilegx_expand_unaligned_store): Ditto.
5809 (TARGET_RETURN_IN_MSB): New.
5810 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
5811 (TARGET_ENDIAN_DEFAULT): New.
5812 (TARGET_BIG_ENDIAN): Handle big endian.
5813 (BYTES_BIG_ENDIAN): Ditto.
5814 (WORDS_BIG_ENDIAN): Ditto.
5815 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
5818 * config/tilegx/tilegx.md (extv): Handle big endian.
5820 (insn_st<n>): Ditto.
5821 (insn_st<n>_add<bitsuffix>): Ditto.
5822 (insn_stnt<n>): Ditto.
5823 (insn_stnt<n>_add<bitsuffix>):Ditto.
5824 (vec_interleave_highv8qi): Handle big endian.
5825 (vec_interleave_highv8qi_be): New.
5826 (vec_interleave_highv8qi_le): New.
5827 (insn_v1int_h): Handle big endian.
5828 (vec_interleave_lowv8qi): Handle big endian.
5829 (vec_interleave_lowv8qi_be): New.
5830 (vec_interleave_lowv8qi_le): New.
5831 (insn_v1int_l): Handle big endian.
5832 (vec_interleave_highv4hi): Handle big endian.
5833 (vec_interleave_highv4hi_be): New.
5834 (vec_interleave_highv4hi_le): New.
5835 (insn_v2int_h): Handle big endian.
5836 (vec_interleave_lowv4hi): Handle big endian.
5837 (vec_interleave_lowv4hi_be): New.
5838 (vec_interleave_lowv4hi_le): New.
5839 (insn_v2int_l): Handle big endian.
5840 (vec_interleave_highv2si): Handle big endian.
5841 (vec_interleave_highv2si_be): New.
5842 (vec_interleave_highv2si_le): New.
5843 (insn_v4int_h): Handle big endian.
5844 (vec_interleave_lowv2si): Handle big endian.
5845 (vec_interleave_lowv2si_be): New.
5846 (vec_interleave_lowv2si_le): New.
5847 (insn_v4int_l): Handle big endian.
5848 * config/tilegx/tilegx.opt (mbig-endian): New option.
5849 (mlittle-endian): New option.
5850 * doc/install.texi: Document tilegxbe-linux.
5851 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
5853 2014-02-24 Martin Jambor <mjambor@suse.cz>
5856 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
5857 there are no parameter descriptors.
5859 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
5861 PR rtl-optimization/60268
5862 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
5863 initialization to ...
5864 (sched_rgn_init): ... here.
5865 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
5867 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5869 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
5872 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
5874 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
5877 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5879 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
5880 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
5882 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5884 * config/microblaze/predicates.md: Add cmp_op predicate.
5885 * config/microblaze/microblaze.md: Add branch_compare instruction
5886 which uses cmp_op predicate and emits cmp insn before branch.
5887 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
5888 to microblaze_expand_conditional_branch and consolidate logic.
5889 (microblaze_expand_conditional_branch): emit branch_compare
5890 insn instead of handling cmp op separate from branch insn.
5892 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5894 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
5897 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5899 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
5900 define_insn with define_expand and new define_insn
5901 *altivec_lve<VI_char>x_internal.
5902 (altivec_stve<VI_char>x): Replace define_insn with define_expand
5903 and new define_insn *altivec_stve<VI_char>x_internal.
5904 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
5906 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
5908 (altivec_expand_stvex_be): New function.
5910 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
5912 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
5913 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
5914 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
5915 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
5917 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
5920 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
5921 instead of emit_move_insn.
5923 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5925 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
5927 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
5928 gen_altivec_vsumsws.
5930 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5932 * config/rs6000/altivec.md (altivec_lvxl): Rename as
5933 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
5934 (altivec_lvxl_<mode>): New define_expand incorporating
5935 -maltivec=be semantics where needed.
5936 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
5937 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
5938 semantics where needed.
5939 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
5940 (altivec_stvx_<mode>): New define_expand incorporating
5941 -maltivec=be semantics where needed.
5942 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
5943 VM2 iterator instead of V4SI.
5944 (altivec_stvxl_<mode>): New define_expand incorporating
5945 -maltivec=be semantics where needed.
5946 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
5947 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
5948 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
5949 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
5950 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
5951 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
5952 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
5953 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
5954 ALTIVEC_BUILTIN_STVXL.
5955 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
5956 (altivec_expand_stvx_be): Likewise.
5957 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
5958 (altivec_expand_lvx_be): Likewise.
5959 (altivec_expand_stvx_be): Likewise.
5960 (altivec_expand_builtin): Add cases for
5961 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
5962 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
5963 (altivec_init_builtins): Add definitions for
5964 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
5965 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
5967 2014-02-21 Catherine Moore <clm@codesourcery.com>
5969 * doc/invoke.texi (mvirt, mno-virt): Document.
5970 * config/mips/mips.opt (mvirt): New option.
5971 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
5973 2014-02-21 Richard Biener <rguenther@suse.de>
5975 PR tree-optimization/60276
5976 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
5977 (STMT_VINFO_MIN_NEG_DIST): New macro.
5978 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
5979 STMT_VINFO_MIN_NEG_DIST.
5980 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
5981 made for negative dependence distances still hold.
5983 2014-02-21 Richard Biener <rguenther@suse.de>
5986 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
5987 DECL_INITIAL for globals not in the current function context.
5989 2014-02-21 Jakub Jelinek <jakub@redhat.com>
5991 PR tree-optimization/56490
5992 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
5993 * tree-ssa-uninit.c: Include params.h.
5994 (compute_control_dep_chain): Add num_calls argument, return false
5995 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
5996 num_calls to recursive call.
5997 (find_predicates): Change dep_chain into normal array,
5998 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
5999 variable and adjust compute_control_dep_chain caller.
6000 (find_def_preds): Likewise.
6002 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
6004 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
6005 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
6007 2014-02-21 Nick Clifton <nickc@redhat.com>
6009 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
6010 (pushhi1): Likewise.
6011 (popqi1): Add mode to pre_dec.
6014 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6016 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
6017 mode for mask of V8SFmode permutation.
6019 2014-02-20 Richard Henderson <rth@redhat.com>
6022 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
6023 a new pseudo for OLDVAL.
6025 2014-02-20 Jakub Jelinek <jakub@redhat.com>
6028 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
6029 gen_reg_rtx if d->testing_p.
6030 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
6031 if d->testing_p and we will certainly return true.
6032 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6035 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6037 * emit-rtl.c (gen_reg_rtx): Assert that
6038 crtl->emit.regno_pointer_align_length is non-zero.
6040 2014-02-20 Richard Henderson <rth@redhat.com>
6043 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6044 on failure the store back into EXPECT.
6046 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6047 Sandra Loosemore <sandra@codesourcery.com>
6049 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6050 * config/nios2/nios2.c (nios2_function_profiler): Add
6051 -fPIC (flag_pic == 2) support.
6052 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6053 (nios2_large_offset_p): New function.
6054 (nios2_unspec_reloc_p): Move up position, update to use
6055 nios2_large_offset_p.
6056 (nios2_unspec_address): Remove function.
6057 (nios2_unspec_offset): New function.
6058 (nios2_large_got_address): New function.
6059 (nios2_got_address): Add large offset support.
6060 (nios2_legitimize_tls_address): Update usage of removed and new
6062 (nios2_symbol_binds_local_p): New function.
6063 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
6064 (nios2_legitimize_address): Update to use nios2_large_offset_p.
6065 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
6066 (nios2_print_operand): Merge H/L processing, add hiadj/lo
6067 processing for (const (unspec ...)).
6068 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
6070 2014-02-20 Richard Biener <rguenther@suse.de>
6072 * tree-cfg.c (replace_uses_by): Mark altered BBs before
6073 doing the substitution.
6074 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
6076 2014-02-20 Martin Jambor <mjambor@suse.cz>
6079 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
6080 info when checking whether lattices are bottom.
6082 2014-02-20 Richard Biener <rguenther@suse.de>
6085 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
6088 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
6091 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
6092 parameter specifying the scaling.
6093 (inline_call): Update.
6094 (want_inline_recursively): Guard division by zero.
6095 (recursive_inlining): Update.
6096 * ipa-inline.h (clone_inlined_nodes): Update.
6098 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6101 * config/i386/i386.c (classify_argument): Pass structures of size
6102 64 bytes or less in register.
6104 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6105 Kirill Yukhin <kirill.yukhin@intel.com>
6107 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
6108 (_mm_rcp28_round_ss): Ditto.
6109 (_mm_rsqrt28_round_sd): Ditto.
6110 (_mm_rsqrt28_round_ss): Ditto.
6111 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
6112 (_mm_rcp14_round_ss): Ditto.
6113 (_mm_rsqrt14_round_sd): Ditto.
6114 (_mm_rsqrt14_round_ss): Ditto.
6115 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
6116 the first input operand, get rid of match_dup.
6117 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
6119 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
6121 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
6122 operand as the first input operand, set type attribute.
6123 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
6125 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
6126 operand as the first input operand, set type attribute.
6128 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6130 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
6133 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
6136 * config/i386/i386.c (construct_container): Remove TFmode check
6137 for X86_64_INTEGER_CLASS.
6139 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
6142 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
6143 only when -Wpsabi is enabled.
6145 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
6148 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
6149 passing arrays in registers are the same as for structs, so remove the
6150 special case for them.
6152 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
6154 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
6155 destination type, extract only the valid bits if the source type is not
6156 integral and has a different mode.
6158 2014-02-19 Richard Biener <rguenther@suse.de>
6161 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
6164 2014-02-19 Richard Biener <rguenther@suse.de>
6167 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
6168 (ipa_modify_call_arguments): Emit an argument load explicitely and
6169 preserve virtual SSA form there and for the replacement call.
6170 Do not update SSA form nor free dominance info.
6172 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6174 * ipa.c (function_and_variable_visibility): Also clear WEAK
6175 flag when disolving COMDAT_GROUP.
6177 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6179 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
6180 * ipa-prop.c (ipa_set_jf_known_type): Return early when
6182 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
6183 do more sanity checks.
6184 (detect_type_change): Return true when giving up early.
6185 (compute_complex_assign_jump_func): Fix type parameter of
6186 ipa_set_ancestor_jf.
6187 (compute_complex_ancestor_jump_func): Likewise.
6188 (update_jump_functions_after_inlining): Fix updating of
6190 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
6192 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6194 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
6195 inline clones when edge disappears.
6197 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6200 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
6201 Split 64-bit moves into 2 patterns. Do not allow the use of
6202 direct move for TDmode in little endian, since the decimal value
6203 has little endian bytes within a word, but the 64-bit pieces are
6204 ordered in a big endian fashion, and normal subreg's of TDmode are
6206 (mov<mode>_64bit_dm): Likewise.
6207 (movtd_64bit_nodm): Likewise.
6209 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6211 PR tree-optimization/60174
6212 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
6213 statement of an SSA_NAME that occurs in an abnormal PHI node.
6215 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6218 * final.c (SEEN_BB): Remove.
6219 (SEEN_NOTE, SEEN_EMITTED): Renumber.
6220 (final_scan_insn): Don't force_source_line on second
6221 NOTE_INSN_BASIC_BLOCK.
6223 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
6226 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
6227 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
6228 (type_natural_mode): Warn ABI change when %zmm register is not
6229 available for AVX512F vector value passing.
6231 2014-02-18 Kai Tietz <ktietz@redhat.com>
6234 * config/i386/i386.c (ix86_expand_prologue): Use value in
6235 rax register as displacement when restoring %r10 or %rax.
6236 Fix wrong offset when restoring both registers.
6238 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6240 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
6241 assertion with conditional return.
6243 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6244 Uros Bizjak <ubizjak@gmail.com>
6247 * config/i386/driver-i386.c (host_detect_local_cpu): If
6248 YMM state is not saved by the OS, also clear has_f16c. Move
6249 CPUID 0x80000001 handling before YMM state saving checking.
6251 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
6253 PR rtl-optimization/58960
6254 * haifa-sched.c (alloc_global_sched_pressure_data): New,
6255 factored out from ...
6256 (sched_init): ... here.
6257 (free_global_sched_pressure_data): New, factored out from ...
6258 (sched_finish): ... here.
6259 * sched-int.h (free_global_sched_pressure_data): Declare.
6260 * sched-rgn.c (nr_regions_initial): New static global.
6261 (haifa_find_rgns): Initialize it.
6262 (schedule_region): Disable sched-pressure for the newly
6265 2014-02-17 Richard Biener <rguenther@suse.de>
6267 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
6268 release SSA defs of pattern stmts.
6270 2014-02-17 Richard Biener <rguenther@suse.de>
6272 * tree-inline.c (expand_call_inline): Release the virtual
6273 operand defined by the call we are about to inline.
6275 2014-02-17 Richard Biener <rguenther@suse.de>
6277 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
6279 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
6280 Ilya Tocar <ilya.tocar@intel.com>
6282 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
6283 arguments order in builtin.
6284 (_mm512_permutexvar_epi64): Ditto.
6285 (_mm512_mask_permutexvar_epi64): Ditto
6286 (_mm512_maskz_permutexvar_epi32): Ditto
6287 (_mm512_permutexvar_epi32): Ditto
6288 (_mm512_mask_permutexvar_epi32): Ditto
6290 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6292 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
6293 (p8_vmrgow): Likewise.
6295 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6297 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
6300 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
6303 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
6304 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
6305 into 64-bit and 32-bit moves. On 64-bit moves, add support for
6306 using direct move instructions on ISA 2.07. Also adjust
6307 instruction length for 64-bit.
6308 (mov<mode>_64bit, TFmode/TDmode): Likewise.
6309 (mov<mode>_32bit, TFmode/TDmode): Likewise.
6311 2014-02-15 Alan Modra <amodra@gmail.com>
6315 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
6316 find_replacement on parts of insn rtl that might be reloaded.
6318 2014-02-15 Richard Biener <rguenther@suse.de>
6320 PR tree-optimization/60183
6321 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
6322 (tree_ssa_phiprop): Calculate and free post-dominators.
6324 2014-02-14 Jeff Law <law@redhat.com>
6326 PR rtl-optimization/60131
6327 * ree.c (get_extended_src_reg): New function.
6328 (combine_reaching_defs): Use it rather than assuming location of REG.
6329 (find_and_remove_re): Verify first operand of extension is
6330 a REG before adding the insns to the copy list.
6332 2014-02-14 Roland McGrath <mcgrathr@google.com>
6334 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
6335 * configure: Regenerated.
6336 * config.in: Regenerated.
6337 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
6338 instead of ASM_SHORT.
6340 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
6341 Richard Earnshaw <rearnsha@arm.com>
6343 PR rtl-optimization/59535
6344 * lra-constraints.c (process_alt_operands): Encourage alternative
6345 when unassigned pseudo class is superset of the alternative class.
6346 (inherit_reload_reg): Don't inherit when optimizing for code size.
6347 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
6348 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
6349 modes not less than 4 for Thumb1.
6351 2014-02-14 Kyle McMartin <kyle@redhat.com>
6354 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
6356 2014-02-14 Richard Biener <rguenther@suse.de>
6358 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
6359 (get_frame_arg): Drop the assert with langhook types_compatible_p.
6360 Do not strip INDIRECT_REFs.
6362 2014-02-14 Richard Biener <rguenther@suse.de>
6365 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
6366 DECL_FUNCTION_SPECIFIC_TARGET.
6367 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
6368 * tree-streamer-out.c (pack_ts_target_option): Remove.
6369 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
6370 (write_ts_function_decl_tree_pointers): Do not stream
6371 DECL_FUNCTION_SPECIFIC_TARGET.
6372 * tree-streamer-in.c (unpack_ts_target_option): Remove.
6373 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
6374 (lto_input_ts_function_decl_tree_pointers): Do not stream
6375 DECL_FUNCTION_SPECIFIC_TARGET.
6377 2014-02-14 Jakub Jelinek <jakub@redhat.com>
6379 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
6380 (get_initial_def_for_induction, vectorizable_induction): Ignore
6381 debug stmts when looking for exit_phi.
6382 (vectorizable_live_operation): Fix up condition.
6384 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6386 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
6387 nreverse() because it changes the content of original tree list.
6389 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6391 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
6392 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
6394 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6396 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
6397 GNU coding standards.
6399 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6402 * dwarf2out.c (gen_subprogram_die): Don't call
6403 add_calling_convention_attribute if subr_die is old_die.
6405 2014-02-13 Sharad Singhai <singhai@google.com>
6407 * doc/optinfo.texi: Fix order of nodes.
6409 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
6411 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
6412 operands[2], not operands[3].
6414 2014-02-13 Richard Biener <rguenther@suse.de>
6417 * doc/install.texi (ISL): Update recommended version to 0.12.2,
6418 mention the possibility of an in-tree build.
6419 (CLooG): Update recommended version to 0.18.1, mention the
6420 possibility of an in-tree build and clarify that the ISL
6421 bundled with CLooG does not work.
6423 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6426 * expr.c (compress_float_constant): If x is a hard register,
6427 extend into a pseudo and then move to x.
6429 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
6431 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
6432 caused by bad second argument to warning_at() with -mhotpatch and
6433 nested functions (e.g. with gfortran).
6435 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
6437 * opts.c (option_name): Remove "enabled by default" rider.
6439 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
6441 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
6443 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
6444 Uros Bizjak <ubizjak@gmail.com>
6447 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
6448 * configure: Regenerated.
6450 2014-02-12 Richard Biener <rguenther@suse.de>
6452 * vec.c (vec_prefix::calculate_allocation): Move as
6453 inline variant to vec.h.
6454 (vec_prefix::calculate_allocation_1): New out-of-line version.
6455 * vec.h (vec_prefix::calculate_allocation_1): Declare.
6456 (vec_prefix::m_has_auto_buf): Rename to ...
6457 (vec_prefix::m_using_auto_storage): ... this.
6458 (vec_prefix::calculate_allocation): Inline the easy cases
6459 and dispatch to calculate_allocation_1 which doesn't need the
6461 (va_heap::reserve): Use gcc_checking_assert.
6462 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
6463 m_using_auto_storage.
6464 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
6466 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
6467 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
6468 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
6470 2014-02-12 Richard Biener <rguenther@suse.de>
6472 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
6473 when we found a dependence.
6475 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
6477 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
6479 (maybe_fold_stmt): ... into this new function.
6480 * omp-low.c (lower_omp): Update comment.
6482 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
6485 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
6488 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
6490 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
6491 identifiers in comments.
6492 (cortexa53_extra_costs): Likewise.
6493 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
6494 (cortexa7_extra_costs): Likewise.
6495 (cortexa12_extra_costs): Likewise.
6496 (cortexa15_extra_costs): Likewise.
6497 (v7m_extra_costs): Likewise.
6499 2014-02-12 Richard Biener <rguenther@suse.de>
6502 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
6503 of posix_memalign being successful.
6504 (lower_stmt): Restrict lowering of posix_memalign to when
6505 -ftree-bit-ccp is enabled.
6507 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6509 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
6511 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
6513 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
6515 PR rtl-optimization/60116
6516 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
6517 other_insn once the combination has been validated.
6519 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
6522 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
6524 * ipa-devirt.c: Include demangle.h
6525 (odr_violation_reported): New static variable.
6526 (add_type_duplicate): Update odr_violations.
6527 (maybe_record_node): Add completep parameter; update it.
6528 (record_target_from_binfo): Add COMPLETEP parameter;
6529 update it as needed.
6530 (possible_polymorphic_call_targets_1): Likewise.
6531 (struct polymorphic_call_target_d): Add nonconstruction_targets;
6532 rename FINAL to COMPLETE.
6533 (record_targets_from_bases): Sanity check we found the binfo;
6534 fix COMPLETEP updating.
6535 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
6536 parameter, fix computing of COMPLETEP.
6537 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
6538 at LTO time do demangling.
6539 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
6540 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
6542 (gimple_get_virt_method_for_binfo): Likewise.
6543 * gimple-fold.h (gimple_get_virt_method_for_binfo,
6544 gimple_get_virt_method_for_vtable): Update prototypes.
6546 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
6549 * genautomata.c (add_presence_absence): Fix typo with
6550 {final_}presence_list.
6552 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
6555 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
6556 for VSX/Altivec vectors that land in GPR registers.
6558 2014-02-11 Richard Henderson <rth@redhat.com>
6559 Jakub Jelinek <jakub@redhat.com>
6562 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
6563 around drhs if type conversion to lacc->type is not useless.
6565 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6567 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
6569 (cortex-a57.cortex-a53): Likewise.
6570 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
6572 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6574 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
6577 2014-02-11 Renlin Li <Renlin.Li@arm.com>
6579 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
6580 add_options_for_arm_vfp3.
6582 2014-02-11 Jeff Law <law@redhat.com>
6585 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
6586 object with an undesirable mode.
6588 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6591 * config/i386/sol2-9.h: New file.
6592 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
6593 *-*-solaris2.9*): Use it.
6595 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6597 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
6598 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
6600 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6602 * config/microblaze/microblaze.c: Extend mcpu version format
6604 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
6606 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
6608 2014-02-10 Richard Henderson <rth@redhat.com>
6611 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
6612 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
6613 ms-abi vs -mno-accumulate-outgoing-args.
6614 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
6615 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
6618 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
6621 * cfgexpand.c (expand_asm_operands): Attach source location to
6622 ASM_INPUT rtx objects.
6623 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
6625 2014-02-10 Nick Clifton <nickc@redhat.com>
6627 * config/mn10300/mn10300.c (popcount): New function.
6628 (mn10300_expand_prologue): Include saved registers in stack usage
6631 2014-02-10 Jeff Law <law@redhat.com>
6634 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
6635 when changing the SET_DEST of a prior insn to avoid an input reload.
6637 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6639 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
6640 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
6641 -mcall-openbsd, or -mcall-linux.
6642 (CC1_ENDIAN_BIG_SPEC): Remove.
6643 (CC1_ENDIAN_LITTLE_SPEC): Remove.
6644 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
6645 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
6646 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
6647 and %cc1_endian_default.
6648 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
6650 2014-02-10 Richard Biener <rguenther@suse.de>
6652 PR tree-optimization/60115
6653 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
6654 MEM_REF handling. Properly verify that the accesses are not
6655 out of the objects bound.
6657 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6659 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
6662 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
6664 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
6665 proper constants and fix formatting.
6666 (possible_polymorphic_call_targets): Fix formatting.
6668 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
6669 Ilya Tocar <ilya.tocar@intel.com>
6671 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
6672 (_mm512_loadu_epi32): Renamed into...
6673 (_mm512_loadu_si512): This.
6674 (_mm512_storeu_epi32): Renamed into...
6675 (_mm512_storeu_si512): This.
6676 (_mm512_maskz_ceil_ps): Removed.
6677 (_mm512_maskz_ceil_pd): Ditto.
6678 (_mm512_maskz_floor_ps): Ditto.
6679 (_mm512_maskz_floor_pd): Ditto.
6680 (_mm512_floor_round_ps): Ditto.
6681 (_mm512_floor_round_pd): Ditto.
6682 (_mm512_ceil_round_ps): Ditto.
6683 (_mm512_ceil_round_pd): Ditto.
6684 (_mm512_mask_floor_round_ps): Ditto.
6685 (_mm512_mask_floor_round_pd): Ditto.
6686 (_mm512_mask_ceil_round_ps): Ditto.
6687 (_mm512_mask_ceil_round_pd): Ditto.
6688 (_mm512_maskz_floor_round_ps): Ditto.
6689 (_mm512_maskz_floor_round_pd): Ditto.
6690 (_mm512_maskz_ceil_round_ps): Ditto.
6691 (_mm512_maskz_ceil_round_pd): Ditto.
6692 (_mm512_expand_pd): Ditto.
6693 (_mm512_expand_ps): Ditto.
6694 * config/i386/i386.c (ix86_builtins): Remove
6695 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
6696 (bdesc_args): Ditto.
6697 * config/i386/predicates.md (const1256_operand): New.
6698 (const_1_to_2_operand): Ditto.
6699 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6700 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6701 (*avx512pf_gatherpf<mode>sf): Ditto.
6702 (avx512pf_gatherpf<mode>df): Ditto.
6703 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6704 (*avx512pf_gatherpf<mode>df): Ditto.
6705 (avx512pf_scatterpf<mode>sf): Ditto.
6706 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6707 (*avx512pf_scatterpf<mode>sf): Ditto.
6708 (avx512pf_scatterpf<mode>df): Ditto.
6709 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6710 (*avx512pf_scatterpf<mode>df): Ditto.
6711 (avx512f_expand<mode>): Removed.
6712 (<shift_insn><mode>3<mask_name>): Change predicate type.
6714 2014-02-08 Jakub Jelinek <jakub@redhat.com>
6716 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
6717 not at the end of datarefs vector use ordered_remove to avoid
6718 reordering datarefs vector.
6721 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
6722 mark local addressable non-static vars as GOVD_PRIVATE
6723 instead of GOVD_LOCAL.
6724 * omp-low.c (lower_omp_for): Move gimple_bind_vars
6725 and BLOCK_VARS of gimple_bind_block to new_stmt rather
6729 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
6730 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
6731 assume_aligned or alloc_align attributes.
6732 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
6733 arguments. Handle also assume_aligned and alloc_align attributes.
6734 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
6735 calls to functions with assume_aligned or alloc_align attributes.
6736 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
6738 2014-02-08 Terry Guo <terry.guo@arm.com>
6740 * doc/invoke.texi: Document ARM -march=armv7e-m.
6742 2014-02-08 Jakub Jelinek <jakub@redhat.com>
6744 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
6745 flag on __cilkrts_rethrow builtin.
6748 * ipa-cp.c (determine_versionability): Fail at -O0
6749 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
6750 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
6753 2014-02-04 Jakub Jelinek <jakub@redhat.com>
6756 * tree-inline.c (copy_forbidden): Fail for
6757 __attribute__((optimize (0))) functions.
6759 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
6761 * varpool.c: Include pointer-set.h.
6762 (varpool_remove_unreferenced_decls): Variables in other partitions
6763 will not be output; be however careful to not lose information
6766 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
6768 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
6769 lookup in the vtable constructor.
6771 2014-02-07 Jeff Law <law@redhat.com>
6774 * config/m68k/m68k.md (ashldi_extsi): Turn into a
6775 define_insn_and_split.
6777 * ipa-inline.c (inline_small_functions): Fix typos.
6779 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6781 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
6782 (s390_can_use_return_insn): Declare.
6783 * config/s390/s390.h (EPILOGUE_USES): Define.
6784 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
6786 (s390_chunkify_start): Handle return JUMP_LABELs.
6787 (s390_early_mach): Emit a main_pool instruction on the entry edge.
6788 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
6789 (s390_can_use_return_insn): New functions.
6790 (s390_fix_long_loop_prediction): Handle conditional returns.
6791 (TARGET_SET_UP_BY_PROLOGUE): Define.
6792 * config/s390/s390.md (ANY_RETURN): New code iterator.
6793 (*creturn, *csimple_return, return, simple_return): New patterns.
6795 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6797 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
6798 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
6799 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
6800 REG_CFA_RESTORE list when deciding not to restore a register.
6802 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6804 * config/s390/s390.c: Include tree-pass.h and context.h.
6805 (s390_early_mach): New function, split out from...
6806 (s390_emit_prologue): ...here.
6807 (pass_data_s390_early_mach): New pass structure.
6808 (pass_s390_early_mach): New class.
6809 (s390_option_override): Create and register early_mach pass.
6810 Move to end of file.
6812 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6814 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
6815 to match for the exit block.
6817 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6819 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
6820 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
6821 Reject misaligned operands.
6823 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6825 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
6827 2014-02-07 Richard Biener <rguenther@suse.de>
6830 * gimple-low.c (lower_builtin_posix_memalign): New function.
6831 (lower_stmt): Call it to lower posix_memalign in a way
6832 to make alignment info accessible.
6834 2014-02-07 Jakub Jelinek <jakub@redhat.com>
6837 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
6838 __builtin_setjmp_receiver.
6840 2014-02-07 Richard Biener <rguenther@suse.de>
6843 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
6844 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
6845 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
6846 Handle BUILT_IN_POSIX_MEMALIGN.
6847 (find_func_clobbers): Likewise.
6848 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
6849 (call_may_clobber_ref_p_1): Likewise.
6851 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6854 * ipa-devirt.c (record_target_from_binfo): Remove overactive
6857 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6860 * lto-cgraph.c (lto_output_node): Use
6861 symtab_get_symbol_partitioning_class.
6862 (lto_output_varpool_node): likewise.
6863 (symtab_get_symbol_partitioning_class): Move here from
6865 * cgraph.h (symbol_partitioning_class): Likewise.
6866 (symtab_get_symbol_partitioning_class): Declare.
6868 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6870 * ggc.h (ggc_internal_cleared_alloc): New macro.
6871 * vec.h (vec_safe_copy): Handle memory stats.
6872 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
6873 * target-globals.c (save_target_globals): Likewise.
6875 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6878 * expr.c (emit_move_resolve_push): Export; be bit more selective
6879 on when to clear alias set.
6880 * expr.h (emit_move_resolve_push): Declare.
6881 * function.h (struct function): Add tail_call_marked.
6882 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
6883 * config/i386/i386-protos.h (ix86_expand_push): Remove.
6884 * config/i386/i386.md (TImode move expander): De not call
6886 (FP push expanders): Preserve memory attributes.
6887 * config/i386/sse.md (push<mode>1): Remove.
6888 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
6889 (ix86_expand_push): Remove.
6890 * config/i386/mmx.md (push<mode>1): Remove.
6892 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6894 PR rtl-optimization/60030
6895 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
6896 lopart with paradoxical subreg before shifting it up by hprec.
6898 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6900 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
6901 Remove extra newline at end of file.
6902 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
6903 (arm_issue_rate): Handle cortexa57.
6904 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
6905 (cortex-a57.cortex-a53): Likewise.
6907 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6910 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
6911 don't record in REG_FRAME_RELATED_EXPR registers not set in that
6913 (arm_expand_prologue): Adjust all callers.
6914 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
6915 info, registers also at the lowest numbered registers side. Use
6916 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
6920 * var-tracking.c (adjust_mems): Before adding a SET to
6921 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
6923 2014-02-06 Alan Modra <amodra@gmail.com>
6926 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
6927 change SDmode to DDmode when lra_in_progress.
6929 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6932 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
6933 free_data_ref on the dr first, and before goto again also set dr
6934 to the next dr. For simd_lane_access, free old datarefs[i] before
6935 overwriting it. For get_vectype_for_scalar_type failure, don't
6936 free_data_ref if simd_lane_access.
6938 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
6941 * tree.h (opts_for_fn): New inline function.
6942 (opt_for_fn): Define.
6943 * config/i386/i386.c (ix86_function_regparm): Use
6944 opt_for_fn (decl, optimize) instead of optimize.
6946 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
6948 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
6949 for SYMBOL_REF in large memory model.
6951 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6953 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
6955 (cortex-a57): Likewise.
6956 (cortex-a57.cortex-a53): Likewise.
6958 2014-02-06 Yury Gribov <y.gribov@samsung.com>
6959 Kugan Vivekanandarajah <kuganv@linaro.org>
6961 * config/arm/arm.c (arm_vector_alignment_reachable): Check
6963 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
6965 2014-02-06 Richard Biener <rguenther@suse.de>
6967 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
6968 set_loop_copy and initialize_original_copy_tables.
6970 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
6972 * config/aarch64/aarch64-simd.md
6973 (aarch64_ashr_simddi): Change QI to SI.
6975 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
6976 Jakub Jelinek <jakub@redhat.com>
6979 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
6982 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6984 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
6985 CODE_FOR_altivec_vpku[hw]um to
6986 CODE_FOR_altivec_vpku[hw]um_direct.
6987 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
6988 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
6989 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
6990 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
6992 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6994 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
6995 generation for -maltivec=be.
6996 (altivec_vsumsws): Simplify redundant test.
6998 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7000 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
7001 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
7002 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
7003 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
7004 gen_altivec_vpkuwum.
7005 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
7007 (altivec_vpks<VI_char>ss): Likewise.
7008 (altivec_vpks<VI_char>us): Likewise.
7009 (altivec_vpku<VI_char>us): Likewise.
7010 (altivec_vpku<VI_char>um): Likewise.
7011 (altivec_vpku<VI_char>um_direct): New (copy of
7012 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
7014 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
7015 target is little endian and -maltivec=be is not specified.
7016 (*altivec_vupkhs<VU_char>_direct): New (copy of
7017 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
7018 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
7019 target is little endian and -maltivec=be is not specified.
7020 (*altivec_vupkls<VU_char>_direct): New (copy of
7021 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
7022 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
7023 little endian and -maltivec=be is not specified.
7024 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
7025 little endian and -maltivec=be is not specified.
7027 2014-02-05 Richard Henderson <rth@redhat.com>
7030 * combine-stack-adj.c: Revert r206943.
7031 * sched-int.h (struct deps_desc): Add last_args_size.
7032 * sched-deps.c (init_deps): Initialize it.
7033 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7034 contain REG_ARGS_SIZE notes.
7036 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7038 * lto-cgraph.c (asm_nodes_output): Make global.
7039 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7040 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7041 (driver_handle_option): Handle OPT_fwpa.
7043 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7046 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7047 a comment typo and formatting issue. If odr_hash hasn't been
7048 created, return vNULL and set *completep to false.
7051 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7052 bb with no successors.
7054 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7057 * doc/invoke.texi (-march): Clarify documentation for ARM.
7061 2014-02-05 Richard Biener <rguenther@suse.de>
7063 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
7064 when not vectorizing because of too many alias checks.
7065 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
7066 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
7068 2014-02-05 Nick Clifton <nickc@redhat.com>
7070 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
7071 accept extended registers in any mode when compiling for the MN10300.
7073 2014-02-05 Yury Gribov <y.gribov@samsung.com>
7075 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
7076 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
7077 sanitization attributes.
7078 (can_inline_edge_p): Likewise.
7079 (sanitize_attrs_match_for_inline_p): New function.
7081 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7083 * ipa-prop.c (detect_type_change): Shor circuit testing of
7084 type changes on THIS pointer.
7086 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
7089 * config/pa/pa.c (legitimize_tls_address): Return original address
7090 if not passed a SYMBOL_REF rtx.
7091 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
7093 (pa_emit_move_sequence): Simplify TLS source operands.
7094 (pa_legitimate_constant_p): Reject all TLS constants.
7095 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
7096 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
7098 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7100 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
7101 groups when we know they are controlled by LTO.
7102 * varasm.c (default_binds_local_p_1): If object is in other partition,
7103 it will be resolved locally.
7105 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7107 * config/host-linux.c (linux_gt_pch_use_address): Don't
7108 use SSIZE_MAX because it is not always defined.
7110 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
7113 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
7114 threshold for pseudo splitting.
7115 (update_ebb_live_info): Process call argument hard registers and
7116 hard registers from insn definition too.
7117 (max_small_class_regs_num): New constant.
7118 (inherit_in_ebb): Update live hard regs through EBBs. Update
7119 reloads_num only for small register classes. Don't split for
7122 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
7125 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
7128 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7130 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
7133 2014-02-04 Marek Polacek <polacek@redhat.com>
7135 * gdbinit.in (pel): Define.
7137 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7139 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
7142 2014-02-04 Richard Biener <rguenther@suse.de>
7145 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
7146 in function context local.
7147 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
7148 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
7149 similar to LTO_imported_decl_ref.
7151 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7153 PR tree-optimization/60002
7154 * cgraphclones.c (build_function_decl_skip_args): Clear
7157 PR tree-optimization/60023
7158 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
7159 false to gsi_replace.
7160 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
7161 has been in some EH region and vec_stmt could throw, add
7162 vec_stmt into the same EH region.
7163 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
7164 has no lhs, ignore it.
7165 * internal-fn.c (expand_MASK_LOAD): Likewise.
7168 * tree-inline.c (copy_forbidden): Fail for
7169 __attribute__((optimize (0))) functions.
7172 * omp-low.c (simd_clone_struct_copy): If from->inbranch
7173 is set, copy one less argument.
7174 (expand_simd_clones): Don't subtract clone_info->inbranch
7175 from simd_clone_struct_alloc argument.
7177 PR rtl-optimization/57915
7178 * recog.c (simplify_while_replacing): If all unary/binary/relational
7179 operation arguments are constant, attempt to simplify those.
7182 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
7183 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
7185 2014-02-04 Richard Biener <rguenther@suse.de>
7187 PR tree-optimization/60012
7188 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
7189 TBAA disambiguation to all DDRs.
7191 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7194 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
7195 (LINK_SPEC): Use it for -shared, -shared-libgcc.
7197 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7200 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
7202 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7204 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
7205 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
7207 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7210 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
7211 to figure out targets of polymorphic calls with known decl.
7212 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7213 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
7214 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
7215 (get_polymorphic_call_info): ... here.
7216 (get_polymorphic_call_info_from_invariant): New function.
7218 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7220 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
7221 lookup via vtable pointer; check for type consistency
7222 and turn inconsitent facts into UNREACHABLE.
7223 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7224 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
7225 type inconsistent querries; return UNREACHABLE instead.
7227 2014-02-03 Richard Henderson <rth@twiddle.net>
7230 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
7231 already processed this node.
7232 (normalize_one_pred_1): Pass along mark_set.
7233 (normalize_one_pred): Create and destroy a pointer_set_t.
7234 (normalize_one_pred_chain): Likewise.
7236 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
7238 PR gcov-profile/58602
7239 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
7241 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7244 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
7245 -fno-devirtualize; try to devirtualize by the knowledge of
7246 virtual table pointer given by aggregate propagation.
7247 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7248 (ipa_print_node_jump_functions): Dump also offset that
7249 is relevant for polymorphic calls.
7250 (determine_known_aggregate_parts): Add arg_type parameter; use it
7251 instead of determining the type from pointer type.
7252 (ipa_compute_jump_functions_for_edge): Update call of
7253 determine_known_aggregate_parts.
7254 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
7255 (gimple_get_virt_method_for_binfo): ... here; simplify using
7256 vtable_pointer_value_to_vtable.
7257 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
7258 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
7259 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
7260 (vtable_pointer_value_to_vtable): Break out from ...; handle also
7262 (vtable_pointer_value_to_binfo): ... here.
7263 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
7265 2014-02-03 Teresa Johnson <tejohnson@google.com>
7267 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
7268 redef of outer loop index variable.
7270 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
7274 * doc/extend.texi (Function Attributes): Typo.
7276 2014-02-03 Cong Hou <congh@google.com>
7278 PR tree-optimization/60000
7279 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
7280 if the vectorized statement is a store. A store statement can only
7281 appear at the end of pattern statements.
7283 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7285 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
7286 (ix86_option_override_internal): Default long double to 64-bit for
7287 32-bit Bionic and to 128-bit for 64-bit Bionic.
7289 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
7290 TARGET_LONG_DOUBLE_128 is true.
7291 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
7293 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
7294 (mlong-double-64): Negate -mlong-double-128.
7295 (mlong-double-128): New option.
7297 * config/i386/i386-c.c (ix86_target_macros): Define
7298 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
7300 * doc/invoke.texi: Document -mlong-double-128.
7302 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7304 PR rtl-optimization/60024
7305 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
7307 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
7309 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
7311 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7313 PR rtl-optimization/57662
7314 * sel-sched.c (code_motion_path_driver): Do not mark already not
7315 existing blocks in the visiting bitmap.
7317 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7319 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
7320 on the insn being emitted.
7322 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
7323 Will Deacon <will.deacon@arm.com>
7325 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
7327 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7329 * config/arm/arm-tables.opt: Regenerate.
7331 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7333 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
7334 for vector types other than V16QImode.
7335 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
7336 define_expand, and call altivec_expand_vec_perm_le when producing
7337 code with little endian element order.
7338 (*altivec_vperm_<mode>_internal): New insn having previous
7339 behavior of altivec_vperm_<mode>.
7340 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
7341 altivec_expand_vec_perm_le when producing code with little endian
7343 (*altivec_vperm_<mode>_uns_internal): New insn having previous
7344 behavior of altivec_vperm_<mode>_uns.
7346 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7348 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
7349 (altivec_vsumsws): Add handling for -maltivec=be with a little
7351 (altivec_vsumsws_direct): New.
7352 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
7353 gen_altivec_vsumsws.
7355 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7357 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
7358 vtable_pointer_value_to_binfo): New functions.
7359 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
7360 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
7362 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
7364 * config/nios2/nios2.md (load_got_register): Initialize GOT
7365 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
7366 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
7368 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7370 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
7371 preserverd by passthrough, do not propagate the type.
7373 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7375 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
7376 (mips_atomic_assign_expand_fenv): New function.
7377 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
7379 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7381 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
7382 (__builtin_mips_set_fcsr): Likewise.
7383 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
7384 MIPS_USI_FTYPE_VOID.
7385 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
7386 (mips16_expand_set_fcsr): Likewise.
7387 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
7388 (mips16_set_fcsr_stub): Likewise.
7389 (mips16_get_fcsr_one_only_stub): New class.
7390 (mips16_set_fcsr_one_only_stub): Likewise.
7391 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
7392 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
7393 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
7394 (hard_float): New availability predicate.
7395 (mips_builtins): Add get_fcsr and set_fcsr.
7396 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
7397 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
7398 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
7399 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
7400 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
7403 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7405 * config/mips/mips.c (mips_one_only_stub): New class.
7406 (mips_need_mips16_rdhwr_p): Replace with...
7407 (mips16_rdhwr_stub): ...this new variable.
7408 (mips16_stub_call_address): New function.
7409 (mips16_rdhwr_one_only_stub): New class.
7410 (mips_expand_thread_pointer): Use mips16_stub_call_address.
7411 (mips_output_mips16_rdhwr): Delete.
7412 (mips_finish_stub): New function.
7413 (mips_code_end): Use it to handle rdhwr stubs.
7415 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
7418 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
7419 when calculating size of integer atomic types.
7421 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
7423 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
7425 2014-02-01 Jakub Jelinek <jakub@redhat.com>
7427 PR tree-optimization/60003
7428 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
7429 * profile.c (branch_prob): Use gimple_call_builtin_p
7430 to check for BUILT_IN_SETJMP_RECEIVER.
7431 * tree-inline.c (copy_bb): Call notice_special_calls.
7433 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
7436 * lra-constraints.c (process_alt_operands): Update reload_sum only
7439 2014-01-31 Richard Henderson <rth@redhat.com>
7442 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
7443 until after else_eh is processed.
7445 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7447 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
7448 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
7449 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
7450 in smmintrin.h, remove them.
7451 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
7452 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
7453 * config/i386/i386.md (ROUND_SAE): Fix value.
7454 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
7455 (const48_operand): New.
7456 * config/i386/subst.md (round), (round_expand): Use
7457 const_4_or_8_to_11_operand.
7458 (round_saeonly), (round_saeonly_expand): Use const48_operand.
7460 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7462 * config/i386/constraints.md (Yk): Swap meaning with k.
7463 * config/i386/i386.md (movhi_internal): Change Yk to k.
7464 (movqi_internal): Ditto.
7465 (*k<logic><mode>): Ditto.
7468 (kandn<mode>): Ditto.
7469 (*<code>hi_1): Ditto.
7470 (*<code>qi_1): Ditto.
7471 (kxnor<mode>): Ditto.
7472 (kortestzhi): Ditto.
7473 (kortestchi): Ditto.
7475 (*one_cmplhi2_1): Ditto.
7476 (*one_cmplqi2_1): Ditto.
7477 * config/i386/sse.md (): Change k to Yk.
7478 (avx512f_load<mode>_mask): Ditto.
7479 (avx512f_blendm<mode>): Ditto.
7480 (avx512f_store<mode>_mask): Ditto.
7481 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
7482 (avx512f_storedqu<mode>_mask): Ditto.
7483 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
7485 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
7486 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
7487 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
7488 (avx512f_maskcmp<mode>3): Ditto.
7489 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
7490 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
7491 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
7492 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
7493 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
7494 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
7495 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
7496 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
7497 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
7498 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
7499 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
7500 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
7501 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
7502 (vec_extract_lo_<mode>_maskm): Ditto.
7503 (vec_extract_hi_<mode>_maskm): Ditto.
7504 (avx512f_vternlog<mode>_mask): Ditto.
7505 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
7506 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
7507 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
7508 (avx512f_<code>v8div16qi2_mask): Ditto.
7509 (avx512f_<code>v8div16qi2_mask_store): Ditto.
7510 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
7511 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
7512 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
7513 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
7514 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7515 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7516 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7517 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7518 (avx512cd_maskb_vec_dupv8di): Ditto.
7519 (avx512cd_maskw_vec_dupv16si): Ditto.
7520 (avx512f_vpermi2var<mode>3_maskz): Ditto.
7521 (avx512f_vpermi2var<mode>3_mask): Ditto.
7522 (avx512f_vpermi2var<mode>3_mask): Ditto.
7523 (avx512f_vpermt2var<mode>3_maskz): Ditto.
7524 (*avx512f_gathersi<mode>): Ditto.
7525 (*avx512f_gathersi<mode>_2): Ditto.
7526 (*avx512f_gatherdi<mode>): Ditto.
7527 (*avx512f_gatherdi<mode>_2): Ditto.
7528 (*avx512f_scattersi<mode>): Ditto.
7529 (*avx512f_scatterdi<mode>): Ditto.
7530 (avx512f_compress<mode>_mask): Ditto.
7531 (avx512f_compressstore<mode>_mask): Ditto.
7532 (avx512f_expand<mode>_mask): Ditto.
7533 * config/i386/subst.md (mask): Change k to Yk.
7534 (mask_scalar_merge): Ditto.
7537 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
7539 * doc/extend.texi (Vector Extensions): Document ?: in C++.
7541 2014-01-31 Richard Biener <rguenther@suse.de>
7544 * builtins.c (fold_builtin_memory_op): Make sure to not
7545 use a floating-point mode or a boolean or enumeral type for
7548 2014-01-30 DJ Delorie <dj@redhat.com>
7550 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
7551 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
7552 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
7553 whenever main() has an epilogue.
7555 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7557 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
7558 unused variable "field".
7559 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
7560 (vsx_mergeh_<mode>): Likewise.
7561 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
7562 (altivec_vmrghh): Likewise.
7563 (altivec_vmrghw): Likewise.
7564 (altivec_vmrglb): Likewise.
7565 (altivec_vmrglh): Likewise.
7566 (altivec_vmrglw): Likewise.
7567 (altivec_vspltb): Add missing uses.
7568 (altivec_vsplth): Likewise.
7569 (altivec_vspltw): Likewise.
7570 (altivec_vspltsf): Likewise.
7572 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7575 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
7576 frame related instructions.
7578 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
7580 PR rtl-optimization/59959
7581 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
7582 any reload of register whose subreg is invalid.
7584 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7586 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
7587 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
7588 Add missing return type - void.
7590 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7592 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
7593 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
7594 remove element index adjustment for endian (now handled in vsx.md
7596 (altivec_expand_vec_perm_const): Use
7597 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
7598 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
7599 (vsx_xxspltw_<mode>): Adjust element index for little endian.
7600 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
7601 define_expand and a new define_insn *altivec_vspltb_internal;
7602 adjust for -maltivec=be on a little endian target.
7603 (altivec_vspltb_direct): New.
7604 (altivec_vsplth): Divide into a define_expand and a new
7605 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
7606 little endian target.
7607 (altivec_vsplth_direct): New.
7608 (altivec_vspltw): Divide into a define_expand and a new
7609 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
7610 little endian target.
7611 (altivec_vspltw_direct): New.
7612 (altivec_vspltsf): Divide into a define_expand and a new
7613 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
7614 a little endian target.
7616 2014-01-30 Richard Biener <rguenther@suse.de>
7618 PR tree-optimization/59993
7619 * tree-ssa-forwprop.c (associate_pointerplus): Check we
7620 can propagate form the earlier stmt and avoid the transform
7621 when the intermediate result is needed.
7623 2014-01-30 Alangi Derick <alangiderick@gmail.com>
7625 * README.Portability: Fix typo.
7627 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
7629 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
7630 comparison_operator with ordered_comparison_operator.
7632 2014-01-30 Nick Clifton <nickc@redhat.com>
7634 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
7635 Rename to mn10300_store_multiple_regs.
7636 * config/mn10300/mn10300.c: Likewise.
7637 * config/mn10300/mn10300.md (store_movm): Fix typo: call
7638 store_multiple_regs.
7639 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
7640 Call mn10300_store_multiple_regs.
7642 2014-01-30 Nick Clifton <nickc@redhat.com>
7643 DJ Delorie <dj@redhat.com>
7645 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
7646 %fp 2 to keep registers after it properly word-aligned.
7647 (rl78_alloc_physical_registers_umul): Handle the case where both
7648 input operands are the same.
7650 2014-01-30 Richard Biener <rguenther@suse.de>
7652 PR tree-optimization/59903
7653 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
7656 2014-01-30 Jason Merrill <jason@redhat.com>
7659 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
7662 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
7664 2014-01-30 Richard Biener <rguenther@suse.de>
7666 PR tree-optimization/59951
7667 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
7669 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
7672 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
7673 SFmode to DFmode case.
7675 2014-01-29 DJ Delorie <dj@redhat.com>
7677 * config/msp430/msp430.opt (-minrt): New.
7678 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
7680 (ENDFILE_SPEC): Likewise.
7682 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
7684 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
7685 (estimate_function_body_sizes): Use it.
7687 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
7690 * dwarf2out.c (is_cxx_auto): New.
7691 (is_base_type): Use it.
7692 (gen_type_die_with_usage): Likewise.
7694 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7696 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
7697 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
7698 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
7699 -maltivec=be with LE targets.
7700 (vsx_mergeh_<mode>): Likewise.
7701 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
7702 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
7703 (altivec_vmrghb): Replace with define_expand and new
7704 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
7705 (altivec_vmrghb_direct): New define_insn.
7706 (altivec_vmrghh): Replace with define_expand and new
7707 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
7708 (altivec_vmrghh_direct): New define_insn.
7709 (altivec_vmrghw): Replace with define_expand and new
7710 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
7711 (altivec_vmrghw_direct): New define_insn.
7712 (*altivec_vmrghsf): Adjust for endianness.
7713 (altivec_vmrglb): Replace with define_expand and new
7714 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
7715 (altivec_vmrglb_direct): New define_insn.
7716 (altivec_vmrglh): Replace with define_expand and new
7717 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
7718 (altivec_vmrglh_direct): New define_insn.
7719 (altivec_vmrglw): Replace with define_expand and new
7720 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
7721 (altivec_vmrglw_direct): New define_insn.
7722 (*altivec_vmrglsf): Adjust for endianness.
7723 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
7724 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
7725 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
7726 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
7727 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
7728 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
7729 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
7730 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
7732 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
7734 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
7735 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
7738 2014-01-29 Richard Biener <rguenther@suse.de>
7740 PR tree-optimization/58742
7741 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
7742 associate_pointerplus_align.
7743 (associate_pointerplus_diff): New function.
7744 (associate_pointerplus): Likewise. Call associate_pointerplus_align
7745 and associate_pointerplus_diff.
7747 2014-01-29 Richard Biener <rguenther@suse.de>
7749 * lto-streamer.h (LTO_major_version): Bump to 3.
7750 (LTO_minor_version): Reset to 0.
7752 2014-01-29 Renlin Li <Renlin.Li@arm.com>
7754 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
7755 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
7756 (arm_file_start): Generate correct asm header for armv7ve.
7757 * config/arm/bpabi.h: Add multilib support for armv7ve.
7758 * config/arm/driver-arm.c: Change the architectures of cortex-a7
7759 and cortex-a15 to armv7ve.
7760 * config/arm/t-aprofile: Add multilib support for armv7ve.
7761 * doc/invoke.texi: Document -march=armv7ve.
7763 2014-01-29 Richard Biener <rguenther@suse.de>
7765 PR tree-optimization/58742
7766 * tree-ssa-forwprop.c (associate_plusminus): Return true
7767 if we changed sth, defer EH cleanup to ...
7768 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
7769 (simplify_mult): New function.
7771 2014-01-29 Jakub Jelinek <jakub@redhat.com>
7774 PR tree-optimization/59920
7775 * tree.c (build_common_builtin_nodes): Remove
7776 __builtin_setjmp_dispatcher initialization.
7777 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
7778 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
7779 instead of gsi_after_labels + manually skipping debug stmts.
7780 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
7781 ignore bbs with IFN_ABNORMAL_DISPATCHER.
7782 * tree-inline.c (copy_edges_for_bb): Remove
7783 can_make_abnormal_goto argument, instead add abnormal_goto_dest
7784 argument. Ignore computed_goto_p stmts. Don't call
7785 make_abnormal_goto_edges. If a call might need abnormal edges
7786 for non-local gotos, see if it already has an edge to
7787 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
7788 with true argument, don't do anything then, otherwise add
7789 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
7790 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
7792 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
7793 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
7794 (lower_stmt): Don't set data->calls_builtin_setjmp.
7795 (lower_builtin_setjmp): Adjust comment.
7796 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
7797 * tree-cfg.c (found_computed_goto): Remove.
7798 (factor_computed_gotos): Remove.
7799 (make_goto_expr_edges): Return bool, true for computed gotos.
7800 Don't call make_abnormal_goto_edges.
7801 (build_gimple_cfg): Don't set found_computed_goto, don't call
7802 factor_computed_gotos.
7803 (computed_goto_p): No longer static.
7804 (make_blocks): Don't set found_computed_goto.
7805 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
7806 (make_edges): If make_goto_expr_edges returns true, push bb
7807 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
7808 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
7809 vector. Record mapping between bbs and OpenMP regions if there
7810 are any, adjust make_gimple_omp_edges caller. Call
7811 handle_abnormal_edges.
7812 (make_abnormal_goto_edges): Remove.
7813 * tree-cfg.h (make_abnormal_goto_edges): Remove.
7814 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
7815 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
7816 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
7817 * internal-fn.def (ABNORMAL_DISPATCHER): New.
7818 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
7819 filling *region also set *region_idx to (*region)->entry->index.
7822 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
7823 For REGs set ORIGINAL_REGNO.
7825 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
7827 * doc/md.texi: Mention that a target shouldn't implement
7828 vec_widen_(s|u)mul_even/odd pair if it is less efficient
7831 2014-01-29 Jakub Jelinek <jakub@redhat.com>
7833 PR tree-optimization/59594
7834 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
7835 a copy of the datarefs vector rather than the vector itself.
7837 2014-01-28 Jason Merrill <jason@redhat.com>
7840 * dwarf2out.c (auto_die): New static.
7841 (gen_type_die_with_usage): Handle C++1y 'auto'.
7842 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
7845 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
7848 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
7849 (SPEC_X32): Likewise.
7850 (SPEC_64): Likewise.
7851 * config/i386/i386.c (ix86_option_override_internal): Turn off
7852 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
7854 (x86_file_start): Output .code16gcc for TARGET_16BIT.
7855 * config/i386/i386.h (TARGET_16BIT): New macro.
7856 (TARGET_16BIT_P): Likewise.
7857 * config/i386/i386.opt: Add m16.
7858 * doc/invoke.texi: Document -m16.
7860 2014-01-28 Jakub Jelinek <jakub@redhat.com>
7862 PR preprocessor/59935
7863 * input.c (location_get_source_line): Bail out on when line number
7864 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
7866 2014-01-28 Richard Biener <rguenther@suse.de>
7868 PR tree-optimization/58742
7869 * tree-ssa-forwprop.c (associate_plusminus): Handle
7870 pointer subtraction of the form (T)(P + A) - (T)P.
7872 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7874 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
7877 2014-01-28 Richard Biener <rguenther@suse.de>
7880 2014-01-28 Richard Biener <rguenther@suse.de>
7882 PR rtl-optimization/45364
7883 PR rtl-optimization/59890
7884 * var-tracking.c (local_get_addr_clear_given_value): Handle
7885 already cleared slot.
7886 (val_reset): Handle not allocated local_get_addr_cache.
7887 (vt_find_locations): Use post-order on the inverted CFG.
7889 2014-01-28 Richard Biener <rguenther@suse.de>
7891 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
7893 2014-01-28 Richard Biener <rguenther@suse.de>
7895 PR rtl-optimization/45364
7896 PR rtl-optimization/59890
7897 * var-tracking.c (local_get_addr_clear_given_value): Handle
7898 already cleared slot.
7899 (val_reset): Handle not allocated local_get_addr_cache.
7900 (vt_find_locations): Use post-order on the inverted CFG.
7902 2014-01-28 Alan Modra <amodra@gmail.com>
7904 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
7905 * configure.ac <recursive call for build != host>: Define
7906 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
7907 and LD_FOR_BUILD too.
7908 * configure: Regenerate.
7910 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
7912 * config/i386/i386.c (get_builtin_code_for_version): Separate
7913 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
7914 Broadwell from Haswell.
7916 2014-01-27 Steve Ellcey <sellcey@mips.com>
7918 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
7919 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
7920 * config/mips/mips.c (mips_option_override): Change setting
7922 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
7923 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
7924 Change from Mask to Var.
7926 2014-01-27 Jeff Law <law@redhat.com>
7928 * ipa-inline.c (inline_small_functions): Fix typo.
7930 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
7932 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
7933 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
7934 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
7935 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
7936 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
7937 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
7938 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
7939 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
7940 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
7941 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
7942 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
7943 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
7944 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
7945 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
7946 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
7947 (_mm512_storeu_epi64): Ditto.
7948 (_mm512_cmpge_epi32_mask): Ditto.
7949 (_mm512_cmpge_epu32_mask): Ditto.
7950 (_mm512_cmpge_epi64_mask): Ditto.
7951 (_mm512_cmpge_epu64_mask): Ditto.
7952 (_mm512_cmple_epi32_mask): Ditto.
7953 (_mm512_cmple_epu32_mask): Ditto.
7954 (_mm512_cmple_epi64_mask): Ditto.
7955 (_mm512_cmple_epu64_mask): Ditto.
7956 (_mm512_cmplt_epi32_mask): Ditto.
7957 (_mm512_cmplt_epu32_mask): Ditto.
7958 (_mm512_cmplt_epi64_mask): Ditto.
7959 (_mm512_cmplt_epu64_mask): Ditto.
7960 (_mm512_cmpneq_epi32_mask): Ditto.
7961 (_mm512_cmpneq_epu32_mask): Ditto.
7962 (_mm512_cmpneq_epi64_mask): Ditto.
7963 (_mm512_cmpneq_epu64_mask): Ditto.
7964 (_mm512_expand_pd): Ditto.
7965 (_mm512_expand_ps): Ditto.
7966 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
7967 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
7968 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
7969 * config/i386/i386.c (ix86_builtins): Add
7970 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
7971 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
7972 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
7973 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
7974 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
7975 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
7976 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
7977 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
7978 IX86_BUILTIN_PMOVUSQW512_MEM.
7979 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
7980 __builtin_ia32_pmovsqd512mem_mask,
7981 __builtin_ia32_pmovqd512mem_mask,
7982 __builtin_ia32_pmovusqw512mem_mask,
7983 __builtin_ia32_pmovsqw512mem_mask,
7984 __builtin_ia32_pmovqw512mem_mask,
7985 __builtin_ia32_pmovusdw512mem_mask,
7986 __builtin_ia32_pmovsdw512mem_mask,
7987 __builtin_ia32_pmovdw512mem_mask,
7988 __builtin_ia32_pmovqb512mem_mask,
7989 __builtin_ia32_pmovusqb512mem_mask,
7990 __builtin_ia32_pmovsqb512mem_mask,
7991 __builtin_ia32_pmovusdb512mem_mask,
7992 __builtin_ia32_pmovsdb512mem_mask,
7993 __builtin_ia32_pmovdb512mem_mask.
7994 (bdesc_args): Add __builtin_ia32_expanddf512,
7995 __builtin_ia32_expandsf512.
7996 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
7997 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
7998 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
7999 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
8000 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
8001 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
8002 (avx512f_<code>v8div16qi2_mask_store): This.
8003 (avx512f_expand<mode>): New.
8005 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
8007 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
8009 (_mm512_mask_prefetch_i64gather_pd): Ditto.
8010 (_mm512_prefetch_i32scatter_pd): Ditto.
8011 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
8012 (_mm512_prefetch_i64scatter_pd): Ditto.
8013 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
8014 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
8015 (_mm512_mask_prefetch_i64gather_ps): Ditto.
8016 (_mm512_prefetch_i32scatter_ps): Ditto.
8017 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
8018 (_mm512_prefetch_i64scatter_ps): Ditto.
8019 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
8020 * config/i386/i386-builtin-types.def: Define
8021 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
8022 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
8023 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
8024 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
8025 IX86_BUILTIN_SCATTERPFQPD.
8026 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
8027 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
8028 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
8029 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
8030 __builtin_ia32_scatterpfqps.
8031 (ix86_expand_builtin): Expand new built-ins.
8032 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8033 fix memory access data type.
8034 (*avx512pf_gatherpf<mode>_mask): Ditto.
8035 (*avx512pf_gatherpf<mode>): Ditto.
8036 (avx512pf_scatterpf<mode>): Ditto.
8037 (*avx512pf_scatterpf<mode>_mask): Ditto.
8038 (*avx512pf_scatterpf<mode>): Ditto.
8039 (GATHER_SCATTER_SF_MEM_MODE): New.
8040 (avx512pf_gatherpf<mode>df): Ditto.
8041 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8042 (*avx512pf_scatterpf<mode>df): Ditto.
8044 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8047 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8048 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8051 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8053 * common/config/arm/arm-common.c
8054 (arm_rewrite_mcpu): Handle multiple names.
8056 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8058 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8060 * gimple-builder.h (create_gimple_tmp): Delete.
8062 2014-01-27 Christian Bruel <christian.bruel@st.com>
8064 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
8067 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
8069 * config/pa/pa.md (call): Generate indirect long calls to non-local
8070 functions when outputing 32-bit code.
8071 (call_value): Likewise except for special call to buggy powf function.
8073 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
8074 portable runtime and PIC indirect calls.
8075 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
8076 and PIC call sequences. Use ldo instead of blr to set return register
8077 in PIC call sequence.
8079 2014-01-25 Walter Lee <walt@tilera.com>
8081 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
8082 avoid clobbering a live register.
8084 2014-01-25 Walter Lee <walt@tilera.com>
8086 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
8087 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
8088 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
8089 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
8091 2014-01-25 Walter Lee <walt@tilera.com>
8093 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
8094 arguments on even registers.
8095 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
8097 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
8098 (BIGGEST_ALIGNMENT): Ditto.
8099 (BIGGEST_FIELD_ALIGNMENT): Ditto.
8101 2014-01-25 Walter Lee <walt@tilera.com>
8103 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
8104 insns before bundling.
8105 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8107 2014-01-25 Walter Lee <walt@tilera.com>
8109 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
8110 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
8111 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
8113 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8115 * config/mips/constraints.md (kl): Delete.
8116 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
8117 define expands, using...
8118 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
8119 instructions for MIPS16.
8120 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
8121 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
8123 2014-01-25 Walter Lee <walt@tilera.com>
8125 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
8129 2014-01-25 Walter Lee <walt@tilera.com>
8131 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
8132 (TARGET_EXPAND_TO_RTL_HOOK): Define.
8134 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8136 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
8139 2014-01-25 Jakub Jelinek <jakub@redhat.com>
8141 * print-rtl.c (in_call_function_usage): New var.
8142 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
8143 EXPR_LIST mode as mode and not as reg note name.
8146 * cfgloopmanip.c (copy_loop_info): If
8147 loop->warned_aggressive_loop_optimizations, make sure
8148 the flag is set in target loop too.
8150 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
8152 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
8154 * builtins.def: Likewise.
8155 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
8156 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
8157 * ira.c (ira_setup_eliminable_regset): Likewise.
8158 * omp-low.c (gate_expand_omp): Likewise.
8159 (execute_lower_omp): Likewise.
8160 (diagnose_sb_0): Likewise.
8161 (gate_diagnose_omp_blocks): Likewise.
8162 (simd_clone_clauses_extract): Likewise.
8165 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8167 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
8168 correction for little endian...
8169 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
8172 2014-01-24 Jeff Law <law@redhat.com>
8174 PR tree-optimization/59919
8175 * tree-vrp.c (find_assert_locations_1): Do not register asserts
8176 for non-returning calls.
8178 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
8180 * common/config/aarch64/aarch64-common.c
8181 (aarch64_rewrite_mcpu): Handle multiple names.
8182 * config/aarch64/aarch64.h
8183 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8185 2014-01-24 Dodji Seketeli <dodji@redhat.com>
8187 * input.c (add_file_to_cache_tab): Handle the case where fopen
8190 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
8193 * config/i386/i386.md (pushsf splitter): Get stack adjustment
8194 from push operand if code of push isn't PRE_DEC.
8196 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8199 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
8200 -mquad-memory-atomic. Update -mquad-memory documentation to say
8201 it is only used for non-atomic loads/stores.
8203 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
8204 -mquad-memory or -mquad-memory-atomic switches.
8206 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
8207 -mquad-memory-atomic to ISA 2.07 support.
8209 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
8210 to separate support of normal quad word memory operations (ldq, stq)
8211 from the atomic quad word memory operations.
8213 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
8214 support to separate non-atomic quad word operations from atomic
8215 quad word operations. Disable non-atomic quad word operations in
8216 little endian mode so that we don't have to swap words after the
8217 load and before the store.
8218 (quad_load_store_p): Add comment about atomic quad word support.
8219 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
8220 options printed with -mdebug=reg.
8222 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
8223 -mquad-memory-atomic as the test for whether we have quad word
8224 atomic instructions.
8225 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
8226 or -mp8-vector are used, allow byte/half-word atomic operations.
8228 * config/rs6000/sync.md (load_lockedti): Insure that the address
8229 is a proper indexed or indirect address for the lqarx instruction.
8230 On little endian systems, swap the hi/lo registers after the lqarx
8232 (load_lockedpti): Use indexed_or_indirect_operand predicate to
8233 insure the address is valid for the lqarx instruction.
8234 (store_conditionalti): Insure that the address is a proper indexed
8235 or indirect address for the stqcrx. instruction. On little endian
8236 systems, swap the hi/lo registers before doing the stqcrx.
8238 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
8239 insure the address is valid for the stqcrx. instruction.
8241 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8242 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
8243 type of quad memory support is available.
8245 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
8248 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
8249 there is a danger of looping.
8251 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
8253 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8254 force flag_ira_loop_pressure if set via command line.
8256 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8258 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
8259 (ashr_simd): New builtin handling DI mode.
8260 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
8261 (aarch64_sshr_simddi): New match pattern.
8262 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
8263 (vshrd_n_s64): Likewise.
8264 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
8266 2014-01-23 Nick Clifton <nickc@redhat.com>
8268 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
8269 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
8270 favour of mcu specific scripts.
8271 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
8274 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8275 Alex Velenko <Alex.Velenko@arm.com>
8277 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
8278 (vaddv_s16): Likewise.
8279 (vaddv_s32): Likewise.
8280 (vaddv_u8): Likewise.
8281 (vaddv_u16): Likewise.
8282 (vaddv_u32): Likewise.
8283 (vaddvq_s8): Likewise.
8284 (vaddvq_s16): Likewise.
8285 (vaddvq_s32): Likewise.
8286 (vaddvq_s64): Likewise.
8287 (vaddvq_u8): Likewise.
8288 (vaddvq_u16): Likewise.
8289 (vaddvq_u32): Likewise.
8290 (vaddvq_u64): Likewise.
8291 (vaddv_f32): Likewise.
8292 (vaddvq_f32): Likewise.
8293 (vaddvq_f64): Likewise.
8294 (vmaxv_f32): Likewise.
8295 (vmaxv_s8): Likewise.
8296 (vmaxv_s16): Likewise.
8297 (vmaxv_s32): Likewise.
8298 (vmaxv_u8): Likewise.
8299 (vmaxv_u16): Likewise.
8300 (vmaxv_u32): Likewise.
8301 (vmaxvq_f32): Likewise.
8302 (vmaxvq_f64): Likewise.
8303 (vmaxvq_s8): Likewise.
8304 (vmaxvq_s16): Likewise.
8305 (vmaxvq_s32): Likewise.
8306 (vmaxvq_u8): Likewise.
8307 (vmaxvq_u16): Likewise.
8308 (vmaxvq_u32): Likewise.
8309 (vmaxnmv_f32): Likewise.
8310 (vmaxnmvq_f32): Likewise.
8311 (vmaxnmvq_f64): Likewise.
8312 (vminv_f32): Likewise.
8313 (vminv_s8): Likewise.
8314 (vminv_s16): Likewise.
8315 (vminv_s32): Likewise.
8316 (vminv_u8): Likewise.
8317 (vminv_u16): Likewise.
8318 (vminv_u32): Likewise.
8319 (vminvq_f32): Likewise.
8320 (vminvq_f64): Likewise.
8321 (vminvq_s8): Likewise.
8322 (vminvq_s16): Likewise.
8323 (vminvq_s32): Likewise.
8324 (vminvq_u8): Likewise.
8325 (vminvq_u16): Likewise.
8326 (vminvq_u32): Likewise.
8327 (vminnmv_f32): Likewise.
8328 (vminnmvq_f32): Likewise.
8329 (vminnmvq_f64): Likewise.
8331 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8333 * config/aarch64/aarch64-simd.md
8334 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
8335 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
8336 (*aarch64_mul3_elt<mode>): Likewise.
8337 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
8338 (*aarch64_mul3_elt_to_64v2df): Likewise.
8339 (*aarch64_mla_elt<mode>): Likewise.
8340 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8341 (*aarch64_mls_elt<mode>): Likewise.
8342 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8343 (*aarch64_fma4_elt<mode>): Likewise.
8344 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8345 (*aarch64_fma4_elt_to_64v2df): Likewise.
8346 (*aarch64_fnma4_elt<mode>): Likewise.
8347 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8348 (*aarch64_fnma4_elt_to_64v2df): Likewise.
8349 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
8350 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
8351 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8352 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8353 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
8354 (aarch64_sqdmull_lane<mode>_internal): Likewise.
8355 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
8357 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
8359 * config/aarch64/aarch64-simd.md
8360 (aarch64_be_checked_get_lane<mode>): New define_expand.
8361 * config/aarch64/aarch64-simd-builtins.def
8362 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
8363 New builtin definition.
8364 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
8365 Use new safe be builtin.
8367 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8369 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
8371 (aarch64_be_st1<mode>): Likewise.
8372 (aarch_ld1<VALL:mode>): Define_expand modified.
8373 (aarch_st1<VALL:mode>): Likewise.
8374 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
8375 (UNSPEC_ST1): Likewise.
8377 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
8379 * config/microblaze/microblaze.md: Add trap insn and attribute
8381 2014-01-23 Dodji Seketeli <dodji@redhat.com>
8383 PR preprocessor/58580
8384 * input.h (location_get_source_line): Take an additional line_size
8386 (void diagnostics_file_cache_fini): Declare new function.
8387 * input.c (struct fcache): New type.
8388 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
8389 New static constants.
8390 (diagnostic_file_cache_init, total_lines_num)
8391 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
8392 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
8393 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
8394 (get_next_line, read_next_line, goto_next_line, read_line_num):
8395 New static function definitions.
8396 (diagnostic_file_cache_fini): New function.
8397 (location_get_source_line): Take an additional output line_len
8398 parameter. Re-write using lookup_or_add_file_to_cache_tab and
8400 * diagnostic.c (diagnostic_finish): Call
8401 diagnostic_file_cache_fini.
8402 (adjust_line): Take an additional input parameter for the length
8403 of the line, rather than calculating it with strlen.
8404 (diagnostic_show_locus): Adjust the use of
8405 location_get_source_line and adjust_line with respect to their new
8406 signature. While displaying a line now, do not stop at the first
8407 null byte. Rather, display the zero byte as a space and keep
8408 going until we reach the size of the line.
8409 * Makefile.in: Add vec.o to OBJS-libcommon
8411 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8412 Ilya Tocar <ilya.tocar@intel.com>
8414 * config/i386/avx512fintrin.h (_mm512_kmov): New.
8415 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
8416 (__builtin_ia32_kmov16): Ditto.
8417 * config/i386/i386.md (UNSPEC_KMOV): New.
8420 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8422 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
8423 (_mm512_storeu_si512): Ditto.
8425 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
8428 * rtl.h (get_referenced_operands): Declare.
8429 * recog.c (get_referenced_operands): New function.
8430 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
8431 operands have been referenced when recording LO_SUM references.
8433 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
8435 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
8437 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8439 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
8440 Enable for generic and recent AMD targets.
8442 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8444 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
8445 ARG_SIZE note when adjustment was eliminated.
8447 2014-01-22 Jeff Law <law@redhat.com>
8449 PR tree-optimization/59597
8450 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
8451 in file. Accept new argument REGISTERING and use it to modify
8452 dump output appropriately.
8453 (register_jump_thread): Corresponding changes.
8454 (mark_threaded_blocks): Reinstate code to cancel unprofitable
8455 thread paths involving joiner blocks. Add code to dump cancelled
8456 jump threading paths.
8458 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
8460 PR rtl-optimization/59477
8461 * lra-constraints.c (inherit_in_ebb): Process call for living hard
8462 regs. Update reloads_num and potential_reload_hard_regs for all insns.
8464 2014-01-22 Tom Tromey <tromey@redhat.com>
8466 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
8468 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
8470 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8472 PR rtl-optimization/59896
8473 * lra-constraints.c (process_alt_operands): Check unused note for
8474 matched operands of insn with no output reloads.
8476 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
8478 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
8479 (mips_move_from_gpr_cost): Likewise.
8481 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8483 PR rtl-optimization/59858
8484 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
8485 ira_class_hard_regs_num.
8486 (process_alt_operands): Increase reject for dying matched operand.
8488 2014-01-21 Jakub Jelinek <jakub@redhat.com>
8491 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
8492 smaller than size, perform several stores or loads and stores
8493 at dst + count - size to store or copy all of size bytes, rather
8494 than just last modesize bytes.
8496 2014-01-20 DJ Delorie <dj@redhat.com>
8498 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
8499 that CLOBBERs are REGs before propogating their values.
8501 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
8504 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
8505 (cgraph_inline_failed_type): New function.
8506 * cgraph.h (DEFCIFCODE): Add type.
8507 (cgraph_inline_failed_type_t): New enum.
8508 (cgraph_inline_failed_type): New prototype.
8509 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
8510 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
8511 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
8512 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
8513 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
8514 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
8515 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
8516 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
8517 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
8518 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
8519 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
8520 OPTIMIZATION_MISMATCH.
8521 * tree-inline.c (expand_call_inline): Emit errors during
8522 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
8524 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8527 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
8528 mode attribute in insn output.
8530 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
8532 * output.h (output_constant): Delete.
8533 * varasm.c (output_constant): Make private.
8535 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
8537 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
8539 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8542 * tree.h (fold_builtin_strcat): New prototype.
8543 * builtins.c (fold_builtin_strcat): No longer static. Add len
8544 argument, if non-NULL, don't call c_strlen. Optimize
8545 directly into __builtin_memcpy instead of __builtin_strcpy.
8546 (fold_builtin_2): Adjust fold_builtin_strcat caller.
8547 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
8549 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8551 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8552 for SImode_address_operand operands, having only a REG argument.
8554 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
8556 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
8557 loader name using mbig-endian.
8558 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
8560 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8562 * doc/invoke.texi (-march): Clarify documentation for AArch64.
8566 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
8568 * config/aarch64/aarch64-protos.h
8569 (aarch64_cannot_change_mode_class_ptr): Declare.
8570 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
8571 aarch64_cannot_change_mode_class_ptr): New.
8572 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
8573 backend hook aarch64_cannot_change_mode_class.
8575 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8577 * common/config/aarch64/aarch64-common.c
8578 (aarch64_handle_option): Don't handle any option order logic here.
8579 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
8580 selected_cpu, warn on architecture version mismatch.
8581 (aarch64_override_options): Fix parsing order for option strings.
8583 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8584 Iain Sandoe <iain@codesourcery.com>
8587 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
8588 warning. Amend comment to reflect current functionality.
8590 2014-01-20 Richard Biener <rguenther@suse.de>
8593 * builtins.c (fold_builtin_strcat): Remove case better handled
8594 by tree-ssa-strlen.c.
8596 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
8598 * config/aarch64/aarch64.opt
8599 (mcpu, march, mtune): Make case-insensitive.
8601 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8604 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8605 if operands[1] is a REG or ZERO_EXTEND of a REG.
8607 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
8609 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
8611 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
8613 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
8614 long non-pic millicode calls.
8616 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8618 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
8620 2014-01-19 Kito Cheng <kito@0xlab.org>
8622 * builtins.c (expand_movstr): Check movstr expand done or fail.
8624 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8625 H.J. Lu <hongjiu.lu@intel.com>
8628 * config/i386/i386.md (*lea<mode>): Zero-extend return register
8629 to DImode for zero-extended addresses.
8631 2014-01-19 Jakub Jelinek <jakub@redhat.com>
8633 PR rtl-optimization/57763
8634 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
8635 on the new indirect jump_insn and increment LABEL_NUSES (label).
8637 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
8641 * config.gcc (x86_archs): New variable.
8642 (x86_64_archs): Likewise.
8643 (x86_cpus): Likewise.
8644 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
8645 --with-arch/--with-cpu= options.
8646 Support --with-arch=/--with-cpu={nehalem,westmere,
8647 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
8649 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8651 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
8652 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
8654 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8656 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
8658 2014-01-18 Jakub Jelinek <jakub@redhat.com>
8661 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
8662 clear cpp_get_options (parse_in)->warn_unused_macros for
8663 ix86_target_macros_internal with cpp_define.
8665 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
8667 * jump.c (delete_related_insns): Keep (use (insn))s.
8668 * reorg.c (redundant_insn): Check for barriers too.
8670 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8672 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
8674 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
8676 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
8677 call to $$dyncall when TARGET_LONG_CALLS is true.
8679 2014-01-17 Jeff Law <law@redhat.com>
8681 * ree.c (combine_set_extension): Temporarily disable test for
8682 changing number of hard registers.
8684 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
8687 * ipa-inline-analysis.c (inline_free_summary):
8688 Do not free summary of aliases.
8690 2014-01-17 Jakub Jelinek <jakub@redhat.com>
8693 * gimplify.c (gimplify_expr): Use create_tmp_var
8694 instead of create_tmp_var_raw. If cond doesn't have
8695 integral type, don't add the IFN_ANNOTATE builtin at all.
8697 2014-01-17 Martin Jambor <mjambor@suse.cz>
8700 * ipa-cp.c (prev_edge_clone): New variable.
8701 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
8702 Also resize prev_edge_clone vector.
8703 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
8704 (ipcp_edge_removal_hook): New function.
8705 (ipcp_driver): Register ipcp_edge_removal_hook.
8707 2014-01-17 Andrew Pinski <apinski@cavium.com>
8708 Steve Ellcey <sellcey@mips.com>
8711 * config/mips/mips.c (mips_print_operand): Check operand mode instead
8714 2014-01-17 Jeff Law <law@redhat.com>
8717 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
8718 so that pass_ccp runs first.
8720 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8722 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
8723 (ix86_adjust_cost): Use !TARGET_XXX.
8724 (do_reorder_for_imul): Likewise.
8725 (swap_top_of_ready_list): Likewise.
8726 (ix86_sched_reorder): Likewise.
8728 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8730 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
8731 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
8732 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
8733 (intel_memset): New. Duplicate slm_memset.
8734 (intel_cost): New. Duplicate slm_cost.
8735 (m_INTEL): New macro.
8736 (processor_target_table): Add "intel".
8737 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
8738 with PROCESSOR_INTEL for "intel".
8739 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
8740 PROCESSOR_SILVERMONT.
8741 (ix86_issue_rate): Likewise.
8742 (ix86_adjust_cost): Likewise.
8743 (ia32_multipass_dfa_lookahead): Likewise.
8744 (swap_top_of_ready_list): Likewise.
8745 (ix86_sched_reorder): Likewise.
8746 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
8747 instead of TARGET_OPT_AGU.
8748 * config/i386/i386.h (TARGET_INTEL): New.
8749 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
8750 (processor_type): Add PROCESSOR_INTEL.
8751 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
8752 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
8754 2014-01-17 Marek Polacek <polacek@redhat.com>
8757 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
8760 2014-01-17 Richard Biener <rguenther@suse.de>
8762 PR tree-optimization/46590
8763 * opts.c (default_options_table): Add entries for
8764 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
8765 all enabled at -O1 but not for -Og.
8766 * common.opt (fbranch-count-reg): Remove Init(1).
8767 (fmove-loop-invariants): Likewise.
8768 (ftree-pta): Likewise.
8770 2014-01-17 Jakub Jelinek <jakub@redhat.com>
8772 * config/i386/i386.c (ix86_data_alignment): For compatibility with
8773 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
8774 decls to at least the GCC 4.8 used alignments.
8777 * tree-nested.c (convert_nonlocal_reference_stmt,
8778 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
8779 of GIMPLE_BIND stmts, adjust associated decls.
8781 2014-01-17 Richard Biener <rguenther@suse.de>
8783 PR tree-optimization/46590
8784 * vec.h (vec<>::bseach): New member function implementing
8785 binary search according to C89 bsearch.
8786 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
8787 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
8788 bitmap pointer again. Make accesses_in_loop a flat array.
8789 (mem_ref_obstack): New global.
8790 (outermost_indep_loop): Adjust for mem_ref->stored changes.
8791 (mark_ref_stored): Likewise.
8792 (ref_indep_loop_p_2): Likewise.
8793 (set_ref_stored_in_loop): New helper function.
8794 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
8795 (memref_free): Adjust.
8796 (record_mem_ref_loc): Simplify.
8797 (gather_mem_refs_stmt): Adjust.
8798 (sort_locs_in_loop_postorder_cmp): New function.
8799 (analyze_memory_references): Sort accesses_in_loop after
8800 loop postorder number.
8801 (find_ref_loc_in_loop_cmp): New function.
8802 (for_all_locs_in_loop): Find relevant cluster of locs in
8803 accesses_in_loop and iterate without recursion.
8804 (execute_sm): Avoid uninit warning.
8805 (struct ref_always_accessed): Simplify.
8806 (ref_always_accessed::operator ()): Likewise.
8807 (ref_always_accessed_p): Likewise.
8808 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
8809 loop postorder numbers here.
8810 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
8813 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
8816 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
8817 on decls for which assemble_alias has been called.
8819 2014-01-17 Nick Clifton <nickc@redhat.com>
8821 * config/msp430/msp430.opt: (mcpu): New option.
8822 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
8823 (msp430_option_override): Parse target_cpu. If the MCU name
8824 matches a generic string, clear target_mcu.
8825 (msp430_attr): Allow numeric interrupt values up to 63.
8826 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
8827 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
8829 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
8831 * config/msp430/msp430.md (popm): Use %J rather than %I.
8832 (addsi3): Use msp430_nonimmediate_operand for operand 2.
8833 (addhi_cy_i): Use immediate_operand for operand 2.
8834 * doc/invoke.texi: Document -mcpu option.
8836 2014-01-17 Richard Biener <rguenther@suse.de>
8838 PR rtl-optimization/38518
8839 * df.h (df_analyze_loop): Declare.
8840 * df-core.c: Include cfgloop.h.
8841 (df_analyze_1): Split out main part of df_analyze.
8842 (df_analyze): Adjust.
8843 (loop_inverted_post_order_compute): New function.
8844 (loop_post_order_compute): Likewise.
8845 (df_analyze_loop): New function avoiding whole-function
8847 * loop-invariant.c (find_defs): Use df_analyze_loop.
8848 (find_invariants): Adjust.
8849 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
8851 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
8853 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
8854 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
8856 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
8858 * ipa-ref.c (ipa_remove_stmt_references): Fix references
8859 traversal when removing references.
8861 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
8864 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
8866 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
8869 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
8870 pushing a reload for an autoinc when we had previously reloaded an
8871 inner part of the address.
8873 2014-01-16 Jakub Jelinek <jakub@redhat.com>
8875 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
8877 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
8878 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
8879 when not giving up or versioning for alias only because of
8881 (vect_analyze_data_ref_dependences): Set to true.
8882 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
8884 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
8885 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
8889 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
8892 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
8893 operand 0 predicate for gathers, use a new pseudo as subtarget.
8895 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
8898 * lra-constraints.c (process_alt_operands): Add printing debug info.
8899 Check absence of input/output reloads for matched operands too.
8901 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
8903 PR rtl-optimization/59835
8904 * ira.c (ira_init_register_move_cost): Increase cost for
8907 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
8909 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
8911 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
8914 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
8915 non-register objects. Use gen_(high/low)part more consistently.
8918 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8921 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
8922 endian support, remove tests for WORDS_BIG_ENDIAN.
8923 (p8_mfvsrd_3_<mode>): Likewise.
8924 (reload_gpr_from_vsx<mode>): Likewise.
8925 (reload_gpr_from_vsxsf): Likewise.
8926 (p8_mfvsrd_4_disf): Likewise.
8928 2014-01-16 Richard Biener <rguenther@suse.de>
8930 PR rtl-optimization/46590
8931 * lcm.c (compute_antinout_edge): Use postorder iteration.
8932 (compute_laterin): Use inverted postorder iteration.
8934 2014-01-16 Nick Clifton <nickc@redhat.com>
8937 * varasm.c (output_constant): Return the number of bytes actually
8939 (output_constructor_array_range): Update the field size with the
8940 number of bytes emitted by output_constant.
8941 (output_constructor_regular_field): Likewise. Also do not
8942 complain if the total number of bytes emitted is now greater
8943 than the expected fieldpos.
8944 * output.h (output_constant): Update prototype and descriptive comment.
8946 2014-01-16 Marek Polacek <polacek@redhat.com>
8949 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
8950 it is error_mark_node.
8952 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
8954 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
8955 VALID_AVX256_REG_OR_OI_MODE.
8957 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
8959 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
8960 current procedure should be profiled.
8962 2014-01-15 Andrew Pinski <apinski@cavium.com>
8964 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
8965 of moving from/to the STACK_REG register class.
8967 2014-01-15 Richard Henderson <rth@redhat.com>
8970 * reginfo.c (global_regs_decl): Globalize.
8971 * rtl.h (global_regs_decl): Declare.
8972 * ira.c (do_reload): Diagnose frame_pointer_needed and it
8973 reserved via global_regs.
8975 2014-01-15 Teresa Johnson <tejohnson@google.com>
8977 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
8979 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8981 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
8982 and vmulosh rather than call gen_vec_widen_smult_*.
8983 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
8984 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
8985 (vec_widen_smult_even_v16qi): Likewise.
8986 (vec_widen_umult_even_v8hi): Likewise.
8987 (vec_widen_smult_even_v8hi): Likewise.
8988 (vec_widen_umult_odd_v16qi): Likewise.
8989 (vec_widen_smult_odd_v16qi): Likewise.
8990 (vec_widen_umult_odd_v8hi): Likewise.
8991 (vec_widen_smult_odd_v8hi): Likewise.
8992 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
8993 vmuloub rather than call gen_vec_widen_umult_*.
8994 (vec_widen_umult_lo_v16qi): Likewise.
8995 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
8996 vmulosb rather than call gen_vec_widen_smult_*.
8997 (vec_widen_smult_lo_v16qi): Likewise.
8998 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
8999 rather than call gen_vec_widen_umult_*.
9000 (vec_widen_umult_lo_v8hi): Likewise.
9001 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
9002 rather than call gen_vec_widen_smult_*.
9003 (vec_widen_smult_lo_v8hi): Likewise.
9005 2014-01-15 Jeff Law <law@redhat.com>
9007 PR tree-optimization/59747
9008 * ree.c (find_and_remove_re): Properly handle case where a second
9009 eliminated extension requires widening a copy created for elimination
9010 of a prior extension.
9011 (combine_set_extension): Ensure that the number of hard regs needed
9012 for a destination register does not change when we widen it.
9014 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
9016 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
9017 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
9018 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
9019 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
9020 (avr-*-rtems*): Likewise.
9021 (bfin*-rtems*): Likewise.
9022 (moxie-*-rtems*): Likewise.
9023 (h8300-*-rtems*): Likewise.
9024 (i[34567]86-*-rtems*): Likewise.
9025 (lm32-*-rtems*): Likewise.
9026 (m32r-*-rtems*): Likewise.
9027 (m68k-*-rtems*): Likewise.
9028 (microblaze*-*-rtems*): Likewise.
9029 (mips*-*-rtems*): Likewise.
9030 (powerpc-*-rtems*): Likewise.
9031 (sh-*-rtems*): Likewise.
9032 (sparc-*-rtems*): Likewise.
9033 (sparc64-*-rtems*): Likewise.
9034 (v850-*-rtems*): Likewise.
9035 (m32c-*-rtems*): Likewise.
9037 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9039 PR rtl-optimization/59511
9040 * ira.c (ira_init_register_move_cost): Use memory costs for some
9041 cases of register move cost calculations.
9042 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9043 instead of BB frequency.
9044 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9045 * lra-assigns.c (find_hard_regno_for): Ditto.
9047 2014-01-15 Richard Biener <rguenther@suse.de>
9049 PR tree-optimization/59822
9050 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9051 (vectorizable_load): Use it to hoist defs of uses of invariant
9052 loads out of the loop.
9054 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9055 Kugan Vivekanandarajah <kuganv@linaro.org>
9058 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9061 2014-01-15 Richard Biener <rguenther@suse.de>
9063 PR rtl-optimization/59802
9064 * lcm.c (compute_available): Use inverted postorder to seed
9065 the initial worklist.
9067 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9070 * config/s390/s390.c (s390_preferred_reload_class): Don't return
9071 ADDR_REGS for invalid symrefs in non-PIC code.
9073 2014-01-15 Jakub Jelinek <jakub@redhat.com>
9076 * builtins.c (determine_block_size): Initialize *probable_max_size
9077 even if len_rtx is CONST_INT.
9079 2014-01-14 Andrew Pinski <apinski@cavium.com>
9081 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
9082 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
9083 (cortexa53_tunings): Likewise.
9084 (aarch64_sched_issue_rate): New function.
9085 (TARGET_SCHED_ISSUE_RATE): Define.
9087 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9089 * ira-costs.c (find_costs_and_classes): Add missed
9090 ira_init_register_move_cost_if_necessary.
9092 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9095 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
9097 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
9100 * config/i386/i386.c (type_natural_mode): Add a bool parameter
9101 to indicate if type is used for function return value. Warn ABI
9102 change if the vector mode isn't available for function return value.
9103 (ix86_function_arg_advance): Pass false to type_natural_mode.
9104 (ix86_function_arg): Likewise.
9105 (ix86_gimplify_va_arg): Likewise.
9106 (function_arg_32): Don't warn ABI change.
9107 (ix86_function_value): Pass true to type_natural_mode.
9108 (ix86_return_in_memory): Likewise.
9109 (ix86_struct_value_rtx): Removed.
9110 (TARGET_STRUCT_VALUE_RTX): Likewise.
9112 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9114 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
9115 converting a conditional jump into a conditional return.
9117 2014-01-14 Richard Biener <rguenther@suse.de>
9119 PR tree-optimization/58921
9120 PR tree-optimization/59006
9121 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
9122 hoisting invariant stmts.
9123 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
9124 invariant loads on the preheader edge if possible.
9126 2014-01-14 Joey Ye <joey.ye@arm.com>
9128 * doc/plugin.texi (Building GCC plugins): Update to C++.
9130 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
9132 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
9133 (_mm_rcp28_round_ss): Ditto.
9134 (_mm_rsqrt28_round_sd): Ditto.
9135 (_mm_rsqrt28_round_ss): Ditto.
9136 (_mm_rcp28_sd): Ditto.
9137 (_mm_rcp28_ss): Ditto.
9138 (_mm_rsqrt28_sd): Ditto.
9139 (_mm_rsqrt28_ss): Ditto.
9140 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
9141 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
9142 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
9143 (IX86_BUILTIN_RCP28SD): Ditto.
9144 (IX86_BUILTIN_RCP28SS): Ditto.
9145 (IX86_BUILTIN_RSQRT28SD): Ditto.
9146 (IX86_BUILTIN_RSQRT28SS): Ditto.
9147 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
9148 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
9149 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
9150 (ix86_expand_special_args_builtin): Expand new FTYPE.
9151 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
9152 (srcp14<mode>): Make insn unary.
9153 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
9154 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
9155 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
9156 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
9157 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
9158 Fix rounding: make it SAE only.
9159 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9161 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9163 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
9164 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
9165 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
9166 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
9167 (round_saeonly_mask_scalar_operand4): Ditto.
9168 (round_saeonly_mask_scalar_op3): Ditto.
9169 (round_saeonly_mask_scalar_op4): Ditto.
9171 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9173 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9174 Implement -maltivec=be for vec_insert and vec_extract.
9176 2014-01-10 DJ Delorie <dj@redhat.com>
9178 * config/msp430/msp430.md (call_internal): Don't allow memory
9179 references with SP as the base register.
9180 (call_value_internal): Likewise.
9181 * config/msp430/constraints.md (Yc): New. For memory references
9182 that don't use SP as a base register.
9184 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
9185 "an integer without a # prefix"
9186 * config/msp430/msp430.md (epilogue_helper): Use it.
9188 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9191 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
9192 AVX512F gather builtins.
9193 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
9194 on gather decls with INTEGER_TYPE masktype.
9195 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
9196 directly into the builtin rather than hoisting it before loop.
9198 PR tree-optimization/59387
9199 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
9200 (scev_const_prop): If folded_casts and type has undefined overflow,
9201 use force_gimple_operand instead of force_gimple_operand_gsi and
9202 for each added stmt if it is assign with
9203 arith_code_with_undefined_signed_overflow, call
9204 rewrite_to_defined_overflow.
9205 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
9206 gimple-fold.h instead.
9207 (arith_code_with_undefined_signed_overflow,
9208 rewrite_to_defined_overflow): Moved to ...
9209 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
9210 rewrite_to_defined_overflow): ... here. No longer static.
9211 Include gimplify-me.h.
9212 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
9213 rewrite_to_defined_overflow): New prototypes.
9215 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9217 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
9219 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
9221 * builtins.c (get_object_alignment_2): Minor tweak.
9222 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
9224 2014-01-13 Christian Bruel <christian.bruel@st.com>
9226 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
9227 optimized non constant lengths.
9229 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9232 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
9233 load as __atomic_load_N if possible.
9235 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9237 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
9239 (rs6000_expand_builtin): Adjust call.
9241 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9244 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
9245 * config/rs6000/rs6000.c: Include target-globals.h.
9246 (rs6000_set_current_function): Instead of doing target_reinit
9247 unconditionally, use save_target_globals_default_opts and
9248 restore_target_globals.
9250 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
9252 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
9253 (rs6000_expand_builtin): Handle mffs and mtfsf.
9254 (rs6000_init_builtins): Define mffs and mtfsf.
9255 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
9256 (rs6000_mffs): New pattern.
9257 (rs6000_mtfsf): New pattern.
9259 2014-01-11 Bin Cheng <bin.cheng@arm.com>
9261 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
9262 Start narrowing with START. Apply candidate-use pair
9263 and check overall cost in narrowing.
9264 (iv_ca_prune): Pass new argument.
9266 2014-01-10 Jeff Law <law@redhat.com>
9269 * ree.c (combine_reaching_defs): Ensure the defining statement
9270 occurs before the extension when optimizing extensions with
9271 different source and destination hard registers.
9273 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9276 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
9277 vtables into the type inheritance graph.
9279 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9281 PR rtl-optimization/59754
9282 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
9283 modes in the REGNO != REGNO case.
9285 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9287 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
9289 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9291 PR tree-optimization/59745
9292 * tree-predcom.c (tree_predictive_commoning_loop): Call
9293 free_affine_expand_cache if giving up because components is NULL.
9295 * target-globals.c (save_target_globals): Allocate < 4KB structs using
9296 GC in payload of target_globals struct instead of allocating them on
9297 the heap and the larger structs separately using GC.
9298 * target-globals.h (struct target_globals): Make regs, hard_regs,
9299 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
9300 of GTY((skip)) and change type to void *.
9301 (reset_target_globals): Cast loads from those fields to corresponding
9304 2014-01-10 Steve Ellcey <sellcey@mips.com>
9307 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
9308 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
9309 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
9311 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
9314 * aarch64-modes.def (CC_Zmode): New flags mode.
9315 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
9316 represents an equality.
9317 (aarch64_get_condition_code): Handle CC_Zmode.
9318 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
9320 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9322 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
9323 extraction in good case.
9325 2014-01-10 Richard Biener <rguenther@suse.de>
9327 PR tree-optimization/59374
9328 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
9329 checking after SLP discovery. Mark stmts not participating
9330 in any SLP instance properly.
9332 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9334 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
9335 when handling a SET rtx.
9337 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9339 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
9340 (cortex-a57): Likewise.
9341 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
9343 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9345 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
9346 non-iwmmxt builtins.
9348 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9352 * ipa-devirt.c record_target_from_binfo): Take as argument
9353 stack of binfos and lookup matching one for virtual inheritance.
9354 (possible_polymorphic_call_targets_1): Update.
9356 2014-01-10 Huacai Chen <chenhc@lemote.com>
9358 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
9359 kernel strings for Loongson-2E/2F/3A.
9361 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9364 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
9365 is_gimple_call before calling gimple_call_internal_p.
9367 2014-01-09 Steve Ellcey <sellcey@mips.com>
9369 * Makefile.in (TREE_FLOW_H): Remove.
9370 (TREE_SSA_H): Add file names from tree-flow.h.
9371 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
9372 * tree.h: Remove tree-flow.h reference.
9373 * hash-table.h: Remove tree-flow.h reference.
9374 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
9375 reference with tree-ssa-loop.h.
9377 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9379 * doc/invoke.texi: Add -maltivec={be,le} options, and document
9380 default element-order behavior for -maltivec.
9381 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
9382 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
9383 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
9384 when targeting big endian, at least for now.
9385 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
9387 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9390 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
9391 var satisfies use_register_for_decl, just take into account type
9392 alignment, rather than decl alignment.
9394 PR tree-optimization/59622
9395 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
9396 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
9397 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
9398 Don't devirtualize for inplace at all. For targets.length () == 1,
9399 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
9401 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9403 * config/i386/i386.md (cpu): Remove the unused btver1.
9405 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9407 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
9409 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9412 * tree-core.h (struct target_globals): New forward declaration.
9413 (struct tree_target_option): Add globals field.
9414 * tree.h (TREE_TARGET_GLOBALS): Define.
9415 (prepare_target_option_nodes_for_pch): New prototype.
9416 * target-globals.h (struct target_globals): Define even if
9418 * tree.c (prepare_target_option_node_for_pch,
9419 prepare_target_option_nodes_for_pch): New functions.
9420 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
9421 * config/i386/i386.c: Include target-globals.h.
9422 (ix86_set_current_function): Instead of doing target_reinit
9423 unconditionally, use save_target_globals_default_opts and
9424 restore_target_globals.
9426 2014-01-09 Richard Biener <rguenther@suse.de>
9428 PR tree-optimization/59715
9429 * tree-cfg.h (split_critical_edges): Declare.
9430 * tree-cfg.c (split_critical_edges): Export.
9431 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
9433 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
9435 * cfgexpand.c (expand_stack_vars): Optionally disable
9436 asan stack protection.
9437 (expand_used_vars): Likewise.
9438 (partition_stack_vars): Likewise.
9439 * asan.c (asan_emit_stack_protection): Optionally disable
9440 after return stack usage.
9441 (instrument_derefs): Optionally disable memory access instrumentation.
9442 (instrument_builtin_call): Likewise.
9443 (instrument_strlen_call): Likewise.
9444 (asan_protect_global): Optionally disable global variables protection.
9445 * doc/invoke.texi: Added doc for new options.
9446 * params.def: Added new options.
9447 * params.h: Likewise.
9449 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9451 PR rtl-optimization/59724
9452 * ifcvt.c (cond_exec_process_if_block): Don't call
9453 flow_find_head_matching_sequence with 0 longest_match.
9454 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
9455 non-active insns if !stop_after.
9456 (try_head_merge_bb): Revert 2014-01-07 changes.
9458 2014-01-08 Jeff Law <law@redhat.com>
9460 * ree.c (get_sub_rtx): New function, extracted from...
9461 (merge_def_and_ext): Here.
9462 (combine_reaching_defs): Use get_sub_rtx.
9464 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
9466 * cgraph.h (varpool_variable_node): Do not choke on null node.
9468 2014-01-08 Catherine Moore <clm@codesourcery.com>
9470 * config/mips/mips.md (simple_return): Attempt to use JRC
9472 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
9474 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9476 PR rtl-optimization/59137
9477 * reorg.c (steal_delay_list_from_target): Call update_block for
9479 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
9481 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9483 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
9484 two duplicate entries.
9486 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9489 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
9491 * config/mips/mips.c (mips_truncated_op_cost): New function.
9492 (mips_rtx_costs): Adjust test for BADDU.
9493 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
9495 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
9497 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
9498 (*baddu_si): ...this new pattern.
9500 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9503 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
9505 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9508 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
9510 (expand_expr, expand_normal): Adjust.
9511 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
9512 inner_reference_p. Use inner_reference_p to expand inner references.
9513 (store_expr): Adjust.
9514 * cfgexpand.c (expand_call_stmt): Adjust.
9516 2014-01-08 Rong Xu <xur@google.com>
9518 * gcov-io.c (gcov_var): Move from gcov-io.h.
9519 (gcov_position): Ditto.
9520 (gcov_is_error): Ditto.
9521 (gcov_rewrite): Ditto.
9522 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
9523 only part to libgcc/libgcov.h.
9525 2014-01-08 Marek Polacek <polacek@redhat.com>
9528 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
9530 2014-01-08 Marek Polacek <polacek@redhat.com>
9533 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
9535 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9537 PR rtl-optimization/59649
9538 * stor-layout.c (get_mode_bounds): For BImode return
9539 0 and STORE_FLAG_VALUE.
9541 2014-01-08 Richard Biener <rguenther@suse.de>
9544 * gimple.h (is_gimple_builtin_call): Remove.
9545 (gimple_builtin_call_types_compatible_p): New.
9546 (gimple_call_builtin_p): New overload.
9547 * gimple.c (is_gimple_builtin_call): Remove.
9548 (validate_call): Rename to ...
9549 (gimple_builtin_call_types_compatible_p): ... this and export. Also
9551 (validate_type): New static function.
9552 (gimple_call_builtin_p): New overload and adjust.
9553 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
9554 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
9555 (gimple_fold_stmt_to_constant_1): Likewise.
9556 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
9558 2014-01-08 Richard Biener <rguenther@suse.de>
9561 * gimplify.c (gimplify_expr): Gimplify register-register type
9562 VIEW_CONVERT_EXPRs to separate stmts.
9564 2014-01-07 Jeff Law <law@redhat.com>
9567 * ree.c (combine_set_extension): Handle case where source
9568 and destination registers in an extension insn are different.
9569 (combine_reaching_defs): Allow source and destination registers
9570 in extension to be different under limited circumstances.
9571 (add_removable_extension): Remove restriction that the
9572 source and destination registers in the extension are the same.
9573 (find_and_remove_re): Emit a copy from the extension's
9574 destination to its source after the defining insn if
9575 the source and destination registers are different.
9578 * ifcvt.c (merge_if_block): If we are merging a block with more than
9579 one successor with a block with no successors, remove any BARRIER
9580 after the second block.
9582 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
9584 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
9586 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
9589 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
9590 for 14-bit register offsets when INT14_OK_STRICT is false.
9592 2014-01-07 Roland Stigge <stigge@antcom.de>
9593 Michael Meissner <meissner@linux.vnet.ibm.com>
9596 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
9597 Only check TFmode for SPE constants. Don't check TImode or TDmode.
9599 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
9601 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
9604 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
9606 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
9607 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
9608 rtx is const0_rtx or not.
9610 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
9613 * target-globals.c (save_target_globals): Remove this_fn_optab
9615 * toplev.c: Include optabs.h.
9616 (target_reinit): Temporarily restore the global options if another
9617 set of options are in force.
9619 2014-01-07 Jakub Jelinek <jakub@redhat.com>
9621 PR rtl-optimization/58668
9622 * cfgcleanup.c (flow_find_cross_jump): Don't count
9623 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
9624 to determine what is counted.
9625 (flow_find_head_matching_sequence): Use active_insn_p to determine
9627 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
9629 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
9630 determine what is counted.
9632 PR tree-optimization/59643
9633 * tree-predcom.c (split_data_refs_to_components): If one dr is
9634 read and one write, determine_offset fails and the write isn't
9635 in the bad component, just put the read into the bad component.
9637 2014-01-07 Mike Stump <mikestump@comcast.net>
9638 Jakub Jelinek <jakub@redhat.com>
9641 * tree-core.h (struct tree_optimization_option): Change optabs
9642 type from unsigned char * to void *.
9643 * optabs.c (init_tree_optimization_optabs): Adjust
9644 TREE_OPTIMIZATION_OPTABS initialization.
9646 2014-01-06 Jakub Jelinek <jakub@redhat.com>
9649 * config/i386/i386.h (struct machine_function): Add
9650 no_drap_save_restore field.
9651 * config/i386/i386.c (ix86_save_reg): Use
9652 !cfun->machine->no_drap_save_restore instead of
9653 crtl->stack_realign_needed.
9654 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
9655 this function clears frame_pointer_needed. Set
9656 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
9657 and DRAP reg is needed.
9659 2014-01-06 Marek Polacek <polacek@redhat.com>
9662 * doc/implement-c.texi: Mention that other integer types are
9663 permitted as bit-field types in strictly conforming mode.
9665 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
9667 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
9670 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
9672 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
9674 2014-01-06 Martin Jambor <mjambor@suse.cz>
9677 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
9679 * ipa-prop.c (ipa_print_node_params): Fix indentation.
9681 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
9685 * var-tracking.c (add_stores): Preserve the value of the source even if
9686 we don't record the store.
9688 2014-01-06 Terry Guo <terry.guo@arm.com>
9690 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
9692 2014-01-05 Iain Sandoe <iain@codesourcery.com>
9695 * config/darwin.c (darwin_function_section): Adjust return values to
9696 correspond to optimisation changes made in r206070.
9698 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
9700 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
9701 from prefetch_block tune setting.
9702 (nocona_cost): Correct size of prefetch block to 64.
9704 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
9706 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
9707 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
9708 used to save the static chain register in the computation of the offset
9709 from which the FP registers need to be restored.
9711 2014-01-04 Jakub Jelinek <jakub@redhat.com>
9713 PR tree-optimization/59519
9714 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
9715 ICE if get_current_def (current_new_name) is already non-NULL, as long
9716 as it is a phi result of some other phi in *new_exit_bb that has
9719 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
9720 or vmovdqu* for misaligned_operand.
9721 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
9722 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
9723 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
9724 aligned_mem for AVX512F masked aligned load and store builtins and for
9727 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
9729 PR tree-optimization/59651
9730 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
9731 Address range for negative step should be added by TYPE_SIZE_UNIT.
9733 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
9735 * config/m68k/m68k.c (handle_move_double): Handle pushes with
9736 overlapping registers also for registers other than the stack pointer.
9738 2014-01-03 Marek Polacek <polacek@redhat.com>
9741 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
9744 2014-01-03 Jakub Jelinek <jakub@redhat.com>
9747 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
9750 * config/i386/i386.md (MODE_SIZE): New mode attribute.
9751 (push splitter): Use <P:MODE_SIZE> instead of
9752 GET_MODE_SIZE (<P:MODE>mode).
9753 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
9754 (mov -1, reg peephole2): Likewise.
9755 * config/i386/sse.md (*mov<mode>_internal,
9756 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
9757 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
9758 *<code><mode>3, *andnot<mode>3<mask_name>,
9759 <mask_codefor><code><mode>3<mask_name>): Likewise.
9760 * config/i386/subst.md (mask_mode512bit_condition,
9761 sd_mask_mode512bit_condition): Likewise.
9763 2014-01-02 Xinliang David Li <davidxl@google.com>
9765 PR tree-optimization/59303
9766 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
9767 (dump_predicates): Better output format.
9768 (pred_equal_p): New function.
9769 (is_neq_relop_p): Ditto.
9770 (is_neq_zero_form_p): Ditto.
9771 (pred_expr_equal_p): Ditto.
9772 (pred_neg_p): Ditto.
9773 (simplify_pred): Ditto.
9774 (simplify_preds_2): Ditto.
9775 (simplify_preds_3): Ditto.
9776 (simplify_preds_4): Ditto.
9777 (simplify_preds): Ditto.
9779 (push_to_worklist): Ditto.
9780 (get_pred_info_from_cmp): Ditto.
9781 (is_degenerated_phi): Ditto.
9782 (normalize_one_pred_1): Ditto.
9783 (normalize_one_pred): Ditto.
9784 (normalize_one_pred_chain): Ditto.
9785 (normalize_preds): Ditto.
9786 (normalize_cond_1): Remove function.
9787 (normalize_cond): Ditto.
9788 (is_gcond_subset_of): Ditto.
9789 (is_subset_of_any): Ditto.
9790 (is_or_set_subset_of): Ditto.
9791 (is_and_set_subset_of): Ditto.
9792 (is_norm_cond_subset_of): Ditto.
9793 (pred_chain_length_cmp): Ditto.
9794 (convert_control_dep_chain_into_preds): Type change.
9795 (find_predicates): Ditto.
9796 (find_def_preds): Ditto.
9797 (destroy_predicates_vecs): Ditto.
9798 (find_matching_predicates_in_rest_chains): Ditto.
9799 (use_pred_not_overlap_with_undef_path_pred): Ditto.
9800 (is_pred_expr_subset): Ditto.
9801 (is_pred_chain_subset_of): Ditto.
9802 (is_included_in): Ditto.
9803 (is_superset_of): Ditto.
9805 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
9807 Update copyright years.
9809 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
9811 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
9812 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
9813 config/arc/arc.md, config/arc/arc.opt,
9814 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
9815 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
9816 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
9817 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
9818 config/linux-protos.h, config/linux.c, config/winnt-c.c,
9819 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
9820 vtable-verify.c, vtable-verify.h: Use the standard form for the
9823 2014-01-02 Tobias Burnus <burnus@net-b.de>
9825 * gcc.c (process_command): Update copyright notice dates.
9826 * gcov-dump.c: Ditto.
9828 * doc/cpp.texi: Bump @copying's copyright year.
9829 * doc/cppinternals.texi: Ditto.
9830 * doc/gcc.texi: Ditto.
9831 * doc/gccint.texi: Ditto.
9832 * doc/gcov.texi: Ditto.
9833 * doc/install.texi: Ditto.
9834 * doc/invoke.texi: Ditto.
9836 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9838 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
9840 2014-01-01 Jakub Jelinek <jakub@redhat.com>
9842 * config/i386/sse.md (*mov<mode>_internal): Guard
9843 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
9845 PR rtl-optimization/59647
9846 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
9847 new_rtx into UNSIGNED_FLOAT rtxes.
9849 Copyright (C) 2014 Free Software Foundation, Inc.
9851 Copying and distribution of this file, with or without modification,
9852 are permitted in any medium without royalty provided the copyright
9853 notice and this notice are preserved.