1 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
4 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
6 * lra-constraints.c (valid_address_p): Move earlier in file.
7 Add a constraint argument to the address_info version.
8 (satisfies_memory_constraint_p): New function.
9 (satisfies_address_constraint_p): Likewise.
10 (process_alt_operands, curr_insn_transform): Use them.
11 (process_address): Pass the constraint to valid_address_p when
12 checking address operands.
14 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
16 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
17 to their respective blocks. Fix inadvertent use of "node".
19 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
21 * emit-rtl.c (init_derived_machine_modes): New functionm, split
23 (init_emit_once): ...here.
24 * rtl.h (init_derived_machine_modes): Declare.
25 * toplev.c (do_compile): Call it even if no_backend.
27 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
28 Mike Stump <mikestump@comcast.net>
29 Richard Sandiford <rdsandiford@googlemail.com>
30 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
32 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
33 (rtx_equal_for_memref_p): Update comment.
34 (adjust_offset_for_component_ref): Use wide-int interfaces.
35 * builtins.c (get_object_alignment_2): Likewise.
36 (c_readstr): Likewise.
37 (target_char_cast): Add comment.
38 (determine_block_size): Use wide-int interfaces.
39 (expand_builtin_signbit): Likewise.
40 (fold_builtin_int_roundingfn): Likewise.
41 (fold_builtin_bitop): Likewise.
42 (fold_builtin_bswap): Likewise.
43 (fold_builtin_logarithm): Use signop.
44 (fold_builtin_pow): Likewise.
45 (fold_builtin_memory_op): Use wide-int interfaces.
46 (fold_builtin_object_size): Likewise.
47 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
48 nb_iterations_estimate.
49 (record_niter_bound): Use wide-int interfaces.
50 (get_estimated_loop_iterations_int): Likewise.
51 (get_estimated_loop_iterations): Likewise.
52 (get_max_loop_iterations): Likewise.
53 * cfgloop.h: Include wide-int.h.
54 (struct nb_iter_bound): Change bound to widest_int.
55 (struct loop): Change nb_iterations_upper_bound and
56 nb_iterations_estimate to widest_int.
57 (record_niter_bound): Switch to use widest_int.
58 (get_estimated_loop_iterations): Likewise.
59 (get_max_loop_iterations): Likewise.
60 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
62 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
63 * combine.c (try_combine): Likewise.
64 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
65 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
67 (aarch64_float_const_representable_p): Likewise.
68 * config/arc/arc.c: Include wide-int.h.
69 (arc_can_use_doloop_p): Use wide-int interfaces.
70 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
71 (vfp3_const_double_index): Likewise.
72 * config/avr/avr.c (avr_out_round): Likewise.
73 (avr_fold_builtin): Likewise.
74 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
75 (bfin_can_use_doloop_p): Likewise.
76 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
77 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
78 * config/i386/i386.c: Include wide-int.h.
79 (ix86_data_alignment): Use wide-int interfaces.
80 (ix86_local_alignment): Likewise.
81 (ix86_emit_swsqrtsf): Update real_from_integer.
82 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
83 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
84 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
85 (zero_constant): Likewise.
86 (input_operand): Likewise.
87 (splat_input_operand): Likewise.
88 (non_logical_cint_operand): Change const_double to const_wide_int.
89 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
90 (easy_altivec_constant): Remove comment.
91 (paired_expand_vector_init): Use CONSTANT_P.
92 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
93 (rs6000_emit_move): Update checks.
94 (rs6000_aggregate_candidate): Use wide-int interfaces.
95 (rs6000_expand_ternop_builtin): Likewise.
96 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
97 (rs6000_assemble_integer): Likewise.
98 (rs6000_hash_constant): Likewise.
99 (output_toc): Likewise.
100 (rs6000_rtx_costs): Likewise.
101 (rs6000_emit_swrsqrt); Update call to real_from_integer.
102 * config/rs6000/rs6000-c.c: Include wide-int.h.
103 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
104 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
105 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
106 Handle CONST_WIDE_INT.
107 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
108 Use tree_fits_uhwi_p.
109 * config/sparc/sparc.c: Include wide-int.h.
110 (sparc_fold_builtin): Use wide-int interfaces.
111 * config/vax/vax.c: Include wide-int.h.
112 (vax_float_literal): Use real_from_integer.
113 * coretypes.h (struct hwivec_def): New.
116 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
117 (equiv_constant): Handle CONST_WIDE_INT.
118 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
119 (cselib_hash_rtx): Handle CONST_WIDE_INT.
120 * dbxout.c (stabstr_U): Use wide-int interfaces.
121 (dbxout_type): Update to use cst_fits_shwi_p.
122 * defaults.h (LOG2_BITS_PER_UNIT): Define.
123 (TARGET_SUPPORTS_WIDE_INT): Add default.
124 * dfp.c: Include wide-int.h.
125 (decimal_real_to_integer2): Use wide-int interfaces and rename to
126 decimal_real_to_integer.
127 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
128 decimal_real_to_integer.
129 * doc/generic.texi (Constant expressions): Update for wide_int.
130 * doc/rtl.texi (const_double): Likewise.
131 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
132 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
133 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
134 (REAL_VALUE_FROM_INT): Remove.
135 (TARGET_SUPPORTS_WIDE_INT): New.
136 * doc/tm.texi: Regenerate.
137 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
138 * double-int.h: Include wide-int.h.
139 (struct wi::int_traits): New.
140 * dwarf2out.c (get_full_len): New.
141 (dw_val_equal_p): Add case dw_val_class_wide_int.
142 (size_of_loc_descr): Likewise.
143 (output_loc_operands): Likewise.
144 (insert_double): Remove.
145 (insert_wide_int): New.
147 (print_die): Add case dw_val_class_wide_int.
148 (attr_checksum): Likewise.
149 (attr_checksum_ordered): Likewise.
150 (same_dw_val_p): Likewise.
151 (size_of_die): Likewise.
152 (value_format): Likewise.
153 (output_die): Likewise.
154 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
156 (clz_loc_descriptor): Use wide-int interfaces.
157 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
158 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
159 (round_up_to_align): Use wide-int interfaces.
160 (field_byte_offset): Likewise.
161 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
162 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
163 CONST_DOUBLE handling. Use wide-int interfaces.
164 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
165 (gen_enumeration_type_die): Use add_AT_wide.
166 (hash_loc_operands): Add case dw_val_class_wide_int.
167 (compare_loc_operands): Likewise.
168 * dwarf2out.h: Include wide-int.h.
170 (enum dw_val_class): Add dw_val_class_wide_int.
171 (struct dw_val_struct): Add val_wide.
172 * emit-rtl.c (const_wide_int_htab): New.
173 (const_wide_int_htab_hash): New.
174 (const_wide_int_htab_eq): New.
175 (lookup_const_wide_int): New.
176 (const_double_htab_hash): Use wide-int interfaces.
177 (const_double_htab_eq): Likewise.
178 (rtx_to_double_int): Conditionally compile for wide-int.
179 (immed_double_int_const): Rename to immed_wide_int_const and
181 (immed_double_const): Conditionally compile for wide-int.
182 (init_emit_once): Use wide-int interfaces.
183 * explow.c (plus_constant): Likewise.
184 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
185 (lshift_value): Use wide-int interfaces.
186 (expand_mult): Likewise.
187 (choose_multiplier): Likewise.
188 (expand_smod_pow2): Likewise.
189 (make_tree): Likewise.
190 * expr.c (convert_modes): Consolidate handling of constants.
191 Use wide-int interfaces.
192 (emit_group_load_1): Add note.
193 (store_expr): Update comment.
194 (get_inner_reference): Use wide-int interfaces.
195 (expand_constructor): Update comment.
196 (expand_expr_real_2): Use wide-int interfaces.
197 (expand_expr_real_1): Likewise.
198 (reduce_to_bit_field_precision): Likewise.
199 (const_vector_from_tree): Likewise.
200 * final.c: Include wide-int-print.h.
201 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
202 * fixed-value.c: Include wide-int.h.
203 (fixed_from_string): Use wide-int interfaces.
204 (fixed_to_decimal): Likewise.
205 (fixed_convert_from_real): Likewise.
206 (real_convert_from_fixed): Likewise.
207 * fold-const.h (mem_ref_offset): Return an offset_int.
208 (div_if_zero_remainder): Remove code parameter.
209 * fold-const.c (div_if_zero_remainder): Remove code parameter.
210 Use wide-int interfaces.
211 (may_negate_without_overflow_p): Use wide-int interfaces.
212 (negate_expr_p): Likewise.
213 (fold_negate_expr): Likewise.
214 (int_const_binop_1): Likewise.
215 (const_binop): Likewise.
216 (fold_convert_const_int_from_int): Likewise.
217 (fold_convert_const_int_from_real): Likewise.
218 (fold_convert_const_int_from_fixed): Likewise.
219 (fold_convert_const_fixed_from_int): Likewise.
220 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
221 (sign_bit_p): Use wide-int interfaces.
222 (make_range_step): Likewise.
223 (build_range_check): Likewise. Pass an integer of the correct type
224 instead of using integer_one_node.
225 (range_predecessor): Pass an integer of the correct type instead
226 of using integer_one_node.
227 (range_successor): Likewise.
228 (merge_ranges): Likewise.
229 (unextend): Use wide-int interfaces.
230 (extract_muldiv_1): Likewise.
231 (fold_div_compare): Likewise.
232 (fold_single_bit_test): Likewise.
233 (fold_sign_changed_comparison): Likewise.
234 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
235 (fold_plusminus_mult_expr): Use wide-int interfaces.
236 (native_encode_int): Likewise.
237 (native_interpret_int): Likewise.
238 (fold_unary_loc): Likewise.
239 (pointer_may_wrap_p): Likewise.
240 (size_low_cst): Likewise.
241 (mask_with_tz): Likewise.
242 (fold_binary_loc): Likewise.
243 (fold_ternary_loc): Likewise.
244 (multiple_of_p): Likewise.
245 (tree_call_nonnegative_warnv_p): Update calls to
246 tree_int_cst_min_precision and real_from_integer.
247 (fold_negate_const): Use wide-int interfaces.
248 (fold_abs_const): Likewise.
249 (fold_relational_const): Use tree_int_cst_lt.
250 (round_up_loc): Use wide-int interfaces.
251 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
252 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
253 * gengtype.c: Remove include of double-int.h.
254 (do_typedef): Use wide-int interfaces.
255 (open_base_files): Add wide-int.h.
256 (main): Add offset_int and widest_int typedefs.
257 * gengtype-lex.l: Handle "^".
258 (CXX_KEYWORD): Add "static".
259 * gengtype-parse.c (require3): New.
260 (require_template_declaration): Handle constant template arguments
261 and nested templates.
262 * gengtype-state.c: Don't include "double-int.h".
263 * genpreds.c (write_one_predicate_function): Update comment.
264 (write_tm_constrs_h): Add check for hval and lval use in
266 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
267 (add_to_sequence): Likewise.
268 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
269 and const_double_operand.
270 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
272 * gimple-fold.c (get_base_constructor): Likewise.
273 (fold_array_ctor_reference): Likewise.
274 (fold_nonarray_ctor_reference): Likewise.
275 (fold_const_aggregate_ref_1): Likewise.
276 (gimple_val_nonnegative_real_p): Likewise.
277 (gimple_fold_indirect_ref): Likewise.
278 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
279 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
280 (struct slsr_cand_d): Change index to be widest_int.
281 (struct incr_info_d): Change incr to be widest_int.
282 (alloc_cand_and_find_basis): Use wide-int interfaces.
283 (slsr_process_phi): Likewise.
284 (backtrace_base_for_ref): Likewise. Return a widest_int.
285 (restructure_reference): Take a widest_int instead of a double_int.
286 (slsr_process_ref): Use wide-int interfaces.
287 (create_mul_ssa_cand): Likewise.
288 (create_mul_imm_cand): Likewise.
289 (create_add_ssa_cand): Likewise.
290 (create_add_imm_cand): Take a widest_int instead of a double_int.
291 (slsr_process_add): Use wide-int interfaces.
292 (slsr_process_cast): Likewise.
293 (slsr_process_copy): Likewise.
294 (dump_candidate): Likewise.
295 (dump_incr_vec): Likewise.
296 (replace_ref): Likewise.
297 (cand_increment): Likewise. Return a widest_int.
298 (cand_abs_increment): Likewise.
299 (replace_mult_candidate): Take a widest_int instead of a double_int.
300 (replace_unconditional_candidate): Use wide-int interfaces.
301 (incr_vec_index): Take a widest_int instead of a double_int.
302 (create_add_on_incoming_edge): Likewise.
303 (create_phi_basis): Use wide-int interfaces.
304 (replace_conditional_candidate): Likewise.
305 (record_increment): Take a widest_int instead of a double_int.
306 (record_phi_increments): Use wide-int interfaces.
307 (phi_incr_cost): Take a widest_int instead of a double_int.
308 (lowest_cost_path): Likewise.
309 (total_savings): Likewise.
310 (analyze_increments): Use wide-int interfaces.
311 (ncd_with_phi): Take a widest_int instead of a double_int.
312 (ncd_of_cand_and_phis): Likewise.
313 (nearest_common_dominator_for_cands): Likewise.
314 (insert_initializers): Use wide-int interfaces.
315 (all_phi_incrs_profitable): Likewise.
316 (replace_one_candidate): Likewise.
317 (replace_profitable_candidates): Likewise.
318 * godump.c: Include wide-int-print.h.
319 (go_output_typedef): Use wide-int interfaces.
320 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
321 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
322 (build_loop_iteration_domains): Likewise.
323 * hooks.h: Include wide-int.h rather than double-int.h.
324 (hook_bool_dint_dint_uint_bool_true): Delete.
325 (hook_bool_wint_wint_uint_bool_true): Declare.
326 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
327 (hook_bool_wint_wint_uint_bool_true): New.
328 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
330 (ubsan_expand_si_overflow_mul_check): Likewise.
331 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
332 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
333 (get_ancestor_addr_info): Likewise.
334 (ipa_modify_call_arguments): Likewise.
335 * loop-doloop.c (doloop_modify): Likewise.
336 (doloop_optimize): Likewise.
337 * loop-iv.c (iv_number_of_iterations): Likewise.
338 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
339 (unroll_loop_constant_iterations): Likewise.
340 (decide_unroll_runtime_iterations): Likewise.
341 (unroll_loop_runtime_iterations): Likewise.
342 (decide_peel_simple): Likewise.
343 (decide_unroll_stupid): Likewise.
344 * lto-streamer-in.c (streamer_read_wi): Add.
345 (input_cfg): Use wide-int interfaces.
346 (lto_input_tree_1): Likewise.
347 * lto-streamer-out.c (streamer_write_wi): Add.
348 (hash_tree): Use wide-int interfaces.
349 (output_cfg): Likewise.
350 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
351 (GTFILES): Add wide-int.h and signop.h.
352 (TAGS): Look for .cc files too.
353 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
354 * optabs.c (expand_subword_shift): Likewise.
355 (expand_doubleword_shift): Likewise.
356 (expand_absneg_bit): Likewise.
357 (expand_copysign_absneg): Likewise.
358 (expand_copysign_bit): Likewise.
359 * postreload.c (reload_cse_simplify_set): Likewise.
360 * predict.c (predict_iv_comparison): Likewise.
361 * pretty-print.h: Include wide-int-print.h.
363 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
364 * print-tree.c: Include wide-int-print.h.
365 (print_node_brief): Use wide-int interfaces.
366 (print_node): Likewise.
367 * read-rtl.c (validate_const_wide_int): New.
368 (read_rtx_code): Add CONST_WIDE_INT case.
369 * real.c: Include wide-int.h.
370 (real_to_integer2): Delete.
371 (real_to_integer): New function, returning a wide_int.
372 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
373 (ten_to_ptwo): Update call to real_from_integer.
374 (real_digit): Likewise.
375 * real.h: Include signop.h, wide-int.h and insn-modes.h.
376 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
377 (REAL_VALUE_TO_INT): Delete.
378 (real_to_integer): Declare a wide-int form.
379 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
380 * recog.c (const_int_operand): Improve comment.
381 (const_scalar_int_operand): New.
382 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
383 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
384 (split_double): Likewise.
385 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
386 (rtx_size): Likewise.
387 (rtx_alloc_stat_v): New.
388 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
389 (cwi_output_hex): New.
390 (iterative_hash_rtx): Handle CONST_WIDE_INT.
391 (cwi_check_failed_bounds): New.
392 * rtl.def (CONST_WIDE_INT): New.
393 * rtl.h: Include <utility> and wide-int.h.
394 (struct hwivec_def): New.
395 (CWI_GET_NUM_ELEM): New.
396 (CWI_PUT_NUM_ELEM): New.
397 (struct rtx_def): Add num_elem and hwiv.
398 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
399 (CASE_CONST_UNIQUE): Likewise.
400 (CASE_CONST_ANY): Likewise.
401 (CONST_SCALAR_INT_P): Likewise.
402 (CONST_WIDE_INT_P): New.
405 (cwi_check_failed_bounds): New.
408 (CONST_WIDE_INT_VEC) New.
409 (CONST_WIDE_INT_NUNITS) New.
410 (CONST_WIDE_INT_ELT) New.
411 (rtx_mode_t): New type.
412 (wi::int_traits <rtx_mode_t>): New.
414 (wi::min_value): New.
415 (wi::max_value): New.
417 (const_wide_int_alloc): New.
418 (immed_wide_int_const): New.
419 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
420 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
421 * signop.h: New file.
422 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
423 (simplify_const_unary_operation): Use wide-int interfaces.
424 (simplify_binary_operation_1): Likewise.
425 (simplify_const_binary_operation): Likewise.
426 (simplify_const_relational_operation): Likewise.
427 (simplify_immed_subreg): Likewise.
428 * stmt.c (expand_case): Likewise.
429 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
430 signop rather than a bool.
431 * stor-layout.c (layout_type): Use wide-int interfaces.
432 (initialize_sizetypes): Update calls to
433 set_min_and_max_values_for_integral_type.
434 (set_min_and_max_values_for_integral_type): Take a signop rather
435 than a bool. Use wide-int interfaces.
436 (fixup_signed_type): Update accordingly. Remove
437 HOST_BITS_PER_DOUBLE_INT limit.
438 (fixup_unsigned_type): Likewise.
439 * system.h (STATIC_CONSTANT_P): New.
440 (STATIC_ASSERT): New.
441 * target.def (can_use_doloop_p): Take widest_ints rather than
443 * target.h: Include wide-int.h rather than double-int.h.
444 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
446 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
447 rather than INT_CST_LT_UNSIGNED.
448 (can_use_doloop_if_innermost): Take widest_ints rather than
450 * tree-affine.c: Include wide-int-print.h.
451 (double_int_ext_for_comb): Delete.
452 (wide_int_ext_for_comb): New.
453 (aff_combination_zero): Use wide-int interfaces.
454 (aff_combination_const): Take a widest_int instead of a double_int.
455 (aff_combination_elt): Use wide-int interfaces.
456 (aff_combination_scale): Take a widest_int instead of a double_int.
457 (aff_combination_add_elt): Likewise.
458 (aff_combination_add_cst): Likewise.
459 (aff_combination_add): Use wide-int interfaces.
460 (aff_combination_convert): Likewise.
461 (tree_to_aff_combination): Likewise.
462 (add_elt_to_tree): Take a widest_int instead of a double_int.
463 (aff_combination_to_tree): Use wide-int interfaces.
464 (aff_combination_remove_elt): Likewise.
465 (aff_combination_add_product): Take a widest_int instead of
467 (aff_combination_mult): Use wide-int interfaces.
468 (aff_combination_expand): Likewise.
469 (double_int_constant_multiple_p): Delete.
470 (wide_int_constant_multiple_p): New.
471 (aff_combination_constant_multiple_p): Take a widest_int pointer
472 instead of a double_int pointer.
473 (print_aff): Use wide-int interfaces.
474 (get_inner_reference_aff): Take a widest_int pointer
475 instead of a double_int pointer.
476 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
477 * tree-affine.h: Include wide-int.h.
478 (struct aff_comb_elt): Change type of coef to widest_int.
479 (struct affine_tree_combination): Change type of offset to widest_int.
480 (double_int_ext_for_comb): Delete.
481 (wide_int_ext_for_comb): New.
482 (aff_combination_const): Use widest_int instead of double_int.
483 (aff_combination_scale): Likewise.
484 (aff_combination_add_elt): Likewise.
485 (aff_combination_constant_multiple_p): Likewise.
486 (get_inner_reference_aff): Likewise.
487 (aff_comb_cannot_overlap_p): Likewise.
488 (aff_combination_zero_p): Use wide-int interfaces.
489 * tree.c: Include tree.h.
490 (init_ttree): Use make_int_cst.
491 (tree_code_size): Removed code for INTEGER_CST case.
492 (tree_size): Add INTEGER_CST case.
493 (make_node_stat): Update comment.
494 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
495 (build_int_cst_type): Use wide-int interfaces.
496 (double_int_to_tree): Likewise.
497 (double_int_fits_to_tree_p): Delete.
498 (force_fit_type_double): Delete.
499 (force_fit_type): New.
500 (int_cst_hash_hash): Use wide-int interfaces.
501 (int_cst_hash_eq): Likewise.
502 (build_int_cst_wide): Delete.
503 (wide_int_to_tree): New.
504 (cache_integer_cst): Use wide-int interfaces.
505 (build_low_bits_mask): Likewise.
506 (cst_and_fits_in_hwi): Likewise.
507 (real_value_from_int_cst): Likewise.
508 (make_int_cst_stat): New.
509 (integer_zerop): Use wide_int interfaces.
510 (integer_onep): Likewise.
511 (integer_all_onesp): Likewise.
512 (integer_pow2p): Likewise.
513 (integer_nonzerop): Likewise.
514 (tree_log2): Likewise.
515 (tree_floor_log2): Likewise.
516 (tree_ctz): Likewise.
517 (int_size_in_bytes): Likewise.
518 (mem_ref_offset): Return an offset_int rather than a double_int.
519 (build_type_attribute_qual_variant): Use wide_int interfaces.
520 (type_hash_eq): Likewise
521 (tree_int_cst_equal): Likewise.
522 (tree_int_cst_lt): Delete.
523 (tree_int_cst_compare): Likewise.
524 (tree_fits_shwi_p): Use wide_int interfaces.
525 (tree_fits_uhwi_p): Likewise.
526 (tree_int_cst_sign_bit): Likewise.
527 (tree_int_cst_sgn): Likewise.
528 (tree_int_cst_min_precision): Take a signop rather than a bool.
529 (simple_cst_equal): Use wide_int interfaces.
530 (compare_tree_int): Likewise.
531 (iterative_hash_expr): Likewise.
532 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
534 (get_type_static_bounds): Use wide_int interfaces.
535 (tree_int_cst_elt_check_failed): New.
536 (build_common_tree_nodes): Reordered to set prec before filling in
538 (int_cst_value): Check cst_and_fits_in_hwi.
539 (widest_int_cst_value): Use wide_int interfaces.
540 (upper_bound_in_type): Likewise.
541 (lower_bound_in_type): Likewise.
542 (num_ending_zeros): Likewise.
543 (drop_tree_overflow): Likewise.
544 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
545 (gen_conditions_for_pow_cst_base): Likewise.
546 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
547 (group_case_labels_stmt): Use wide-int interfaces.
548 (verify_gimple_assign_binary): Likewise.
549 (print_loop): Likewise.
550 * tree-chrec.c (tree_fold_binomial): Likewise.
551 * tree-core.h (struct tree_base): Add int_length.
552 (struct tree_int_cst): Change rep of value.
553 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
554 (dr_may_alias_p): Likewise.
555 (max_stmt_executions_tree): Likewise.
556 * tree.def (INTEGER_CST): Update comment.
557 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
558 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
559 * tree-dump.c: Include wide-int.h and wide-int-print.h.
560 (dequeue_and_dump): Use wide-int interfaces.
561 * tree.h: Include wide-int.h.
562 (NULL_TREE): Moved to earlier loc in file.
563 (TREE_INT_CST_ELT_CHECK): New.
564 (tree_int_cst_elt_check_failed): New.
566 (TREE_INT_CST): Delete.
567 (TREE_INT_CST_LOW): Use wide-int interfaces.
568 (TREE_INT_CST_HIGH): Delete.
569 (TREE_INT_CST_NUNITS): New.
570 (TREE_INT_CST_EXT_NUNITS): Likewise.
571 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
572 (TREE_INT_CST_ELT): Likewise.
573 (INT_CST_LT): Delete.
574 (tree_int_cst_elt_check): New (two forms).
575 (type_code_size): Update comment.
576 (make_int_cst_stat, make_int_cst): New.
577 (tree_to_double_int): Delete.
578 (double_int_fits_to_tree_p): Delete.
579 (force_fit_type_double): Delete.
580 (build_int_cstu): Replace with out-of-line function.
581 (build_int_cst_wide): Delete.
582 (tree_int_cst_lt): Define inline.
583 (tree_int_cst_le): New.
584 (tree_int_cst_compare): Define inline.
585 (tree_int_cst_min_precision): Take a signop rather than a bool.
586 (wi::int_traits <const_tree>): New.
587 (wi::int_traits <tree>): New.
588 (wi::extended_tree): New.
589 (wi::int_traits <wi::extended_tree>): New.
590 (wi::to_widest): New.
591 (wi::to_offset): New.
592 (wi::fits_to_tree_p): New.
593 (wi::min_value): New.
594 (wi::max_value): New.
595 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
596 (copy_tree_body_r): Likewise.
597 * tree-object-size.c (compute_object_offset): Likewise.
598 (addr_object_size): Likewise.
599 * tree-predcom.c: Include wide-int-print.h.
600 (struct dref_d): Change type of offset to widest_int.
601 (dump_dref): Call wide-int printer.
602 (aff_combination_dr_offset): Use wide-int interfaces.
603 (determine_offset): Take a widest_int pointer rather than a
605 (split_data_refs_to_components): Use wide-int interfaces.
606 (suitable_component_p): Likewise.
607 (order_drefs): Likewise.
608 (add_ref_to_chain): Likewise.
609 (valid_initializer_p): Likewise.
610 (determine_roots_comp): Likewise.
611 * tree-pretty-print.c: Include wide-int-print.h.
612 (dump_generic_node): Use wide-int interfaces.
613 * tree-sra.c (sra_ipa_modify_expr): Likewise.
614 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
615 (move_fixed_address_to_symbol): Likewise.
616 (move_hint_to_base): Likewise.
617 (move_pointer_to_base): Likewise.
618 (move_variant_to_index): Likewise.
619 (most_expensive_mult_to_index): Likewise.
620 (addr_to_parts): Likewise.
621 (copy_ref_info): Likewise.
622 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
623 (indirect_refs_may_alias_p): Likewise.
624 (stmt_kills_ref_p_1): Likewise.
625 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
626 * tree-ssa-ccp.c: Update comment at top of file. Include
628 (struct prop_value_d): Change type of mask to widest_int.
629 (extend_mask): New function.
630 (dump_lattice_value): Use wide-int interfaces.
631 (get_default_value): Likewise.
632 (set_constant_value): Likewise.
633 (set_value_varying): Likewise.
634 (valid_lattice_transition): Likewise.
635 (set_lattice_value): Likewise.
636 (value_to_double_int): Delete.
637 (value_to_wide_int): New.
638 (get_value_from_alignment): Use wide-int interfaces.
639 (get_value_for_expr): Likewise.
640 (do_dbg_cnt): Likewise.
641 (ccp_finalize): Likewise.
642 (ccp_lattice_meet): Likewise.
643 (bit_value_unop_1): Use widest_ints rather than double_ints.
644 (bit_value_binop_1): Likewise.
645 (bit_value_unop): Use wide-int interfaces.
646 (bit_value_binop): Likewise.
647 (bit_value_assume_aligned): Likewise.
648 (evaluate_stmt): Likewise.
649 (ccp_fold_stmt): Likewise.
650 (visit_cond_stmt): Likewise.
651 (ccp_visit_stmt): Likewise.
652 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
653 (constant_pointer_difference): Likewise.
654 (associate_pointerplus): Likewise.
655 (combine_conversions): Likewise.
656 * tree-ssa-loop.h: Include wide-int.h.
657 (struct tree_niter_desc): Change type of max to widest_int.
658 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
659 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
660 (remove_redundant_iv_tests): Likewise.
661 (canonicalize_loop_induction_variables): Likewise.
662 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
663 (constant_multiple_of): Take a widest_int pointer instead of
664 a double_int pointer.
665 (get_computation_aff): Use wide-int interfaces.
666 (ptr_difference_cost): Likewise.
667 (difference_cost): Likewise.
668 (get_loop_invariant_expr_id): Likewise.
669 (get_computation_cost_at): Likewise.
670 (iv_elimination_compare_lt): Likewise.
671 (may_eliminate_iv): Likewise.
672 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
673 instead of double_int.
674 (max_loop_iterations): Likewise.
675 (max_stmt_executions): Likewise.
676 (estimated_stmt_executions): Likewise.
677 * tree-ssa-loop-niter.c: Include wide-int-print.h.
678 (split_to_var_and_offset): Use wide-int interfaces.
679 (determine_value_range): Likewise.
680 (bound_difference_of_offsetted_base): Likewise.
681 (bounds_add): Take a widest_int instead of a double_int.
682 (number_of_iterations_ne_max): Use wide-int interfaces.
683 (number_of_iterations_ne): Likewise.
684 (number_of_iterations_lt_to_ne): Likewise.
685 (assert_loop_rolls_lt): Likewise.
686 (number_of_iterations_lt): Likewise.
687 (number_of_iterations_le): Likewise.
688 (number_of_iterations_cond): Likewise.
689 (number_of_iterations_exit): Likewise.
690 (finite_loop_p): Likewise.
691 (derive_constant_upper_bound_assign): Likewise.
692 (derive_constant_upper_bound): Return a widest_int.
693 (derive_constant_upper_bound_ops): Likewise.
694 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
695 (record_estimate): Take a widest_int rather than a double_int.
696 (record_nonwrapping_iv): Use wide-int interfaces.
697 (double_int_cmp): Delete.
699 (bound_index): Take a widest_int rather than a double_int.
700 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
701 (maybe_lower_iteration_bound): Likewise.
702 (estimate_numbers_of_iterations_loop): Likewise.
703 (estimated_loop_iterations): Take a widest_int pointer than than
704 a double_int pointer.
705 (estimated_loop_iterations_int): Use wide-int interfaces.
706 (max_loop_iterations): Take a widest_int pointer than than
707 a double_int pointer.
708 (max_loop_iterations_int): Use wide-int interfaces.
709 (max_stmt_executions): Take a widest_int pointer than than
710 a double_int pointer.
711 (estimated_stmt_executions): Likewise.
712 (n_of_executions_at_most): Use wide-int interfaces.
713 (scev_probably_wraps_p): Likewise.
714 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
716 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
718 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
719 double_ints. Adjust for trailing_wide_ints <3> representation.
720 (set_nonzero_bits): Likewise.
721 (get_range_info): Return wide_ints rather than double_ints.
722 Adjust for trailing_wide_ints <3> representation.
723 (get_nonzero_bits): Likewise.
724 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
726 * tree-ssanames.h (struct range_info_def): Replace min, max and
727 nonzero_bits with a trailing_wide_ints <3>.
728 (set_range_info): Use wide_int_refs rather than double_ints.
729 (set_nonzero_bits): Likewise.
730 (get_range_info): Return wide_ints rather than double_ints.
731 (get_nonzero_bits): Likewise.
732 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
733 * tree-ssa-pre.c (phi_translate_1): Likewise.
734 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
735 (acceptable_pow_call): Likewise.
736 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
738 (vn_reference_fold_indirect): Likewise.
739 (vn_reference_maybe_forwprop_address): Likewise.
740 (valueize_refs_1): Likewise.
741 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
742 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
743 tree_int_cst_lt and tree_int_cst_le.
744 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
746 (streamer_alloc_tree): Likewise.
747 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
748 (streamer_write_tree_header): Likewise.
749 (streamer_write_integer_cst): Likewise.
750 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
751 (build_constructors): Likewise.
752 (array_value_type): Likewise.
753 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
754 (vect_check_gather): Likewise.
755 * tree-vect-generic.c (build_replicated_const): Likewise.
756 (expand_vector_divmod): Likewise.
757 * tree-vect-loop.c (vect_transform_loop): Likewise.
758 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
759 (vect_do_peeling_for_alignment): Likewise.
760 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
761 * tree-vrp.c: Include wide-int.h.
762 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
763 (extract_range_from_assert): Use wide-int interfaces.
764 (vrp_int_const_binop): Likewise.
765 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
767 (ranges_from_anti_range): Use wide-int interfaces.
768 (quad_int_cmp): Delete.
769 (quad_int_pair_sort): Likewise.
770 (extract_range_from_binary_expr_1): Use wide-int interfaces.
771 (extract_range_from_unary_expr_1): Likewise.
772 (adjust_range_with_scev): Likewise.
773 (masked_increment): Take and return wide_ints rather than double_ints.
774 (register_edge_assert_for_2): Use wide-int interfaces.
775 (check_array_ref): Likewise.
776 (search_for_addr_array): Likewise.
777 (maybe_set_nonzero_bits): Likewise.
778 (union_ranges): Pass an integer of the correct type instead of
779 using integer_one_node.
780 (intersect_ranges): Likewise.
781 (simplify_truth_ops_using_ranges): Likewise.
782 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
783 (range_fits_type_p): Likewise.
784 (simplify_cond_using_ranges): Likewise. Take a signop rather than
786 (simplify_conversion_using_ranges): Use wide-int interfaces.
787 (simplify_float_conversion_using_ranges): Likewise.
788 (vrp_finalize): Likewise.
789 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
790 (gimple_stringops_transform): Likewise.
791 * varasm.c (decode_addr_const): Likewise.
792 (const_hash_1): Likewise.
793 (const_rtx_hash_1): Likewise
794 (output_constant): Likewise.
795 (array_size_for_constructor): Likewise.
796 (output_constructor_regular_field): Likewise.
797 (output_constructor_bitfield): Likewise.
798 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
799 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
801 * gencheck.c: Define BITS_PER_UNIT.
804 * wide-int-print.cc: New.
805 * wide-int-print.h: New.
807 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
809 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
811 2014-05-06 Richard Biener <rguenther@suse.de>
813 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
814 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
815 (TODO_verify_all): Adjust.
816 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
817 TODO_verify_stmts and TODO_verify_rtl_sharing.
818 * bb-reorder.c: Likewise.
819 * cfgexpand.c: Likewise.
822 * function.c: Likewise.
823 * fwprop.c: Likewise.
825 * gimple-ssa-isolate-paths.c: Likewise.
826 * gimple-ssa-strength-reduction.c: Likewise.
827 * ipa-split.c: Likewise.
828 * loop-init.c: Likewise.
829 * loop-unroll.c: Likewise.
830 * lower-subreg.c: Likewise.
831 * modulo-sched.c: Likewise.
832 * postreload-gcse.c: Likewise.
833 * predict.c: Likewise.
835 * sched-rgn.c: Likewise.
836 * store-motion.c: Likewise.
837 * tracer.c: Likewise.
838 * trans-mem.c: Likewise.
839 * tree-call-cdce.c: Likewise.
840 * tree-cfg.c: Likewise.
841 * tree-cfgcleanup.c: Likewise.
842 * tree-complex.c: Likewise.
843 * tree-eh.c: Likewise.
844 * tree-emutls.c: Likewise.
845 * tree-if-conv.c: Likewise.
846 * tree-into-ssa.c: Likewise.
847 * tree-loop-distribution.c: Likewise.
848 * tree-object-size.c: Likewise.
849 * tree-parloops.c: Likewise.
850 * tree-pass.h: Likewise.
851 * tree-sra.c: Likewise.
852 * tree-ssa-ccp.c: Likewise.
853 * tree-ssa-copy.c: Likewise.
854 * tree-ssa-copyrename.c: Likewise.
855 * tree-ssa-dce.c: Likewise.
856 * tree-ssa-dom.c: Likewise.
857 * tree-ssa-dse.c: Likewise.
858 * tree-ssa-forwprop.c: Likewise.
859 * tree-ssa-ifcombine.c: Likewise.
860 * tree-ssa-loop-ch.c: Likewise.
861 * tree-ssa-loop-ivcanon.c: Likewise.
862 * tree-ssa-loop.c: Likewise.
863 * tree-ssa-math-opts.c: Likewise.
864 * tree-ssa-phiopt.c: Likewise.
865 * tree-ssa-phiprop.c: Likewise.
866 * tree-ssa-pre.c: Likewise.
867 * tree-ssa-reassoc.c: Likewise.
868 * tree-ssa-sink.c: Likewise.
869 * tree-ssa-strlen.c: Likewise.
870 * tree-ssa-tail-merge.c: Likewise.
871 * tree-ssa-uncprop.c: Likewise.
872 * tree-switch-conversion.c: Likewise.
873 * tree-tailcall.c: Likewise.
874 * tree-vect-generic.c: Likewise.
875 * tree-vectorizer.c: Likewise.
876 * tree-vrp.c: Likewise.
878 * var-tracking.c: Likewise.
879 * bt-load.c: Likewise.
880 * cfgcleanup.c: Likewise.
881 * combine-stack-adj.c: Likewise.
882 * combine.c: Likewise.
883 * compare-elim.c: Likewise.
884 * config/epiphany/resolve-sw-modes.c: Likewise.
885 * config/i386/i386.c: Likewise.
886 * config/mips/mips.c: Likewise.
887 * config/s390/s390.c: Likewise.
888 * config/sh/sh_treg_combine.cc: Likewise.
889 * config/sparc/sparc.c: Likewise.
894 * mode-switching.c: Likewise.
895 * passes.c: Likewise.
896 * postreload.c: Likewise.
898 * reg-stack.c: Likewise.
899 * regcprop.c: Likewise.
900 * regrename.c: Likewise.
903 2014-05-06 Richard Biener <rguenther@suse.de>
906 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
907 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
909 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
912 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
914 2014-05-05 Radovan Obradovic <robradovic@mips.com>
915 Tom de Vries <tom@codesourcery.com>
917 * target.def (call_fusage_contains_non_callee_clobbers): New DEFHOOKPOD.
918 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
920 (@node Miscellaneous Register Hooks): New node.
921 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
922 * doc/tm.texi: Regenerate.
924 2014-05-05 Marek Polacek <polacek@redhat.com>
927 * opts.c (common_handle_option): Call error_at instead of warning_at.
929 2014-05-05 Richard Biener <rguenther@suse.de>
931 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
932 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
933 under the TODO_verify_il umbrella.
935 2014-05-05 Richard Biener <rguenther@suse.de>
937 * passes.c (execute_function_todo): Move TODO_verify_flow under
938 the TODO_verify_ul umbrella.
940 2014-05-05 Richard Biener <rguenther@suse.de>
943 * fold-const.c (fold_binary_loc): Consistently avoid
944 canonicalizing X & CST away from a CST that is the mask
947 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
949 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
950 int argument to enum machine_mode.
951 (picochip_class_max_nregs): Ditto.
952 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
953 (picochip_class_max_nregs): Ditto.
955 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
957 * target.def: Add new target hook.
958 * doc/tm.texi: Regenerate.
959 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
960 * targhooks.c (default_keep_leaf_when_profiled): New function.
962 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
963 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
965 2014-05-05 Bin Cheng <bin.cheng@arm.com>
967 PR tree-optimization/60363
968 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
969 (copy_phi_args): New parameters. Call get_value_locus_in_path.
970 (update_destination_phis): New parameter.
971 (create_edge_and_update_destination_phis): Ditto.
972 (ssa_fix_duplicate_block_edges): Pass new arguments.
973 (thread_single_edge): Ditto.
975 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
977 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
978 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
979 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
980 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
981 Use RS6000_BTM_HARD_FLOAT.
982 (BU_MISC_2): Likewise.
983 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
984 RS6000_BTM_HARD_FLOAT.
985 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
987 (rs6000_invalid_builtin): Add hard floating builtin support.
988 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
990 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
992 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
994 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
995 Add missing function* argument.
997 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
999 * lra-constraints.c (valid_address_p): Move earlier in file.
1000 Add a constraint argument to the address_info version.
1001 (satisfies_memory_constraint_p): New function.
1002 (satisfies_address_constraint_p): Likewise.
1003 (process_alt_operands, curr_insn_transform): Use them.
1004 (process_address): Pass the constraint to valid_address_p when
1005 checking address operands.
1007 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1009 * config/mips/mips.c (mips_isa_rev): New variable.
1010 (mips_set_architecture): Set it.
1011 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1013 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1014 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1015 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1016 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1017 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
1018 conditions in terms of mips_isa_rev.
1019 (mips_isa_rev): Declare.
1021 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1023 * config/sh/sh-mem.cc: Use tabs instead of spaces.
1024 (prob_unlikely, prob_likely): Make variables const.
1026 2014-05-03 Denis Chertykov <chertykov@gmail.com>
1028 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
1030 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1032 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
1034 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1036 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
1037 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
1038 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
1040 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
1041 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
1043 Replace usage of ROUND_REG with sh_round_reg.
1044 Use CEIL instead of ROUND_ADVANCE.
1046 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1049 * config/sh/sh.c: Include stdlib headers before everything else.
1051 2014-05-02 Jakub Jelinek <jakub@redhat.com>
1053 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
1054 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
1055 (gimplify_adjust_omp_clauses): Simd region is never
1056 directly nested in combined parallel. Instead, for linear
1057 with copyin/copyout, if in combined for simd loop, make decl
1058 firstprivate/lastprivate on OMP_FOR.
1059 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
1060 expand_omp_for_static_chunk): When setting endvar, also set
1061 fd->loop.v to the same value.
1063 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1065 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
1067 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
1069 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
1072 2014-05-02 Marek Polacek <polacek@redhat.com>
1074 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
1076 2014-02-26 Kito Cheng <kito@0xlab.org>
1078 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
1079 to a C expression marco.
1080 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
1081 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
1082 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
1083 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
1084 HONOR_REG_ALLOC_ORDER.
1085 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
1087 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1089 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
1091 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1093 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
1095 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
1097 * tree-if-conv.c (is_cond_scalar_reduction): New function.
1098 (convert_scalar_cond_reduction): Likewise.
1099 (predicate_scalar_phi): Add recognition and transformation
1100 of simple conditioanl reduction to be vectorizable.
1102 2014-05-01 Marek Polacek <polacek@redhat.com>
1105 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
1107 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
1109 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
1110 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
1111 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
1112 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
1113 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
1114 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
1115 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
1116 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
1118 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
1120 * config/arc/arc.opt (mlra): Move comment above option name
1121 to avoid mis-parsing as language options.
1123 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1125 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
1126 * config/sol2.h: ... here.
1127 * config/sol2-10.h: Remove.
1129 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
1130 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
1131 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
1132 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
1133 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
1134 * config/sol2.h: ... here.
1135 (SECTION_NAME_FORMAT): Don't redefine.
1136 (STARTFILE_ARCH32_SPEC): Rename to ...
1137 (STARTFILE_ARCH_SPEC): ... this.
1138 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
1139 * config/sparc/sol2.h: ... here.
1140 (SECTION_NAME_FORMAT): Don't undef.
1141 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
1142 (SUBTARGET_EXTRA_SPECS): Remove.
1143 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
1145 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
1146 (MD_STARTFILE_PREFIX): Remove.
1147 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
1148 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
1149 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
1150 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
1151 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
1152 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
1153 * config/i386/sol2.h: ... here.
1154 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
1155 * config/i386/sol2-bi.h: Remove.
1156 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
1157 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
1159 * config/i386/t-sol2-64: Rename to ...
1160 * config/i386/t-sol2: ... this.
1161 * config/sparc/t-sol2-64: Rename to ...
1162 * config/sparc/t-sol2: ... this.
1164 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
1165 sol2_tm_file_head, sol2_tm_file_tail.
1166 Include ${cpu_type}/sol2.h before sol2.h.
1168 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
1169 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
1170 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
1171 Reflect i386/t-sol2-64 renaming.
1172 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
1173 Reflect sparc/t-sol2-64 renaming.
1175 2014-04-30 Richard Biener <rguenther@suse.de>
1177 * passes.c (execute_function_todo): Move TODO_verify_stmts
1178 and TODO_verify_ssa under the TODO_verify_il umbrella.
1179 * tree-ssa.h (verify_ssa): Adjust prototype.
1180 * tree-ssa.c (verify_ssa): Add parameter to tell whether
1181 we should verify SSA operands.
1182 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
1183 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
1184 whether we should verify whether not throwing stmts have EH info.
1185 * graphite-scop-detection.c (create_sese_edges): Adjust.
1186 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
1187 * tree-eh.c (lower_try_finally_switch): Do not add the
1188 default case label twice.
1190 2014-04-30 Marek Polacek <polacek@redhat.com>
1192 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
1193 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
1194 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
1195 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
1197 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
1199 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
1200 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
1201 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
1202 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
1203 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
1204 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
1205 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
1206 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
1208 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1210 * tree-cfg.c (dump_function_to_file): Dump the return type of
1211 functions, in a line to itself before the function body, mimicking
1212 the layout of a C function.
1214 2014-04-29 Jakub Jelinek <jakub@redhat.com>
1216 PR tree-optimization/60971
1217 * tree-tailcall.c (process_assignment): Reject conversions which
1220 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
1222 * calls.c (initialize_argument_information): Always treat
1223 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
1224 (expand_call): Likewise.
1225 (emit_library_call_calue_1): Likewise.
1226 * expr.c (PUSH_ARGS_REVERSED): Do not define.
1227 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
1230 2014-04-29 Nick Clifton <nickc@redhat.com>
1232 * config/msp430/msp430.md (umulsidi): Fix typo.
1233 (mulhisi3): Enable even inside interrupt handlers.
1234 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
1235 bigger return address pushed in large mode.
1237 2014-04-29 Nick Clifton <nickc@redhat.com>
1239 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
1240 (arc_init_reg_tables): Use a machine_mode enum to iterate over
1242 * config/m32r/m32r.c (init_reg_tables): Likewise.
1243 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
1244 enum to hold the modes.
1246 2014-04-29 Richard Biener <rguenther@suse.de>
1248 * dominance.c (free_dominance_info): Add overload with
1250 (dom_info_state): Likewise.
1251 (dom_info_available_p): Likewise.
1252 * basic-block.h (free_dominance_info, dom_info_state,
1253 dom_info_available_p): Declare overloads.
1254 * passes.c (execute_function_todo): Verify that verifiers
1255 don't change dominator info state. Drop dominator info
1256 for IPA pass invocations.
1257 * cgraph.c (release_function_body): Restore asserts that
1258 dominator information is released.
1260 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
1262 * doc/invoke.texi: Fix typo.
1263 * tree-vrp.c: Fix typos.
1264 * gimple.c (infer_nonnull_range): Reorder operands of an &&
1267 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1269 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
1271 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1273 * config/aarch64/aarch64-builtins.c
1274 (aarch64_types_storestruct_lane_qualifiers): New.
1275 (TYPES_STORESTRUCT_LANE): Likewise.
1276 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
1277 (st3_lane): Likewise.
1278 (st4_lane): Likewise.
1279 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
1280 (vec_store_lanesci_lane<mode>): Likewise.
1281 (vec_store_lanesxi_lane<mode>): Likewise.
1282 (aarch64_st2_lane<VQ:mode>): Likewise.
1283 (aarch64_st3_lane<VQ:mode>): Likewise.
1284 (aarch64_st4_lane<VQ:mode>): Likewise.
1285 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
1286 * config/aarch64/arm_neon.h
1287 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
1288 use new macro arguments.
1289 (__ST3_LANE_FUNC): Likewise.
1290 (__ST4_LANE_FUNC): Likewise.
1291 * config/aarch64/iterators.md (V_TWO_ELEM): New.
1292 (V_THREE_ELEM): Likewise.
1293 (V_FOUR_ELEM): Likewise.
1295 2014-04-28 David Malcolm <dmalcolm@redhat.com>
1297 * doc/gimple.texi: Replace the description of the now-defunct
1298 union gimple_statement_d with a diagram showing the
1299 gimple_statement_base class hierarchy and its relationships to
1300 the GSS_ and GIMPLE_ enums.
1302 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1304 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
1305 * config/aarch64/aarch64.c
1306 (aarch64_cannot_change_mode_class): Weaken conditions.
1307 (aarch64_modes_tieable_p): New.
1308 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
1310 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
1312 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
1313 (loadsync_<mode>): Change mode.
1314 (load_quadpti, store_quadpti): New.
1315 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
1316 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
1318 2014-04-28 Martin Jambor <mjambor@suse.cz>
1320 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
1321 same alias type as the original statement.
1322 (subreplacement_assignment_data): New type.
1323 (handle_unscalarized_data_in_subtree): New type of parameter,
1324 generate new memory accesses with same alias type as the original
1326 (load_assign_lhs_subreplacements): Likewise.
1327 (sra_modify_constructor_assign): Generate new memory accesses with
1328 same alias type as the original statement.
1330 2014-04-28 Richard Biener <rguenther@suse.de>
1332 * tree-pass.h (TODO_verify_il): Define.
1333 (TODO_verify_all): Complete properly.
1334 * passes.c (execute_function_todo): Move existing loop-closed
1335 SSA verification under TODO_verify_il.
1336 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
1337 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
1338 Fix tree sharing issue.
1340 2014-04-28 Richard Biener <rguenther@suse.de>
1343 * builtins.def (DEF_C11_BUILTIN): Add.
1344 (BUILT_IN_ALIGNED_ALLOC): Likewise.
1345 * coretypes.h (enum function_class): Add function_c11_misc.
1346 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1347 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
1348 (call_may_clobber_ref_p_1): Likewise.
1349 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
1350 (mark_all_reaching_defs_necessary_1): Likewise.
1351 (propagate_necessity): Likewise.
1352 (eliminate_unnecessary_stmts): Likewise.
1353 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
1355 2014-04-28 Richard Biener <rguenther@suse.de>
1357 * tree-vrp.c (vrp_var_may_overflow): Remove.
1358 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
1359 with overflow immediately bump to one before that value and
1360 let iteration figure out overflow status.
1362 2014-04-28 Richard Biener <rguenther@suse.de>
1364 * configure.ac: Do valgrind header checks unconditionally.
1365 Add --enable-valgrind-annotations.
1366 * system.h: Guard valgrind header inclusion with
1367 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
1368 * alloc-pool.c (pool_alloc, pool_free): Use
1369 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
1370 to guard possibly dead code.
1371 * config.in: Regenerated.
1372 * configure: Likewise.
1374 2014-04-28 Jeff Law <law@redhat.com>
1376 PR tree-optimization/60902
1377 * tree-ssa-threadedge.c
1378 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
1379 over real defs when invalidating outputs from statements that do not
1380 produce useful outputs for threading.
1382 2014-04-28 Richard Biener <rguenther@suse.de>
1384 PR tree-optimization/60979
1385 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
1386 SCOPs that end in a block with a successor with abnormal
1389 2014-04-28 Richard Biener <rguenther@suse.de>
1391 * tree-pass.h (execute_pass_list): Adjust prototype.
1392 * passes.c (pass_manager::execute_early_local_passes):
1394 (do_per_function): Change callback signature, push all actual
1395 work to the callbals.
1396 (do_per_function_toporder): Likewise.
1397 (execute_function_dump): Adjust.
1398 (execute_function_todo): Likewise.
1399 (clear_last_verified): Likewise.
1400 (verify_curr_properties): Likewise.
1401 (update_properties_after_pass): Likewise.
1402 (execute_pass_list_1): Split out from ...
1403 (execute_pass_list): ... here. Adjust.
1404 (execute_ipa_pass_list): Likewise.
1405 * cgraphunit.c (cgraph_add_new_function): Adjust.
1406 (analyze_function): Likewise.
1407 (expand_function): Likewise.
1408 * cgraph.c (release_function_body): Free dominance info
1409 here instead of asserting it was magically freed elsewhere.
1411 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
1413 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
1414 * configure: Regenerate.
1415 * config/sparc/sparc.opt (muser-mode): New option.
1416 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
1418 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
1419 * doc/invoke.texi (SPARC options): Document -muser-mode.
1421 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
1423 * cselib.c (find_slot_memmode): Delete.
1424 (cselib_hasher): Change compare_type to a struct.
1425 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
1427 (preserve_constants_and_equivs): Adjust for new compare_type.
1428 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
1429 (wrap_constant): Delete.
1430 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
1432 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
1434 * doc/install.texi (Building with profile feedback): Remove
1437 2014-04-26 Tom de Vries <tom@codesourcery.com>
1439 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
1442 2014-04-25 Cary Coutant <ccoutant@google.com>
1445 * dwarf2out.c (should_move_die_to_comdat): A type definition
1446 can contain a subprogram definition, but don't move it to a
1448 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
1449 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
1451 (clone_tree_hash): Rename to...
1452 (clone_tree_partial): ...this; change callers. Copy
1453 DW_TAG_subprogram DIEs as declarations.
1454 (copy_decls_walk): Don't copy children of a declaration into a
1457 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
1460 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
1463 2014-04-25 Jiong Wang <jiong.wang@arm.com>
1465 * config/arm/predicates.md (call_insn_operand): Add long_call check.
1466 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
1468 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
1471 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1473 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
1476 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1478 PR tree-optimization/60930
1479 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
1480 creating a multiply candidate by folding two constant
1481 multiplicands when the result overflows.
1483 2014-04-25 Jakub Jelinek <jakub@redhat.com>
1485 PR tree-optimization/60960
1486 * tree-vect-generic.c (expand_vector_operation): Only call
1487 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
1489 2014-04-25 Tom de Vries <tom@codesourcery.com>
1491 * expr.c (clobber_reg_mode): New function.
1492 * expr.h (clobber_reg): New function.
1494 2014-04-25 Tom de Vries <tom@codesourcery.com>
1496 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
1499 2014-04-25 Radovan Obradovic <robradovic@mips.com>
1500 Tom de Vries <tom@codesourcery.com>
1502 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
1504 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
1505 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
1506 new argument to find_all_hard_reg_sets call.
1508 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1510 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
1511 Use HOST_WIDE_INT_C for mask literal.
1512 (aarch_rev16_shleft_mask_imm_p): Likewise.
1514 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
1517 * config/sparc/sparc.md (ashlsi3_extend): Delete.
1519 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
1521 PR preprocessor/56540
1522 * config/i386/i386-c.c (ix86_target_macros): Define
1523 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
1525 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1527 * configure.ac (tga_func): Remove.
1528 (LIB_TLS_SPEC): Remove.
1529 * configure: Regenerate.
1530 * config.in: Regenerate.
1531 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
1533 2014-04-25 Richard Biener <rguenther@suse.de>
1536 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
1537 call stmt use/clobber sets during stmt walk instead of
1538 walking the possibly incomplete set of caller edges.
1540 2014-04-25 Richard Biener <rguenther@suse.de>
1543 * passes.c (apply_ipa_transforms): Inline into only caller ...
1544 (execute_one_pass): ... here. Properly bring in function
1545 bodies for nodes we want to apply IPA transforms to.
1547 2014-04-24 Cong Hou <congh@google.com>
1549 PR tree-optimization/60896
1550 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
1551 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
1552 (vect_mark_pattern_stmts): Set the def type of all statements in
1553 PATTERN_DEF_SEQ as vect_internal_def.
1555 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
1557 * doc/extend.texi (PowerPC Built-in Functions): Document new
1558 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
1560 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
1562 * config/rs6000/predicates.md (const_0_to_3_operand): New
1563 predicate to match 0..3 integer constants.
1565 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
1566 to support adding miscellaneous builtin functions.
1567 (BU_DFP_MISC_2): Likewise.
1568 (BU_P7_MISC_1): Likewise.
1569 (BU_P7_MISC_2): Likewise.
1570 (BU_P8V_MISC_3): Likewise.
1571 (BU_MISC_1): Likewise.
1572 (BU_MISC_2): Likewise.
1573 (DIVWE): Add extended divide builtin functions.
1576 (DIVWEUO): Likewise.
1580 (DIVDEUO): Likewise.
1581 (DXEX): Add decimal floating-point builtin functions.
1584 (DDEDPDQ): Likewise.
1586 (DENBCDQ): Likewise.
1593 (CDTBCD): Add new BCD builtin functions.
1597 (BCDADD_LT): Likewise.
1598 (BCDADD_EQ): Likewise.
1599 (BCDADD_GT): Likewise.
1600 (BCDADD_OV): Likewise.
1602 (BCDSUB_LT): Likewise.
1603 (BCDSUB_EQ): Likewise.
1604 (BCDSUB_GT): Likewise.
1605 (BCDSUB_OV): Likewise.
1606 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
1607 (UNPACK_TD): Likewise.
1608 (PACK_TF): Likewise.
1609 (UNPACK_TF): Likewise.
1610 (UNPACK_TF_0): Likewise.
1611 (UNPACK_TF_1): Likewise.
1612 (PACK_V1TI): Likewise.
1613 (UNPACK_V1TI): Likewise.
1615 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
1616 support for decimal floating point builtin functions.
1617 (rs6000_expand_ternop_builtin): Add checks for the new builtin
1618 functions that take constant arguments.
1619 (rs6000_invalid_builtin): Add decimal floating point builtin
1621 (rs6000_init_builtins): Setup long double, _Decimal64, and
1622 _Decimal128 types for new builtin functions.
1623 (builtin_function_type): Set the unsigned flags appropriately for
1624 the new builtin functions.
1625 (rs6000_opt_masks): Add support for decimal floating point builtin
1628 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
1629 floating point builtin functions.
1630 (RS6000_BTM_COMMON): Likewise.
1631 (RS6000_BTI_long_double): Likewise.
1632 (RS6000_BTI_dfloat64): Likewise.
1633 (RS6000_BTI_dfloat128): Likewise.
1634 (long_double_type_internal_node): Likewise.
1635 (dfloat64_type_internal_node): Likewise.
1636 (dfloat128_type_internal_node): Likewise.
1638 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
1639 2.07 bcd arithmetic instructions.
1640 (UNSPEC_BCDSUB): Likewise.
1641 (UNSPEC_BCD_OVERFLOW): Likewise.
1642 (UNSPEC_BCD_ADD_SUB): Likewise.
1643 (bcd_add_sub): Likewise.
1644 (BCD_TEST): Likewise.
1645 (bcd<bcd_add_sub>): Likewise.
1646 (bcd<bcd_add_sub>_test): Likewise.
1647 (bcd<bcd_add_sub>_test2): Likewise.
1648 (bcd<bcd_add_sub>_<code>): Likewise.
1649 (peephole2 for combined bcd ops): Likewise.
1651 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
1652 decimal floating point builtin functions.
1653 (UNSPEC_DENBCD): Likewise.
1654 (UNSPEC_DXEX): Likewise.
1655 (UNSPEC_DIEX): Likewise.
1656 (UNSPEC_DSCLI): Likewise.
1657 (UNSPEC_DSCRI): Likewise.
1658 (D64_D128): Likewise.
1659 (dfp_suffix): Likewise.
1660 (dfp_ddedpd_<mode>): Likewise.
1661 (dfp_denbcd_<mode>): Likewise.
1662 (dfp_dxex_<mode>): Likewise.
1663 (dfp_diex_<mode>): Likewise.
1664 (dfp_dscli_<mode>): Likewise.
1665 (dfp_dscri_<mode>): Likewise.
1667 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
1669 (UNSPEC_CDTBCD): Likewise.
1670 (UNSPEC_CBCDTD): Likewise.
1671 (UNSPEC_DIVE): Add support for new extended divide builtin
1673 (UNSPEC_DIVEO): Likewise.
1674 (UNSPEC_DIVEU): Likewise.
1675 (UNSPEC_DIVEUO): Likewise.
1676 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
1677 pack/unpack 128-bit types.
1678 (UNSPEC_PACK_128BIT): Likewise.
1679 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
1680 (udiv<mode>3): Use idiv_ldiv mode attribute.
1681 (div<mode>3): Likewise.
1682 (addg6s): Add new BCD builtin functions.
1685 (UNSPEC_DIV_EXTEND): Add support for new extended divide
1687 (div_extend): Likewise.
1688 (div<div_extend>_<mode>"): Likewise.
1689 (FP128_64): Add support for new builtin functions to pack/unpack
1691 (unpack<mode>): Likewise.
1692 (unpacktf_0): Likewise.
1693 (unpacktf_1): Likewise.
1694 (unpack<mode>_dm): Likewise.
1695 (unpack<mode>_nodm): Likewise.
1696 (pack<mode>): Likewise.
1697 (unpackv1ti): Likewise.
1698 (packv1ti): Likewise.
1700 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
1702 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
1705 2014-04-24 Jakub Jelinek <jakub@redhat.com>
1707 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
1708 * gimplify.c (omp_is_private): Change last argument's type to int.
1709 Only diagnose lastprivate if the simd argument is 1, only diagnose
1710 linear if the simd argument is 2.
1711 (gimplify_omp_for): Adjust omp_is_private callers. When adding
1712 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
1713 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
1714 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
1715 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
1716 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
1717 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
1718 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
1719 * tree-nested.c (convert_nonlocal_omp_clauses,
1720 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
1722 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
1725 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
1728 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
1730 * flag-types.h (enum ivar_visibility): Add.
1732 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
1734 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
1735 function * argument.
1737 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
1739 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
1741 2014-04-24 Radovan Obradovic <robradovic@mips.com>
1742 Tom de Vries <tom@codesourcery.com>
1744 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
1745 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
1747 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
1748 * emit-rtl.c (try_split): Same.
1750 2014-04-24 Radovan Obradovic <robradovic@mips.com>
1751 Tom de Vries <tom@codesourcery.com>
1753 * common.opt (fuse-caller-save): New option.
1755 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
1757 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
1760 2014-04-24 Richard Biener <rguenther@suse.de>
1762 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
1763 during TER and instead use the sepops interface for expanding
1764 non-GIMPLE_SINGLE_RHS.
1766 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1768 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
1769 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
1771 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1773 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
1774 assembler 64-bit option.
1775 * configure: Regenerate.
1777 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1779 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
1780 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
1781 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
1782 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
1783 (TARGET_CRYPTO): Take TARGET_SIMD into account.
1785 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1787 * config/aarch64/aarch64-builtins.c
1788 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
1789 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
1790 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
1791 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
1793 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
1794 (Vrevsuff): New mode attribute.
1796 2014-04-24 Terry Guo <terry.guo@arm.com>
1798 * config/arm/arm.h (machine_function): Define variable
1799 after_arm_reorg here.
1800 * config/arm/arm.c (after_arm_reorg): Remove the definition.
1801 (arm_split_constant): Update the way to access variable
1804 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
1806 2014-04-23 Tom de Vries <tom@codesourcery.com>
1808 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
1810 2014-04-23 David Malcolm <dmalcolm@redhat.com>
1812 * is-a.h: Update comments to reflect the following changes to the
1813 "pointerness" of the API, making the template parameter match the
1814 return type, allowing use of is-a.h with typedefs of pointers.
1815 (is_a_helper::cast): Return a T rather then a pointer to a T, so
1816 that the return type matches the parameter to the is_a_helper.
1818 (dyn_cast): Likewise.
1820 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
1821 pointer from the is-a.h API.
1823 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
1824 (is_a_helper <cgraph_node *>::test): ...this, matching change to
1826 (is_a_helper <varpool_node>::test): Likewise, convert to...
1827 (is_a_helper <varpool_node *>::test): ...this.
1829 (varpool_first_variable): Update for removal of implicit pointer
1830 from the is-a.h API.
1831 (varpool_next_variable): Likewise.
1832 (varpool_first_static_initializer): Likewise.
1833 (varpool_next_static_initializer): Likewise.
1834 (varpool_first_defined_variable): Likewise.
1835 (varpool_next_defined_variable): Likewise.
1836 (cgraph_first_defined_function): Likewise.
1837 (cgraph_next_defined_function): Likewise.
1838 (cgraph_first_function): Likewise.
1839 (cgraph_next_function): Likewise.
1840 (cgraph_first_function_with_gimple_body): Likewise.
1841 (cgraph_next_function_with_gimple_body): Likewise.
1842 (cgraph_alias_target): Likewise.
1843 (varpool_alias_target): Likewise.
1844 (cgraph_function_or_thunk_node): Likewise.
1845 (varpool_variable_node): Likewise.
1846 (symtab_real_symbol_p): Likewise.
1847 * cgraphunit.c (referred_to_p): Likewise.
1848 (analyze_functions): Likewise.
1849 (handle_alias_pairs): Likewise.
1850 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1851 * gimple-ssa.h (gimple_vuse_op): Likewise.
1852 (gimple_vdef_op): Likewise.
1853 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
1854 * gimple.c (gimple_build_asm_1): Likewise.
1855 (gimple_build_try): Likewise.
1856 (gimple_build_resx): Likewise.
1857 (gimple_build_eh_dispatch): Likewise.
1858 (gimple_build_omp_for): Likewise.
1859 (gimple_omp_for_set_clauses): Likewise.
1861 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
1862 (is_a_helper <gimple_statement_asm *>::test): ...this.
1863 (is_a_helper <gimple_statement_bind>::test): Convert to...
1864 (is_a_helper <gimple_statement_bind *>::test): ...this.
1865 (is_a_helper <gimple_statement_call>::test): Convert to...
1866 (is_a_helper <gimple_statement_call *>::test): ...this.
1867 (is_a_helper <gimple_statement_catch>::test): Convert to...
1868 (is_a_helper <gimple_statement_catch *>::test): ...this.
1869 (is_a_helper <gimple_statement_resx>::test): Convert to...
1870 (is_a_helper <gimple_statement_resx *>::test): ...this.
1871 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
1872 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
1873 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
1874 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
1875 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
1876 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
1877 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
1878 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
1879 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
1880 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
1881 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
1882 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
1883 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
1884 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
1885 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
1886 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
1887 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
1888 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
1889 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
1890 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
1891 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
1892 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
1893 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
1894 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
1895 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
1896 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
1897 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
1898 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
1899 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
1900 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
1901 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
1902 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
1903 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
1904 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
1905 (is_a_helper <gimple_statement_phi>::test): Convert to...
1906 (is_a_helper <gimple_statement_phi *>::test): ...this.
1907 (is_a_helper <gimple_statement_transaction>::test): Convert to...
1908 (is_a_helper <gimple_statement_transaction *>::test): ...this.
1909 (is_a_helper <gimple_statement_try>::test): Convert to...
1910 (is_a_helper <gimple_statement_try *>::test): ...this.
1911 (is_a_helper <gimple_statement_wce>::test): Convert to...
1912 (is_a_helper <gimple_statement_wce *>::test): ...this.
1913 (is_a_helper <const gimple_statement_asm>::test): Convert to...
1914 (is_a_helper <const gimple_statement_asm *>::test): ...this.
1915 (is_a_helper <const gimple_statement_bind>::test): Convert to...
1916 (is_a_helper <const gimple_statement_bind *>::test): ...this.
1917 (is_a_helper <const gimple_statement_call>::test): Convert to...
1918 (is_a_helper <const gimple_statement_call *>::test): ...this.
1919 (is_a_helper <const gimple_statement_catch>::test): Convert to...
1920 (is_a_helper <const gimple_statement_catch *>::test): ...this.
1921 (is_a_helper <const gimple_statement_resx>::test): Convert to...
1922 (is_a_helper <const gimple_statement_resx *>::test): ...this.
1923 (is_a_helper <const gimple_statement_eh_dispatch>::test):
1925 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
1927 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
1929 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
1930 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
1932 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
1934 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
1936 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
1938 (is_a_helper <const gimple_statement_omp_return>::test): Convert
1940 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
1941 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
1943 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
1944 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
1946 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
1947 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
1948 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
1949 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
1951 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
1952 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
1954 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
1955 (is_a_helper <const gimple_statement_omp_target>::test): Convert
1957 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
1958 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
1960 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
1961 (is_a_helper <const gimple_statement_omp_single>::test): Convert
1963 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
1964 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
1966 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
1967 (is_a_helper <const gimple_statement_omp_task>::test): Convert
1969 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
1970 (is_a_helper <const gimple_statement_phi>::test): Convert to...
1971 (is_a_helper <const gimple_statement_phi *>::test): ...this.
1972 (is_a_helper <const gimple_statement_transaction>::test): Convert
1974 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
1975 (is_a_helper <const gimple_statement_with_ops>::test): Convert
1977 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
1978 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
1979 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
1980 (is_a_helper <const gimple_statement_with_memory_ops>::test):
1982 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
1984 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
1986 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
1988 (gimple_use_ops): Update for removal of implicit pointer from the
1990 (gimple_set_use_ops): Likewise.
1991 (gimple_vuse): Likewise.
1992 (gimple_vdef): Likewise.
1993 (gimple_vuse_ptr): Likewise.
1994 (gimple_vdef_ptr): Likewise.
1995 (gimple_set_vuse): Likewise.
1996 (gimple_set_vdef): Likewise.
1997 (gimple_omp_return_set_lhs): Likewise.
1998 (gimple_omp_return_lhs): Likewise.
1999 (gimple_omp_return_lhs_ptr): Likewise.
2000 (gimple_call_fntype): Likewise.
2001 (gimple_call_set_fntype): Likewise.
2002 (gimple_call_set_internal_fn): Likewise.
2003 (gimple_call_use_set): Likewise.
2004 (gimple_call_clobber_set): Likewise.
2005 (gimple_bind_vars): Likewise.
2006 (gimple_bind_set_vars): Likewise.
2007 (gimple_bind_body_ptr): Likewise.
2008 (gimple_bind_set_body): Likewise.
2009 (gimple_bind_add_stmt): Likewise.
2010 (gimple_bind_block): Likewise.
2011 (gimple_bind_set_block): Likewise.
2012 (gimple_asm_ninputs): Likewise.
2013 (gimple_asm_noutputs): Likewise.
2014 (gimple_asm_nclobbers): Likewise.
2015 (gimple_asm_nlabels): Likewise.
2016 (gimple_asm_input_op): Likewise.
2017 (gimple_asm_input_op_ptr): Likewise.
2018 (gimple_asm_output_op): Likewise.
2019 (gimple_asm_output_op_ptr): Likewise.
2020 (gimple_asm_set_output_op): Likewise.
2021 (gimple_asm_clobber_op): Likewise.
2022 (gimple_asm_set_clobber_op): Likewise.
2023 (gimple_asm_label_op): Likewise.
2024 (gimple_asm_set_label_op): Likewise.
2025 (gimple_asm_string): Likewise.
2026 (gimple_catch_types): Likewise.
2027 (gimple_catch_types_ptr): Likewise.
2028 (gimple_catch_handler_ptr): Likewise.
2029 (gimple_catch_set_types): Likewise.
2030 (gimple_catch_set_handler): Likewise.
2031 (gimple_eh_filter_types): Likewise.
2032 (gimple_eh_filter_types_ptr): Likewise.
2033 (gimple_eh_filter_failure_ptr): Likewise.
2034 (gimple_eh_filter_set_types): Likewise.
2035 (gimple_eh_filter_set_failure): Likewise.
2036 (gimple_eh_must_not_throw_fndecl): Likewise.
2037 (gimple_eh_must_not_throw_set_fndecl): Likewise.
2038 (gimple_eh_else_n_body_ptr): Likewise.
2039 (gimple_eh_else_e_body_ptr): Likewise.
2040 (gimple_eh_else_set_n_body): Likewise.
2041 (gimple_eh_else_set_e_body): Likewise.
2042 (gimple_try_eval_ptr): Likewise.
2043 (gimple_try_cleanup_ptr): Likewise.
2044 (gimple_try_set_eval): Likewise.
2045 (gimple_try_set_cleanup): Likewise.
2046 (gimple_wce_cleanup_ptr): Likewise.
2047 (gimple_wce_set_cleanup): Likewise.
2048 (gimple_phi_capacity): Likewise.
2049 (gimple_phi_num_args): Likewise.
2050 (gimple_phi_result): Likewise.
2051 (gimple_phi_result_ptr): Likewise.
2052 (gimple_phi_set_result): Likewise.
2053 (gimple_phi_arg): Likewise.
2054 (gimple_phi_set_arg): Likewise.
2055 (gimple_resx_region): Likewise.
2056 (gimple_resx_set_region): Likewise.
2057 (gimple_eh_dispatch_region): Likewise.
2058 (gimple_eh_dispatch_set_region): Likewise.
2059 (gimple_omp_critical_name): Likewise.
2060 (gimple_omp_critical_name_ptr): Likewise.
2061 (gimple_omp_critical_set_name): Likewise.
2062 (gimple_omp_for_clauses): Likewise.
2063 (gimple_omp_for_clauses_ptr): Likewise.
2064 (gimple_omp_for_set_clauses): Likewise.
2065 (gimple_omp_for_collapse): Likewise.
2066 (gimple_omp_for_index): Likewise.
2067 (gimple_omp_for_index_ptr): Likewise.
2068 (gimple_omp_for_set_index): Likewise.
2069 (gimple_omp_for_initial): Likewise.
2070 (gimple_omp_for_initial_ptr): Likewise.
2071 (gimple_omp_for_set_initial): Likewise.
2072 (gimple_omp_for_final): Likewise.
2073 (gimple_omp_for_final_ptr): Likewise.
2074 (gimple_omp_for_set_final): Likewise.
2075 (gimple_omp_for_incr): Likewise.
2076 (gimple_omp_for_incr_ptr): Likewise.
2077 (gimple_omp_for_set_incr): Likewise.
2078 (gimple_omp_for_pre_body_ptr): Likewise.
2079 (gimple_omp_for_set_pre_body): Likewise.
2080 (gimple_omp_parallel_clauses): Likewise.
2081 (gimple_omp_parallel_clauses_ptr): Likewise.
2082 (gimple_omp_parallel_set_clauses): Likewise.
2083 (gimple_omp_parallel_child_fn): Likewise.
2084 (gimple_omp_parallel_child_fn_ptr): Likewise.
2085 (gimple_omp_parallel_set_child_fn): Likewise.
2086 (gimple_omp_parallel_data_arg): Likewise.
2087 (gimple_omp_parallel_data_arg_ptr): Likewise.
2088 (gimple_omp_parallel_set_data_arg): Likewise.
2089 (gimple_omp_task_clauses): Likewise.
2090 (gimple_omp_task_clauses_ptr): Likewise.
2091 (gimple_omp_task_set_clauses): Likewise.
2092 (gimple_omp_task_child_fn): Likewise.
2093 (gimple_omp_task_child_fn_ptr): Likewise.
2094 (gimple_omp_task_set_child_fn): Likewise.
2095 (gimple_omp_task_data_arg): Likewise.
2096 (gimple_omp_task_data_arg_ptr): Likewise.
2097 (gimple_omp_task_set_data_arg): Likewise.
2098 (gimple_omp_taskreg_clauses): Likewise.
2099 (gimple_omp_taskreg_clauses_ptr): Likewise.
2100 (gimple_omp_taskreg_set_clauses): Likewise.
2101 (gimple_omp_taskreg_child_fn): Likewise.
2102 (gimple_omp_taskreg_child_fn_ptr): Likewise.
2103 (gimple_omp_taskreg_set_child_fn): Likewise.
2104 (gimple_omp_taskreg_data_arg): Likewise.
2105 (gimple_omp_taskreg_data_arg_ptr): Likewise.
2106 (gimple_omp_taskreg_set_data_arg): Likewise.
2107 (gimple_omp_task_copy_fn): Likewise.
2108 (gimple_omp_task_copy_fn_ptr): Likewise.
2109 (gimple_omp_task_set_copy_fn): Likewise.
2110 (gimple_omp_task_arg_size): Likewise.
2111 (gimple_omp_task_arg_size_ptr): Likewise.
2112 (gimple_omp_task_set_arg_size): Likewise.
2113 (gimple_omp_task_arg_align): Likewise.
2114 (gimple_omp_task_arg_align_ptr): Likewise.
2115 (gimple_omp_task_set_arg_align): Likewise.
2116 (gimple_omp_single_clauses): Likewise.
2117 (gimple_omp_single_clauses_ptr): Likewise.
2118 (gimple_omp_single_set_clauses): Likewise.
2119 (gimple_omp_target_clauses): Likewise.
2120 (gimple_omp_target_clauses_ptr): Likewise.
2121 (gimple_omp_target_set_clauses): Likewise.
2122 (gimple_omp_target_child_fn): Likewise.
2123 (gimple_omp_target_child_fn_ptr): Likewise.
2124 (gimple_omp_target_set_child_fn): Likewise.
2125 (gimple_omp_target_data_arg): Likewise.
2126 (gimple_omp_target_data_arg_ptr): Likewise.
2127 (gimple_omp_target_set_data_arg): Likewise.
2128 (gimple_omp_teams_clauses): Likewise.
2129 (gimple_omp_teams_clauses_ptr): Likewise.
2130 (gimple_omp_teams_set_clauses): Likewise.
2131 (gimple_omp_sections_clauses): Likewise.
2132 (gimple_omp_sections_clauses_ptr): Likewise.
2133 (gimple_omp_sections_set_clauses): Likewise.
2134 (gimple_omp_sections_control): Likewise.
2135 (gimple_omp_sections_control_ptr): Likewise.
2136 (gimple_omp_sections_set_control): Likewise.
2137 (gimple_omp_for_set_cond): Likewise.
2138 (gimple_omp_for_cond): Likewise.
2139 (gimple_omp_atomic_store_set_val): Likewise.
2140 (gimple_omp_atomic_store_val): Likewise.
2141 (gimple_omp_atomic_store_val_ptr): Likewise.
2142 (gimple_omp_atomic_load_set_lhs): Likewise.
2143 (gimple_omp_atomic_load_lhs): Likewise.
2144 (gimple_omp_atomic_load_lhs_ptr): Likewise.
2145 (gimple_omp_atomic_load_set_rhs): Likewise.
2146 (gimple_omp_atomic_load_rhs): Likewise.
2147 (gimple_omp_atomic_load_rhs_ptr): Likewise.
2148 (gimple_omp_continue_control_def): Likewise.
2149 (gimple_omp_continue_control_def_ptr): Likewise.
2150 (gimple_omp_continue_set_control_def): Likewise.
2151 (gimple_omp_continue_control_use): Likewise.
2152 (gimple_omp_continue_control_use_ptr): Likewise.
2153 (gimple_omp_continue_set_control_use): Likewise.
2154 (gimple_transaction_body_ptr): Likewise.
2155 (gimple_transaction_label): Likewise.
2156 (gimple_transaction_label_ptr): Likewise.
2157 (gimple_transaction_set_body): Likewise.
2158 (gimple_transaction_set_label): Likewise.
2160 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
2161 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2162 * ipa-ref.c (ipa_record_reference): Likewise.
2163 * ipa-reference.c (analyze_function): Likewise.
2164 (ipa_reference_write_optimization_summary): Likewise.
2165 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
2166 (address_taken_from_non_vtable_p): Likewise.
2167 (comdat_can_be_unshared_p_1): Likewise.
2168 * lto-cgraph.c (lto_output_ref): Likewise.
2169 (add_references): Likewise.
2170 (compute_ltrans_boundary): Likewise.
2171 (output_symtab): Likewise.
2172 (input_ref): Likewise.
2173 (input_cgraph_1): Likewise.
2174 (output_cgraph_opt_summary): Likewise.
2175 * lto-streamer-out.c (lto_output): Likewise.
2176 (output_symbol_p): Likewise.
2177 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
2178 (lsei_start_function_in_partition): Likewise.
2179 (lsei_next_variable_in_partition): Likewise.
2180 (lsei_start_variable_in_partition): Likewise.
2181 * symtab.c (insert_to_assembler_name_hash): Likewise.
2182 (unlink_from_assembler_name_hash): Likewise.
2183 (symtab_unregister_node): Likewise.
2184 (symtab_remove_node): Likewise.
2185 (dump_symtab_node): Likewise.
2186 (verify_symtab_base): Likewise.
2187 (verify_symtab_node): Likewise.
2188 (symtab_make_decl_local): Likewise.
2189 (symtab_alias_ultimate_target): Likewise.
2190 (symtab_resolve_alias): Likewise.
2191 (symtab_get_symbol_partitioning_class): Likewise.
2192 * tree-phinodes.c (allocate_phi_node): Likewise.
2193 (reserve_phi_args_for_new_edge): Likewise.
2194 (remove_phi_args): Likewise.
2195 * varpool.c (varpool_node_for_asm): Likewise.
2196 (varpool_remove_unreferenced_decls): Likewise.
2198 2014-04-23 Jeff Law <law@redhat.com>
2200 PR tree-optimization/60902
2201 * tree-ssa-threadedge.c
2202 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
2203 invalidate outputs from statements that do not produce useful
2204 outputs for threading.
2206 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2208 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
2209 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
2210 machine descriptions for Stack Smashing Protector.
2212 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
2214 * aarch64.md (<optab>_rol<mode>3): New pattern.
2215 (<optab>_rolsi3_uxtw): Likewise.
2216 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
2218 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
2220 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
2221 (arm_cortex_a12_tune): Likewise.
2223 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2225 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
2227 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2229 * config/arm/arm.md (arm_rev16si2): New pattern.
2230 (arm_rev16si2_alt): Likewise.
2231 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
2233 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2235 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
2236 (rev16<mode>2_alt): Likewise.
2237 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
2238 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
2239 (aarch_rev16_shleft_mask_imm_p): Likewise.
2240 (aarch_rev16_p_1): Likewise.
2241 (aarch_rev16_p): Likewise.
2242 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
2243 (aarch_rev16_shright_mask_imm_p): Likewise.
2244 (aarch_rev16_shleft_mask_imm_p): Likewise.
2246 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2248 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
2249 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
2251 (cortex_a53_extra_costs): Likewise.
2252 (cortex_a57_extra_costs): Likewise.
2253 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
2254 (cortexa7_extra_costs): Likewise.
2255 (cortexa8_extra_costs): Likewise.
2256 (cortexa12_extra_costs): Likewise.
2257 (cortexa15_extra_costs): Likewise.
2258 (v7m_extra_costs): Likewise.
2259 (arm_new_rtx_costs): Handle BSWAP.
2261 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2263 * config/arm/arm.c (cortexa8_extra_costs): New table.
2264 (arm_cortex_a8_tune): New tuning struct.
2265 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
2267 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2269 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
2271 2014-04-23 Richard Biener <rguenther@suse.de>
2273 * Makefile.in (OBJS): Remove loop-unswitch.o.
2274 * loop-unswitch.c: Delete.
2275 * tree-pass.h (make_pass_rtl_unswitch): Remove.
2276 * passes.def (pass_rtl_unswitch): Likewise.
2277 * loop-init.c (gate_rtl_unswitch): Likewise.
2278 (rtl_unswitch): Likewise.
2279 (pass_data_rtl_unswitch): Likewise.
2280 (pass_rtl_unswitch): Likewise.
2281 (make_pass_rtl_unswitch): Likewise.
2282 * rtl.h (reversed_condition): Likewise.
2283 (compare_and_jump_seq): Likewise.
2284 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
2286 * loop-unroll.c (compare_and_jump_seq): Likewise.
2288 2014-04-23 Richard Biener <rguenther@suse.de>
2290 PR tree-optimization/60903
2291 * tree-ssa-loop-im.c (analyze_memory_references): Remove
2292 commented code block.
2293 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
2294 loop flags to newly created BBs and edges.
2296 2014-04-23 Nick Clifton <nickc@redhat.com>
2298 * config/msp430/msp430.c (msp430_handle_option): Move function
2300 (msp430_option_override): Simplify mcu and mcpu option handling.
2301 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
2302 support for -mhwmult command line option.
2303 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
2304 -mhwmult command line option.
2305 (msp430_hwmult_enabled): Delete.
2306 (msp43o_output_labelref): Add support for -mhwmult command line
2308 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
2309 (umulsidi3): Likewise.
2310 * config/msp430/msp430.opt (mmcu): Add Report attribute.
2311 (mcpu, mlarge, msmall): Likewise.
2312 (mhwmult): New option.
2313 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
2315 (msp430_is_f5_mcu): Remove prototype.
2316 (msp430_use_f5_series_hwmult): Add prototype.
2317 * config/msp430/msp430-opts.h: New file.
2318 * common/config/msp430: New directory.
2319 * common/config/msp430/msp430-common.c: New file.
2320 * config.gcc (msp430): Remove target_has_targetm_common.
2321 * doc/invoke.texi: Document -mhwmult command line option.
2323 2014-04-23 Nick Clifton <nickc@redhat.com>
2325 * config/i386/cygwin.h (ENDFILE_SPEC): Include
2326 default-manifest.o if it can be found in the search path.
2327 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
2329 2014-04-23 Terry Guo <terry.guo@arm.com>
2331 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
2333 2014-04-23 Richard Biener <rguenther@suse.de>
2336 * tree-inline.c (declare_return_variable): Use mark_addressable.
2338 2014-04-23 Richard Biener <rguenther@suse.de>
2341 * loop-init.c (loop_optimizer_init): Make sure to apply
2342 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
2344 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2347 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
2349 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
2350 if flag_sanitize_undefined_trap_on_error.
2351 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
2352 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
2353 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
2354 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
2355 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
2356 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
2357 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
2358 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
2359 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
2360 * ubsan.c (ubsan_instrument_unreachable): Return
2361 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
2362 (ubsan_expand_null_ifn): Emit __builtin_trap ()
2363 if flag_sanitize_undefined_trap_on_error and
2364 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
2365 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
2366 instrument_bool_enum_load): Emit __builtin_trap () if
2367 flag_sanitize_undefined_trap_on_error and
2368 __builtin_handle_*_abort () if !flag_sanitize_recover.
2369 * doc/invoke.texi (-fsanitize-recover,
2370 -fsanitize-undefined-trap-on-error): Document.
2372 2014-04-22 Christian Bruel <christian.bruel@st.com>
2374 * config/sh/sh.md (mov<mode>): Replace movQIHI.
2375 Force immediates to SImode.
2377 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
2379 * config/nios2/nios2.md (UNSPEC_ROUND): New.
2381 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
2382 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
2383 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
2384 (nios2_fpu_insn): Add entry for round.
2385 (N2FPU_NO_ERRNO_P): Define.
2386 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
2388 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
2390 2014-04-22 Richard Henderson <rth@redhat.com>
2392 * config/aarch64/aarch64 (addti3, subti3): New expanders.
2393 (add<GPI>3_compare0): Remove leading * from name.
2394 (add<GPI>3_carryin): Likewise.
2395 (sub<GPI>3_compare0): Likewise.
2396 (sub<GPI>3_carryin): Likewise.
2397 (<su_optab>mulditi3): New expander.
2398 (multi3): New expander.
2399 (madd<GPI>): Remove leading * from name.
2401 2014-04-22 Martin Jambor <mjambor@suse.cz>
2403 * cgraphclones.c (cgraph_function_versioning): Copy
2404 ipa_transforms_to_apply instead of asserting it is empty.
2406 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
2409 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
2410 on count_exp to get mode.
2412 2014-04-22 Andrew Pinski <apinski@cavium.com>
2414 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
2415 Handle TLS for ILP32.
2416 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
2417 (tlsie_small_<mode>): this and handle PTR.
2418 (tlsie_small_sidi): New pattern.
2419 (tlsle_small): Change to an expand to handle ILP32.
2420 (tlsle_small_<mode>): New pattern.
2421 (tlsdesc_small): Rename to ...
2422 (tlsdesc_small_<mode>): this and handle PTR.
2424 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2426 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
2428 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2430 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2431 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
2432 (aarch64_types_signed_poly_qualifiers): Likewise.
2433 (aarch64_types_unsigned_signed_qualifiers): Likewise.
2434 (aarch64_types_poly_signed_qualifiers): Likewise.
2435 (TYPES_REINTERP_SS): Type macro added.
2436 (TYPES_REINTERP_SU): Likewise.
2437 (TYPES_REINTERP_SP): Likewise.
2438 (TYPES_REINTERP_US): Likewise.
2439 (TYPES_REINTERP_PS): Likewise.
2440 (aarch64_fold_builtin): New expression folding added.
2441 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
2442 Declarations removed.
2443 (REINTERP_SS): Declarations added.
2444 (REINTERP_US): Likewise.
2445 (REINTERP_PS): Likewise.
2446 (REINTERP_SU): Likewise.
2447 (REINTERP_SP): Likewise.
2448 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
2449 (vreinterpretq_p8_f64): Likewise.
2450 (vreinterpret_p16_f64): Likewise.
2451 (vreinterpretq_p16_f64): Likewise.
2452 (vreinterpret_f32_f64): Likewise.
2453 (vreinterpretq_f32_f64): Likewise.
2454 (vreinterpret_f64_f32): Likewise.
2455 (vreinterpret_f64_p8): Likewise.
2456 (vreinterpret_f64_p16): Likewise.
2457 (vreinterpret_f64_s8): Likewise.
2458 (vreinterpret_f64_s16): Likewise.
2459 (vreinterpret_f64_s32): Likewise.
2460 (vreinterpret_f64_s64): Likewise.
2461 (vreinterpret_f64_u8): Likewise.
2462 (vreinterpret_f64_u16): Likewise.
2463 (vreinterpret_f64_u32): Likewise.
2464 (vreinterpret_f64_u64): Likewise.
2465 (vreinterpretq_f64_f32): Likewise.
2466 (vreinterpretq_f64_p8): Likewise.
2467 (vreinterpretq_f64_p16): Likewise.
2468 (vreinterpretq_f64_s8): Likewise.
2469 (vreinterpretq_f64_s16): Likewise.
2470 (vreinterpretq_f64_s32): Likewise.
2471 (vreinterpretq_f64_s64): Likewise.
2472 (vreinterpretq_f64_u8): Likewise.
2473 (vreinterpretq_f64_u16): Likewise.
2474 (vreinterpretq_f64_u32): Likewise.
2475 (vreinterpretq_f64_u64): Likewise.
2476 (vreinterpret_s64_f64): Likewise.
2477 (vreinterpretq_s64_f64): Likewise.
2478 (vreinterpret_u64_f64): Likewise.
2479 (vreinterpretq_u64_f64): Likewise.
2480 (vreinterpret_s8_f64): Likewise.
2481 (vreinterpretq_s8_f64): Likewise.
2482 (vreinterpret_s16_f64): Likewise.
2483 (vreinterpretq_s16_f64): Likewise.
2484 (vreinterpret_s32_f64): Likewise.
2485 (vreinterpretq_s32_f64): Likewise.
2486 (vreinterpret_u8_f64): Likewise.
2487 (vreinterpretq_u8_f64): Likewise.
2488 (vreinterpret_u16_f64): Likewise.
2489 (vreinterpretq_u16_f64): Likewise.
2490 (vreinterpret_u32_f64): Likewise.
2491 (vreinterpretq_u32_f64): Likewise.
2493 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2495 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2496 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
2497 (vreinterpret_p8_s8): Likewise.
2498 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
2499 (vreinterpret_p8_s16): Likewise.
2500 (vreinterpret_p8_s32): Likewise.
2501 (vreinterpret_p8_s64): Likewise.
2502 (vreinterpret_p8_f32): Likewise.
2503 (vreinterpret_p8_u8): Likewise.
2504 (vreinterpret_p8_u16): Likewise.
2505 (vreinterpret_p8_u32): Likewise.
2506 (vreinterpret_p8_u64): Likewise.
2507 (vreinterpret_p8_p16): Likewise.
2508 (vreinterpretq_p8_s8): Likewise.
2509 (vreinterpretq_p8_s16): Likewise.
2510 (vreinterpretq_p8_s32): Likewise.
2511 (vreinterpretq_p8_s64): Likewise.
2512 (vreinterpretq_p8_f32): Likewise.
2513 (vreinterpretq_p8_u8): Likewise.
2514 (vreinterpretq_p8_u16): Likewise.
2515 (vreinterpretq_p8_u32): Likewise.
2516 (vreinterpretq_p8_u64): Likewise.
2517 (vreinterpretq_p8_p16): Likewise.
2518 (vreinterpret_p16_s8): Likewise.
2519 (vreinterpret_p16_s16): Likewise.
2520 (vreinterpret_p16_s32): Likewise.
2521 (vreinterpret_p16_s64): Likewise.
2522 (vreinterpret_p16_f32): Likewise.
2523 (vreinterpret_p16_u8): Likewise.
2524 (vreinterpret_p16_u16): Likewise.
2525 (vreinterpret_p16_u32): Likewise.
2526 (vreinterpret_p16_u64): Likewise.
2527 (vreinterpret_p16_p8): Likewise.
2528 (vreinterpretq_p16_s8): Likewise.
2529 (vreinterpretq_p16_s16): Likewise.
2530 (vreinterpretq_p16_s32): Likewise.
2531 (vreinterpretq_p16_s64): Likewise.
2532 (vreinterpretq_p16_f32): Likewise.
2533 (vreinterpretq_p16_u8): Likewise.
2534 (vreinterpretq_p16_u16): Likewise.
2535 (vreinterpretq_p16_u32): Likewise.
2536 (vreinterpretq_p16_u64): Likewise.
2537 (vreinterpretq_p16_p8): Likewise.
2538 (vreinterpret_f32_s8): Likewise.
2539 (vreinterpret_f32_s16): Likewise.
2540 (vreinterpret_f32_s32): Likewise.
2541 (vreinterpret_f32_s64): Likewise.
2542 (vreinterpret_f32_u8): Likewise.
2543 (vreinterpret_f32_u16): Likewise.
2544 (vreinterpret_f32_u32): Likewise.
2545 (vreinterpret_f32_u64): Likewise.
2546 (vreinterpret_f32_p8): Likewise.
2547 (vreinterpret_f32_p16): Likewise.
2548 (vreinterpretq_f32_s8): Likewise.
2549 (vreinterpretq_f32_s16): Likewise.
2550 (vreinterpretq_f32_s32): Likewise.
2551 (vreinterpretq_f32_s64): Likewise.
2552 (vreinterpretq_f32_u8): Likewise.
2553 (vreinterpretq_f32_u16): Likewise.
2554 (vreinterpretq_f32_u32): Likewise.
2555 (vreinterpretq_f32_u64): Likewise.
2556 (vreinterpretq_f32_p8): Likewise.
2557 (vreinterpretq_f32_p16): Likewise.
2558 (vreinterpret_s64_s8): Likewise.
2559 (vreinterpret_s64_s16): Likewise.
2560 (vreinterpret_s64_s32): Likewise.
2561 (vreinterpret_s64_f32): Likewise.
2562 (vreinterpret_s64_u8): Likewise.
2563 (vreinterpret_s64_u16): Likewise.
2564 (vreinterpret_s64_u32): Likewise.
2565 (vreinterpret_s64_u64): Likewise.
2566 (vreinterpret_s64_p8): Likewise.
2567 (vreinterpret_s64_p16): Likewise.
2568 (vreinterpretq_s64_s8): Likewise.
2569 (vreinterpretq_s64_s16): Likewise.
2570 (vreinterpretq_s64_s32): Likewise.
2571 (vreinterpretq_s64_f32): Likewise.
2572 (vreinterpretq_s64_u8): Likewise.
2573 (vreinterpretq_s64_u16): Likewise.
2574 (vreinterpretq_s64_u32): Likewise.
2575 (vreinterpretq_s64_u64): Likewise.
2576 (vreinterpretq_s64_p8): Likewise.
2577 (vreinterpretq_s64_p16): Likewise.
2578 (vreinterpret_u64_s8): Likewise.
2579 (vreinterpret_u64_s16): Likewise.
2580 (vreinterpret_u64_s32): Likewise.
2581 (vreinterpret_u64_s64): Likewise.
2582 (vreinterpret_u64_f32): Likewise.
2583 (vreinterpret_u64_u8): Likewise.
2584 (vreinterpret_u64_u16): Likewise.
2585 (vreinterpret_u64_u32): Likewise.
2586 (vreinterpret_u64_p8): Likewise.
2587 (vreinterpret_u64_p16): Likewise.
2588 (vreinterpretq_u64_s8): Likewise.
2589 (vreinterpretq_u64_s16): Likewise.
2590 (vreinterpretq_u64_s32): Likewise.
2591 (vreinterpretq_u64_s64): Likewise.
2592 (vreinterpretq_u64_f32): Likewise.
2593 (vreinterpretq_u64_u8): Likewise.
2594 (vreinterpretq_u64_u16): Likewise.
2595 (vreinterpretq_u64_u32): Likewise.
2596 (vreinterpretq_u64_p8): Likewise.
2597 (vreinterpretq_u64_p16): Likewise.
2598 (vreinterpret_s8_s16): Likewise.
2599 (vreinterpret_s8_s32): Likewise.
2600 (vreinterpret_s8_s64): Likewise.
2601 (vreinterpret_s8_f32): Likewise.
2602 (vreinterpret_s8_u8): Likewise.
2603 (vreinterpret_s8_u16): Likewise.
2604 (vreinterpret_s8_u32): Likewise.
2605 (vreinterpret_s8_u64): Likewise.
2606 (vreinterpret_s8_p8): Likewise.
2607 (vreinterpret_s8_p16): Likewise.
2608 (vreinterpretq_s8_s16): Likewise.
2609 (vreinterpretq_s8_s32): Likewise.
2610 (vreinterpretq_s8_s64): Likewise.
2611 (vreinterpretq_s8_f32): Likewise.
2612 (vreinterpretq_s8_u8): Likewise.
2613 (vreinterpretq_s8_u16): Likewise.
2614 (vreinterpretq_s8_u32): Likewise.
2615 (vreinterpretq_s8_u64): Likewise.
2616 (vreinterpretq_s8_p8): Likewise.
2617 (vreinterpretq_s8_p16): Likewise.
2618 (vreinterpret_s16_s8): Likewise.
2619 (vreinterpret_s16_s32): Likewise.
2620 (vreinterpret_s16_s64): Likewise.
2621 (vreinterpret_s16_f32): Likewise.
2622 (vreinterpret_s16_u8): Likewise.
2623 (vreinterpret_s16_u16): Likewise.
2624 (vreinterpret_s16_u32): Likewise.
2625 (vreinterpret_s16_u64): Likewise.
2626 (vreinterpret_s16_p8): Likewise.
2627 (vreinterpret_s16_p16): Likewise.
2628 (vreinterpretq_s16_s8): Likewise.
2629 (vreinterpretq_s16_s32): Likewise.
2630 (vreinterpretq_s16_s64): Likewise.
2631 (vreinterpretq_s16_f32): Likewise.
2632 (vreinterpretq_s16_u8): Likewise.
2633 (vreinterpretq_s16_u16): Likewise.
2634 (vreinterpretq_s16_u32): Likewise.
2635 (vreinterpretq_s16_u64): Likewise.
2636 (vreinterpretq_s16_p8): Likewise.
2637 (vreinterpretq_s16_p16): Likewise.
2638 (vreinterpret_s32_s8): Likewise.
2639 (vreinterpret_s32_s16): Likewise.
2640 (vreinterpret_s32_s64): Likewise.
2641 (vreinterpret_s32_f32): Likewise.
2642 (vreinterpret_s32_u8): Likewise.
2643 (vreinterpret_s32_u16): Likewise.
2644 (vreinterpret_s32_u32): Likewise.
2645 (vreinterpret_s32_u64): Likewise.
2646 (vreinterpret_s32_p8): Likewise.
2647 (vreinterpret_s32_p16): Likewise.
2648 (vreinterpretq_s32_s8): Likewise.
2649 (vreinterpretq_s32_s16): Likewise.
2650 (vreinterpretq_s32_s64): Likewise.
2651 (vreinterpretq_s32_f32): Likewise.
2652 (vreinterpretq_s32_u8): Likewise.
2653 (vreinterpretq_s32_u16): Likewise.
2654 (vreinterpretq_s32_u32): Likewise.
2655 (vreinterpretq_s32_u64): Likewise.
2656 (vreinterpretq_s32_p8): Likewise.
2657 (vreinterpretq_s32_p16): Likewise.
2658 (vreinterpret_u8_s8): Likewise.
2659 (vreinterpret_u8_s16): Likewise.
2660 (vreinterpret_u8_s32): Likewise.
2661 (vreinterpret_u8_s64): Likewise.
2662 (vreinterpret_u8_f32): Likewise.
2663 (vreinterpret_u8_u16): Likewise.
2664 (vreinterpret_u8_u32): Likewise.
2665 (vreinterpret_u8_u64): Likewise.
2666 (vreinterpret_u8_p8): Likewise.
2667 (vreinterpret_u8_p16): Likewise.
2668 (vreinterpretq_u8_s8): Likewise.
2669 (vreinterpretq_u8_s16): Likewise.
2670 (vreinterpretq_u8_s32): Likewise.
2671 (vreinterpretq_u8_s64): Likewise.
2672 (vreinterpretq_u8_f32): Likewise.
2673 (vreinterpretq_u8_u16): Likewise.
2674 (vreinterpretq_u8_u32): Likewise.
2675 (vreinterpretq_u8_u64): Likewise.
2676 (vreinterpretq_u8_p8): Likewise.
2677 (vreinterpretq_u8_p16): Likewise.
2678 (vreinterpret_u16_s8): Likewise.
2679 (vreinterpret_u16_s16): Likewise.
2680 (vreinterpret_u16_s32): Likewise.
2681 (vreinterpret_u16_s64): Likewise.
2682 (vreinterpret_u16_f32): Likewise.
2683 (vreinterpret_u16_u8): Likewise.
2684 (vreinterpret_u16_u32): Likewise.
2685 (vreinterpret_u16_u64): Likewise.
2686 (vreinterpret_u16_p8): Likewise.
2687 (vreinterpret_u16_p16): Likewise.
2688 (vreinterpretq_u16_s8): Likewise.
2689 (vreinterpretq_u16_s16): Likewise.
2690 (vreinterpretq_u16_s32): Likewise.
2691 (vreinterpretq_u16_s64): Likewise.
2692 (vreinterpretq_u16_f32): Likewise.
2693 (vreinterpretq_u16_u8): Likewise.
2694 (vreinterpretq_u16_u32): Likewise.
2695 (vreinterpretq_u16_u64): Likewise.
2696 (vreinterpretq_u16_p8): Likewise.
2697 (vreinterpretq_u16_p16): Likewise.
2698 (vreinterpret_u32_s8): Likewise.
2699 (vreinterpret_u32_s16): Likewise.
2700 (vreinterpret_u32_s32): Likewise.
2701 (vreinterpret_u32_s64): Likewise.
2702 (vreinterpret_u32_f32): Likewise.
2703 (vreinterpret_u32_u8): Likewise.
2704 (vreinterpret_u32_u16): Likewise.
2705 (vreinterpret_u32_u64): Likewise.
2706 (vreinterpret_u32_p8): Likewise.
2707 (vreinterpret_u32_p16): Likewise.
2708 (vreinterpretq_u32_s8): Likewise.
2709 (vreinterpretq_u32_s16): Likewise.
2710 (vreinterpretq_u32_s32): Likewise.
2711 (vreinterpretq_u32_s64): Likewise.
2712 (vreinterpretq_u32_f32): Likewise.
2713 (vreinterpretq_u32_u8): Likewise.
2714 (vreinterpretq_u32_u16): Likewise.
2715 (vreinterpretq_u32_u64): Likewise.
2716 (vreinterpretq_u32_p8): Likewise.
2717 (vreinterpretq_u32_p16): Likewise.
2719 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2721 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
2723 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
2726 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
2727 (vqnegd_s64): Likewise.
2728 (vqabs_s64): Likewise.
2729 (vqabsd_s64): Likewise.
2731 2014-04-22 Richard Henderson <rth@redhat.com>
2733 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
2734 computation to the top of the loop.
2736 2014-04-22 Renlin <renlin.li@arm.com>
2737 Jiong Wang <jiong.wang@arm.com>
2739 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
2740 * config/aarch64/aarch64.c (aarch64_layout_frame)
2741 (aarch64_initial_elimination_offset): Likewise.
2743 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
2745 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
2748 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
2750 * machmode.h (bitwise_mode_for_mode): Declare.
2751 * stor-layout.h (bitwise_type_for_mode): Likewise.
2752 * stor-layout.c (bitwise_mode_for_mode): New function.
2753 (bitwise_type_for_mode): Likewise.
2754 * builtins.c (fold_builtin_memory_op): Use it instead of
2755 int_mode_for_mode and build_nonstandard_integer_type.
2757 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2759 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
2760 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
2761 (*-*-solaris2*): Simplify.
2762 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
2763 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
2764 *-*-solaris2.9* handling.
2766 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
2768 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
2769 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
2771 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
2772 * configure: Regenerate.
2774 * config/i386/sol2-9.h: Remove.
2776 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
2777 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
2778 Remove Solaris 9 references.
2780 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
2782 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
2783 (floatuns<GPI:mode><GPF:mode>2): Remove.
2784 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
2785 and floatuns conversions.
2786 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
2787 and floatuns conversions.
2788 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
2789 (w1,w2): New mode attributes for inequal width conversions.
2791 2014-04-22 Renlin Li <Renlin.Li@arm.com>
2793 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
2794 the output asm format.
2796 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
2798 * config/aarch64/aarch64-simd.md
2799 (aarch64_cm<optab>di): Always split.
2800 (*aarch64_cm<optab>di): New.
2801 (aarch64_cmtstdi): Always split.
2802 (*aarch64_cmtstdi): New.
2804 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2806 PR tree-optimization/60823
2807 * omp-low.c (ipa_simd_modify_function_body): Go through
2808 all SSA_NAMEs and for those refering to vector arguments
2809 which are going to be replaced adjust SSA_NAME_VAR and,
2810 if it is a default definition, change it into a non-default
2811 definition assigned at the beginning of function from new_decl.
2812 (ipa_simd_modify_stmt_ops): Rewritten.
2813 * tree-dfa.c (set_ssa_default_def): When removing default def,
2814 check for NULL loc instead of NULL *loc.
2816 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2818 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
2819 restrictions on core registers for DImode values in Thumb2.
2821 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2823 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
2824 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
2826 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2828 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
2829 (*iordi_notzesidi_di): Likewise.
2830 (*iordi_notsesidi_di): Likewise.
2832 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2834 * config/arm/arm-protos.h (tune_params): New struct members.
2835 * config/arm/arm.c: Initialise tune_params per processor.
2836 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
2837 for speed, based on new tune_params.
2839 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2841 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
2843 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
2845 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
2847 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
2848 * config/aarch64/arm_neon.h (vrnd_f64): Added.
2849 (vrnda_f64): Likewise.
2850 (vrndi_f64): Likewise.
2851 (vrndm_f64): Likewise.
2852 (vrndn_f64): Likewise.
2853 (vrndp_f64): Likewise.
2854 (vrndx_f64): Likewise.
2856 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2858 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
2859 GET_MODE_SIZE argument is enum machine_mode.
2861 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2864 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
2865 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
2867 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
2870 * asan.c (asan_emit_stack_protection): Force the base to align to
2871 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
2872 appropriate bits if STRICT_ALIGNMENT.
2873 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
2875 (expand_used_vars): Leave a space in the stack frame for alignment
2876 if STRICT_ALIGNMENT.
2878 2014-04-21 David Malcolm <dmalcolm@redhat.com>
2880 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
2882 (gimple_store_p): Likewise.
2883 (gimple_assign_load_p): Likewise.
2884 (gimple_assign_cast_p): Likewise.
2885 (gimple_clobber_p): Likewise.
2887 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
2888 rather than a gimple.
2889 (gimple_assign_cast_p): Likewise.
2891 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
2894 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
2895 If mode is DDmode and TARGET_E500_DOUBLE allow move.
2897 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
2898 more debug information for E500 if -mdebug=reg.
2900 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
2903 * config/i386/i386.c (ix86_expand_builtin)
2904 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
2905 register for target RTX.
2906 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
2908 2014-04-18 Cong Hou <congh@google.com>
2910 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
2911 the widen-mult pattern by handling two operands with different sizes,
2912 and operands whose size is smaller than half of the result type.
2914 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
2916 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
2917 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
2918 (do_estimate_edge_time): Compute it.
2919 * ipa-inline.c (want_inline_small_function_p): Bypass
2920 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
2922 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
2924 * ipa-inline.c (spec_rem): New static variable.
2925 (dump_overall_stats): New function.
2926 (dump_inline_stats): New function.
2928 2014-04-18 Richard Henderson <rth@redhat.com>
2930 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
2931 to GET_MODE_SIZE, not a reg_class_t.
2933 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2935 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
2936 (vsx_xxmrglw_<mode>): Likewise.
2938 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2941 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
2942 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
2943 (rs6000_init_hard_regno_mode_ok): Likewise.
2945 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
2947 * ipa-inline.c (inline_small_functions): Account only non-cold
2949 * doc/invoke.texi (inline-unit-growth): Update documentation.
2951 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
2953 * config/rs6000/rs6000.md (addti3, subti3): New.
2955 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
2958 * config/i386/i386.c (ix86_expand_clear): Remove outdated
2959 comment. Check optimize_insn_for_size_p instead of
2960 optimize_insn_for_speed_p.
2962 2014-04-17 Martin Jambor <mjambor@suse.cz>
2964 * gimple-iterator.c (gsi_start_edge): New function.
2965 * gimple-iterator.h (gsi_start_edge): Declare.
2966 * tree-sra.c (single_non_eh_succ): New function.
2967 (disqualify_ops_if_throwing_stmt): Renamed to
2968 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
2969 having one non-EH successor BB.
2970 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
2971 generate loads into replacements.
2972 (sra_modify_assign): Likewise and and also use the simple path for
2974 (sra_modify_function_body): Commit statements on edges.
2976 2014-04-17 Richard Biener <rguenther@suse.de>
2979 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
2980 comparison results and add clarifying comment.
2982 2014-04-17 Jakub Jelinek <jakub@redhat.com>
2984 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
2985 (blank_mode): Initialize it.
2986 (emit_mode_size_inline, emit_mode_nunits_inline,
2987 emit_mode_inner_inline): New functions.
2988 (emit_insn_modes_h): Call them and surround their output with
2989 #if GCC_VERSION >= 4001 ... #endif.
2990 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
2991 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
2992 mode_* arrays if the argument is __builtin_constant_p.
2993 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
2994 is enum machine_mode.
2996 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2998 * passes.c (opt_pass::execute): Adjust.
2999 (pass_manager::execute_pass_mode_switching): Likewise.
3000 (early_local_passes::execute): Likewise.
3001 (execute_one_pass): Pass cfun to the pass's execute method.
3002 * tree-pass.h (opt_pass::execute): Add function * argument.
3003 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3004 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3005 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3006 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3007 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3008 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
3009 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
3010 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3011 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3012 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3013 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3014 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3015 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3016 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3017 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3018 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3019 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3020 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3021 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3022 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3023 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3024 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3025 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3026 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3027 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3028 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3029 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3030 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3031 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3034 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3036 * passes.c (opt_pass::gate): Take function * argument.
3037 (gate_all_early_local_passes): Merge into
3038 (early_local_passes::gate): this.
3039 (gate_all_early_optimizations): Merge into
3040 (all_early_optimizations::gate): this.
3041 (gate_all_optimizations): Mege into
3042 (all_optimizations::gate): this.
3043 (gate_all_optimizations_g): Merge into
3044 (all_optimizations_g::gate): this.
3045 (gate_rest_of_compilation): Mege into
3046 (rest_of_compilation::gate): this.
3047 (gate_postreload): Merge into
3048 (postreload::gate): this.
3049 (dump_one_pass): Pass cfun to the pass's gate method.
3050 (execute_ipa_summary_passes): Likewise.
3051 (execute_one_pass): Likewise.
3052 (ipa_write_summaries_2): Likewise.
3053 (ipa_write_optimization_summaries_1): Likewise.
3054 (ipa_read_summaries_1): Likewise.
3055 (ipa_read_optimization_summaries_1): Likewise.
3056 (execute_ipa_stmt_fixups): Likewise.
3057 * tree-pass.h (opt_pass::gate): Add function * argument.
3058 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
3059 combine-stack-adj.c, combine.c, compare-elim.c,
3060 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3061 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
3062 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
3063 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
3064 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3065 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3066 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3067 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
3068 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
3069 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
3070 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3071 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3072 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
3073 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3074 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3075 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3076 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3077 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3078 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3079 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3080 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3081 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3082 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3083 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
3084 var-tracking.c, vtable-verify.c, web.c: Adjust.
3086 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3088 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
3089 * configure: Regenerate.
3091 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3093 * passes.c (dump_one_pass): don't check pass->has_gate.
3094 (execute_ipa_summary_passes): Likewise.
3095 (execute_one_pass): Likewise.
3096 (ipa_write_summaries_2): Likewise.
3097 (ipa_write_optimization_summaries_1): Likewise.
3098 (ipa_read_optimization_summaries_1): Likewise.
3099 (execute_ipa_stmt_fixups): Likewise.
3100 * tree-pass.h (pass_data::has_gate): Remove.
3101 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3102 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3103 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3104 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3105 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3106 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
3107 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
3108 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
3109 gimple-low.c, gimple-ssa-isolate-paths.c,
3110 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
3111 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3112 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3113 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
3114 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
3115 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
3116 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
3117 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3118 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
3119 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3120 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3121 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3122 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3123 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3124 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3125 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3126 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3127 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3128 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3129 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3130 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3131 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3134 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3136 * pass_manager.h (pass_manager::register_dump_files_1): Remove
3138 * passes.c (pass_manager::register_dump_files_1): Merge into
3139 (pass_manager::register_dump_files): this, and remove its handling of
3140 properties since the pass always has the properties anyway.
3141 (pass_manager::pass_manager): Adjust.
3143 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3145 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
3146 * passes.c (pass_manager::register_dump_files_1): Remove dead code
3147 dealing with properties.
3148 (pass_manager::register_dump_files): Adjust.
3150 2014-03-20 Mark Wielaard <mjw@redhat.com>
3152 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
3153 then represent the bound as normal constant value.
3155 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3158 Forward port from 4.8 branch
3159 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
3161 * config/i386/bmiintrin.h (_blsi_u32): New.
3165 (_blsmsk_u32): Ditto.
3166 (_blsmsk_u64): Ditto.
3167 (_tzcnt_u32): Ditto.
3168 (_tzcnt_u64): Ditto.
3170 2014-04-17 Kito Cheng <kito@0xlab.org>
3172 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
3174 2014-04-17 Richard Biener <rguenther@suse.de>
3177 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
3178 boolean results for comparisons.
3180 2014-04-17 Richard Biener <rguenther@suse.de>
3182 PR tree-optimization/60836
3183 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
3184 initial PHI args to be gimple values.
3186 2014-04-17 Richard Biener <rguenther@suse.de>
3188 PR tree-optimization/60841
3189 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
3190 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
3191 of stmts to SLP build.
3192 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
3193 (vect_analyze_slp): Likewise.
3194 (vect_analyze_slp_instance): Likewise.
3195 (vect_build_slp_tree): Limit overall SLP tree growth.
3196 * tree-vectorizer.h (vect_analyze_data_refs,
3197 vect_analyze_slp): Adjust prototypes.
3199 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3201 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
3204 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3206 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
3207 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
3208 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
3209 for TARGET_SLOW_PSHUFB
3211 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3213 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
3214 * config/i386/i386.c (intel_cost): Ditto.
3216 2014-04-17 Joey Ye <joey.ye@arm.com>
3218 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
3220 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3222 * opts.c (common_handle_option): Disable -fipa-reference coorectly
3225 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3227 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
3228 (type_all_derivations_known_p): New predicate.
3229 (type_all_ctors_visible_p): New predicate.
3230 (type_possibly_instantiated_p): New predicate.
3231 (get_odr_type): Compute all_derivations_known.
3232 (dump_odr_type): Dump the flag.
3233 (maybe_record_type): Cleanup.
3234 (record_target_from_binfo): Add bases_to_consider array;
3235 record bases for types w/o instances and skip CXX destructor.
3236 (possible_polymorphic_call_targets_1): Add bases_to_consider
3237 and consider_construction parameters; check if type may have instance.
3238 (get_polymorphic_call_info): Set maybe_in_construction to true
3239 when we know nothing.
3240 (record_targets_from_bases): Skip CXX destructors; they are
3241 never called for types in construction.
3242 (possible_polymorphic_call_targets): Do not record target when
3243 type may not have instance.
3245 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3248 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
3249 external aliases alive, too.
3251 2014-04-16 Andrew Pinski <apinski@cavium.com>
3253 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
3256 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3258 * final.c (compute_alignments): Do not apply loop alignment to a block
3259 falling through to the exit.
3261 2014-04-16 Catherine Moore <clm@codesourcery.com>
3263 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
3264 Adjust constraints for microMIPS store patterns.
3266 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3268 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
3270 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3272 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
3273 (append_use): Run at -O0.
3274 (append_vdef): Likewise.
3275 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
3276 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
3278 2014-04-16 Jakub Jelinek <jakub@redhat.com>
3280 PR tree-optimization/60844
3281 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
3282 (propagate_op_to_single_use, remove_visited_stmt_chain,
3283 linearize_expr, repropagate_negates, reassociate_bb): Use it
3284 instead of gsi_remove.
3286 2014-04-16 Martin Jambor <mjambor@suse.cz>
3288 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
3289 ipa_transforms_to_apply.
3290 (cgraph_function_versioning): Assert that old_node has empty
3291 ipa_transforms_to_apply.
3292 * trans-mem.c (ipa_tm_create_version): Likewise.
3293 * tree-inline.c (tree_function_versioning): Do not duplicate
3294 ipa_transforms_to_apply.
3296 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3299 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
3301 Pass necessary as flags on 64-bit Solaris/x86.
3302 Use lowercase relocs for x86_64-*-*.
3303 * configure: Regenerate.
3305 2014-04-15 Jan Hubicka <jh@suse.cz>
3307 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
3308 (maybe_record_node, likely_target_p): Use it.
3310 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3313 Revert following patch
3315 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3318 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3319 software floating point or no floating point registers, do not
3320 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3321 in GPRs that occurs after we tested for GPRs that would never be
3324 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3325 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3326 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3327 specifically allow DDmode, since that does not use the SPE SIMD
3330 2014-03-21 Mark Wielaard <mjw@redhat.com>
3332 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
3333 as unsigned or int depending on type and value used.
3335 2014-04-15 Richard Biener <rguenther@suse.de>
3337 PR rtl-optimization/56965
3338 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
3339 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
3341 * alias.c (true_dependence_1): Do not call
3342 nonoverlapping_component_refs_p.
3343 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
3344 nonoverlapping_component_refs_p.
3345 (indirect_refs_may_alias_p): Likewise.
3347 2014-04-15 Teresa Johnson <tejohnson@google.com>
3349 * cfg.c (dump_bb_info): Fix flags check.
3350 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
3352 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3354 PR rtl-optimization/60663
3355 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
3358 2014-04-15 Richard Biener <rguenther@suse.de>
3360 * lto-streamer.h (LTO_major_version): Bump to 4.
3362 2014-04-15 Richard Biener <rguenther@suse.de>
3364 * common.opt (lto_partition_model): New enum.
3365 (flto-partition=): Merge separate options with a single with argument,
3366 add -flto-partition=one support.
3367 * flag-types.h (enum lto_partition_model): Declare.
3368 * opts.c (finish_options): Remove duplicate -flto-partition=
3370 * lto-wrapper.c (run_gcc): Adjust.
3372 2014-04-15 Richard Biener <rguenther@suse.de>
3374 * alias.c (ncr_compar): New function.
3375 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
3377 2014-04-15 Richard Biener <rguenther@suse.de>
3379 * alias.c (record_component_aliases): Do not walk BINFOs.
3381 2014-04-15 Richard Biener <rguenther@suse.de>
3383 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
3384 Add struct function argument and adjust.
3385 (find_func_aliases_for_call): Likewise.
3386 (find_func_aliases): Likewise.
3387 (find_func_clobbers): Likewise.
3388 (intra_create_variable_infos): Likewise.
3389 (compute_points_to_sets): Likewise.
3390 (ipa_pta_execute): Adjust. Do not push/pop cfun.
3392 2014-04-15 Richard Biener <rguenther@suse.de>
3394 * tree.c (iterative_hash_expr): Use enum tree_code_class
3395 to store TREE_CODE_CLASS.
3396 (tree_block): Likewise.
3397 (tree_set_block): Likewise.
3398 * tree.h (fold_build_pointer_plus_loc): Use
3399 convert_to_ptrofftype_loc.
3401 2014-04-15 Jakub Jelinek <jakub@redhat.com>
3404 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
3407 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
3409 * cfgloop.h (struct loop): Move force_vectorize down.
3410 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
3411 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
3412 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
3413 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
3414 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
3415 * tree-core.h (enum annot_expr_kind): Add new kind values.
3416 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
3417 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
3419 * tree.def (ANNOTATE_EXPR): Tweak comment.
3421 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3423 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
3426 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
3428 * tree.h (TYPE_IDENTIFIER): Declare.
3429 * tree.c (subrange_type_for_debug_p): Use it.
3430 * godump.c (go_format_type): Likewise.
3431 * dwarf2out.c (is_cxx_auto, modified_type_die,
3432 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
3433 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
3435 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3438 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
3440 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
3442 * config/i386/i386.c (examine_argument): Return bool. Return true if
3443 parameter should be passed in memory.
3444 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
3445 (construct_container): Update calls to examine_argument.
3446 (function_arg_advance_64): Ditto.
3447 (return_in_memory_32): Merge with ix86_return_in_memory.
3448 (return_in_memory_64): Ditto.
3449 (return_in_memory_ms_64): Ditto.
3451 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3453 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
3454 * coverage.c (coverage_compute_profile_id): Handle externally visible
3457 2014-04-14 Martin Jambor <mjambor@suse.cz>
3459 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
3460 DECL_DISREGARD_INLINE_LIMITS functions.
3462 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3465 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
3467 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3470 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
3471 optimize_insn_for_speed_p instead of
3472 optimize_function_for_speed_p.
3474 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
3476 * doc/invoke.texi (free): Document AArch64.
3478 2014-04-14 Richard Biener <rguenther@suse.de>
3480 PR tree-optimization/60042
3481 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
3482 (insert_into_preds_of_block): Do not prevent PHI insertion
3483 for REFERENCE exprs here ...
3484 (eliminate_dom_walker::before_dom_children): ... but prevent
3485 their use here under similar conditions when applied to the
3486 IL after PRE optimizations.
3488 2014-04-14 Richard Biener <rguenther@suse.de>
3490 * passes.def: Move early points-to after early SRA.
3492 2014-04-14 Richard Biener <rguenther@suse.de>
3494 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
3495 check for which sign-changes we allow when forwarding
3496 a converted value into a switch.
3498 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3500 * stor-layout.c (place_field): Finalize non-constant offset for the
3503 2014-04-14 Richard Biener <rguenther@suse.de>
3505 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
3507 (expand_switch_using_bit_tests_p): Likewise.
3508 (process_switch): Compute and pass on speed_p based on the
3510 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
3511 optimize_bb_for_speed_p.
3513 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3515 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
3516 * function.h (struct function): Rename has_force_vect_loops into
3517 has_force_vectorize_loops.
3518 * lto-streamer-in.c (input_cfg): Adjust for renaming.
3519 (input_struct_function_base): Likewise.
3520 * lto-streamer-out.c (output_cfg): Likewise.
3521 (output_struct_function_base): Likewise.
3522 * omp-low.c (expand_omp_simd): Likewise.
3523 * tree-cfg.c (move_sese_region_to_fn): Likewise.
3524 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
3525 (version_loop_for_if_conversion): Likewise.
3526 (tree_if_conversion): Likewise.
3527 (main_tree_if_conversion): Likewise.
3528 (gate_tree_if_conversion): Likewise.
3529 * tree-inline.c (copy_loops): Likewise.
3530 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
3531 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
3532 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
3533 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
3534 * tree-vectorizer.c (vectorize_loops): Likewise.
3535 * tree-vectorizer.h (unlimited_cost_model): Likewise.
3537 2014-04-14 Richard Biener <rguenther@suse.de>
3540 * lto-streamer-out.c (wrap_refs): New function.
3541 (lto_output): Wrap symbol references in global initializes in
3542 type-preserving MEM_REFs.
3544 2014-04-14 Christian Bruel <christian.bruel@st.com>
3546 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
3548 2014-04-14 Christian Bruel <christian.bruel@st.com>
3550 * config/sh/sh.md (setmemqi): New expand pattern.
3551 * config/sh/sh.h (CLEAR_RATIO): Define.
3552 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
3553 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
3555 2014-04-14 Richard Biener <rguenther@suse.de>
3558 * fold-const.c (negate_expr_p): Don't negate directional rounding
3560 (fold_negate_expr): Likewise.
3562 2014-04-14 Richard Biener <rguenther@suse.de>
3564 PR tree-optimization/59817
3565 PR tree-optimization/60453
3566 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
3567 recursion to catch all CHRECs in the scalar evolution and restrict
3568 the predicate for the remains appropriately.
3570 2014-04-12 Catherine Moore <clm@codesourcery.com>
3572 * config/mips/constraints.md: Add new register constraint "kb".
3573 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
3574 (*movhi_internal): Likewise.
3575 (*movqi_internal): Likewise.
3576 * config/mips/mips.h (M16_STORE_REGS): New register class.
3577 (REG_CLASS_NAMES): Add M16_STORE_REGS.
3578 (REG_CLASS_CONTENTS): Likewise.
3579 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
3581 2014-04-11 Tobias Burnus <burnus@net-b.de>
3584 * doc/invoke.texi (-Wformat-signedness): Document it.
3585 (Wformat=2): Mention that this enables -Wformat-signedness.
3587 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3589 * common/config/epiphany/epiphany-common.c
3590 (epiphany_option_optimization_table): Enable section anchors by
3591 default at -O1 or higher.
3592 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
3593 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
3594 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
3595 carries no extra cost.
3596 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
3597 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
3598 * config/epiphany/predicates.md (memclob_operand): New predicate.
3599 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
3600 Use memclob_operand predicate and X constraint for operand 3.
3602 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3604 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
3605 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
3608 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3610 PR rtl-optimization/60651
3611 * mode-switching.c (optimize_mode_switching): Make sure to emit
3612 sets of a lower numbered entity before sets of a higher numbered
3613 entity to a mode of the same or lower priority.
3614 When creating a seginfo for a basic block that starts with a code
3615 label, move the insertion point past the code label.
3616 (new_seginfo): Document and enforce requirement that
3617 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
3618 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
3619 * doc/tm.texi: Regenerate.
3621 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
3624 * config/arc/arc.c (arc_save_restore): Fix assert typo.
3626 2013-04-11 Jakub Jelinek <jakub@redhat.com>
3628 * BASE-VER: Set to 4.10.0.
3630 2014-04-11 Tobias Burnus <burnus@net-b.de>
3633 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
3634 * doc/gcc.texi (Service): Update description in the @menu
3635 * doc/invoke.texi (Option Summary): Remove misplaced and
3638 2014-04-11 Steve Ellcey <sellcey@mips.com>
3639 Jakub Jelinek <jakub@redhat.com>
3642 * expr.c (convert_move): Use emit_store_flag_force instead of
3643 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
3646 2014-04-11 Richard Biener <rguenther@suse.de>
3649 * varasm.c (assemble_alias): Avoid endless error reporting
3650 recursion by setting TREE_ASM_WRITTEN.
3652 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3654 * config/s390/s390.md: Add a splitter for NOT rtx.
3656 2014-04-11 Jakub Jelinek <jakub@redhat.com>
3658 PR rtl-optimization/60663
3659 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
3661 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
3662 Jakub Jelinek <jakub@redhat.com>
3665 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
3666 flag from decl_node to node.
3668 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3671 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
3672 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
3673 ameliorating the cases where it can be.
3675 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
3678 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
3680 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3681 (loadsync_<mode>): Change mode.
3682 (load_quadpti, store_quadpti): New.
3683 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3684 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3685 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
3687 2014-04-09 Cong Hou <congh@google.com>
3690 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
3693 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3695 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
3696 instead of vnor to exploit possible fusion opportunity in the
3698 (altivec_expand_vec_perm_const_le): Likewise.
3700 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
3702 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3703 (loadsync_<mode>): Change mode.
3704 (load_quadpti, store_quadpti): New.
3705 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3706 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3708 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
3711 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
3712 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
3713 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
3715 2014-04-08 Richard Biener <rguenther@suse.de>
3718 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
3719 a 64bit widest int print double-int similar to on HWI64 hosts.
3721 2014-04-08 Richard Biener <rguenther@suse.de>
3723 PR tree-optimization/60785
3724 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
3725 default defs properly.
3727 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
3729 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
3730 (Weffc++): Likewise.
3732 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
3734 * ipa-devirt.c (maybe_record_node): When node is not recorded,
3735 set completep to false rather than true.
3737 2014-04-07 Douglas B Rupp <rupp@adacore.com>
3740 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
3741 ARM_TARGET2_DWARF_FORMAT.
3743 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
3746 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
3747 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
3750 2014-04-07 Richard Biener <rguenther@suse.de>
3752 PR tree-optimization/60766
3753 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
3754 (may_eliminate_iv): Convert cand_value_at result to desired type.
3756 2014-04-07 Jason Merrill <jason@redhat.com>
3759 * common.opt (-fno-gnu-unique): Add.
3760 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
3762 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3764 * haifa-sched.c: Fix outdated function reference and minor
3765 grammar errors in introductory comment.
3767 2014-04-07 Richard Biener <rguenther@suse.de>
3770 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
3772 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
3774 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
3777 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
3778 size accounting for thunks.
3779 (pa_asm_output_mi_thunk): Use final_start_function() and
3780 final_end_function() to output function start and end directives.
3782 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3784 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
3785 device specific ISA/ feature information. Remove short_sp and
3786 errata_skip ds. Add avr_device_specific_features enum to have device
3788 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
3789 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
3790 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
3791 updated device specific info.
3792 * config/avr/avr-mcus.def: Merge device specific details to
3793 dev_attribute field.
3794 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
3796 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
3797 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
3798 assembler if RMW isa supported by current device.
3799 * config/avr/genmultilib.awk: Update as device info structure changed.
3800 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
3802 2014-04-04 Cong Hou <congh@google.com>
3804 PR tree-optimization/60656
3805 * tree-vect-stmts.c (supportable_widening_operation):
3806 Fix a bug that elements in a vector with vect_used_by_reduction
3807 property are incorrectly reordered when the operation on it is not
3808 consistant with the one in reduction operation.
3810 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
3812 PR rtl-optimization/60155
3813 * gcse.c (record_set_data): New function.
3814 (single_set_gcse): New function.
3815 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
3816 (hoist_code): Likewise.
3817 (get_pressure_class_and_nregs): Likewise.
3819 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
3821 * explow.c (probe_stack_range): Emit a final optimization blockage.
3823 2014-04-04 Anthony Green <green@moxielogic.com>
3825 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
3828 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
3831 * lto-cgraph.c (input_overwrite_node): Check that partitioning
3832 flags are set only during streaming.
3833 * ipa.c (process_references, walk_polymorphic_call_targets,
3834 symtab_remove_unreachable_nodes): Drop bodies of always inline
3835 after early inlining.
3836 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
3838 2014-04-04 Jakub Jelinek <jakub@redhat.com>
3839 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3842 * dwarf2out.c (const_ok_for_output_1): Reject expressions
3845 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3848 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
3850 (cortex_a53_fdivd): Likewise.
3852 2014-04-04 Martin Jambor <mjambor@suse.cz>
3855 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
3857 * cgraph.c (clone_of_p): Also return true if thunks match.
3858 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
3859 cgraph_function_or_thunk_node and an obsolete comment.
3860 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
3862 (build_function_decl_skip_args): Likewise.
3863 (set_new_clone_decl_and_node_flags): New function.
3864 (duplicate_thunk_for_node): Likewise.
3865 (redirect_edge_duplicating_thunks): Likewise.
3866 (cgraph_clone_node): New parameter args_to_skip, pass it to
3867 redirect_edge_duplicating_thunks which is called instead of
3868 cgraph_redirect_edge_callee.
3869 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
3870 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
3872 2014-04-04 Jeff Law <law@redhat.com>
3875 * config/arm/predicates.md (const_int_I_operand): New predicate.
3876 (const_int_M_operand): Similarly.
3877 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
3879 (insv_t2, extv_reg, extzv_t2): Likewise.
3880 (load_multiple_with_writeback): Similarly for const_int_I_operand.
3881 (pop_multiple_with_writeback_and_return): Likewise.
3882 (vfp_pop_multiple_with_writeback): Likewise
3884 2014-04-04 Richard Biener <rguenther@suse.de>
3887 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
3888 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
3890 * gimplify.h (gimple_add_tmp_var_fn): Declare.
3891 * gimplify.c (gimple_add_tmp_var_fn): New function.
3892 * gimple-expr.h (create_tmp_reg_fn): Declare.
3893 * gimple-expr.c (create_tmp_reg_fn): New function.
3894 * gimple-low.c (record_vars_into): Don't change cfun.
3895 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
3896 code generation without cfun.
3898 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
3901 * Makefile.in (install-driver): Fix shell scripting.
3903 2014-04-03 Cong Hou <congh@google.com>
3905 PR tree-optimization/60505
3906 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
3907 threshold of number of iterations below which no vectorization
3909 * tree-vect-loop.c (new_loop_vec_info):
3910 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
3911 * tree-vect-loop.c (vect_analyze_loop_operations):
3912 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
3913 * tree-vect-loop.c (vect_transform_loop):
3914 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
3915 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
3916 of iterations of the loop and see if we should build the epilogue.
3918 2014-04-03 Richard Biener <rguenther@suse.de>
3920 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
3921 (streamer_tree_cache_create): Adjust.
3922 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
3923 to allow optional nodes array.
3924 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
3925 (streamer_tree_cache_append): Likewise.
3926 (streamer_tree_cache_create): Create nodes array optionally
3927 as specified by parameter.
3928 * lto-streamer-out.c (create_output_block): Avoid maintaining
3929 the node array in the writer cache.
3930 (DFS_write_tree): Remove assertion.
3931 (produce_asm_for_decls): Free the out decl state hash table early.
3932 * lto-streamer-in.c (lto_data_in_create): Adjust for
3933 streamer_tree_cache_create prototype change.
3935 2014-04-03 Richard Biener <rguenther@suse.de>
3937 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
3938 set TREE_CHAIN to NULL_TREE.
3940 2014-04-03 Richard Biener <rguenther@suse.de>
3942 PR tree-optimization/60740
3943 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
3944 over all GIMPLE_COND operands.
3946 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
3948 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
3949 (Weffc++): Remove Scott's numbering, merge lists and reference
3952 2014-04-03 Nick Clifton <nickc@redhat.com>
3954 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
3957 2014-04-03 Martin Jambor <mjambor@suse.cz>
3959 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
3960 mention gcc_unreachable before failing.
3961 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
3964 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
3967 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
3968 inconsistent code and instead mark the context inconsistent.
3969 (possible_polymorphic_call_targets): For inconsistent contexts
3970 return empty complete list.
3972 2014-04-02 Anthony Green <green@moxielogic.com>
3974 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
3975 (extendqisi2, extendhisi2): Define.
3976 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
3977 (WCHAR_TYPE): Change to unsigned int.
3979 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3981 PR tree-optimization/60733
3982 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
3983 insertion point for PHI candidates to be the end of the feeding
3984 block for the PHI argument.
3986 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
3988 PR rtl-optimization/60650
3989 * lra-constraints.c (process_alt_operands): Decrease reject for
3990 earlyclobber matching.
3992 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3994 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
3996 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3998 * config/spu/spu.c (pad_bb): Do not crash when the last
3999 insn is CODE_FOR_blockage.
4001 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4003 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
4004 lies outside the target mode.
4006 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4009 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4010 software floating point or no floating point registers, do not
4011 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4012 in GPRs that occurs after we tested for GPRs that would never be
4015 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4016 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4017 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4018 specifically allow DDmode, since that does not use the SPE SIMD
4021 2014-04-02 Richard Biener <rguenther@suse.de>
4024 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4025 MODE_INTs. Properly use negv_optab.
4026 (expand_abs): Likewise.
4028 2014-04-02 Richard Biener <rguenther@suse.de>
4031 * Makefile.in (install-driver): Guard extra installs with special
4034 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4036 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4039 2014-04-01 Richard Henderson <rth@redhat.com>
4042 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
4043 alternative enabled before register allocation.
4045 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
4047 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
4048 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
4050 (nios2_large_got_address): Remove unneeded 'sym' parameter.
4051 (nios2_got_address): Update nios2_large_got_address call site.
4052 (nios2_delegitimize_address): New function.
4053 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
4054 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
4055 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
4057 2014-04-01 Martin Husemann <martin@duskware.de>
4059 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
4062 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
4064 PR rtl-optimization/60604
4065 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
4066 check from register_operand.
4067 (register_operand): Redefine in terms of general_operand.
4068 (nonmemory_operand): Use register_operand for the non-constant cases.
4070 2014-04-01 Richard Biener <rguenther@suse.de>
4072 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
4074 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
4076 * doc/invoke.texi (mapp-regs): Clarify.
4078 2014-03-31 Ulrich Drepper <drepper@gmail.com>
4080 * config/i386/avx512fintrin.h (__v32hi): Define type.
4081 (__v64qi): Likewise.
4082 (_mm512_set1_epi8): Define.
4083 (_mm512_set1_epi16): Define.
4084 (_mm512_set4_epi32): Define.
4085 (_mm512_set4_epi64): Define.
4086 (_mm512_set4_pd): Define.
4087 (_mm512_set4_ps): Define.
4088 (_mm512_setr4_epi64): Define.
4089 (_mm512_setr4_epi32): Define.
4090 (_mm512_setr4_pd): Define.
4091 (_mm512_setr4_ps): Define.
4092 (_mm512_setzero_epi32): Define.
4094 2014-03-31 Martin Jambor <mjambor@suse.cz>
4097 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
4098 callsite_arguments_match_p. Updated all callers. Also check types of
4099 corresponding formal parameters and actual arguments.
4100 (not_all_callers_have_enough_arguments_p) Renamed to
4101 some_callers_have_mismatched_arguments_p.
4103 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
4105 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
4107 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
4110 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
4113 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
4115 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
4116 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
4118 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
4119 Use FMAMODE_NOVF512 mode iterator.
4120 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
4121 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
4122 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
4123 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
4125 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
4126 Use VF_128_256 mode iterator.
4127 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
4130 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4132 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
4133 static chain if needed.
4135 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4138 * lra-constraints.c (index_part_to_reg): New.
4139 (process_address): Use it.
4141 2014-03-27 Jeff Law <law@redhat.com>
4142 Jakub Jelinek <jakub@redhat.com>
4145 * expr.c (do_tablejump): Use simplify_gen_binary rather than
4146 gen_rtx_{PLUS,MULT} to build up the address expression.
4148 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
4149 creating non-canonical RTL.
4151 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4154 * ipa-inline.c (want_inline_small_function_p): Short circuit large
4155 functions; reorganize to make cheap checks first.
4156 (inline_small_functions): Do not estimate growth when dumping;
4158 * ipa-inline.h (inline_summary): Add min_size.
4159 (growth_likely_positive): New function.
4160 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
4161 (set_cond_stmt_execution_predicate): Cleanup.
4162 (estimate_edge_size_and_time): Compute min_size.
4163 (estimate_calls_size_and_time): Likewise.
4164 (estimate_node_size_and_time): Likewise.
4165 (inline_update_overall_summary): Update min_size.
4166 (do_estimate_edge_time): Likewise.
4167 (do_estimate_edge_size): Update.
4168 (do_estimate_edge_hints): Update.
4169 (growth_likely_positive): New function.
4171 2014-03-28 Jakub Jelinek <jakub@redhat.com>
4174 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
4175 also if addr has VOIDmode.
4177 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4179 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
4180 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
4182 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
4183 instructions as well as AdvancedSIMD loads.
4185 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4187 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
4188 Use crypto_aese type.
4189 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
4190 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
4191 crypto_aese, crypto_aesmc. Move to types.md.
4192 * config/arm/types.md (crypto_aes): Split into crypto_aese,
4194 * config/arm/iterators.md (crypto_type): Likewise.
4196 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4198 * cgraph.c: Include expr.h and tree-dfa.h.
4199 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
4202 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4205 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
4206 regs from checking multi-reg pseudos.
4208 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4210 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
4212 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4214 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
4215 if it would clobber the stack pointer, even temporarily.
4217 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
4219 * mode-switching.c: Make small adjustments to the top comment.
4221 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4223 * config/rs6000/constraints.md (wD constraint): New constraint to
4224 match the constant integer to get the top DImode/DFmode out of a
4225 vector in a VSX register.
4227 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
4228 match the constant integer to get the top DImode/DFmode out of a
4229 vector in a VSX register.
4231 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
4234 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
4237 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
4238 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
4240 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
4241 Optimize vec_extract of 64-bit values, where the value being
4242 extracted is in the top word, where we can use scalar
4243 instructions. Add direct move and store support. Combine the big
4244 endian/little endian vector select load support into a single insn.
4245 (vsx_extract_<mode>_internal1): Likewise.
4246 (vsx_extract_<mode>_internal2): Likewise.
4247 (vsx_extract_<mode>_load): Likewise.
4248 (vsx_extract_<mode>_store): Likewise.
4249 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
4250 combined into vsx_extract_<mode>_load.
4251 (vsx_extract_<mode>_one_le): Likewise.
4253 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
4254 define the top 64-bit vector element.
4256 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
4259 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4260 Document vec_vbpermq builtin.
4263 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
4264 enable use of xxsldwi and xxpermdi builtin functions.
4265 (vec_xxpermdi): Likewise.
4267 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4268 Document use of vec_xxsldwi and vec_xxpermdi builtins.
4270 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
4272 PR rtl-optimization/60650
4273 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
4275 (find_spills_for): New.
4276 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
4277 Spill all pseudos on the second iteration.
4279 2014-03-27 Marek Polacek <polacek@redhat.com>
4282 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
4285 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4287 * config/s390/s390.c (s390_can_use_return_insn): Check for
4288 call-saved FPRs on 31 bit.
4290 2014-03-27 Jakub Jelinek <jakub@redhat.com>
4293 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
4294 if they need regimplification, just drop them instead of
4295 calling gimple_regimplify_operands on them.
4297 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
4300 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
4301 (aarch64_frame_pointer_required): Adjust logic.
4302 (aarch64_can_eliminate): Adjust logic.
4303 (aarch64_override_options_after_change): Adjust logic.
4305 2014-03-27 Dehao Chen <dehao@google.com>
4307 * ipa-inline.c (early_inliner): Update node's inline info.
4309 2014-03-26 Dehao Chen <dehao@google.com>
4311 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
4312 compiler inserted conditional jumps for NAN float check.
4314 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4316 * ubsan.h (ubsan_create_data): Change second argument's type
4317 to const location_t *.
4318 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
4320 (ubsan_create_data): Change second argument to const location_t *PLOC.
4321 Create Loc field whenever PLOC is non-NULL.
4322 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
4323 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
4327 * real.c (real_to_integer2): Change type of low to UHWI.
4329 2014-03-26 Tobias Burnus <burnus@net-b.de>
4331 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
4332 (CILK_SELF_SPECS): New define.
4333 (driver_self_specs): Use it.
4335 2014-03-26 Richard Biener <rguenther@suse.de>
4337 * tree-pretty-print.c (percent_K_format): Implement special
4338 case for LTO and its stripped down BLOCK tree.
4340 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4343 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
4345 * tree-vrp.c (simplify_internal_call_using_ranges): If only
4346 one range is range_int_cst_p, but not both, at least optimize
4347 addition/subtraction of 0 and multiplication by 0 or 1.
4348 * gimple-fold.c (gimple_fold_call): Fold
4349 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
4350 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
4351 INTEGER_CSTs, try to fold at least x * 0 and y - y.
4353 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
4355 PR rtl-optimization/60452
4356 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
4357 <case REG>: Return 1 for invalid offsets from the frame pointer.
4359 2014-03-26 Marek Polacek <polacek@redhat.com>
4362 * doc/extend.texi (C Extensions): Mention variable-length arrays in
4365 2014-03-26 Marek Polacek <polacek@redhat.com>
4368 * doc/extend.texi (Designated Inits): Describe what happens to omitted
4371 2014-03-26 Marek Polacek <polacek@redhat.com>
4374 * ira-color.c (update_conflict_hard_regno_costs): Perform the
4375 multiplication in unsigned type.
4377 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4379 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
4381 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4383 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
4385 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4388 * cif-code.def (UNREACHABLE) New code.
4389 * ipa-inline.c (inline_small_functions): Skip edges to
4390 __builtlin_unreachable.
4391 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
4392 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
4393 predicate to __bulitin_unreachable.
4394 (set_cond_stmt_execution_predicate): Fix issue when
4395 invert_tree_comparison returns ERROR_MARK.
4396 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
4397 propagate to inline clones.
4398 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
4400 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
4401 * cgraphclones.c (cgraph_clone_node): If call destination is already
4402 ureachable, do not redirect it back.
4403 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
4406 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4408 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
4409 Do not modify inline clones.
4411 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4413 * config/i386/i386.md (general_sext_operand): New mode attr.
4414 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
4415 don't generate (sign_extend (const_int)).
4416 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
4417 operands[2]. Use We constraint instead of <i> and
4418 <general_sext_operand> predicate instead of <general_operand>.
4419 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
4420 * config/i386/constraints.md (We): New constraint.
4421 * config/i386/predicates.md (x86_64_sext_operand,
4422 sext_operand): New predicates.
4424 2014-03-25 Martin Jambor <mjambor@suse.cz>
4427 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
4428 inconsistent devirtualizations to __builtin_unreachable.
4430 2014-03-25 Marek Polacek <polacek@redhat.com>
4433 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
4435 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
4437 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
4438 order of elements for big-endian.
4440 2014-03-25 Richard Biener <rguenther@suse.de>
4443 * gimplify-me.c (gimple_regimplify_operands): Update the
4446 2014-03-25 Martin Jambor <mjambor@suse.cz>
4449 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
4450 (lto_output_varpool_node): Likewise.
4451 (input_overwrite_node): Likewise.
4452 (input_varpool_node): Likewise.
4454 2014-03-25 Richard Biener <rguenther@suse.de>
4456 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
4457 (run_gcc): Likewise.
4459 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4461 * combine.c (simplify_compare_const): Add MODE argument.
4462 Handle mode_width 0 as very large mode_width.
4463 (try_combine, simplify_comparison): Adjust callers.
4465 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
4466 type to avoid signed integer overflow.
4467 * explow.c (plus_constant): Likewise.
4469 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
4471 * doc/generic.texi: Correct typos.
4473 2014-03-24 Tobias Burnus <burnus@net-b.de>
4475 * doc/invoke.texi (-flto): Expand section about
4476 using static libraries with LTO.
4478 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4480 PR rtl-optimization/60501
4481 * optabs.def (addptr3_optab): New optab.
4482 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
4483 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
4484 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
4486 * lra.c (emit_add3_insn): Use the addptr pattern if available.
4488 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
4490 2014-03-24 Ulrich Drepper <drepper@gmail.com>
4492 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
4495 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
4496 (_mm256_undefined_ps): Define.
4497 (_mm256_undefined_pd): Define.
4498 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
4499 (_mm_undefined_pd): Define.
4500 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
4501 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
4502 (_mm512_undefined_ps): Define.
4503 (_mm512_undefined_pd): Define.
4504 Use _mm*_undefined_*.
4505 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
4507 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
4509 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
4510 (lshr_simd): DI mode added.
4511 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
4512 (aarch64_ushr_simddi): Likewise.
4513 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
4514 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
4515 (vshrd_n_u64): Likewise.
4517 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4519 * Makefile.in (s-macro_list): Depend on cc1.
4521 2014-03-23 Teresa Johnson <tejohnson@google.com>
4523 * ipa-utils.c (ipa_print_order): Use specified dump file.
4525 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
4527 PR rtl-optimization/60601
4528 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
4530 * gcc.c (eval_spec_function): Initialize save_growing_value.
4532 2014-03-22 Jakub Jelinek <jakub@redhat.com>
4535 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
4536 code == MINUS_EXPR, never swap op0 with op1.
4538 * toplev.c (init_local_tick): Avoid signed integer multiplication
4540 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
4541 shift by first operand's bitsize.
4543 2014-03-21 Jakub Jelinek <jakub@redhat.com>
4546 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
4548 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
4549 TARGET_ISA_64BIT_P(x).
4551 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4553 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
4554 pattern for vector nor instead of subtract from splat(-1).
4555 (altivec_expand_vec_perm_const_le): Likewise.
4557 2014-03-21 Richard Henderson <rth@twiddle.net>
4560 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
4561 related insns after epilogue_completed.
4563 2014-03-21 Martin Jambor <mjambor@suse.cz>
4566 * cgraph.h (symtab_node): New flag body_removed.
4567 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
4568 when removing bodies.
4569 * symtab.c (dump_symtab_base): Dump body_removed flag.
4570 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
4571 had their bodies removed.
4573 2014-03-21 Martin Jambor <mjambor@suse.cz>
4576 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
4579 2014-03-21 Richard Biener <rguenther@suse.de>
4581 PR tree-optimization/60577
4582 * tree-core.h (struct tree_base): Document nothrow_flag use
4583 in VAR_DECL_NONALIASED.
4584 * tree.h (VAR_DECL_NONALIASED): New.
4585 (may_be_aliased): Adjust.
4586 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
4588 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4590 * expr.c (expand_expr_real_1): Remove outdated comment.
4592 2014-03-20 Jakub Jelinek <jakub@redhat.com>
4595 * ira.c (adjust_cleared_regs): Call copy_rtx on
4596 *reg_equiv[REGNO (loc)].src_p before passing it to
4597 simplify_replace_fn_rtx.
4600 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
4601 into CONST, put pic register as first operand of PLUS. Use
4602 gen_const_mem for both 32-bit and 64-bit PIC got loads.
4604 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4606 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
4608 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4610 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
4611 around for store forwarding issue in the FPU on the UT699.
4612 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
4613 loads and operations if -mfix-ut699 is specified.
4614 (divtf3_hq): Tweak attribute.
4615 (sqrttf2_hq): Likewise.
4617 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4619 * calls.c (store_one_arg): Remove incorrect const qualification on the
4620 type of the temporary.
4621 * cfgexpand.c (expand_return): Likewise.
4622 * expr.c (expand_constructor): Likewise.
4623 (expand_expr_real_1): Likewise.
4625 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4627 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
4630 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
4633 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
4635 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
4637 * config/arm/aarch-common-protos.h
4638 (alu_cost_table): Fix spelling of "extend".
4639 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
4641 2014-03-19 Richard Biener <rguenther@suse.de>
4644 * tree-core.h (tree_type_common): Re-order pointer members
4645 to reduce recursion depth during GC walks.
4647 2014-03-19 Marek Polacek <polacek@redhat.com>
4650 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
4651 before accessing it.
4653 2014-03-19 Richard Biener <rguenther@suse.de>
4656 * lto-streamer-in.c (input_function): In WPA stage do not drop
4659 2014-03-19 Jakub Jelinek <jakub@redhat.com>
4661 PR tree-optimization/60559
4662 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
4663 with build_zero_cst assignment.
4665 2014-03-18 Kai Tietz <ktietz@redhat.com>
4667 PR rtl-optimization/56356
4668 * sdbout.c (sdbout_parms): Verify that parms'
4669 incoming argument is valid.
4670 (sdbout_reg_parms): Likewise.
4672 2014-03-18 Richard Henderson <rth@redhat.com>
4675 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
4676 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
4677 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
4679 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
4681 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
4682 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
4683 Italicize plugin event names in description. Explain that
4684 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
4685 Remind that no GCC functions should be called after PLUGIN_FINISH.
4686 Explain what pragmas with expansion are.
4688 2014-03-18 Martin Liska <mliska@suse.cz>
4690 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
4691 gimple call statement is update.
4692 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
4693 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
4695 2014-03-18 Jakub Jelinek <jakub@redhat.com>
4698 * ubsan.c (ubsan_instrument_unreachable): Call
4699 initialize_sanitizer_builtins.
4700 (ubsan_pass): Likewise.
4703 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
4704 varpool_finalize_decl instead of rest_of_decl_compilation.
4706 2014-03-18 Richard Biener <rguenther@suse.de>
4708 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
4709 by using bitmap_and_compl instead of bitmap_and_compl_into.
4710 (df_rd_transfer_function): Likewise.
4712 2014-03-18 Richard Biener <rguenther@suse.de>
4714 * doc/lto.texi (fresolution): Fix typo.
4716 2014-03-18 Richard Biener <rguenther@suse.de>
4718 * doc/invoke.texi (flto): Update for changes in 4.9.
4720 2014-03-18 Richard Biener <rguenther@suse.de>
4722 * doc/loop.texi: Remove section on the removed lambda framework.
4723 Update loop docs with recent changes in preserving loop structure.
4725 2014-03-18 Richard Biener <rguenther@suse.de>
4727 * doc/lto.texi (-fresolution): Document.
4729 2014-03-18 Richard Biener <rguenther@suse.de>
4731 * doc/contrib.texi: Adjust my name.
4733 2014-03-18 Jakub Jelinek <jakub@redhat.com>
4736 * internal-fn.c: Include diagnostic-core.h.
4737 (expand_BUILTIN_EXPECT): New function.
4738 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
4739 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
4740 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
4741 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
4743 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
4744 Revert 3 argument __builtin_expect code.
4745 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
4746 * gimple-fold.c (gimple_fold_call): Likewise.
4747 * tree.h (fold_builtin_expect): New prototype.
4748 * builtins.c (build_builtin_expect_predicate): Add predictor
4749 argument, if non-NULL, create 3 argument __builtin_expect.
4750 (fold_builtin_expect): No longer static. Add ARG2 argument,
4751 pass it through to build_builtin_expect_predicate.
4752 (fold_builtin_2): Adjust caller.
4753 (fold_builtin_3): Handle BUILT_IN_EXPECT.
4754 * internal-fn.def (BUILTIN_EXPECT): New.
4756 2014-03-18 Tobias Burnus <burnus@net-b.de>
4759 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
4760 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
4761 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
4763 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
4766 * predict.c (combine_predictions_for_bb): Fix up formatting.
4767 (expr_expected_value_1, expr_expected_value): Add predictor argument,
4768 fill what it points to if non-NULL.
4769 (tree_predict_by_opcode): Adjust caller, use the predictor.
4770 * predict.def (PRED_COMPARE_AND_SWAP): Add.
4772 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
4774 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
4775 proper constant for the store mode.
4777 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
4779 * symtab.c (change_decl_assembler_name): Fix transparent alias
4782 2014-03-16 Renlin Li <Renlin.Li@arm.com>
4784 * config/aarch64/aarch64.c: Correct the comments about the
4785 aarch64 stack layout.
4787 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
4789 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
4790 check for GF_OMP_FOR_KIND_FOR.
4792 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
4794 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
4795 ymm and zmm register names.
4797 2014-03-17 Jakub Jelinek <jakub@redhat.com>
4800 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
4801 note creation for the 2010-08-31 changes.
4803 2014-03-17 Marek Polacek <polacek@redhat.com>
4806 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
4807 as -fno-tree-loop-vectorize.
4808 (expand_omp_simd): Likewise.
4810 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
4812 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
4813 (eligible_for_call_delay): New prototype.
4814 * config/sparc/sparc.c (tls_call_delay): Rename into...
4815 (eligible_for_call_delay): ...this. Return false if the instruction
4816 cannot be put in the delay slot of a branch.
4817 (eligible_for_restore_insn): Simplify.
4818 (eligible_for_return_delay): Return false if the instruction cannot be
4819 put in the delay slot of a branch and simplify.
4820 (eligible_for_sibcall_delay): Return false if the instruction cannot be
4821 put in the delay slot of a branch.
4822 * config/sparc/sparc.md (fix_ut699): New attribute.
4823 (tls_call_delay): Delete.
4824 (in_call_delay): Reimplement.
4825 (eligible_for_sibcall_delay): Rename into...
4826 (in_sibcall_delay): ...this.
4827 (eligible_for_return_delay): Rename into...
4828 (in_return_delay): ...this.
4829 (in_branch_delay): Reimplement.
4830 (in_uncond_branch_delay): Delete.
4831 (in_annul_branch_delay): Delete.
4833 2014-03-14 Richard Henderson <rth@redhat.com>
4836 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
4837 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
4838 (*floathi<X87MODEF>2_i387_with_temp): Remove.
4839 (floathi splitters): Remove.
4840 (float<SWI48x>xf2): New pattern.
4841 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
4842 code that tried to handle DImode for 32-bit, but which was excluded
4843 by the pattern's condition. Drop allocation of stack temporary.
4844 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
4845 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
4846 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
4847 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
4848 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
4849 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
4850 (*float<SWI48><MODEF>2_sse_interunit): Remove.
4851 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
4852 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
4853 (*float<SWI48x><X87MODEF>2_i387): Remove.
4854 (all float _with_temp splitters): Remove.
4855 (*float<SWI48x><MODEF>2_i387): New pattern.
4856 (*float<SWI48><MODEF>2_sse): New pattern.
4857 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
4858 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
4860 2014-03-14 Jakub Jelinek <jakub@redhat.com>
4861 Marek Polacek <polacek@redhat.com>
4864 * common.opt (dump_base_name_prefixed): New Variable.
4865 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
4866 if x_dump_base_name_prefixed is already set, set it at the end.
4868 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
4870 PR rtl-optimization/60508
4871 * lra-constraints.c (get_reload_reg): Add new parameter
4873 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
4874 Pass the new parameter values.
4876 2014-03-14 Richard Biener <rguenther@suse.de>
4878 * common.opt: Revert unintented changes from r205065.
4881 2014-03-14 Richard Biener <rguenther@suse.de>
4884 * cfghooks.c (split_block): Properly adjust all loops the
4885 block was a latch of.
4887 2014-03-14 Martin Jambor <mjambor@suse.cz>
4890 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
4893 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
4896 * config/avr/avr.c (avr_set_current_function): Pass function name
4897 through default_strip_name_encoding before sanity checking instead
4898 of skipping the first char of the assembler name.
4900 2014-03-13 Richard Henderson <rth@redhat.com>
4903 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
4904 (ix86_force_to_memory, ix86_free_from_memory): Remove.
4905 * config/i386/i386-protos.h: Likewise.
4906 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
4907 in the expander instead of a splitter.
4908 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
4909 any possibility of requiring a memory.
4910 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
4911 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
4912 (fp branch splitters): Update for ix86_split_fp_branch.
4913 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
4914 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
4915 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
4916 (*fop_<MODEF>_2_i387): Remove f/r alternative.
4917 (*fop_<MODEF>_3_i387): Likewise.
4918 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
4919 (splitters for the fop_* register patterns): Remove.
4920 (fscalexf4_i387): Rename from *fscalexf4_i387.
4921 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
4923 2014-03-13 Jakub Jelinek <jakub@redhat.com>
4925 PR tree-optimization/59779
4926 * tree-dfa.c (get_ref_base_and_extent): Use double_int
4927 type for bitsize and maxsize instead of HOST_WIDE_INT.
4929 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
4931 PR rtl-optimization/57320
4932 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
4933 the CFG after thread_prologue_and_epilogue_insns.
4935 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
4937 PR rtl-optimization/57189
4938 * lra-constraints.c (process_alt_operands): Disfavor spilling
4941 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
4943 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
4945 2014-03-13 Jakub Jelinek <jakub@redhat.com>
4947 PR tree-optimization/59025
4949 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
4950 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
4952 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
4955 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
4956 calls of avr_out_plus_1.
4958 2014-03-13 Bin Cheng <bin.cheng@arm.com>
4960 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
4961 BB's single pred and update the father loop's latch info later.
4963 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
4965 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
4969 (VEC_base): Likewise.
4970 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
4971 registers, we need to swap double words in little endian mode.
4973 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
4974 to be a container mode for 128-bit integer operations added in ISA
4975 2.07. Unlike TImode and PTImode, the preferred register set is
4976 the Altivec/VMX registers for the 128-bit operations.
4978 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
4980 (rs6000_split_128bit_ok_p): Likewise.
4982 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
4983 macros for creating ISA 2.07 normal and overloaded builtin
4984 functions with 3 arguments.
4985 (BU_P8V_OVERLOAD_3): Likewise.
4986 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
4987 for use as overloaded functions.
4988 (VPERM_1TI_UNS): Likewise.
4989 (VSEL_1TI): Likewise.
4990 (VSEL_1TI_UNS): Likewise.
4991 (ST_INTERNAL_1ti): Likewise.
4992 (LD_INTERNAL_1ti): Likewise.
4993 (XXSEL_1TI): Likewise.
4994 (XXSEL_1TI_UNS): Likewise.
4995 (VPERM_1TI): Likewise.
4996 (VPERM_1TI_UNS): Likewise.
4997 (XXPERMDI_1TI): Likewise.
4998 (SET_1TI): Likewise.
4999 (LXVD2X_V1TI): Likewise.
5000 (STXVD2X_V1TI): Likewise.
5001 (VEC_INIT_V1TI): Likewise.
5002 (VEC_SET_V1TI): Likewise.
5003 (VEC_EXT_V1TI): Likewise.
5004 (EQV_V1TI): Likewise.
5005 (NAND_V1TI): Likewise.
5006 (ORC_V1TI): Likewise.
5007 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
5008 added in ISA 2.07. Add both normal 'altivec' builtins, and the
5010 (VADDUQM): Likewise.
5011 (VSUBCUQ): Likewise.
5012 (VADDEUQM): Likewise.
5013 (VADDECUQ): Likewise.
5014 (VSUBEUQM): Likewise.
5015 (VSUBECUQ): Likewise.
5017 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5018 __int128_t and __uint128_t types.
5019 (__uint128_type): Likewise.
5020 (altivec_categorize_keyword): Add support for vector __int128_t,
5021 vector __uint128_t, vector __int128, and vector unsigned __int128
5022 as a container type for TImode operations that need to be done in
5023 VSX/Altivec registers.
5024 (rs6000_macro_to_expand): Likewise.
5025 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5026 to support 128-bit integer instructions vaddcuq, vadduqm,
5027 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5028 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5030 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5031 for V1TImode, and set up preferences to use VSX/Altivec registers.
5032 Setup VSX reload handlers.
5033 (rs6000_debug_reg_global): Likewise.
5034 (rs6000_init_hard_regno_mode_ok): Likewise.
5035 (rs6000_preferred_simd_mode): Likewise.
5036 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5037 (easy_altivec_constant): Likewise.
5038 (output_vec_const_move): Likewise.
5039 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5041 (rs6000_expand_vector_extract): Likewise.
5042 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
5044 (rs6000_const_vec): Add support for V1TImode.
5045 (rs6000_emit_le_vsx_load): Swap double words when loading or
5046 storing TImode/V1TImode.
5047 (rs6000_emit_le_vsx_store): Likewise.
5048 (rs6000_emit_le_vsx_move): Likewise.
5049 (rs6000_emit_move): Add support for V1TImode.
5050 (altivec_expand_ld_builtin): Likewise.
5051 (altivec_expand_st_builtin): Likewise.
5052 (altivec_expand_vec_init_builtin): Likewise.
5053 (altivec_expand_builtin): Likewise.
5054 (rs6000_init_builtins): Add support for V1TImode type. Add
5055 support for ISA 2.07 128-bit integer builtins. Define type names
5056 for the VSX/Altivec vector types.
5057 (altivec_init_builtins): Add support for overloaded vector
5058 functions with V1TImode type.
5059 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
5060 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
5062 (rs6000_split_128bit_ok_p): Likewise.
5063 (rs6000_handle_altivec_attribute): Create V1TImode from vector
5064 __int128_t and vector __uint128_t.
5066 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
5067 and mode attributes.
5074 (VS_scalar): Likewise.
5075 (VS_double): Likewise.
5076 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
5078 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
5079 we support the ISA 2.07 128-bit integer arithmetic instructions.
5080 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
5081 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
5082 and TImode types for use with the builtin functions.
5083 (V1TI_type_node): Likewise.
5084 (unsigned_V1TI_type_node): Likewise.
5085 (intTI_type_internal_node): Likewise.
5086 (uintTI_type_internal_node): Likewise.
5088 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
5089 128-bit builtin functions.
5090 (UNSPEC_VADDEUQM): Likewise.
5091 (UNSPEC_VADDECUQ): Likewise.
5092 (UNSPEC_VSUBCUQ): Likewise.
5093 (UNSPEC_VSUBEUQM): Likewise.
5094 (UNSPEC_VSUBECUQ): Likewise.
5095 (VM): Add V1TImode to vector mode iterators.
5097 (VI_unit): Likewise.
5098 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
5099 (altivec_vaddcuq): Likewise.
5100 (altivec_vsubuqm): Likewise.
5101 (altivec_vsubcuq): Likewise.
5102 (altivec_vaddeuqm): Likewise.
5103 (altivec_vaddecuq): Likewise.
5104 (altivec_vsubeuqm): Likewise.
5105 (altivec_vsubecuq): Likewise.
5107 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
5109 (BOOL_128): Likewise.
5110 (BOOL_REGS_OUTPUT): Likewise.
5111 (BOOL_REGS_OP1): Likewise.
5112 (BOOL_REGS_OP2): Likewise.
5113 (BOOL_REGS_UNARY): Likewise.
5114 (BOOL_REGS_AND_CR0): Likewise.
5116 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
5117 128-bit integer builtin support.
5118 (vec_vadduqm): Likewise.
5119 (vec_vaddecuq): Likewise.
5120 (vec_vaddeuqm): Likewise.
5121 (vec_vsubecuq): Likewise.
5122 (vec_vsubeuqm): Likewise.
5123 (vec_vsubcuq): Likewise.
5124 (vec_vsubuqm): Likewise.
5126 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5127 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
5128 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
5129 128-bit integer add/subtract to ISA 2.07.
5131 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
5133 * config/arc/arc.c (arc_predicate_delay_insns):
5134 Fix third argument passed to conditionalize_nonjump.
5136 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
5138 * config/aarch64/aarch64-builtins.c
5139 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
5140 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
5141 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
5142 instead of __builtin_lfloor.
5143 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
5145 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5147 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
5148 (tree_ssa_ifcombine_bb_1): New function.
5149 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
5150 is an empty forwarder block to then_bb or vice versa and then_bb
5151 and else_bb are effectively swapped.
5153 2014-03-12 Christian Bruel <christian.bruel@st.com>
5156 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
5157 REG_CFA_DEF_CFA note.
5158 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
5159 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
5161 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
5163 PR tree-optimization/60454
5164 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
5166 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5168 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
5169 Do not define target_cpu_default2 to generic.
5170 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
5171 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
5172 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
5174 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5175 Marc Glisse <marc.glisse@inria.fr>
5177 PR tree-optimization/60502
5178 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
5179 instead of build_low_bits_mask.
5181 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5184 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
5185 if there are multiple uses, but op doesn't live on E edge.
5186 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
5187 clobber stmts before __builtin_unreachable.
5189 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
5191 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
5192 hard_frame_pointer_rtx.
5193 * cse.c (cse_insn): Remove volatile check.
5194 * cselib.c (cselib_process_insn): Likewise.
5195 * dse.c (scan_insn): Likewise.
5197 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5199 * config/arc/arc.c (conditionalize_nonjump): New function,
5201 (arc_ifcvt): ... this.
5202 (arc_predicate_delay_insns): Use it.
5204 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5206 * config/arc/predicates.md (extend_operand): During/after reload,
5207 allow const_int_operand.
5208 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
5209 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
5210 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
5212 (umulsi3_highpart_i): Likewise.
5214 2014-03-11 Richard Biener <rguenther@suse.de>
5216 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
5217 Add asserts to guard possible wrong-code bugs.
5219 2014-03-11 Richard Biener <rguenther@suse.de>
5221 PR tree-optimization/60429
5222 PR tree-optimization/60485
5223 * tree-ssa-structalias.c (set_union_with_increment): Properly
5224 take into account all fields that overlap the shifted vars.
5225 (do_sd_constraint): Likewise.
5226 (do_ds_constraint): Likewise.
5227 (get_constraint_for_ptr_offset): Likewise.
5229 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
5231 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
5232 (nios2_compute_frame_layout):
5233 Add calculation of cfun->machine->fp_save_offset.
5234 (nios2_expand_prologue): Correct setting of frame pointer register
5236 (nios2_expand_epilogue): Update recovery of stack pointer from
5237 frame pointer accordingly.
5238 (nios2_initial_elimination_offset): Update calculation of offset
5239 for eliminating to HARD_FRAME_POINTER_REGNUM.
5241 2014-03-10 Jakub Jelinek <jakub@redhat.com>
5244 * ipa.c (symtab_remove_unreachable_nodes): Don't call
5245 cgraph_get_create_node on VAR_DECLs.
5247 2014-03-10 Richard Biener <rguenther@suse.de>
5250 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
5252 2014-03-08 Douglas B Rupp <rupp@gnat.com>
5254 * config/vms/vms.opt (vms_float_format): New variable.
5256 2014-03-08 Tobias Burnus <burnus@net-b.de>
5258 * doc/invoke.texi (-fcilkplus): Update implementation status.
5260 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
5261 Richard Biener <rguenther@suse.de>
5263 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
5264 consistently accross all TUs.
5265 (run_gcc): Enable -fshort-double automatically at link at link-time
5266 and disallow override.
5268 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
5271 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
5272 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
5273 if they can't be used.
5275 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5277 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
5278 for Solaris 11/x86 ld.
5279 * configure: Regenerate.
5281 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5283 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
5284 (LIB_TLS_SPEC): Save as ld_tls_libs.
5285 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
5286 (HAVE_AS_IX86_TLSLDM): New test.
5287 * configure, config.in: Regenerate.
5288 * config/i386/i386.c (legitimize_tls_address): Fall back to
5289 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
5290 cannot support TLS_MODEL_LOCAL_DYNAMIC.
5291 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
5292 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
5294 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
5296 * common.opt (fira-loop-pressure): Mark as optimization.
5298 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
5300 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
5301 an OpenMP mappable type.
5303 2014-03-06 Matthias Klose <doko@ubuntu.com>
5305 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
5306 MULTILIB_OSDIRNAMES is not defined.
5308 2014-03-06 Jakub Jelinek <jakub@redhat.com>
5309 Meador Inge <meadori@codesourcery.com>
5312 * config/arm/arm.c (arm_tls_symbol_p): Remove.
5313 (arm_legitimize_address): Call legitimize_tls_address for any
5314 arm_tls_referenced_p expression, handle constant addend. Call it
5315 before testing for !TARGET_ARM.
5316 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
5318 2014-03-06 Richard Biener <rguenther@suse.de>
5324 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5326 * tree-streamer.c (record_common_node): Assert we don't record
5327 nodes with type double.
5328 (preload_common_node): Skip type double, complex double and double
5329 pointer since it is now frontend dependent due to fshort-double option.
5331 2014-03-06 Richard Biener <rguenther@suse.de>
5333 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
5334 or -fno-lto is specified and the linker has full plugin support.
5335 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
5336 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
5337 * lto-wrapper.c (merge_and_complain): Merge compile-time
5338 optimization levels.
5339 (run_gcc): And pass it through to the link options.
5341 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
5345 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5347 * cselib.c (remove_useless_values): Skip to avoid quadratic
5348 behavior if the condition moved from...
5349 (cselib_process_insn): ... here holds.
5351 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5354 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
5355 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
5358 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
5359 (TM_H): Add x86-tune.def.
5361 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5363 * config/aarch64/aarch64.c (generic_tunings):
5364 Use cortexa57_extra_costs.
5366 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5369 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
5370 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
5371 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
5374 2014-03-04 Heiher <r@hev.cc>
5376 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
5377 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
5379 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
5381 * config/i386/predicates.md (const2356_operand): Change to ...
5382 (const2367_operand): ... this.
5383 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
5385 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5386 (*avx512pf_scatterpf<mode>sf): Ditto.
5387 (avx512pf_scatterpf<mode>df): Ditto.
5388 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5389 (*avx512pf_scatterpf<mode>df): Ditto.
5390 * config/i386/i386.c (ix86_expand_builtin): Update
5391 incorrect hint operand error message.
5393 2014-03-04 Richard Biener <rguenther@suse.de>
5395 * lto-section-in.c (lto_get_section_data): Fix const cast.
5397 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5399 * tree-streamer.c (record_common_node): Assert we don't record
5400 nodes with type double.
5401 (preload_common_node): Skip type double, complex double and double
5402 pointer since it is now frontend dependent due to fshort-double option.
5404 2014-03-04 Richard Biener <rguenther@suse.de>
5407 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
5408 (lto_input_toplevel_asms): Likewise.
5409 * lto-section-in.c (lto_get_section_data): Instead do it here
5412 2014-03-04 Richard Biener <rguenther@suse.de>
5414 PR tree-optimization/60382
5415 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
5416 dead PHIs a reduction.
5418 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
5420 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
5422 (_mm_prefetch): Move out of GCC target("sse") pragma.
5423 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
5424 GCC target("prfchw") pragma.
5425 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
5427 * config/i386/i386.c (ix86_option_override_internal): Enable
5428 -mprfchw with -mprefetchwt1.
5430 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5432 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
5435 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5437 * opts.h (CL_PCH_IGNORE): Define.
5438 * targhooks.c (option_affects_pch_p):
5439 Return false for options that have CL_PCH_IGNORE set.
5440 * opt-functions.awk: Process PchIgnore.
5441 * doc/options.texi: Document PchIgnore.
5443 * config/arc/arc.opt (misize): Add PchIgnore property.
5445 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5447 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
5448 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
5449 constraint on constants to permit them being loaded into
5450 GENERAL_REGS or BASE_REGS.
5452 2014-03-03 Nick Clifton <nickc@redhat.com>
5454 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
5455 anti-cacnonical alternatives.
5456 (negandhi3_real): New pattern.
5457 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
5459 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5461 * config/avr/avr-mcus.def: Remove atxmega16x1.
5462 * config/avr/avr-tables.opt: Regenerate.
5463 * config/avr/t-multilib: Regenerate.
5464 * doc/avr-mmcu.texi: Regenerate.
5466 2014-03-03 Tobias Grosser <tobias@grosser.es>
5467 Mircea Namolaru <mircea.namolaru@inria.fr>
5469 PR tree-optimization/58028
5470 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
5473 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5475 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
5476 not handled by recognizers.
5478 2014-03-03 Jakub Jelinek <jakub@redhat.com>
5481 * function.c (expand_function_end): Don't emit
5482 clobber_return_register sequence if clobber_after is a BARRIER.
5483 * cfgexpand.c (construct_exit_block): Append instructions before
5484 return_label to prev_bb.
5486 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5488 * config/rs6000/constraints.md: Document reserved use of "wc".
5490 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5493 * ipa.c (function_and_variable_visibility): When dissolving comdat
5494 group, also set all symbols to local.
5496 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5501 2013-12-14 Jan Hubicka <jh@suse.cz>
5503 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
5505 2014-03-02 Jon Beniston <jon@beniston.com>
5511 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
5512 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
5513 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
5514 (simple_return, *simple_return): New patterns
5515 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
5516 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
5518 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
5520 * dwarf2out.c (gen_subprogram_die): Tidy.
5522 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
5525 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
5526 (*mov_t_msb_neg_negc): ... this new insn.
5528 2014-02-28 Jason Merrill <jason@redhat.com>
5531 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
5534 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
5537 * dwarf2out.c (decltype_auto_die): New static.
5538 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
5539 (gen_type_die_with_usage): Handle 'decltype(auto)'.
5540 (is_cxx_auto): Likewise.
5542 2014-02-28 Ian Bolton <ian.bolton@arm.com>
5544 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
5545 we are not using general regs only.
5547 2014-02-28 Richard Biener <rguenther@suse.de>
5550 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
5551 previous fix and only allow to remove trivial pre-headers
5552 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
5553 (remove_forwarder_block): Properly update the latch of a loop.
5555 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5558 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
5559 (cselib_preserved_hash_table): New.
5560 (preserve_constants_and_equivs): Move preserved vals to it.
5561 (cselib_find_slot): Look it up first.
5562 (cselib_init): Initialize it.
5563 (cselib_finish): Release it.
5564 (dump_cselib_table): Dump it.
5566 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5569 * cselib.c (remove_useless_values): Skip to avoid quadratic
5570 behavior if the condition moved from...
5571 (cselib_process_insn): ... here holds.
5573 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5576 * var-tracking.c (vt_initialize): Apply the same condition to
5577 preserve the CFA base value.
5579 2014-02-28 Joey Ye <joey.ye@arm.com>
5582 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
5583 if reload in progress or completed.
5585 2014-02-28 Tobias Burnus <burnus@net-b.de>
5588 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
5591 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
5593 * doc/tm.texi.in (Condition Code Status): Update documention for
5594 relative locations of cc0-setter and cc0-user.
5596 2014-02-27 Jeff Law <law@redhat.com>
5598 PR rtl-optimization/52714
5599 * combine.c (try_combine): When splitting an unrecognized PARALLEL
5600 into two independent simple sets, if I3 is a jump, ensure the
5601 pattern we place into I3 is a (set (pc) ...).
5603 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
5604 Jeff Law <law@redhat.com>
5606 PR rtl-optimization/49847
5607 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
5608 are in different blocks.
5609 * doc/tm.texi (Condition Code Status): Update documention for
5610 relative locations of cc0-setter and cc0-user.
5612 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
5615 * lra.c (lra_emit_add): Check SUBREG too.
5617 2014-02-27 Andreas Schwab <schwab@suse.de>
5619 * config/m68k/m68k.c (m68k_option_override): Disable
5620 -flive-range-shrinkage for classic m68k.
5621 (m68k_override_options_after_change): Likewise.
5623 2014-02-27 Marek Polacek <polacek@redhat.com>
5626 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
5627 -Wmaybe-uninitialized.
5629 2014-02-27 Alan Modra <amodra@gmail.com>
5632 * reload1.c (emit_input_reload_insns): When reload_override_in,
5633 set old to rl->in_reg when rl->in_reg is a subreg.
5635 2014-02-26 Richard Biener <rguenther@suse.de>
5638 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
5640 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
5642 * common/config/i386/predicates.md (const1256_operand): Remove.
5643 (const2356_operand): New.
5644 (const_1_to_2_operand): Remove.
5645 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5646 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5647 (*avx512pf_gatherpf<mode>sf): Ditto.
5648 (avx512pf_gatherpf<mode>df): Ditto.
5649 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5650 (*avx512pf_gatherpf<mode>df): Ditto.
5651 (avx512pf_scatterpf<mode>sf): Ditto.
5652 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5653 (*avx512pf_scatterpf<mode>sf): Ditto.
5654 (avx512pf_scatterpf<mode>df): Ditto.
5655 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5656 (*avx512pf_scatterpf<mode>df): Ditto.
5657 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
5659 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
5661 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
5662 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
5663 (_mm512_mask_testn_epi64_mask): Move to ...
5664 * config/i386/avx512cdintrin.h: Here.
5665 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
5666 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
5667 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
5668 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
5669 TARGET_AVX512F from TARGET_AVX512CD.
5671 2014-02-26 Richard Biener <rguenther@suse.de>
5674 * ipa.c (walk_polymorphic_call_targets): Properly guard
5675 call to inline_update_overall_summary.
5677 2014-02-26 Bin Cheng <bin.cheng@arm.com>
5680 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
5681 and latches only if requested. Fix latch if it is removed.
5682 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
5683 LOOPS_HAVE_PREHEADERS.
5685 2014-02-25 Andrew Pinski <apinski@cavium.com>
5687 * builtins.c (expand_builtin_thread_pointer): Create a new target
5688 when the target is NULL.
5690 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
5692 PR rtl-optimization/60317
5693 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
5694 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
5695 * lra-assigns.c: Include params.h.
5696 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
5697 other reload pseudos considerations.
5699 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5701 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
5702 to use canonical form for nor<mode>3.
5704 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5707 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
5710 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
5712 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
5713 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
5714 (ix86_handle_option): Handle OPT_mprefetchwt1.
5715 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
5716 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
5718 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
5719 OPTION_MASK_ISA_PREFETCHWT1.
5720 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
5721 (PTA_PREFETCHWT1): New.
5722 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
5723 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
5724 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
5725 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
5726 (*prefetch_avx512pf_<mode>_: Change into ...
5727 (*prefetch_prefetchwt1_<mode>: This.
5728 * config/i386/i386.opt (mprefetchwt1): New.
5729 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
5730 (_mm_prefetch): Handle intent to write.
5731 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
5733 2014-02-25 Richard Biener <rguenther@suse.de>
5736 * emit-rtl.c (mem_attrs_htab): Remove.
5737 (mem_attrs_htab_hash): Likewise.
5738 (mem_attrs_htab_eq): Likewise.
5739 (set_mem_attrs): Always allocate new mem-attrs when something changed.
5740 (init_emit_once): Do not allocate mem_attrs_htab.
5742 2014-02-25 Richard Biener <rguenther@suse.de>
5745 * lto-opts.c (lto_write_options): Output non-explicit conservative
5746 -fwrapv, -fno-trapv and -fno-strict-overflow.
5747 * lto-wrapper.c (merge_and_complain): Handle merging those options.
5748 (run_gcc): And pass them through.
5750 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
5752 * sel-sched.c (calculate_new_fences): New parameter ptime.
5753 Calculate it as a maximum over all fence cycles.
5754 (sel_sched_region_2): Adjust the call to calculate_new_fences.
5755 Print the final schedule timing when sched_verbose.
5757 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
5759 PR rtl-optimization/60292
5760 * sel-sched.c (fill_vec_av_set): Do not reset target availability
5761 bit fot the fence instruction.
5763 2014-02-24 Alangi Derick <alangiderick@gmail.com>
5765 * calls.h: Fix typo in comment.
5767 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
5769 * config/pa/pa.c (pa_output_move_double): Don't valididate when
5770 adjusting offsetable addresses.
5772 2014-02-24 Guozhi Wei <carrot@google.com>
5774 * sparseset.h (sparseset_pop): Fix the wrong index.
5776 2014-02-24 Walter Lee <walt@tilera.com>
5778 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
5779 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
5781 * common/config/tilegx/tilegx-common.c
5782 (TARGET_DEFAULT_TARGET_FLAGS): Define.
5783 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
5785 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
5786 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
5787 (tilegx_gimplify_va_arg_expr): Handle big endian.
5788 (tilegx_expand_unaligned_load): Ditto.
5789 (tilegx_expand_unaligned_store): Ditto.
5790 (TARGET_RETURN_IN_MSB): New.
5791 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
5792 (TARGET_ENDIAN_DEFAULT): New.
5793 (TARGET_BIG_ENDIAN): Handle big endian.
5794 (BYTES_BIG_ENDIAN): Ditto.
5795 (WORDS_BIG_ENDIAN): Ditto.
5796 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
5799 * config/tilegx/tilegx.md (extv): Handle big endian.
5801 (insn_st<n>): Ditto.
5802 (insn_st<n>_add<bitsuffix>): Ditto.
5803 (insn_stnt<n>): Ditto.
5804 (insn_stnt<n>_add<bitsuffix>):Ditto.
5805 (vec_interleave_highv8qi): Handle big endian.
5806 (vec_interleave_highv8qi_be): New.
5807 (vec_interleave_highv8qi_le): New.
5808 (insn_v1int_h): Handle big endian.
5809 (vec_interleave_lowv8qi): Handle big endian.
5810 (vec_interleave_lowv8qi_be): New.
5811 (vec_interleave_lowv8qi_le): New.
5812 (insn_v1int_l): Handle big endian.
5813 (vec_interleave_highv4hi): Handle big endian.
5814 (vec_interleave_highv4hi_be): New.
5815 (vec_interleave_highv4hi_le): New.
5816 (insn_v2int_h): Handle big endian.
5817 (vec_interleave_lowv4hi): Handle big endian.
5818 (vec_interleave_lowv4hi_be): New.
5819 (vec_interleave_lowv4hi_le): New.
5820 (insn_v2int_l): Handle big endian.
5821 (vec_interleave_highv2si): Handle big endian.
5822 (vec_interleave_highv2si_be): New.
5823 (vec_interleave_highv2si_le): New.
5824 (insn_v4int_h): Handle big endian.
5825 (vec_interleave_lowv2si): Handle big endian.
5826 (vec_interleave_lowv2si_be): New.
5827 (vec_interleave_lowv2si_le): New.
5828 (insn_v4int_l): Handle big endian.
5829 * config/tilegx/tilegx.opt (mbig-endian): New option.
5830 (mlittle-endian): New option.
5831 * doc/install.texi: Document tilegxbe-linux.
5832 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
5834 2014-02-24 Martin Jambor <mjambor@suse.cz>
5837 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
5838 there are no parameter descriptors.
5840 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
5842 PR rtl-optimization/60268
5843 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
5844 initialization to ...
5845 (sched_rgn_init): ... here.
5846 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
5848 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5850 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
5853 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
5855 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
5858 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5860 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
5861 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
5863 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5865 * config/microblaze/predicates.md: Add cmp_op predicate.
5866 * config/microblaze/microblaze.md: Add branch_compare instruction
5867 which uses cmp_op predicate and emits cmp insn before branch.
5868 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
5869 to microblaze_expand_conditional_branch and consolidate logic.
5870 (microblaze_expand_conditional_branch): emit branch_compare
5871 insn instead of handling cmp op separate from branch insn.
5873 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5875 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
5878 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5880 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
5881 define_insn with define_expand and new define_insn
5882 *altivec_lve<VI_char>x_internal.
5883 (altivec_stve<VI_char>x): Replace define_insn with define_expand
5884 and new define_insn *altivec_stve<VI_char>x_internal.
5885 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
5887 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
5889 (altivec_expand_stvex_be): New function.
5891 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
5893 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
5894 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
5895 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
5896 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
5898 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
5901 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
5902 instead of emit_move_insn.
5904 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5906 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
5908 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
5909 gen_altivec_vsumsws.
5911 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5913 * config/rs6000/altivec.md (altivec_lvxl): Rename as
5914 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
5915 (altivec_lvxl_<mode>): New define_expand incorporating
5916 -maltivec=be semantics where needed.
5917 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
5918 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
5919 semantics where needed.
5920 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
5921 (altivec_stvx_<mode>): New define_expand incorporating
5922 -maltivec=be semantics where needed.
5923 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
5924 VM2 iterator instead of V4SI.
5925 (altivec_stvxl_<mode>): New define_expand incorporating
5926 -maltivec=be semantics where needed.
5927 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
5928 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
5929 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
5930 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
5931 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
5932 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
5933 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
5934 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
5935 ALTIVEC_BUILTIN_STVXL.
5936 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
5937 (altivec_expand_stvx_be): Likewise.
5938 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
5939 (altivec_expand_lvx_be): Likewise.
5940 (altivec_expand_stvx_be): Likewise.
5941 (altivec_expand_builtin): Add cases for
5942 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
5943 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
5944 (altivec_init_builtins): Add definitions for
5945 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
5946 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
5948 2014-02-21 Catherine Moore <clm@codesourcery.com>
5950 * doc/invoke.texi (mvirt, mno-virt): Document.
5951 * config/mips/mips.opt (mvirt): New option.
5952 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
5954 2014-02-21 Richard Biener <rguenther@suse.de>
5956 PR tree-optimization/60276
5957 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
5958 (STMT_VINFO_MIN_NEG_DIST): New macro.
5959 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
5960 STMT_VINFO_MIN_NEG_DIST.
5961 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
5962 made for negative dependence distances still hold.
5964 2014-02-21 Richard Biener <rguenther@suse.de>
5967 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
5968 DECL_INITIAL for globals not in the current function context.
5970 2014-02-21 Jakub Jelinek <jakub@redhat.com>
5972 PR tree-optimization/56490
5973 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
5974 * tree-ssa-uninit.c: Include params.h.
5975 (compute_control_dep_chain): Add num_calls argument, return false
5976 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
5977 num_calls to recursive call.
5978 (find_predicates): Change dep_chain into normal array,
5979 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
5980 variable and adjust compute_control_dep_chain caller.
5981 (find_def_preds): Likewise.
5983 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
5985 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
5986 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
5988 2014-02-21 Nick Clifton <nickc@redhat.com>
5990 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
5991 (pushhi1): Likewise.
5992 (popqi1): Add mode to pre_dec.
5995 2014-02-21 Jakub Jelinek <jakub@redhat.com>
5997 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
5998 mode for mask of V8SFmode permutation.
6000 2014-02-20 Richard Henderson <rth@redhat.com>
6003 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
6004 a new pseudo for OLDVAL.
6006 2014-02-20 Jakub Jelinek <jakub@redhat.com>
6009 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
6010 gen_reg_rtx if d->testing_p.
6011 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
6012 if d->testing_p and we will certainly return true.
6013 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6016 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6018 * emit-rtl.c (gen_reg_rtx): Assert that
6019 crtl->emit.regno_pointer_align_length is non-zero.
6021 2014-02-20 Richard Henderson <rth@redhat.com>
6024 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6025 on failure the store back into EXPECT.
6027 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6028 Sandra Loosemore <sandra@codesourcery.com>
6030 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6031 * config/nios2/nios2.c (nios2_function_profiler): Add
6032 -fPIC (flag_pic == 2) support.
6033 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6034 (nios2_large_offset_p): New function.
6035 (nios2_unspec_reloc_p): Move up position, update to use
6036 nios2_large_offset_p.
6037 (nios2_unspec_address): Remove function.
6038 (nios2_unspec_offset): New function.
6039 (nios2_large_got_address): New function.
6040 (nios2_got_address): Add large offset support.
6041 (nios2_legitimize_tls_address): Update usage of removed and new
6043 (nios2_symbol_binds_local_p): New function.
6044 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
6045 (nios2_legitimize_address): Update to use nios2_large_offset_p.
6046 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
6047 (nios2_print_operand): Merge H/L processing, add hiadj/lo
6048 processing for (const (unspec ...)).
6049 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
6051 2014-02-20 Richard Biener <rguenther@suse.de>
6053 * tree-cfg.c (replace_uses_by): Mark altered BBs before
6054 doing the substitution.
6055 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
6057 2014-02-20 Martin Jambor <mjambor@suse.cz>
6060 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
6061 info when checking whether lattices are bottom.
6063 2014-02-20 Richard Biener <rguenther@suse.de>
6066 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
6069 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
6072 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
6073 parameter specifying the scaling.
6074 (inline_call): Update.
6075 (want_inline_recursively): Guard division by zero.
6076 (recursive_inlining): Update.
6077 * ipa-inline.h (clone_inlined_nodes): Update.
6079 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6082 * config/i386/i386.c (classify_argument): Pass structures of size
6083 64 bytes or less in register.
6085 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6086 Kirill Yukhin <kirill.yukhin@intel.com>
6088 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
6089 (_mm_rcp28_round_ss): Ditto.
6090 (_mm_rsqrt28_round_sd): Ditto.
6091 (_mm_rsqrt28_round_ss): Ditto.
6092 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
6093 (_mm_rcp14_round_ss): Ditto.
6094 (_mm_rsqrt14_round_sd): Ditto.
6095 (_mm_rsqrt14_round_ss): Ditto.
6096 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
6097 the first input operand, get rid of match_dup.
6098 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
6100 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
6102 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
6103 operand as the first input operand, set type attribute.
6104 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
6106 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
6107 operand as the first input operand, set type attribute.
6109 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6111 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
6114 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
6117 * config/i386/i386.c (construct_container): Remove TFmode check
6118 for X86_64_INTEGER_CLASS.
6120 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
6123 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
6124 only when -Wpsabi is enabled.
6126 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
6129 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
6130 passing arrays in registers are the same as for structs, so remove the
6131 special case for them.
6133 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
6135 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
6136 destination type, extract only the valid bits if the source type is not
6137 integral and has a different mode.
6139 2014-02-19 Richard Biener <rguenther@suse.de>
6142 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
6145 2014-02-19 Richard Biener <rguenther@suse.de>
6148 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
6149 (ipa_modify_call_arguments): Emit an argument load explicitely and
6150 preserve virtual SSA form there and for the replacement call.
6151 Do not update SSA form nor free dominance info.
6153 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6155 * ipa.c (function_and_variable_visibility): Also clear WEAK
6156 flag when disolving COMDAT_GROUP.
6158 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6160 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
6161 * ipa-prop.c (ipa_set_jf_known_type): Return early when
6163 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
6164 do more sanity checks.
6165 (detect_type_change): Return true when giving up early.
6166 (compute_complex_assign_jump_func): Fix type parameter of
6167 ipa_set_ancestor_jf.
6168 (compute_complex_ancestor_jump_func): Likewise.
6169 (update_jump_functions_after_inlining): Fix updating of
6171 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
6173 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6175 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
6176 inline clones when edge disappears.
6178 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6181 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
6182 Split 64-bit moves into 2 patterns. Do not allow the use of
6183 direct move for TDmode in little endian, since the decimal value
6184 has little endian bytes within a word, but the 64-bit pieces are
6185 ordered in a big endian fashion, and normal subreg's of TDmode are
6187 (mov<mode>_64bit_dm): Likewise.
6188 (movtd_64bit_nodm): Likewise.
6190 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6192 PR tree-optimization/60174
6193 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
6194 statement of an SSA_NAME that occurs in an abnormal PHI node.
6196 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6199 * final.c (SEEN_BB): Remove.
6200 (SEEN_NOTE, SEEN_EMITTED): Renumber.
6201 (final_scan_insn): Don't force_source_line on second
6202 NOTE_INSN_BASIC_BLOCK.
6204 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
6207 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
6208 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
6209 (type_natural_mode): Warn ABI change when %zmm register is not
6210 available for AVX512F vector value passing.
6212 2014-02-18 Kai Tietz <ktietz@redhat.com>
6215 * config/i386/i386.c (ix86_expand_prologue): Use value in
6216 rax register as displacement when restoring %r10 or %rax.
6217 Fix wrong offset when restoring both registers.
6219 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6221 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
6222 assertion with conditional return.
6224 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6225 Uros Bizjak <ubizjak@gmail.com>
6228 * config/i386/driver-i386.c (host_detect_local_cpu): If
6229 YMM state is not saved by the OS, also clear has_f16c. Move
6230 CPUID 0x80000001 handling before YMM state saving checking.
6232 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
6234 PR rtl-optimization/58960
6235 * haifa-sched.c (alloc_global_sched_pressure_data): New,
6236 factored out from ...
6237 (sched_init): ... here.
6238 (free_global_sched_pressure_data): New, factored out from ...
6239 (sched_finish): ... here.
6240 * sched-int.h (free_global_sched_pressure_data): Declare.
6241 * sched-rgn.c (nr_regions_initial): New static global.
6242 (haifa_find_rgns): Initialize it.
6243 (schedule_region): Disable sched-pressure for the newly
6246 2014-02-17 Richard Biener <rguenther@suse.de>
6248 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
6249 release SSA defs of pattern stmts.
6251 2014-02-17 Richard Biener <rguenther@suse.de>
6253 * tree-inline.c (expand_call_inline): Release the virtual
6254 operand defined by the call we are about to inline.
6256 2014-02-17 Richard Biener <rguenther@suse.de>
6258 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
6260 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
6261 Ilya Tocar <ilya.tocar@intel.com>
6263 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
6264 arguments order in builtin.
6265 (_mm512_permutexvar_epi64): Ditto.
6266 (_mm512_mask_permutexvar_epi64): Ditto
6267 (_mm512_maskz_permutexvar_epi32): Ditto
6268 (_mm512_permutexvar_epi32): Ditto
6269 (_mm512_mask_permutexvar_epi32): Ditto
6271 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6273 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
6274 (p8_vmrgow): Likewise.
6276 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6278 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
6281 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
6284 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
6285 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
6286 into 64-bit and 32-bit moves. On 64-bit moves, add support for
6287 using direct move instructions on ISA 2.07. Also adjust
6288 instruction length for 64-bit.
6289 (mov<mode>_64bit, TFmode/TDmode): Likewise.
6290 (mov<mode>_32bit, TFmode/TDmode): Likewise.
6292 2014-02-15 Alan Modra <amodra@gmail.com>
6296 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
6297 find_replacement on parts of insn rtl that might be reloaded.
6299 2014-02-15 Richard Biener <rguenther@suse.de>
6301 PR tree-optimization/60183
6302 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
6303 (tree_ssa_phiprop): Calculate and free post-dominators.
6305 2014-02-14 Jeff Law <law@redhat.com>
6307 PR rtl-optimization/60131
6308 * ree.c (get_extended_src_reg): New function.
6309 (combine_reaching_defs): Use it rather than assuming location of REG.
6310 (find_and_remove_re): Verify first operand of extension is
6311 a REG before adding the insns to the copy list.
6313 2014-02-14 Roland McGrath <mcgrathr@google.com>
6315 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
6316 * configure: Regenerated.
6317 * config.in: Regenerated.
6318 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
6319 instead of ASM_SHORT.
6321 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
6322 Richard Earnshaw <rearnsha@arm.com>
6324 PR rtl-optimization/59535
6325 * lra-constraints.c (process_alt_operands): Encourage alternative
6326 when unassigned pseudo class is superset of the alternative class.
6327 (inherit_reload_reg): Don't inherit when optimizing for code size.
6328 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
6329 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
6330 modes not less than 4 for Thumb1.
6332 2014-02-14 Kyle McMartin <kyle@redhat.com>
6335 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
6337 2014-02-14 Richard Biener <rguenther@suse.de>
6339 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
6340 (get_frame_arg): Drop the assert with langhook types_compatible_p.
6341 Do not strip INDIRECT_REFs.
6343 2014-02-14 Richard Biener <rguenther@suse.de>
6346 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
6347 DECL_FUNCTION_SPECIFIC_TARGET.
6348 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
6349 * tree-streamer-out.c (pack_ts_target_option): Remove.
6350 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
6351 (write_ts_function_decl_tree_pointers): Do not stream
6352 DECL_FUNCTION_SPECIFIC_TARGET.
6353 * tree-streamer-in.c (unpack_ts_target_option): Remove.
6354 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
6355 (lto_input_ts_function_decl_tree_pointers): Do not stream
6356 DECL_FUNCTION_SPECIFIC_TARGET.
6358 2014-02-14 Jakub Jelinek <jakub@redhat.com>
6360 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
6361 (get_initial_def_for_induction, vectorizable_induction): Ignore
6362 debug stmts when looking for exit_phi.
6363 (vectorizable_live_operation): Fix up condition.
6365 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6367 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
6368 nreverse() because it changes the content of original tree list.
6370 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6372 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
6373 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
6375 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6377 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
6378 GNU coding standards.
6380 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6383 * dwarf2out.c (gen_subprogram_die): Don't call
6384 add_calling_convention_attribute if subr_die is old_die.
6386 2014-02-13 Sharad Singhai <singhai@google.com>
6388 * doc/optinfo.texi: Fix order of nodes.
6390 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
6392 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
6393 operands[2], not operands[3].
6395 2014-02-13 Richard Biener <rguenther@suse.de>
6398 * doc/install.texi (ISL): Update recommended version to 0.12.2,
6399 mention the possibility of an in-tree build.
6400 (CLooG): Update recommended version to 0.18.1, mention the
6401 possibility of an in-tree build and clarify that the ISL
6402 bundled with CLooG does not work.
6404 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6407 * expr.c (compress_float_constant): If x is a hard register,
6408 extend into a pseudo and then move to x.
6410 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
6412 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
6413 caused by bad second argument to warning_at() with -mhotpatch and
6414 nested functions (e.g. with gfortran).
6416 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
6418 * opts.c (option_name): Remove "enabled by default" rider.
6420 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
6422 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
6424 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
6425 Uros Bizjak <ubizjak@gmail.com>
6428 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
6429 * configure: Regenerated.
6431 2014-02-12 Richard Biener <rguenther@suse.de>
6433 * vec.c (vec_prefix::calculate_allocation): Move as
6434 inline variant to vec.h.
6435 (vec_prefix::calculate_allocation_1): New out-of-line version.
6436 * vec.h (vec_prefix::calculate_allocation_1): Declare.
6437 (vec_prefix::m_has_auto_buf): Rename to ...
6438 (vec_prefix::m_using_auto_storage): ... this.
6439 (vec_prefix::calculate_allocation): Inline the easy cases
6440 and dispatch to calculate_allocation_1 which doesn't need the
6442 (va_heap::reserve): Use gcc_checking_assert.
6443 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
6444 m_using_auto_storage.
6445 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
6447 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
6448 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
6449 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
6451 2014-02-12 Richard Biener <rguenther@suse.de>
6453 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
6454 when we found a dependence.
6456 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
6458 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
6460 (maybe_fold_stmt): ... into this new function.
6461 * omp-low.c (lower_omp): Update comment.
6463 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
6466 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
6469 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
6471 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
6472 identifiers in comments.
6473 (cortexa53_extra_costs): Likewise.
6474 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
6475 (cortexa7_extra_costs): Likewise.
6476 (cortexa12_extra_costs): Likewise.
6477 (cortexa15_extra_costs): Likewise.
6478 (v7m_extra_costs): Likewise.
6480 2014-02-12 Richard Biener <rguenther@suse.de>
6483 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
6484 of posix_memalign being successful.
6485 (lower_stmt): Restrict lowering of posix_memalign to when
6486 -ftree-bit-ccp is enabled.
6488 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6490 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
6492 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
6494 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
6496 PR rtl-optimization/60116
6497 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
6498 other_insn once the combination has been validated.
6500 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
6503 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
6505 * ipa-devirt.c: Include demangle.h
6506 (odr_violation_reported): New static variable.
6507 (add_type_duplicate): Update odr_violations.
6508 (maybe_record_node): Add completep parameter; update it.
6509 (record_target_from_binfo): Add COMPLETEP parameter;
6510 update it as needed.
6511 (possible_polymorphic_call_targets_1): Likewise.
6512 (struct polymorphic_call_target_d): Add nonconstruction_targets;
6513 rename FINAL to COMPLETE.
6514 (record_targets_from_bases): Sanity check we found the binfo;
6515 fix COMPLETEP updating.
6516 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
6517 parameter, fix computing of COMPLETEP.
6518 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
6519 at LTO time do demangling.
6520 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
6521 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
6523 (gimple_get_virt_method_for_binfo): Likewise.
6524 * gimple-fold.h (gimple_get_virt_method_for_binfo,
6525 gimple_get_virt_method_for_vtable): Update prototypes.
6527 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
6530 * genautomata.c (add_presence_absence): Fix typo with
6531 {final_}presence_list.
6533 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
6536 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
6537 for VSX/Altivec vectors that land in GPR registers.
6539 2014-02-11 Richard Henderson <rth@redhat.com>
6540 Jakub Jelinek <jakub@redhat.com>
6543 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
6544 around drhs if type conversion to lacc->type is not useless.
6546 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6548 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
6550 (cortex-a57.cortex-a53): Likewise.
6551 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
6553 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6555 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
6558 2014-02-11 Renlin Li <Renlin.Li@arm.com>
6560 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
6561 add_options_for_arm_vfp3.
6563 2014-02-11 Jeff Law <law@redhat.com>
6566 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
6567 object with an undesirable mode.
6569 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6572 * config/i386/sol2-9.h: New file.
6573 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
6574 *-*-solaris2.9*): Use it.
6576 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6578 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
6579 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
6581 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6583 * config/microblaze/microblaze.c: Extend mcpu version format
6585 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
6587 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
6589 2014-02-10 Richard Henderson <rth@redhat.com>
6592 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
6593 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
6594 ms-abi vs -mno-accumulate-outgoing-args.
6595 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
6596 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
6599 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
6602 * cfgexpand.c (expand_asm_operands): Attach source location to
6603 ASM_INPUT rtx objects.
6604 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
6606 2014-02-10 Nick Clifton <nickc@redhat.com>
6608 * config/mn10300/mn10300.c (popcount): New function.
6609 (mn10300_expand_prologue): Include saved registers in stack usage
6612 2014-02-10 Jeff Law <law@redhat.com>
6615 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
6616 when changing the SET_DEST of a prior insn to avoid an input reload.
6618 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6620 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
6621 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
6622 -mcall-openbsd, or -mcall-linux.
6623 (CC1_ENDIAN_BIG_SPEC): Remove.
6624 (CC1_ENDIAN_LITTLE_SPEC): Remove.
6625 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
6626 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
6627 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
6628 and %cc1_endian_default.
6629 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
6631 2014-02-10 Richard Biener <rguenther@suse.de>
6633 PR tree-optimization/60115
6634 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
6635 MEM_REF handling. Properly verify that the accesses are not
6636 out of the objects bound.
6638 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6640 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
6643 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
6645 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
6646 proper constants and fix formatting.
6647 (possible_polymorphic_call_targets): Fix formatting.
6649 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
6650 Ilya Tocar <ilya.tocar@intel.com>
6652 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
6653 (_mm512_loadu_epi32): Renamed into...
6654 (_mm512_loadu_si512): This.
6655 (_mm512_storeu_epi32): Renamed into...
6656 (_mm512_storeu_si512): This.
6657 (_mm512_maskz_ceil_ps): Removed.
6658 (_mm512_maskz_ceil_pd): Ditto.
6659 (_mm512_maskz_floor_ps): Ditto.
6660 (_mm512_maskz_floor_pd): Ditto.
6661 (_mm512_floor_round_ps): Ditto.
6662 (_mm512_floor_round_pd): Ditto.
6663 (_mm512_ceil_round_ps): Ditto.
6664 (_mm512_ceil_round_pd): Ditto.
6665 (_mm512_mask_floor_round_ps): Ditto.
6666 (_mm512_mask_floor_round_pd): Ditto.
6667 (_mm512_mask_ceil_round_ps): Ditto.
6668 (_mm512_mask_ceil_round_pd): Ditto.
6669 (_mm512_maskz_floor_round_ps): Ditto.
6670 (_mm512_maskz_floor_round_pd): Ditto.
6671 (_mm512_maskz_ceil_round_ps): Ditto.
6672 (_mm512_maskz_ceil_round_pd): Ditto.
6673 (_mm512_expand_pd): Ditto.
6674 (_mm512_expand_ps): Ditto.
6675 * config/i386/i386.c (ix86_builtins): Remove
6676 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
6677 (bdesc_args): Ditto.
6678 * config/i386/predicates.md (const1256_operand): New.
6679 (const_1_to_2_operand): Ditto.
6680 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6681 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6682 (*avx512pf_gatherpf<mode>sf): Ditto.
6683 (avx512pf_gatherpf<mode>df): Ditto.
6684 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6685 (*avx512pf_gatherpf<mode>df): Ditto.
6686 (avx512pf_scatterpf<mode>sf): Ditto.
6687 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6688 (*avx512pf_scatterpf<mode>sf): Ditto.
6689 (avx512pf_scatterpf<mode>df): Ditto.
6690 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6691 (*avx512pf_scatterpf<mode>df): Ditto.
6692 (avx512f_expand<mode>): Removed.
6693 (<shift_insn><mode>3<mask_name>): Change predicate type.
6695 2014-02-08 Jakub Jelinek <jakub@redhat.com>
6697 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
6698 not at the end of datarefs vector use ordered_remove to avoid
6699 reordering datarefs vector.
6702 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
6703 mark local addressable non-static vars as GOVD_PRIVATE
6704 instead of GOVD_LOCAL.
6705 * omp-low.c (lower_omp_for): Move gimple_bind_vars
6706 and BLOCK_VARS of gimple_bind_block to new_stmt rather
6710 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
6711 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
6712 assume_aligned or alloc_align attributes.
6713 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
6714 arguments. Handle also assume_aligned and alloc_align attributes.
6715 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
6716 calls to functions with assume_aligned or alloc_align attributes.
6717 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
6719 2014-02-08 Terry Guo <terry.guo@arm.com>
6721 * doc/invoke.texi: Document ARM -march=armv7e-m.
6723 2014-02-08 Jakub Jelinek <jakub@redhat.com>
6725 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
6726 flag on __cilkrts_rethrow builtin.
6729 * ipa-cp.c (determine_versionability): Fail at -O0
6730 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
6731 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
6734 2014-02-04 Jakub Jelinek <jakub@redhat.com>
6737 * tree-inline.c (copy_forbidden): Fail for
6738 __attribute__((optimize (0))) functions.
6740 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
6742 * varpool.c: Include pointer-set.h.
6743 (varpool_remove_unreferenced_decls): Variables in other partitions
6744 will not be output; be however careful to not lose information
6747 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
6749 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
6750 lookup in the vtable constructor.
6752 2014-02-07 Jeff Law <law@redhat.com>
6755 * config/m68k/m68k.md (ashldi_extsi): Turn into a
6756 define_insn_and_split.
6758 * ipa-inline.c (inline_small_functions): Fix typos.
6760 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6762 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
6763 (s390_can_use_return_insn): Declare.
6764 * config/s390/s390.h (EPILOGUE_USES): Define.
6765 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
6767 (s390_chunkify_start): Handle return JUMP_LABELs.
6768 (s390_early_mach): Emit a main_pool instruction on the entry edge.
6769 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
6770 (s390_can_use_return_insn): New functions.
6771 (s390_fix_long_loop_prediction): Handle conditional returns.
6772 (TARGET_SET_UP_BY_PROLOGUE): Define.
6773 * config/s390/s390.md (ANY_RETURN): New code iterator.
6774 (*creturn, *csimple_return, return, simple_return): New patterns.
6776 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6778 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
6779 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
6780 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
6781 REG_CFA_RESTORE list when deciding not to restore a register.
6783 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6785 * config/s390/s390.c: Include tree-pass.h and context.h.
6786 (s390_early_mach): New function, split out from...
6787 (s390_emit_prologue): ...here.
6788 (pass_data_s390_early_mach): New pass structure.
6789 (pass_s390_early_mach): New class.
6790 (s390_option_override): Create and register early_mach pass.
6791 Move to end of file.
6793 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6795 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
6796 to match for the exit block.
6798 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6800 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
6801 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
6802 Reject misaligned operands.
6804 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6806 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
6808 2014-02-07 Richard Biener <rguenther@suse.de>
6811 * gimple-low.c (lower_builtin_posix_memalign): New function.
6812 (lower_stmt): Call it to lower posix_memalign in a way
6813 to make alignment info accessible.
6815 2014-02-07 Jakub Jelinek <jakub@redhat.com>
6818 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
6819 __builtin_setjmp_receiver.
6821 2014-02-07 Richard Biener <rguenther@suse.de>
6824 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
6825 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
6826 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
6827 Handle BUILT_IN_POSIX_MEMALIGN.
6828 (find_func_clobbers): Likewise.
6829 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
6830 (call_may_clobber_ref_p_1): Likewise.
6832 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6835 * ipa-devirt.c (record_target_from_binfo): Remove overactive
6838 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6841 * lto-cgraph.c (lto_output_node): Use
6842 symtab_get_symbol_partitioning_class.
6843 (lto_output_varpool_node): likewise.
6844 (symtab_get_symbol_partitioning_class): Move here from
6846 * cgraph.h (symbol_partitioning_class): Likewise.
6847 (symtab_get_symbol_partitioning_class): Declare.
6849 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6851 * ggc.h (ggc_internal_cleared_alloc): New macro.
6852 * vec.h (vec_safe_copy): Handle memory stats.
6853 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
6854 * target-globals.c (save_target_globals): Likewise.
6856 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6859 * expr.c (emit_move_resolve_push): Export; be bit more selective
6860 on when to clear alias set.
6861 * expr.h (emit_move_resolve_push): Declare.
6862 * function.h (struct function): Add tail_call_marked.
6863 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
6864 * config/i386/i386-protos.h (ix86_expand_push): Remove.
6865 * config/i386/i386.md (TImode move expander): De not call
6867 (FP push expanders): Preserve memory attributes.
6868 * config/i386/sse.md (push<mode>1): Remove.
6869 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
6870 (ix86_expand_push): Remove.
6871 * config/i386/mmx.md (push<mode>1): Remove.
6873 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6875 PR rtl-optimization/60030
6876 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
6877 lopart with paradoxical subreg before shifting it up by hprec.
6879 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6881 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
6882 Remove extra newline at end of file.
6883 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
6884 (arm_issue_rate): Handle cortexa57.
6885 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
6886 (cortex-a57.cortex-a53): Likewise.
6888 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6891 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
6892 don't record in REG_FRAME_RELATED_EXPR registers not set in that
6894 (arm_expand_prologue): Adjust all callers.
6895 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
6896 info, registers also at the lowest numbered registers side. Use
6897 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
6901 * var-tracking.c (adjust_mems): Before adding a SET to
6902 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
6904 2014-02-06 Alan Modra <amodra@gmail.com>
6907 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
6908 change SDmode to DDmode when lra_in_progress.
6910 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6913 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
6914 free_data_ref on the dr first, and before goto again also set dr
6915 to the next dr. For simd_lane_access, free old datarefs[i] before
6916 overwriting it. For get_vectype_for_scalar_type failure, don't
6917 free_data_ref if simd_lane_access.
6919 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
6922 * tree.h (opts_for_fn): New inline function.
6923 (opt_for_fn): Define.
6924 * config/i386/i386.c (ix86_function_regparm): Use
6925 opt_for_fn (decl, optimize) instead of optimize.
6927 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
6929 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
6930 for SYMBOL_REF in large memory model.
6932 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6934 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
6936 (cortex-a57): Likewise.
6937 (cortex-a57.cortex-a53): Likewise.
6939 2014-02-06 Yury Gribov <y.gribov@samsung.com>
6940 Kugan Vivekanandarajah <kuganv@linaro.org>
6942 * config/arm/arm.c (arm_vector_alignment_reachable): Check
6944 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
6946 2014-02-06 Richard Biener <rguenther@suse.de>
6948 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
6949 set_loop_copy and initialize_original_copy_tables.
6951 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
6953 * config/aarch64/aarch64-simd.md
6954 (aarch64_ashr_simddi): Change QI to SI.
6956 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
6957 Jakub Jelinek <jakub@redhat.com>
6960 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
6963 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6965 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
6966 CODE_FOR_altivec_vpku[hw]um to
6967 CODE_FOR_altivec_vpku[hw]um_direct.
6968 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
6969 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
6970 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
6971 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
6973 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6975 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
6976 generation for -maltivec=be.
6977 (altivec_vsumsws): Simplify redundant test.
6979 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6981 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
6982 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
6983 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
6984 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
6985 gen_altivec_vpkuwum.
6986 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
6988 (altivec_vpks<VI_char>ss): Likewise.
6989 (altivec_vpks<VI_char>us): Likewise.
6990 (altivec_vpku<VI_char>us): Likewise.
6991 (altivec_vpku<VI_char>um): Likewise.
6992 (altivec_vpku<VI_char>um_direct): New (copy of
6993 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
6995 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
6996 target is little endian and -maltivec=be is not specified.
6997 (*altivec_vupkhs<VU_char>_direct): New (copy of
6998 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
6999 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
7000 target is little endian and -maltivec=be is not specified.
7001 (*altivec_vupkls<VU_char>_direct): New (copy of
7002 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
7003 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
7004 little endian and -maltivec=be is not specified.
7005 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
7006 little endian and -maltivec=be is not specified.
7008 2014-02-05 Richard Henderson <rth@redhat.com>
7011 * combine-stack-adj.c: Revert r206943.
7012 * sched-int.h (struct deps_desc): Add last_args_size.
7013 * sched-deps.c (init_deps): Initialize it.
7014 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7015 contain REG_ARGS_SIZE notes.
7017 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7019 * lto-cgraph.c (asm_nodes_output): Make global.
7020 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7021 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7022 (driver_handle_option): Handle OPT_fwpa.
7024 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7027 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7028 a comment typo and formatting issue. If odr_hash hasn't been
7029 created, return vNULL and set *completep to false.
7032 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7033 bb with no successors.
7035 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7038 * doc/invoke.texi (-march): Clarify documentation for ARM.
7042 2014-02-05 Richard Biener <rguenther@suse.de>
7044 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
7045 when not vectorizing because of too many alias checks.
7046 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
7047 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
7049 2014-02-05 Nick Clifton <nickc@redhat.com>
7051 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
7052 accept extended registers in any mode when compiling for the MN10300.
7054 2014-02-05 Yury Gribov <y.gribov@samsung.com>
7056 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
7057 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
7058 sanitization attributes.
7059 (can_inline_edge_p): Likewise.
7060 (sanitize_attrs_match_for_inline_p): New function.
7062 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7064 * ipa-prop.c (detect_type_change): Shor circuit testing of
7065 type changes on THIS pointer.
7067 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
7070 * config/pa/pa.c (legitimize_tls_address): Return original address
7071 if not passed a SYMBOL_REF rtx.
7072 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
7074 (pa_emit_move_sequence): Simplify TLS source operands.
7075 (pa_legitimate_constant_p): Reject all TLS constants.
7076 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
7077 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
7079 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7081 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
7082 groups when we know they are controlled by LTO.
7083 * varasm.c (default_binds_local_p_1): If object is in other partition,
7084 it will be resolved locally.
7086 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7088 * config/host-linux.c (linux_gt_pch_use_address): Don't
7089 use SSIZE_MAX because it is not always defined.
7091 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
7094 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
7095 threshold for pseudo splitting.
7096 (update_ebb_live_info): Process call argument hard registers and
7097 hard registers from insn definition too.
7098 (max_small_class_regs_num): New constant.
7099 (inherit_in_ebb): Update live hard regs through EBBs. Update
7100 reloads_num only for small register classes. Don't split for
7103 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
7106 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
7109 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7111 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
7114 2014-02-04 Marek Polacek <polacek@redhat.com>
7116 * gdbinit.in (pel): Define.
7118 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7120 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
7123 2014-02-04 Richard Biener <rguenther@suse.de>
7126 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
7127 in function context local.
7128 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
7129 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
7130 similar to LTO_imported_decl_ref.
7132 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7134 PR tree-optimization/60002
7135 * cgraphclones.c (build_function_decl_skip_args): Clear
7138 PR tree-optimization/60023
7139 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
7140 false to gsi_replace.
7141 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
7142 has been in some EH region and vec_stmt could throw, add
7143 vec_stmt into the same EH region.
7144 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
7145 has no lhs, ignore it.
7146 * internal-fn.c (expand_MASK_LOAD): Likewise.
7149 * tree-inline.c (copy_forbidden): Fail for
7150 __attribute__((optimize (0))) functions.
7153 * omp-low.c (simd_clone_struct_copy): If from->inbranch
7154 is set, copy one less argument.
7155 (expand_simd_clones): Don't subtract clone_info->inbranch
7156 from simd_clone_struct_alloc argument.
7158 PR rtl-optimization/57915
7159 * recog.c (simplify_while_replacing): If all unary/binary/relational
7160 operation arguments are constant, attempt to simplify those.
7163 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
7164 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
7166 2014-02-04 Richard Biener <rguenther@suse.de>
7168 PR tree-optimization/60012
7169 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
7170 TBAA disambiguation to all DDRs.
7172 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7175 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
7176 (LINK_SPEC): Use it for -shared, -shared-libgcc.
7178 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7181 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
7183 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7185 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
7186 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
7188 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7191 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
7192 to figure out targets of polymorphic calls with known decl.
7193 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7194 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
7195 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
7196 (get_polymorphic_call_info): ... here.
7197 (get_polymorphic_call_info_from_invariant): New function.
7199 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7201 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
7202 lookup via vtable pointer; check for type consistency
7203 and turn inconsitent facts into UNREACHABLE.
7204 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7205 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
7206 type inconsistent querries; return UNREACHABLE instead.
7208 2014-02-03 Richard Henderson <rth@twiddle.net>
7211 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
7212 already processed this node.
7213 (normalize_one_pred_1): Pass along mark_set.
7214 (normalize_one_pred): Create and destroy a pointer_set_t.
7215 (normalize_one_pred_chain): Likewise.
7217 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
7219 PR gcov-profile/58602
7220 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
7222 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7225 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
7226 -fno-devirtualize; try to devirtualize by the knowledge of
7227 virtual table pointer given by aggregate propagation.
7228 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7229 (ipa_print_node_jump_functions): Dump also offset that
7230 is relevant for polymorphic calls.
7231 (determine_known_aggregate_parts): Add arg_type parameter; use it
7232 instead of determining the type from pointer type.
7233 (ipa_compute_jump_functions_for_edge): Update call of
7234 determine_known_aggregate_parts.
7235 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
7236 (gimple_get_virt_method_for_binfo): ... here; simplify using
7237 vtable_pointer_value_to_vtable.
7238 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
7239 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
7240 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
7241 (vtable_pointer_value_to_vtable): Break out from ...; handle also
7243 (vtable_pointer_value_to_binfo): ... here.
7244 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
7246 2014-02-03 Teresa Johnson <tejohnson@google.com>
7248 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
7249 redef of outer loop index variable.
7251 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
7255 * doc/extend.texi (Function Attributes): Typo.
7257 2014-02-03 Cong Hou <congh@google.com>
7259 PR tree-optimization/60000
7260 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
7261 if the vectorized statement is a store. A store statement can only
7262 appear at the end of pattern statements.
7264 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7266 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
7267 (ix86_option_override_internal): Default long double to 64-bit for
7268 32-bit Bionic and to 128-bit for 64-bit Bionic.
7270 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
7271 TARGET_LONG_DOUBLE_128 is true.
7272 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
7274 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
7275 (mlong-double-64): Negate -mlong-double-128.
7276 (mlong-double-128): New option.
7278 * config/i386/i386-c.c (ix86_target_macros): Define
7279 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
7281 * doc/invoke.texi: Document -mlong-double-128.
7283 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7285 PR rtl-optimization/60024
7286 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
7288 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
7290 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
7292 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7294 PR rtl-optimization/57662
7295 * sel-sched.c (code_motion_path_driver): Do not mark already not
7296 existing blocks in the visiting bitmap.
7298 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7300 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
7301 on the insn being emitted.
7303 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
7304 Will Deacon <will.deacon@arm.com>
7306 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
7308 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7310 * config/arm/arm-tables.opt: Regenerate.
7312 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7314 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
7315 for vector types other than V16QImode.
7316 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
7317 define_expand, and call altivec_expand_vec_perm_le when producing
7318 code with little endian element order.
7319 (*altivec_vperm_<mode>_internal): New insn having previous
7320 behavior of altivec_vperm_<mode>.
7321 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
7322 altivec_expand_vec_perm_le when producing code with little endian
7324 (*altivec_vperm_<mode>_uns_internal): New insn having previous
7325 behavior of altivec_vperm_<mode>_uns.
7327 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7329 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
7330 (altivec_vsumsws): Add handling for -maltivec=be with a little
7332 (altivec_vsumsws_direct): New.
7333 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
7334 gen_altivec_vsumsws.
7336 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7338 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
7339 vtable_pointer_value_to_binfo): New functions.
7340 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
7341 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
7343 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
7345 * config/nios2/nios2.md (load_got_register): Initialize GOT
7346 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
7347 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
7349 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7351 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
7352 preserverd by passthrough, do not propagate the type.
7354 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7356 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
7357 (mips_atomic_assign_expand_fenv): New function.
7358 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
7360 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7362 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
7363 (__builtin_mips_set_fcsr): Likewise.
7364 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
7365 MIPS_USI_FTYPE_VOID.
7366 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
7367 (mips16_expand_set_fcsr): Likewise.
7368 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
7369 (mips16_set_fcsr_stub): Likewise.
7370 (mips16_get_fcsr_one_only_stub): New class.
7371 (mips16_set_fcsr_one_only_stub): Likewise.
7372 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
7373 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
7374 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
7375 (hard_float): New availability predicate.
7376 (mips_builtins): Add get_fcsr and set_fcsr.
7377 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
7378 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
7379 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
7380 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
7381 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
7384 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7386 * config/mips/mips.c (mips_one_only_stub): New class.
7387 (mips_need_mips16_rdhwr_p): Replace with...
7388 (mips16_rdhwr_stub): ...this new variable.
7389 (mips16_stub_call_address): New function.
7390 (mips16_rdhwr_one_only_stub): New class.
7391 (mips_expand_thread_pointer): Use mips16_stub_call_address.
7392 (mips_output_mips16_rdhwr): Delete.
7393 (mips_finish_stub): New function.
7394 (mips_code_end): Use it to handle rdhwr stubs.
7396 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
7399 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
7400 when calculating size of integer atomic types.
7402 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
7404 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
7406 2014-02-01 Jakub Jelinek <jakub@redhat.com>
7408 PR tree-optimization/60003
7409 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
7410 * profile.c (branch_prob): Use gimple_call_builtin_p
7411 to check for BUILT_IN_SETJMP_RECEIVER.
7412 * tree-inline.c (copy_bb): Call notice_special_calls.
7414 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
7417 * lra-constraints.c (process_alt_operands): Update reload_sum only
7420 2014-01-31 Richard Henderson <rth@redhat.com>
7423 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
7424 until after else_eh is processed.
7426 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7428 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
7429 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
7430 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
7431 in smmintrin.h, remove them.
7432 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
7433 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
7434 * config/i386/i386.md (ROUND_SAE): Fix value.
7435 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
7436 (const48_operand): New.
7437 * config/i386/subst.md (round), (round_expand): Use
7438 const_4_or_8_to_11_operand.
7439 (round_saeonly), (round_saeonly_expand): Use const48_operand.
7441 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7443 * config/i386/constraints.md (Yk): Swap meaning with k.
7444 * config/i386/i386.md (movhi_internal): Change Yk to k.
7445 (movqi_internal): Ditto.
7446 (*k<logic><mode>): Ditto.
7449 (kandn<mode>): Ditto.
7450 (*<code>hi_1): Ditto.
7451 (*<code>qi_1): Ditto.
7452 (kxnor<mode>): Ditto.
7453 (kortestzhi): Ditto.
7454 (kortestchi): Ditto.
7456 (*one_cmplhi2_1): Ditto.
7457 (*one_cmplqi2_1): Ditto.
7458 * config/i386/sse.md (): Change k to Yk.
7459 (avx512f_load<mode>_mask): Ditto.
7460 (avx512f_blendm<mode>): Ditto.
7461 (avx512f_store<mode>_mask): Ditto.
7462 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
7463 (avx512f_storedqu<mode>_mask): Ditto.
7464 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
7466 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
7467 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
7468 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
7469 (avx512f_maskcmp<mode>3): Ditto.
7470 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
7471 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
7472 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
7473 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
7474 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
7475 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
7476 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
7477 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
7478 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
7479 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
7480 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
7481 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
7482 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
7483 (vec_extract_lo_<mode>_maskm): Ditto.
7484 (vec_extract_hi_<mode>_maskm): Ditto.
7485 (avx512f_vternlog<mode>_mask): Ditto.
7486 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
7487 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
7488 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
7489 (avx512f_<code>v8div16qi2_mask): Ditto.
7490 (avx512f_<code>v8div16qi2_mask_store): Ditto.
7491 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
7492 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
7493 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
7494 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
7495 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7496 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7497 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7498 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7499 (avx512cd_maskb_vec_dupv8di): Ditto.
7500 (avx512cd_maskw_vec_dupv16si): Ditto.
7501 (avx512f_vpermi2var<mode>3_maskz): Ditto.
7502 (avx512f_vpermi2var<mode>3_mask): Ditto.
7503 (avx512f_vpermi2var<mode>3_mask): Ditto.
7504 (avx512f_vpermt2var<mode>3_maskz): Ditto.
7505 (*avx512f_gathersi<mode>): Ditto.
7506 (*avx512f_gathersi<mode>_2): Ditto.
7507 (*avx512f_gatherdi<mode>): Ditto.
7508 (*avx512f_gatherdi<mode>_2): Ditto.
7509 (*avx512f_scattersi<mode>): Ditto.
7510 (*avx512f_scatterdi<mode>): Ditto.
7511 (avx512f_compress<mode>_mask): Ditto.
7512 (avx512f_compressstore<mode>_mask): Ditto.
7513 (avx512f_expand<mode>_mask): Ditto.
7514 * config/i386/subst.md (mask): Change k to Yk.
7515 (mask_scalar_merge): Ditto.
7518 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
7520 * doc/extend.texi (Vector Extensions): Document ?: in C++.
7522 2014-01-31 Richard Biener <rguenther@suse.de>
7525 * builtins.c (fold_builtin_memory_op): Make sure to not
7526 use a floating-point mode or a boolean or enumeral type for
7529 2014-01-30 DJ Delorie <dj@redhat.com>
7531 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
7532 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
7533 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
7534 whenever main() has an epilogue.
7536 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7538 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
7539 unused variable "field".
7540 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
7541 (vsx_mergeh_<mode>): Likewise.
7542 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
7543 (altivec_vmrghh): Likewise.
7544 (altivec_vmrghw): Likewise.
7545 (altivec_vmrglb): Likewise.
7546 (altivec_vmrglh): Likewise.
7547 (altivec_vmrglw): Likewise.
7548 (altivec_vspltb): Add missing uses.
7549 (altivec_vsplth): Likewise.
7550 (altivec_vspltw): Likewise.
7551 (altivec_vspltsf): Likewise.
7553 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7556 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
7557 frame related instructions.
7559 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
7561 PR rtl-optimization/59959
7562 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
7563 any reload of register whose subreg is invalid.
7565 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7567 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
7568 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
7569 Add missing return type - void.
7571 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7573 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
7574 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
7575 remove element index adjustment for endian (now handled in vsx.md
7577 (altivec_expand_vec_perm_const): Use
7578 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
7579 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
7580 (vsx_xxspltw_<mode>): Adjust element index for little endian.
7581 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
7582 define_expand and a new define_insn *altivec_vspltb_internal;
7583 adjust for -maltivec=be on a little endian target.
7584 (altivec_vspltb_direct): New.
7585 (altivec_vsplth): Divide into a define_expand and a new
7586 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
7587 little endian target.
7588 (altivec_vsplth_direct): New.
7589 (altivec_vspltw): Divide into a define_expand and a new
7590 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
7591 little endian target.
7592 (altivec_vspltw_direct): New.
7593 (altivec_vspltsf): Divide into a define_expand and a new
7594 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
7595 a little endian target.
7597 2014-01-30 Richard Biener <rguenther@suse.de>
7599 PR tree-optimization/59993
7600 * tree-ssa-forwprop.c (associate_pointerplus): Check we
7601 can propagate form the earlier stmt and avoid the transform
7602 when the intermediate result is needed.
7604 2014-01-30 Alangi Derick <alangiderick@gmail.com>
7606 * README.Portability: Fix typo.
7608 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
7610 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
7611 comparison_operator with ordered_comparison_operator.
7613 2014-01-30 Nick Clifton <nickc@redhat.com>
7615 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
7616 Rename to mn10300_store_multiple_regs.
7617 * config/mn10300/mn10300.c: Likewise.
7618 * config/mn10300/mn10300.md (store_movm): Fix typo: call
7619 store_multiple_regs.
7620 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
7621 Call mn10300_store_multiple_regs.
7623 2014-01-30 Nick Clifton <nickc@redhat.com>
7624 DJ Delorie <dj@redhat.com>
7626 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
7627 %fp 2 to keep registers after it properly word-aligned.
7628 (rl78_alloc_physical_registers_umul): Handle the case where both
7629 input operands are the same.
7631 2014-01-30 Richard Biener <rguenther@suse.de>
7633 PR tree-optimization/59903
7634 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
7637 2014-01-30 Jason Merrill <jason@redhat.com>
7640 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
7643 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
7645 2014-01-30 Richard Biener <rguenther@suse.de>
7647 PR tree-optimization/59951
7648 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
7650 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
7653 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
7654 SFmode to DFmode case.
7656 2014-01-29 DJ Delorie <dj@redhat.com>
7658 * config/msp430/msp430.opt (-minrt): New.
7659 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
7661 (ENDFILE_SPEC): Likewise.
7663 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
7665 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
7666 (estimate_function_body_sizes): Use it.
7668 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
7671 * dwarf2out.c (is_cxx_auto): New.
7672 (is_base_type): Use it.
7673 (gen_type_die_with_usage): Likewise.
7675 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7677 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
7678 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
7679 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
7680 -maltivec=be with LE targets.
7681 (vsx_mergeh_<mode>): Likewise.
7682 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
7683 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
7684 (altivec_vmrghb): Replace with define_expand and new
7685 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
7686 (altivec_vmrghb_direct): New define_insn.
7687 (altivec_vmrghh): Replace with define_expand and new
7688 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
7689 (altivec_vmrghh_direct): New define_insn.
7690 (altivec_vmrghw): Replace with define_expand and new
7691 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
7692 (altivec_vmrghw_direct): New define_insn.
7693 (*altivec_vmrghsf): Adjust for endianness.
7694 (altivec_vmrglb): Replace with define_expand and new
7695 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
7696 (altivec_vmrglb_direct): New define_insn.
7697 (altivec_vmrglh): Replace with define_expand and new
7698 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
7699 (altivec_vmrglh_direct): New define_insn.
7700 (altivec_vmrglw): Replace with define_expand and new
7701 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
7702 (altivec_vmrglw_direct): New define_insn.
7703 (*altivec_vmrglsf): Adjust for endianness.
7704 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
7705 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
7706 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
7707 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
7708 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
7709 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
7710 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
7711 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
7713 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
7715 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
7716 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
7719 2014-01-29 Richard Biener <rguenther@suse.de>
7721 PR tree-optimization/58742
7722 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
7723 associate_pointerplus_align.
7724 (associate_pointerplus_diff): New function.
7725 (associate_pointerplus): Likewise. Call associate_pointerplus_align
7726 and associate_pointerplus_diff.
7728 2014-01-29 Richard Biener <rguenther@suse.de>
7730 * lto-streamer.h (LTO_major_version): Bump to 3.
7731 (LTO_minor_version): Reset to 0.
7733 2014-01-29 Renlin Li <Renlin.Li@arm.com>
7735 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
7736 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
7737 (arm_file_start): Generate correct asm header for armv7ve.
7738 * config/arm/bpabi.h: Add multilib support for armv7ve.
7739 * config/arm/driver-arm.c: Change the architectures of cortex-a7
7740 and cortex-a15 to armv7ve.
7741 * config/arm/t-aprofile: Add multilib support for armv7ve.
7742 * doc/invoke.texi: Document -march=armv7ve.
7744 2014-01-29 Richard Biener <rguenther@suse.de>
7746 PR tree-optimization/58742
7747 * tree-ssa-forwprop.c (associate_plusminus): Return true
7748 if we changed sth, defer EH cleanup to ...
7749 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
7750 (simplify_mult): New function.
7752 2014-01-29 Jakub Jelinek <jakub@redhat.com>
7755 PR tree-optimization/59920
7756 * tree.c (build_common_builtin_nodes): Remove
7757 __builtin_setjmp_dispatcher initialization.
7758 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
7759 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
7760 instead of gsi_after_labels + manually skipping debug stmts.
7761 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
7762 ignore bbs with IFN_ABNORMAL_DISPATCHER.
7763 * tree-inline.c (copy_edges_for_bb): Remove
7764 can_make_abnormal_goto argument, instead add abnormal_goto_dest
7765 argument. Ignore computed_goto_p stmts. Don't call
7766 make_abnormal_goto_edges. If a call might need abnormal edges
7767 for non-local gotos, see if it already has an edge to
7768 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
7769 with true argument, don't do anything then, otherwise add
7770 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
7771 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
7773 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
7774 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
7775 (lower_stmt): Don't set data->calls_builtin_setjmp.
7776 (lower_builtin_setjmp): Adjust comment.
7777 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
7778 * tree-cfg.c (found_computed_goto): Remove.
7779 (factor_computed_gotos): Remove.
7780 (make_goto_expr_edges): Return bool, true for computed gotos.
7781 Don't call make_abnormal_goto_edges.
7782 (build_gimple_cfg): Don't set found_computed_goto, don't call
7783 factor_computed_gotos.
7784 (computed_goto_p): No longer static.
7785 (make_blocks): Don't set found_computed_goto.
7786 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
7787 (make_edges): If make_goto_expr_edges returns true, push bb
7788 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
7789 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
7790 vector. Record mapping between bbs and OpenMP regions if there
7791 are any, adjust make_gimple_omp_edges caller. Call
7792 handle_abnormal_edges.
7793 (make_abnormal_goto_edges): Remove.
7794 * tree-cfg.h (make_abnormal_goto_edges): Remove.
7795 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
7796 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
7797 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
7798 * internal-fn.def (ABNORMAL_DISPATCHER): New.
7799 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
7800 filling *region also set *region_idx to (*region)->entry->index.
7803 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
7804 For REGs set ORIGINAL_REGNO.
7806 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
7808 * doc/md.texi: Mention that a target shouldn't implement
7809 vec_widen_(s|u)mul_even/odd pair if it is less efficient
7812 2014-01-29 Jakub Jelinek <jakub@redhat.com>
7814 PR tree-optimization/59594
7815 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
7816 a copy of the datarefs vector rather than the vector itself.
7818 2014-01-28 Jason Merrill <jason@redhat.com>
7821 * dwarf2out.c (auto_die): New static.
7822 (gen_type_die_with_usage): Handle C++1y 'auto'.
7823 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
7826 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
7829 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
7830 (SPEC_X32): Likewise.
7831 (SPEC_64): Likewise.
7832 * config/i386/i386.c (ix86_option_override_internal): Turn off
7833 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
7835 (x86_file_start): Output .code16gcc for TARGET_16BIT.
7836 * config/i386/i386.h (TARGET_16BIT): New macro.
7837 (TARGET_16BIT_P): Likewise.
7838 * config/i386/i386.opt: Add m16.
7839 * doc/invoke.texi: Document -m16.
7841 2014-01-28 Jakub Jelinek <jakub@redhat.com>
7843 PR preprocessor/59935
7844 * input.c (location_get_source_line): Bail out on when line number
7845 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
7847 2014-01-28 Richard Biener <rguenther@suse.de>
7849 PR tree-optimization/58742
7850 * tree-ssa-forwprop.c (associate_plusminus): Handle
7851 pointer subtraction of the form (T)(P + A) - (T)P.
7853 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7855 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
7858 2014-01-28 Richard Biener <rguenther@suse.de>
7861 2014-01-28 Richard Biener <rguenther@suse.de>
7863 PR rtl-optimization/45364
7864 PR rtl-optimization/59890
7865 * var-tracking.c (local_get_addr_clear_given_value): Handle
7866 already cleared slot.
7867 (val_reset): Handle not allocated local_get_addr_cache.
7868 (vt_find_locations): Use post-order on the inverted CFG.
7870 2014-01-28 Richard Biener <rguenther@suse.de>
7872 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
7874 2014-01-28 Richard Biener <rguenther@suse.de>
7876 PR rtl-optimization/45364
7877 PR rtl-optimization/59890
7878 * var-tracking.c (local_get_addr_clear_given_value): Handle
7879 already cleared slot.
7880 (val_reset): Handle not allocated local_get_addr_cache.
7881 (vt_find_locations): Use post-order on the inverted CFG.
7883 2014-01-28 Alan Modra <amodra@gmail.com>
7885 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
7886 * configure.ac <recursive call for build != host>: Define
7887 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
7888 and LD_FOR_BUILD too.
7889 * configure: Regenerate.
7891 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
7893 * config/i386/i386.c (get_builtin_code_for_version): Separate
7894 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
7895 Broadwell from Haswell.
7897 2014-01-27 Steve Ellcey <sellcey@mips.com>
7899 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
7900 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
7901 * config/mips/mips.c (mips_option_override): Change setting
7903 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
7904 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
7905 Change from Mask to Var.
7907 2014-01-27 Jeff Law <law@redhat.com>
7909 * ipa-inline.c (inline_small_functions): Fix typo.
7911 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
7913 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
7914 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
7915 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
7916 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
7917 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
7918 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
7919 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
7920 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
7921 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
7922 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
7923 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
7924 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
7925 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
7926 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
7927 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
7928 (_mm512_storeu_epi64): Ditto.
7929 (_mm512_cmpge_epi32_mask): Ditto.
7930 (_mm512_cmpge_epu32_mask): Ditto.
7931 (_mm512_cmpge_epi64_mask): Ditto.
7932 (_mm512_cmpge_epu64_mask): Ditto.
7933 (_mm512_cmple_epi32_mask): Ditto.
7934 (_mm512_cmple_epu32_mask): Ditto.
7935 (_mm512_cmple_epi64_mask): Ditto.
7936 (_mm512_cmple_epu64_mask): Ditto.
7937 (_mm512_cmplt_epi32_mask): Ditto.
7938 (_mm512_cmplt_epu32_mask): Ditto.
7939 (_mm512_cmplt_epi64_mask): Ditto.
7940 (_mm512_cmplt_epu64_mask): Ditto.
7941 (_mm512_cmpneq_epi32_mask): Ditto.
7942 (_mm512_cmpneq_epu32_mask): Ditto.
7943 (_mm512_cmpneq_epi64_mask): Ditto.
7944 (_mm512_cmpneq_epu64_mask): Ditto.
7945 (_mm512_expand_pd): Ditto.
7946 (_mm512_expand_ps): Ditto.
7947 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
7948 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
7949 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
7950 * config/i386/i386.c (ix86_builtins): Add
7951 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
7952 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
7953 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
7954 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
7955 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
7956 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
7957 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
7958 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
7959 IX86_BUILTIN_PMOVUSQW512_MEM.
7960 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
7961 __builtin_ia32_pmovsqd512mem_mask,
7962 __builtin_ia32_pmovqd512mem_mask,
7963 __builtin_ia32_pmovusqw512mem_mask,
7964 __builtin_ia32_pmovsqw512mem_mask,
7965 __builtin_ia32_pmovqw512mem_mask,
7966 __builtin_ia32_pmovusdw512mem_mask,
7967 __builtin_ia32_pmovsdw512mem_mask,
7968 __builtin_ia32_pmovdw512mem_mask,
7969 __builtin_ia32_pmovqb512mem_mask,
7970 __builtin_ia32_pmovusqb512mem_mask,
7971 __builtin_ia32_pmovsqb512mem_mask,
7972 __builtin_ia32_pmovusdb512mem_mask,
7973 __builtin_ia32_pmovsdb512mem_mask,
7974 __builtin_ia32_pmovdb512mem_mask.
7975 (bdesc_args): Add __builtin_ia32_expanddf512,
7976 __builtin_ia32_expandsf512.
7977 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
7978 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
7979 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
7980 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
7981 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
7982 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
7983 (avx512f_<code>v8div16qi2_mask_store): This.
7984 (avx512f_expand<mode>): New.
7986 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
7988 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
7990 (_mm512_mask_prefetch_i64gather_pd): Ditto.
7991 (_mm512_prefetch_i32scatter_pd): Ditto.
7992 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
7993 (_mm512_prefetch_i64scatter_pd): Ditto.
7994 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
7995 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
7996 (_mm512_mask_prefetch_i64gather_ps): Ditto.
7997 (_mm512_prefetch_i32scatter_ps): Ditto.
7998 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
7999 (_mm512_prefetch_i64scatter_ps): Ditto.
8000 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
8001 * config/i386/i386-builtin-types.def: Define
8002 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
8003 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
8004 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
8005 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
8006 IX86_BUILTIN_SCATTERPFQPD.
8007 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
8008 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
8009 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
8010 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
8011 __builtin_ia32_scatterpfqps.
8012 (ix86_expand_builtin): Expand new built-ins.
8013 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8014 fix memory access data type.
8015 (*avx512pf_gatherpf<mode>_mask): Ditto.
8016 (*avx512pf_gatherpf<mode>): Ditto.
8017 (avx512pf_scatterpf<mode>): Ditto.
8018 (*avx512pf_scatterpf<mode>_mask): Ditto.
8019 (*avx512pf_scatterpf<mode>): Ditto.
8020 (GATHER_SCATTER_SF_MEM_MODE): New.
8021 (avx512pf_gatherpf<mode>df): Ditto.
8022 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8023 (*avx512pf_scatterpf<mode>df): Ditto.
8025 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8028 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8029 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8032 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8034 * common/config/arm/arm-common.c
8035 (arm_rewrite_mcpu): Handle multiple names.
8037 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8039 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8041 * gimple-builder.h (create_gimple_tmp): Delete.
8043 2014-01-27 Christian Bruel <christian.bruel@st.com>
8045 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
8048 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
8050 * config/pa/pa.md (call): Generate indirect long calls to non-local
8051 functions when outputing 32-bit code.
8052 (call_value): Likewise except for special call to buggy powf function.
8054 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
8055 portable runtime and PIC indirect calls.
8056 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
8057 and PIC call sequences. Use ldo instead of blr to set return register
8058 in PIC call sequence.
8060 2014-01-25 Walter Lee <walt@tilera.com>
8062 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
8063 avoid clobbering a live register.
8065 2014-01-25 Walter Lee <walt@tilera.com>
8067 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
8068 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
8069 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
8070 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
8072 2014-01-25 Walter Lee <walt@tilera.com>
8074 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
8075 arguments on even registers.
8076 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
8078 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
8079 (BIGGEST_ALIGNMENT): Ditto.
8080 (BIGGEST_FIELD_ALIGNMENT): Ditto.
8082 2014-01-25 Walter Lee <walt@tilera.com>
8084 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
8085 insns before bundling.
8086 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8088 2014-01-25 Walter Lee <walt@tilera.com>
8090 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
8091 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
8092 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
8094 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8096 * config/mips/constraints.md (kl): Delete.
8097 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
8098 define expands, using...
8099 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
8100 instructions for MIPS16.
8101 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
8102 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
8104 2014-01-25 Walter Lee <walt@tilera.com>
8106 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
8110 2014-01-25 Walter Lee <walt@tilera.com>
8112 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
8113 (TARGET_EXPAND_TO_RTL_HOOK): Define.
8115 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8117 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
8120 2014-01-25 Jakub Jelinek <jakub@redhat.com>
8122 * print-rtl.c (in_call_function_usage): New var.
8123 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
8124 EXPR_LIST mode as mode and not as reg note name.
8127 * cfgloopmanip.c (copy_loop_info): If
8128 loop->warned_aggressive_loop_optimizations, make sure
8129 the flag is set in target loop too.
8131 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
8133 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
8135 * builtins.def: Likewise.
8136 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
8137 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
8138 * ira.c (ira_setup_eliminable_regset): Likewise.
8139 * omp-low.c (gate_expand_omp): Likewise.
8140 (execute_lower_omp): Likewise.
8141 (diagnose_sb_0): Likewise.
8142 (gate_diagnose_omp_blocks): Likewise.
8143 (simd_clone_clauses_extract): Likewise.
8146 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8148 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
8149 correction for little endian...
8150 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
8153 2014-01-24 Jeff Law <law@redhat.com>
8155 PR tree-optimization/59919
8156 * tree-vrp.c (find_assert_locations_1): Do not register asserts
8157 for non-returning calls.
8159 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
8161 * common/config/aarch64/aarch64-common.c
8162 (aarch64_rewrite_mcpu): Handle multiple names.
8163 * config/aarch64/aarch64.h
8164 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8166 2014-01-24 Dodji Seketeli <dodji@redhat.com>
8168 * input.c (add_file_to_cache_tab): Handle the case where fopen
8171 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
8174 * config/i386/i386.md (pushsf splitter): Get stack adjustment
8175 from push operand if code of push isn't PRE_DEC.
8177 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8180 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
8181 -mquad-memory-atomic. Update -mquad-memory documentation to say
8182 it is only used for non-atomic loads/stores.
8184 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
8185 -mquad-memory or -mquad-memory-atomic switches.
8187 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
8188 -mquad-memory-atomic to ISA 2.07 support.
8190 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
8191 to separate support of normal quad word memory operations (ldq, stq)
8192 from the atomic quad word memory operations.
8194 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
8195 support to separate non-atomic quad word operations from atomic
8196 quad word operations. Disable non-atomic quad word operations in
8197 little endian mode so that we don't have to swap words after the
8198 load and before the store.
8199 (quad_load_store_p): Add comment about atomic quad word support.
8200 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
8201 options printed with -mdebug=reg.
8203 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
8204 -mquad-memory-atomic as the test for whether we have quad word
8205 atomic instructions.
8206 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
8207 or -mp8-vector are used, allow byte/half-word atomic operations.
8209 * config/rs6000/sync.md (load_lockedti): Insure that the address
8210 is a proper indexed or indirect address for the lqarx instruction.
8211 On little endian systems, swap the hi/lo registers after the lqarx
8213 (load_lockedpti): Use indexed_or_indirect_operand predicate to
8214 insure the address is valid for the lqarx instruction.
8215 (store_conditionalti): Insure that the address is a proper indexed
8216 or indirect address for the stqcrx. instruction. On little endian
8217 systems, swap the hi/lo registers before doing the stqcrx.
8219 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
8220 insure the address is valid for the stqcrx. instruction.
8222 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8223 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
8224 type of quad memory support is available.
8226 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
8229 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
8230 there is a danger of looping.
8232 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
8234 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8235 force flag_ira_loop_pressure if set via command line.
8237 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8239 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
8240 (ashr_simd): New builtin handling DI mode.
8241 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
8242 (aarch64_sshr_simddi): New match pattern.
8243 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
8244 (vshrd_n_s64): Likewise.
8245 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
8247 2014-01-23 Nick Clifton <nickc@redhat.com>
8249 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
8250 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
8251 favour of mcu specific scripts.
8252 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
8255 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8256 Alex Velenko <Alex.Velenko@arm.com>
8258 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
8259 (vaddv_s16): Likewise.
8260 (vaddv_s32): Likewise.
8261 (vaddv_u8): Likewise.
8262 (vaddv_u16): Likewise.
8263 (vaddv_u32): Likewise.
8264 (vaddvq_s8): Likewise.
8265 (vaddvq_s16): Likewise.
8266 (vaddvq_s32): Likewise.
8267 (vaddvq_s64): Likewise.
8268 (vaddvq_u8): Likewise.
8269 (vaddvq_u16): Likewise.
8270 (vaddvq_u32): Likewise.
8271 (vaddvq_u64): Likewise.
8272 (vaddv_f32): Likewise.
8273 (vaddvq_f32): Likewise.
8274 (vaddvq_f64): Likewise.
8275 (vmaxv_f32): Likewise.
8276 (vmaxv_s8): Likewise.
8277 (vmaxv_s16): Likewise.
8278 (vmaxv_s32): Likewise.
8279 (vmaxv_u8): Likewise.
8280 (vmaxv_u16): Likewise.
8281 (vmaxv_u32): Likewise.
8282 (vmaxvq_f32): Likewise.
8283 (vmaxvq_f64): Likewise.
8284 (vmaxvq_s8): Likewise.
8285 (vmaxvq_s16): Likewise.
8286 (vmaxvq_s32): Likewise.
8287 (vmaxvq_u8): Likewise.
8288 (vmaxvq_u16): Likewise.
8289 (vmaxvq_u32): Likewise.
8290 (vmaxnmv_f32): Likewise.
8291 (vmaxnmvq_f32): Likewise.
8292 (vmaxnmvq_f64): Likewise.
8293 (vminv_f32): Likewise.
8294 (vminv_s8): Likewise.
8295 (vminv_s16): Likewise.
8296 (vminv_s32): Likewise.
8297 (vminv_u8): Likewise.
8298 (vminv_u16): Likewise.
8299 (vminv_u32): Likewise.
8300 (vminvq_f32): Likewise.
8301 (vminvq_f64): Likewise.
8302 (vminvq_s8): Likewise.
8303 (vminvq_s16): Likewise.
8304 (vminvq_s32): Likewise.
8305 (vminvq_u8): Likewise.
8306 (vminvq_u16): Likewise.
8307 (vminvq_u32): Likewise.
8308 (vminnmv_f32): Likewise.
8309 (vminnmvq_f32): Likewise.
8310 (vminnmvq_f64): Likewise.
8312 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8314 * config/aarch64/aarch64-simd.md
8315 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
8316 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
8317 (*aarch64_mul3_elt<mode>): Likewise.
8318 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
8319 (*aarch64_mul3_elt_to_64v2df): Likewise.
8320 (*aarch64_mla_elt<mode>): Likewise.
8321 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8322 (*aarch64_mls_elt<mode>): Likewise.
8323 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8324 (*aarch64_fma4_elt<mode>): Likewise.
8325 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8326 (*aarch64_fma4_elt_to_64v2df): Likewise.
8327 (*aarch64_fnma4_elt<mode>): Likewise.
8328 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8329 (*aarch64_fnma4_elt_to_64v2df): Likewise.
8330 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
8331 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
8332 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8333 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8334 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
8335 (aarch64_sqdmull_lane<mode>_internal): Likewise.
8336 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
8338 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
8340 * config/aarch64/aarch64-simd.md
8341 (aarch64_be_checked_get_lane<mode>): New define_expand.
8342 * config/aarch64/aarch64-simd-builtins.def
8343 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
8344 New builtin definition.
8345 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
8346 Use new safe be builtin.
8348 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8350 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
8352 (aarch64_be_st1<mode>): Likewise.
8353 (aarch_ld1<VALL:mode>): Define_expand modified.
8354 (aarch_st1<VALL:mode>): Likewise.
8355 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
8356 (UNSPEC_ST1): Likewise.
8358 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
8360 * config/microblaze/microblaze.md: Add trap insn and attribute
8362 2014-01-23 Dodji Seketeli <dodji@redhat.com>
8364 PR preprocessor/58580
8365 * input.h (location_get_source_line): Take an additional line_size
8367 (void diagnostics_file_cache_fini): Declare new function.
8368 * input.c (struct fcache): New type.
8369 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
8370 New static constants.
8371 (diagnostic_file_cache_init, total_lines_num)
8372 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
8373 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
8374 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
8375 (get_next_line, read_next_line, goto_next_line, read_line_num):
8376 New static function definitions.
8377 (diagnostic_file_cache_fini): New function.
8378 (location_get_source_line): Take an additional output line_len
8379 parameter. Re-write using lookup_or_add_file_to_cache_tab and
8381 * diagnostic.c (diagnostic_finish): Call
8382 diagnostic_file_cache_fini.
8383 (adjust_line): Take an additional input parameter for the length
8384 of the line, rather than calculating it with strlen.
8385 (diagnostic_show_locus): Adjust the use of
8386 location_get_source_line and adjust_line with respect to their new
8387 signature. While displaying a line now, do not stop at the first
8388 null byte. Rather, display the zero byte as a space and keep
8389 going until we reach the size of the line.
8390 * Makefile.in: Add vec.o to OBJS-libcommon
8392 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8393 Ilya Tocar <ilya.tocar@intel.com>
8395 * config/i386/avx512fintrin.h (_mm512_kmov): New.
8396 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
8397 (__builtin_ia32_kmov16): Ditto.
8398 * config/i386/i386.md (UNSPEC_KMOV): New.
8401 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8403 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
8404 (_mm512_storeu_si512): Ditto.
8406 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
8409 * rtl.h (get_referenced_operands): Declare.
8410 * recog.c (get_referenced_operands): New function.
8411 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
8412 operands have been referenced when recording LO_SUM references.
8414 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
8416 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
8418 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8420 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
8421 Enable for generic and recent AMD targets.
8423 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8425 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
8426 ARG_SIZE note when adjustment was eliminated.
8428 2014-01-22 Jeff Law <law@redhat.com>
8430 PR tree-optimization/59597
8431 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
8432 in file. Accept new argument REGISTERING and use it to modify
8433 dump output appropriately.
8434 (register_jump_thread): Corresponding changes.
8435 (mark_threaded_blocks): Reinstate code to cancel unprofitable
8436 thread paths involving joiner blocks. Add code to dump cancelled
8437 jump threading paths.
8439 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
8441 PR rtl-optimization/59477
8442 * lra-constraints.c (inherit_in_ebb): Process call for living hard
8443 regs. Update reloads_num and potential_reload_hard_regs for all insns.
8445 2014-01-22 Tom Tromey <tromey@redhat.com>
8447 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
8449 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
8451 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8453 PR rtl-optimization/59896
8454 * lra-constraints.c (process_alt_operands): Check unused note for
8455 matched operands of insn with no output reloads.
8457 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
8459 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
8460 (mips_move_from_gpr_cost): Likewise.
8462 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8464 PR rtl-optimization/59858
8465 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
8466 ira_class_hard_regs_num.
8467 (process_alt_operands): Increase reject for dying matched operand.
8469 2014-01-21 Jakub Jelinek <jakub@redhat.com>
8472 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
8473 smaller than size, perform several stores or loads and stores
8474 at dst + count - size to store or copy all of size bytes, rather
8475 than just last modesize bytes.
8477 2014-01-20 DJ Delorie <dj@redhat.com>
8479 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
8480 that CLOBBERs are REGs before propogating their values.
8482 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
8485 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
8486 (cgraph_inline_failed_type): New function.
8487 * cgraph.h (DEFCIFCODE): Add type.
8488 (cgraph_inline_failed_type_t): New enum.
8489 (cgraph_inline_failed_type): New prototype.
8490 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
8491 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
8492 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
8493 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
8494 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
8495 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
8496 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
8497 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
8498 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
8499 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
8500 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
8501 OPTIMIZATION_MISMATCH.
8502 * tree-inline.c (expand_call_inline): Emit errors during
8503 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
8505 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8508 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
8509 mode attribute in insn output.
8511 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
8513 * output.h (output_constant): Delete.
8514 * varasm.c (output_constant): Make private.
8516 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
8518 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
8520 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8523 * tree.h (fold_builtin_strcat): New prototype.
8524 * builtins.c (fold_builtin_strcat): No longer static. Add len
8525 argument, if non-NULL, don't call c_strlen. Optimize
8526 directly into __builtin_memcpy instead of __builtin_strcpy.
8527 (fold_builtin_2): Adjust fold_builtin_strcat caller.
8528 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
8530 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8532 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8533 for SImode_address_operand operands, having only a REG argument.
8535 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
8537 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
8538 loader name using mbig-endian.
8539 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
8541 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8543 * doc/invoke.texi (-march): Clarify documentation for AArch64.
8547 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
8549 * config/aarch64/aarch64-protos.h
8550 (aarch64_cannot_change_mode_class_ptr): Declare.
8551 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
8552 aarch64_cannot_change_mode_class_ptr): New.
8553 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
8554 backend hook aarch64_cannot_change_mode_class.
8556 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8558 * common/config/aarch64/aarch64-common.c
8559 (aarch64_handle_option): Don't handle any option order logic here.
8560 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
8561 selected_cpu, warn on architecture version mismatch.
8562 (aarch64_override_options): Fix parsing order for option strings.
8564 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8565 Iain Sandoe <iain@codesourcery.com>
8568 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
8569 warning. Amend comment to reflect current functionality.
8571 2014-01-20 Richard Biener <rguenther@suse.de>
8574 * builtins.c (fold_builtin_strcat): Remove case better handled
8575 by tree-ssa-strlen.c.
8577 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
8579 * config/aarch64/aarch64.opt
8580 (mcpu, march, mtune): Make case-insensitive.
8582 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8585 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8586 if operands[1] is a REG or ZERO_EXTEND of a REG.
8588 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
8590 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
8592 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
8594 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
8595 long non-pic millicode calls.
8597 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8599 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
8601 2014-01-19 Kito Cheng <kito@0xlab.org>
8603 * builtins.c (expand_movstr): Check movstr expand done or fail.
8605 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8606 H.J. Lu <hongjiu.lu@intel.com>
8609 * config/i386/i386.md (*lea<mode>): Zero-extend return register
8610 to DImode for zero-extended addresses.
8612 2014-01-19 Jakub Jelinek <jakub@redhat.com>
8614 PR rtl-optimization/57763
8615 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
8616 on the new indirect jump_insn and increment LABEL_NUSES (label).
8618 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
8622 * config.gcc (x86_archs): New variable.
8623 (x86_64_archs): Likewise.
8624 (x86_cpus): Likewise.
8625 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
8626 --with-arch/--with-cpu= options.
8627 Support --with-arch=/--with-cpu={nehalem,westmere,
8628 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
8630 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8632 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
8633 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
8635 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8637 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
8639 2014-01-18 Jakub Jelinek <jakub@redhat.com>
8642 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
8643 clear cpp_get_options (parse_in)->warn_unused_macros for
8644 ix86_target_macros_internal with cpp_define.
8646 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
8648 * jump.c (delete_related_insns): Keep (use (insn))s.
8649 * reorg.c (redundant_insn): Check for barriers too.
8651 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8653 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
8655 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
8657 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
8658 call to $$dyncall when TARGET_LONG_CALLS is true.
8660 2014-01-17 Jeff Law <law@redhat.com>
8662 * ree.c (combine_set_extension): Temporarily disable test for
8663 changing number of hard registers.
8665 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
8668 * ipa-inline-analysis.c (inline_free_summary):
8669 Do not free summary of aliases.
8671 2014-01-17 Jakub Jelinek <jakub@redhat.com>
8674 * gimplify.c (gimplify_expr): Use create_tmp_var
8675 instead of create_tmp_var_raw. If cond doesn't have
8676 integral type, don't add the IFN_ANNOTATE builtin at all.
8678 2014-01-17 Martin Jambor <mjambor@suse.cz>
8681 * ipa-cp.c (prev_edge_clone): New variable.
8682 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
8683 Also resize prev_edge_clone vector.
8684 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
8685 (ipcp_edge_removal_hook): New function.
8686 (ipcp_driver): Register ipcp_edge_removal_hook.
8688 2014-01-17 Andrew Pinski <apinski@cavium.com>
8689 Steve Ellcey <sellcey@mips.com>
8692 * config/mips/mips.c (mips_print_operand): Check operand mode instead
8695 2014-01-17 Jeff Law <law@redhat.com>
8698 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
8699 so that pass_ccp runs first.
8701 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8703 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
8704 (ix86_adjust_cost): Use !TARGET_XXX.
8705 (do_reorder_for_imul): Likewise.
8706 (swap_top_of_ready_list): Likewise.
8707 (ix86_sched_reorder): Likewise.
8709 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8711 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
8712 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
8713 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
8714 (intel_memset): New. Duplicate slm_memset.
8715 (intel_cost): New. Duplicate slm_cost.
8716 (m_INTEL): New macro.
8717 (processor_target_table): Add "intel".
8718 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
8719 with PROCESSOR_INTEL for "intel".
8720 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
8721 PROCESSOR_SILVERMONT.
8722 (ix86_issue_rate): Likewise.
8723 (ix86_adjust_cost): Likewise.
8724 (ia32_multipass_dfa_lookahead): Likewise.
8725 (swap_top_of_ready_list): Likewise.
8726 (ix86_sched_reorder): Likewise.
8727 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
8728 instead of TARGET_OPT_AGU.
8729 * config/i386/i386.h (TARGET_INTEL): New.
8730 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
8731 (processor_type): Add PROCESSOR_INTEL.
8732 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
8733 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
8735 2014-01-17 Marek Polacek <polacek@redhat.com>
8738 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
8741 2014-01-17 Richard Biener <rguenther@suse.de>
8743 PR tree-optimization/46590
8744 * opts.c (default_options_table): Add entries for
8745 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
8746 all enabled at -O1 but not for -Og.
8747 * common.opt (fbranch-count-reg): Remove Init(1).
8748 (fmove-loop-invariants): Likewise.
8749 (ftree-pta): Likewise.
8751 2014-01-17 Jakub Jelinek <jakub@redhat.com>
8753 * config/i386/i386.c (ix86_data_alignment): For compatibility with
8754 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
8755 decls to at least the GCC 4.8 used alignments.
8758 * tree-nested.c (convert_nonlocal_reference_stmt,
8759 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
8760 of GIMPLE_BIND stmts, adjust associated decls.
8762 2014-01-17 Richard Biener <rguenther@suse.de>
8764 PR tree-optimization/46590
8765 * vec.h (vec<>::bseach): New member function implementing
8766 binary search according to C89 bsearch.
8767 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
8768 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
8769 bitmap pointer again. Make accesses_in_loop a flat array.
8770 (mem_ref_obstack): New global.
8771 (outermost_indep_loop): Adjust for mem_ref->stored changes.
8772 (mark_ref_stored): Likewise.
8773 (ref_indep_loop_p_2): Likewise.
8774 (set_ref_stored_in_loop): New helper function.
8775 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
8776 (memref_free): Adjust.
8777 (record_mem_ref_loc): Simplify.
8778 (gather_mem_refs_stmt): Adjust.
8779 (sort_locs_in_loop_postorder_cmp): New function.
8780 (analyze_memory_references): Sort accesses_in_loop after
8781 loop postorder number.
8782 (find_ref_loc_in_loop_cmp): New function.
8783 (for_all_locs_in_loop): Find relevant cluster of locs in
8784 accesses_in_loop and iterate without recursion.
8785 (execute_sm): Avoid uninit warning.
8786 (struct ref_always_accessed): Simplify.
8787 (ref_always_accessed::operator ()): Likewise.
8788 (ref_always_accessed_p): Likewise.
8789 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
8790 loop postorder numbers here.
8791 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
8794 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
8797 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
8798 on decls for which assemble_alias has been called.
8800 2014-01-17 Nick Clifton <nickc@redhat.com>
8802 * config/msp430/msp430.opt: (mcpu): New option.
8803 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
8804 (msp430_option_override): Parse target_cpu. If the MCU name
8805 matches a generic string, clear target_mcu.
8806 (msp430_attr): Allow numeric interrupt values up to 63.
8807 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
8808 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
8810 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
8812 * config/msp430/msp430.md (popm): Use %J rather than %I.
8813 (addsi3): Use msp430_nonimmediate_operand for operand 2.
8814 (addhi_cy_i): Use immediate_operand for operand 2.
8815 * doc/invoke.texi: Document -mcpu option.
8817 2014-01-17 Richard Biener <rguenther@suse.de>
8819 PR rtl-optimization/38518
8820 * df.h (df_analyze_loop): Declare.
8821 * df-core.c: Include cfgloop.h.
8822 (df_analyze_1): Split out main part of df_analyze.
8823 (df_analyze): Adjust.
8824 (loop_inverted_post_order_compute): New function.
8825 (loop_post_order_compute): Likewise.
8826 (df_analyze_loop): New function avoiding whole-function
8828 * loop-invariant.c (find_defs): Use df_analyze_loop.
8829 (find_invariants): Adjust.
8830 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
8832 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
8834 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
8835 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
8837 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
8839 * ipa-ref.c (ipa_remove_stmt_references): Fix references
8840 traversal when removing references.
8842 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
8845 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
8847 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
8850 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
8851 pushing a reload for an autoinc when we had previously reloaded an
8852 inner part of the address.
8854 2014-01-16 Jakub Jelinek <jakub@redhat.com>
8856 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
8858 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
8859 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
8860 when not giving up or versioning for alias only because of
8862 (vect_analyze_data_ref_dependences): Set to true.
8863 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
8865 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
8866 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
8870 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
8873 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
8874 operand 0 predicate for gathers, use a new pseudo as subtarget.
8876 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
8879 * lra-constraints.c (process_alt_operands): Add printing debug info.
8880 Check absence of input/output reloads for matched operands too.
8882 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
8884 PR rtl-optimization/59835
8885 * ira.c (ira_init_register_move_cost): Increase cost for
8888 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
8890 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
8892 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
8895 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
8896 non-register objects. Use gen_(high/low)part more consistently.
8899 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8902 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
8903 endian support, remove tests for WORDS_BIG_ENDIAN.
8904 (p8_mfvsrd_3_<mode>): Likewise.
8905 (reload_gpr_from_vsx<mode>): Likewise.
8906 (reload_gpr_from_vsxsf): Likewise.
8907 (p8_mfvsrd_4_disf): Likewise.
8909 2014-01-16 Richard Biener <rguenther@suse.de>
8911 PR rtl-optimization/46590
8912 * lcm.c (compute_antinout_edge): Use postorder iteration.
8913 (compute_laterin): Use inverted postorder iteration.
8915 2014-01-16 Nick Clifton <nickc@redhat.com>
8918 * varasm.c (output_constant): Return the number of bytes actually
8920 (output_constructor_array_range): Update the field size with the
8921 number of bytes emitted by output_constant.
8922 (output_constructor_regular_field): Likewise. Also do not
8923 complain if the total number of bytes emitted is now greater
8924 than the expected fieldpos.
8925 * output.h (output_constant): Update prototype and descriptive comment.
8927 2014-01-16 Marek Polacek <polacek@redhat.com>
8930 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
8931 it is error_mark_node.
8933 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
8935 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
8936 VALID_AVX256_REG_OR_OI_MODE.
8938 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
8940 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
8941 current procedure should be profiled.
8943 2014-01-15 Andrew Pinski <apinski@cavium.com>
8945 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
8946 of moving from/to the STACK_REG register class.
8948 2014-01-15 Richard Henderson <rth@redhat.com>
8951 * reginfo.c (global_regs_decl): Globalize.
8952 * rtl.h (global_regs_decl): Declare.
8953 * ira.c (do_reload): Diagnose frame_pointer_needed and it
8954 reserved via global_regs.
8956 2014-01-15 Teresa Johnson <tejohnson@google.com>
8958 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
8960 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8962 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
8963 and vmulosh rather than call gen_vec_widen_smult_*.
8964 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
8965 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
8966 (vec_widen_smult_even_v16qi): Likewise.
8967 (vec_widen_umult_even_v8hi): Likewise.
8968 (vec_widen_smult_even_v8hi): Likewise.
8969 (vec_widen_umult_odd_v16qi): Likewise.
8970 (vec_widen_smult_odd_v16qi): Likewise.
8971 (vec_widen_umult_odd_v8hi): Likewise.
8972 (vec_widen_smult_odd_v8hi): Likewise.
8973 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
8974 vmuloub rather than call gen_vec_widen_umult_*.
8975 (vec_widen_umult_lo_v16qi): Likewise.
8976 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
8977 vmulosb rather than call gen_vec_widen_smult_*.
8978 (vec_widen_smult_lo_v16qi): Likewise.
8979 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
8980 rather than call gen_vec_widen_umult_*.
8981 (vec_widen_umult_lo_v8hi): Likewise.
8982 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
8983 rather than call gen_vec_widen_smult_*.
8984 (vec_widen_smult_lo_v8hi): Likewise.
8986 2014-01-15 Jeff Law <law@redhat.com>
8988 PR tree-optimization/59747
8989 * ree.c (find_and_remove_re): Properly handle case where a second
8990 eliminated extension requires widening a copy created for elimination
8991 of a prior extension.
8992 (combine_set_extension): Ensure that the number of hard regs needed
8993 for a destination register does not change when we widen it.
8995 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
8997 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
8998 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
8999 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
9000 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
9001 (avr-*-rtems*): Likewise.
9002 (bfin*-rtems*): Likewise.
9003 (moxie-*-rtems*): Likewise.
9004 (h8300-*-rtems*): Likewise.
9005 (i[34567]86-*-rtems*): Likewise.
9006 (lm32-*-rtems*): Likewise.
9007 (m32r-*-rtems*): Likewise.
9008 (m68k-*-rtems*): Likewise.
9009 (microblaze*-*-rtems*): Likewise.
9010 (mips*-*-rtems*): Likewise.
9011 (powerpc-*-rtems*): Likewise.
9012 (sh-*-rtems*): Likewise.
9013 (sparc-*-rtems*): Likewise.
9014 (sparc64-*-rtems*): Likewise.
9015 (v850-*-rtems*): Likewise.
9016 (m32c-*-rtems*): Likewise.
9018 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9020 PR rtl-optimization/59511
9021 * ira.c (ira_init_register_move_cost): Use memory costs for some
9022 cases of register move cost calculations.
9023 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9024 instead of BB frequency.
9025 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9026 * lra-assigns.c (find_hard_regno_for): Ditto.
9028 2014-01-15 Richard Biener <rguenther@suse.de>
9030 PR tree-optimization/59822
9031 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9032 (vectorizable_load): Use it to hoist defs of uses of invariant
9033 loads out of the loop.
9035 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9036 Kugan Vivekanandarajah <kuganv@linaro.org>
9039 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9042 2014-01-15 Richard Biener <rguenther@suse.de>
9044 PR rtl-optimization/59802
9045 * lcm.c (compute_available): Use inverted postorder to seed
9046 the initial worklist.
9048 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9051 * config/s390/s390.c (s390_preferred_reload_class): Don't return
9052 ADDR_REGS for invalid symrefs in non-PIC code.
9054 2014-01-15 Jakub Jelinek <jakub@redhat.com>
9057 * builtins.c (determine_block_size): Initialize *probable_max_size
9058 even if len_rtx is CONST_INT.
9060 2014-01-14 Andrew Pinski <apinski@cavium.com>
9062 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
9063 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
9064 (cortexa53_tunings): Likewise.
9065 (aarch64_sched_issue_rate): New function.
9066 (TARGET_SCHED_ISSUE_RATE): Define.
9068 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9070 * ira-costs.c (find_costs_and_classes): Add missed
9071 ira_init_register_move_cost_if_necessary.
9073 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9076 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
9078 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
9081 * config/i386/i386.c (type_natural_mode): Add a bool parameter
9082 to indicate if type is used for function return value. Warn ABI
9083 change if the vector mode isn't available for function return value.
9084 (ix86_function_arg_advance): Pass false to type_natural_mode.
9085 (ix86_function_arg): Likewise.
9086 (ix86_gimplify_va_arg): Likewise.
9087 (function_arg_32): Don't warn ABI change.
9088 (ix86_function_value): Pass true to type_natural_mode.
9089 (ix86_return_in_memory): Likewise.
9090 (ix86_struct_value_rtx): Removed.
9091 (TARGET_STRUCT_VALUE_RTX): Likewise.
9093 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9095 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
9096 converting a conditional jump into a conditional return.
9098 2014-01-14 Richard Biener <rguenther@suse.de>
9100 PR tree-optimization/58921
9101 PR tree-optimization/59006
9102 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
9103 hoisting invariant stmts.
9104 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
9105 invariant loads on the preheader edge if possible.
9107 2014-01-14 Joey Ye <joey.ye@arm.com>
9109 * doc/plugin.texi (Building GCC plugins): Update to C++.
9111 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
9113 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
9114 (_mm_rcp28_round_ss): Ditto.
9115 (_mm_rsqrt28_round_sd): Ditto.
9116 (_mm_rsqrt28_round_ss): Ditto.
9117 (_mm_rcp28_sd): Ditto.
9118 (_mm_rcp28_ss): Ditto.
9119 (_mm_rsqrt28_sd): Ditto.
9120 (_mm_rsqrt28_ss): Ditto.
9121 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
9122 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
9123 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
9124 (IX86_BUILTIN_RCP28SD): Ditto.
9125 (IX86_BUILTIN_RCP28SS): Ditto.
9126 (IX86_BUILTIN_RSQRT28SD): Ditto.
9127 (IX86_BUILTIN_RSQRT28SS): Ditto.
9128 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
9129 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
9130 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
9131 (ix86_expand_special_args_builtin): Expand new FTYPE.
9132 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
9133 (srcp14<mode>): Make insn unary.
9134 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
9135 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
9136 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
9137 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
9138 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
9139 Fix rounding: make it SAE only.
9140 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9142 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9144 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
9145 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
9146 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
9147 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
9148 (round_saeonly_mask_scalar_operand4): Ditto.
9149 (round_saeonly_mask_scalar_op3): Ditto.
9150 (round_saeonly_mask_scalar_op4): Ditto.
9152 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9154 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9155 Implement -maltivec=be for vec_insert and vec_extract.
9157 2014-01-10 DJ Delorie <dj@redhat.com>
9159 * config/msp430/msp430.md (call_internal): Don't allow memory
9160 references with SP as the base register.
9161 (call_value_internal): Likewise.
9162 * config/msp430/constraints.md (Yc): New. For memory references
9163 that don't use SP as a base register.
9165 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
9166 "an integer without a # prefix"
9167 * config/msp430/msp430.md (epilogue_helper): Use it.
9169 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9172 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
9173 AVX512F gather builtins.
9174 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
9175 on gather decls with INTEGER_TYPE masktype.
9176 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
9177 directly into the builtin rather than hoisting it before loop.
9179 PR tree-optimization/59387
9180 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
9181 (scev_const_prop): If folded_casts and type has undefined overflow,
9182 use force_gimple_operand instead of force_gimple_operand_gsi and
9183 for each added stmt if it is assign with
9184 arith_code_with_undefined_signed_overflow, call
9185 rewrite_to_defined_overflow.
9186 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
9187 gimple-fold.h instead.
9188 (arith_code_with_undefined_signed_overflow,
9189 rewrite_to_defined_overflow): Moved to ...
9190 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
9191 rewrite_to_defined_overflow): ... here. No longer static.
9192 Include gimplify-me.h.
9193 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
9194 rewrite_to_defined_overflow): New prototypes.
9196 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9198 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
9200 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
9202 * builtins.c (get_object_alignment_2): Minor tweak.
9203 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
9205 2014-01-13 Christian Bruel <christian.bruel@st.com>
9207 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
9208 optimized non constant lengths.
9210 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9213 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
9214 load as __atomic_load_N if possible.
9216 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9218 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
9220 (rs6000_expand_builtin): Adjust call.
9222 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9225 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
9226 * config/rs6000/rs6000.c: Include target-globals.h.
9227 (rs6000_set_current_function): Instead of doing target_reinit
9228 unconditionally, use save_target_globals_default_opts and
9229 restore_target_globals.
9231 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
9233 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
9234 (rs6000_expand_builtin): Handle mffs and mtfsf.
9235 (rs6000_init_builtins): Define mffs and mtfsf.
9236 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
9237 (rs6000_mffs): New pattern.
9238 (rs6000_mtfsf): New pattern.
9240 2014-01-11 Bin Cheng <bin.cheng@arm.com>
9242 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
9243 Start narrowing with START. Apply candidate-use pair
9244 and check overall cost in narrowing.
9245 (iv_ca_prune): Pass new argument.
9247 2014-01-10 Jeff Law <law@redhat.com>
9250 * ree.c (combine_reaching_defs): Ensure the defining statement
9251 occurs before the extension when optimizing extensions with
9252 different source and destination hard registers.
9254 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9257 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
9258 vtables into the type inheritance graph.
9260 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9262 PR rtl-optimization/59754
9263 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
9264 modes in the REGNO != REGNO case.
9266 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9268 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
9270 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9272 PR tree-optimization/59745
9273 * tree-predcom.c (tree_predictive_commoning_loop): Call
9274 free_affine_expand_cache if giving up because components is NULL.
9276 * target-globals.c (save_target_globals): Allocate < 4KB structs using
9277 GC in payload of target_globals struct instead of allocating them on
9278 the heap and the larger structs separately using GC.
9279 * target-globals.h (struct target_globals): Make regs, hard_regs,
9280 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
9281 of GTY((skip)) and change type to void *.
9282 (reset_target_globals): Cast loads from those fields to corresponding
9285 2014-01-10 Steve Ellcey <sellcey@mips.com>
9288 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
9289 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
9290 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
9292 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
9295 * aarch64-modes.def (CC_Zmode): New flags mode.
9296 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
9297 represents an equality.
9298 (aarch64_get_condition_code): Handle CC_Zmode.
9299 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
9301 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9303 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
9304 extraction in good case.
9306 2014-01-10 Richard Biener <rguenther@suse.de>
9308 PR tree-optimization/59374
9309 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
9310 checking after SLP discovery. Mark stmts not participating
9311 in any SLP instance properly.
9313 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9315 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
9316 when handling a SET rtx.
9318 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9320 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
9321 (cortex-a57): Likewise.
9322 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
9324 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9326 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
9327 non-iwmmxt builtins.
9329 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9333 * ipa-devirt.c record_target_from_binfo): Take as argument
9334 stack of binfos and lookup matching one for virtual inheritance.
9335 (possible_polymorphic_call_targets_1): Update.
9337 2014-01-10 Huacai Chen <chenhc@lemote.com>
9339 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
9340 kernel strings for Loongson-2E/2F/3A.
9342 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9345 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
9346 is_gimple_call before calling gimple_call_internal_p.
9348 2014-01-09 Steve Ellcey <sellcey@mips.com>
9350 * Makefile.in (TREE_FLOW_H): Remove.
9351 (TREE_SSA_H): Add file names from tree-flow.h.
9352 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
9353 * tree.h: Remove tree-flow.h reference.
9354 * hash-table.h: Remove tree-flow.h reference.
9355 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
9356 reference with tree-ssa-loop.h.
9358 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9360 * doc/invoke.texi: Add -maltivec={be,le} options, and document
9361 default element-order behavior for -maltivec.
9362 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
9363 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
9364 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
9365 when targeting big endian, at least for now.
9366 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
9368 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9371 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
9372 var satisfies use_register_for_decl, just take into account type
9373 alignment, rather than decl alignment.
9375 PR tree-optimization/59622
9376 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
9377 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
9378 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
9379 Don't devirtualize for inplace at all. For targets.length () == 1,
9380 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
9382 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9384 * config/i386/i386.md (cpu): Remove the unused btver1.
9386 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9388 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
9390 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9393 * tree-core.h (struct target_globals): New forward declaration.
9394 (struct tree_target_option): Add globals field.
9395 * tree.h (TREE_TARGET_GLOBALS): Define.
9396 (prepare_target_option_nodes_for_pch): New prototype.
9397 * target-globals.h (struct target_globals): Define even if
9399 * tree.c (prepare_target_option_node_for_pch,
9400 prepare_target_option_nodes_for_pch): New functions.
9401 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
9402 * config/i386/i386.c: Include target-globals.h.
9403 (ix86_set_current_function): Instead of doing target_reinit
9404 unconditionally, use save_target_globals_default_opts and
9405 restore_target_globals.
9407 2014-01-09 Richard Biener <rguenther@suse.de>
9409 PR tree-optimization/59715
9410 * tree-cfg.h (split_critical_edges): Declare.
9411 * tree-cfg.c (split_critical_edges): Export.
9412 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
9414 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
9416 * cfgexpand.c (expand_stack_vars): Optionally disable
9417 asan stack protection.
9418 (expand_used_vars): Likewise.
9419 (partition_stack_vars): Likewise.
9420 * asan.c (asan_emit_stack_protection): Optionally disable
9421 after return stack usage.
9422 (instrument_derefs): Optionally disable memory access instrumentation.
9423 (instrument_builtin_call): Likewise.
9424 (instrument_strlen_call): Likewise.
9425 (asan_protect_global): Optionally disable global variables protection.
9426 * doc/invoke.texi: Added doc for new options.
9427 * params.def: Added new options.
9428 * params.h: Likewise.
9430 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9432 PR rtl-optimization/59724
9433 * ifcvt.c (cond_exec_process_if_block): Don't call
9434 flow_find_head_matching_sequence with 0 longest_match.
9435 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
9436 non-active insns if !stop_after.
9437 (try_head_merge_bb): Revert 2014-01-07 changes.
9439 2014-01-08 Jeff Law <law@redhat.com>
9441 * ree.c (get_sub_rtx): New function, extracted from...
9442 (merge_def_and_ext): Here.
9443 (combine_reaching_defs): Use get_sub_rtx.
9445 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
9447 * cgraph.h (varpool_variable_node): Do not choke on null node.
9449 2014-01-08 Catherine Moore <clm@codesourcery.com>
9451 * config/mips/mips.md (simple_return): Attempt to use JRC
9453 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
9455 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9457 PR rtl-optimization/59137
9458 * reorg.c (steal_delay_list_from_target): Call update_block for
9460 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
9462 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9464 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
9465 two duplicate entries.
9467 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9470 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
9472 * config/mips/mips.c (mips_truncated_op_cost): New function.
9473 (mips_rtx_costs): Adjust test for BADDU.
9474 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
9476 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
9478 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
9479 (*baddu_si): ...this new pattern.
9481 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9484 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
9486 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9489 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
9491 (expand_expr, expand_normal): Adjust.
9492 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
9493 inner_reference_p. Use inner_reference_p to expand inner references.
9494 (store_expr): Adjust.
9495 * cfgexpand.c (expand_call_stmt): Adjust.
9497 2014-01-08 Rong Xu <xur@google.com>
9499 * gcov-io.c (gcov_var): Move from gcov-io.h.
9500 (gcov_position): Ditto.
9501 (gcov_is_error): Ditto.
9502 (gcov_rewrite): Ditto.
9503 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
9504 only part to libgcc/libgcov.h.
9506 2014-01-08 Marek Polacek <polacek@redhat.com>
9509 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
9511 2014-01-08 Marek Polacek <polacek@redhat.com>
9514 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
9516 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9518 PR rtl-optimization/59649
9519 * stor-layout.c (get_mode_bounds): For BImode return
9520 0 and STORE_FLAG_VALUE.
9522 2014-01-08 Richard Biener <rguenther@suse.de>
9525 * gimple.h (is_gimple_builtin_call): Remove.
9526 (gimple_builtin_call_types_compatible_p): New.
9527 (gimple_call_builtin_p): New overload.
9528 * gimple.c (is_gimple_builtin_call): Remove.
9529 (validate_call): Rename to ...
9530 (gimple_builtin_call_types_compatible_p): ... this and export. Also
9532 (validate_type): New static function.
9533 (gimple_call_builtin_p): New overload and adjust.
9534 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
9535 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
9536 (gimple_fold_stmt_to_constant_1): Likewise.
9537 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
9539 2014-01-08 Richard Biener <rguenther@suse.de>
9542 * gimplify.c (gimplify_expr): Gimplify register-register type
9543 VIEW_CONVERT_EXPRs to separate stmts.
9545 2014-01-07 Jeff Law <law@redhat.com>
9548 * ree.c (combine_set_extension): Handle case where source
9549 and destination registers in an extension insn are different.
9550 (combine_reaching_defs): Allow source and destination registers
9551 in extension to be different under limited circumstances.
9552 (add_removable_extension): Remove restriction that the
9553 source and destination registers in the extension are the same.
9554 (find_and_remove_re): Emit a copy from the extension's
9555 destination to its source after the defining insn if
9556 the source and destination registers are different.
9559 * ifcvt.c (merge_if_block): If we are merging a block with more than
9560 one successor with a block with no successors, remove any BARRIER
9561 after the second block.
9563 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
9565 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
9567 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
9570 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
9571 for 14-bit register offsets when INT14_OK_STRICT is false.
9573 2014-01-07 Roland Stigge <stigge@antcom.de>
9574 Michael Meissner <meissner@linux.vnet.ibm.com>
9577 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
9578 Only check TFmode for SPE constants. Don't check TImode or TDmode.
9580 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
9582 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
9585 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
9587 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
9588 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
9589 rtx is const0_rtx or not.
9591 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
9594 * target-globals.c (save_target_globals): Remove this_fn_optab
9596 * toplev.c: Include optabs.h.
9597 (target_reinit): Temporarily restore the global options if another
9598 set of options are in force.
9600 2014-01-07 Jakub Jelinek <jakub@redhat.com>
9602 PR rtl-optimization/58668
9603 * cfgcleanup.c (flow_find_cross_jump): Don't count
9604 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
9605 to determine what is counted.
9606 (flow_find_head_matching_sequence): Use active_insn_p to determine
9608 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
9610 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
9611 determine what is counted.
9613 PR tree-optimization/59643
9614 * tree-predcom.c (split_data_refs_to_components): If one dr is
9615 read and one write, determine_offset fails and the write isn't
9616 in the bad component, just put the read into the bad component.
9618 2014-01-07 Mike Stump <mikestump@comcast.net>
9619 Jakub Jelinek <jakub@redhat.com>
9622 * tree-core.h (struct tree_optimization_option): Change optabs
9623 type from unsigned char * to void *.
9624 * optabs.c (init_tree_optimization_optabs): Adjust
9625 TREE_OPTIMIZATION_OPTABS initialization.
9627 2014-01-06 Jakub Jelinek <jakub@redhat.com>
9630 * config/i386/i386.h (struct machine_function): Add
9631 no_drap_save_restore field.
9632 * config/i386/i386.c (ix86_save_reg): Use
9633 !cfun->machine->no_drap_save_restore instead of
9634 crtl->stack_realign_needed.
9635 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
9636 this function clears frame_pointer_needed. Set
9637 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
9638 and DRAP reg is needed.
9640 2014-01-06 Marek Polacek <polacek@redhat.com>
9643 * doc/implement-c.texi: Mention that other integer types are
9644 permitted as bit-field types in strictly conforming mode.
9646 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
9648 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
9651 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
9653 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
9655 2014-01-06 Martin Jambor <mjambor@suse.cz>
9658 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
9660 * ipa-prop.c (ipa_print_node_params): Fix indentation.
9662 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
9666 * var-tracking.c (add_stores): Preserve the value of the source even if
9667 we don't record the store.
9669 2014-01-06 Terry Guo <terry.guo@arm.com>
9671 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
9673 2014-01-05 Iain Sandoe <iain@codesourcery.com>
9676 * config/darwin.c (darwin_function_section): Adjust return values to
9677 correspond to optimisation changes made in r206070.
9679 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
9681 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
9682 from prefetch_block tune setting.
9683 (nocona_cost): Correct size of prefetch block to 64.
9685 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
9687 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
9688 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
9689 used to save the static chain register in the computation of the offset
9690 from which the FP registers need to be restored.
9692 2014-01-04 Jakub Jelinek <jakub@redhat.com>
9694 PR tree-optimization/59519
9695 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
9696 ICE if get_current_def (current_new_name) is already non-NULL, as long
9697 as it is a phi result of some other phi in *new_exit_bb that has
9700 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
9701 or vmovdqu* for misaligned_operand.
9702 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
9703 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
9704 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
9705 aligned_mem for AVX512F masked aligned load and store builtins and for
9708 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
9710 PR tree-optimization/59651
9711 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
9712 Address range for negative step should be added by TYPE_SIZE_UNIT.
9714 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
9716 * config/m68k/m68k.c (handle_move_double): Handle pushes with
9717 overlapping registers also for registers other than the stack pointer.
9719 2014-01-03 Marek Polacek <polacek@redhat.com>
9722 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
9725 2014-01-03 Jakub Jelinek <jakub@redhat.com>
9728 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
9731 * config/i386/i386.md (MODE_SIZE): New mode attribute.
9732 (push splitter): Use <P:MODE_SIZE> instead of
9733 GET_MODE_SIZE (<P:MODE>mode).
9734 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
9735 (mov -1, reg peephole2): Likewise.
9736 * config/i386/sse.md (*mov<mode>_internal,
9737 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
9738 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
9739 *<code><mode>3, *andnot<mode>3<mask_name>,
9740 <mask_codefor><code><mode>3<mask_name>): Likewise.
9741 * config/i386/subst.md (mask_mode512bit_condition,
9742 sd_mask_mode512bit_condition): Likewise.
9744 2014-01-02 Xinliang David Li <davidxl@google.com>
9746 PR tree-optimization/59303
9747 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
9748 (dump_predicates): Better output format.
9749 (pred_equal_p): New function.
9750 (is_neq_relop_p): Ditto.
9751 (is_neq_zero_form_p): Ditto.
9752 (pred_expr_equal_p): Ditto.
9753 (pred_neg_p): Ditto.
9754 (simplify_pred): Ditto.
9755 (simplify_preds_2): Ditto.
9756 (simplify_preds_3): Ditto.
9757 (simplify_preds_4): Ditto.
9758 (simplify_preds): Ditto.
9760 (push_to_worklist): Ditto.
9761 (get_pred_info_from_cmp): Ditto.
9762 (is_degenerated_phi): Ditto.
9763 (normalize_one_pred_1): Ditto.
9764 (normalize_one_pred): Ditto.
9765 (normalize_one_pred_chain): Ditto.
9766 (normalize_preds): Ditto.
9767 (normalize_cond_1): Remove function.
9768 (normalize_cond): Ditto.
9769 (is_gcond_subset_of): Ditto.
9770 (is_subset_of_any): Ditto.
9771 (is_or_set_subset_of): Ditto.
9772 (is_and_set_subset_of): Ditto.
9773 (is_norm_cond_subset_of): Ditto.
9774 (pred_chain_length_cmp): Ditto.
9775 (convert_control_dep_chain_into_preds): Type change.
9776 (find_predicates): Ditto.
9777 (find_def_preds): Ditto.
9778 (destroy_predicates_vecs): Ditto.
9779 (find_matching_predicates_in_rest_chains): Ditto.
9780 (use_pred_not_overlap_with_undef_path_pred): Ditto.
9781 (is_pred_expr_subset): Ditto.
9782 (is_pred_chain_subset_of): Ditto.
9783 (is_included_in): Ditto.
9784 (is_superset_of): Ditto.
9786 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
9788 Update copyright years.
9790 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
9792 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
9793 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
9794 config/arc/arc.md, config/arc/arc.opt,
9795 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
9796 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
9797 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
9798 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
9799 config/linux-protos.h, config/linux.c, config/winnt-c.c,
9800 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
9801 vtable-verify.c, vtable-verify.h: Use the standard form for the
9804 2014-01-02 Tobias Burnus <burnus@net-b.de>
9806 * gcc.c (process_command): Update copyright notice dates.
9807 * gcov-dump.c: Ditto.
9809 * doc/cpp.texi: Bump @copying's copyright year.
9810 * doc/cppinternals.texi: Ditto.
9811 * doc/gcc.texi: Ditto.
9812 * doc/gccint.texi: Ditto.
9813 * doc/gcov.texi: Ditto.
9814 * doc/install.texi: Ditto.
9815 * doc/invoke.texi: Ditto.
9817 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9819 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
9821 2014-01-01 Jakub Jelinek <jakub@redhat.com>
9823 * config/i386/sse.md (*mov<mode>_internal): Guard
9824 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
9826 PR rtl-optimization/59647
9827 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
9828 new_rtx into UNSIGNED_FLOAT rtxes.
9830 Copyright (C) 2014 Free Software Foundation, Inc.
9832 Copying and distribution of this file, with or without modification,
9833 are permitted in any medium without royalty provided the copyright
9834 notice and this notice are preserved.