1 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
3 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
4 to their respective blocks. Fix inadvertent use of "node".
6 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
8 * emit-rtl.c (init_derived_machine_modes): New functionm, split
10 (init_emit_once): ...here.
11 * rtl.h (init_derived_machine_modes): Declare.
12 * toplev.c (do_compile): Call it even if no_backend.
14 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
15 Mike Stump <mikestump@comcast.net>
16 Richard Sandiford <rdsandiford@googlemail.com>
17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
20 (rtx_equal_for_memref_p): Update comment.
21 (adjust_offset_for_component_ref): Use wide-int interfaces.
22 * builtins.c (get_object_alignment_2): Likewise.
23 (c_readstr): Likewise.
24 (target_char_cast): Add comment.
25 (determine_block_size): Use wide-int interfaces.
26 (expand_builtin_signbit): Likewise.
27 (fold_builtin_int_roundingfn): Likewise.
28 (fold_builtin_bitop): Likewise.
29 (fold_builtin_bswap): Likewise.
30 (fold_builtin_logarithm): Use signop.
31 (fold_builtin_pow): Likewise.
32 (fold_builtin_memory_op): Use wide-int interfaces.
33 (fold_builtin_object_size): Likewise.
34 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
35 nb_iterations_estimate.
36 (record_niter_bound): Use wide-int interfaces.
37 (get_estimated_loop_iterations_int): Likewise.
38 (get_estimated_loop_iterations): Likewise.
39 (get_max_loop_iterations): Likewise.
40 * cfgloop.h: Include wide-int.h.
41 (struct nb_iter_bound): Change bound to widest_int.
42 (struct loop): Change nb_iterations_upper_bound and
43 nb_iterations_estimate to widest_int.
44 (record_niter_bound): Switch to use widest_int.
45 (get_estimated_loop_iterations): Likewise.
46 (get_max_loop_iterations): Likewise.
47 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
49 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
50 * combine.c (try_combine): Likewise.
51 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
52 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
54 (aarch64_float_const_representable_p): Likewise.
55 * config/arc/arc.c: Include wide-int.h.
56 (arc_can_use_doloop_p): Use wide-int interfaces.
57 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
58 (vfp3_const_double_index): Likewise.
59 * config/avr/avr.c (avr_out_round): Likewise.
60 (avr_fold_builtin): Likewise.
61 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
62 (bfin_can_use_doloop_p): Likewise.
63 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
64 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
65 * config/i386/i386.c: Include wide-int.h.
66 (ix86_data_alignment): Use wide-int interfaces.
67 (ix86_local_alignment): Likewise.
68 (ix86_emit_swsqrtsf): Update real_from_integer.
69 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
70 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
71 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
72 (zero_constant): Likewise.
73 (input_operand): Likewise.
74 (splat_input_operand): Likewise.
75 (non_logical_cint_operand): Change const_double to const_wide_int.
76 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
77 (easy_altivec_constant): Remove comment.
78 (paired_expand_vector_init): Use CONSTANT_P.
79 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
80 (rs6000_emit_move): Update checks.
81 (rs6000_aggregate_candidate): Use wide-int interfaces.
82 (rs6000_expand_ternop_builtin): Likewise.
83 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
84 (rs6000_assemble_integer): Likewise.
85 (rs6000_hash_constant): Likewise.
86 (output_toc): Likewise.
87 (rs6000_rtx_costs): Likewise.
88 (rs6000_emit_swrsqrt); Update call to real_from_integer.
89 * config/rs6000/rs6000-c.c: Include wide-int.h.
90 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
91 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
92 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
93 Handle CONST_WIDE_INT.
94 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
96 * config/sparc/sparc.c: Include wide-int.h.
97 (sparc_fold_builtin): Use wide-int interfaces.
98 * config/vax/vax.c: Include wide-int.h.
99 (vax_float_literal): Use real_from_integer.
100 * coretypes.h (struct hwivec_def): New.
103 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
104 (equiv_constant): Handle CONST_WIDE_INT.
105 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
106 (cselib_hash_rtx): Handle CONST_WIDE_INT.
107 * dbxout.c (stabstr_U): Use wide-int interfaces.
108 (dbxout_type): Update to use cst_fits_shwi_p.
109 * defaults.h (LOG2_BITS_PER_UNIT): Define.
110 (TARGET_SUPPORTS_WIDE_INT): Add default.
111 * dfp.c: Include wide-int.h.
112 (decimal_real_to_integer2): Use wide-int interfaces and rename to
113 decimal_real_to_integer.
114 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
115 decimal_real_to_integer.
116 * doc/generic.texi (Constant expressions): Update for wide_int.
117 * doc/rtl.texi (const_double): Likewise.
118 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
119 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
120 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
121 (REAL_VALUE_FROM_INT): Remove.
122 (TARGET_SUPPORTS_WIDE_INT): New.
123 * doc/tm.texi: Regenerate.
124 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
125 * double-int.h: Include wide-int.h.
126 (struct wi::int_traits): New.
127 * dwarf2out.c (get_full_len): New.
128 (dw_val_equal_p): Add case dw_val_class_wide_int.
129 (size_of_loc_descr): Likewise.
130 (output_loc_operands): Likewise.
131 (insert_double): Remove.
132 (insert_wide_int): New.
134 (print_die): Add case dw_val_class_wide_int.
135 (attr_checksum): Likewise.
136 (attr_checksum_ordered): Likewise.
137 (same_dw_val_p): Likewise.
138 (size_of_die): Likewise.
139 (value_format): Likewise.
140 (output_die): Likewise.
141 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
143 (clz_loc_descriptor): Use wide-int interfaces.
144 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
145 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
146 (round_up_to_align): Use wide-int interfaces.
147 (field_byte_offset): Likewise.
148 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
149 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
150 CONST_DOUBLE handling. Use wide-int interfaces.
151 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
152 (gen_enumeration_type_die): Use add_AT_wide.
153 (hash_loc_operands): Add case dw_val_class_wide_int.
154 (compare_loc_operands): Likewise.
155 * dwarf2out.h: Include wide-int.h.
157 (enum dw_val_class): Add dw_val_class_wide_int.
158 (struct dw_val_struct): Add val_wide.
159 * emit-rtl.c (const_wide_int_htab): New.
160 (const_wide_int_htab_hash): New.
161 (const_wide_int_htab_eq): New.
162 (lookup_const_wide_int): New.
163 (const_double_htab_hash): Use wide-int interfaces.
164 (const_double_htab_eq): Likewise.
165 (rtx_to_double_int): Conditionally compile for wide-int.
166 (immed_double_int_const): Rename to immed_wide_int_const and
168 (immed_double_const): Conditionally compile for wide-int.
169 (init_emit_once): Use wide-int interfaces.
170 * explow.c (plus_constant): Likewise.
171 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
172 (lshift_value): Use wide-int interfaces.
173 (expand_mult): Likewise.
174 (choose_multiplier): Likewise.
175 (expand_smod_pow2): Likewise.
176 (make_tree): Likewise.
177 * expr.c (convert_modes): Consolidate handling of constants.
178 Use wide-int interfaces.
179 (emit_group_load_1): Add note.
180 (store_expr): Update comment.
181 (get_inner_reference): Use wide-int interfaces.
182 (expand_constructor): Update comment.
183 (expand_expr_real_2): Use wide-int interfaces.
184 (expand_expr_real_1): Likewise.
185 (reduce_to_bit_field_precision): Likewise.
186 (const_vector_from_tree): Likewise.
187 * final.c: Include wide-int-print.h.
188 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
189 * fixed-value.c: Include wide-int.h.
190 (fixed_from_string): Use wide-int interfaces.
191 (fixed_to_decimal): Likewise.
192 (fixed_convert_from_real): Likewise.
193 (real_convert_from_fixed): Likewise.
194 * fold-const.h (mem_ref_offset): Return an offset_int.
195 (div_if_zero_remainder): Remove code parameter.
196 * fold-const.c (div_if_zero_remainder): Remove code parameter.
197 Use wide-int interfaces.
198 (may_negate_without_overflow_p): Use wide-int interfaces.
199 (negate_expr_p): Likewise.
200 (fold_negate_expr): Likewise.
201 (int_const_binop_1): Likewise.
202 (const_binop): Likewise.
203 (fold_convert_const_int_from_int): Likewise.
204 (fold_convert_const_int_from_real): Likewise.
205 (fold_convert_const_int_from_fixed): Likewise.
206 (fold_convert_const_fixed_from_int): Likewise.
207 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
208 (sign_bit_p): Use wide-int interfaces.
209 (make_range_step): Likewise.
210 (build_range_check): Likewise. Pass an integer of the correct type
211 instead of using integer_one_node.
212 (range_predecessor): Pass an integer of the correct type instead
213 of using integer_one_node.
214 (range_successor): Likewise.
215 (merge_ranges): Likewise.
216 (unextend): Use wide-int interfaces.
217 (extract_muldiv_1): Likewise.
218 (fold_div_compare): Likewise.
219 (fold_single_bit_test): Likewise.
220 (fold_sign_changed_comparison): Likewise.
221 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
222 (fold_plusminus_mult_expr): Use wide-int interfaces.
223 (native_encode_int): Likewise.
224 (native_interpret_int): Likewise.
225 (fold_unary_loc): Likewise.
226 (pointer_may_wrap_p): Likewise.
227 (size_low_cst): Likewise.
228 (mask_with_tz): Likewise.
229 (fold_binary_loc): Likewise.
230 (fold_ternary_loc): Likewise.
231 (multiple_of_p): Likewise.
232 (tree_call_nonnegative_warnv_p): Update calls to
233 tree_int_cst_min_precision and real_from_integer.
234 (fold_negate_const): Use wide-int interfaces.
235 (fold_abs_const): Likewise.
236 (fold_relational_const): Use tree_int_cst_lt.
237 (round_up_loc): Use wide-int interfaces.
238 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
239 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
240 * gengtype.c: Remove include of double-int.h.
241 (do_typedef): Use wide-int interfaces.
242 (open_base_files): Add wide-int.h.
243 (main): Add offset_int and widest_int typedefs.
244 * gengtype-lex.l: Handle "^".
245 (CXX_KEYWORD): Add "static".
246 * gengtype-parse.c (require3): New.
247 (require_template_declaration): Handle constant template arguments
248 and nested templates.
249 * gengtype-state.c: Don't include "double-int.h".
250 * genpreds.c (write_one_predicate_function): Update comment.
251 (write_tm_constrs_h): Add check for hval and lval use in
253 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
254 (add_to_sequence): Likewise.
255 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
256 and const_double_operand.
257 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
259 * gimple-fold.c (get_base_constructor): Likewise.
260 (fold_array_ctor_reference): Likewise.
261 (fold_nonarray_ctor_reference): Likewise.
262 (fold_const_aggregate_ref_1): Likewise.
263 (gimple_val_nonnegative_real_p): Likewise.
264 (gimple_fold_indirect_ref): Likewise.
265 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
266 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
267 (struct slsr_cand_d): Change index to be widest_int.
268 (struct incr_info_d): Change incr to be widest_int.
269 (alloc_cand_and_find_basis): Use wide-int interfaces.
270 (slsr_process_phi): Likewise.
271 (backtrace_base_for_ref): Likewise. Return a widest_int.
272 (restructure_reference): Take a widest_int instead of a double_int.
273 (slsr_process_ref): Use wide-int interfaces.
274 (create_mul_ssa_cand): Likewise.
275 (create_mul_imm_cand): Likewise.
276 (create_add_ssa_cand): Likewise.
277 (create_add_imm_cand): Take a widest_int instead of a double_int.
278 (slsr_process_add): Use wide-int interfaces.
279 (slsr_process_cast): Likewise.
280 (slsr_process_copy): Likewise.
281 (dump_candidate): Likewise.
282 (dump_incr_vec): Likewise.
283 (replace_ref): Likewise.
284 (cand_increment): Likewise. Return a widest_int.
285 (cand_abs_increment): Likewise.
286 (replace_mult_candidate): Take a widest_int instead of a double_int.
287 (replace_unconditional_candidate): Use wide-int interfaces.
288 (incr_vec_index): Take a widest_int instead of a double_int.
289 (create_add_on_incoming_edge): Likewise.
290 (create_phi_basis): Use wide-int interfaces.
291 (replace_conditional_candidate): Likewise.
292 (record_increment): Take a widest_int instead of a double_int.
293 (record_phi_increments): Use wide-int interfaces.
294 (phi_incr_cost): Take a widest_int instead of a double_int.
295 (lowest_cost_path): Likewise.
296 (total_savings): Likewise.
297 (analyze_increments): Use wide-int interfaces.
298 (ncd_with_phi): Take a widest_int instead of a double_int.
299 (ncd_of_cand_and_phis): Likewise.
300 (nearest_common_dominator_for_cands): Likewise.
301 (insert_initializers): Use wide-int interfaces.
302 (all_phi_incrs_profitable): Likewise.
303 (replace_one_candidate): Likewise.
304 (replace_profitable_candidates): Likewise.
305 * godump.c: Include wide-int-print.h.
306 (go_output_typedef): Use wide-int interfaces.
307 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
308 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
309 (build_loop_iteration_domains): Likewise.
310 * hooks.h: Include wide-int.h rather than double-int.h.
311 (hook_bool_dint_dint_uint_bool_true): Delete.
312 (hook_bool_wint_wint_uint_bool_true): Declare.
313 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
314 (hook_bool_wint_wint_uint_bool_true): New.
315 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
317 (ubsan_expand_si_overflow_mul_check): Likewise.
318 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
319 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
320 (get_ancestor_addr_info): Likewise.
321 (ipa_modify_call_arguments): Likewise.
322 * loop-doloop.c (doloop_modify): Likewise.
323 (doloop_optimize): Likewise.
324 * loop-iv.c (iv_number_of_iterations): Likewise.
325 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
326 (unroll_loop_constant_iterations): Likewise.
327 (decide_unroll_runtime_iterations): Likewise.
328 (unroll_loop_runtime_iterations): Likewise.
329 (decide_peel_simple): Likewise.
330 (decide_unroll_stupid): Likewise.
331 * lto-streamer-in.c (streamer_read_wi): Add.
332 (input_cfg): Use wide-int interfaces.
333 (lto_input_tree_1): Likewise.
334 * lto-streamer-out.c (streamer_write_wi): Add.
335 (hash_tree): Use wide-int interfaces.
336 (output_cfg): Likewise.
337 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
338 (GTFILES): Add wide-int.h and signop.h.
339 (TAGS): Look for .cc files too.
340 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
341 * optabs.c (expand_subword_shift): Likewise.
342 (expand_doubleword_shift): Likewise.
343 (expand_absneg_bit): Likewise.
344 (expand_copysign_absneg): Likewise.
345 (expand_copysign_bit): Likewise.
346 * postreload.c (reload_cse_simplify_set): Likewise.
347 * predict.c (predict_iv_comparison): Likewise.
348 * pretty-print.h: Include wide-int-print.h.
350 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
351 * print-tree.c: Include wide-int-print.h.
352 (print_node_brief): Use wide-int interfaces.
353 (print_node): Likewise.
354 * read-rtl.c (validate_const_wide_int): New.
355 (read_rtx_code): Add CONST_WIDE_INT case.
356 * real.c: Include wide-int.h.
357 (real_to_integer2): Delete.
358 (real_to_integer): New function, returning a wide_int.
359 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
360 (ten_to_ptwo): Update call to real_from_integer.
361 (real_digit): Likewise.
362 * real.h: Include signop.h, wide-int.h and insn-modes.h.
363 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
364 (REAL_VALUE_TO_INT): Delete.
365 (real_to_integer): Declare a wide-int form.
366 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
367 * recog.c (const_int_operand): Improve comment.
368 (const_scalar_int_operand): New.
369 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
370 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
371 (split_double): Likewise.
372 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
373 (rtx_size): Likewise.
374 (rtx_alloc_stat_v): New.
375 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
376 (cwi_output_hex): New.
377 (iterative_hash_rtx): Handle CONST_WIDE_INT.
378 (cwi_check_failed_bounds): New.
379 * rtl.def (CONST_WIDE_INT): New.
380 * rtl.h: Include <utility> and wide-int.h.
381 (struct hwivec_def): New.
382 (CWI_GET_NUM_ELEM): New.
383 (CWI_PUT_NUM_ELEM): New.
384 (struct rtx_def): Add num_elem and hwiv.
385 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
386 (CASE_CONST_UNIQUE): Likewise.
387 (CASE_CONST_ANY): Likewise.
388 (CONST_SCALAR_INT_P): Likewise.
389 (CONST_WIDE_INT_P): New.
392 (cwi_check_failed_bounds): New.
395 (CONST_WIDE_INT_VEC) New.
396 (CONST_WIDE_INT_NUNITS) New.
397 (CONST_WIDE_INT_ELT) New.
398 (rtx_mode_t): New type.
399 (wi::int_traits <rtx_mode_t>): New.
401 (wi::min_value): New.
402 (wi::max_value): New.
404 (const_wide_int_alloc): New.
405 (immed_wide_int_const): New.
406 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
407 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
408 * signop.h: New file.
409 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
410 (simplify_const_unary_operation): Use wide-int interfaces.
411 (simplify_binary_operation_1): Likewise.
412 (simplify_const_binary_operation): Likewise.
413 (simplify_const_relational_operation): Likewise.
414 (simplify_immed_subreg): Likewise.
415 * stmt.c (expand_case): Likewise.
416 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
417 signop rather than a bool.
418 * stor-layout.c (layout_type): Use wide-int interfaces.
419 (initialize_sizetypes): Update calls to
420 set_min_and_max_values_for_integral_type.
421 (set_min_and_max_values_for_integral_type): Take a signop rather
422 than a bool. Use wide-int interfaces.
423 (fixup_signed_type): Update accordingly. Remove
424 HOST_BITS_PER_DOUBLE_INT limit.
425 (fixup_unsigned_type): Likewise.
426 * system.h (STATIC_CONSTANT_P): New.
427 (STATIC_ASSERT): New.
428 * target.def (can_use_doloop_p): Take widest_ints rather than
430 * target.h: Include wide-int.h rather than double-int.h.
431 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
433 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
434 rather than INT_CST_LT_UNSIGNED.
435 (can_use_doloop_if_innermost): Take widest_ints rather than
437 * tree-affine.c: Include wide-int-print.h.
438 (double_int_ext_for_comb): Delete.
439 (wide_int_ext_for_comb): New.
440 (aff_combination_zero): Use wide-int interfaces.
441 (aff_combination_const): Take a widest_int instead of a double_int.
442 (aff_combination_elt): Use wide-int interfaces.
443 (aff_combination_scale): Take a widest_int instead of a double_int.
444 (aff_combination_add_elt): Likewise.
445 (aff_combination_add_cst): Likewise.
446 (aff_combination_add): Use wide-int interfaces.
447 (aff_combination_convert): Likewise.
448 (tree_to_aff_combination): Likewise.
449 (add_elt_to_tree): Take a widest_int instead of a double_int.
450 (aff_combination_to_tree): Use wide-int interfaces.
451 (aff_combination_remove_elt): Likewise.
452 (aff_combination_add_product): Take a widest_int instead of
454 (aff_combination_mult): Use wide-int interfaces.
455 (aff_combination_expand): Likewise.
456 (double_int_constant_multiple_p): Delete.
457 (wide_int_constant_multiple_p): New.
458 (aff_combination_constant_multiple_p): Take a widest_int pointer
459 instead of a double_int pointer.
460 (print_aff): Use wide-int interfaces.
461 (get_inner_reference_aff): Take a widest_int pointer
462 instead of a double_int pointer.
463 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
464 * tree-affine.h: Include wide-int.h.
465 (struct aff_comb_elt): Change type of coef to widest_int.
466 (struct affine_tree_combination): Change type of offset to widest_int.
467 (double_int_ext_for_comb): Delete.
468 (wide_int_ext_for_comb): New.
469 (aff_combination_const): Use widest_int instead of double_int.
470 (aff_combination_scale): Likewise.
471 (aff_combination_add_elt): Likewise.
472 (aff_combination_constant_multiple_p): Likewise.
473 (get_inner_reference_aff): Likewise.
474 (aff_comb_cannot_overlap_p): Likewise.
475 (aff_combination_zero_p): Use wide-int interfaces.
476 * tree.c: Include tree.h.
477 (init_ttree): Use make_int_cst.
478 (tree_code_size): Removed code for INTEGER_CST case.
479 (tree_size): Add INTEGER_CST case.
480 (make_node_stat): Update comment.
481 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
482 (build_int_cst_type): Use wide-int interfaces.
483 (double_int_to_tree): Likewise.
484 (double_int_fits_to_tree_p): Delete.
485 (force_fit_type_double): Delete.
486 (force_fit_type): New.
487 (int_cst_hash_hash): Use wide-int interfaces.
488 (int_cst_hash_eq): Likewise.
489 (build_int_cst_wide): Delete.
490 (wide_int_to_tree): New.
491 (cache_integer_cst): Use wide-int interfaces.
492 (build_low_bits_mask): Likewise.
493 (cst_and_fits_in_hwi): Likewise.
494 (real_value_from_int_cst): Likewise.
495 (make_int_cst_stat): New.
496 (integer_zerop): Use wide_int interfaces.
497 (integer_onep): Likewise.
498 (integer_all_onesp): Likewise.
499 (integer_pow2p): Likewise.
500 (integer_nonzerop): Likewise.
501 (tree_log2): Likewise.
502 (tree_floor_log2): Likewise.
503 (tree_ctz): Likewise.
504 (int_size_in_bytes): Likewise.
505 (mem_ref_offset): Return an offset_int rather than a double_int.
506 (build_type_attribute_qual_variant): Use wide_int interfaces.
507 (type_hash_eq): Likewise
508 (tree_int_cst_equal): Likewise.
509 (tree_int_cst_lt): Delete.
510 (tree_int_cst_compare): Likewise.
511 (tree_fits_shwi_p): Use wide_int interfaces.
512 (tree_fits_uhwi_p): Likewise.
513 (tree_int_cst_sign_bit): Likewise.
514 (tree_int_cst_sgn): Likewise.
515 (tree_int_cst_min_precision): Take a signop rather than a bool.
516 (simple_cst_equal): Use wide_int interfaces.
517 (compare_tree_int): Likewise.
518 (iterative_hash_expr): Likewise.
519 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
521 (get_type_static_bounds): Use wide_int interfaces.
522 (tree_int_cst_elt_check_failed): New.
523 (build_common_tree_nodes): Reordered to set prec before filling in
525 (int_cst_value): Check cst_and_fits_in_hwi.
526 (widest_int_cst_value): Use wide_int interfaces.
527 (upper_bound_in_type): Likewise.
528 (lower_bound_in_type): Likewise.
529 (num_ending_zeros): Likewise.
530 (drop_tree_overflow): Likewise.
531 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
532 (gen_conditions_for_pow_cst_base): Likewise.
533 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
534 (group_case_labels_stmt): Use wide-int interfaces.
535 (verify_gimple_assign_binary): Likewise.
536 (print_loop): Likewise.
537 * tree-chrec.c (tree_fold_binomial): Likewise.
538 * tree-core.h (struct tree_base): Add int_length.
539 (struct tree_int_cst): Change rep of value.
540 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
541 (dr_may_alias_p): Likewise.
542 (max_stmt_executions_tree): Likewise.
543 * tree.def (INTEGER_CST): Update comment.
544 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
545 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
546 * tree-dump.c: Include wide-int.h and wide-int-print.h.
547 (dequeue_and_dump): Use wide-int interfaces.
548 * tree.h: Include wide-int.h.
549 (NULL_TREE): Moved to earlier loc in file.
550 (TREE_INT_CST_ELT_CHECK): New.
551 (tree_int_cst_elt_check_failed): New.
553 (TREE_INT_CST): Delete.
554 (TREE_INT_CST_LOW): Use wide-int interfaces.
555 (TREE_INT_CST_HIGH): Delete.
556 (TREE_INT_CST_NUNITS): New.
557 (TREE_INT_CST_EXT_NUNITS): Likewise.
558 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
559 (TREE_INT_CST_ELT): Likewise.
560 (INT_CST_LT): Delete.
561 (tree_int_cst_elt_check): New (two forms).
562 (type_code_size): Update comment.
563 (make_int_cst_stat, make_int_cst): New.
564 (tree_to_double_int): Delete.
565 (double_int_fits_to_tree_p): Delete.
566 (force_fit_type_double): Delete.
567 (build_int_cstu): Replace with out-of-line function.
568 (build_int_cst_wide): Delete.
569 (tree_int_cst_lt): Define inline.
570 (tree_int_cst_le): New.
571 (tree_int_cst_compare): Define inline.
572 (tree_int_cst_min_precision): Take a signop rather than a bool.
573 (wi::int_traits <const_tree>): New.
574 (wi::int_traits <tree>): New.
575 (wi::extended_tree): New.
576 (wi::int_traits <wi::extended_tree>): New.
577 (wi::to_widest): New.
578 (wi::to_offset): New.
579 (wi::fits_to_tree_p): New.
580 (wi::min_value): New.
581 (wi::max_value): New.
582 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
583 (copy_tree_body_r): Likewise.
584 * tree-object-size.c (compute_object_offset): Likewise.
585 (addr_object_size): Likewise.
586 * tree-predcom.c: Include wide-int-print.h.
587 (struct dref_d): Change type of offset to widest_int.
588 (dump_dref): Call wide-int printer.
589 (aff_combination_dr_offset): Use wide-int interfaces.
590 (determine_offset): Take a widest_int pointer rather than a
592 (split_data_refs_to_components): Use wide-int interfaces.
593 (suitable_component_p): Likewise.
594 (order_drefs): Likewise.
595 (add_ref_to_chain): Likewise.
596 (valid_initializer_p): Likewise.
597 (determine_roots_comp): Likewise.
598 * tree-pretty-print.c: Include wide-int-print.h.
599 (dump_generic_node): Use wide-int interfaces.
600 * tree-sra.c (sra_ipa_modify_expr): Likewise.
601 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
602 (move_fixed_address_to_symbol): Likewise.
603 (move_hint_to_base): Likewise.
604 (move_pointer_to_base): Likewise.
605 (move_variant_to_index): Likewise.
606 (most_expensive_mult_to_index): Likewise.
607 (addr_to_parts): Likewise.
608 (copy_ref_info): Likewise.
609 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
610 (indirect_refs_may_alias_p): Likewise.
611 (stmt_kills_ref_p_1): Likewise.
612 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
613 * tree-ssa-ccp.c: Update comment at top of file. Include
615 (struct prop_value_d): Change type of mask to widest_int.
616 (extend_mask): New function.
617 (dump_lattice_value): Use wide-int interfaces.
618 (get_default_value): Likewise.
619 (set_constant_value): Likewise.
620 (set_value_varying): Likewise.
621 (valid_lattice_transition): Likewise.
622 (set_lattice_value): Likewise.
623 (value_to_double_int): Delete.
624 (value_to_wide_int): New.
625 (get_value_from_alignment): Use wide-int interfaces.
626 (get_value_for_expr): Likewise.
627 (do_dbg_cnt): Likewise.
628 (ccp_finalize): Likewise.
629 (ccp_lattice_meet): Likewise.
630 (bit_value_unop_1): Use widest_ints rather than double_ints.
631 (bit_value_binop_1): Likewise.
632 (bit_value_unop): Use wide-int interfaces.
633 (bit_value_binop): Likewise.
634 (bit_value_assume_aligned): Likewise.
635 (evaluate_stmt): Likewise.
636 (ccp_fold_stmt): Likewise.
637 (visit_cond_stmt): Likewise.
638 (ccp_visit_stmt): Likewise.
639 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
640 (constant_pointer_difference): Likewise.
641 (associate_pointerplus): Likewise.
642 (combine_conversions): Likewise.
643 * tree-ssa-loop.h: Include wide-int.h.
644 (struct tree_niter_desc): Change type of max to widest_int.
645 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
646 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
647 (remove_redundant_iv_tests): Likewise.
648 (canonicalize_loop_induction_variables): Likewise.
649 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
650 (constant_multiple_of): Take a widest_int pointer instead of
651 a double_int pointer.
652 (get_computation_aff): Use wide-int interfaces.
653 (ptr_difference_cost): Likewise.
654 (difference_cost): Likewise.
655 (get_loop_invariant_expr_id): Likewise.
656 (get_computation_cost_at): Likewise.
657 (iv_elimination_compare_lt): Likewise.
658 (may_eliminate_iv): Likewise.
659 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
660 instead of double_int.
661 (max_loop_iterations): Likewise.
662 (max_stmt_executions): Likewise.
663 (estimated_stmt_executions): Likewise.
664 * tree-ssa-loop-niter.c: Include wide-int-print.h.
665 (split_to_var_and_offset): Use wide-int interfaces.
666 (determine_value_range): Likewise.
667 (bound_difference_of_offsetted_base): Likewise.
668 (bounds_add): Take a widest_int instead of a double_int.
669 (number_of_iterations_ne_max): Use wide-int interfaces.
670 (number_of_iterations_ne): Likewise.
671 (number_of_iterations_lt_to_ne): Likewise.
672 (assert_loop_rolls_lt): Likewise.
673 (number_of_iterations_lt): Likewise.
674 (number_of_iterations_le): Likewise.
675 (number_of_iterations_cond): Likewise.
676 (number_of_iterations_exit): Likewise.
677 (finite_loop_p): Likewise.
678 (derive_constant_upper_bound_assign): Likewise.
679 (derive_constant_upper_bound): Return a widest_int.
680 (derive_constant_upper_bound_ops): Likewise.
681 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
682 (record_estimate): Take a widest_int rather than a double_int.
683 (record_nonwrapping_iv): Use wide-int interfaces.
684 (double_int_cmp): Delete.
686 (bound_index): Take a widest_int rather than a double_int.
687 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
688 (maybe_lower_iteration_bound): Likewise.
689 (estimate_numbers_of_iterations_loop): Likewise.
690 (estimated_loop_iterations): Take a widest_int pointer than than
691 a double_int pointer.
692 (estimated_loop_iterations_int): Use wide-int interfaces.
693 (max_loop_iterations): Take a widest_int pointer than than
694 a double_int pointer.
695 (max_loop_iterations_int): Use wide-int interfaces.
696 (max_stmt_executions): Take a widest_int pointer than than
697 a double_int pointer.
698 (estimated_stmt_executions): Likewise.
699 (n_of_executions_at_most): Use wide-int interfaces.
700 (scev_probably_wraps_p): Likewise.
701 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
703 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
705 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
706 double_ints. Adjust for trailing_wide_ints <3> representation.
707 (set_nonzero_bits): Likewise.
708 (get_range_info): Return wide_ints rather than double_ints.
709 Adjust for trailing_wide_ints <3> representation.
710 (get_nonzero_bits): Likewise.
711 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
713 * tree-ssanames.h (struct range_info_def): Replace min, max and
714 nonzero_bits with a trailing_wide_ints <3>.
715 (set_range_info): Use wide_int_refs rather than double_ints.
716 (set_nonzero_bits): Likewise.
717 (get_range_info): Return wide_ints rather than double_ints.
718 (get_nonzero_bits): Likewise.
719 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
720 * tree-ssa-pre.c (phi_translate_1): Likewise.
721 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
722 (acceptable_pow_call): Likewise.
723 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
725 (vn_reference_fold_indirect): Likewise.
726 (vn_reference_maybe_forwprop_address): Likewise.
727 (valueize_refs_1): Likewise.
728 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
729 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
730 tree_int_cst_lt and tree_int_cst_le.
731 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
733 (streamer_alloc_tree): Likewise.
734 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
735 (streamer_write_tree_header): Likewise.
736 (streamer_write_integer_cst): Likewise.
737 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
738 (build_constructors): Likewise.
739 (array_value_type): Likewise.
740 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
741 (vect_check_gather): Likewise.
742 * tree-vect-generic.c (build_replicated_const): Likewise.
743 (expand_vector_divmod): Likewise.
744 * tree-vect-loop.c (vect_transform_loop): Likewise.
745 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
746 (vect_do_peeling_for_alignment): Likewise.
747 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
748 * tree-vrp.c: Include wide-int.h.
749 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
750 (extract_range_from_assert): Use wide-int interfaces.
751 (vrp_int_const_binop): Likewise.
752 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
754 (ranges_from_anti_range): Use wide-int interfaces.
755 (quad_int_cmp): Delete.
756 (quad_int_pair_sort): Likewise.
757 (extract_range_from_binary_expr_1): Use wide-int interfaces.
758 (extract_range_from_unary_expr_1): Likewise.
759 (adjust_range_with_scev): Likewise.
760 (masked_increment): Take and return wide_ints rather than double_ints.
761 (register_edge_assert_for_2): Use wide-int interfaces.
762 (check_array_ref): Likewise.
763 (search_for_addr_array): Likewise.
764 (maybe_set_nonzero_bits): Likewise.
765 (union_ranges): Pass an integer of the correct type instead of
766 using integer_one_node.
767 (intersect_ranges): Likewise.
768 (simplify_truth_ops_using_ranges): Likewise.
769 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
770 (range_fits_type_p): Likewise.
771 (simplify_cond_using_ranges): Likewise. Take a signop rather than
773 (simplify_conversion_using_ranges): Use wide-int interfaces.
774 (simplify_float_conversion_using_ranges): Likewise.
775 (vrp_finalize): Likewise.
776 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
777 (gimple_stringops_transform): Likewise.
778 * varasm.c (decode_addr_const): Likewise.
779 (const_hash_1): Likewise.
780 (const_rtx_hash_1): Likewise
781 (output_constant): Likewise.
782 (array_size_for_constructor): Likewise.
783 (output_constructor_regular_field): Likewise.
784 (output_constructor_bitfield): Likewise.
785 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
786 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
788 * gencheck.c: Define BITS_PER_UNIT.
791 * wide-int-print.cc: New.
792 * wide-int-print.h: New.
794 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
796 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
798 2014-05-06 Richard Biener <rguenther@suse.de>
800 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
801 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
802 (TODO_verify_all): Adjust.
803 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
804 TODO_verify_stmts and TODO_verify_rtl_sharing.
805 * bb-reorder.c: Likewise.
806 * cfgexpand.c: Likewise.
809 * function.c: Likewise.
810 * fwprop.c: Likewise.
812 * gimple-ssa-isolate-paths.c: Likewise.
813 * gimple-ssa-strength-reduction.c: Likewise.
814 * ipa-split.c: Likewise.
815 * loop-init.c: Likewise.
816 * loop-unroll.c: Likewise.
817 * lower-subreg.c: Likewise.
818 * modulo-sched.c: Likewise.
819 * postreload-gcse.c: Likewise.
820 * predict.c: Likewise.
822 * sched-rgn.c: Likewise.
823 * store-motion.c: Likewise.
824 * tracer.c: Likewise.
825 * trans-mem.c: Likewise.
826 * tree-call-cdce.c: Likewise.
827 * tree-cfg.c: Likewise.
828 * tree-cfgcleanup.c: Likewise.
829 * tree-complex.c: Likewise.
830 * tree-eh.c: Likewise.
831 * tree-emutls.c: Likewise.
832 * tree-if-conv.c: Likewise.
833 * tree-into-ssa.c: Likewise.
834 * tree-loop-distribution.c: Likewise.
835 * tree-object-size.c: Likewise.
836 * tree-parloops.c: Likewise.
837 * tree-pass.h: Likewise.
838 * tree-sra.c: Likewise.
839 * tree-ssa-ccp.c: Likewise.
840 * tree-ssa-copy.c: Likewise.
841 * tree-ssa-copyrename.c: Likewise.
842 * tree-ssa-dce.c: Likewise.
843 * tree-ssa-dom.c: Likewise.
844 * tree-ssa-dse.c: Likewise.
845 * tree-ssa-forwprop.c: Likewise.
846 * tree-ssa-ifcombine.c: Likewise.
847 * tree-ssa-loop-ch.c: Likewise.
848 * tree-ssa-loop-ivcanon.c: Likewise.
849 * tree-ssa-loop.c: Likewise.
850 * tree-ssa-math-opts.c: Likewise.
851 * tree-ssa-phiopt.c: Likewise.
852 * tree-ssa-phiprop.c: Likewise.
853 * tree-ssa-pre.c: Likewise.
854 * tree-ssa-reassoc.c: Likewise.
855 * tree-ssa-sink.c: Likewise.
856 * tree-ssa-strlen.c: Likewise.
857 * tree-ssa-tail-merge.c: Likewise.
858 * tree-ssa-uncprop.c: Likewise.
859 * tree-switch-conversion.c: Likewise.
860 * tree-tailcall.c: Likewise.
861 * tree-vect-generic.c: Likewise.
862 * tree-vectorizer.c: Likewise.
863 * tree-vrp.c: Likewise.
865 * var-tracking.c: Likewise.
866 * bt-load.c: Likewise.
867 * cfgcleanup.c: Likewise.
868 * combine-stack-adj.c: Likewise.
869 * combine.c: Likewise.
870 * compare-elim.c: Likewise.
871 * config/epiphany/resolve-sw-modes.c: Likewise.
872 * config/i386/i386.c: Likewise.
873 * config/mips/mips.c: Likewise.
874 * config/s390/s390.c: Likewise.
875 * config/sh/sh_treg_combine.cc: Likewise.
876 * config/sparc/sparc.c: Likewise.
881 * mode-switching.c: Likewise.
882 * passes.c: Likewise.
883 * postreload.c: Likewise.
885 * reg-stack.c: Likewise.
886 * regcprop.c: Likewise.
887 * regrename.c: Likewise.
890 2014-05-06 Richard Biener <rguenther@suse.de>
893 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
894 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
896 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
899 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
901 2014-05-05 Radovan Obradovic <robradovic@mips.com>
902 Tom de Vries <tom@codesourcery.com>
904 * target.def (call_fusage_contains_non_callee_clobbers): New DEFHOOKPOD.
905 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
907 (@node Miscellaneous Register Hooks): New node.
908 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
909 * doc/tm.texi: Regenerate.
911 2014-05-05 Marek Polacek <polacek@redhat.com>
914 * opts.c (common_handle_option): Call error_at instead of warning_at.
916 2014-05-05 Richard Biener <rguenther@suse.de>
918 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
919 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
920 under the TODO_verify_il umbrella.
922 2014-05-05 Richard Biener <rguenther@suse.de>
924 * passes.c (execute_function_todo): Move TODO_verify_flow under
925 the TODO_verify_ul umbrella.
927 2014-05-05 Richard Biener <rguenther@suse.de>
930 * fold-const.c (fold_binary_loc): Consistently avoid
931 canonicalizing X & CST away from a CST that is the mask
934 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
936 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
937 int argument to enum machine_mode.
938 (picochip_class_max_nregs): Ditto.
939 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
940 (picochip_class_max_nregs): Ditto.
942 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
944 * target.def: Add new target hook.
945 * doc/tm.texi: Regenerate.
946 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
947 * targhooks.c (default_keep_leaf_when_profiled): New function.
949 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
950 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
952 2014-05-05 Bin Cheng <bin.cheng@arm.com>
954 PR tree-optimization/60363
955 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
956 (copy_phi_args): New parameters. Call get_value_locus_in_path.
957 (update_destination_phis): New parameter.
958 (create_edge_and_update_destination_phis): Ditto.
959 (ssa_fix_duplicate_block_edges): Pass new arguments.
960 (thread_single_edge): Ditto.
962 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
964 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
965 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
966 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
967 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
968 Use RS6000_BTM_HARD_FLOAT.
969 (BU_MISC_2): Likewise.
970 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
971 RS6000_BTM_HARD_FLOAT.
972 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
974 (rs6000_invalid_builtin): Add hard floating builtin support.
975 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
977 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
979 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
981 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
982 Add missing function* argument.
984 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
986 * lra-constraints.c (valid_address_p): Move earlier in file.
987 Add a constraint argument to the address_info version.
988 (satisfies_memory_constraint_p): New function.
989 (satisfies_address_constraint_p): Likewise.
990 (process_alt_operands, curr_insn_transform): Use them.
991 (process_address): Pass the constraint to valid_address_p when
992 checking address operands.
994 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
996 * config/mips/mips.c (mips_isa_rev): New variable.
997 (mips_set_architecture): Set it.
998 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1000 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1001 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1002 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1003 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1004 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
1005 conditions in terms of mips_isa_rev.
1006 (mips_isa_rev): Declare.
1008 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1010 * config/sh/sh-mem.cc: Use tabs instead of spaces.
1011 (prob_unlikely, prob_likely): Make variables const.
1013 2014-05-03 Denis Chertykov <chertykov@gmail.com>
1015 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
1017 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1019 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
1021 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1023 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
1024 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
1025 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
1027 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
1028 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
1030 Replace usage of ROUND_REG with sh_round_reg.
1031 Use CEIL instead of ROUND_ADVANCE.
1033 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1036 * config/sh/sh.c: Include stdlib headers before everything else.
1038 2014-05-02 Jakub Jelinek <jakub@redhat.com>
1040 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
1041 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
1042 (gimplify_adjust_omp_clauses): Simd region is never
1043 directly nested in combined parallel. Instead, for linear
1044 with copyin/copyout, if in combined for simd loop, make decl
1045 firstprivate/lastprivate on OMP_FOR.
1046 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
1047 expand_omp_for_static_chunk): When setting endvar, also set
1048 fd->loop.v to the same value.
1050 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1052 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
1054 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
1056 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
1059 2014-05-02 Marek Polacek <polacek@redhat.com>
1061 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
1063 2014-02-26 Kito Cheng <kito@0xlab.org>
1065 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
1066 to a C expression marco.
1067 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
1068 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
1069 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
1070 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
1071 HONOR_REG_ALLOC_ORDER.
1072 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
1074 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1076 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
1078 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1080 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
1082 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
1084 * tree-if-conv.c (is_cond_scalar_reduction): New function.
1085 (convert_scalar_cond_reduction): Likewise.
1086 (predicate_scalar_phi): Add recognition and transformation
1087 of simple conditioanl reduction to be vectorizable.
1089 2014-05-01 Marek Polacek <polacek@redhat.com>
1092 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
1094 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
1096 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
1097 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
1098 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
1099 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
1100 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
1101 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
1102 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
1103 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
1105 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
1107 * config/arc/arc.opt (mlra): Move comment above option name
1108 to avoid mis-parsing as language options.
1110 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1112 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
1113 * config/sol2.h: ... here.
1114 * config/sol2-10.h: Remove.
1116 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
1117 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
1118 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
1119 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
1120 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
1121 * config/sol2.h: ... here.
1122 (SECTION_NAME_FORMAT): Don't redefine.
1123 (STARTFILE_ARCH32_SPEC): Rename to ...
1124 (STARTFILE_ARCH_SPEC): ... this.
1125 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
1126 * config/sparc/sol2.h: ... here.
1127 (SECTION_NAME_FORMAT): Don't undef.
1128 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
1129 (SUBTARGET_EXTRA_SPECS): Remove.
1130 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
1132 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
1133 (MD_STARTFILE_PREFIX): Remove.
1134 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
1135 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
1136 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
1137 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
1138 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
1139 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
1140 * config/i386/sol2.h: ... here.
1141 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
1142 * config/i386/sol2-bi.h: Remove.
1143 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
1144 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
1146 * config/i386/t-sol2-64: Rename to ...
1147 * config/i386/t-sol2: ... this.
1148 * config/sparc/t-sol2-64: Rename to ...
1149 * config/sparc/t-sol2: ... this.
1151 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
1152 sol2_tm_file_head, sol2_tm_file_tail.
1153 Include ${cpu_type}/sol2.h before sol2.h.
1155 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
1156 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
1157 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
1158 Reflect i386/t-sol2-64 renaming.
1159 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
1160 Reflect sparc/t-sol2-64 renaming.
1162 2014-04-30 Richard Biener <rguenther@suse.de>
1164 * passes.c (execute_function_todo): Move TODO_verify_stmts
1165 and TODO_verify_ssa under the TODO_verify_il umbrella.
1166 * tree-ssa.h (verify_ssa): Adjust prototype.
1167 * tree-ssa.c (verify_ssa): Add parameter to tell whether
1168 we should verify SSA operands.
1169 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
1170 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
1171 whether we should verify whether not throwing stmts have EH info.
1172 * graphite-scop-detection.c (create_sese_edges): Adjust.
1173 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
1174 * tree-eh.c (lower_try_finally_switch): Do not add the
1175 default case label twice.
1177 2014-04-30 Marek Polacek <polacek@redhat.com>
1179 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
1180 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
1181 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
1182 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
1184 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
1186 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
1187 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
1188 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
1189 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
1190 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
1191 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
1192 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
1193 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
1195 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1197 * tree-cfg.c (dump_function_to_file): Dump the return type of
1198 functions, in a line to itself before the function body, mimicking
1199 the layout of a C function.
1201 2014-04-29 Jakub Jelinek <jakub@redhat.com>
1203 PR tree-optimization/60971
1204 * tree-tailcall.c (process_assignment): Reject conversions which
1207 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
1209 * calls.c (initialize_argument_information): Always treat
1210 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
1211 (expand_call): Likewise.
1212 (emit_library_call_calue_1): Likewise.
1213 * expr.c (PUSH_ARGS_REVERSED): Do not define.
1214 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
1217 2014-04-29 Nick Clifton <nickc@redhat.com>
1219 * config/msp430/msp430.md (umulsidi): Fix typo.
1220 (mulhisi3): Enable even inside interrupt handlers.
1221 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
1222 bigger return address pushed in large mode.
1224 2014-04-29 Nick Clifton <nickc@redhat.com>
1226 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
1227 (arc_init_reg_tables): Use a machine_mode enum to iterate over
1229 * config/m32r/m32r.c (init_reg_tables): Likewise.
1230 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
1231 enum to hold the modes.
1233 2014-04-29 Richard Biener <rguenther@suse.de>
1235 * dominance.c (free_dominance_info): Add overload with
1237 (dom_info_state): Likewise.
1238 (dom_info_available_p): Likewise.
1239 * basic-block.h (free_dominance_info, dom_info_state,
1240 dom_info_available_p): Declare overloads.
1241 * passes.c (execute_function_todo): Verify that verifiers
1242 don't change dominator info state. Drop dominator info
1243 for IPA pass invocations.
1244 * cgraph.c (release_function_body): Restore asserts that
1245 dominator information is released.
1247 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
1249 * doc/invoke.texi: Fix typo.
1250 * tree-vrp.c: Fix typos.
1251 * gimple.c (infer_nonnull_range): Reorder operands of an &&
1254 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1256 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
1258 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1260 * config/aarch64/aarch64-builtins.c
1261 (aarch64_types_storestruct_lane_qualifiers): New.
1262 (TYPES_STORESTRUCT_LANE): Likewise.
1263 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
1264 (st3_lane): Likewise.
1265 (st4_lane): Likewise.
1266 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
1267 (vec_store_lanesci_lane<mode>): Likewise.
1268 (vec_store_lanesxi_lane<mode>): Likewise.
1269 (aarch64_st2_lane<VQ:mode>): Likewise.
1270 (aarch64_st3_lane<VQ:mode>): Likewise.
1271 (aarch64_st4_lane<VQ:mode>): Likewise.
1272 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
1273 * config/aarch64/arm_neon.h
1274 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
1275 use new macro arguments.
1276 (__ST3_LANE_FUNC): Likewise.
1277 (__ST4_LANE_FUNC): Likewise.
1278 * config/aarch64/iterators.md (V_TWO_ELEM): New.
1279 (V_THREE_ELEM): Likewise.
1280 (V_FOUR_ELEM): Likewise.
1282 2014-04-28 David Malcolm <dmalcolm@redhat.com>
1284 * doc/gimple.texi: Replace the description of the now-defunct
1285 union gimple_statement_d with a diagram showing the
1286 gimple_statement_base class hierarchy and its relationships to
1287 the GSS_ and GIMPLE_ enums.
1289 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1291 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
1292 * config/aarch64/aarch64.c
1293 (aarch64_cannot_change_mode_class): Weaken conditions.
1294 (aarch64_modes_tieable_p): New.
1295 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
1297 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
1299 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
1300 (loadsync_<mode>): Change mode.
1301 (load_quadpti, store_quadpti): New.
1302 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
1303 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
1305 2014-04-28 Martin Jambor <mjambor@suse.cz>
1307 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
1308 same alias type as the original statement.
1309 (subreplacement_assignment_data): New type.
1310 (handle_unscalarized_data_in_subtree): New type of parameter,
1311 generate new memory accesses with same alias type as the original
1313 (load_assign_lhs_subreplacements): Likewise.
1314 (sra_modify_constructor_assign): Generate new memory accesses with
1315 same alias type as the original statement.
1317 2014-04-28 Richard Biener <rguenther@suse.de>
1319 * tree-pass.h (TODO_verify_il): Define.
1320 (TODO_verify_all): Complete properly.
1321 * passes.c (execute_function_todo): Move existing loop-closed
1322 SSA verification under TODO_verify_il.
1323 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
1324 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
1325 Fix tree sharing issue.
1327 2014-04-28 Richard Biener <rguenther@suse.de>
1330 * builtins.def (DEF_C11_BUILTIN): Add.
1331 (BUILT_IN_ALIGNED_ALLOC): Likewise.
1332 * coretypes.h (enum function_class): Add function_c11_misc.
1333 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1334 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
1335 (call_may_clobber_ref_p_1): Likewise.
1336 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
1337 (mark_all_reaching_defs_necessary_1): Likewise.
1338 (propagate_necessity): Likewise.
1339 (eliminate_unnecessary_stmts): Likewise.
1340 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
1342 2014-04-28 Richard Biener <rguenther@suse.de>
1344 * tree-vrp.c (vrp_var_may_overflow): Remove.
1345 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
1346 with overflow immediately bump to one before that value and
1347 let iteration figure out overflow status.
1349 2014-04-28 Richard Biener <rguenther@suse.de>
1351 * configure.ac: Do valgrind header checks unconditionally.
1352 Add --enable-valgrind-annotations.
1353 * system.h: Guard valgrind header inclusion with
1354 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
1355 * alloc-pool.c (pool_alloc, pool_free): Use
1356 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
1357 to guard possibly dead code.
1358 * config.in: Regenerated.
1359 * configure: Likewise.
1361 2014-04-28 Jeff Law <law@redhat.com>
1363 PR tree-optimization/60902
1364 * tree-ssa-threadedge.c
1365 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
1366 over real defs when invalidating outputs from statements that do not
1367 produce useful outputs for threading.
1369 2014-04-28 Richard Biener <rguenther@suse.de>
1371 PR tree-optimization/60979
1372 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
1373 SCOPs that end in a block with a successor with abnormal
1376 2014-04-28 Richard Biener <rguenther@suse.de>
1378 * tree-pass.h (execute_pass_list): Adjust prototype.
1379 * passes.c (pass_manager::execute_early_local_passes):
1381 (do_per_function): Change callback signature, push all actual
1382 work to the callbals.
1383 (do_per_function_toporder): Likewise.
1384 (execute_function_dump): Adjust.
1385 (execute_function_todo): Likewise.
1386 (clear_last_verified): Likewise.
1387 (verify_curr_properties): Likewise.
1388 (update_properties_after_pass): Likewise.
1389 (execute_pass_list_1): Split out from ...
1390 (execute_pass_list): ... here. Adjust.
1391 (execute_ipa_pass_list): Likewise.
1392 * cgraphunit.c (cgraph_add_new_function): Adjust.
1393 (analyze_function): Likewise.
1394 (expand_function): Likewise.
1395 * cgraph.c (release_function_body): Free dominance info
1396 here instead of asserting it was magically freed elsewhere.
1398 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
1400 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
1401 * configure: Regenerate.
1402 * config/sparc/sparc.opt (muser-mode): New option.
1403 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
1405 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
1406 * doc/invoke.texi (SPARC options): Document -muser-mode.
1408 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
1410 * cselib.c (find_slot_memmode): Delete.
1411 (cselib_hasher): Change compare_type to a struct.
1412 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
1414 (preserve_constants_and_equivs): Adjust for new compare_type.
1415 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
1416 (wrap_constant): Delete.
1417 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
1419 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
1421 * doc/install.texi (Building with profile feedback): Remove
1424 2014-04-26 Tom de Vries <tom@codesourcery.com>
1426 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
1429 2014-04-25 Cary Coutant <ccoutant@google.com>
1432 * dwarf2out.c (should_move_die_to_comdat): A type definition
1433 can contain a subprogram definition, but don't move it to a
1435 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
1436 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
1438 (clone_tree_hash): Rename to...
1439 (clone_tree_partial): ...this; change callers. Copy
1440 DW_TAG_subprogram DIEs as declarations.
1441 (copy_decls_walk): Don't copy children of a declaration into a
1444 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
1447 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
1450 2014-04-25 Jiong Wang <jiong.wang@arm.com>
1452 * config/arm/predicates.md (call_insn_operand): Add long_call check.
1453 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
1455 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
1458 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1460 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
1463 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1465 PR tree-optimization/60930
1466 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
1467 creating a multiply candidate by folding two constant
1468 multiplicands when the result overflows.
1470 2014-04-25 Jakub Jelinek <jakub@redhat.com>
1472 PR tree-optimization/60960
1473 * tree-vect-generic.c (expand_vector_operation): Only call
1474 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
1476 2014-04-25 Tom de Vries <tom@codesourcery.com>
1478 * expr.c (clobber_reg_mode): New function.
1479 * expr.h (clobber_reg): New function.
1481 2014-04-25 Tom de Vries <tom@codesourcery.com>
1483 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
1486 2014-04-25 Radovan Obradovic <robradovic@mips.com>
1487 Tom de Vries <tom@codesourcery.com>
1489 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
1491 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
1492 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
1493 new argument to find_all_hard_reg_sets call.
1495 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1497 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
1498 Use HOST_WIDE_INT_C for mask literal.
1499 (aarch_rev16_shleft_mask_imm_p): Likewise.
1501 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
1504 * config/sparc/sparc.md (ashlsi3_extend): Delete.
1506 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
1508 PR preprocessor/56540
1509 * config/i386/i386-c.c (ix86_target_macros): Define
1510 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
1512 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1514 * configure.ac (tga_func): Remove.
1515 (LIB_TLS_SPEC): Remove.
1516 * configure: Regenerate.
1517 * config.in: Regenerate.
1518 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
1520 2014-04-25 Richard Biener <rguenther@suse.de>
1523 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
1524 call stmt use/clobber sets during stmt walk instead of
1525 walking the possibly incomplete set of caller edges.
1527 2014-04-25 Richard Biener <rguenther@suse.de>
1530 * passes.c (apply_ipa_transforms): Inline into only caller ...
1531 (execute_one_pass): ... here. Properly bring in function
1532 bodies for nodes we want to apply IPA transforms to.
1534 2014-04-24 Cong Hou <congh@google.com>
1536 PR tree-optimization/60896
1537 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
1538 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
1539 (vect_mark_pattern_stmts): Set the def type of all statements in
1540 PATTERN_DEF_SEQ as vect_internal_def.
1542 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
1544 * doc/extend.texi (PowerPC Built-in Functions): Document new
1545 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
1547 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
1549 * config/rs6000/predicates.md (const_0_to_3_operand): New
1550 predicate to match 0..3 integer constants.
1552 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
1553 to support adding miscellaneous builtin functions.
1554 (BU_DFP_MISC_2): Likewise.
1555 (BU_P7_MISC_1): Likewise.
1556 (BU_P7_MISC_2): Likewise.
1557 (BU_P8V_MISC_3): Likewise.
1558 (BU_MISC_1): Likewise.
1559 (BU_MISC_2): Likewise.
1560 (DIVWE): Add extended divide builtin functions.
1563 (DIVWEUO): Likewise.
1567 (DIVDEUO): Likewise.
1568 (DXEX): Add decimal floating-point builtin functions.
1571 (DDEDPDQ): Likewise.
1573 (DENBCDQ): Likewise.
1580 (CDTBCD): Add new BCD builtin functions.
1584 (BCDADD_LT): Likewise.
1585 (BCDADD_EQ): Likewise.
1586 (BCDADD_GT): Likewise.
1587 (BCDADD_OV): Likewise.
1589 (BCDSUB_LT): Likewise.
1590 (BCDSUB_EQ): Likewise.
1591 (BCDSUB_GT): Likewise.
1592 (BCDSUB_OV): Likewise.
1593 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
1594 (UNPACK_TD): Likewise.
1595 (PACK_TF): Likewise.
1596 (UNPACK_TF): Likewise.
1597 (UNPACK_TF_0): Likewise.
1598 (UNPACK_TF_1): Likewise.
1599 (PACK_V1TI): Likewise.
1600 (UNPACK_V1TI): Likewise.
1602 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
1603 support for decimal floating point builtin functions.
1604 (rs6000_expand_ternop_builtin): Add checks for the new builtin
1605 functions that take constant arguments.
1606 (rs6000_invalid_builtin): Add decimal floating point builtin
1608 (rs6000_init_builtins): Setup long double, _Decimal64, and
1609 _Decimal128 types for new builtin functions.
1610 (builtin_function_type): Set the unsigned flags appropriately for
1611 the new builtin functions.
1612 (rs6000_opt_masks): Add support for decimal floating point builtin
1615 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
1616 floating point builtin functions.
1617 (RS6000_BTM_COMMON): Likewise.
1618 (RS6000_BTI_long_double): Likewise.
1619 (RS6000_BTI_dfloat64): Likewise.
1620 (RS6000_BTI_dfloat128): Likewise.
1621 (long_double_type_internal_node): Likewise.
1622 (dfloat64_type_internal_node): Likewise.
1623 (dfloat128_type_internal_node): Likewise.
1625 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
1626 2.07 bcd arithmetic instructions.
1627 (UNSPEC_BCDSUB): Likewise.
1628 (UNSPEC_BCD_OVERFLOW): Likewise.
1629 (UNSPEC_BCD_ADD_SUB): Likewise.
1630 (bcd_add_sub): Likewise.
1631 (BCD_TEST): Likewise.
1632 (bcd<bcd_add_sub>): Likewise.
1633 (bcd<bcd_add_sub>_test): Likewise.
1634 (bcd<bcd_add_sub>_test2): Likewise.
1635 (bcd<bcd_add_sub>_<code>): Likewise.
1636 (peephole2 for combined bcd ops): Likewise.
1638 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
1639 decimal floating point builtin functions.
1640 (UNSPEC_DENBCD): Likewise.
1641 (UNSPEC_DXEX): Likewise.
1642 (UNSPEC_DIEX): Likewise.
1643 (UNSPEC_DSCLI): Likewise.
1644 (UNSPEC_DSCRI): Likewise.
1645 (D64_D128): Likewise.
1646 (dfp_suffix): Likewise.
1647 (dfp_ddedpd_<mode>): Likewise.
1648 (dfp_denbcd_<mode>): Likewise.
1649 (dfp_dxex_<mode>): Likewise.
1650 (dfp_diex_<mode>): Likewise.
1651 (dfp_dscli_<mode>): Likewise.
1652 (dfp_dscri_<mode>): Likewise.
1654 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
1656 (UNSPEC_CDTBCD): Likewise.
1657 (UNSPEC_CBCDTD): Likewise.
1658 (UNSPEC_DIVE): Add support for new extended divide builtin
1660 (UNSPEC_DIVEO): Likewise.
1661 (UNSPEC_DIVEU): Likewise.
1662 (UNSPEC_DIVEUO): Likewise.
1663 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
1664 pack/unpack 128-bit types.
1665 (UNSPEC_PACK_128BIT): Likewise.
1666 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
1667 (udiv<mode>3): Use idiv_ldiv mode attribute.
1668 (div<mode>3): Likewise.
1669 (addg6s): Add new BCD builtin functions.
1672 (UNSPEC_DIV_EXTEND): Add support for new extended divide
1674 (div_extend): Likewise.
1675 (div<div_extend>_<mode>"): Likewise.
1676 (FP128_64): Add support for new builtin functions to pack/unpack
1678 (unpack<mode>): Likewise.
1679 (unpacktf_0): Likewise.
1680 (unpacktf_1): Likewise.
1681 (unpack<mode>_dm): Likewise.
1682 (unpack<mode>_nodm): Likewise.
1683 (pack<mode>): Likewise.
1684 (unpackv1ti): Likewise.
1685 (packv1ti): Likewise.
1687 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
1689 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
1692 2014-04-24 Jakub Jelinek <jakub@redhat.com>
1694 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
1695 * gimplify.c (omp_is_private): Change last argument's type to int.
1696 Only diagnose lastprivate if the simd argument is 1, only diagnose
1697 linear if the simd argument is 2.
1698 (gimplify_omp_for): Adjust omp_is_private callers. When adding
1699 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
1700 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
1701 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
1702 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
1703 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
1704 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
1705 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
1706 * tree-nested.c (convert_nonlocal_omp_clauses,
1707 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
1709 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
1712 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
1715 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
1717 * flag-types.h (enum ivar_visibility): Add.
1719 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
1721 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
1722 function * argument.
1724 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
1726 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
1728 2014-04-24 Radovan Obradovic <robradovic@mips.com>
1729 Tom de Vries <tom@codesourcery.com>
1731 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
1732 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
1734 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
1735 * emit-rtl.c (try_split): Same.
1737 2014-04-24 Radovan Obradovic <robradovic@mips.com>
1738 Tom de Vries <tom@codesourcery.com>
1740 * common.opt (fuse-caller-save): New option.
1742 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
1744 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
1747 2014-04-24 Richard Biener <rguenther@suse.de>
1749 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
1750 during TER and instead use the sepops interface for expanding
1751 non-GIMPLE_SINGLE_RHS.
1753 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1755 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
1756 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
1758 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1760 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
1761 assembler 64-bit option.
1762 * configure: Regenerate.
1764 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1766 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
1767 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
1768 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
1769 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
1770 (TARGET_CRYPTO): Take TARGET_SIMD into account.
1772 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1774 * config/aarch64/aarch64-builtins.c
1775 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
1776 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
1777 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
1778 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
1780 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
1781 (Vrevsuff): New mode attribute.
1783 2014-04-24 Terry Guo <terry.guo@arm.com>
1785 * config/arm/arm.h (machine_function): Define variable
1786 after_arm_reorg here.
1787 * config/arm/arm.c (after_arm_reorg): Remove the definition.
1788 (arm_split_constant): Update the way to access variable
1791 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
1793 2014-04-23 Tom de Vries <tom@codesourcery.com>
1795 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
1797 2014-04-23 David Malcolm <dmalcolm@redhat.com>
1799 * is-a.h: Update comments to reflect the following changes to the
1800 "pointerness" of the API, making the template parameter match the
1801 return type, allowing use of is-a.h with typedefs of pointers.
1802 (is_a_helper::cast): Return a T rather then a pointer to a T, so
1803 that the return type matches the parameter to the is_a_helper.
1805 (dyn_cast): Likewise.
1807 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
1808 pointer from the is-a.h API.
1810 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
1811 (is_a_helper <cgraph_node *>::test): ...this, matching change to
1813 (is_a_helper <varpool_node>::test): Likewise, convert to...
1814 (is_a_helper <varpool_node *>::test): ...this.
1816 (varpool_first_variable): Update for removal of implicit pointer
1817 from the is-a.h API.
1818 (varpool_next_variable): Likewise.
1819 (varpool_first_static_initializer): Likewise.
1820 (varpool_next_static_initializer): Likewise.
1821 (varpool_first_defined_variable): Likewise.
1822 (varpool_next_defined_variable): Likewise.
1823 (cgraph_first_defined_function): Likewise.
1824 (cgraph_next_defined_function): Likewise.
1825 (cgraph_first_function): Likewise.
1826 (cgraph_next_function): Likewise.
1827 (cgraph_first_function_with_gimple_body): Likewise.
1828 (cgraph_next_function_with_gimple_body): Likewise.
1829 (cgraph_alias_target): Likewise.
1830 (varpool_alias_target): Likewise.
1831 (cgraph_function_or_thunk_node): Likewise.
1832 (varpool_variable_node): Likewise.
1833 (symtab_real_symbol_p): Likewise.
1834 * cgraphunit.c (referred_to_p): Likewise.
1835 (analyze_functions): Likewise.
1836 (handle_alias_pairs): Likewise.
1837 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1838 * gimple-ssa.h (gimple_vuse_op): Likewise.
1839 (gimple_vdef_op): Likewise.
1840 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
1841 * gimple.c (gimple_build_asm_1): Likewise.
1842 (gimple_build_try): Likewise.
1843 (gimple_build_resx): Likewise.
1844 (gimple_build_eh_dispatch): Likewise.
1845 (gimple_build_omp_for): Likewise.
1846 (gimple_omp_for_set_clauses): Likewise.
1848 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
1849 (is_a_helper <gimple_statement_asm *>::test): ...this.
1850 (is_a_helper <gimple_statement_bind>::test): Convert to...
1851 (is_a_helper <gimple_statement_bind *>::test): ...this.
1852 (is_a_helper <gimple_statement_call>::test): Convert to...
1853 (is_a_helper <gimple_statement_call *>::test): ...this.
1854 (is_a_helper <gimple_statement_catch>::test): Convert to...
1855 (is_a_helper <gimple_statement_catch *>::test): ...this.
1856 (is_a_helper <gimple_statement_resx>::test): Convert to...
1857 (is_a_helper <gimple_statement_resx *>::test): ...this.
1858 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
1859 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
1860 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
1861 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
1862 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
1863 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
1864 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
1865 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
1866 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
1867 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
1868 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
1869 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
1870 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
1871 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
1872 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
1873 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
1874 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
1875 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
1876 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
1877 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
1878 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
1879 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
1880 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
1881 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
1882 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
1883 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
1884 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
1885 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
1886 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
1887 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
1888 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
1889 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
1890 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
1891 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
1892 (is_a_helper <gimple_statement_phi>::test): Convert to...
1893 (is_a_helper <gimple_statement_phi *>::test): ...this.
1894 (is_a_helper <gimple_statement_transaction>::test): Convert to...
1895 (is_a_helper <gimple_statement_transaction *>::test): ...this.
1896 (is_a_helper <gimple_statement_try>::test): Convert to...
1897 (is_a_helper <gimple_statement_try *>::test): ...this.
1898 (is_a_helper <gimple_statement_wce>::test): Convert to...
1899 (is_a_helper <gimple_statement_wce *>::test): ...this.
1900 (is_a_helper <const gimple_statement_asm>::test): Convert to...
1901 (is_a_helper <const gimple_statement_asm *>::test): ...this.
1902 (is_a_helper <const gimple_statement_bind>::test): Convert to...
1903 (is_a_helper <const gimple_statement_bind *>::test): ...this.
1904 (is_a_helper <const gimple_statement_call>::test): Convert to...
1905 (is_a_helper <const gimple_statement_call *>::test): ...this.
1906 (is_a_helper <const gimple_statement_catch>::test): Convert to...
1907 (is_a_helper <const gimple_statement_catch *>::test): ...this.
1908 (is_a_helper <const gimple_statement_resx>::test): Convert to...
1909 (is_a_helper <const gimple_statement_resx *>::test): ...this.
1910 (is_a_helper <const gimple_statement_eh_dispatch>::test):
1912 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
1914 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
1916 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
1917 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
1919 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
1921 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
1923 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
1925 (is_a_helper <const gimple_statement_omp_return>::test): Convert
1927 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
1928 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
1930 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
1931 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
1933 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
1934 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
1935 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
1936 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
1938 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
1939 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
1941 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
1942 (is_a_helper <const gimple_statement_omp_target>::test): Convert
1944 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
1945 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
1947 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
1948 (is_a_helper <const gimple_statement_omp_single>::test): Convert
1950 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
1951 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
1953 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
1954 (is_a_helper <const gimple_statement_omp_task>::test): Convert
1956 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
1957 (is_a_helper <const gimple_statement_phi>::test): Convert to...
1958 (is_a_helper <const gimple_statement_phi *>::test): ...this.
1959 (is_a_helper <const gimple_statement_transaction>::test): Convert
1961 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
1962 (is_a_helper <const gimple_statement_with_ops>::test): Convert
1964 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
1965 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
1966 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
1967 (is_a_helper <const gimple_statement_with_memory_ops>::test):
1969 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
1971 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
1973 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
1975 (gimple_use_ops): Update for removal of implicit pointer from the
1977 (gimple_set_use_ops): Likewise.
1978 (gimple_vuse): Likewise.
1979 (gimple_vdef): Likewise.
1980 (gimple_vuse_ptr): Likewise.
1981 (gimple_vdef_ptr): Likewise.
1982 (gimple_set_vuse): Likewise.
1983 (gimple_set_vdef): Likewise.
1984 (gimple_omp_return_set_lhs): Likewise.
1985 (gimple_omp_return_lhs): Likewise.
1986 (gimple_omp_return_lhs_ptr): Likewise.
1987 (gimple_call_fntype): Likewise.
1988 (gimple_call_set_fntype): Likewise.
1989 (gimple_call_set_internal_fn): Likewise.
1990 (gimple_call_use_set): Likewise.
1991 (gimple_call_clobber_set): Likewise.
1992 (gimple_bind_vars): Likewise.
1993 (gimple_bind_set_vars): Likewise.
1994 (gimple_bind_body_ptr): Likewise.
1995 (gimple_bind_set_body): Likewise.
1996 (gimple_bind_add_stmt): Likewise.
1997 (gimple_bind_block): Likewise.
1998 (gimple_bind_set_block): Likewise.
1999 (gimple_asm_ninputs): Likewise.
2000 (gimple_asm_noutputs): Likewise.
2001 (gimple_asm_nclobbers): Likewise.
2002 (gimple_asm_nlabels): Likewise.
2003 (gimple_asm_input_op): Likewise.
2004 (gimple_asm_input_op_ptr): Likewise.
2005 (gimple_asm_output_op): Likewise.
2006 (gimple_asm_output_op_ptr): Likewise.
2007 (gimple_asm_set_output_op): Likewise.
2008 (gimple_asm_clobber_op): Likewise.
2009 (gimple_asm_set_clobber_op): Likewise.
2010 (gimple_asm_label_op): Likewise.
2011 (gimple_asm_set_label_op): Likewise.
2012 (gimple_asm_string): Likewise.
2013 (gimple_catch_types): Likewise.
2014 (gimple_catch_types_ptr): Likewise.
2015 (gimple_catch_handler_ptr): Likewise.
2016 (gimple_catch_set_types): Likewise.
2017 (gimple_catch_set_handler): Likewise.
2018 (gimple_eh_filter_types): Likewise.
2019 (gimple_eh_filter_types_ptr): Likewise.
2020 (gimple_eh_filter_failure_ptr): Likewise.
2021 (gimple_eh_filter_set_types): Likewise.
2022 (gimple_eh_filter_set_failure): Likewise.
2023 (gimple_eh_must_not_throw_fndecl): Likewise.
2024 (gimple_eh_must_not_throw_set_fndecl): Likewise.
2025 (gimple_eh_else_n_body_ptr): Likewise.
2026 (gimple_eh_else_e_body_ptr): Likewise.
2027 (gimple_eh_else_set_n_body): Likewise.
2028 (gimple_eh_else_set_e_body): Likewise.
2029 (gimple_try_eval_ptr): Likewise.
2030 (gimple_try_cleanup_ptr): Likewise.
2031 (gimple_try_set_eval): Likewise.
2032 (gimple_try_set_cleanup): Likewise.
2033 (gimple_wce_cleanup_ptr): Likewise.
2034 (gimple_wce_set_cleanup): Likewise.
2035 (gimple_phi_capacity): Likewise.
2036 (gimple_phi_num_args): Likewise.
2037 (gimple_phi_result): Likewise.
2038 (gimple_phi_result_ptr): Likewise.
2039 (gimple_phi_set_result): Likewise.
2040 (gimple_phi_arg): Likewise.
2041 (gimple_phi_set_arg): Likewise.
2042 (gimple_resx_region): Likewise.
2043 (gimple_resx_set_region): Likewise.
2044 (gimple_eh_dispatch_region): Likewise.
2045 (gimple_eh_dispatch_set_region): Likewise.
2046 (gimple_omp_critical_name): Likewise.
2047 (gimple_omp_critical_name_ptr): Likewise.
2048 (gimple_omp_critical_set_name): Likewise.
2049 (gimple_omp_for_clauses): Likewise.
2050 (gimple_omp_for_clauses_ptr): Likewise.
2051 (gimple_omp_for_set_clauses): Likewise.
2052 (gimple_omp_for_collapse): Likewise.
2053 (gimple_omp_for_index): Likewise.
2054 (gimple_omp_for_index_ptr): Likewise.
2055 (gimple_omp_for_set_index): Likewise.
2056 (gimple_omp_for_initial): Likewise.
2057 (gimple_omp_for_initial_ptr): Likewise.
2058 (gimple_omp_for_set_initial): Likewise.
2059 (gimple_omp_for_final): Likewise.
2060 (gimple_omp_for_final_ptr): Likewise.
2061 (gimple_omp_for_set_final): Likewise.
2062 (gimple_omp_for_incr): Likewise.
2063 (gimple_omp_for_incr_ptr): Likewise.
2064 (gimple_omp_for_set_incr): Likewise.
2065 (gimple_omp_for_pre_body_ptr): Likewise.
2066 (gimple_omp_for_set_pre_body): Likewise.
2067 (gimple_omp_parallel_clauses): Likewise.
2068 (gimple_omp_parallel_clauses_ptr): Likewise.
2069 (gimple_omp_parallel_set_clauses): Likewise.
2070 (gimple_omp_parallel_child_fn): Likewise.
2071 (gimple_omp_parallel_child_fn_ptr): Likewise.
2072 (gimple_omp_parallel_set_child_fn): Likewise.
2073 (gimple_omp_parallel_data_arg): Likewise.
2074 (gimple_omp_parallel_data_arg_ptr): Likewise.
2075 (gimple_omp_parallel_set_data_arg): Likewise.
2076 (gimple_omp_task_clauses): Likewise.
2077 (gimple_omp_task_clauses_ptr): Likewise.
2078 (gimple_omp_task_set_clauses): Likewise.
2079 (gimple_omp_task_child_fn): Likewise.
2080 (gimple_omp_task_child_fn_ptr): Likewise.
2081 (gimple_omp_task_set_child_fn): Likewise.
2082 (gimple_omp_task_data_arg): Likewise.
2083 (gimple_omp_task_data_arg_ptr): Likewise.
2084 (gimple_omp_task_set_data_arg): Likewise.
2085 (gimple_omp_taskreg_clauses): Likewise.
2086 (gimple_omp_taskreg_clauses_ptr): Likewise.
2087 (gimple_omp_taskreg_set_clauses): Likewise.
2088 (gimple_omp_taskreg_child_fn): Likewise.
2089 (gimple_omp_taskreg_child_fn_ptr): Likewise.
2090 (gimple_omp_taskreg_set_child_fn): Likewise.
2091 (gimple_omp_taskreg_data_arg): Likewise.
2092 (gimple_omp_taskreg_data_arg_ptr): Likewise.
2093 (gimple_omp_taskreg_set_data_arg): Likewise.
2094 (gimple_omp_task_copy_fn): Likewise.
2095 (gimple_omp_task_copy_fn_ptr): Likewise.
2096 (gimple_omp_task_set_copy_fn): Likewise.
2097 (gimple_omp_task_arg_size): Likewise.
2098 (gimple_omp_task_arg_size_ptr): Likewise.
2099 (gimple_omp_task_set_arg_size): Likewise.
2100 (gimple_omp_task_arg_align): Likewise.
2101 (gimple_omp_task_arg_align_ptr): Likewise.
2102 (gimple_omp_task_set_arg_align): Likewise.
2103 (gimple_omp_single_clauses): Likewise.
2104 (gimple_omp_single_clauses_ptr): Likewise.
2105 (gimple_omp_single_set_clauses): Likewise.
2106 (gimple_omp_target_clauses): Likewise.
2107 (gimple_omp_target_clauses_ptr): Likewise.
2108 (gimple_omp_target_set_clauses): Likewise.
2109 (gimple_omp_target_child_fn): Likewise.
2110 (gimple_omp_target_child_fn_ptr): Likewise.
2111 (gimple_omp_target_set_child_fn): Likewise.
2112 (gimple_omp_target_data_arg): Likewise.
2113 (gimple_omp_target_data_arg_ptr): Likewise.
2114 (gimple_omp_target_set_data_arg): Likewise.
2115 (gimple_omp_teams_clauses): Likewise.
2116 (gimple_omp_teams_clauses_ptr): Likewise.
2117 (gimple_omp_teams_set_clauses): Likewise.
2118 (gimple_omp_sections_clauses): Likewise.
2119 (gimple_omp_sections_clauses_ptr): Likewise.
2120 (gimple_omp_sections_set_clauses): Likewise.
2121 (gimple_omp_sections_control): Likewise.
2122 (gimple_omp_sections_control_ptr): Likewise.
2123 (gimple_omp_sections_set_control): Likewise.
2124 (gimple_omp_for_set_cond): Likewise.
2125 (gimple_omp_for_cond): Likewise.
2126 (gimple_omp_atomic_store_set_val): Likewise.
2127 (gimple_omp_atomic_store_val): Likewise.
2128 (gimple_omp_atomic_store_val_ptr): Likewise.
2129 (gimple_omp_atomic_load_set_lhs): Likewise.
2130 (gimple_omp_atomic_load_lhs): Likewise.
2131 (gimple_omp_atomic_load_lhs_ptr): Likewise.
2132 (gimple_omp_atomic_load_set_rhs): Likewise.
2133 (gimple_omp_atomic_load_rhs): Likewise.
2134 (gimple_omp_atomic_load_rhs_ptr): Likewise.
2135 (gimple_omp_continue_control_def): Likewise.
2136 (gimple_omp_continue_control_def_ptr): Likewise.
2137 (gimple_omp_continue_set_control_def): Likewise.
2138 (gimple_omp_continue_control_use): Likewise.
2139 (gimple_omp_continue_control_use_ptr): Likewise.
2140 (gimple_omp_continue_set_control_use): Likewise.
2141 (gimple_transaction_body_ptr): Likewise.
2142 (gimple_transaction_label): Likewise.
2143 (gimple_transaction_label_ptr): Likewise.
2144 (gimple_transaction_set_body): Likewise.
2145 (gimple_transaction_set_label): Likewise.
2147 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
2148 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2149 * ipa-ref.c (ipa_record_reference): Likewise.
2150 * ipa-reference.c (analyze_function): Likewise.
2151 (ipa_reference_write_optimization_summary): Likewise.
2152 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
2153 (address_taken_from_non_vtable_p): Likewise.
2154 (comdat_can_be_unshared_p_1): Likewise.
2155 * lto-cgraph.c (lto_output_ref): Likewise.
2156 (add_references): Likewise.
2157 (compute_ltrans_boundary): Likewise.
2158 (output_symtab): Likewise.
2159 (input_ref): Likewise.
2160 (input_cgraph_1): Likewise.
2161 (output_cgraph_opt_summary): Likewise.
2162 * lto-streamer-out.c (lto_output): Likewise.
2163 (output_symbol_p): Likewise.
2164 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
2165 (lsei_start_function_in_partition): Likewise.
2166 (lsei_next_variable_in_partition): Likewise.
2167 (lsei_start_variable_in_partition): Likewise.
2168 * symtab.c (insert_to_assembler_name_hash): Likewise.
2169 (unlink_from_assembler_name_hash): Likewise.
2170 (symtab_unregister_node): Likewise.
2171 (symtab_remove_node): Likewise.
2172 (dump_symtab_node): Likewise.
2173 (verify_symtab_base): Likewise.
2174 (verify_symtab_node): Likewise.
2175 (symtab_make_decl_local): Likewise.
2176 (symtab_alias_ultimate_target): Likewise.
2177 (symtab_resolve_alias): Likewise.
2178 (symtab_get_symbol_partitioning_class): Likewise.
2179 * tree-phinodes.c (allocate_phi_node): Likewise.
2180 (reserve_phi_args_for_new_edge): Likewise.
2181 (remove_phi_args): Likewise.
2182 * varpool.c (varpool_node_for_asm): Likewise.
2183 (varpool_remove_unreferenced_decls): Likewise.
2185 2014-04-23 Jeff Law <law@redhat.com>
2187 PR tree-optimization/60902
2188 * tree-ssa-threadedge.c
2189 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
2190 invalidate outputs from statements that do not produce useful
2191 outputs for threading.
2193 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2195 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
2196 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
2197 machine descriptions for Stack Smashing Protector.
2199 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
2201 * aarch64.md (<optab>_rol<mode>3): New pattern.
2202 (<optab>_rolsi3_uxtw): Likewise.
2203 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
2205 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
2207 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
2208 (arm_cortex_a12_tune): Likewise.
2210 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2212 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
2214 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2216 * config/arm/arm.md (arm_rev16si2): New pattern.
2217 (arm_rev16si2_alt): Likewise.
2218 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
2220 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2222 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
2223 (rev16<mode>2_alt): Likewise.
2224 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
2225 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
2226 (aarch_rev16_shleft_mask_imm_p): Likewise.
2227 (aarch_rev16_p_1): Likewise.
2228 (aarch_rev16_p): Likewise.
2229 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
2230 (aarch_rev16_shright_mask_imm_p): Likewise.
2231 (aarch_rev16_shleft_mask_imm_p): Likewise.
2233 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2235 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
2236 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
2238 (cortex_a53_extra_costs): Likewise.
2239 (cortex_a57_extra_costs): Likewise.
2240 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
2241 (cortexa7_extra_costs): Likewise.
2242 (cortexa8_extra_costs): Likewise.
2243 (cortexa12_extra_costs): Likewise.
2244 (cortexa15_extra_costs): Likewise.
2245 (v7m_extra_costs): Likewise.
2246 (arm_new_rtx_costs): Handle BSWAP.
2248 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2250 * config/arm/arm.c (cortexa8_extra_costs): New table.
2251 (arm_cortex_a8_tune): New tuning struct.
2252 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
2254 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2256 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
2258 2014-04-23 Richard Biener <rguenther@suse.de>
2260 * Makefile.in (OBJS): Remove loop-unswitch.o.
2261 * loop-unswitch.c: Delete.
2262 * tree-pass.h (make_pass_rtl_unswitch): Remove.
2263 * passes.def (pass_rtl_unswitch): Likewise.
2264 * loop-init.c (gate_rtl_unswitch): Likewise.
2265 (rtl_unswitch): Likewise.
2266 (pass_data_rtl_unswitch): Likewise.
2267 (pass_rtl_unswitch): Likewise.
2268 (make_pass_rtl_unswitch): Likewise.
2269 * rtl.h (reversed_condition): Likewise.
2270 (compare_and_jump_seq): Likewise.
2271 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
2273 * loop-unroll.c (compare_and_jump_seq): Likewise.
2275 2014-04-23 Richard Biener <rguenther@suse.de>
2277 PR tree-optimization/60903
2278 * tree-ssa-loop-im.c (analyze_memory_references): Remove
2279 commented code block.
2280 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
2281 loop flags to newly created BBs and edges.
2283 2014-04-23 Nick Clifton <nickc@redhat.com>
2285 * config/msp430/msp430.c (msp430_handle_option): Move function
2287 (msp430_option_override): Simplify mcu and mcpu option handling.
2288 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
2289 support for -mhwmult command line option.
2290 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
2291 -mhwmult command line option.
2292 (msp430_hwmult_enabled): Delete.
2293 (msp43o_output_labelref): Add support for -mhwmult command line
2295 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
2296 (umulsidi3): Likewise.
2297 * config/msp430/msp430.opt (mmcu): Add Report attribute.
2298 (mcpu, mlarge, msmall): Likewise.
2299 (mhwmult): New option.
2300 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
2302 (msp430_is_f5_mcu): Remove prototype.
2303 (msp430_use_f5_series_hwmult): Add prototype.
2304 * config/msp430/msp430-opts.h: New file.
2305 * common/config/msp430: New directory.
2306 * common/config/msp430/msp430-common.c: New file.
2307 * config.gcc (msp430): Remove target_has_targetm_common.
2308 * doc/invoke.texi: Document -mhwmult command line option.
2310 2014-04-23 Nick Clifton <nickc@redhat.com>
2312 * config/i386/cygwin.h (ENDFILE_SPEC): Include
2313 default-manifest.o if it can be found in the search path.
2314 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
2316 2014-04-23 Terry Guo <terry.guo@arm.com>
2318 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
2320 2014-04-23 Richard Biener <rguenther@suse.de>
2323 * tree-inline.c (declare_return_variable): Use mark_addressable.
2325 2014-04-23 Richard Biener <rguenther@suse.de>
2328 * loop-init.c (loop_optimizer_init): Make sure to apply
2329 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
2331 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2334 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
2336 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
2337 if flag_sanitize_undefined_trap_on_error.
2338 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
2339 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
2340 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
2341 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
2342 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
2343 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
2344 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
2345 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
2346 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
2347 * ubsan.c (ubsan_instrument_unreachable): Return
2348 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
2349 (ubsan_expand_null_ifn): Emit __builtin_trap ()
2350 if flag_sanitize_undefined_trap_on_error and
2351 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
2352 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
2353 instrument_bool_enum_load): Emit __builtin_trap () if
2354 flag_sanitize_undefined_trap_on_error and
2355 __builtin_handle_*_abort () if !flag_sanitize_recover.
2356 * doc/invoke.texi (-fsanitize-recover,
2357 -fsanitize-undefined-trap-on-error): Document.
2359 2014-04-22 Christian Bruel <christian.bruel@st.com>
2361 * config/sh/sh.md (mov<mode>): Replace movQIHI.
2362 Force immediates to SImode.
2364 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
2366 * config/nios2/nios2.md (UNSPEC_ROUND): New.
2368 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
2369 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
2370 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
2371 (nios2_fpu_insn): Add entry for round.
2372 (N2FPU_NO_ERRNO_P): Define.
2373 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
2375 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
2377 2014-04-22 Richard Henderson <rth@redhat.com>
2379 * config/aarch64/aarch64 (addti3, subti3): New expanders.
2380 (add<GPI>3_compare0): Remove leading * from name.
2381 (add<GPI>3_carryin): Likewise.
2382 (sub<GPI>3_compare0): Likewise.
2383 (sub<GPI>3_carryin): Likewise.
2384 (<su_optab>mulditi3): New expander.
2385 (multi3): New expander.
2386 (madd<GPI>): Remove leading * from name.
2388 2014-04-22 Martin Jambor <mjambor@suse.cz>
2390 * cgraphclones.c (cgraph_function_versioning): Copy
2391 ipa_transforms_to_apply instead of asserting it is empty.
2393 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
2396 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
2397 on count_exp to get mode.
2399 2014-04-22 Andrew Pinski <apinski@cavium.com>
2401 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
2402 Handle TLS for ILP32.
2403 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
2404 (tlsie_small_<mode>): this and handle PTR.
2405 (tlsie_small_sidi): New pattern.
2406 (tlsle_small): Change to an expand to handle ILP32.
2407 (tlsle_small_<mode>): New pattern.
2408 (tlsdesc_small): Rename to ...
2409 (tlsdesc_small_<mode>): this and handle PTR.
2411 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2413 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
2415 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2417 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2418 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
2419 (aarch64_types_signed_poly_qualifiers): Likewise.
2420 (aarch64_types_unsigned_signed_qualifiers): Likewise.
2421 (aarch64_types_poly_signed_qualifiers): Likewise.
2422 (TYPES_REINTERP_SS): Type macro added.
2423 (TYPES_REINTERP_SU): Likewise.
2424 (TYPES_REINTERP_SP): Likewise.
2425 (TYPES_REINTERP_US): Likewise.
2426 (TYPES_REINTERP_PS): Likewise.
2427 (aarch64_fold_builtin): New expression folding added.
2428 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
2429 Declarations removed.
2430 (REINTERP_SS): Declarations added.
2431 (REINTERP_US): Likewise.
2432 (REINTERP_PS): Likewise.
2433 (REINTERP_SU): Likewise.
2434 (REINTERP_SP): Likewise.
2435 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
2436 (vreinterpretq_p8_f64): Likewise.
2437 (vreinterpret_p16_f64): Likewise.
2438 (vreinterpretq_p16_f64): Likewise.
2439 (vreinterpret_f32_f64): Likewise.
2440 (vreinterpretq_f32_f64): Likewise.
2441 (vreinterpret_f64_f32): Likewise.
2442 (vreinterpret_f64_p8): Likewise.
2443 (vreinterpret_f64_p16): Likewise.
2444 (vreinterpret_f64_s8): Likewise.
2445 (vreinterpret_f64_s16): Likewise.
2446 (vreinterpret_f64_s32): Likewise.
2447 (vreinterpret_f64_s64): Likewise.
2448 (vreinterpret_f64_u8): Likewise.
2449 (vreinterpret_f64_u16): Likewise.
2450 (vreinterpret_f64_u32): Likewise.
2451 (vreinterpret_f64_u64): Likewise.
2452 (vreinterpretq_f64_f32): Likewise.
2453 (vreinterpretq_f64_p8): Likewise.
2454 (vreinterpretq_f64_p16): Likewise.
2455 (vreinterpretq_f64_s8): Likewise.
2456 (vreinterpretq_f64_s16): Likewise.
2457 (vreinterpretq_f64_s32): Likewise.
2458 (vreinterpretq_f64_s64): Likewise.
2459 (vreinterpretq_f64_u8): Likewise.
2460 (vreinterpretq_f64_u16): Likewise.
2461 (vreinterpretq_f64_u32): Likewise.
2462 (vreinterpretq_f64_u64): Likewise.
2463 (vreinterpret_s64_f64): Likewise.
2464 (vreinterpretq_s64_f64): Likewise.
2465 (vreinterpret_u64_f64): Likewise.
2466 (vreinterpretq_u64_f64): Likewise.
2467 (vreinterpret_s8_f64): Likewise.
2468 (vreinterpretq_s8_f64): Likewise.
2469 (vreinterpret_s16_f64): Likewise.
2470 (vreinterpretq_s16_f64): Likewise.
2471 (vreinterpret_s32_f64): Likewise.
2472 (vreinterpretq_s32_f64): Likewise.
2473 (vreinterpret_u8_f64): Likewise.
2474 (vreinterpretq_u8_f64): Likewise.
2475 (vreinterpret_u16_f64): Likewise.
2476 (vreinterpretq_u16_f64): Likewise.
2477 (vreinterpret_u32_f64): Likewise.
2478 (vreinterpretq_u32_f64): Likewise.
2480 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2482 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2483 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
2484 (vreinterpret_p8_s8): Likewise.
2485 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
2486 (vreinterpret_p8_s16): Likewise.
2487 (vreinterpret_p8_s32): Likewise.
2488 (vreinterpret_p8_s64): Likewise.
2489 (vreinterpret_p8_f32): Likewise.
2490 (vreinterpret_p8_u8): Likewise.
2491 (vreinterpret_p8_u16): Likewise.
2492 (vreinterpret_p8_u32): Likewise.
2493 (vreinterpret_p8_u64): Likewise.
2494 (vreinterpret_p8_p16): Likewise.
2495 (vreinterpretq_p8_s8): Likewise.
2496 (vreinterpretq_p8_s16): Likewise.
2497 (vreinterpretq_p8_s32): Likewise.
2498 (vreinterpretq_p8_s64): Likewise.
2499 (vreinterpretq_p8_f32): Likewise.
2500 (vreinterpretq_p8_u8): Likewise.
2501 (vreinterpretq_p8_u16): Likewise.
2502 (vreinterpretq_p8_u32): Likewise.
2503 (vreinterpretq_p8_u64): Likewise.
2504 (vreinterpretq_p8_p16): Likewise.
2505 (vreinterpret_p16_s8): Likewise.
2506 (vreinterpret_p16_s16): Likewise.
2507 (vreinterpret_p16_s32): Likewise.
2508 (vreinterpret_p16_s64): Likewise.
2509 (vreinterpret_p16_f32): Likewise.
2510 (vreinterpret_p16_u8): Likewise.
2511 (vreinterpret_p16_u16): Likewise.
2512 (vreinterpret_p16_u32): Likewise.
2513 (vreinterpret_p16_u64): Likewise.
2514 (vreinterpret_p16_p8): Likewise.
2515 (vreinterpretq_p16_s8): Likewise.
2516 (vreinterpretq_p16_s16): Likewise.
2517 (vreinterpretq_p16_s32): Likewise.
2518 (vreinterpretq_p16_s64): Likewise.
2519 (vreinterpretq_p16_f32): Likewise.
2520 (vreinterpretq_p16_u8): Likewise.
2521 (vreinterpretq_p16_u16): Likewise.
2522 (vreinterpretq_p16_u32): Likewise.
2523 (vreinterpretq_p16_u64): Likewise.
2524 (vreinterpretq_p16_p8): Likewise.
2525 (vreinterpret_f32_s8): Likewise.
2526 (vreinterpret_f32_s16): Likewise.
2527 (vreinterpret_f32_s32): Likewise.
2528 (vreinterpret_f32_s64): Likewise.
2529 (vreinterpret_f32_u8): Likewise.
2530 (vreinterpret_f32_u16): Likewise.
2531 (vreinterpret_f32_u32): Likewise.
2532 (vreinterpret_f32_u64): Likewise.
2533 (vreinterpret_f32_p8): Likewise.
2534 (vreinterpret_f32_p16): Likewise.
2535 (vreinterpretq_f32_s8): Likewise.
2536 (vreinterpretq_f32_s16): Likewise.
2537 (vreinterpretq_f32_s32): Likewise.
2538 (vreinterpretq_f32_s64): Likewise.
2539 (vreinterpretq_f32_u8): Likewise.
2540 (vreinterpretq_f32_u16): Likewise.
2541 (vreinterpretq_f32_u32): Likewise.
2542 (vreinterpretq_f32_u64): Likewise.
2543 (vreinterpretq_f32_p8): Likewise.
2544 (vreinterpretq_f32_p16): Likewise.
2545 (vreinterpret_s64_s8): Likewise.
2546 (vreinterpret_s64_s16): Likewise.
2547 (vreinterpret_s64_s32): Likewise.
2548 (vreinterpret_s64_f32): Likewise.
2549 (vreinterpret_s64_u8): Likewise.
2550 (vreinterpret_s64_u16): Likewise.
2551 (vreinterpret_s64_u32): Likewise.
2552 (vreinterpret_s64_u64): Likewise.
2553 (vreinterpret_s64_p8): Likewise.
2554 (vreinterpret_s64_p16): Likewise.
2555 (vreinterpretq_s64_s8): Likewise.
2556 (vreinterpretq_s64_s16): Likewise.
2557 (vreinterpretq_s64_s32): Likewise.
2558 (vreinterpretq_s64_f32): Likewise.
2559 (vreinterpretq_s64_u8): Likewise.
2560 (vreinterpretq_s64_u16): Likewise.
2561 (vreinterpretq_s64_u32): Likewise.
2562 (vreinterpretq_s64_u64): Likewise.
2563 (vreinterpretq_s64_p8): Likewise.
2564 (vreinterpretq_s64_p16): Likewise.
2565 (vreinterpret_u64_s8): Likewise.
2566 (vreinterpret_u64_s16): Likewise.
2567 (vreinterpret_u64_s32): Likewise.
2568 (vreinterpret_u64_s64): Likewise.
2569 (vreinterpret_u64_f32): Likewise.
2570 (vreinterpret_u64_u8): Likewise.
2571 (vreinterpret_u64_u16): Likewise.
2572 (vreinterpret_u64_u32): Likewise.
2573 (vreinterpret_u64_p8): Likewise.
2574 (vreinterpret_u64_p16): Likewise.
2575 (vreinterpretq_u64_s8): Likewise.
2576 (vreinterpretq_u64_s16): Likewise.
2577 (vreinterpretq_u64_s32): Likewise.
2578 (vreinterpretq_u64_s64): Likewise.
2579 (vreinterpretq_u64_f32): Likewise.
2580 (vreinterpretq_u64_u8): Likewise.
2581 (vreinterpretq_u64_u16): Likewise.
2582 (vreinterpretq_u64_u32): Likewise.
2583 (vreinterpretq_u64_p8): Likewise.
2584 (vreinterpretq_u64_p16): Likewise.
2585 (vreinterpret_s8_s16): Likewise.
2586 (vreinterpret_s8_s32): Likewise.
2587 (vreinterpret_s8_s64): Likewise.
2588 (vreinterpret_s8_f32): Likewise.
2589 (vreinterpret_s8_u8): Likewise.
2590 (vreinterpret_s8_u16): Likewise.
2591 (vreinterpret_s8_u32): Likewise.
2592 (vreinterpret_s8_u64): Likewise.
2593 (vreinterpret_s8_p8): Likewise.
2594 (vreinterpret_s8_p16): Likewise.
2595 (vreinterpretq_s8_s16): Likewise.
2596 (vreinterpretq_s8_s32): Likewise.
2597 (vreinterpretq_s8_s64): Likewise.
2598 (vreinterpretq_s8_f32): Likewise.
2599 (vreinterpretq_s8_u8): Likewise.
2600 (vreinterpretq_s8_u16): Likewise.
2601 (vreinterpretq_s8_u32): Likewise.
2602 (vreinterpretq_s8_u64): Likewise.
2603 (vreinterpretq_s8_p8): Likewise.
2604 (vreinterpretq_s8_p16): Likewise.
2605 (vreinterpret_s16_s8): Likewise.
2606 (vreinterpret_s16_s32): Likewise.
2607 (vreinterpret_s16_s64): Likewise.
2608 (vreinterpret_s16_f32): Likewise.
2609 (vreinterpret_s16_u8): Likewise.
2610 (vreinterpret_s16_u16): Likewise.
2611 (vreinterpret_s16_u32): Likewise.
2612 (vreinterpret_s16_u64): Likewise.
2613 (vreinterpret_s16_p8): Likewise.
2614 (vreinterpret_s16_p16): Likewise.
2615 (vreinterpretq_s16_s8): Likewise.
2616 (vreinterpretq_s16_s32): Likewise.
2617 (vreinterpretq_s16_s64): Likewise.
2618 (vreinterpretq_s16_f32): Likewise.
2619 (vreinterpretq_s16_u8): Likewise.
2620 (vreinterpretq_s16_u16): Likewise.
2621 (vreinterpretq_s16_u32): Likewise.
2622 (vreinterpretq_s16_u64): Likewise.
2623 (vreinterpretq_s16_p8): Likewise.
2624 (vreinterpretq_s16_p16): Likewise.
2625 (vreinterpret_s32_s8): Likewise.
2626 (vreinterpret_s32_s16): Likewise.
2627 (vreinterpret_s32_s64): Likewise.
2628 (vreinterpret_s32_f32): Likewise.
2629 (vreinterpret_s32_u8): Likewise.
2630 (vreinterpret_s32_u16): Likewise.
2631 (vreinterpret_s32_u32): Likewise.
2632 (vreinterpret_s32_u64): Likewise.
2633 (vreinterpret_s32_p8): Likewise.
2634 (vreinterpret_s32_p16): Likewise.
2635 (vreinterpretq_s32_s8): Likewise.
2636 (vreinterpretq_s32_s16): Likewise.
2637 (vreinterpretq_s32_s64): Likewise.
2638 (vreinterpretq_s32_f32): Likewise.
2639 (vreinterpretq_s32_u8): Likewise.
2640 (vreinterpretq_s32_u16): Likewise.
2641 (vreinterpretq_s32_u32): Likewise.
2642 (vreinterpretq_s32_u64): Likewise.
2643 (vreinterpretq_s32_p8): Likewise.
2644 (vreinterpretq_s32_p16): Likewise.
2645 (vreinterpret_u8_s8): Likewise.
2646 (vreinterpret_u8_s16): Likewise.
2647 (vreinterpret_u8_s32): Likewise.
2648 (vreinterpret_u8_s64): Likewise.
2649 (vreinterpret_u8_f32): Likewise.
2650 (vreinterpret_u8_u16): Likewise.
2651 (vreinterpret_u8_u32): Likewise.
2652 (vreinterpret_u8_u64): Likewise.
2653 (vreinterpret_u8_p8): Likewise.
2654 (vreinterpret_u8_p16): Likewise.
2655 (vreinterpretq_u8_s8): Likewise.
2656 (vreinterpretq_u8_s16): Likewise.
2657 (vreinterpretq_u8_s32): Likewise.
2658 (vreinterpretq_u8_s64): Likewise.
2659 (vreinterpretq_u8_f32): Likewise.
2660 (vreinterpretq_u8_u16): Likewise.
2661 (vreinterpretq_u8_u32): Likewise.
2662 (vreinterpretq_u8_u64): Likewise.
2663 (vreinterpretq_u8_p8): Likewise.
2664 (vreinterpretq_u8_p16): Likewise.
2665 (vreinterpret_u16_s8): Likewise.
2666 (vreinterpret_u16_s16): Likewise.
2667 (vreinterpret_u16_s32): Likewise.
2668 (vreinterpret_u16_s64): Likewise.
2669 (vreinterpret_u16_f32): Likewise.
2670 (vreinterpret_u16_u8): Likewise.
2671 (vreinterpret_u16_u32): Likewise.
2672 (vreinterpret_u16_u64): Likewise.
2673 (vreinterpret_u16_p8): Likewise.
2674 (vreinterpret_u16_p16): Likewise.
2675 (vreinterpretq_u16_s8): Likewise.
2676 (vreinterpretq_u16_s16): Likewise.
2677 (vreinterpretq_u16_s32): Likewise.
2678 (vreinterpretq_u16_s64): Likewise.
2679 (vreinterpretq_u16_f32): Likewise.
2680 (vreinterpretq_u16_u8): Likewise.
2681 (vreinterpretq_u16_u32): Likewise.
2682 (vreinterpretq_u16_u64): Likewise.
2683 (vreinterpretq_u16_p8): Likewise.
2684 (vreinterpretq_u16_p16): Likewise.
2685 (vreinterpret_u32_s8): Likewise.
2686 (vreinterpret_u32_s16): Likewise.
2687 (vreinterpret_u32_s32): Likewise.
2688 (vreinterpret_u32_s64): Likewise.
2689 (vreinterpret_u32_f32): Likewise.
2690 (vreinterpret_u32_u8): Likewise.
2691 (vreinterpret_u32_u16): Likewise.
2692 (vreinterpret_u32_u64): Likewise.
2693 (vreinterpret_u32_p8): Likewise.
2694 (vreinterpret_u32_p16): Likewise.
2695 (vreinterpretq_u32_s8): Likewise.
2696 (vreinterpretq_u32_s16): Likewise.
2697 (vreinterpretq_u32_s32): Likewise.
2698 (vreinterpretq_u32_s64): Likewise.
2699 (vreinterpretq_u32_f32): Likewise.
2700 (vreinterpretq_u32_u8): Likewise.
2701 (vreinterpretq_u32_u16): Likewise.
2702 (vreinterpretq_u32_u64): Likewise.
2703 (vreinterpretq_u32_p8): Likewise.
2704 (vreinterpretq_u32_p16): Likewise.
2706 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2708 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
2710 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
2713 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
2714 (vqnegd_s64): Likewise.
2715 (vqabs_s64): Likewise.
2716 (vqabsd_s64): Likewise.
2718 2014-04-22 Richard Henderson <rth@redhat.com>
2720 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
2721 computation to the top of the loop.
2723 2014-04-22 Renlin <renlin.li@arm.com>
2724 Jiong Wang <jiong.wang@arm.com>
2726 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
2727 * config/aarch64/aarch64.c (aarch64_layout_frame)
2728 (aarch64_initial_elimination_offset): Likewise.
2730 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
2732 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
2735 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
2737 * machmode.h (bitwise_mode_for_mode): Declare.
2738 * stor-layout.h (bitwise_type_for_mode): Likewise.
2739 * stor-layout.c (bitwise_mode_for_mode): New function.
2740 (bitwise_type_for_mode): Likewise.
2741 * builtins.c (fold_builtin_memory_op): Use it instead of
2742 int_mode_for_mode and build_nonstandard_integer_type.
2744 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2746 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
2747 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
2748 (*-*-solaris2*): Simplify.
2749 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
2750 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
2751 *-*-solaris2.9* handling.
2753 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
2755 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
2756 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
2758 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
2759 * configure: Regenerate.
2761 * config/i386/sol2-9.h: Remove.
2763 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
2764 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
2765 Remove Solaris 9 references.
2767 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
2769 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
2770 (floatuns<GPI:mode><GPF:mode>2): Remove.
2771 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
2772 and floatuns conversions.
2773 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
2774 and floatuns conversions.
2775 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
2776 (w1,w2): New mode attributes for inequal width conversions.
2778 2014-04-22 Renlin Li <Renlin.Li@arm.com>
2780 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
2781 the output asm format.
2783 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
2785 * config/aarch64/aarch64-simd.md
2786 (aarch64_cm<optab>di): Always split.
2787 (*aarch64_cm<optab>di): New.
2788 (aarch64_cmtstdi): Always split.
2789 (*aarch64_cmtstdi): New.
2791 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2793 PR tree-optimization/60823
2794 * omp-low.c (ipa_simd_modify_function_body): Go through
2795 all SSA_NAMEs and for those refering to vector arguments
2796 which are going to be replaced adjust SSA_NAME_VAR and,
2797 if it is a default definition, change it into a non-default
2798 definition assigned at the beginning of function from new_decl.
2799 (ipa_simd_modify_stmt_ops): Rewritten.
2800 * tree-dfa.c (set_ssa_default_def): When removing default def,
2801 check for NULL loc instead of NULL *loc.
2803 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2805 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
2806 restrictions on core registers for DImode values in Thumb2.
2808 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2810 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
2811 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
2813 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2815 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
2816 (*iordi_notzesidi_di): Likewise.
2817 (*iordi_notsesidi_di): Likewise.
2819 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2821 * config/arm/arm-protos.h (tune_params): New struct members.
2822 * config/arm/arm.c: Initialise tune_params per processor.
2823 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
2824 for speed, based on new tune_params.
2826 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2828 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
2830 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
2832 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
2834 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
2835 * config/aarch64/arm_neon.h (vrnd_f64): Added.
2836 (vrnda_f64): Likewise.
2837 (vrndi_f64): Likewise.
2838 (vrndm_f64): Likewise.
2839 (vrndn_f64): Likewise.
2840 (vrndp_f64): Likewise.
2841 (vrndx_f64): Likewise.
2843 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2845 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
2846 GET_MODE_SIZE argument is enum machine_mode.
2848 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2851 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
2852 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
2854 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
2857 * asan.c (asan_emit_stack_protection): Force the base to align to
2858 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
2859 appropriate bits if STRICT_ALIGNMENT.
2860 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
2862 (expand_used_vars): Leave a space in the stack frame for alignment
2863 if STRICT_ALIGNMENT.
2865 2014-04-21 David Malcolm <dmalcolm@redhat.com>
2867 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
2869 (gimple_store_p): Likewise.
2870 (gimple_assign_load_p): Likewise.
2871 (gimple_assign_cast_p): Likewise.
2872 (gimple_clobber_p): Likewise.
2874 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
2875 rather than a gimple.
2876 (gimple_assign_cast_p): Likewise.
2878 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
2881 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
2882 If mode is DDmode and TARGET_E500_DOUBLE allow move.
2884 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
2885 more debug information for E500 if -mdebug=reg.
2887 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
2890 * config/i386/i386.c (ix86_expand_builtin)
2891 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
2892 register for target RTX.
2893 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
2895 2014-04-18 Cong Hou <congh@google.com>
2897 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
2898 the widen-mult pattern by handling two operands with different sizes,
2899 and operands whose size is smaller than half of the result type.
2901 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
2903 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
2904 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
2905 (do_estimate_edge_time): Compute it.
2906 * ipa-inline.c (want_inline_small_function_p): Bypass
2907 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
2909 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
2911 * ipa-inline.c (spec_rem): New static variable.
2912 (dump_overall_stats): New function.
2913 (dump_inline_stats): New function.
2915 2014-04-18 Richard Henderson <rth@redhat.com>
2917 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
2918 to GET_MODE_SIZE, not a reg_class_t.
2920 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2922 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
2923 (vsx_xxmrglw_<mode>): Likewise.
2925 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2928 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
2929 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
2930 (rs6000_init_hard_regno_mode_ok): Likewise.
2932 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
2934 * ipa-inline.c (inline_small_functions): Account only non-cold
2936 * doc/invoke.texi (inline-unit-growth): Update documentation.
2938 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
2940 * config/rs6000/rs6000.md (addti3, subti3): New.
2942 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
2945 * config/i386/i386.c (ix86_expand_clear): Remove outdated
2946 comment. Check optimize_insn_for_size_p instead of
2947 optimize_insn_for_speed_p.
2949 2014-04-17 Martin Jambor <mjambor@suse.cz>
2951 * gimple-iterator.c (gsi_start_edge): New function.
2952 * gimple-iterator.h (gsi_start_edge): Declare.
2953 * tree-sra.c (single_non_eh_succ): New function.
2954 (disqualify_ops_if_throwing_stmt): Renamed to
2955 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
2956 having one non-EH successor BB.
2957 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
2958 generate loads into replacements.
2959 (sra_modify_assign): Likewise and and also use the simple path for
2961 (sra_modify_function_body): Commit statements on edges.
2963 2014-04-17 Richard Biener <rguenther@suse.de>
2966 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
2967 comparison results and add clarifying comment.
2969 2014-04-17 Jakub Jelinek <jakub@redhat.com>
2971 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
2972 (blank_mode): Initialize it.
2973 (emit_mode_size_inline, emit_mode_nunits_inline,
2974 emit_mode_inner_inline): New functions.
2975 (emit_insn_modes_h): Call them and surround their output with
2976 #if GCC_VERSION >= 4001 ... #endif.
2977 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
2978 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
2979 mode_* arrays if the argument is __builtin_constant_p.
2980 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
2981 is enum machine_mode.
2983 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2985 * passes.c (opt_pass::execute): Adjust.
2986 (pass_manager::execute_pass_mode_switching): Likewise.
2987 (early_local_passes::execute): Likewise.
2988 (execute_one_pass): Pass cfun to the pass's execute method.
2989 * tree-pass.h (opt_pass::execute): Add function * argument.
2990 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
2991 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
2992 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
2993 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2994 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
2995 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
2996 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
2997 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
2998 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
2999 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3000 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3001 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3002 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3003 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3004 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3005 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3006 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3007 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3008 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3009 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3010 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3011 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3012 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3013 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3014 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3015 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3016 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3017 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3018 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3021 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3023 * passes.c (opt_pass::gate): Take function * argument.
3024 (gate_all_early_local_passes): Merge into
3025 (early_local_passes::gate): this.
3026 (gate_all_early_optimizations): Merge into
3027 (all_early_optimizations::gate): this.
3028 (gate_all_optimizations): Mege into
3029 (all_optimizations::gate): this.
3030 (gate_all_optimizations_g): Merge into
3031 (all_optimizations_g::gate): this.
3032 (gate_rest_of_compilation): Mege into
3033 (rest_of_compilation::gate): this.
3034 (gate_postreload): Merge into
3035 (postreload::gate): this.
3036 (dump_one_pass): Pass cfun to the pass's gate method.
3037 (execute_ipa_summary_passes): Likewise.
3038 (execute_one_pass): Likewise.
3039 (ipa_write_summaries_2): Likewise.
3040 (ipa_write_optimization_summaries_1): Likewise.
3041 (ipa_read_summaries_1): Likewise.
3042 (ipa_read_optimization_summaries_1): Likewise.
3043 (execute_ipa_stmt_fixups): Likewise.
3044 * tree-pass.h (opt_pass::gate): Add function * argument.
3045 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
3046 combine-stack-adj.c, combine.c, compare-elim.c,
3047 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3048 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
3049 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
3050 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
3051 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3052 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3053 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3054 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
3055 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
3056 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
3057 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3058 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3059 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
3060 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3061 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3062 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3063 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3064 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3065 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3066 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3067 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3068 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3069 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3070 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
3071 var-tracking.c, vtable-verify.c, web.c: Adjust.
3073 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3075 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
3076 * configure: Regenerate.
3078 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3080 * passes.c (dump_one_pass): don't check pass->has_gate.
3081 (execute_ipa_summary_passes): Likewise.
3082 (execute_one_pass): Likewise.
3083 (ipa_write_summaries_2): Likewise.
3084 (ipa_write_optimization_summaries_1): Likewise.
3085 (ipa_read_optimization_summaries_1): Likewise.
3086 (execute_ipa_stmt_fixups): Likewise.
3087 * tree-pass.h (pass_data::has_gate): Remove.
3088 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3089 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3090 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3091 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3092 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3093 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
3094 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
3095 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
3096 gimple-low.c, gimple-ssa-isolate-paths.c,
3097 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
3098 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3099 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3100 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
3101 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
3102 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
3103 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
3104 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3105 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
3106 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3107 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3108 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3109 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3110 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3111 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3112 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3113 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3114 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3115 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3116 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3117 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3118 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3121 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3123 * pass_manager.h (pass_manager::register_dump_files_1): Remove
3125 * passes.c (pass_manager::register_dump_files_1): Merge into
3126 (pass_manager::register_dump_files): this, and remove its handling of
3127 properties since the pass always has the properties anyway.
3128 (pass_manager::pass_manager): Adjust.
3130 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3132 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
3133 * passes.c (pass_manager::register_dump_files_1): Remove dead code
3134 dealing with properties.
3135 (pass_manager::register_dump_files): Adjust.
3137 2014-03-20 Mark Wielaard <mjw@redhat.com>
3139 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
3140 then represent the bound as normal constant value.
3142 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3145 Forward port from 4.8 branch
3146 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
3148 * config/i386/bmiintrin.h (_blsi_u32): New.
3152 (_blsmsk_u32): Ditto.
3153 (_blsmsk_u64): Ditto.
3154 (_tzcnt_u32): Ditto.
3155 (_tzcnt_u64): Ditto.
3157 2014-04-17 Kito Cheng <kito@0xlab.org>
3159 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
3161 2014-04-17 Richard Biener <rguenther@suse.de>
3164 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
3165 boolean results for comparisons.
3167 2014-04-17 Richard Biener <rguenther@suse.de>
3169 PR tree-optimization/60836
3170 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
3171 initial PHI args to be gimple values.
3173 2014-04-17 Richard Biener <rguenther@suse.de>
3175 PR tree-optimization/60841
3176 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
3177 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
3178 of stmts to SLP build.
3179 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
3180 (vect_analyze_slp): Likewise.
3181 (vect_analyze_slp_instance): Likewise.
3182 (vect_build_slp_tree): Limit overall SLP tree growth.
3183 * tree-vectorizer.h (vect_analyze_data_refs,
3184 vect_analyze_slp): Adjust prototypes.
3186 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3188 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
3191 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3193 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
3194 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
3195 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
3196 for TARGET_SLOW_PSHUFB
3198 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3200 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
3201 * config/i386/i386.c (intel_cost): Ditto.
3203 2014-04-17 Joey Ye <joey.ye@arm.com>
3205 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
3207 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3209 * opts.c (common_handle_option): Disable -fipa-reference coorectly
3212 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3214 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
3215 (type_all_derivations_known_p): New predicate.
3216 (type_all_ctors_visible_p): New predicate.
3217 (type_possibly_instantiated_p): New predicate.
3218 (get_odr_type): Compute all_derivations_known.
3219 (dump_odr_type): Dump the flag.
3220 (maybe_record_type): Cleanup.
3221 (record_target_from_binfo): Add bases_to_consider array;
3222 record bases for types w/o instances and skip CXX destructor.
3223 (possible_polymorphic_call_targets_1): Add bases_to_consider
3224 and consider_construction parameters; check if type may have instance.
3225 (get_polymorphic_call_info): Set maybe_in_construction to true
3226 when we know nothing.
3227 (record_targets_from_bases): Skip CXX destructors; they are
3228 never called for types in construction.
3229 (possible_polymorphic_call_targets): Do not record target when
3230 type may not have instance.
3232 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3235 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
3236 external aliases alive, too.
3238 2014-04-16 Andrew Pinski <apinski@cavium.com>
3240 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
3243 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3245 * final.c (compute_alignments): Do not apply loop alignment to a block
3246 falling through to the exit.
3248 2014-04-16 Catherine Moore <clm@codesourcery.com>
3250 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
3251 Adjust constraints for microMIPS store patterns.
3253 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3255 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
3257 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3259 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
3260 (append_use): Run at -O0.
3261 (append_vdef): Likewise.
3262 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
3263 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
3265 2014-04-16 Jakub Jelinek <jakub@redhat.com>
3267 PR tree-optimization/60844
3268 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
3269 (propagate_op_to_single_use, remove_visited_stmt_chain,
3270 linearize_expr, repropagate_negates, reassociate_bb): Use it
3271 instead of gsi_remove.
3273 2014-04-16 Martin Jambor <mjambor@suse.cz>
3275 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
3276 ipa_transforms_to_apply.
3277 (cgraph_function_versioning): Assert that old_node has empty
3278 ipa_transforms_to_apply.
3279 * trans-mem.c (ipa_tm_create_version): Likewise.
3280 * tree-inline.c (tree_function_versioning): Do not duplicate
3281 ipa_transforms_to_apply.
3283 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3286 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
3288 Pass necessary as flags on 64-bit Solaris/x86.
3289 Use lowercase relocs for x86_64-*-*.
3290 * configure: Regenerate.
3292 2014-04-15 Jan Hubicka <jh@suse.cz>
3294 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
3295 (maybe_record_node, likely_target_p): Use it.
3297 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3300 Revert following patch
3302 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3305 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3306 software floating point or no floating point registers, do not
3307 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3308 in GPRs that occurs after we tested for GPRs that would never be
3311 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3312 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3313 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3314 specifically allow DDmode, since that does not use the SPE SIMD
3317 2014-03-21 Mark Wielaard <mjw@redhat.com>
3319 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
3320 as unsigned or int depending on type and value used.
3322 2014-04-15 Richard Biener <rguenther@suse.de>
3324 PR rtl-optimization/56965
3325 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
3326 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
3328 * alias.c (true_dependence_1): Do not call
3329 nonoverlapping_component_refs_p.
3330 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
3331 nonoverlapping_component_refs_p.
3332 (indirect_refs_may_alias_p): Likewise.
3334 2014-04-15 Teresa Johnson <tejohnson@google.com>
3336 * cfg.c (dump_bb_info): Fix flags check.
3337 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
3339 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3341 PR rtl-optimization/60663
3342 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
3345 2014-04-15 Richard Biener <rguenther@suse.de>
3347 * lto-streamer.h (LTO_major_version): Bump to 4.
3349 2014-04-15 Richard Biener <rguenther@suse.de>
3351 * common.opt (lto_partition_model): New enum.
3352 (flto-partition=): Merge separate options with a single with argument,
3353 add -flto-partition=one support.
3354 * flag-types.h (enum lto_partition_model): Declare.
3355 * opts.c (finish_options): Remove duplicate -flto-partition=
3357 * lto-wrapper.c (run_gcc): Adjust.
3359 2014-04-15 Richard Biener <rguenther@suse.de>
3361 * alias.c (ncr_compar): New function.
3362 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
3364 2014-04-15 Richard Biener <rguenther@suse.de>
3366 * alias.c (record_component_aliases): Do not walk BINFOs.
3368 2014-04-15 Richard Biener <rguenther@suse.de>
3370 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
3371 Add struct function argument and adjust.
3372 (find_func_aliases_for_call): Likewise.
3373 (find_func_aliases): Likewise.
3374 (find_func_clobbers): Likewise.
3375 (intra_create_variable_infos): Likewise.
3376 (compute_points_to_sets): Likewise.
3377 (ipa_pta_execute): Adjust. Do not push/pop cfun.
3379 2014-04-15 Richard Biener <rguenther@suse.de>
3381 * tree.c (iterative_hash_expr): Use enum tree_code_class
3382 to store TREE_CODE_CLASS.
3383 (tree_block): Likewise.
3384 (tree_set_block): Likewise.
3385 * tree.h (fold_build_pointer_plus_loc): Use
3386 convert_to_ptrofftype_loc.
3388 2014-04-15 Jakub Jelinek <jakub@redhat.com>
3391 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
3394 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
3396 * cfgloop.h (struct loop): Move force_vectorize down.
3397 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
3398 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
3399 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
3400 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
3401 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
3402 * tree-core.h (enum annot_expr_kind): Add new kind values.
3403 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
3404 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
3406 * tree.def (ANNOTATE_EXPR): Tweak comment.
3408 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3410 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
3413 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
3415 * tree.h (TYPE_IDENTIFIER): Declare.
3416 * tree.c (subrange_type_for_debug_p): Use it.
3417 * godump.c (go_format_type): Likewise.
3418 * dwarf2out.c (is_cxx_auto, modified_type_die,
3419 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
3420 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
3422 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3425 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
3427 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
3429 * config/i386/i386.c (examine_argument): Return bool. Return true if
3430 parameter should be passed in memory.
3431 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
3432 (construct_container): Update calls to examine_argument.
3433 (function_arg_advance_64): Ditto.
3434 (return_in_memory_32): Merge with ix86_return_in_memory.
3435 (return_in_memory_64): Ditto.
3436 (return_in_memory_ms_64): Ditto.
3438 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3440 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
3441 * coverage.c (coverage_compute_profile_id): Handle externally visible
3444 2014-04-14 Martin Jambor <mjambor@suse.cz>
3446 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
3447 DECL_DISREGARD_INLINE_LIMITS functions.
3449 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3452 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
3454 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3457 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
3458 optimize_insn_for_speed_p instead of
3459 optimize_function_for_speed_p.
3461 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
3463 * doc/invoke.texi (free): Document AArch64.
3465 2014-04-14 Richard Biener <rguenther@suse.de>
3467 PR tree-optimization/60042
3468 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
3469 (insert_into_preds_of_block): Do not prevent PHI insertion
3470 for REFERENCE exprs here ...
3471 (eliminate_dom_walker::before_dom_children): ... but prevent
3472 their use here under similar conditions when applied to the
3473 IL after PRE optimizations.
3475 2014-04-14 Richard Biener <rguenther@suse.de>
3477 * passes.def: Move early points-to after early SRA.
3479 2014-04-14 Richard Biener <rguenther@suse.de>
3481 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
3482 check for which sign-changes we allow when forwarding
3483 a converted value into a switch.
3485 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3487 * stor-layout.c (place_field): Finalize non-constant offset for the
3490 2014-04-14 Richard Biener <rguenther@suse.de>
3492 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
3494 (expand_switch_using_bit_tests_p): Likewise.
3495 (process_switch): Compute and pass on speed_p based on the
3497 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
3498 optimize_bb_for_speed_p.
3500 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3502 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
3503 * function.h (struct function): Rename has_force_vect_loops into
3504 has_force_vectorize_loops.
3505 * lto-streamer-in.c (input_cfg): Adjust for renaming.
3506 (input_struct_function_base): Likewise.
3507 * lto-streamer-out.c (output_cfg): Likewise.
3508 (output_struct_function_base): Likewise.
3509 * omp-low.c (expand_omp_simd): Likewise.
3510 * tree-cfg.c (move_sese_region_to_fn): Likewise.
3511 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
3512 (version_loop_for_if_conversion): Likewise.
3513 (tree_if_conversion): Likewise.
3514 (main_tree_if_conversion): Likewise.
3515 (gate_tree_if_conversion): Likewise.
3516 * tree-inline.c (copy_loops): Likewise.
3517 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
3518 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
3519 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
3520 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
3521 * tree-vectorizer.c (vectorize_loops): Likewise.
3522 * tree-vectorizer.h (unlimited_cost_model): Likewise.
3524 2014-04-14 Richard Biener <rguenther@suse.de>
3527 * lto-streamer-out.c (wrap_refs): New function.
3528 (lto_output): Wrap symbol references in global initializes in
3529 type-preserving MEM_REFs.
3531 2014-04-14 Christian Bruel <christian.bruel@st.com>
3533 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
3535 2014-04-14 Christian Bruel <christian.bruel@st.com>
3537 * config/sh/sh.md (setmemqi): New expand pattern.
3538 * config/sh/sh.h (CLEAR_RATIO): Define.
3539 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
3540 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
3542 2014-04-14 Richard Biener <rguenther@suse.de>
3545 * fold-const.c (negate_expr_p): Don't negate directional rounding
3547 (fold_negate_expr): Likewise.
3549 2014-04-14 Richard Biener <rguenther@suse.de>
3551 PR tree-optimization/59817
3552 PR tree-optimization/60453
3553 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
3554 recursion to catch all CHRECs in the scalar evolution and restrict
3555 the predicate for the remains appropriately.
3557 2014-04-12 Catherine Moore <clm@codesourcery.com>
3559 * config/mips/constraints.md: Add new register constraint "kb".
3560 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
3561 (*movhi_internal): Likewise.
3562 (*movqi_internal): Likewise.
3563 * config/mips/mips.h (M16_STORE_REGS): New register class.
3564 (REG_CLASS_NAMES): Add M16_STORE_REGS.
3565 (REG_CLASS_CONTENTS): Likewise.
3566 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
3568 2014-04-11 Tobias Burnus <burnus@net-b.de>
3571 * doc/invoke.texi (-Wformat-signedness): Document it.
3572 (Wformat=2): Mention that this enables -Wformat-signedness.
3574 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3576 * common/config/epiphany/epiphany-common.c
3577 (epiphany_option_optimization_table): Enable section anchors by
3578 default at -O1 or higher.
3579 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
3580 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
3581 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
3582 carries no extra cost.
3583 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
3584 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
3585 * config/epiphany/predicates.md (memclob_operand): New predicate.
3586 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
3587 Use memclob_operand predicate and X constraint for operand 3.
3589 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3591 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
3592 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
3595 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3597 PR rtl-optimization/60651
3598 * mode-switching.c (optimize_mode_switching): Make sure to emit
3599 sets of a lower numbered entity before sets of a higher numbered
3600 entity to a mode of the same or lower priority.
3601 When creating a seginfo for a basic block that starts with a code
3602 label, move the insertion point past the code label.
3603 (new_seginfo): Document and enforce requirement that
3604 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
3605 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
3606 * doc/tm.texi: Regenerate.
3608 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
3611 * config/arc/arc.c (arc_save_restore): Fix assert typo.
3613 2013-04-11 Jakub Jelinek <jakub@redhat.com>
3615 * BASE-VER: Set to 4.10.0.
3617 2014-04-11 Tobias Burnus <burnus@net-b.de>
3620 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
3621 * doc/gcc.texi (Service): Update description in the @menu
3622 * doc/invoke.texi (Option Summary): Remove misplaced and
3625 2014-04-11 Steve Ellcey <sellcey@mips.com>
3626 Jakub Jelinek <jakub@redhat.com>
3629 * expr.c (convert_move): Use emit_store_flag_force instead of
3630 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
3633 2014-04-11 Richard Biener <rguenther@suse.de>
3636 * varasm.c (assemble_alias): Avoid endless error reporting
3637 recursion by setting TREE_ASM_WRITTEN.
3639 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3641 * config/s390/s390.md: Add a splitter for NOT rtx.
3643 2014-04-11 Jakub Jelinek <jakub@redhat.com>
3645 PR rtl-optimization/60663
3646 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
3648 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
3649 Jakub Jelinek <jakub@redhat.com>
3652 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
3653 flag from decl_node to node.
3655 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3658 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
3659 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
3660 ameliorating the cases where it can be.
3662 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
3665 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
3667 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3668 (loadsync_<mode>): Change mode.
3669 (load_quadpti, store_quadpti): New.
3670 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3671 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3672 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
3674 2014-04-09 Cong Hou <congh@google.com>
3677 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
3680 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3682 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
3683 instead of vnor to exploit possible fusion opportunity in the
3685 (altivec_expand_vec_perm_const_le): Likewise.
3687 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
3689 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3690 (loadsync_<mode>): Change mode.
3691 (load_quadpti, store_quadpti): New.
3692 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3693 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3695 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
3698 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
3699 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
3700 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
3702 2014-04-08 Richard Biener <rguenther@suse.de>
3705 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
3706 a 64bit widest int print double-int similar to on HWI64 hosts.
3708 2014-04-08 Richard Biener <rguenther@suse.de>
3710 PR tree-optimization/60785
3711 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
3712 default defs properly.
3714 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
3716 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
3717 (Weffc++): Likewise.
3719 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
3721 * ipa-devirt.c (maybe_record_node): When node is not recorded,
3722 set completep to false rather than true.
3724 2014-04-07 Douglas B Rupp <rupp@adacore.com>
3727 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
3728 ARM_TARGET2_DWARF_FORMAT.
3730 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
3733 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
3734 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
3737 2014-04-07 Richard Biener <rguenther@suse.de>
3739 PR tree-optimization/60766
3740 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
3741 (may_eliminate_iv): Convert cand_value_at result to desired type.
3743 2014-04-07 Jason Merrill <jason@redhat.com>
3746 * common.opt (-fno-gnu-unique): Add.
3747 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
3749 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3751 * haifa-sched.c: Fix outdated function reference and minor
3752 grammar errors in introductory comment.
3754 2014-04-07 Richard Biener <rguenther@suse.de>
3757 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
3759 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
3761 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
3764 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
3765 size accounting for thunks.
3766 (pa_asm_output_mi_thunk): Use final_start_function() and
3767 final_end_function() to output function start and end directives.
3769 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3771 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
3772 device specific ISA/ feature information. Remove short_sp and
3773 errata_skip ds. Add avr_device_specific_features enum to have device
3775 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
3776 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
3777 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
3778 updated device specific info.
3779 * config/avr/avr-mcus.def: Merge device specific details to
3780 dev_attribute field.
3781 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
3783 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
3784 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
3785 assembler if RMW isa supported by current device.
3786 * config/avr/genmultilib.awk: Update as device info structure changed.
3787 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
3789 2014-04-04 Cong Hou <congh@google.com>
3791 PR tree-optimization/60656
3792 * tree-vect-stmts.c (supportable_widening_operation):
3793 Fix a bug that elements in a vector with vect_used_by_reduction
3794 property are incorrectly reordered when the operation on it is not
3795 consistant with the one in reduction operation.
3797 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
3799 PR rtl-optimization/60155
3800 * gcse.c (record_set_data): New function.
3801 (single_set_gcse): New function.
3802 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
3803 (hoist_code): Likewise.
3804 (get_pressure_class_and_nregs): Likewise.
3806 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
3808 * explow.c (probe_stack_range): Emit a final optimization blockage.
3810 2014-04-04 Anthony Green <green@moxielogic.com>
3812 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
3815 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
3818 * lto-cgraph.c (input_overwrite_node): Check that partitioning
3819 flags are set only during streaming.
3820 * ipa.c (process_references, walk_polymorphic_call_targets,
3821 symtab_remove_unreachable_nodes): Drop bodies of always inline
3822 after early inlining.
3823 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
3825 2014-04-04 Jakub Jelinek <jakub@redhat.com>
3826 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3829 * dwarf2out.c (const_ok_for_output_1): Reject expressions
3832 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3835 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
3837 (cortex_a53_fdivd): Likewise.
3839 2014-04-04 Martin Jambor <mjambor@suse.cz>
3842 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
3844 * cgraph.c (clone_of_p): Also return true if thunks match.
3845 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
3846 cgraph_function_or_thunk_node and an obsolete comment.
3847 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
3849 (build_function_decl_skip_args): Likewise.
3850 (set_new_clone_decl_and_node_flags): New function.
3851 (duplicate_thunk_for_node): Likewise.
3852 (redirect_edge_duplicating_thunks): Likewise.
3853 (cgraph_clone_node): New parameter args_to_skip, pass it to
3854 redirect_edge_duplicating_thunks which is called instead of
3855 cgraph_redirect_edge_callee.
3856 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
3857 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
3859 2014-04-04 Jeff Law <law@redhat.com>
3862 * config/arm/predicates.md (const_int_I_operand): New predicate.
3863 (const_int_M_operand): Similarly.
3864 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
3866 (insv_t2, extv_reg, extzv_t2): Likewise.
3867 (load_multiple_with_writeback): Similarly for const_int_I_operand.
3868 (pop_multiple_with_writeback_and_return): Likewise.
3869 (vfp_pop_multiple_with_writeback): Likewise
3871 2014-04-04 Richard Biener <rguenther@suse.de>
3874 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
3875 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
3877 * gimplify.h (gimple_add_tmp_var_fn): Declare.
3878 * gimplify.c (gimple_add_tmp_var_fn): New function.
3879 * gimple-expr.h (create_tmp_reg_fn): Declare.
3880 * gimple-expr.c (create_tmp_reg_fn): New function.
3881 * gimple-low.c (record_vars_into): Don't change cfun.
3882 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
3883 code generation without cfun.
3885 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
3888 * Makefile.in (install-driver): Fix shell scripting.
3890 2014-04-03 Cong Hou <congh@google.com>
3892 PR tree-optimization/60505
3893 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
3894 threshold of number of iterations below which no vectorization
3896 * tree-vect-loop.c (new_loop_vec_info):
3897 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
3898 * tree-vect-loop.c (vect_analyze_loop_operations):
3899 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
3900 * tree-vect-loop.c (vect_transform_loop):
3901 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
3902 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
3903 of iterations of the loop and see if we should build the epilogue.
3905 2014-04-03 Richard Biener <rguenther@suse.de>
3907 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
3908 (streamer_tree_cache_create): Adjust.
3909 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
3910 to allow optional nodes array.
3911 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
3912 (streamer_tree_cache_append): Likewise.
3913 (streamer_tree_cache_create): Create nodes array optionally
3914 as specified by parameter.
3915 * lto-streamer-out.c (create_output_block): Avoid maintaining
3916 the node array in the writer cache.
3917 (DFS_write_tree): Remove assertion.
3918 (produce_asm_for_decls): Free the out decl state hash table early.
3919 * lto-streamer-in.c (lto_data_in_create): Adjust for
3920 streamer_tree_cache_create prototype change.
3922 2014-04-03 Richard Biener <rguenther@suse.de>
3924 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
3925 set TREE_CHAIN to NULL_TREE.
3927 2014-04-03 Richard Biener <rguenther@suse.de>
3929 PR tree-optimization/60740
3930 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
3931 over all GIMPLE_COND operands.
3933 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
3935 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
3936 (Weffc++): Remove Scott's numbering, merge lists and reference
3939 2014-04-03 Nick Clifton <nickc@redhat.com>
3941 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
3944 2014-04-03 Martin Jambor <mjambor@suse.cz>
3946 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
3947 mention gcc_unreachable before failing.
3948 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
3951 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
3954 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
3955 inconsistent code and instead mark the context inconsistent.
3956 (possible_polymorphic_call_targets): For inconsistent contexts
3957 return empty complete list.
3959 2014-04-02 Anthony Green <green@moxielogic.com>
3961 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
3962 (extendqisi2, extendhisi2): Define.
3963 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
3964 (WCHAR_TYPE): Change to unsigned int.
3966 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3968 PR tree-optimization/60733
3969 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
3970 insertion point for PHI candidates to be the end of the feeding
3971 block for the PHI argument.
3973 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
3975 PR rtl-optimization/60650
3976 * lra-constraints.c (process_alt_operands): Decrease reject for
3977 earlyclobber matching.
3979 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3981 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
3983 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3985 * config/spu/spu.c (pad_bb): Do not crash when the last
3986 insn is CODE_FOR_blockage.
3988 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3990 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
3991 lies outside the target mode.
3993 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3996 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3997 software floating point or no floating point registers, do not
3998 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3999 in GPRs that occurs after we tested for GPRs that would never be
4002 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4003 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4004 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4005 specifically allow DDmode, since that does not use the SPE SIMD
4008 2014-04-02 Richard Biener <rguenther@suse.de>
4011 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4012 MODE_INTs. Properly use negv_optab.
4013 (expand_abs): Likewise.
4015 2014-04-02 Richard Biener <rguenther@suse.de>
4018 * Makefile.in (install-driver): Guard extra installs with special
4021 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4023 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4026 2014-04-01 Richard Henderson <rth@redhat.com>
4029 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
4030 alternative enabled before register allocation.
4032 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
4034 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
4035 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
4037 (nios2_large_got_address): Remove unneeded 'sym' parameter.
4038 (nios2_got_address): Update nios2_large_got_address call site.
4039 (nios2_delegitimize_address): New function.
4040 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
4041 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
4042 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
4044 2014-04-01 Martin Husemann <martin@duskware.de>
4046 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
4049 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
4051 PR rtl-optimization/60604
4052 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
4053 check from register_operand.
4054 (register_operand): Redefine in terms of general_operand.
4055 (nonmemory_operand): Use register_operand for the non-constant cases.
4057 2014-04-01 Richard Biener <rguenther@suse.de>
4059 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
4061 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
4063 * doc/invoke.texi (mapp-regs): Clarify.
4065 2014-03-31 Ulrich Drepper <drepper@gmail.com>
4067 * config/i386/avx512fintrin.h (__v32hi): Define type.
4068 (__v64qi): Likewise.
4069 (_mm512_set1_epi8): Define.
4070 (_mm512_set1_epi16): Define.
4071 (_mm512_set4_epi32): Define.
4072 (_mm512_set4_epi64): Define.
4073 (_mm512_set4_pd): Define.
4074 (_mm512_set4_ps): Define.
4075 (_mm512_setr4_epi64): Define.
4076 (_mm512_setr4_epi32): Define.
4077 (_mm512_setr4_pd): Define.
4078 (_mm512_setr4_ps): Define.
4079 (_mm512_setzero_epi32): Define.
4081 2014-03-31 Martin Jambor <mjambor@suse.cz>
4084 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
4085 callsite_arguments_match_p. Updated all callers. Also check types of
4086 corresponding formal parameters and actual arguments.
4087 (not_all_callers_have_enough_arguments_p) Renamed to
4088 some_callers_have_mismatched_arguments_p.
4090 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
4092 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
4094 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
4097 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
4100 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
4102 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
4103 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
4105 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
4106 Use FMAMODE_NOVF512 mode iterator.
4107 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
4108 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
4109 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
4110 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
4112 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
4113 Use VF_128_256 mode iterator.
4114 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
4117 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4119 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
4120 static chain if needed.
4122 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4125 * lra-constraints.c (index_part_to_reg): New.
4126 (process_address): Use it.
4128 2014-03-27 Jeff Law <law@redhat.com>
4129 Jakub Jelinek <jakub@redhat.com>
4132 * expr.c (do_tablejump): Use simplify_gen_binary rather than
4133 gen_rtx_{PLUS,MULT} to build up the address expression.
4135 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
4136 creating non-canonical RTL.
4138 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4141 * ipa-inline.c (want_inline_small_function_p): Short circuit large
4142 functions; reorganize to make cheap checks first.
4143 (inline_small_functions): Do not estimate growth when dumping;
4145 * ipa-inline.h (inline_summary): Add min_size.
4146 (growth_likely_positive): New function.
4147 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
4148 (set_cond_stmt_execution_predicate): Cleanup.
4149 (estimate_edge_size_and_time): Compute min_size.
4150 (estimate_calls_size_and_time): Likewise.
4151 (estimate_node_size_and_time): Likewise.
4152 (inline_update_overall_summary): Update min_size.
4153 (do_estimate_edge_time): Likewise.
4154 (do_estimate_edge_size): Update.
4155 (do_estimate_edge_hints): Update.
4156 (growth_likely_positive): New function.
4158 2014-03-28 Jakub Jelinek <jakub@redhat.com>
4161 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
4162 also if addr has VOIDmode.
4164 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4166 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
4167 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
4169 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
4170 instructions as well as AdvancedSIMD loads.
4172 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4174 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
4175 Use crypto_aese type.
4176 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
4177 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
4178 crypto_aese, crypto_aesmc. Move to types.md.
4179 * config/arm/types.md (crypto_aes): Split into crypto_aese,
4181 * config/arm/iterators.md (crypto_type): Likewise.
4183 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4185 * cgraph.c: Include expr.h and tree-dfa.h.
4186 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
4189 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4192 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
4193 regs from checking multi-reg pseudos.
4195 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4197 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
4199 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4201 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
4202 if it would clobber the stack pointer, even temporarily.
4204 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
4206 * mode-switching.c: Make small adjustments to the top comment.
4208 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4210 * config/rs6000/constraints.md (wD constraint): New constraint to
4211 match the constant integer to get the top DImode/DFmode out of a
4212 vector in a VSX register.
4214 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
4215 match the constant integer to get the top DImode/DFmode out of a
4216 vector in a VSX register.
4218 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
4221 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
4224 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
4225 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
4227 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
4228 Optimize vec_extract of 64-bit values, where the value being
4229 extracted is in the top word, where we can use scalar
4230 instructions. Add direct move and store support. Combine the big
4231 endian/little endian vector select load support into a single insn.
4232 (vsx_extract_<mode>_internal1): Likewise.
4233 (vsx_extract_<mode>_internal2): Likewise.
4234 (vsx_extract_<mode>_load): Likewise.
4235 (vsx_extract_<mode>_store): Likewise.
4236 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
4237 combined into vsx_extract_<mode>_load.
4238 (vsx_extract_<mode>_one_le): Likewise.
4240 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
4241 define the top 64-bit vector element.
4243 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
4246 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4247 Document vec_vbpermq builtin.
4250 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
4251 enable use of xxsldwi and xxpermdi builtin functions.
4252 (vec_xxpermdi): Likewise.
4254 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4255 Document use of vec_xxsldwi and vec_xxpermdi builtins.
4257 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
4259 PR rtl-optimization/60650
4260 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
4262 (find_spills_for): New.
4263 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
4264 Spill all pseudos on the second iteration.
4266 2014-03-27 Marek Polacek <polacek@redhat.com>
4269 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
4272 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4274 * config/s390/s390.c (s390_can_use_return_insn): Check for
4275 call-saved FPRs on 31 bit.
4277 2014-03-27 Jakub Jelinek <jakub@redhat.com>
4280 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
4281 if they need regimplification, just drop them instead of
4282 calling gimple_regimplify_operands on them.
4284 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
4287 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
4288 (aarch64_frame_pointer_required): Adjust logic.
4289 (aarch64_can_eliminate): Adjust logic.
4290 (aarch64_override_options_after_change): Adjust logic.
4292 2014-03-27 Dehao Chen <dehao@google.com>
4294 * ipa-inline.c (early_inliner): Update node's inline info.
4296 2014-03-26 Dehao Chen <dehao@google.com>
4298 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
4299 compiler inserted conditional jumps for NAN float check.
4301 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4303 * ubsan.h (ubsan_create_data): Change second argument's type
4304 to const location_t *.
4305 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
4307 (ubsan_create_data): Change second argument to const location_t *PLOC.
4308 Create Loc field whenever PLOC is non-NULL.
4309 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
4310 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
4314 * real.c (real_to_integer2): Change type of low to UHWI.
4316 2014-03-26 Tobias Burnus <burnus@net-b.de>
4318 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
4319 (CILK_SELF_SPECS): New define.
4320 (driver_self_specs): Use it.
4322 2014-03-26 Richard Biener <rguenther@suse.de>
4324 * tree-pretty-print.c (percent_K_format): Implement special
4325 case for LTO and its stripped down BLOCK tree.
4327 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4330 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
4332 * tree-vrp.c (simplify_internal_call_using_ranges): If only
4333 one range is range_int_cst_p, but not both, at least optimize
4334 addition/subtraction of 0 and multiplication by 0 or 1.
4335 * gimple-fold.c (gimple_fold_call): Fold
4336 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
4337 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
4338 INTEGER_CSTs, try to fold at least x * 0 and y - y.
4340 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
4342 PR rtl-optimization/60452
4343 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
4344 <case REG>: Return 1 for invalid offsets from the frame pointer.
4346 2014-03-26 Marek Polacek <polacek@redhat.com>
4349 * doc/extend.texi (C Extensions): Mention variable-length arrays in
4352 2014-03-26 Marek Polacek <polacek@redhat.com>
4355 * doc/extend.texi (Designated Inits): Describe what happens to omitted
4358 2014-03-26 Marek Polacek <polacek@redhat.com>
4361 * ira-color.c (update_conflict_hard_regno_costs): Perform the
4362 multiplication in unsigned type.
4364 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4366 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
4368 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4370 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
4372 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4375 * cif-code.def (UNREACHABLE) New code.
4376 * ipa-inline.c (inline_small_functions): Skip edges to
4377 __builtlin_unreachable.
4378 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
4379 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
4380 predicate to __bulitin_unreachable.
4381 (set_cond_stmt_execution_predicate): Fix issue when
4382 invert_tree_comparison returns ERROR_MARK.
4383 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
4384 propagate to inline clones.
4385 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
4387 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
4388 * cgraphclones.c (cgraph_clone_node): If call destination is already
4389 ureachable, do not redirect it back.
4390 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
4393 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4395 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
4396 Do not modify inline clones.
4398 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4400 * config/i386/i386.md (general_sext_operand): New mode attr.
4401 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
4402 don't generate (sign_extend (const_int)).
4403 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
4404 operands[2]. Use We constraint instead of <i> and
4405 <general_sext_operand> predicate instead of <general_operand>.
4406 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
4407 * config/i386/constraints.md (We): New constraint.
4408 * config/i386/predicates.md (x86_64_sext_operand,
4409 sext_operand): New predicates.
4411 2014-03-25 Martin Jambor <mjambor@suse.cz>
4414 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
4415 inconsistent devirtualizations to __builtin_unreachable.
4417 2014-03-25 Marek Polacek <polacek@redhat.com>
4420 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
4422 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
4424 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
4425 order of elements for big-endian.
4427 2014-03-25 Richard Biener <rguenther@suse.de>
4430 * gimplify-me.c (gimple_regimplify_operands): Update the
4433 2014-03-25 Martin Jambor <mjambor@suse.cz>
4436 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
4437 (lto_output_varpool_node): Likewise.
4438 (input_overwrite_node): Likewise.
4439 (input_varpool_node): Likewise.
4441 2014-03-25 Richard Biener <rguenther@suse.de>
4443 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
4444 (run_gcc): Likewise.
4446 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4448 * combine.c (simplify_compare_const): Add MODE argument.
4449 Handle mode_width 0 as very large mode_width.
4450 (try_combine, simplify_comparison): Adjust callers.
4452 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
4453 type to avoid signed integer overflow.
4454 * explow.c (plus_constant): Likewise.
4456 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
4458 * doc/generic.texi: Correct typos.
4460 2014-03-24 Tobias Burnus <burnus@net-b.de>
4462 * doc/invoke.texi (-flto): Expand section about
4463 using static libraries with LTO.
4465 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4467 PR rtl-optimization/60501
4468 * optabs.def (addptr3_optab): New optab.
4469 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
4470 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
4471 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
4473 * lra.c (emit_add3_insn): Use the addptr pattern if available.
4475 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
4477 2014-03-24 Ulrich Drepper <drepper@gmail.com>
4479 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
4482 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
4483 (_mm256_undefined_ps): Define.
4484 (_mm256_undefined_pd): Define.
4485 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
4486 (_mm_undefined_pd): Define.
4487 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
4488 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
4489 (_mm512_undefined_ps): Define.
4490 (_mm512_undefined_pd): Define.
4491 Use _mm*_undefined_*.
4492 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
4494 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
4496 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
4497 (lshr_simd): DI mode added.
4498 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
4499 (aarch64_ushr_simddi): Likewise.
4500 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
4501 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
4502 (vshrd_n_u64): Likewise.
4504 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4506 * Makefile.in (s-macro_list): Depend on cc1.
4508 2014-03-23 Teresa Johnson <tejohnson@google.com>
4510 * ipa-utils.c (ipa_print_order): Use specified dump file.
4512 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
4514 PR rtl-optimization/60601
4515 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
4517 * gcc.c (eval_spec_function): Initialize save_growing_value.
4519 2014-03-22 Jakub Jelinek <jakub@redhat.com>
4522 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
4523 code == MINUS_EXPR, never swap op0 with op1.
4525 * toplev.c (init_local_tick): Avoid signed integer multiplication
4527 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
4528 shift by first operand's bitsize.
4530 2014-03-21 Jakub Jelinek <jakub@redhat.com>
4533 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
4535 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
4536 TARGET_ISA_64BIT_P(x).
4538 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4540 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
4541 pattern for vector nor instead of subtract from splat(-1).
4542 (altivec_expand_vec_perm_const_le): Likewise.
4544 2014-03-21 Richard Henderson <rth@twiddle.net>
4547 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
4548 related insns after epilogue_completed.
4550 2014-03-21 Martin Jambor <mjambor@suse.cz>
4553 * cgraph.h (symtab_node): New flag body_removed.
4554 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
4555 when removing bodies.
4556 * symtab.c (dump_symtab_base): Dump body_removed flag.
4557 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
4558 had their bodies removed.
4560 2014-03-21 Martin Jambor <mjambor@suse.cz>
4563 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
4566 2014-03-21 Richard Biener <rguenther@suse.de>
4568 PR tree-optimization/60577
4569 * tree-core.h (struct tree_base): Document nothrow_flag use
4570 in VAR_DECL_NONALIASED.
4571 * tree.h (VAR_DECL_NONALIASED): New.
4572 (may_be_aliased): Adjust.
4573 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
4575 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4577 * expr.c (expand_expr_real_1): Remove outdated comment.
4579 2014-03-20 Jakub Jelinek <jakub@redhat.com>
4582 * ira.c (adjust_cleared_regs): Call copy_rtx on
4583 *reg_equiv[REGNO (loc)].src_p before passing it to
4584 simplify_replace_fn_rtx.
4587 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
4588 into CONST, put pic register as first operand of PLUS. Use
4589 gen_const_mem for both 32-bit and 64-bit PIC got loads.
4591 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4593 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
4595 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4597 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
4598 around for store forwarding issue in the FPU on the UT699.
4599 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
4600 loads and operations if -mfix-ut699 is specified.
4601 (divtf3_hq): Tweak attribute.
4602 (sqrttf2_hq): Likewise.
4604 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4606 * calls.c (store_one_arg): Remove incorrect const qualification on the
4607 type of the temporary.
4608 * cfgexpand.c (expand_return): Likewise.
4609 * expr.c (expand_constructor): Likewise.
4610 (expand_expr_real_1): Likewise.
4612 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4614 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
4617 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
4620 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
4622 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
4624 * config/arm/aarch-common-protos.h
4625 (alu_cost_table): Fix spelling of "extend".
4626 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
4628 2014-03-19 Richard Biener <rguenther@suse.de>
4631 * tree-core.h (tree_type_common): Re-order pointer members
4632 to reduce recursion depth during GC walks.
4634 2014-03-19 Marek Polacek <polacek@redhat.com>
4637 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
4638 before accessing it.
4640 2014-03-19 Richard Biener <rguenther@suse.de>
4643 * lto-streamer-in.c (input_function): In WPA stage do not drop
4646 2014-03-19 Jakub Jelinek <jakub@redhat.com>
4648 PR tree-optimization/60559
4649 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
4650 with build_zero_cst assignment.
4652 2014-03-18 Kai Tietz <ktietz@redhat.com>
4654 PR rtl-optimization/56356
4655 * sdbout.c (sdbout_parms): Verify that parms'
4656 incoming argument is valid.
4657 (sdbout_reg_parms): Likewise.
4659 2014-03-18 Richard Henderson <rth@redhat.com>
4662 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
4663 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
4664 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
4666 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
4668 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
4669 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
4670 Italicize plugin event names in description. Explain that
4671 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
4672 Remind that no GCC functions should be called after PLUGIN_FINISH.
4673 Explain what pragmas with expansion are.
4675 2014-03-18 Martin Liska <mliska@suse.cz>
4677 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
4678 gimple call statement is update.
4679 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
4680 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
4682 2014-03-18 Jakub Jelinek <jakub@redhat.com>
4685 * ubsan.c (ubsan_instrument_unreachable): Call
4686 initialize_sanitizer_builtins.
4687 (ubsan_pass): Likewise.
4690 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
4691 varpool_finalize_decl instead of rest_of_decl_compilation.
4693 2014-03-18 Richard Biener <rguenther@suse.de>
4695 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
4696 by using bitmap_and_compl instead of bitmap_and_compl_into.
4697 (df_rd_transfer_function): Likewise.
4699 2014-03-18 Richard Biener <rguenther@suse.de>
4701 * doc/lto.texi (fresolution): Fix typo.
4703 2014-03-18 Richard Biener <rguenther@suse.de>
4705 * doc/invoke.texi (flto): Update for changes in 4.9.
4707 2014-03-18 Richard Biener <rguenther@suse.de>
4709 * doc/loop.texi: Remove section on the removed lambda framework.
4710 Update loop docs with recent changes in preserving loop structure.
4712 2014-03-18 Richard Biener <rguenther@suse.de>
4714 * doc/lto.texi (-fresolution): Document.
4716 2014-03-18 Richard Biener <rguenther@suse.de>
4718 * doc/contrib.texi: Adjust my name.
4720 2014-03-18 Jakub Jelinek <jakub@redhat.com>
4723 * internal-fn.c: Include diagnostic-core.h.
4724 (expand_BUILTIN_EXPECT): New function.
4725 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
4726 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
4727 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
4728 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
4730 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
4731 Revert 3 argument __builtin_expect code.
4732 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
4733 * gimple-fold.c (gimple_fold_call): Likewise.
4734 * tree.h (fold_builtin_expect): New prototype.
4735 * builtins.c (build_builtin_expect_predicate): Add predictor
4736 argument, if non-NULL, create 3 argument __builtin_expect.
4737 (fold_builtin_expect): No longer static. Add ARG2 argument,
4738 pass it through to build_builtin_expect_predicate.
4739 (fold_builtin_2): Adjust caller.
4740 (fold_builtin_3): Handle BUILT_IN_EXPECT.
4741 * internal-fn.def (BUILTIN_EXPECT): New.
4743 2014-03-18 Tobias Burnus <burnus@net-b.de>
4746 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
4747 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
4748 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
4750 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
4753 * predict.c (combine_predictions_for_bb): Fix up formatting.
4754 (expr_expected_value_1, expr_expected_value): Add predictor argument,
4755 fill what it points to if non-NULL.
4756 (tree_predict_by_opcode): Adjust caller, use the predictor.
4757 * predict.def (PRED_COMPARE_AND_SWAP): Add.
4759 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
4761 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
4762 proper constant for the store mode.
4764 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
4766 * symtab.c (change_decl_assembler_name): Fix transparent alias
4769 2014-03-16 Renlin Li <Renlin.Li@arm.com>
4771 * config/aarch64/aarch64.c: Correct the comments about the
4772 aarch64 stack layout.
4774 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
4776 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
4777 check for GF_OMP_FOR_KIND_FOR.
4779 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
4781 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
4782 ymm and zmm register names.
4784 2014-03-17 Jakub Jelinek <jakub@redhat.com>
4787 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
4788 note creation for the 2010-08-31 changes.
4790 2014-03-17 Marek Polacek <polacek@redhat.com>
4793 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
4794 as -fno-tree-loop-vectorize.
4795 (expand_omp_simd): Likewise.
4797 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
4799 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
4800 (eligible_for_call_delay): New prototype.
4801 * config/sparc/sparc.c (tls_call_delay): Rename into...
4802 (eligible_for_call_delay): ...this. Return false if the instruction
4803 cannot be put in the delay slot of a branch.
4804 (eligible_for_restore_insn): Simplify.
4805 (eligible_for_return_delay): Return false if the instruction cannot be
4806 put in the delay slot of a branch and simplify.
4807 (eligible_for_sibcall_delay): Return false if the instruction cannot be
4808 put in the delay slot of a branch.
4809 * config/sparc/sparc.md (fix_ut699): New attribute.
4810 (tls_call_delay): Delete.
4811 (in_call_delay): Reimplement.
4812 (eligible_for_sibcall_delay): Rename into...
4813 (in_sibcall_delay): ...this.
4814 (eligible_for_return_delay): Rename into...
4815 (in_return_delay): ...this.
4816 (in_branch_delay): Reimplement.
4817 (in_uncond_branch_delay): Delete.
4818 (in_annul_branch_delay): Delete.
4820 2014-03-14 Richard Henderson <rth@redhat.com>
4823 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
4824 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
4825 (*floathi<X87MODEF>2_i387_with_temp): Remove.
4826 (floathi splitters): Remove.
4827 (float<SWI48x>xf2): New pattern.
4828 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
4829 code that tried to handle DImode for 32-bit, but which was excluded
4830 by the pattern's condition. Drop allocation of stack temporary.
4831 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
4832 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
4833 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
4834 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
4835 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
4836 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
4837 (*float<SWI48><MODEF>2_sse_interunit): Remove.
4838 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
4839 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
4840 (*float<SWI48x><X87MODEF>2_i387): Remove.
4841 (all float _with_temp splitters): Remove.
4842 (*float<SWI48x><MODEF>2_i387): New pattern.
4843 (*float<SWI48><MODEF>2_sse): New pattern.
4844 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
4845 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
4847 2014-03-14 Jakub Jelinek <jakub@redhat.com>
4848 Marek Polacek <polacek@redhat.com>
4851 * common.opt (dump_base_name_prefixed): New Variable.
4852 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
4853 if x_dump_base_name_prefixed is already set, set it at the end.
4855 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
4857 PR rtl-optimization/60508
4858 * lra-constraints.c (get_reload_reg): Add new parameter
4860 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
4861 Pass the new parameter values.
4863 2014-03-14 Richard Biener <rguenther@suse.de>
4865 * common.opt: Revert unintented changes from r205065.
4868 2014-03-14 Richard Biener <rguenther@suse.de>
4871 * cfghooks.c (split_block): Properly adjust all loops the
4872 block was a latch of.
4874 2014-03-14 Martin Jambor <mjambor@suse.cz>
4877 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
4880 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
4883 * config/avr/avr.c (avr_set_current_function): Pass function name
4884 through default_strip_name_encoding before sanity checking instead
4885 of skipping the first char of the assembler name.
4887 2014-03-13 Richard Henderson <rth@redhat.com>
4890 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
4891 (ix86_force_to_memory, ix86_free_from_memory): Remove.
4892 * config/i386/i386-protos.h: Likewise.
4893 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
4894 in the expander instead of a splitter.
4895 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
4896 any possibility of requiring a memory.
4897 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
4898 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
4899 (fp branch splitters): Update for ix86_split_fp_branch.
4900 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
4901 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
4902 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
4903 (*fop_<MODEF>_2_i387): Remove f/r alternative.
4904 (*fop_<MODEF>_3_i387): Likewise.
4905 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
4906 (splitters for the fop_* register patterns): Remove.
4907 (fscalexf4_i387): Rename from *fscalexf4_i387.
4908 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
4910 2014-03-13 Jakub Jelinek <jakub@redhat.com>
4912 PR tree-optimization/59779
4913 * tree-dfa.c (get_ref_base_and_extent): Use double_int
4914 type for bitsize and maxsize instead of HOST_WIDE_INT.
4916 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
4918 PR rtl-optimization/57320
4919 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
4920 the CFG after thread_prologue_and_epilogue_insns.
4922 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
4924 PR rtl-optimization/57189
4925 * lra-constraints.c (process_alt_operands): Disfavor spilling
4928 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
4930 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
4932 2014-03-13 Jakub Jelinek <jakub@redhat.com>
4934 PR tree-optimization/59025
4936 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
4937 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
4939 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
4942 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
4943 calls of avr_out_plus_1.
4945 2014-03-13 Bin Cheng <bin.cheng@arm.com>
4947 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
4948 BB's single pred and update the father loop's latch info later.
4950 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
4952 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
4956 (VEC_base): Likewise.
4957 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
4958 registers, we need to swap double words in little endian mode.
4960 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
4961 to be a container mode for 128-bit integer operations added in ISA
4962 2.07. Unlike TImode and PTImode, the preferred register set is
4963 the Altivec/VMX registers for the 128-bit operations.
4965 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
4967 (rs6000_split_128bit_ok_p): Likewise.
4969 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
4970 macros for creating ISA 2.07 normal and overloaded builtin
4971 functions with 3 arguments.
4972 (BU_P8V_OVERLOAD_3): Likewise.
4973 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
4974 for use as overloaded functions.
4975 (VPERM_1TI_UNS): Likewise.
4976 (VSEL_1TI): Likewise.
4977 (VSEL_1TI_UNS): Likewise.
4978 (ST_INTERNAL_1ti): Likewise.
4979 (LD_INTERNAL_1ti): Likewise.
4980 (XXSEL_1TI): Likewise.
4981 (XXSEL_1TI_UNS): Likewise.
4982 (VPERM_1TI): Likewise.
4983 (VPERM_1TI_UNS): Likewise.
4984 (XXPERMDI_1TI): Likewise.
4985 (SET_1TI): Likewise.
4986 (LXVD2X_V1TI): Likewise.
4987 (STXVD2X_V1TI): Likewise.
4988 (VEC_INIT_V1TI): Likewise.
4989 (VEC_SET_V1TI): Likewise.
4990 (VEC_EXT_V1TI): Likewise.
4991 (EQV_V1TI): Likewise.
4992 (NAND_V1TI): Likewise.
4993 (ORC_V1TI): Likewise.
4994 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
4995 added in ISA 2.07. Add both normal 'altivec' builtins, and the
4997 (VADDUQM): Likewise.
4998 (VSUBCUQ): Likewise.
4999 (VADDEUQM): Likewise.
5000 (VADDECUQ): Likewise.
5001 (VSUBEUQM): Likewise.
5002 (VSUBECUQ): Likewise.
5004 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5005 __int128_t and __uint128_t types.
5006 (__uint128_type): Likewise.
5007 (altivec_categorize_keyword): Add support for vector __int128_t,
5008 vector __uint128_t, vector __int128, and vector unsigned __int128
5009 as a container type for TImode operations that need to be done in
5010 VSX/Altivec registers.
5011 (rs6000_macro_to_expand): Likewise.
5012 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5013 to support 128-bit integer instructions vaddcuq, vadduqm,
5014 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5015 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5017 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5018 for V1TImode, and set up preferences to use VSX/Altivec registers.
5019 Setup VSX reload handlers.
5020 (rs6000_debug_reg_global): Likewise.
5021 (rs6000_init_hard_regno_mode_ok): Likewise.
5022 (rs6000_preferred_simd_mode): Likewise.
5023 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5024 (easy_altivec_constant): Likewise.
5025 (output_vec_const_move): Likewise.
5026 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5028 (rs6000_expand_vector_extract): Likewise.
5029 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
5031 (rs6000_const_vec): Add support for V1TImode.
5032 (rs6000_emit_le_vsx_load): Swap double words when loading or
5033 storing TImode/V1TImode.
5034 (rs6000_emit_le_vsx_store): Likewise.
5035 (rs6000_emit_le_vsx_move): Likewise.
5036 (rs6000_emit_move): Add support for V1TImode.
5037 (altivec_expand_ld_builtin): Likewise.
5038 (altivec_expand_st_builtin): Likewise.
5039 (altivec_expand_vec_init_builtin): Likewise.
5040 (altivec_expand_builtin): Likewise.
5041 (rs6000_init_builtins): Add support for V1TImode type. Add
5042 support for ISA 2.07 128-bit integer builtins. Define type names
5043 for the VSX/Altivec vector types.
5044 (altivec_init_builtins): Add support for overloaded vector
5045 functions with V1TImode type.
5046 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
5047 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
5049 (rs6000_split_128bit_ok_p): Likewise.
5050 (rs6000_handle_altivec_attribute): Create V1TImode from vector
5051 __int128_t and vector __uint128_t.
5053 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
5054 and mode attributes.
5061 (VS_scalar): Likewise.
5062 (VS_double): Likewise.
5063 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
5065 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
5066 we support the ISA 2.07 128-bit integer arithmetic instructions.
5067 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
5068 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
5069 and TImode types for use with the builtin functions.
5070 (V1TI_type_node): Likewise.
5071 (unsigned_V1TI_type_node): Likewise.
5072 (intTI_type_internal_node): Likewise.
5073 (uintTI_type_internal_node): Likewise.
5075 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
5076 128-bit builtin functions.
5077 (UNSPEC_VADDEUQM): Likewise.
5078 (UNSPEC_VADDECUQ): Likewise.
5079 (UNSPEC_VSUBCUQ): Likewise.
5080 (UNSPEC_VSUBEUQM): Likewise.
5081 (UNSPEC_VSUBECUQ): Likewise.
5082 (VM): Add V1TImode to vector mode iterators.
5084 (VI_unit): Likewise.
5085 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
5086 (altivec_vaddcuq): Likewise.
5087 (altivec_vsubuqm): Likewise.
5088 (altivec_vsubcuq): Likewise.
5089 (altivec_vaddeuqm): Likewise.
5090 (altivec_vaddecuq): Likewise.
5091 (altivec_vsubeuqm): Likewise.
5092 (altivec_vsubecuq): Likewise.
5094 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
5096 (BOOL_128): Likewise.
5097 (BOOL_REGS_OUTPUT): Likewise.
5098 (BOOL_REGS_OP1): Likewise.
5099 (BOOL_REGS_OP2): Likewise.
5100 (BOOL_REGS_UNARY): Likewise.
5101 (BOOL_REGS_AND_CR0): Likewise.
5103 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
5104 128-bit integer builtin support.
5105 (vec_vadduqm): Likewise.
5106 (vec_vaddecuq): Likewise.
5107 (vec_vaddeuqm): Likewise.
5108 (vec_vsubecuq): Likewise.
5109 (vec_vsubeuqm): Likewise.
5110 (vec_vsubcuq): Likewise.
5111 (vec_vsubuqm): Likewise.
5113 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5114 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
5115 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
5116 128-bit integer add/subtract to ISA 2.07.
5118 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
5120 * config/arc/arc.c (arc_predicate_delay_insns):
5121 Fix third argument passed to conditionalize_nonjump.
5123 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
5125 * config/aarch64/aarch64-builtins.c
5126 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
5127 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
5128 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
5129 instead of __builtin_lfloor.
5130 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
5132 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5134 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
5135 (tree_ssa_ifcombine_bb_1): New function.
5136 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
5137 is an empty forwarder block to then_bb or vice versa and then_bb
5138 and else_bb are effectively swapped.
5140 2014-03-12 Christian Bruel <christian.bruel@st.com>
5143 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
5144 REG_CFA_DEF_CFA note.
5145 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
5146 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
5148 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
5150 PR tree-optimization/60454
5151 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
5153 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5155 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
5156 Do not define target_cpu_default2 to generic.
5157 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
5158 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
5159 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
5161 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5162 Marc Glisse <marc.glisse@inria.fr>
5164 PR tree-optimization/60502
5165 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
5166 instead of build_low_bits_mask.
5168 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5171 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
5172 if there are multiple uses, but op doesn't live on E edge.
5173 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
5174 clobber stmts before __builtin_unreachable.
5176 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
5178 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
5179 hard_frame_pointer_rtx.
5180 * cse.c (cse_insn): Remove volatile check.
5181 * cselib.c (cselib_process_insn): Likewise.
5182 * dse.c (scan_insn): Likewise.
5184 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5186 * config/arc/arc.c (conditionalize_nonjump): New function,
5188 (arc_ifcvt): ... this.
5189 (arc_predicate_delay_insns): Use it.
5191 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5193 * config/arc/predicates.md (extend_operand): During/after reload,
5194 allow const_int_operand.
5195 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
5196 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
5197 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
5199 (umulsi3_highpart_i): Likewise.
5201 2014-03-11 Richard Biener <rguenther@suse.de>
5203 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
5204 Add asserts to guard possible wrong-code bugs.
5206 2014-03-11 Richard Biener <rguenther@suse.de>
5208 PR tree-optimization/60429
5209 PR tree-optimization/60485
5210 * tree-ssa-structalias.c (set_union_with_increment): Properly
5211 take into account all fields that overlap the shifted vars.
5212 (do_sd_constraint): Likewise.
5213 (do_ds_constraint): Likewise.
5214 (get_constraint_for_ptr_offset): Likewise.
5216 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
5218 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
5219 (nios2_compute_frame_layout):
5220 Add calculation of cfun->machine->fp_save_offset.
5221 (nios2_expand_prologue): Correct setting of frame pointer register
5223 (nios2_expand_epilogue): Update recovery of stack pointer from
5224 frame pointer accordingly.
5225 (nios2_initial_elimination_offset): Update calculation of offset
5226 for eliminating to HARD_FRAME_POINTER_REGNUM.
5228 2014-03-10 Jakub Jelinek <jakub@redhat.com>
5231 * ipa.c (symtab_remove_unreachable_nodes): Don't call
5232 cgraph_get_create_node on VAR_DECLs.
5234 2014-03-10 Richard Biener <rguenther@suse.de>
5237 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
5239 2014-03-08 Douglas B Rupp <rupp@gnat.com>
5241 * config/vms/vms.opt (vms_float_format): New variable.
5243 2014-03-08 Tobias Burnus <burnus@net-b.de>
5245 * doc/invoke.texi (-fcilkplus): Update implementation status.
5247 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
5248 Richard Biener <rguenther@suse.de>
5250 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
5251 consistently accross all TUs.
5252 (run_gcc): Enable -fshort-double automatically at link at link-time
5253 and disallow override.
5255 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
5258 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
5259 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
5260 if they can't be used.
5262 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5264 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
5265 for Solaris 11/x86 ld.
5266 * configure: Regenerate.
5268 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5270 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
5271 (LIB_TLS_SPEC): Save as ld_tls_libs.
5272 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
5273 (HAVE_AS_IX86_TLSLDM): New test.
5274 * configure, config.in: Regenerate.
5275 * config/i386/i386.c (legitimize_tls_address): Fall back to
5276 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
5277 cannot support TLS_MODEL_LOCAL_DYNAMIC.
5278 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
5279 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
5281 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
5283 * common.opt (fira-loop-pressure): Mark as optimization.
5285 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
5287 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
5288 an OpenMP mappable type.
5290 2014-03-06 Matthias Klose <doko@ubuntu.com>
5292 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
5293 MULTILIB_OSDIRNAMES is not defined.
5295 2014-03-06 Jakub Jelinek <jakub@redhat.com>
5296 Meador Inge <meadori@codesourcery.com>
5299 * config/arm/arm.c (arm_tls_symbol_p): Remove.
5300 (arm_legitimize_address): Call legitimize_tls_address for any
5301 arm_tls_referenced_p expression, handle constant addend. Call it
5302 before testing for !TARGET_ARM.
5303 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
5305 2014-03-06 Richard Biener <rguenther@suse.de>
5311 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5313 * tree-streamer.c (record_common_node): Assert we don't record
5314 nodes with type double.
5315 (preload_common_node): Skip type double, complex double and double
5316 pointer since it is now frontend dependent due to fshort-double option.
5318 2014-03-06 Richard Biener <rguenther@suse.de>
5320 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
5321 or -fno-lto is specified and the linker has full plugin support.
5322 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
5323 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
5324 * lto-wrapper.c (merge_and_complain): Merge compile-time
5325 optimization levels.
5326 (run_gcc): And pass it through to the link options.
5328 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
5332 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5334 * cselib.c (remove_useless_values): Skip to avoid quadratic
5335 behavior if the condition moved from...
5336 (cselib_process_insn): ... here holds.
5338 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5341 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
5342 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
5345 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
5346 (TM_H): Add x86-tune.def.
5348 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5350 * config/aarch64/aarch64.c (generic_tunings):
5351 Use cortexa57_extra_costs.
5353 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5356 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
5357 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
5358 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
5361 2014-03-04 Heiher <r@hev.cc>
5363 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
5364 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
5366 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
5368 * config/i386/predicates.md (const2356_operand): Change to ...
5369 (const2367_operand): ... this.
5370 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
5372 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5373 (*avx512pf_scatterpf<mode>sf): Ditto.
5374 (avx512pf_scatterpf<mode>df): Ditto.
5375 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5376 (*avx512pf_scatterpf<mode>df): Ditto.
5377 * config/i386/i386.c (ix86_expand_builtin): Update
5378 incorrect hint operand error message.
5380 2014-03-04 Richard Biener <rguenther@suse.de>
5382 * lto-section-in.c (lto_get_section_data): Fix const cast.
5384 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5386 * tree-streamer.c (record_common_node): Assert we don't record
5387 nodes with type double.
5388 (preload_common_node): Skip type double, complex double and double
5389 pointer since it is now frontend dependent due to fshort-double option.
5391 2014-03-04 Richard Biener <rguenther@suse.de>
5394 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
5395 (lto_input_toplevel_asms): Likewise.
5396 * lto-section-in.c (lto_get_section_data): Instead do it here
5399 2014-03-04 Richard Biener <rguenther@suse.de>
5401 PR tree-optimization/60382
5402 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
5403 dead PHIs a reduction.
5405 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
5407 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
5409 (_mm_prefetch): Move out of GCC target("sse") pragma.
5410 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
5411 GCC target("prfchw") pragma.
5412 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
5414 * config/i386/i386.c (ix86_option_override_internal): Enable
5415 -mprfchw with -mprefetchwt1.
5417 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5419 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
5422 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5424 * opts.h (CL_PCH_IGNORE): Define.
5425 * targhooks.c (option_affects_pch_p):
5426 Return false for options that have CL_PCH_IGNORE set.
5427 * opt-functions.awk: Process PchIgnore.
5428 * doc/options.texi: Document PchIgnore.
5430 * config/arc/arc.opt (misize): Add PchIgnore property.
5432 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5434 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
5435 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
5436 constraint on constants to permit them being loaded into
5437 GENERAL_REGS or BASE_REGS.
5439 2014-03-03 Nick Clifton <nickc@redhat.com>
5441 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
5442 anti-cacnonical alternatives.
5443 (negandhi3_real): New pattern.
5444 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
5446 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5448 * config/avr/avr-mcus.def: Remove atxmega16x1.
5449 * config/avr/avr-tables.opt: Regenerate.
5450 * config/avr/t-multilib: Regenerate.
5451 * doc/avr-mmcu.texi: Regenerate.
5453 2014-03-03 Tobias Grosser <tobias@grosser.es>
5454 Mircea Namolaru <mircea.namolaru@inria.fr>
5456 PR tree-optimization/58028
5457 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
5460 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5462 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
5463 not handled by recognizers.
5465 2014-03-03 Jakub Jelinek <jakub@redhat.com>
5468 * function.c (expand_function_end): Don't emit
5469 clobber_return_register sequence if clobber_after is a BARRIER.
5470 * cfgexpand.c (construct_exit_block): Append instructions before
5471 return_label to prev_bb.
5473 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5475 * config/rs6000/constraints.md: Document reserved use of "wc".
5477 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5480 * ipa.c (function_and_variable_visibility): When dissolving comdat
5481 group, also set all symbols to local.
5483 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5488 2013-12-14 Jan Hubicka <jh@suse.cz>
5490 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
5492 2014-03-02 Jon Beniston <jon@beniston.com>
5498 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
5499 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
5500 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
5501 (simple_return, *simple_return): New patterns
5502 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
5503 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
5505 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
5507 * dwarf2out.c (gen_subprogram_die): Tidy.
5509 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
5512 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
5513 (*mov_t_msb_neg_negc): ... this new insn.
5515 2014-02-28 Jason Merrill <jason@redhat.com>
5518 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
5521 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
5524 * dwarf2out.c (decltype_auto_die): New static.
5525 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
5526 (gen_type_die_with_usage): Handle 'decltype(auto)'.
5527 (is_cxx_auto): Likewise.
5529 2014-02-28 Ian Bolton <ian.bolton@arm.com>
5531 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
5532 we are not using general regs only.
5534 2014-02-28 Richard Biener <rguenther@suse.de>
5537 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
5538 previous fix and only allow to remove trivial pre-headers
5539 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
5540 (remove_forwarder_block): Properly update the latch of a loop.
5542 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5545 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
5546 (cselib_preserved_hash_table): New.
5547 (preserve_constants_and_equivs): Move preserved vals to it.
5548 (cselib_find_slot): Look it up first.
5549 (cselib_init): Initialize it.
5550 (cselib_finish): Release it.
5551 (dump_cselib_table): Dump it.
5553 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5556 * cselib.c (remove_useless_values): Skip to avoid quadratic
5557 behavior if the condition moved from...
5558 (cselib_process_insn): ... here holds.
5560 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5563 * var-tracking.c (vt_initialize): Apply the same condition to
5564 preserve the CFA base value.
5566 2014-02-28 Joey Ye <joey.ye@arm.com>
5569 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
5570 if reload in progress or completed.
5572 2014-02-28 Tobias Burnus <burnus@net-b.de>
5575 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
5578 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
5580 * doc/tm.texi.in (Condition Code Status): Update documention for
5581 relative locations of cc0-setter and cc0-user.
5583 2014-02-27 Jeff Law <law@redhat.com>
5585 PR rtl-optimization/52714
5586 * combine.c (try_combine): When splitting an unrecognized PARALLEL
5587 into two independent simple sets, if I3 is a jump, ensure the
5588 pattern we place into I3 is a (set (pc) ...).
5590 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
5591 Jeff Law <law@redhat.com>
5593 PR rtl-optimization/49847
5594 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
5595 are in different blocks.
5596 * doc/tm.texi (Condition Code Status): Update documention for
5597 relative locations of cc0-setter and cc0-user.
5599 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
5602 * lra.c (lra_emit_add): Check SUBREG too.
5604 2014-02-27 Andreas Schwab <schwab@suse.de>
5606 * config/m68k/m68k.c (m68k_option_override): Disable
5607 -flive-range-shrinkage for classic m68k.
5608 (m68k_override_options_after_change): Likewise.
5610 2014-02-27 Marek Polacek <polacek@redhat.com>
5613 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
5614 -Wmaybe-uninitialized.
5616 2014-02-27 Alan Modra <amodra@gmail.com>
5619 * reload1.c (emit_input_reload_insns): When reload_override_in,
5620 set old to rl->in_reg when rl->in_reg is a subreg.
5622 2014-02-26 Richard Biener <rguenther@suse.de>
5625 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
5627 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
5629 * common/config/i386/predicates.md (const1256_operand): Remove.
5630 (const2356_operand): New.
5631 (const_1_to_2_operand): Remove.
5632 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5633 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5634 (*avx512pf_gatherpf<mode>sf): Ditto.
5635 (avx512pf_gatherpf<mode>df): Ditto.
5636 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5637 (*avx512pf_gatherpf<mode>df): Ditto.
5638 (avx512pf_scatterpf<mode>sf): Ditto.
5639 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5640 (*avx512pf_scatterpf<mode>sf): Ditto.
5641 (avx512pf_scatterpf<mode>df): Ditto.
5642 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5643 (*avx512pf_scatterpf<mode>df): Ditto.
5644 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
5646 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
5648 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
5649 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
5650 (_mm512_mask_testn_epi64_mask): Move to ...
5651 * config/i386/avx512cdintrin.h: Here.
5652 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
5653 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
5654 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
5655 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
5656 TARGET_AVX512F from TARGET_AVX512CD.
5658 2014-02-26 Richard Biener <rguenther@suse.de>
5661 * ipa.c (walk_polymorphic_call_targets): Properly guard
5662 call to inline_update_overall_summary.
5664 2014-02-26 Bin Cheng <bin.cheng@arm.com>
5667 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
5668 and latches only if requested. Fix latch if it is removed.
5669 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
5670 LOOPS_HAVE_PREHEADERS.
5672 2014-02-25 Andrew Pinski <apinski@cavium.com>
5674 * builtins.c (expand_builtin_thread_pointer): Create a new target
5675 when the target is NULL.
5677 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
5679 PR rtl-optimization/60317
5680 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
5681 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
5682 * lra-assigns.c: Include params.h.
5683 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
5684 other reload pseudos considerations.
5686 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5688 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
5689 to use canonical form for nor<mode>3.
5691 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5694 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
5697 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
5699 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
5700 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
5701 (ix86_handle_option): Handle OPT_mprefetchwt1.
5702 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
5703 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
5705 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
5706 OPTION_MASK_ISA_PREFETCHWT1.
5707 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
5708 (PTA_PREFETCHWT1): New.
5709 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
5710 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
5711 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
5712 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
5713 (*prefetch_avx512pf_<mode>_: Change into ...
5714 (*prefetch_prefetchwt1_<mode>: This.
5715 * config/i386/i386.opt (mprefetchwt1): New.
5716 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
5717 (_mm_prefetch): Handle intent to write.
5718 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
5720 2014-02-25 Richard Biener <rguenther@suse.de>
5723 * emit-rtl.c (mem_attrs_htab): Remove.
5724 (mem_attrs_htab_hash): Likewise.
5725 (mem_attrs_htab_eq): Likewise.
5726 (set_mem_attrs): Always allocate new mem-attrs when something changed.
5727 (init_emit_once): Do not allocate mem_attrs_htab.
5729 2014-02-25 Richard Biener <rguenther@suse.de>
5732 * lto-opts.c (lto_write_options): Output non-explicit conservative
5733 -fwrapv, -fno-trapv and -fno-strict-overflow.
5734 * lto-wrapper.c (merge_and_complain): Handle merging those options.
5735 (run_gcc): And pass them through.
5737 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
5739 * sel-sched.c (calculate_new_fences): New parameter ptime.
5740 Calculate it as a maximum over all fence cycles.
5741 (sel_sched_region_2): Adjust the call to calculate_new_fences.
5742 Print the final schedule timing when sched_verbose.
5744 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
5746 PR rtl-optimization/60292
5747 * sel-sched.c (fill_vec_av_set): Do not reset target availability
5748 bit fot the fence instruction.
5750 2014-02-24 Alangi Derick <alangiderick@gmail.com>
5752 * calls.h: Fix typo in comment.
5754 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
5756 * config/pa/pa.c (pa_output_move_double): Don't valididate when
5757 adjusting offsetable addresses.
5759 2014-02-24 Guozhi Wei <carrot@google.com>
5761 * sparseset.h (sparseset_pop): Fix the wrong index.
5763 2014-02-24 Walter Lee <walt@tilera.com>
5765 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
5766 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
5768 * common/config/tilegx/tilegx-common.c
5769 (TARGET_DEFAULT_TARGET_FLAGS): Define.
5770 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
5772 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
5773 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
5774 (tilegx_gimplify_va_arg_expr): Handle big endian.
5775 (tilegx_expand_unaligned_load): Ditto.
5776 (tilegx_expand_unaligned_store): Ditto.
5777 (TARGET_RETURN_IN_MSB): New.
5778 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
5779 (TARGET_ENDIAN_DEFAULT): New.
5780 (TARGET_BIG_ENDIAN): Handle big endian.
5781 (BYTES_BIG_ENDIAN): Ditto.
5782 (WORDS_BIG_ENDIAN): Ditto.
5783 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
5786 * config/tilegx/tilegx.md (extv): Handle big endian.
5788 (insn_st<n>): Ditto.
5789 (insn_st<n>_add<bitsuffix>): Ditto.
5790 (insn_stnt<n>): Ditto.
5791 (insn_stnt<n>_add<bitsuffix>):Ditto.
5792 (vec_interleave_highv8qi): Handle big endian.
5793 (vec_interleave_highv8qi_be): New.
5794 (vec_interleave_highv8qi_le): New.
5795 (insn_v1int_h): Handle big endian.
5796 (vec_interleave_lowv8qi): Handle big endian.
5797 (vec_interleave_lowv8qi_be): New.
5798 (vec_interleave_lowv8qi_le): New.
5799 (insn_v1int_l): Handle big endian.
5800 (vec_interleave_highv4hi): Handle big endian.
5801 (vec_interleave_highv4hi_be): New.
5802 (vec_interleave_highv4hi_le): New.
5803 (insn_v2int_h): Handle big endian.
5804 (vec_interleave_lowv4hi): Handle big endian.
5805 (vec_interleave_lowv4hi_be): New.
5806 (vec_interleave_lowv4hi_le): New.
5807 (insn_v2int_l): Handle big endian.
5808 (vec_interleave_highv2si): Handle big endian.
5809 (vec_interleave_highv2si_be): New.
5810 (vec_interleave_highv2si_le): New.
5811 (insn_v4int_h): Handle big endian.
5812 (vec_interleave_lowv2si): Handle big endian.
5813 (vec_interleave_lowv2si_be): New.
5814 (vec_interleave_lowv2si_le): New.
5815 (insn_v4int_l): Handle big endian.
5816 * config/tilegx/tilegx.opt (mbig-endian): New option.
5817 (mlittle-endian): New option.
5818 * doc/install.texi: Document tilegxbe-linux.
5819 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
5821 2014-02-24 Martin Jambor <mjambor@suse.cz>
5824 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
5825 there are no parameter descriptors.
5827 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
5829 PR rtl-optimization/60268
5830 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
5831 initialization to ...
5832 (sched_rgn_init): ... here.
5833 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
5835 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5837 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
5840 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
5842 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
5845 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5847 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
5848 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
5850 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5852 * config/microblaze/predicates.md: Add cmp_op predicate.
5853 * config/microblaze/microblaze.md: Add branch_compare instruction
5854 which uses cmp_op predicate and emits cmp insn before branch.
5855 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
5856 to microblaze_expand_conditional_branch and consolidate logic.
5857 (microblaze_expand_conditional_branch): emit branch_compare
5858 insn instead of handling cmp op separate from branch insn.
5860 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5862 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
5865 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5867 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
5868 define_insn with define_expand and new define_insn
5869 *altivec_lve<VI_char>x_internal.
5870 (altivec_stve<VI_char>x): Replace define_insn with define_expand
5871 and new define_insn *altivec_stve<VI_char>x_internal.
5872 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
5874 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
5876 (altivec_expand_stvex_be): New function.
5878 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
5880 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
5881 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
5882 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
5883 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
5885 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
5888 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
5889 instead of emit_move_insn.
5891 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5893 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
5895 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
5896 gen_altivec_vsumsws.
5898 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5900 * config/rs6000/altivec.md (altivec_lvxl): Rename as
5901 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
5902 (altivec_lvxl_<mode>): New define_expand incorporating
5903 -maltivec=be semantics where needed.
5904 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
5905 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
5906 semantics where needed.
5907 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
5908 (altivec_stvx_<mode>): New define_expand incorporating
5909 -maltivec=be semantics where needed.
5910 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
5911 VM2 iterator instead of V4SI.
5912 (altivec_stvxl_<mode>): New define_expand incorporating
5913 -maltivec=be semantics where needed.
5914 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
5915 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
5916 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
5917 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
5918 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
5919 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
5920 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
5921 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
5922 ALTIVEC_BUILTIN_STVXL.
5923 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
5924 (altivec_expand_stvx_be): Likewise.
5925 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
5926 (altivec_expand_lvx_be): Likewise.
5927 (altivec_expand_stvx_be): Likewise.
5928 (altivec_expand_builtin): Add cases for
5929 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
5930 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
5931 (altivec_init_builtins): Add definitions for
5932 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
5933 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
5935 2014-02-21 Catherine Moore <clm@codesourcery.com>
5937 * doc/invoke.texi (mvirt, mno-virt): Document.
5938 * config/mips/mips.opt (mvirt): New option.
5939 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
5941 2014-02-21 Richard Biener <rguenther@suse.de>
5943 PR tree-optimization/60276
5944 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
5945 (STMT_VINFO_MIN_NEG_DIST): New macro.
5946 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
5947 STMT_VINFO_MIN_NEG_DIST.
5948 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
5949 made for negative dependence distances still hold.
5951 2014-02-21 Richard Biener <rguenther@suse.de>
5954 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
5955 DECL_INITIAL for globals not in the current function context.
5957 2014-02-21 Jakub Jelinek <jakub@redhat.com>
5959 PR tree-optimization/56490
5960 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
5961 * tree-ssa-uninit.c: Include params.h.
5962 (compute_control_dep_chain): Add num_calls argument, return false
5963 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
5964 num_calls to recursive call.
5965 (find_predicates): Change dep_chain into normal array,
5966 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
5967 variable and adjust compute_control_dep_chain caller.
5968 (find_def_preds): Likewise.
5970 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
5972 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
5973 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
5975 2014-02-21 Nick Clifton <nickc@redhat.com>
5977 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
5978 (pushhi1): Likewise.
5979 (popqi1): Add mode to pre_dec.
5982 2014-02-21 Jakub Jelinek <jakub@redhat.com>
5984 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
5985 mode for mask of V8SFmode permutation.
5987 2014-02-20 Richard Henderson <rth@redhat.com>
5990 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
5991 a new pseudo for OLDVAL.
5993 2014-02-20 Jakub Jelinek <jakub@redhat.com>
5996 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
5997 gen_reg_rtx if d->testing_p.
5998 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
5999 if d->testing_p and we will certainly return true.
6000 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6003 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6005 * emit-rtl.c (gen_reg_rtx): Assert that
6006 crtl->emit.regno_pointer_align_length is non-zero.
6008 2014-02-20 Richard Henderson <rth@redhat.com>
6011 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6012 on failure the store back into EXPECT.
6014 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6015 Sandra Loosemore <sandra@codesourcery.com>
6017 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6018 * config/nios2/nios2.c (nios2_function_profiler): Add
6019 -fPIC (flag_pic == 2) support.
6020 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6021 (nios2_large_offset_p): New function.
6022 (nios2_unspec_reloc_p): Move up position, update to use
6023 nios2_large_offset_p.
6024 (nios2_unspec_address): Remove function.
6025 (nios2_unspec_offset): New function.
6026 (nios2_large_got_address): New function.
6027 (nios2_got_address): Add large offset support.
6028 (nios2_legitimize_tls_address): Update usage of removed and new
6030 (nios2_symbol_binds_local_p): New function.
6031 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
6032 (nios2_legitimize_address): Update to use nios2_large_offset_p.
6033 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
6034 (nios2_print_operand): Merge H/L processing, add hiadj/lo
6035 processing for (const (unspec ...)).
6036 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
6038 2014-02-20 Richard Biener <rguenther@suse.de>
6040 * tree-cfg.c (replace_uses_by): Mark altered BBs before
6041 doing the substitution.
6042 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
6044 2014-02-20 Martin Jambor <mjambor@suse.cz>
6047 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
6048 info when checking whether lattices are bottom.
6050 2014-02-20 Richard Biener <rguenther@suse.de>
6053 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
6056 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
6059 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
6060 parameter specifying the scaling.
6061 (inline_call): Update.
6062 (want_inline_recursively): Guard division by zero.
6063 (recursive_inlining): Update.
6064 * ipa-inline.h (clone_inlined_nodes): Update.
6066 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6069 * config/i386/i386.c (classify_argument): Pass structures of size
6070 64 bytes or less in register.
6072 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6073 Kirill Yukhin <kirill.yukhin@intel.com>
6075 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
6076 (_mm_rcp28_round_ss): Ditto.
6077 (_mm_rsqrt28_round_sd): Ditto.
6078 (_mm_rsqrt28_round_ss): Ditto.
6079 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
6080 (_mm_rcp14_round_ss): Ditto.
6081 (_mm_rsqrt14_round_sd): Ditto.
6082 (_mm_rsqrt14_round_ss): Ditto.
6083 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
6084 the first input operand, get rid of match_dup.
6085 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
6087 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
6089 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
6090 operand as the first input operand, set type attribute.
6091 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
6093 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
6094 operand as the first input operand, set type attribute.
6096 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6098 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
6101 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
6104 * config/i386/i386.c (construct_container): Remove TFmode check
6105 for X86_64_INTEGER_CLASS.
6107 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
6110 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
6111 only when -Wpsabi is enabled.
6113 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
6116 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
6117 passing arrays in registers are the same as for structs, so remove the
6118 special case for them.
6120 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
6122 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
6123 destination type, extract only the valid bits if the source type is not
6124 integral and has a different mode.
6126 2014-02-19 Richard Biener <rguenther@suse.de>
6129 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
6132 2014-02-19 Richard Biener <rguenther@suse.de>
6135 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
6136 (ipa_modify_call_arguments): Emit an argument load explicitely and
6137 preserve virtual SSA form there and for the replacement call.
6138 Do not update SSA form nor free dominance info.
6140 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6142 * ipa.c (function_and_variable_visibility): Also clear WEAK
6143 flag when disolving COMDAT_GROUP.
6145 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6147 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
6148 * ipa-prop.c (ipa_set_jf_known_type): Return early when
6150 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
6151 do more sanity checks.
6152 (detect_type_change): Return true when giving up early.
6153 (compute_complex_assign_jump_func): Fix type parameter of
6154 ipa_set_ancestor_jf.
6155 (compute_complex_ancestor_jump_func): Likewise.
6156 (update_jump_functions_after_inlining): Fix updating of
6158 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
6160 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6162 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
6163 inline clones when edge disappears.
6165 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6168 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
6169 Split 64-bit moves into 2 patterns. Do not allow the use of
6170 direct move for TDmode in little endian, since the decimal value
6171 has little endian bytes within a word, but the 64-bit pieces are
6172 ordered in a big endian fashion, and normal subreg's of TDmode are
6174 (mov<mode>_64bit_dm): Likewise.
6175 (movtd_64bit_nodm): Likewise.
6177 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6179 PR tree-optimization/60174
6180 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
6181 statement of an SSA_NAME that occurs in an abnormal PHI node.
6183 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6186 * final.c (SEEN_BB): Remove.
6187 (SEEN_NOTE, SEEN_EMITTED): Renumber.
6188 (final_scan_insn): Don't force_source_line on second
6189 NOTE_INSN_BASIC_BLOCK.
6191 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
6194 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
6195 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
6196 (type_natural_mode): Warn ABI change when %zmm register is not
6197 available for AVX512F vector value passing.
6199 2014-02-18 Kai Tietz <ktietz@redhat.com>
6202 * config/i386/i386.c (ix86_expand_prologue): Use value in
6203 rax register as displacement when restoring %r10 or %rax.
6204 Fix wrong offset when restoring both registers.
6206 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6208 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
6209 assertion with conditional return.
6211 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6212 Uros Bizjak <ubizjak@gmail.com>
6215 * config/i386/driver-i386.c (host_detect_local_cpu): If
6216 YMM state is not saved by the OS, also clear has_f16c. Move
6217 CPUID 0x80000001 handling before YMM state saving checking.
6219 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
6221 PR rtl-optimization/58960
6222 * haifa-sched.c (alloc_global_sched_pressure_data): New,
6223 factored out from ...
6224 (sched_init): ... here.
6225 (free_global_sched_pressure_data): New, factored out from ...
6226 (sched_finish): ... here.
6227 * sched-int.h (free_global_sched_pressure_data): Declare.
6228 * sched-rgn.c (nr_regions_initial): New static global.
6229 (haifa_find_rgns): Initialize it.
6230 (schedule_region): Disable sched-pressure for the newly
6233 2014-02-17 Richard Biener <rguenther@suse.de>
6235 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
6236 release SSA defs of pattern stmts.
6238 2014-02-17 Richard Biener <rguenther@suse.de>
6240 * tree-inline.c (expand_call_inline): Release the virtual
6241 operand defined by the call we are about to inline.
6243 2014-02-17 Richard Biener <rguenther@suse.de>
6245 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
6247 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
6248 Ilya Tocar <ilya.tocar@intel.com>
6250 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
6251 arguments order in builtin.
6252 (_mm512_permutexvar_epi64): Ditto.
6253 (_mm512_mask_permutexvar_epi64): Ditto
6254 (_mm512_maskz_permutexvar_epi32): Ditto
6255 (_mm512_permutexvar_epi32): Ditto
6256 (_mm512_mask_permutexvar_epi32): Ditto
6258 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6260 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
6261 (p8_vmrgow): Likewise.
6263 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6265 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
6268 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
6271 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
6272 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
6273 into 64-bit and 32-bit moves. On 64-bit moves, add support for
6274 using direct move instructions on ISA 2.07. Also adjust
6275 instruction length for 64-bit.
6276 (mov<mode>_64bit, TFmode/TDmode): Likewise.
6277 (mov<mode>_32bit, TFmode/TDmode): Likewise.
6279 2014-02-15 Alan Modra <amodra@gmail.com>
6283 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
6284 find_replacement on parts of insn rtl that might be reloaded.
6286 2014-02-15 Richard Biener <rguenther@suse.de>
6288 PR tree-optimization/60183
6289 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
6290 (tree_ssa_phiprop): Calculate and free post-dominators.
6292 2014-02-14 Jeff Law <law@redhat.com>
6294 PR rtl-optimization/60131
6295 * ree.c (get_extended_src_reg): New function.
6296 (combine_reaching_defs): Use it rather than assuming location of REG.
6297 (find_and_remove_re): Verify first operand of extension is
6298 a REG before adding the insns to the copy list.
6300 2014-02-14 Roland McGrath <mcgrathr@google.com>
6302 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
6303 * configure: Regenerated.
6304 * config.in: Regenerated.
6305 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
6306 instead of ASM_SHORT.
6308 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
6309 Richard Earnshaw <rearnsha@arm.com>
6311 PR rtl-optimization/59535
6312 * lra-constraints.c (process_alt_operands): Encourage alternative
6313 when unassigned pseudo class is superset of the alternative class.
6314 (inherit_reload_reg): Don't inherit when optimizing for code size.
6315 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
6316 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
6317 modes not less than 4 for Thumb1.
6319 2014-02-14 Kyle McMartin <kyle@redhat.com>
6322 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
6324 2014-02-14 Richard Biener <rguenther@suse.de>
6326 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
6327 (get_frame_arg): Drop the assert with langhook types_compatible_p.
6328 Do not strip INDIRECT_REFs.
6330 2014-02-14 Richard Biener <rguenther@suse.de>
6333 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
6334 DECL_FUNCTION_SPECIFIC_TARGET.
6335 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
6336 * tree-streamer-out.c (pack_ts_target_option): Remove.
6337 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
6338 (write_ts_function_decl_tree_pointers): Do not stream
6339 DECL_FUNCTION_SPECIFIC_TARGET.
6340 * tree-streamer-in.c (unpack_ts_target_option): Remove.
6341 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
6342 (lto_input_ts_function_decl_tree_pointers): Do not stream
6343 DECL_FUNCTION_SPECIFIC_TARGET.
6345 2014-02-14 Jakub Jelinek <jakub@redhat.com>
6347 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
6348 (get_initial_def_for_induction, vectorizable_induction): Ignore
6349 debug stmts when looking for exit_phi.
6350 (vectorizable_live_operation): Fix up condition.
6352 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6354 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
6355 nreverse() because it changes the content of original tree list.
6357 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6359 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
6360 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
6362 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6364 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
6365 GNU coding standards.
6367 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6370 * dwarf2out.c (gen_subprogram_die): Don't call
6371 add_calling_convention_attribute if subr_die is old_die.
6373 2014-02-13 Sharad Singhai <singhai@google.com>
6375 * doc/optinfo.texi: Fix order of nodes.
6377 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
6379 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
6380 operands[2], not operands[3].
6382 2014-02-13 Richard Biener <rguenther@suse.de>
6385 * doc/install.texi (ISL): Update recommended version to 0.12.2,
6386 mention the possibility of an in-tree build.
6387 (CLooG): Update recommended version to 0.18.1, mention the
6388 possibility of an in-tree build and clarify that the ISL
6389 bundled with CLooG does not work.
6391 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6394 * expr.c (compress_float_constant): If x is a hard register,
6395 extend into a pseudo and then move to x.
6397 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
6399 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
6400 caused by bad second argument to warning_at() with -mhotpatch and
6401 nested functions (e.g. with gfortran).
6403 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
6405 * opts.c (option_name): Remove "enabled by default" rider.
6407 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
6409 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
6411 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
6412 Uros Bizjak <ubizjak@gmail.com>
6415 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
6416 * configure: Regenerated.
6418 2014-02-12 Richard Biener <rguenther@suse.de>
6420 * vec.c (vec_prefix::calculate_allocation): Move as
6421 inline variant to vec.h.
6422 (vec_prefix::calculate_allocation_1): New out-of-line version.
6423 * vec.h (vec_prefix::calculate_allocation_1): Declare.
6424 (vec_prefix::m_has_auto_buf): Rename to ...
6425 (vec_prefix::m_using_auto_storage): ... this.
6426 (vec_prefix::calculate_allocation): Inline the easy cases
6427 and dispatch to calculate_allocation_1 which doesn't need the
6429 (va_heap::reserve): Use gcc_checking_assert.
6430 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
6431 m_using_auto_storage.
6432 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
6434 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
6435 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
6436 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
6438 2014-02-12 Richard Biener <rguenther@suse.de>
6440 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
6441 when we found a dependence.
6443 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
6445 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
6447 (maybe_fold_stmt): ... into this new function.
6448 * omp-low.c (lower_omp): Update comment.
6450 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
6453 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
6456 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
6458 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
6459 identifiers in comments.
6460 (cortexa53_extra_costs): Likewise.
6461 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
6462 (cortexa7_extra_costs): Likewise.
6463 (cortexa12_extra_costs): Likewise.
6464 (cortexa15_extra_costs): Likewise.
6465 (v7m_extra_costs): Likewise.
6467 2014-02-12 Richard Biener <rguenther@suse.de>
6470 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
6471 of posix_memalign being successful.
6472 (lower_stmt): Restrict lowering of posix_memalign to when
6473 -ftree-bit-ccp is enabled.
6475 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6477 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
6479 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
6481 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
6483 PR rtl-optimization/60116
6484 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
6485 other_insn once the combination has been validated.
6487 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
6490 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
6492 * ipa-devirt.c: Include demangle.h
6493 (odr_violation_reported): New static variable.
6494 (add_type_duplicate): Update odr_violations.
6495 (maybe_record_node): Add completep parameter; update it.
6496 (record_target_from_binfo): Add COMPLETEP parameter;
6497 update it as needed.
6498 (possible_polymorphic_call_targets_1): Likewise.
6499 (struct polymorphic_call_target_d): Add nonconstruction_targets;
6500 rename FINAL to COMPLETE.
6501 (record_targets_from_bases): Sanity check we found the binfo;
6502 fix COMPLETEP updating.
6503 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
6504 parameter, fix computing of COMPLETEP.
6505 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
6506 at LTO time do demangling.
6507 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
6508 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
6510 (gimple_get_virt_method_for_binfo): Likewise.
6511 * gimple-fold.h (gimple_get_virt_method_for_binfo,
6512 gimple_get_virt_method_for_vtable): Update prototypes.
6514 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
6517 * genautomata.c (add_presence_absence): Fix typo with
6518 {final_}presence_list.
6520 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
6523 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
6524 for VSX/Altivec vectors that land in GPR registers.
6526 2014-02-11 Richard Henderson <rth@redhat.com>
6527 Jakub Jelinek <jakub@redhat.com>
6530 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
6531 around drhs if type conversion to lacc->type is not useless.
6533 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6535 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
6537 (cortex-a57.cortex-a53): Likewise.
6538 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
6540 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6542 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
6545 2014-02-11 Renlin Li <Renlin.Li@arm.com>
6547 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
6548 add_options_for_arm_vfp3.
6550 2014-02-11 Jeff Law <law@redhat.com>
6553 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
6554 object with an undesirable mode.
6556 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6559 * config/i386/sol2-9.h: New file.
6560 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
6561 *-*-solaris2.9*): Use it.
6563 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6565 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
6566 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
6568 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6570 * config/microblaze/microblaze.c: Extend mcpu version format
6572 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
6574 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
6576 2014-02-10 Richard Henderson <rth@redhat.com>
6579 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
6580 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
6581 ms-abi vs -mno-accumulate-outgoing-args.
6582 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
6583 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
6586 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
6589 * cfgexpand.c (expand_asm_operands): Attach source location to
6590 ASM_INPUT rtx objects.
6591 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
6593 2014-02-10 Nick Clifton <nickc@redhat.com>
6595 * config/mn10300/mn10300.c (popcount): New function.
6596 (mn10300_expand_prologue): Include saved registers in stack usage
6599 2014-02-10 Jeff Law <law@redhat.com>
6602 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
6603 when changing the SET_DEST of a prior insn to avoid an input reload.
6605 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6607 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
6608 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
6609 -mcall-openbsd, or -mcall-linux.
6610 (CC1_ENDIAN_BIG_SPEC): Remove.
6611 (CC1_ENDIAN_LITTLE_SPEC): Remove.
6612 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
6613 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
6614 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
6615 and %cc1_endian_default.
6616 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
6618 2014-02-10 Richard Biener <rguenther@suse.de>
6620 PR tree-optimization/60115
6621 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
6622 MEM_REF handling. Properly verify that the accesses are not
6623 out of the objects bound.
6625 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6627 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
6630 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
6632 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
6633 proper constants and fix formatting.
6634 (possible_polymorphic_call_targets): Fix formatting.
6636 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
6637 Ilya Tocar <ilya.tocar@intel.com>
6639 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
6640 (_mm512_loadu_epi32): Renamed into...
6641 (_mm512_loadu_si512): This.
6642 (_mm512_storeu_epi32): Renamed into...
6643 (_mm512_storeu_si512): This.
6644 (_mm512_maskz_ceil_ps): Removed.
6645 (_mm512_maskz_ceil_pd): Ditto.
6646 (_mm512_maskz_floor_ps): Ditto.
6647 (_mm512_maskz_floor_pd): Ditto.
6648 (_mm512_floor_round_ps): Ditto.
6649 (_mm512_floor_round_pd): Ditto.
6650 (_mm512_ceil_round_ps): Ditto.
6651 (_mm512_ceil_round_pd): Ditto.
6652 (_mm512_mask_floor_round_ps): Ditto.
6653 (_mm512_mask_floor_round_pd): Ditto.
6654 (_mm512_mask_ceil_round_ps): Ditto.
6655 (_mm512_mask_ceil_round_pd): Ditto.
6656 (_mm512_maskz_floor_round_ps): Ditto.
6657 (_mm512_maskz_floor_round_pd): Ditto.
6658 (_mm512_maskz_ceil_round_ps): Ditto.
6659 (_mm512_maskz_ceil_round_pd): Ditto.
6660 (_mm512_expand_pd): Ditto.
6661 (_mm512_expand_ps): Ditto.
6662 * config/i386/i386.c (ix86_builtins): Remove
6663 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
6664 (bdesc_args): Ditto.
6665 * config/i386/predicates.md (const1256_operand): New.
6666 (const_1_to_2_operand): Ditto.
6667 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6668 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6669 (*avx512pf_gatherpf<mode>sf): Ditto.
6670 (avx512pf_gatherpf<mode>df): Ditto.
6671 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6672 (*avx512pf_gatherpf<mode>df): Ditto.
6673 (avx512pf_scatterpf<mode>sf): Ditto.
6674 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6675 (*avx512pf_scatterpf<mode>sf): Ditto.
6676 (avx512pf_scatterpf<mode>df): Ditto.
6677 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6678 (*avx512pf_scatterpf<mode>df): Ditto.
6679 (avx512f_expand<mode>): Removed.
6680 (<shift_insn><mode>3<mask_name>): Change predicate type.
6682 2014-02-08 Jakub Jelinek <jakub@redhat.com>
6684 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
6685 not at the end of datarefs vector use ordered_remove to avoid
6686 reordering datarefs vector.
6689 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
6690 mark local addressable non-static vars as GOVD_PRIVATE
6691 instead of GOVD_LOCAL.
6692 * omp-low.c (lower_omp_for): Move gimple_bind_vars
6693 and BLOCK_VARS of gimple_bind_block to new_stmt rather
6697 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
6698 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
6699 assume_aligned or alloc_align attributes.
6700 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
6701 arguments. Handle also assume_aligned and alloc_align attributes.
6702 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
6703 calls to functions with assume_aligned or alloc_align attributes.
6704 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
6706 2014-02-08 Terry Guo <terry.guo@arm.com>
6708 * doc/invoke.texi: Document ARM -march=armv7e-m.
6710 2014-02-08 Jakub Jelinek <jakub@redhat.com>
6712 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
6713 flag on __cilkrts_rethrow builtin.
6716 * ipa-cp.c (determine_versionability): Fail at -O0
6717 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
6718 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
6721 2014-02-04 Jakub Jelinek <jakub@redhat.com>
6724 * tree-inline.c (copy_forbidden): Fail for
6725 __attribute__((optimize (0))) functions.
6727 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
6729 * varpool.c: Include pointer-set.h.
6730 (varpool_remove_unreferenced_decls): Variables in other partitions
6731 will not be output; be however careful to not lose information
6734 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
6736 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
6737 lookup in the vtable constructor.
6739 2014-02-07 Jeff Law <law@redhat.com>
6742 * config/m68k/m68k.md (ashldi_extsi): Turn into a
6743 define_insn_and_split.
6745 * ipa-inline.c (inline_small_functions): Fix typos.
6747 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6749 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
6750 (s390_can_use_return_insn): Declare.
6751 * config/s390/s390.h (EPILOGUE_USES): Define.
6752 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
6754 (s390_chunkify_start): Handle return JUMP_LABELs.
6755 (s390_early_mach): Emit a main_pool instruction on the entry edge.
6756 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
6757 (s390_can_use_return_insn): New functions.
6758 (s390_fix_long_loop_prediction): Handle conditional returns.
6759 (TARGET_SET_UP_BY_PROLOGUE): Define.
6760 * config/s390/s390.md (ANY_RETURN): New code iterator.
6761 (*creturn, *csimple_return, return, simple_return): New patterns.
6763 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6765 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
6766 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
6767 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
6768 REG_CFA_RESTORE list when deciding not to restore a register.
6770 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6772 * config/s390/s390.c: Include tree-pass.h and context.h.
6773 (s390_early_mach): New function, split out from...
6774 (s390_emit_prologue): ...here.
6775 (pass_data_s390_early_mach): New pass structure.
6776 (pass_s390_early_mach): New class.
6777 (s390_option_override): Create and register early_mach pass.
6778 Move to end of file.
6780 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6782 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
6783 to match for the exit block.
6785 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6787 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
6788 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
6789 Reject misaligned operands.
6791 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6793 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
6795 2014-02-07 Richard Biener <rguenther@suse.de>
6798 * gimple-low.c (lower_builtin_posix_memalign): New function.
6799 (lower_stmt): Call it to lower posix_memalign in a way
6800 to make alignment info accessible.
6802 2014-02-07 Jakub Jelinek <jakub@redhat.com>
6805 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
6806 __builtin_setjmp_receiver.
6808 2014-02-07 Richard Biener <rguenther@suse.de>
6811 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
6812 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
6813 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
6814 Handle BUILT_IN_POSIX_MEMALIGN.
6815 (find_func_clobbers): Likewise.
6816 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
6817 (call_may_clobber_ref_p_1): Likewise.
6819 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6822 * ipa-devirt.c (record_target_from_binfo): Remove overactive
6825 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6828 * lto-cgraph.c (lto_output_node): Use
6829 symtab_get_symbol_partitioning_class.
6830 (lto_output_varpool_node): likewise.
6831 (symtab_get_symbol_partitioning_class): Move here from
6833 * cgraph.h (symbol_partitioning_class): Likewise.
6834 (symtab_get_symbol_partitioning_class): Declare.
6836 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6838 * ggc.h (ggc_internal_cleared_alloc): New macro.
6839 * vec.h (vec_safe_copy): Handle memory stats.
6840 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
6841 * target-globals.c (save_target_globals): Likewise.
6843 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6846 * expr.c (emit_move_resolve_push): Export; be bit more selective
6847 on when to clear alias set.
6848 * expr.h (emit_move_resolve_push): Declare.
6849 * function.h (struct function): Add tail_call_marked.
6850 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
6851 * config/i386/i386-protos.h (ix86_expand_push): Remove.
6852 * config/i386/i386.md (TImode move expander): De not call
6854 (FP push expanders): Preserve memory attributes.
6855 * config/i386/sse.md (push<mode>1): Remove.
6856 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
6857 (ix86_expand_push): Remove.
6858 * config/i386/mmx.md (push<mode>1): Remove.
6860 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6862 PR rtl-optimization/60030
6863 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
6864 lopart with paradoxical subreg before shifting it up by hprec.
6866 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6868 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
6869 Remove extra newline at end of file.
6870 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
6871 (arm_issue_rate): Handle cortexa57.
6872 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
6873 (cortex-a57.cortex-a53): Likewise.
6875 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6878 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
6879 don't record in REG_FRAME_RELATED_EXPR registers not set in that
6881 (arm_expand_prologue): Adjust all callers.
6882 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
6883 info, registers also at the lowest numbered registers side. Use
6884 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
6888 * var-tracking.c (adjust_mems): Before adding a SET to
6889 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
6891 2014-02-06 Alan Modra <amodra@gmail.com>
6894 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
6895 change SDmode to DDmode when lra_in_progress.
6897 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6900 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
6901 free_data_ref on the dr first, and before goto again also set dr
6902 to the next dr. For simd_lane_access, free old datarefs[i] before
6903 overwriting it. For get_vectype_for_scalar_type failure, don't
6904 free_data_ref if simd_lane_access.
6906 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
6909 * tree.h (opts_for_fn): New inline function.
6910 (opt_for_fn): Define.
6911 * config/i386/i386.c (ix86_function_regparm): Use
6912 opt_for_fn (decl, optimize) instead of optimize.
6914 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
6916 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
6917 for SYMBOL_REF in large memory model.
6919 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6921 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
6923 (cortex-a57): Likewise.
6924 (cortex-a57.cortex-a53): Likewise.
6926 2014-02-06 Yury Gribov <y.gribov@samsung.com>
6927 Kugan Vivekanandarajah <kuganv@linaro.org>
6929 * config/arm/arm.c (arm_vector_alignment_reachable): Check
6931 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
6933 2014-02-06 Richard Biener <rguenther@suse.de>
6935 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
6936 set_loop_copy and initialize_original_copy_tables.
6938 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
6940 * config/aarch64/aarch64-simd.md
6941 (aarch64_ashr_simddi): Change QI to SI.
6943 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
6944 Jakub Jelinek <jakub@redhat.com>
6947 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
6950 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6952 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
6953 CODE_FOR_altivec_vpku[hw]um to
6954 CODE_FOR_altivec_vpku[hw]um_direct.
6955 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
6956 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
6957 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
6958 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
6960 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6962 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
6963 generation for -maltivec=be.
6964 (altivec_vsumsws): Simplify redundant test.
6966 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6968 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
6969 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
6970 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
6971 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
6972 gen_altivec_vpkuwum.
6973 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
6975 (altivec_vpks<VI_char>ss): Likewise.
6976 (altivec_vpks<VI_char>us): Likewise.
6977 (altivec_vpku<VI_char>us): Likewise.
6978 (altivec_vpku<VI_char>um): Likewise.
6979 (altivec_vpku<VI_char>um_direct): New (copy of
6980 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
6982 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
6983 target is little endian and -maltivec=be is not specified.
6984 (*altivec_vupkhs<VU_char>_direct): New (copy of
6985 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
6986 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
6987 target is little endian and -maltivec=be is not specified.
6988 (*altivec_vupkls<VU_char>_direct): New (copy of
6989 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
6990 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
6991 little endian and -maltivec=be is not specified.
6992 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
6993 little endian and -maltivec=be is not specified.
6995 2014-02-05 Richard Henderson <rth@redhat.com>
6998 * combine-stack-adj.c: Revert r206943.
6999 * sched-int.h (struct deps_desc): Add last_args_size.
7000 * sched-deps.c (init_deps): Initialize it.
7001 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7002 contain REG_ARGS_SIZE notes.
7004 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7006 * lto-cgraph.c (asm_nodes_output): Make global.
7007 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7008 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7009 (driver_handle_option): Handle OPT_fwpa.
7011 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7014 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7015 a comment typo and formatting issue. If odr_hash hasn't been
7016 created, return vNULL and set *completep to false.
7019 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7020 bb with no successors.
7022 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7025 * doc/invoke.texi (-march): Clarify documentation for ARM.
7029 2014-02-05 Richard Biener <rguenther@suse.de>
7031 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
7032 when not vectorizing because of too many alias checks.
7033 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
7034 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
7036 2014-02-05 Nick Clifton <nickc@redhat.com>
7038 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
7039 accept extended registers in any mode when compiling for the MN10300.
7041 2014-02-05 Yury Gribov <y.gribov@samsung.com>
7043 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
7044 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
7045 sanitization attributes.
7046 (can_inline_edge_p): Likewise.
7047 (sanitize_attrs_match_for_inline_p): New function.
7049 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7051 * ipa-prop.c (detect_type_change): Shor circuit testing of
7052 type changes on THIS pointer.
7054 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
7057 * config/pa/pa.c (legitimize_tls_address): Return original address
7058 if not passed a SYMBOL_REF rtx.
7059 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
7061 (pa_emit_move_sequence): Simplify TLS source operands.
7062 (pa_legitimate_constant_p): Reject all TLS constants.
7063 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
7064 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
7066 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7068 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
7069 groups when we know they are controlled by LTO.
7070 * varasm.c (default_binds_local_p_1): If object is in other partition,
7071 it will be resolved locally.
7073 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7075 * config/host-linux.c (linux_gt_pch_use_address): Don't
7076 use SSIZE_MAX because it is not always defined.
7078 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
7081 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
7082 threshold for pseudo splitting.
7083 (update_ebb_live_info): Process call argument hard registers and
7084 hard registers from insn definition too.
7085 (max_small_class_regs_num): New constant.
7086 (inherit_in_ebb): Update live hard regs through EBBs. Update
7087 reloads_num only for small register classes. Don't split for
7090 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
7093 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
7096 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7098 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
7101 2014-02-04 Marek Polacek <polacek@redhat.com>
7103 * gdbinit.in (pel): Define.
7105 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7107 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
7110 2014-02-04 Richard Biener <rguenther@suse.de>
7113 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
7114 in function context local.
7115 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
7116 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
7117 similar to LTO_imported_decl_ref.
7119 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7121 PR tree-optimization/60002
7122 * cgraphclones.c (build_function_decl_skip_args): Clear
7125 PR tree-optimization/60023
7126 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
7127 false to gsi_replace.
7128 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
7129 has been in some EH region and vec_stmt could throw, add
7130 vec_stmt into the same EH region.
7131 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
7132 has no lhs, ignore it.
7133 * internal-fn.c (expand_MASK_LOAD): Likewise.
7136 * tree-inline.c (copy_forbidden): Fail for
7137 __attribute__((optimize (0))) functions.
7140 * omp-low.c (simd_clone_struct_copy): If from->inbranch
7141 is set, copy one less argument.
7142 (expand_simd_clones): Don't subtract clone_info->inbranch
7143 from simd_clone_struct_alloc argument.
7145 PR rtl-optimization/57915
7146 * recog.c (simplify_while_replacing): If all unary/binary/relational
7147 operation arguments are constant, attempt to simplify those.
7150 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
7151 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
7153 2014-02-04 Richard Biener <rguenther@suse.de>
7155 PR tree-optimization/60012
7156 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
7157 TBAA disambiguation to all DDRs.
7159 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7162 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
7163 (LINK_SPEC): Use it for -shared, -shared-libgcc.
7165 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7168 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
7170 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7172 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
7173 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
7175 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7178 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
7179 to figure out targets of polymorphic calls with known decl.
7180 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7181 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
7182 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
7183 (get_polymorphic_call_info): ... here.
7184 (get_polymorphic_call_info_from_invariant): New function.
7186 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7188 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
7189 lookup via vtable pointer; check for type consistency
7190 and turn inconsitent facts into UNREACHABLE.
7191 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7192 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
7193 type inconsistent querries; return UNREACHABLE instead.
7195 2014-02-03 Richard Henderson <rth@twiddle.net>
7198 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
7199 already processed this node.
7200 (normalize_one_pred_1): Pass along mark_set.
7201 (normalize_one_pred): Create and destroy a pointer_set_t.
7202 (normalize_one_pred_chain): Likewise.
7204 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
7206 PR gcov-profile/58602
7207 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
7209 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7212 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
7213 -fno-devirtualize; try to devirtualize by the knowledge of
7214 virtual table pointer given by aggregate propagation.
7215 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7216 (ipa_print_node_jump_functions): Dump also offset that
7217 is relevant for polymorphic calls.
7218 (determine_known_aggregate_parts): Add arg_type parameter; use it
7219 instead of determining the type from pointer type.
7220 (ipa_compute_jump_functions_for_edge): Update call of
7221 determine_known_aggregate_parts.
7222 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
7223 (gimple_get_virt_method_for_binfo): ... here; simplify using
7224 vtable_pointer_value_to_vtable.
7225 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
7226 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
7227 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
7228 (vtable_pointer_value_to_vtable): Break out from ...; handle also
7230 (vtable_pointer_value_to_binfo): ... here.
7231 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
7233 2014-02-03 Teresa Johnson <tejohnson@google.com>
7235 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
7236 redef of outer loop index variable.
7238 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
7242 * doc/extend.texi (Function Attributes): Typo.
7244 2014-02-03 Cong Hou <congh@google.com>
7246 PR tree-optimization/60000
7247 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
7248 if the vectorized statement is a store. A store statement can only
7249 appear at the end of pattern statements.
7251 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7253 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
7254 (ix86_option_override_internal): Default long double to 64-bit for
7255 32-bit Bionic and to 128-bit for 64-bit Bionic.
7257 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
7258 TARGET_LONG_DOUBLE_128 is true.
7259 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
7261 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
7262 (mlong-double-64): Negate -mlong-double-128.
7263 (mlong-double-128): New option.
7265 * config/i386/i386-c.c (ix86_target_macros): Define
7266 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
7268 * doc/invoke.texi: Document -mlong-double-128.
7270 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7272 PR rtl-optimization/60024
7273 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
7275 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
7277 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
7279 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7281 PR rtl-optimization/57662
7282 * sel-sched.c (code_motion_path_driver): Do not mark already not
7283 existing blocks in the visiting bitmap.
7285 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7287 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
7288 on the insn being emitted.
7290 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
7291 Will Deacon <will.deacon@arm.com>
7293 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
7295 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7297 * config/arm/arm-tables.opt: Regenerate.
7299 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7301 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
7302 for vector types other than V16QImode.
7303 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
7304 define_expand, and call altivec_expand_vec_perm_le when producing
7305 code with little endian element order.
7306 (*altivec_vperm_<mode>_internal): New insn having previous
7307 behavior of altivec_vperm_<mode>.
7308 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
7309 altivec_expand_vec_perm_le when producing code with little endian
7311 (*altivec_vperm_<mode>_uns_internal): New insn having previous
7312 behavior of altivec_vperm_<mode>_uns.
7314 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7316 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
7317 (altivec_vsumsws): Add handling for -maltivec=be with a little
7319 (altivec_vsumsws_direct): New.
7320 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
7321 gen_altivec_vsumsws.
7323 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7325 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
7326 vtable_pointer_value_to_binfo): New functions.
7327 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
7328 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
7330 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
7332 * config/nios2/nios2.md (load_got_register): Initialize GOT
7333 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
7334 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
7336 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7338 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
7339 preserverd by passthrough, do not propagate the type.
7341 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7343 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
7344 (mips_atomic_assign_expand_fenv): New function.
7345 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
7347 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7349 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
7350 (__builtin_mips_set_fcsr): Likewise.
7351 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
7352 MIPS_USI_FTYPE_VOID.
7353 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
7354 (mips16_expand_set_fcsr): Likewise.
7355 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
7356 (mips16_set_fcsr_stub): Likewise.
7357 (mips16_get_fcsr_one_only_stub): New class.
7358 (mips16_set_fcsr_one_only_stub): Likewise.
7359 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
7360 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
7361 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
7362 (hard_float): New availability predicate.
7363 (mips_builtins): Add get_fcsr and set_fcsr.
7364 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
7365 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
7366 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
7367 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
7368 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
7371 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7373 * config/mips/mips.c (mips_one_only_stub): New class.
7374 (mips_need_mips16_rdhwr_p): Replace with...
7375 (mips16_rdhwr_stub): ...this new variable.
7376 (mips16_stub_call_address): New function.
7377 (mips16_rdhwr_one_only_stub): New class.
7378 (mips_expand_thread_pointer): Use mips16_stub_call_address.
7379 (mips_output_mips16_rdhwr): Delete.
7380 (mips_finish_stub): New function.
7381 (mips_code_end): Use it to handle rdhwr stubs.
7383 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
7386 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
7387 when calculating size of integer atomic types.
7389 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
7391 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
7393 2014-02-01 Jakub Jelinek <jakub@redhat.com>
7395 PR tree-optimization/60003
7396 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
7397 * profile.c (branch_prob): Use gimple_call_builtin_p
7398 to check for BUILT_IN_SETJMP_RECEIVER.
7399 * tree-inline.c (copy_bb): Call notice_special_calls.
7401 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
7404 * lra-constraints.c (process_alt_operands): Update reload_sum only
7407 2014-01-31 Richard Henderson <rth@redhat.com>
7410 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
7411 until after else_eh is processed.
7413 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7415 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
7416 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
7417 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
7418 in smmintrin.h, remove them.
7419 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
7420 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
7421 * config/i386/i386.md (ROUND_SAE): Fix value.
7422 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
7423 (const48_operand): New.
7424 * config/i386/subst.md (round), (round_expand): Use
7425 const_4_or_8_to_11_operand.
7426 (round_saeonly), (round_saeonly_expand): Use const48_operand.
7428 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7430 * config/i386/constraints.md (Yk): Swap meaning with k.
7431 * config/i386/i386.md (movhi_internal): Change Yk to k.
7432 (movqi_internal): Ditto.
7433 (*k<logic><mode>): Ditto.
7436 (kandn<mode>): Ditto.
7437 (*<code>hi_1): Ditto.
7438 (*<code>qi_1): Ditto.
7439 (kxnor<mode>): Ditto.
7440 (kortestzhi): Ditto.
7441 (kortestchi): Ditto.
7443 (*one_cmplhi2_1): Ditto.
7444 (*one_cmplqi2_1): Ditto.
7445 * config/i386/sse.md (): Change k to Yk.
7446 (avx512f_load<mode>_mask): Ditto.
7447 (avx512f_blendm<mode>): Ditto.
7448 (avx512f_store<mode>_mask): Ditto.
7449 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
7450 (avx512f_storedqu<mode>_mask): Ditto.
7451 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
7453 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
7454 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
7455 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
7456 (avx512f_maskcmp<mode>3): Ditto.
7457 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
7458 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
7459 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
7460 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
7461 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
7462 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
7463 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
7464 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
7465 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
7466 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
7467 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
7468 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
7469 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
7470 (vec_extract_lo_<mode>_maskm): Ditto.
7471 (vec_extract_hi_<mode>_maskm): Ditto.
7472 (avx512f_vternlog<mode>_mask): Ditto.
7473 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
7474 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
7475 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
7476 (avx512f_<code>v8div16qi2_mask): Ditto.
7477 (avx512f_<code>v8div16qi2_mask_store): Ditto.
7478 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
7479 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
7480 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
7481 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
7482 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7483 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7484 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7485 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7486 (avx512cd_maskb_vec_dupv8di): Ditto.
7487 (avx512cd_maskw_vec_dupv16si): Ditto.
7488 (avx512f_vpermi2var<mode>3_maskz): Ditto.
7489 (avx512f_vpermi2var<mode>3_mask): Ditto.
7490 (avx512f_vpermi2var<mode>3_mask): Ditto.
7491 (avx512f_vpermt2var<mode>3_maskz): Ditto.
7492 (*avx512f_gathersi<mode>): Ditto.
7493 (*avx512f_gathersi<mode>_2): Ditto.
7494 (*avx512f_gatherdi<mode>): Ditto.
7495 (*avx512f_gatherdi<mode>_2): Ditto.
7496 (*avx512f_scattersi<mode>): Ditto.
7497 (*avx512f_scatterdi<mode>): Ditto.
7498 (avx512f_compress<mode>_mask): Ditto.
7499 (avx512f_compressstore<mode>_mask): Ditto.
7500 (avx512f_expand<mode>_mask): Ditto.
7501 * config/i386/subst.md (mask): Change k to Yk.
7502 (mask_scalar_merge): Ditto.
7505 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
7507 * doc/extend.texi (Vector Extensions): Document ?: in C++.
7509 2014-01-31 Richard Biener <rguenther@suse.de>
7512 * builtins.c (fold_builtin_memory_op): Make sure to not
7513 use a floating-point mode or a boolean or enumeral type for
7516 2014-01-30 DJ Delorie <dj@redhat.com>
7518 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
7519 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
7520 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
7521 whenever main() has an epilogue.
7523 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7525 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
7526 unused variable "field".
7527 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
7528 (vsx_mergeh_<mode>): Likewise.
7529 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
7530 (altivec_vmrghh): Likewise.
7531 (altivec_vmrghw): Likewise.
7532 (altivec_vmrglb): Likewise.
7533 (altivec_vmrglh): Likewise.
7534 (altivec_vmrglw): Likewise.
7535 (altivec_vspltb): Add missing uses.
7536 (altivec_vsplth): Likewise.
7537 (altivec_vspltw): Likewise.
7538 (altivec_vspltsf): Likewise.
7540 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7543 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
7544 frame related instructions.
7546 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
7548 PR rtl-optimization/59959
7549 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
7550 any reload of register whose subreg is invalid.
7552 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7554 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
7555 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
7556 Add missing return type - void.
7558 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7560 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
7561 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
7562 remove element index adjustment for endian (now handled in vsx.md
7564 (altivec_expand_vec_perm_const): Use
7565 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
7566 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
7567 (vsx_xxspltw_<mode>): Adjust element index for little endian.
7568 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
7569 define_expand and a new define_insn *altivec_vspltb_internal;
7570 adjust for -maltivec=be on a little endian target.
7571 (altivec_vspltb_direct): New.
7572 (altivec_vsplth): Divide into a define_expand and a new
7573 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
7574 little endian target.
7575 (altivec_vsplth_direct): New.
7576 (altivec_vspltw): Divide into a define_expand and a new
7577 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
7578 little endian target.
7579 (altivec_vspltw_direct): New.
7580 (altivec_vspltsf): Divide into a define_expand and a new
7581 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
7582 a little endian target.
7584 2014-01-30 Richard Biener <rguenther@suse.de>
7586 PR tree-optimization/59993
7587 * tree-ssa-forwprop.c (associate_pointerplus): Check we
7588 can propagate form the earlier stmt and avoid the transform
7589 when the intermediate result is needed.
7591 2014-01-30 Alangi Derick <alangiderick@gmail.com>
7593 * README.Portability: Fix typo.
7595 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
7597 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
7598 comparison_operator with ordered_comparison_operator.
7600 2014-01-30 Nick Clifton <nickc@redhat.com>
7602 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
7603 Rename to mn10300_store_multiple_regs.
7604 * config/mn10300/mn10300.c: Likewise.
7605 * config/mn10300/mn10300.md (store_movm): Fix typo: call
7606 store_multiple_regs.
7607 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
7608 Call mn10300_store_multiple_regs.
7610 2014-01-30 Nick Clifton <nickc@redhat.com>
7611 DJ Delorie <dj@redhat.com>
7613 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
7614 %fp 2 to keep registers after it properly word-aligned.
7615 (rl78_alloc_physical_registers_umul): Handle the case where both
7616 input operands are the same.
7618 2014-01-30 Richard Biener <rguenther@suse.de>
7620 PR tree-optimization/59903
7621 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
7624 2014-01-30 Jason Merrill <jason@redhat.com>
7627 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
7630 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
7632 2014-01-30 Richard Biener <rguenther@suse.de>
7634 PR tree-optimization/59951
7635 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
7637 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
7640 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
7641 SFmode to DFmode case.
7643 2014-01-29 DJ Delorie <dj@redhat.com>
7645 * config/msp430/msp430.opt (-minrt): New.
7646 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
7648 (ENDFILE_SPEC): Likewise.
7650 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
7652 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
7653 (estimate_function_body_sizes): Use it.
7655 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
7658 * dwarf2out.c (is_cxx_auto): New.
7659 (is_base_type): Use it.
7660 (gen_type_die_with_usage): Likewise.
7662 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7664 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
7665 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
7666 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
7667 -maltivec=be with LE targets.
7668 (vsx_mergeh_<mode>): Likewise.
7669 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
7670 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
7671 (altivec_vmrghb): Replace with define_expand and new
7672 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
7673 (altivec_vmrghb_direct): New define_insn.
7674 (altivec_vmrghh): Replace with define_expand and new
7675 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
7676 (altivec_vmrghh_direct): New define_insn.
7677 (altivec_vmrghw): Replace with define_expand and new
7678 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
7679 (altivec_vmrghw_direct): New define_insn.
7680 (*altivec_vmrghsf): Adjust for endianness.
7681 (altivec_vmrglb): Replace with define_expand and new
7682 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
7683 (altivec_vmrglb_direct): New define_insn.
7684 (altivec_vmrglh): Replace with define_expand and new
7685 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
7686 (altivec_vmrglh_direct): New define_insn.
7687 (altivec_vmrglw): Replace with define_expand and new
7688 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
7689 (altivec_vmrglw_direct): New define_insn.
7690 (*altivec_vmrglsf): Adjust for endianness.
7691 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
7692 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
7693 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
7694 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
7695 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
7696 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
7697 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
7698 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
7700 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
7702 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
7703 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
7706 2014-01-29 Richard Biener <rguenther@suse.de>
7708 PR tree-optimization/58742
7709 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
7710 associate_pointerplus_align.
7711 (associate_pointerplus_diff): New function.
7712 (associate_pointerplus): Likewise. Call associate_pointerplus_align
7713 and associate_pointerplus_diff.
7715 2014-01-29 Richard Biener <rguenther@suse.de>
7717 * lto-streamer.h (LTO_major_version): Bump to 3.
7718 (LTO_minor_version): Reset to 0.
7720 2014-01-29 Renlin Li <Renlin.Li@arm.com>
7722 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
7723 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
7724 (arm_file_start): Generate correct asm header for armv7ve.
7725 * config/arm/bpabi.h: Add multilib support for armv7ve.
7726 * config/arm/driver-arm.c: Change the architectures of cortex-a7
7727 and cortex-a15 to armv7ve.
7728 * config/arm/t-aprofile: Add multilib support for armv7ve.
7729 * doc/invoke.texi: Document -march=armv7ve.
7731 2014-01-29 Richard Biener <rguenther@suse.de>
7733 PR tree-optimization/58742
7734 * tree-ssa-forwprop.c (associate_plusminus): Return true
7735 if we changed sth, defer EH cleanup to ...
7736 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
7737 (simplify_mult): New function.
7739 2014-01-29 Jakub Jelinek <jakub@redhat.com>
7742 PR tree-optimization/59920
7743 * tree.c (build_common_builtin_nodes): Remove
7744 __builtin_setjmp_dispatcher initialization.
7745 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
7746 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
7747 instead of gsi_after_labels + manually skipping debug stmts.
7748 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
7749 ignore bbs with IFN_ABNORMAL_DISPATCHER.
7750 * tree-inline.c (copy_edges_for_bb): Remove
7751 can_make_abnormal_goto argument, instead add abnormal_goto_dest
7752 argument. Ignore computed_goto_p stmts. Don't call
7753 make_abnormal_goto_edges. If a call might need abnormal edges
7754 for non-local gotos, see if it already has an edge to
7755 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
7756 with true argument, don't do anything then, otherwise add
7757 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
7758 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
7760 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
7761 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
7762 (lower_stmt): Don't set data->calls_builtin_setjmp.
7763 (lower_builtin_setjmp): Adjust comment.
7764 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
7765 * tree-cfg.c (found_computed_goto): Remove.
7766 (factor_computed_gotos): Remove.
7767 (make_goto_expr_edges): Return bool, true for computed gotos.
7768 Don't call make_abnormal_goto_edges.
7769 (build_gimple_cfg): Don't set found_computed_goto, don't call
7770 factor_computed_gotos.
7771 (computed_goto_p): No longer static.
7772 (make_blocks): Don't set found_computed_goto.
7773 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
7774 (make_edges): If make_goto_expr_edges returns true, push bb
7775 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
7776 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
7777 vector. Record mapping between bbs and OpenMP regions if there
7778 are any, adjust make_gimple_omp_edges caller. Call
7779 handle_abnormal_edges.
7780 (make_abnormal_goto_edges): Remove.
7781 * tree-cfg.h (make_abnormal_goto_edges): Remove.
7782 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
7783 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
7784 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
7785 * internal-fn.def (ABNORMAL_DISPATCHER): New.
7786 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
7787 filling *region also set *region_idx to (*region)->entry->index.
7790 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
7791 For REGs set ORIGINAL_REGNO.
7793 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
7795 * doc/md.texi: Mention that a target shouldn't implement
7796 vec_widen_(s|u)mul_even/odd pair if it is less efficient
7799 2014-01-29 Jakub Jelinek <jakub@redhat.com>
7801 PR tree-optimization/59594
7802 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
7803 a copy of the datarefs vector rather than the vector itself.
7805 2014-01-28 Jason Merrill <jason@redhat.com>
7808 * dwarf2out.c (auto_die): New static.
7809 (gen_type_die_with_usage): Handle C++1y 'auto'.
7810 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
7813 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
7816 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
7817 (SPEC_X32): Likewise.
7818 (SPEC_64): Likewise.
7819 * config/i386/i386.c (ix86_option_override_internal): Turn off
7820 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
7822 (x86_file_start): Output .code16gcc for TARGET_16BIT.
7823 * config/i386/i386.h (TARGET_16BIT): New macro.
7824 (TARGET_16BIT_P): Likewise.
7825 * config/i386/i386.opt: Add m16.
7826 * doc/invoke.texi: Document -m16.
7828 2014-01-28 Jakub Jelinek <jakub@redhat.com>
7830 PR preprocessor/59935
7831 * input.c (location_get_source_line): Bail out on when line number
7832 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
7834 2014-01-28 Richard Biener <rguenther@suse.de>
7836 PR tree-optimization/58742
7837 * tree-ssa-forwprop.c (associate_plusminus): Handle
7838 pointer subtraction of the form (T)(P + A) - (T)P.
7840 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7842 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
7845 2014-01-28 Richard Biener <rguenther@suse.de>
7848 2014-01-28 Richard Biener <rguenther@suse.de>
7850 PR rtl-optimization/45364
7851 PR rtl-optimization/59890
7852 * var-tracking.c (local_get_addr_clear_given_value): Handle
7853 already cleared slot.
7854 (val_reset): Handle not allocated local_get_addr_cache.
7855 (vt_find_locations): Use post-order on the inverted CFG.
7857 2014-01-28 Richard Biener <rguenther@suse.de>
7859 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
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 Alan Modra <amodra@gmail.com>
7872 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
7873 * configure.ac <recursive call for build != host>: Define
7874 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
7875 and LD_FOR_BUILD too.
7876 * configure: Regenerate.
7878 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
7880 * config/i386/i386.c (get_builtin_code_for_version): Separate
7881 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
7882 Broadwell from Haswell.
7884 2014-01-27 Steve Ellcey <sellcey@mips.com>
7886 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
7887 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
7888 * config/mips/mips.c (mips_option_override): Change setting
7890 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
7891 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
7892 Change from Mask to Var.
7894 2014-01-27 Jeff Law <law@redhat.com>
7896 * ipa-inline.c (inline_small_functions): Fix typo.
7898 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
7900 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
7901 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
7902 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
7903 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
7904 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
7905 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
7906 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
7907 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
7908 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
7909 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
7910 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
7911 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
7912 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
7913 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
7914 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
7915 (_mm512_storeu_epi64): Ditto.
7916 (_mm512_cmpge_epi32_mask): Ditto.
7917 (_mm512_cmpge_epu32_mask): Ditto.
7918 (_mm512_cmpge_epi64_mask): Ditto.
7919 (_mm512_cmpge_epu64_mask): Ditto.
7920 (_mm512_cmple_epi32_mask): Ditto.
7921 (_mm512_cmple_epu32_mask): Ditto.
7922 (_mm512_cmple_epi64_mask): Ditto.
7923 (_mm512_cmple_epu64_mask): Ditto.
7924 (_mm512_cmplt_epi32_mask): Ditto.
7925 (_mm512_cmplt_epu32_mask): Ditto.
7926 (_mm512_cmplt_epi64_mask): Ditto.
7927 (_mm512_cmplt_epu64_mask): Ditto.
7928 (_mm512_cmpneq_epi32_mask): Ditto.
7929 (_mm512_cmpneq_epu32_mask): Ditto.
7930 (_mm512_cmpneq_epi64_mask): Ditto.
7931 (_mm512_cmpneq_epu64_mask): Ditto.
7932 (_mm512_expand_pd): Ditto.
7933 (_mm512_expand_ps): Ditto.
7934 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
7935 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
7936 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
7937 * config/i386/i386.c (ix86_builtins): Add
7938 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
7939 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
7940 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
7941 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
7942 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
7943 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
7944 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
7945 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
7946 IX86_BUILTIN_PMOVUSQW512_MEM.
7947 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
7948 __builtin_ia32_pmovsqd512mem_mask,
7949 __builtin_ia32_pmovqd512mem_mask,
7950 __builtin_ia32_pmovusqw512mem_mask,
7951 __builtin_ia32_pmovsqw512mem_mask,
7952 __builtin_ia32_pmovqw512mem_mask,
7953 __builtin_ia32_pmovusdw512mem_mask,
7954 __builtin_ia32_pmovsdw512mem_mask,
7955 __builtin_ia32_pmovdw512mem_mask,
7956 __builtin_ia32_pmovqb512mem_mask,
7957 __builtin_ia32_pmovusqb512mem_mask,
7958 __builtin_ia32_pmovsqb512mem_mask,
7959 __builtin_ia32_pmovusdb512mem_mask,
7960 __builtin_ia32_pmovsdb512mem_mask,
7961 __builtin_ia32_pmovdb512mem_mask.
7962 (bdesc_args): Add __builtin_ia32_expanddf512,
7963 __builtin_ia32_expandsf512.
7964 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
7965 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
7966 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
7967 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
7968 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
7969 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
7970 (avx512f_<code>v8div16qi2_mask_store): This.
7971 (avx512f_expand<mode>): New.
7973 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
7975 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
7977 (_mm512_mask_prefetch_i64gather_pd): Ditto.
7978 (_mm512_prefetch_i32scatter_pd): Ditto.
7979 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
7980 (_mm512_prefetch_i64scatter_pd): Ditto.
7981 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
7982 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
7983 (_mm512_mask_prefetch_i64gather_ps): Ditto.
7984 (_mm512_prefetch_i32scatter_ps): Ditto.
7985 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
7986 (_mm512_prefetch_i64scatter_ps): Ditto.
7987 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
7988 * config/i386/i386-builtin-types.def: Define
7989 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
7990 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
7991 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
7992 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
7993 IX86_BUILTIN_SCATTERPFQPD.
7994 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
7995 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
7996 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
7997 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
7998 __builtin_ia32_scatterpfqps.
7999 (ix86_expand_builtin): Expand new built-ins.
8000 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8001 fix memory access data type.
8002 (*avx512pf_gatherpf<mode>_mask): Ditto.
8003 (*avx512pf_gatherpf<mode>): Ditto.
8004 (avx512pf_scatterpf<mode>): Ditto.
8005 (*avx512pf_scatterpf<mode>_mask): Ditto.
8006 (*avx512pf_scatterpf<mode>): Ditto.
8007 (GATHER_SCATTER_SF_MEM_MODE): New.
8008 (avx512pf_gatherpf<mode>df): Ditto.
8009 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8010 (*avx512pf_scatterpf<mode>df): Ditto.
8012 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8015 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8016 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8019 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8021 * common/config/arm/arm-common.c
8022 (arm_rewrite_mcpu): Handle multiple names.
8024 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8026 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8028 * gimple-builder.h (create_gimple_tmp): Delete.
8030 2014-01-27 Christian Bruel <christian.bruel@st.com>
8032 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
8035 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
8037 * config/pa/pa.md (call): Generate indirect long calls to non-local
8038 functions when outputing 32-bit code.
8039 (call_value): Likewise except for special call to buggy powf function.
8041 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
8042 portable runtime and PIC indirect calls.
8043 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
8044 and PIC call sequences. Use ldo instead of blr to set return register
8045 in PIC call sequence.
8047 2014-01-25 Walter Lee <walt@tilera.com>
8049 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
8050 avoid clobbering a live register.
8052 2014-01-25 Walter Lee <walt@tilera.com>
8054 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
8055 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
8056 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
8057 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
8059 2014-01-25 Walter Lee <walt@tilera.com>
8061 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
8062 arguments on even registers.
8063 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
8065 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
8066 (BIGGEST_ALIGNMENT): Ditto.
8067 (BIGGEST_FIELD_ALIGNMENT): Ditto.
8069 2014-01-25 Walter Lee <walt@tilera.com>
8071 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
8072 insns before bundling.
8073 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8075 2014-01-25 Walter Lee <walt@tilera.com>
8077 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
8078 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
8079 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
8081 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8083 * config/mips/constraints.md (kl): Delete.
8084 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
8085 define expands, using...
8086 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
8087 instructions for MIPS16.
8088 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
8089 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
8091 2014-01-25 Walter Lee <walt@tilera.com>
8093 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
8097 2014-01-25 Walter Lee <walt@tilera.com>
8099 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
8100 (TARGET_EXPAND_TO_RTL_HOOK): Define.
8102 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8104 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
8107 2014-01-25 Jakub Jelinek <jakub@redhat.com>
8109 * print-rtl.c (in_call_function_usage): New var.
8110 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
8111 EXPR_LIST mode as mode and not as reg note name.
8114 * cfgloopmanip.c (copy_loop_info): If
8115 loop->warned_aggressive_loop_optimizations, make sure
8116 the flag is set in target loop too.
8118 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
8120 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
8122 * builtins.def: Likewise.
8123 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
8124 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
8125 * ira.c (ira_setup_eliminable_regset): Likewise.
8126 * omp-low.c (gate_expand_omp): Likewise.
8127 (execute_lower_omp): Likewise.
8128 (diagnose_sb_0): Likewise.
8129 (gate_diagnose_omp_blocks): Likewise.
8130 (simd_clone_clauses_extract): Likewise.
8133 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8135 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
8136 correction for little endian...
8137 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
8140 2014-01-24 Jeff Law <law@redhat.com>
8142 PR tree-optimization/59919
8143 * tree-vrp.c (find_assert_locations_1): Do not register asserts
8144 for non-returning calls.
8146 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
8148 * common/config/aarch64/aarch64-common.c
8149 (aarch64_rewrite_mcpu): Handle multiple names.
8150 * config/aarch64/aarch64.h
8151 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8153 2014-01-24 Dodji Seketeli <dodji@redhat.com>
8155 * input.c (add_file_to_cache_tab): Handle the case where fopen
8158 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
8161 * config/i386/i386.md (pushsf splitter): Get stack adjustment
8162 from push operand if code of push isn't PRE_DEC.
8164 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8167 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
8168 -mquad-memory-atomic. Update -mquad-memory documentation to say
8169 it is only used for non-atomic loads/stores.
8171 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
8172 -mquad-memory or -mquad-memory-atomic switches.
8174 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
8175 -mquad-memory-atomic to ISA 2.07 support.
8177 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
8178 to separate support of normal quad word memory operations (ldq, stq)
8179 from the atomic quad word memory operations.
8181 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
8182 support to separate non-atomic quad word operations from atomic
8183 quad word operations. Disable non-atomic quad word operations in
8184 little endian mode so that we don't have to swap words after the
8185 load and before the store.
8186 (quad_load_store_p): Add comment about atomic quad word support.
8187 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
8188 options printed with -mdebug=reg.
8190 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
8191 -mquad-memory-atomic as the test for whether we have quad word
8192 atomic instructions.
8193 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
8194 or -mp8-vector are used, allow byte/half-word atomic operations.
8196 * config/rs6000/sync.md (load_lockedti): Insure that the address
8197 is a proper indexed or indirect address for the lqarx instruction.
8198 On little endian systems, swap the hi/lo registers after the lqarx
8200 (load_lockedpti): Use indexed_or_indirect_operand predicate to
8201 insure the address is valid for the lqarx instruction.
8202 (store_conditionalti): Insure that the address is a proper indexed
8203 or indirect address for the stqcrx. instruction. On little endian
8204 systems, swap the hi/lo registers before doing the stqcrx.
8206 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
8207 insure the address is valid for the stqcrx. instruction.
8209 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8210 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
8211 type of quad memory support is available.
8213 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
8216 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
8217 there is a danger of looping.
8219 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
8221 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8222 force flag_ira_loop_pressure if set via command line.
8224 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8226 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
8227 (ashr_simd): New builtin handling DI mode.
8228 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
8229 (aarch64_sshr_simddi): New match pattern.
8230 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
8231 (vshrd_n_s64): Likewise.
8232 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
8234 2014-01-23 Nick Clifton <nickc@redhat.com>
8236 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
8237 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
8238 favour of mcu specific scripts.
8239 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
8242 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8243 Alex Velenko <Alex.Velenko@arm.com>
8245 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
8246 (vaddv_s16): Likewise.
8247 (vaddv_s32): Likewise.
8248 (vaddv_u8): Likewise.
8249 (vaddv_u16): Likewise.
8250 (vaddv_u32): Likewise.
8251 (vaddvq_s8): Likewise.
8252 (vaddvq_s16): Likewise.
8253 (vaddvq_s32): Likewise.
8254 (vaddvq_s64): Likewise.
8255 (vaddvq_u8): Likewise.
8256 (vaddvq_u16): Likewise.
8257 (vaddvq_u32): Likewise.
8258 (vaddvq_u64): Likewise.
8259 (vaddv_f32): Likewise.
8260 (vaddvq_f32): Likewise.
8261 (vaddvq_f64): Likewise.
8262 (vmaxv_f32): Likewise.
8263 (vmaxv_s8): Likewise.
8264 (vmaxv_s16): Likewise.
8265 (vmaxv_s32): Likewise.
8266 (vmaxv_u8): Likewise.
8267 (vmaxv_u16): Likewise.
8268 (vmaxv_u32): Likewise.
8269 (vmaxvq_f32): Likewise.
8270 (vmaxvq_f64): Likewise.
8271 (vmaxvq_s8): Likewise.
8272 (vmaxvq_s16): Likewise.
8273 (vmaxvq_s32): Likewise.
8274 (vmaxvq_u8): Likewise.
8275 (vmaxvq_u16): Likewise.
8276 (vmaxvq_u32): Likewise.
8277 (vmaxnmv_f32): Likewise.
8278 (vmaxnmvq_f32): Likewise.
8279 (vmaxnmvq_f64): Likewise.
8280 (vminv_f32): Likewise.
8281 (vminv_s8): Likewise.
8282 (vminv_s16): Likewise.
8283 (vminv_s32): Likewise.
8284 (vminv_u8): Likewise.
8285 (vminv_u16): Likewise.
8286 (vminv_u32): Likewise.
8287 (vminvq_f32): Likewise.
8288 (vminvq_f64): Likewise.
8289 (vminvq_s8): Likewise.
8290 (vminvq_s16): Likewise.
8291 (vminvq_s32): Likewise.
8292 (vminvq_u8): Likewise.
8293 (vminvq_u16): Likewise.
8294 (vminvq_u32): Likewise.
8295 (vminnmv_f32): Likewise.
8296 (vminnmvq_f32): Likewise.
8297 (vminnmvq_f64): Likewise.
8299 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8301 * config/aarch64/aarch64-simd.md
8302 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
8303 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
8304 (*aarch64_mul3_elt<mode>): Likewise.
8305 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
8306 (*aarch64_mul3_elt_to_64v2df): Likewise.
8307 (*aarch64_mla_elt<mode>): Likewise.
8308 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8309 (*aarch64_mls_elt<mode>): Likewise.
8310 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8311 (*aarch64_fma4_elt<mode>): Likewise.
8312 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8313 (*aarch64_fma4_elt_to_64v2df): Likewise.
8314 (*aarch64_fnma4_elt<mode>): Likewise.
8315 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8316 (*aarch64_fnma4_elt_to_64v2df): Likewise.
8317 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
8318 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
8319 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8320 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8321 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
8322 (aarch64_sqdmull_lane<mode>_internal): Likewise.
8323 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
8325 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
8327 * config/aarch64/aarch64-simd.md
8328 (aarch64_be_checked_get_lane<mode>): New define_expand.
8329 * config/aarch64/aarch64-simd-builtins.def
8330 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
8331 New builtin definition.
8332 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
8333 Use new safe be builtin.
8335 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8337 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
8339 (aarch64_be_st1<mode>): Likewise.
8340 (aarch_ld1<VALL:mode>): Define_expand modified.
8341 (aarch_st1<VALL:mode>): Likewise.
8342 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
8343 (UNSPEC_ST1): Likewise.
8345 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
8347 * config/microblaze/microblaze.md: Add trap insn and attribute
8349 2014-01-23 Dodji Seketeli <dodji@redhat.com>
8351 PR preprocessor/58580
8352 * input.h (location_get_source_line): Take an additional line_size
8354 (void diagnostics_file_cache_fini): Declare new function.
8355 * input.c (struct fcache): New type.
8356 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
8357 New static constants.
8358 (diagnostic_file_cache_init, total_lines_num)
8359 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
8360 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
8361 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
8362 (get_next_line, read_next_line, goto_next_line, read_line_num):
8363 New static function definitions.
8364 (diagnostic_file_cache_fini): New function.
8365 (location_get_source_line): Take an additional output line_len
8366 parameter. Re-write using lookup_or_add_file_to_cache_tab and
8368 * diagnostic.c (diagnostic_finish): Call
8369 diagnostic_file_cache_fini.
8370 (adjust_line): Take an additional input parameter for the length
8371 of the line, rather than calculating it with strlen.
8372 (diagnostic_show_locus): Adjust the use of
8373 location_get_source_line and adjust_line with respect to their new
8374 signature. While displaying a line now, do not stop at the first
8375 null byte. Rather, display the zero byte as a space and keep
8376 going until we reach the size of the line.
8377 * Makefile.in: Add vec.o to OBJS-libcommon
8379 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8380 Ilya Tocar <ilya.tocar@intel.com>
8382 * config/i386/avx512fintrin.h (_mm512_kmov): New.
8383 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
8384 (__builtin_ia32_kmov16): Ditto.
8385 * config/i386/i386.md (UNSPEC_KMOV): New.
8388 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8390 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
8391 (_mm512_storeu_si512): Ditto.
8393 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
8396 * rtl.h (get_referenced_operands): Declare.
8397 * recog.c (get_referenced_operands): New function.
8398 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
8399 operands have been referenced when recording LO_SUM references.
8401 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
8403 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
8405 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8407 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
8408 Enable for generic and recent AMD targets.
8410 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8412 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
8413 ARG_SIZE note when adjustment was eliminated.
8415 2014-01-22 Jeff Law <law@redhat.com>
8417 PR tree-optimization/59597
8418 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
8419 in file. Accept new argument REGISTERING and use it to modify
8420 dump output appropriately.
8421 (register_jump_thread): Corresponding changes.
8422 (mark_threaded_blocks): Reinstate code to cancel unprofitable
8423 thread paths involving joiner blocks. Add code to dump cancelled
8424 jump threading paths.
8426 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
8428 PR rtl-optimization/59477
8429 * lra-constraints.c (inherit_in_ebb): Process call for living hard
8430 regs. Update reloads_num and potential_reload_hard_regs for all insns.
8432 2014-01-22 Tom Tromey <tromey@redhat.com>
8434 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
8436 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
8438 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8440 PR rtl-optimization/59896
8441 * lra-constraints.c (process_alt_operands): Check unused note for
8442 matched operands of insn with no output reloads.
8444 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
8446 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
8447 (mips_move_from_gpr_cost): Likewise.
8449 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8451 PR rtl-optimization/59858
8452 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
8453 ira_class_hard_regs_num.
8454 (process_alt_operands): Increase reject for dying matched operand.
8456 2014-01-21 Jakub Jelinek <jakub@redhat.com>
8459 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
8460 smaller than size, perform several stores or loads and stores
8461 at dst + count - size to store or copy all of size bytes, rather
8462 than just last modesize bytes.
8464 2014-01-20 DJ Delorie <dj@redhat.com>
8466 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
8467 that CLOBBERs are REGs before propogating their values.
8469 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
8472 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
8473 (cgraph_inline_failed_type): New function.
8474 * cgraph.h (DEFCIFCODE): Add type.
8475 (cgraph_inline_failed_type_t): New enum.
8476 (cgraph_inline_failed_type): New prototype.
8477 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
8478 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
8479 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
8480 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
8481 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
8482 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
8483 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
8484 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
8485 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
8486 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
8487 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
8488 OPTIMIZATION_MISMATCH.
8489 * tree-inline.c (expand_call_inline): Emit errors during
8490 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
8492 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8495 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
8496 mode attribute in insn output.
8498 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
8500 * output.h (output_constant): Delete.
8501 * varasm.c (output_constant): Make private.
8503 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
8505 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
8507 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8510 * tree.h (fold_builtin_strcat): New prototype.
8511 * builtins.c (fold_builtin_strcat): No longer static. Add len
8512 argument, if non-NULL, don't call c_strlen. Optimize
8513 directly into __builtin_memcpy instead of __builtin_strcpy.
8514 (fold_builtin_2): Adjust fold_builtin_strcat caller.
8515 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
8517 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8519 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8520 for SImode_address_operand operands, having only a REG argument.
8522 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
8524 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
8525 loader name using mbig-endian.
8526 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
8528 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8530 * doc/invoke.texi (-march): Clarify documentation for AArch64.
8534 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
8536 * config/aarch64/aarch64-protos.h
8537 (aarch64_cannot_change_mode_class_ptr): Declare.
8538 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
8539 aarch64_cannot_change_mode_class_ptr): New.
8540 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
8541 backend hook aarch64_cannot_change_mode_class.
8543 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8545 * common/config/aarch64/aarch64-common.c
8546 (aarch64_handle_option): Don't handle any option order logic here.
8547 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
8548 selected_cpu, warn on architecture version mismatch.
8549 (aarch64_override_options): Fix parsing order for option strings.
8551 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8552 Iain Sandoe <iain@codesourcery.com>
8555 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
8556 warning. Amend comment to reflect current functionality.
8558 2014-01-20 Richard Biener <rguenther@suse.de>
8561 * builtins.c (fold_builtin_strcat): Remove case better handled
8562 by tree-ssa-strlen.c.
8564 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
8566 * config/aarch64/aarch64.opt
8567 (mcpu, march, mtune): Make case-insensitive.
8569 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8572 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8573 if operands[1] is a REG or ZERO_EXTEND of a REG.
8575 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
8577 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
8579 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
8581 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
8582 long non-pic millicode calls.
8584 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8586 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
8588 2014-01-19 Kito Cheng <kito@0xlab.org>
8590 * builtins.c (expand_movstr): Check movstr expand done or fail.
8592 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8593 H.J. Lu <hongjiu.lu@intel.com>
8596 * config/i386/i386.md (*lea<mode>): Zero-extend return register
8597 to DImode for zero-extended addresses.
8599 2014-01-19 Jakub Jelinek <jakub@redhat.com>
8601 PR rtl-optimization/57763
8602 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
8603 on the new indirect jump_insn and increment LABEL_NUSES (label).
8605 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
8609 * config.gcc (x86_archs): New variable.
8610 (x86_64_archs): Likewise.
8611 (x86_cpus): Likewise.
8612 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
8613 --with-arch/--with-cpu= options.
8614 Support --with-arch=/--with-cpu={nehalem,westmere,
8615 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
8617 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8619 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
8620 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
8622 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8624 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
8626 2014-01-18 Jakub Jelinek <jakub@redhat.com>
8629 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
8630 clear cpp_get_options (parse_in)->warn_unused_macros for
8631 ix86_target_macros_internal with cpp_define.
8633 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
8635 * jump.c (delete_related_insns): Keep (use (insn))s.
8636 * reorg.c (redundant_insn): Check for barriers too.
8638 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8640 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
8642 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
8644 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
8645 call to $$dyncall when TARGET_LONG_CALLS is true.
8647 2014-01-17 Jeff Law <law@redhat.com>
8649 * ree.c (combine_set_extension): Temporarily disable test for
8650 changing number of hard registers.
8652 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
8655 * ipa-inline-analysis.c (inline_free_summary):
8656 Do not free summary of aliases.
8658 2014-01-17 Jakub Jelinek <jakub@redhat.com>
8661 * gimplify.c (gimplify_expr): Use create_tmp_var
8662 instead of create_tmp_var_raw. If cond doesn't have
8663 integral type, don't add the IFN_ANNOTATE builtin at all.
8665 2014-01-17 Martin Jambor <mjambor@suse.cz>
8668 * ipa-cp.c (prev_edge_clone): New variable.
8669 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
8670 Also resize prev_edge_clone vector.
8671 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
8672 (ipcp_edge_removal_hook): New function.
8673 (ipcp_driver): Register ipcp_edge_removal_hook.
8675 2014-01-17 Andrew Pinski <apinski@cavium.com>
8676 Steve Ellcey <sellcey@mips.com>
8679 * config/mips/mips.c (mips_print_operand): Check operand mode instead
8682 2014-01-17 Jeff Law <law@redhat.com>
8685 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
8686 so that pass_ccp runs first.
8688 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8690 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
8691 (ix86_adjust_cost): Use !TARGET_XXX.
8692 (do_reorder_for_imul): Likewise.
8693 (swap_top_of_ready_list): Likewise.
8694 (ix86_sched_reorder): Likewise.
8696 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8698 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
8699 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
8700 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
8701 (intel_memset): New. Duplicate slm_memset.
8702 (intel_cost): New. Duplicate slm_cost.
8703 (m_INTEL): New macro.
8704 (processor_target_table): Add "intel".
8705 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
8706 with PROCESSOR_INTEL for "intel".
8707 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
8708 PROCESSOR_SILVERMONT.
8709 (ix86_issue_rate): Likewise.
8710 (ix86_adjust_cost): Likewise.
8711 (ia32_multipass_dfa_lookahead): Likewise.
8712 (swap_top_of_ready_list): Likewise.
8713 (ix86_sched_reorder): Likewise.
8714 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
8715 instead of TARGET_OPT_AGU.
8716 * config/i386/i386.h (TARGET_INTEL): New.
8717 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
8718 (processor_type): Add PROCESSOR_INTEL.
8719 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
8720 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
8722 2014-01-17 Marek Polacek <polacek@redhat.com>
8725 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
8728 2014-01-17 Richard Biener <rguenther@suse.de>
8730 PR tree-optimization/46590
8731 * opts.c (default_options_table): Add entries for
8732 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
8733 all enabled at -O1 but not for -Og.
8734 * common.opt (fbranch-count-reg): Remove Init(1).
8735 (fmove-loop-invariants): Likewise.
8736 (ftree-pta): Likewise.
8738 2014-01-17 Jakub Jelinek <jakub@redhat.com>
8740 * config/i386/i386.c (ix86_data_alignment): For compatibility with
8741 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
8742 decls to at least the GCC 4.8 used alignments.
8745 * tree-nested.c (convert_nonlocal_reference_stmt,
8746 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
8747 of GIMPLE_BIND stmts, adjust associated decls.
8749 2014-01-17 Richard Biener <rguenther@suse.de>
8751 PR tree-optimization/46590
8752 * vec.h (vec<>::bseach): New member function implementing
8753 binary search according to C89 bsearch.
8754 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
8755 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
8756 bitmap pointer again. Make accesses_in_loop a flat array.
8757 (mem_ref_obstack): New global.
8758 (outermost_indep_loop): Adjust for mem_ref->stored changes.
8759 (mark_ref_stored): Likewise.
8760 (ref_indep_loop_p_2): Likewise.
8761 (set_ref_stored_in_loop): New helper function.
8762 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
8763 (memref_free): Adjust.
8764 (record_mem_ref_loc): Simplify.
8765 (gather_mem_refs_stmt): Adjust.
8766 (sort_locs_in_loop_postorder_cmp): New function.
8767 (analyze_memory_references): Sort accesses_in_loop after
8768 loop postorder number.
8769 (find_ref_loc_in_loop_cmp): New function.
8770 (for_all_locs_in_loop): Find relevant cluster of locs in
8771 accesses_in_loop and iterate without recursion.
8772 (execute_sm): Avoid uninit warning.
8773 (struct ref_always_accessed): Simplify.
8774 (ref_always_accessed::operator ()): Likewise.
8775 (ref_always_accessed_p): Likewise.
8776 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
8777 loop postorder numbers here.
8778 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
8781 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
8784 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
8785 on decls for which assemble_alias has been called.
8787 2014-01-17 Nick Clifton <nickc@redhat.com>
8789 * config/msp430/msp430.opt: (mcpu): New option.
8790 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
8791 (msp430_option_override): Parse target_cpu. If the MCU name
8792 matches a generic string, clear target_mcu.
8793 (msp430_attr): Allow numeric interrupt values up to 63.
8794 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
8795 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
8797 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
8799 * config/msp430/msp430.md (popm): Use %J rather than %I.
8800 (addsi3): Use msp430_nonimmediate_operand for operand 2.
8801 (addhi_cy_i): Use immediate_operand for operand 2.
8802 * doc/invoke.texi: Document -mcpu option.
8804 2014-01-17 Richard Biener <rguenther@suse.de>
8806 PR rtl-optimization/38518
8807 * df.h (df_analyze_loop): Declare.
8808 * df-core.c: Include cfgloop.h.
8809 (df_analyze_1): Split out main part of df_analyze.
8810 (df_analyze): Adjust.
8811 (loop_inverted_post_order_compute): New function.
8812 (loop_post_order_compute): Likewise.
8813 (df_analyze_loop): New function avoiding whole-function
8815 * loop-invariant.c (find_defs): Use df_analyze_loop.
8816 (find_invariants): Adjust.
8817 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
8819 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
8821 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
8822 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
8824 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
8826 * ipa-ref.c (ipa_remove_stmt_references): Fix references
8827 traversal when removing references.
8829 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
8832 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
8834 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
8837 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
8838 pushing a reload for an autoinc when we had previously reloaded an
8839 inner part of the address.
8841 2014-01-16 Jakub Jelinek <jakub@redhat.com>
8843 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
8845 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
8846 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
8847 when not giving up or versioning for alias only because of
8849 (vect_analyze_data_ref_dependences): Set to true.
8850 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
8852 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
8853 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
8857 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
8860 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
8861 operand 0 predicate for gathers, use a new pseudo as subtarget.
8863 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
8866 * lra-constraints.c (process_alt_operands): Add printing debug info.
8867 Check absence of input/output reloads for matched operands too.
8869 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
8871 PR rtl-optimization/59835
8872 * ira.c (ira_init_register_move_cost): Increase cost for
8875 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
8877 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
8879 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
8882 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
8883 non-register objects. Use gen_(high/low)part more consistently.
8886 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8889 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
8890 endian support, remove tests for WORDS_BIG_ENDIAN.
8891 (p8_mfvsrd_3_<mode>): Likewise.
8892 (reload_gpr_from_vsx<mode>): Likewise.
8893 (reload_gpr_from_vsxsf): Likewise.
8894 (p8_mfvsrd_4_disf): Likewise.
8896 2014-01-16 Richard Biener <rguenther@suse.de>
8898 PR rtl-optimization/46590
8899 * lcm.c (compute_antinout_edge): Use postorder iteration.
8900 (compute_laterin): Use inverted postorder iteration.
8902 2014-01-16 Nick Clifton <nickc@redhat.com>
8905 * varasm.c (output_constant): Return the number of bytes actually
8907 (output_constructor_array_range): Update the field size with the
8908 number of bytes emitted by output_constant.
8909 (output_constructor_regular_field): Likewise. Also do not
8910 complain if the total number of bytes emitted is now greater
8911 than the expected fieldpos.
8912 * output.h (output_constant): Update prototype and descriptive comment.
8914 2014-01-16 Marek Polacek <polacek@redhat.com>
8917 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
8918 it is error_mark_node.
8920 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
8922 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
8923 VALID_AVX256_REG_OR_OI_MODE.
8925 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
8927 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
8928 current procedure should be profiled.
8930 2014-01-15 Andrew Pinski <apinski@cavium.com>
8932 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
8933 of moving from/to the STACK_REG register class.
8935 2014-01-15 Richard Henderson <rth@redhat.com>
8938 * reginfo.c (global_regs_decl): Globalize.
8939 * rtl.h (global_regs_decl): Declare.
8940 * ira.c (do_reload): Diagnose frame_pointer_needed and it
8941 reserved via global_regs.
8943 2014-01-15 Teresa Johnson <tejohnson@google.com>
8945 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
8947 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8949 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
8950 and vmulosh rather than call gen_vec_widen_smult_*.
8951 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
8952 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
8953 (vec_widen_smult_even_v16qi): Likewise.
8954 (vec_widen_umult_even_v8hi): Likewise.
8955 (vec_widen_smult_even_v8hi): Likewise.
8956 (vec_widen_umult_odd_v16qi): Likewise.
8957 (vec_widen_smult_odd_v16qi): Likewise.
8958 (vec_widen_umult_odd_v8hi): Likewise.
8959 (vec_widen_smult_odd_v8hi): Likewise.
8960 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
8961 vmuloub rather than call gen_vec_widen_umult_*.
8962 (vec_widen_umult_lo_v16qi): Likewise.
8963 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
8964 vmulosb rather than call gen_vec_widen_smult_*.
8965 (vec_widen_smult_lo_v16qi): Likewise.
8966 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
8967 rather than call gen_vec_widen_umult_*.
8968 (vec_widen_umult_lo_v8hi): Likewise.
8969 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
8970 rather than call gen_vec_widen_smult_*.
8971 (vec_widen_smult_lo_v8hi): Likewise.
8973 2014-01-15 Jeff Law <law@redhat.com>
8975 PR tree-optimization/59747
8976 * ree.c (find_and_remove_re): Properly handle case where a second
8977 eliminated extension requires widening a copy created for elimination
8978 of a prior extension.
8979 (combine_set_extension): Ensure that the number of hard regs needed
8980 for a destination register does not change when we widen it.
8982 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
8984 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
8985 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
8986 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
8987 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
8988 (avr-*-rtems*): Likewise.
8989 (bfin*-rtems*): Likewise.
8990 (moxie-*-rtems*): Likewise.
8991 (h8300-*-rtems*): Likewise.
8992 (i[34567]86-*-rtems*): Likewise.
8993 (lm32-*-rtems*): Likewise.
8994 (m32r-*-rtems*): Likewise.
8995 (m68k-*-rtems*): Likewise.
8996 (microblaze*-*-rtems*): Likewise.
8997 (mips*-*-rtems*): Likewise.
8998 (powerpc-*-rtems*): Likewise.
8999 (sh-*-rtems*): Likewise.
9000 (sparc-*-rtems*): Likewise.
9001 (sparc64-*-rtems*): Likewise.
9002 (v850-*-rtems*): Likewise.
9003 (m32c-*-rtems*): Likewise.
9005 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9007 PR rtl-optimization/59511
9008 * ira.c (ira_init_register_move_cost): Use memory costs for some
9009 cases of register move cost calculations.
9010 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9011 instead of BB frequency.
9012 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9013 * lra-assigns.c (find_hard_regno_for): Ditto.
9015 2014-01-15 Richard Biener <rguenther@suse.de>
9017 PR tree-optimization/59822
9018 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9019 (vectorizable_load): Use it to hoist defs of uses of invariant
9020 loads out of the loop.
9022 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9023 Kugan Vivekanandarajah <kuganv@linaro.org>
9026 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9029 2014-01-15 Richard Biener <rguenther@suse.de>
9031 PR rtl-optimization/59802
9032 * lcm.c (compute_available): Use inverted postorder to seed
9033 the initial worklist.
9035 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9038 * config/s390/s390.c (s390_preferred_reload_class): Don't return
9039 ADDR_REGS for invalid symrefs in non-PIC code.
9041 2014-01-15 Jakub Jelinek <jakub@redhat.com>
9044 * builtins.c (determine_block_size): Initialize *probable_max_size
9045 even if len_rtx is CONST_INT.
9047 2014-01-14 Andrew Pinski <apinski@cavium.com>
9049 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
9050 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
9051 (cortexa53_tunings): Likewise.
9052 (aarch64_sched_issue_rate): New function.
9053 (TARGET_SCHED_ISSUE_RATE): Define.
9055 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9057 * ira-costs.c (find_costs_and_classes): Add missed
9058 ira_init_register_move_cost_if_necessary.
9060 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9063 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
9065 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
9068 * config/i386/i386.c (type_natural_mode): Add a bool parameter
9069 to indicate if type is used for function return value. Warn ABI
9070 change if the vector mode isn't available for function return value.
9071 (ix86_function_arg_advance): Pass false to type_natural_mode.
9072 (ix86_function_arg): Likewise.
9073 (ix86_gimplify_va_arg): Likewise.
9074 (function_arg_32): Don't warn ABI change.
9075 (ix86_function_value): Pass true to type_natural_mode.
9076 (ix86_return_in_memory): Likewise.
9077 (ix86_struct_value_rtx): Removed.
9078 (TARGET_STRUCT_VALUE_RTX): Likewise.
9080 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9082 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
9083 converting a conditional jump into a conditional return.
9085 2014-01-14 Richard Biener <rguenther@suse.de>
9087 PR tree-optimization/58921
9088 PR tree-optimization/59006
9089 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
9090 hoisting invariant stmts.
9091 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
9092 invariant loads on the preheader edge if possible.
9094 2014-01-14 Joey Ye <joey.ye@arm.com>
9096 * doc/plugin.texi (Building GCC plugins): Update to C++.
9098 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
9100 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
9101 (_mm_rcp28_round_ss): Ditto.
9102 (_mm_rsqrt28_round_sd): Ditto.
9103 (_mm_rsqrt28_round_ss): Ditto.
9104 (_mm_rcp28_sd): Ditto.
9105 (_mm_rcp28_ss): Ditto.
9106 (_mm_rsqrt28_sd): Ditto.
9107 (_mm_rsqrt28_ss): Ditto.
9108 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
9109 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
9110 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
9111 (IX86_BUILTIN_RCP28SD): Ditto.
9112 (IX86_BUILTIN_RCP28SS): Ditto.
9113 (IX86_BUILTIN_RSQRT28SD): Ditto.
9114 (IX86_BUILTIN_RSQRT28SS): Ditto.
9115 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
9116 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
9117 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
9118 (ix86_expand_special_args_builtin): Expand new FTYPE.
9119 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
9120 (srcp14<mode>): Make insn unary.
9121 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
9122 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
9123 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
9124 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
9125 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
9126 Fix rounding: make it SAE only.
9127 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9129 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9131 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
9132 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
9133 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
9134 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
9135 (round_saeonly_mask_scalar_operand4): Ditto.
9136 (round_saeonly_mask_scalar_op3): Ditto.
9137 (round_saeonly_mask_scalar_op4): Ditto.
9139 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9141 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9142 Implement -maltivec=be for vec_insert and vec_extract.
9144 2014-01-10 DJ Delorie <dj@redhat.com>
9146 * config/msp430/msp430.md (call_internal): Don't allow memory
9147 references with SP as the base register.
9148 (call_value_internal): Likewise.
9149 * config/msp430/constraints.md (Yc): New. For memory references
9150 that don't use SP as a base register.
9152 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
9153 "an integer without a # prefix"
9154 * config/msp430/msp430.md (epilogue_helper): Use it.
9156 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9159 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
9160 AVX512F gather builtins.
9161 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
9162 on gather decls with INTEGER_TYPE masktype.
9163 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
9164 directly into the builtin rather than hoisting it before loop.
9166 PR tree-optimization/59387
9167 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
9168 (scev_const_prop): If folded_casts and type has undefined overflow,
9169 use force_gimple_operand instead of force_gimple_operand_gsi and
9170 for each added stmt if it is assign with
9171 arith_code_with_undefined_signed_overflow, call
9172 rewrite_to_defined_overflow.
9173 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
9174 gimple-fold.h instead.
9175 (arith_code_with_undefined_signed_overflow,
9176 rewrite_to_defined_overflow): Moved to ...
9177 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
9178 rewrite_to_defined_overflow): ... here. No longer static.
9179 Include gimplify-me.h.
9180 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
9181 rewrite_to_defined_overflow): New prototypes.
9183 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9185 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
9187 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
9189 * builtins.c (get_object_alignment_2): Minor tweak.
9190 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
9192 2014-01-13 Christian Bruel <christian.bruel@st.com>
9194 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
9195 optimized non constant lengths.
9197 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9200 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
9201 load as __atomic_load_N if possible.
9203 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9205 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
9207 (rs6000_expand_builtin): Adjust call.
9209 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9212 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
9213 * config/rs6000/rs6000.c: Include target-globals.h.
9214 (rs6000_set_current_function): Instead of doing target_reinit
9215 unconditionally, use save_target_globals_default_opts and
9216 restore_target_globals.
9218 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
9220 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
9221 (rs6000_expand_builtin): Handle mffs and mtfsf.
9222 (rs6000_init_builtins): Define mffs and mtfsf.
9223 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
9224 (rs6000_mffs): New pattern.
9225 (rs6000_mtfsf): New pattern.
9227 2014-01-11 Bin Cheng <bin.cheng@arm.com>
9229 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
9230 Start narrowing with START. Apply candidate-use pair
9231 and check overall cost in narrowing.
9232 (iv_ca_prune): Pass new argument.
9234 2014-01-10 Jeff Law <law@redhat.com>
9237 * ree.c (combine_reaching_defs): Ensure the defining statement
9238 occurs before the extension when optimizing extensions with
9239 different source and destination hard registers.
9241 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9244 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
9245 vtables into the type inheritance graph.
9247 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9249 PR rtl-optimization/59754
9250 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
9251 modes in the REGNO != REGNO case.
9253 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9255 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
9257 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9259 PR tree-optimization/59745
9260 * tree-predcom.c (tree_predictive_commoning_loop): Call
9261 free_affine_expand_cache if giving up because components is NULL.
9263 * target-globals.c (save_target_globals): Allocate < 4KB structs using
9264 GC in payload of target_globals struct instead of allocating them on
9265 the heap and the larger structs separately using GC.
9266 * target-globals.h (struct target_globals): Make regs, hard_regs,
9267 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
9268 of GTY((skip)) and change type to void *.
9269 (reset_target_globals): Cast loads from those fields to corresponding
9272 2014-01-10 Steve Ellcey <sellcey@mips.com>
9275 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
9276 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
9277 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
9279 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
9282 * aarch64-modes.def (CC_Zmode): New flags mode.
9283 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
9284 represents an equality.
9285 (aarch64_get_condition_code): Handle CC_Zmode.
9286 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
9288 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9290 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
9291 extraction in good case.
9293 2014-01-10 Richard Biener <rguenther@suse.de>
9295 PR tree-optimization/59374
9296 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
9297 checking after SLP discovery. Mark stmts not participating
9298 in any SLP instance properly.
9300 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9302 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
9303 when handling a SET rtx.
9305 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9307 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
9308 (cortex-a57): Likewise.
9309 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
9311 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9313 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
9314 non-iwmmxt builtins.
9316 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9320 * ipa-devirt.c record_target_from_binfo): Take as argument
9321 stack of binfos and lookup matching one for virtual inheritance.
9322 (possible_polymorphic_call_targets_1): Update.
9324 2014-01-10 Huacai Chen <chenhc@lemote.com>
9326 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
9327 kernel strings for Loongson-2E/2F/3A.
9329 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9332 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
9333 is_gimple_call before calling gimple_call_internal_p.
9335 2014-01-09 Steve Ellcey <sellcey@mips.com>
9337 * Makefile.in (TREE_FLOW_H): Remove.
9338 (TREE_SSA_H): Add file names from tree-flow.h.
9339 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
9340 * tree.h: Remove tree-flow.h reference.
9341 * hash-table.h: Remove tree-flow.h reference.
9342 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
9343 reference with tree-ssa-loop.h.
9345 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9347 * doc/invoke.texi: Add -maltivec={be,le} options, and document
9348 default element-order behavior for -maltivec.
9349 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
9350 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
9351 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
9352 when targeting big endian, at least for now.
9353 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
9355 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9358 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
9359 var satisfies use_register_for_decl, just take into account type
9360 alignment, rather than decl alignment.
9362 PR tree-optimization/59622
9363 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
9364 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
9365 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
9366 Don't devirtualize for inplace at all. For targets.length () == 1,
9367 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
9369 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9371 * config/i386/i386.md (cpu): Remove the unused btver1.
9373 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9375 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
9377 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9380 * tree-core.h (struct target_globals): New forward declaration.
9381 (struct tree_target_option): Add globals field.
9382 * tree.h (TREE_TARGET_GLOBALS): Define.
9383 (prepare_target_option_nodes_for_pch): New prototype.
9384 * target-globals.h (struct target_globals): Define even if
9386 * tree.c (prepare_target_option_node_for_pch,
9387 prepare_target_option_nodes_for_pch): New functions.
9388 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
9389 * config/i386/i386.c: Include target-globals.h.
9390 (ix86_set_current_function): Instead of doing target_reinit
9391 unconditionally, use save_target_globals_default_opts and
9392 restore_target_globals.
9394 2014-01-09 Richard Biener <rguenther@suse.de>
9396 PR tree-optimization/59715
9397 * tree-cfg.h (split_critical_edges): Declare.
9398 * tree-cfg.c (split_critical_edges): Export.
9399 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
9401 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
9403 * cfgexpand.c (expand_stack_vars): Optionally disable
9404 asan stack protection.
9405 (expand_used_vars): Likewise.
9406 (partition_stack_vars): Likewise.
9407 * asan.c (asan_emit_stack_protection): Optionally disable
9408 after return stack usage.
9409 (instrument_derefs): Optionally disable memory access instrumentation.
9410 (instrument_builtin_call): Likewise.
9411 (instrument_strlen_call): Likewise.
9412 (asan_protect_global): Optionally disable global variables protection.
9413 * doc/invoke.texi: Added doc for new options.
9414 * params.def: Added new options.
9415 * params.h: Likewise.
9417 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9419 PR rtl-optimization/59724
9420 * ifcvt.c (cond_exec_process_if_block): Don't call
9421 flow_find_head_matching_sequence with 0 longest_match.
9422 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
9423 non-active insns if !stop_after.
9424 (try_head_merge_bb): Revert 2014-01-07 changes.
9426 2014-01-08 Jeff Law <law@redhat.com>
9428 * ree.c (get_sub_rtx): New function, extracted from...
9429 (merge_def_and_ext): Here.
9430 (combine_reaching_defs): Use get_sub_rtx.
9432 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
9434 * cgraph.h (varpool_variable_node): Do not choke on null node.
9436 2014-01-08 Catherine Moore <clm@codesourcery.com>
9438 * config/mips/mips.md (simple_return): Attempt to use JRC
9440 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
9442 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9444 PR rtl-optimization/59137
9445 * reorg.c (steal_delay_list_from_target): Call update_block for
9447 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
9449 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9451 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
9452 two duplicate entries.
9454 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9457 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
9459 * config/mips/mips.c (mips_truncated_op_cost): New function.
9460 (mips_rtx_costs): Adjust test for BADDU.
9461 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
9463 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
9465 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
9466 (*baddu_si): ...this new pattern.
9468 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9471 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
9473 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9476 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
9478 (expand_expr, expand_normal): Adjust.
9479 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
9480 inner_reference_p. Use inner_reference_p to expand inner references.
9481 (store_expr): Adjust.
9482 * cfgexpand.c (expand_call_stmt): Adjust.
9484 2014-01-08 Rong Xu <xur@google.com>
9486 * gcov-io.c (gcov_var): Move from gcov-io.h.
9487 (gcov_position): Ditto.
9488 (gcov_is_error): Ditto.
9489 (gcov_rewrite): Ditto.
9490 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
9491 only part to libgcc/libgcov.h.
9493 2014-01-08 Marek Polacek <polacek@redhat.com>
9496 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
9498 2014-01-08 Marek Polacek <polacek@redhat.com>
9501 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
9503 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9505 PR rtl-optimization/59649
9506 * stor-layout.c (get_mode_bounds): For BImode return
9507 0 and STORE_FLAG_VALUE.
9509 2014-01-08 Richard Biener <rguenther@suse.de>
9512 * gimple.h (is_gimple_builtin_call): Remove.
9513 (gimple_builtin_call_types_compatible_p): New.
9514 (gimple_call_builtin_p): New overload.
9515 * gimple.c (is_gimple_builtin_call): Remove.
9516 (validate_call): Rename to ...
9517 (gimple_builtin_call_types_compatible_p): ... this and export. Also
9519 (validate_type): New static function.
9520 (gimple_call_builtin_p): New overload and adjust.
9521 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
9522 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
9523 (gimple_fold_stmt_to_constant_1): Likewise.
9524 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
9526 2014-01-08 Richard Biener <rguenther@suse.de>
9529 * gimplify.c (gimplify_expr): Gimplify register-register type
9530 VIEW_CONVERT_EXPRs to separate stmts.
9532 2014-01-07 Jeff Law <law@redhat.com>
9535 * ree.c (combine_set_extension): Handle case where source
9536 and destination registers in an extension insn are different.
9537 (combine_reaching_defs): Allow source and destination registers
9538 in extension to be different under limited circumstances.
9539 (add_removable_extension): Remove restriction that the
9540 source and destination registers in the extension are the same.
9541 (find_and_remove_re): Emit a copy from the extension's
9542 destination to its source after the defining insn if
9543 the source and destination registers are different.
9546 * ifcvt.c (merge_if_block): If we are merging a block with more than
9547 one successor with a block with no successors, remove any BARRIER
9548 after the second block.
9550 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
9552 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
9554 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
9557 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
9558 for 14-bit register offsets when INT14_OK_STRICT is false.
9560 2014-01-07 Roland Stigge <stigge@antcom.de>
9561 Michael Meissner <meissner@linux.vnet.ibm.com>
9564 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
9565 Only check TFmode for SPE constants. Don't check TImode or TDmode.
9567 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
9569 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
9572 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
9574 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
9575 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
9576 rtx is const0_rtx or not.
9578 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
9581 * target-globals.c (save_target_globals): Remove this_fn_optab
9583 * toplev.c: Include optabs.h.
9584 (target_reinit): Temporarily restore the global options if another
9585 set of options are in force.
9587 2014-01-07 Jakub Jelinek <jakub@redhat.com>
9589 PR rtl-optimization/58668
9590 * cfgcleanup.c (flow_find_cross_jump): Don't count
9591 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
9592 to determine what is counted.
9593 (flow_find_head_matching_sequence): Use active_insn_p to determine
9595 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
9597 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
9598 determine what is counted.
9600 PR tree-optimization/59643
9601 * tree-predcom.c (split_data_refs_to_components): If one dr is
9602 read and one write, determine_offset fails and the write isn't
9603 in the bad component, just put the read into the bad component.
9605 2014-01-07 Mike Stump <mikestump@comcast.net>
9606 Jakub Jelinek <jakub@redhat.com>
9609 * tree-core.h (struct tree_optimization_option): Change optabs
9610 type from unsigned char * to void *.
9611 * optabs.c (init_tree_optimization_optabs): Adjust
9612 TREE_OPTIMIZATION_OPTABS initialization.
9614 2014-01-06 Jakub Jelinek <jakub@redhat.com>
9617 * config/i386/i386.h (struct machine_function): Add
9618 no_drap_save_restore field.
9619 * config/i386/i386.c (ix86_save_reg): Use
9620 !cfun->machine->no_drap_save_restore instead of
9621 crtl->stack_realign_needed.
9622 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
9623 this function clears frame_pointer_needed. Set
9624 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
9625 and DRAP reg is needed.
9627 2014-01-06 Marek Polacek <polacek@redhat.com>
9630 * doc/implement-c.texi: Mention that other integer types are
9631 permitted as bit-field types in strictly conforming mode.
9633 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
9635 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
9638 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
9640 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
9642 2014-01-06 Martin Jambor <mjambor@suse.cz>
9645 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
9647 * ipa-prop.c (ipa_print_node_params): Fix indentation.
9649 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
9653 * var-tracking.c (add_stores): Preserve the value of the source even if
9654 we don't record the store.
9656 2014-01-06 Terry Guo <terry.guo@arm.com>
9658 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
9660 2014-01-05 Iain Sandoe <iain@codesourcery.com>
9663 * config/darwin.c (darwin_function_section): Adjust return values to
9664 correspond to optimisation changes made in r206070.
9666 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
9668 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
9669 from prefetch_block tune setting.
9670 (nocona_cost): Correct size of prefetch block to 64.
9672 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
9674 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
9675 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
9676 used to save the static chain register in the computation of the offset
9677 from which the FP registers need to be restored.
9679 2014-01-04 Jakub Jelinek <jakub@redhat.com>
9681 PR tree-optimization/59519
9682 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
9683 ICE if get_current_def (current_new_name) is already non-NULL, as long
9684 as it is a phi result of some other phi in *new_exit_bb that has
9687 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
9688 or vmovdqu* for misaligned_operand.
9689 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
9690 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
9691 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
9692 aligned_mem for AVX512F masked aligned load and store builtins and for
9695 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
9697 PR tree-optimization/59651
9698 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
9699 Address range for negative step should be added by TYPE_SIZE_UNIT.
9701 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
9703 * config/m68k/m68k.c (handle_move_double): Handle pushes with
9704 overlapping registers also for registers other than the stack pointer.
9706 2014-01-03 Marek Polacek <polacek@redhat.com>
9709 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
9712 2014-01-03 Jakub Jelinek <jakub@redhat.com>
9715 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
9718 * config/i386/i386.md (MODE_SIZE): New mode attribute.
9719 (push splitter): Use <P:MODE_SIZE> instead of
9720 GET_MODE_SIZE (<P:MODE>mode).
9721 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
9722 (mov -1, reg peephole2): Likewise.
9723 * config/i386/sse.md (*mov<mode>_internal,
9724 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
9725 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
9726 *<code><mode>3, *andnot<mode>3<mask_name>,
9727 <mask_codefor><code><mode>3<mask_name>): Likewise.
9728 * config/i386/subst.md (mask_mode512bit_condition,
9729 sd_mask_mode512bit_condition): Likewise.
9731 2014-01-02 Xinliang David Li <davidxl@google.com>
9733 PR tree-optimization/59303
9734 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
9735 (dump_predicates): Better output format.
9736 (pred_equal_p): New function.
9737 (is_neq_relop_p): Ditto.
9738 (is_neq_zero_form_p): Ditto.
9739 (pred_expr_equal_p): Ditto.
9740 (pred_neg_p): Ditto.
9741 (simplify_pred): Ditto.
9742 (simplify_preds_2): Ditto.
9743 (simplify_preds_3): Ditto.
9744 (simplify_preds_4): Ditto.
9745 (simplify_preds): Ditto.
9747 (push_to_worklist): Ditto.
9748 (get_pred_info_from_cmp): Ditto.
9749 (is_degenerated_phi): Ditto.
9750 (normalize_one_pred_1): Ditto.
9751 (normalize_one_pred): Ditto.
9752 (normalize_one_pred_chain): Ditto.
9753 (normalize_preds): Ditto.
9754 (normalize_cond_1): Remove function.
9755 (normalize_cond): Ditto.
9756 (is_gcond_subset_of): Ditto.
9757 (is_subset_of_any): Ditto.
9758 (is_or_set_subset_of): Ditto.
9759 (is_and_set_subset_of): Ditto.
9760 (is_norm_cond_subset_of): Ditto.
9761 (pred_chain_length_cmp): Ditto.
9762 (convert_control_dep_chain_into_preds): Type change.
9763 (find_predicates): Ditto.
9764 (find_def_preds): Ditto.
9765 (destroy_predicates_vecs): Ditto.
9766 (find_matching_predicates_in_rest_chains): Ditto.
9767 (use_pred_not_overlap_with_undef_path_pred): Ditto.
9768 (is_pred_expr_subset): Ditto.
9769 (is_pred_chain_subset_of): Ditto.
9770 (is_included_in): Ditto.
9771 (is_superset_of): Ditto.
9773 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
9775 Update copyright years.
9777 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
9779 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
9780 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
9781 config/arc/arc.md, config/arc/arc.opt,
9782 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
9783 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
9784 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
9785 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
9786 config/linux-protos.h, config/linux.c, config/winnt-c.c,
9787 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
9788 vtable-verify.c, vtable-verify.h: Use the standard form for the
9791 2014-01-02 Tobias Burnus <burnus@net-b.de>
9793 * gcc.c (process_command): Update copyright notice dates.
9794 * gcov-dump.c: Ditto.
9796 * doc/cpp.texi: Bump @copying's copyright year.
9797 * doc/cppinternals.texi: Ditto.
9798 * doc/gcc.texi: Ditto.
9799 * doc/gccint.texi: Ditto.
9800 * doc/gcov.texi: Ditto.
9801 * doc/install.texi: Ditto.
9802 * doc/invoke.texi: Ditto.
9804 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9806 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
9808 2014-01-01 Jakub Jelinek <jakub@redhat.com>
9810 * config/i386/sse.md (*mov<mode>_internal): Guard
9811 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
9813 PR rtl-optimization/59647
9814 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
9815 new_rtx into UNSIGNED_FLOAT rtxes.
9817 Copyright (C) 2014 Free Software Foundation, Inc.
9819 Copying and distribution of this file, with or without modification,
9820 are permitted in any medium without royalty provided the copyright
9821 notice and this notice are preserved.