1 2014-05-07 David Malcolm <dmalcolm@redhat.com>
3 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
4 const_gimple, rather than a gimple.
5 (gimple_call_builtin_p): Likewise, for the three variants.
7 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
8 (gimple_call_builtin_p): Likewise, for the three variants.
10 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
12 PR tree-optimization/61095
13 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
15 2014-05-07 Richard Biener <rguenther@suse.de>
17 PR tree-optimization/61034
18 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
19 (maybe_skip_until): Use translate to take into account
20 lattices when trying to do disambiguations.
21 (get_continuation_for_phi_1): Likewise.
22 (get_continuation_for_phi): Adjust for added translate
24 (walk_non_aliased_vuses): Likewise.
25 * tree-ssa-alias.h (get_continuation_for_phi): Adjust
27 (walk_non_aliased_vuses): Likewise.
28 (call_may_clobber_ref_p_1): Declare.
29 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also
30 disambiguate against calls. Stop early if we are
31 only supposed to disambiguate.
32 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
34 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
36 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
37 Emit an error when the function has arguments.
39 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
41 * cfgloop.h (unswitch_loops): Remove.
42 * doc/passes.texi: Remove references to loop-unswitch.c
43 * timevar.def (TV_LOOP_UNSWITCH): Remove.
45 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
47 * tree-vect-data-refs.c (vect_grouped_load_supported): New
48 check for loads group of length 3.
49 (vect_permute_load_chain): New permutations for loads group of
51 * tree-vect-stmts.c (vect_model_load_cost): Change cost
52 of vec_perm_shuffle for the new permutations.
54 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
56 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
57 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
58 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
59 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
60 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
61 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
62 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
63 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
65 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
67 * loop-unswitch.c: Delete.
69 2014-05-07 Richard Biener <rguenther@suse.de>
71 * config.gcc: Always set need_64bit_hwint to yes.
73 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
75 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
76 of using optimize_size.
78 2014-05-06 Mike Stump <mikestump@comcast.net>
80 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
82 2014-05-06 Joseph Myers <joseph@codesourcery.com>
84 * config/i386/sse.md (*mov<mode>_internal)
85 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
86 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
87 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
88 (*<code><mode>3, *andnot<mode>3<mask_name>)
89 (<mask_codefor><code><mode>3<mask_name>): Only consider
90 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
92 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
95 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
97 * lra-constraints.c (valid_address_p): Move earlier in file.
98 Add a constraint argument to the address_info version.
99 (satisfies_memory_constraint_p): New function.
100 (satisfies_address_constraint_p): Likewise.
101 (process_alt_operands, curr_insn_transform): Use them.
102 (process_address): Pass the constraint to valid_address_p when
103 checking address operands.
105 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
107 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
108 to their respective blocks. Fix inadvertent use of "node".
110 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
112 * emit-rtl.c (init_derived_machine_modes): New functionm, split
114 (init_emit_once): ...here.
115 * rtl.h (init_derived_machine_modes): Declare.
116 * toplev.c (do_compile): Call it even if no_backend.
118 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
119 Mike Stump <mikestump@comcast.net>
120 Richard Sandiford <rdsandiford@googlemail.com>
121 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
123 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
124 (rtx_equal_for_memref_p): Update comment.
125 (adjust_offset_for_component_ref): Use wide-int interfaces.
126 * builtins.c (get_object_alignment_2): Likewise.
127 (c_readstr): Likewise.
128 (target_char_cast): Add comment.
129 (determine_block_size): Use wide-int interfaces.
130 (expand_builtin_signbit): Likewise.
131 (fold_builtin_int_roundingfn): Likewise.
132 (fold_builtin_bitop): Likewise.
133 (fold_builtin_bswap): Likewise.
134 (fold_builtin_logarithm): Use signop.
135 (fold_builtin_pow): Likewise.
136 (fold_builtin_memory_op): Use wide-int interfaces.
137 (fold_builtin_object_size): Likewise.
138 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
139 nb_iterations_estimate.
140 (record_niter_bound): Use wide-int interfaces.
141 (get_estimated_loop_iterations_int): Likewise.
142 (get_estimated_loop_iterations): Likewise.
143 (get_max_loop_iterations): Likewise.
144 * cfgloop.h: Include wide-int.h.
145 (struct nb_iter_bound): Change bound to widest_int.
146 (struct loop): Change nb_iterations_upper_bound and
147 nb_iterations_estimate to widest_int.
148 (record_niter_bound): Switch to use widest_int.
149 (get_estimated_loop_iterations): Likewise.
150 (get_max_loop_iterations): Likewise.
151 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
153 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
154 * combine.c (try_combine): Likewise.
155 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
156 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
158 (aarch64_float_const_representable_p): Likewise.
159 * config/arc/arc.c: Include wide-int.h.
160 (arc_can_use_doloop_p): Use wide-int interfaces.
161 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
162 (vfp3_const_double_index): Likewise.
163 * config/avr/avr.c (avr_out_round): Likewise.
164 (avr_fold_builtin): Likewise.
165 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
166 (bfin_can_use_doloop_p): Likewise.
167 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
168 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
169 * config/i386/i386.c: Include wide-int.h.
170 (ix86_data_alignment): Use wide-int interfaces.
171 (ix86_local_alignment): Likewise.
172 (ix86_emit_swsqrtsf): Update real_from_integer.
173 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
174 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
175 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
176 (zero_constant): Likewise.
177 (input_operand): Likewise.
178 (splat_input_operand): Likewise.
179 (non_logical_cint_operand): Change const_double to const_wide_int.
180 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
181 (easy_altivec_constant): Remove comment.
182 (paired_expand_vector_init): Use CONSTANT_P.
183 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
184 (rs6000_emit_move): Update checks.
185 (rs6000_aggregate_candidate): Use wide-int interfaces.
186 (rs6000_expand_ternop_builtin): Likewise.
187 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
188 (rs6000_assemble_integer): Likewise.
189 (rs6000_hash_constant): Likewise.
190 (output_toc): Likewise.
191 (rs6000_rtx_costs): Likewise.
192 (rs6000_emit_swrsqrt); Update call to real_from_integer.
193 * config/rs6000/rs6000-c.c: Include wide-int.h.
194 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
195 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
196 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
197 Handle CONST_WIDE_INT.
198 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
199 Use tree_fits_uhwi_p.
200 * config/sparc/sparc.c: Include wide-int.h.
201 (sparc_fold_builtin): Use wide-int interfaces.
202 * config/vax/vax.c: Include wide-int.h.
203 (vax_float_literal): Use real_from_integer.
204 * coretypes.h (struct hwivec_def): New.
207 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
208 (equiv_constant): Handle CONST_WIDE_INT.
209 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
210 (cselib_hash_rtx): Handle CONST_WIDE_INT.
211 * dbxout.c (stabstr_U): Use wide-int interfaces.
212 (dbxout_type): Update to use cst_fits_shwi_p.
213 * defaults.h (LOG2_BITS_PER_UNIT): Define.
214 (TARGET_SUPPORTS_WIDE_INT): Add default.
215 * dfp.c: Include wide-int.h.
216 (decimal_real_to_integer2): Use wide-int interfaces and rename to
217 decimal_real_to_integer.
218 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
219 decimal_real_to_integer.
220 * doc/generic.texi (Constant expressions): Update for wide_int.
221 * doc/rtl.texi (const_double): Likewise.
222 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
223 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
224 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
225 (REAL_VALUE_FROM_INT): Remove.
226 (TARGET_SUPPORTS_WIDE_INT): New.
227 * doc/tm.texi: Regenerate.
228 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
229 * double-int.h: Include wide-int.h.
230 (struct wi::int_traits): New.
231 * dwarf2out.c (get_full_len): New.
232 (dw_val_equal_p): Add case dw_val_class_wide_int.
233 (size_of_loc_descr): Likewise.
234 (output_loc_operands): Likewise.
235 (insert_double): Remove.
236 (insert_wide_int): New.
238 (print_die): Add case dw_val_class_wide_int.
239 (attr_checksum): Likewise.
240 (attr_checksum_ordered): Likewise.
241 (same_dw_val_p): Likewise.
242 (size_of_die): Likewise.
243 (value_format): Likewise.
244 (output_die): Likewise.
245 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
247 (clz_loc_descriptor): Use wide-int interfaces.
248 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
249 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
250 (round_up_to_align): Use wide-int interfaces.
251 (field_byte_offset): Likewise.
252 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
253 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
254 CONST_DOUBLE handling. Use wide-int interfaces.
255 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
256 (gen_enumeration_type_die): Use add_AT_wide.
257 (hash_loc_operands): Add case dw_val_class_wide_int.
258 (compare_loc_operands): Likewise.
259 * dwarf2out.h: Include wide-int.h.
261 (enum dw_val_class): Add dw_val_class_wide_int.
262 (struct dw_val_struct): Add val_wide.
263 * emit-rtl.c (const_wide_int_htab): New.
264 (const_wide_int_htab_hash): New.
265 (const_wide_int_htab_eq): New.
266 (lookup_const_wide_int): New.
267 (const_double_htab_hash): Use wide-int interfaces.
268 (const_double_htab_eq): Likewise.
269 (rtx_to_double_int): Conditionally compile for wide-int.
270 (immed_double_int_const): Rename to immed_wide_int_const and
272 (immed_double_const): Conditionally compile for wide-int.
273 (init_emit_once): Use wide-int interfaces.
274 * explow.c (plus_constant): Likewise.
275 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
276 (lshift_value): Use wide-int interfaces.
277 (expand_mult): Likewise.
278 (choose_multiplier): Likewise.
279 (expand_smod_pow2): Likewise.
280 (make_tree): Likewise.
281 * expr.c (convert_modes): Consolidate handling of constants.
282 Use wide-int interfaces.
283 (emit_group_load_1): Add note.
284 (store_expr): Update comment.
285 (get_inner_reference): Use wide-int interfaces.
286 (expand_constructor): Update comment.
287 (expand_expr_real_2): Use wide-int interfaces.
288 (expand_expr_real_1): Likewise.
289 (reduce_to_bit_field_precision): Likewise.
290 (const_vector_from_tree): Likewise.
291 * final.c: Include wide-int-print.h.
292 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
293 * fixed-value.c: Include wide-int.h.
294 (fixed_from_string): Use wide-int interfaces.
295 (fixed_to_decimal): Likewise.
296 (fixed_convert_from_real): Likewise.
297 (real_convert_from_fixed): Likewise.
298 * fold-const.h (mem_ref_offset): Return an offset_int.
299 (div_if_zero_remainder): Remove code parameter.
300 * fold-const.c (div_if_zero_remainder): Remove code parameter.
301 Use wide-int interfaces.
302 (may_negate_without_overflow_p): Use wide-int interfaces.
303 (negate_expr_p): Likewise.
304 (fold_negate_expr): Likewise.
305 (int_const_binop_1): Likewise.
306 (const_binop): Likewise.
307 (fold_convert_const_int_from_int): Likewise.
308 (fold_convert_const_int_from_real): Likewise.
309 (fold_convert_const_int_from_fixed): Likewise.
310 (fold_convert_const_fixed_from_int): Likewise.
311 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
312 (sign_bit_p): Use wide-int interfaces.
313 (make_range_step): Likewise.
314 (build_range_check): Likewise. Pass an integer of the correct type
315 instead of using integer_one_node.
316 (range_predecessor): Pass an integer of the correct type instead
317 of using integer_one_node.
318 (range_successor): Likewise.
319 (merge_ranges): Likewise.
320 (unextend): Use wide-int interfaces.
321 (extract_muldiv_1): Likewise.
322 (fold_div_compare): Likewise.
323 (fold_single_bit_test): Likewise.
324 (fold_sign_changed_comparison): Likewise.
325 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
326 (fold_plusminus_mult_expr): Use wide-int interfaces.
327 (native_encode_int): Likewise.
328 (native_interpret_int): Likewise.
329 (fold_unary_loc): Likewise.
330 (pointer_may_wrap_p): Likewise.
331 (size_low_cst): Likewise.
332 (mask_with_tz): Likewise.
333 (fold_binary_loc): Likewise.
334 (fold_ternary_loc): Likewise.
335 (multiple_of_p): Likewise.
336 (tree_call_nonnegative_warnv_p): Update calls to
337 tree_int_cst_min_precision and real_from_integer.
338 (fold_negate_const): Use wide-int interfaces.
339 (fold_abs_const): Likewise.
340 (fold_relational_const): Use tree_int_cst_lt.
341 (round_up_loc): Use wide-int interfaces.
342 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
343 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
344 * gengtype.c: Remove include of double-int.h.
345 (do_typedef): Use wide-int interfaces.
346 (open_base_files): Add wide-int.h.
347 (main): Add offset_int and widest_int typedefs.
348 * gengtype-lex.l: Handle "^".
349 (CXX_KEYWORD): Add "static".
350 * gengtype-parse.c (require3): New.
351 (require_template_declaration): Handle constant template arguments
352 and nested templates.
353 * gengtype-state.c: Don't include "double-int.h".
354 * genpreds.c (write_one_predicate_function): Update comment.
355 (write_tm_constrs_h): Add check for hval and lval use in
357 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
358 (add_to_sequence): Likewise.
359 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
360 and const_double_operand.
361 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
363 * gimple-fold.c (get_base_constructor): Likewise.
364 (fold_array_ctor_reference): Likewise.
365 (fold_nonarray_ctor_reference): Likewise.
366 (fold_const_aggregate_ref_1): Likewise.
367 (gimple_val_nonnegative_real_p): Likewise.
368 (gimple_fold_indirect_ref): Likewise.
369 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
370 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
371 (struct slsr_cand_d): Change index to be widest_int.
372 (struct incr_info_d): Change incr to be widest_int.
373 (alloc_cand_and_find_basis): Use wide-int interfaces.
374 (slsr_process_phi): Likewise.
375 (backtrace_base_for_ref): Likewise. Return a widest_int.
376 (restructure_reference): Take a widest_int instead of a double_int.
377 (slsr_process_ref): Use wide-int interfaces.
378 (create_mul_ssa_cand): Likewise.
379 (create_mul_imm_cand): Likewise.
380 (create_add_ssa_cand): Likewise.
381 (create_add_imm_cand): Take a widest_int instead of a double_int.
382 (slsr_process_add): Use wide-int interfaces.
383 (slsr_process_cast): Likewise.
384 (slsr_process_copy): Likewise.
385 (dump_candidate): Likewise.
386 (dump_incr_vec): Likewise.
387 (replace_ref): Likewise.
388 (cand_increment): Likewise. Return a widest_int.
389 (cand_abs_increment): Likewise.
390 (replace_mult_candidate): Take a widest_int instead of a double_int.
391 (replace_unconditional_candidate): Use wide-int interfaces.
392 (incr_vec_index): Take a widest_int instead of a double_int.
393 (create_add_on_incoming_edge): Likewise.
394 (create_phi_basis): Use wide-int interfaces.
395 (replace_conditional_candidate): Likewise.
396 (record_increment): Take a widest_int instead of a double_int.
397 (record_phi_increments): Use wide-int interfaces.
398 (phi_incr_cost): Take a widest_int instead of a double_int.
399 (lowest_cost_path): Likewise.
400 (total_savings): Likewise.
401 (analyze_increments): Use wide-int interfaces.
402 (ncd_with_phi): Take a widest_int instead of a double_int.
403 (ncd_of_cand_and_phis): Likewise.
404 (nearest_common_dominator_for_cands): Likewise.
405 (insert_initializers): Use wide-int interfaces.
406 (all_phi_incrs_profitable): Likewise.
407 (replace_one_candidate): Likewise.
408 (replace_profitable_candidates): Likewise.
409 * godump.c: Include wide-int-print.h.
410 (go_output_typedef): Use wide-int interfaces.
411 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
412 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
413 (build_loop_iteration_domains): Likewise.
414 * hooks.h: Include wide-int.h rather than double-int.h.
415 (hook_bool_dint_dint_uint_bool_true): Delete.
416 (hook_bool_wint_wint_uint_bool_true): Declare.
417 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
418 (hook_bool_wint_wint_uint_bool_true): New.
419 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
421 (ubsan_expand_si_overflow_mul_check): Likewise.
422 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
423 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
424 (get_ancestor_addr_info): Likewise.
425 (ipa_modify_call_arguments): Likewise.
426 * loop-doloop.c (doloop_modify): Likewise.
427 (doloop_optimize): Likewise.
428 * loop-iv.c (iv_number_of_iterations): Likewise.
429 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
430 (unroll_loop_constant_iterations): Likewise.
431 (decide_unroll_runtime_iterations): Likewise.
432 (unroll_loop_runtime_iterations): Likewise.
433 (decide_peel_simple): Likewise.
434 (decide_unroll_stupid): Likewise.
435 * lto-streamer-in.c (streamer_read_wi): Add.
436 (input_cfg): Use wide-int interfaces.
437 (lto_input_tree_1): Likewise.
438 * lto-streamer-out.c (streamer_write_wi): Add.
439 (hash_tree): Use wide-int interfaces.
440 (output_cfg): Likewise.
441 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
442 (GTFILES): Add wide-int.h and signop.h.
443 (TAGS): Look for .cc files too.
444 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
445 * optabs.c (expand_subword_shift): Likewise.
446 (expand_doubleword_shift): Likewise.
447 (expand_absneg_bit): Likewise.
448 (expand_copysign_absneg): Likewise.
449 (expand_copysign_bit): Likewise.
450 * postreload.c (reload_cse_simplify_set): Likewise.
451 * predict.c (predict_iv_comparison): Likewise.
452 * pretty-print.h: Include wide-int-print.h.
454 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
455 * print-tree.c: Include wide-int-print.h.
456 (print_node_brief): Use wide-int interfaces.
457 (print_node): Likewise.
458 * read-rtl.c (validate_const_wide_int): New.
459 (read_rtx_code): Add CONST_WIDE_INT case.
460 * real.c: Include wide-int.h.
461 (real_to_integer2): Delete.
462 (real_to_integer): New function, returning a wide_int.
463 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
464 (ten_to_ptwo): Update call to real_from_integer.
465 (real_digit): Likewise.
466 * real.h: Include signop.h, wide-int.h and insn-modes.h.
467 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
468 (REAL_VALUE_TO_INT): Delete.
469 (real_to_integer): Declare a wide-int form.
470 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
471 * recog.c (const_int_operand): Improve comment.
472 (const_scalar_int_operand): New.
473 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
474 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
475 (split_double): Likewise.
476 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
477 (rtx_size): Likewise.
478 (rtx_alloc_stat_v): New.
479 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
480 (cwi_output_hex): New.
481 (iterative_hash_rtx): Handle CONST_WIDE_INT.
482 (cwi_check_failed_bounds): New.
483 * rtl.def (CONST_WIDE_INT): New.
484 * rtl.h: Include <utility> and wide-int.h.
485 (struct hwivec_def): New.
486 (CWI_GET_NUM_ELEM): New.
487 (CWI_PUT_NUM_ELEM): New.
488 (struct rtx_def): Add num_elem and hwiv.
489 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
490 (CASE_CONST_UNIQUE): Likewise.
491 (CASE_CONST_ANY): Likewise.
492 (CONST_SCALAR_INT_P): Likewise.
493 (CONST_WIDE_INT_P): New.
496 (cwi_check_failed_bounds): New.
499 (CONST_WIDE_INT_VEC) New.
500 (CONST_WIDE_INT_NUNITS) New.
501 (CONST_WIDE_INT_ELT) New.
502 (rtx_mode_t): New type.
503 (wi::int_traits <rtx_mode_t>): New.
505 (wi::min_value): New.
506 (wi::max_value): New.
508 (const_wide_int_alloc): New.
509 (immed_wide_int_const): New.
510 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
511 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
512 * signop.h: New file.
513 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
514 (simplify_const_unary_operation): Use wide-int interfaces.
515 (simplify_binary_operation_1): Likewise.
516 (simplify_const_binary_operation): Likewise.
517 (simplify_const_relational_operation): Likewise.
518 (simplify_immed_subreg): Likewise.
519 * stmt.c (expand_case): Likewise.
520 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
521 signop rather than a bool.
522 * stor-layout.c (layout_type): Use wide-int interfaces.
523 (initialize_sizetypes): Update calls to
524 set_min_and_max_values_for_integral_type.
525 (set_min_and_max_values_for_integral_type): Take a signop rather
526 than a bool. Use wide-int interfaces.
527 (fixup_signed_type): Update accordingly. Remove
528 HOST_BITS_PER_DOUBLE_INT limit.
529 (fixup_unsigned_type): Likewise.
530 * system.h (STATIC_CONSTANT_P): New.
531 (STATIC_ASSERT): New.
532 * target.def (can_use_doloop_p): Take widest_ints rather than
534 * target.h: Include wide-int.h rather than double-int.h.
535 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
537 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
538 rather than INT_CST_LT_UNSIGNED.
539 (can_use_doloop_if_innermost): Take widest_ints rather than
541 * tree-affine.c: Include wide-int-print.h.
542 (double_int_ext_for_comb): Delete.
543 (wide_int_ext_for_comb): New.
544 (aff_combination_zero): Use wide-int interfaces.
545 (aff_combination_const): Take a widest_int instead of a double_int.
546 (aff_combination_elt): Use wide-int interfaces.
547 (aff_combination_scale): Take a widest_int instead of a double_int.
548 (aff_combination_add_elt): Likewise.
549 (aff_combination_add_cst): Likewise.
550 (aff_combination_add): Use wide-int interfaces.
551 (aff_combination_convert): Likewise.
552 (tree_to_aff_combination): Likewise.
553 (add_elt_to_tree): Take a widest_int instead of a double_int.
554 (aff_combination_to_tree): Use wide-int interfaces.
555 (aff_combination_remove_elt): Likewise.
556 (aff_combination_add_product): Take a widest_int instead of
558 (aff_combination_mult): Use wide-int interfaces.
559 (aff_combination_expand): Likewise.
560 (double_int_constant_multiple_p): Delete.
561 (wide_int_constant_multiple_p): New.
562 (aff_combination_constant_multiple_p): Take a widest_int pointer
563 instead of a double_int pointer.
564 (print_aff): Use wide-int interfaces.
565 (get_inner_reference_aff): Take a widest_int pointer
566 instead of a double_int pointer.
567 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
568 * tree-affine.h: Include wide-int.h.
569 (struct aff_comb_elt): Change type of coef to widest_int.
570 (struct affine_tree_combination): Change type of offset to widest_int.
571 (double_int_ext_for_comb): Delete.
572 (wide_int_ext_for_comb): New.
573 (aff_combination_const): Use widest_int instead of double_int.
574 (aff_combination_scale): Likewise.
575 (aff_combination_add_elt): Likewise.
576 (aff_combination_constant_multiple_p): Likewise.
577 (get_inner_reference_aff): Likewise.
578 (aff_comb_cannot_overlap_p): Likewise.
579 (aff_combination_zero_p): Use wide-int interfaces.
580 * tree.c: Include tree.h.
581 (init_ttree): Use make_int_cst.
582 (tree_code_size): Removed code for INTEGER_CST case.
583 (tree_size): Add INTEGER_CST case.
584 (make_node_stat): Update comment.
585 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
586 (build_int_cst_type): Use wide-int interfaces.
587 (double_int_to_tree): Likewise.
588 (double_int_fits_to_tree_p): Delete.
589 (force_fit_type_double): Delete.
590 (force_fit_type): New.
591 (int_cst_hash_hash): Use wide-int interfaces.
592 (int_cst_hash_eq): Likewise.
593 (build_int_cst_wide): Delete.
594 (wide_int_to_tree): New.
595 (cache_integer_cst): Use wide-int interfaces.
596 (build_low_bits_mask): Likewise.
597 (cst_and_fits_in_hwi): Likewise.
598 (real_value_from_int_cst): Likewise.
599 (make_int_cst_stat): New.
600 (integer_zerop): Use wide_int interfaces.
601 (integer_onep): Likewise.
602 (integer_all_onesp): Likewise.
603 (integer_pow2p): Likewise.
604 (integer_nonzerop): Likewise.
605 (tree_log2): Likewise.
606 (tree_floor_log2): Likewise.
607 (tree_ctz): Likewise.
608 (int_size_in_bytes): Likewise.
609 (mem_ref_offset): Return an offset_int rather than a double_int.
610 (build_type_attribute_qual_variant): Use wide_int interfaces.
611 (type_hash_eq): Likewise
612 (tree_int_cst_equal): Likewise.
613 (tree_int_cst_lt): Delete.
614 (tree_int_cst_compare): Likewise.
615 (tree_fits_shwi_p): Use wide_int interfaces.
616 (tree_fits_uhwi_p): Likewise.
617 (tree_int_cst_sign_bit): Likewise.
618 (tree_int_cst_sgn): Likewise.
619 (tree_int_cst_min_precision): Take a signop rather than a bool.
620 (simple_cst_equal): Use wide_int interfaces.
621 (compare_tree_int): Likewise.
622 (iterative_hash_expr): Likewise.
623 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
625 (get_type_static_bounds): Use wide_int interfaces.
626 (tree_int_cst_elt_check_failed): New.
627 (build_common_tree_nodes): Reordered to set prec before filling in
629 (int_cst_value): Check cst_and_fits_in_hwi.
630 (widest_int_cst_value): Use wide_int interfaces.
631 (upper_bound_in_type): Likewise.
632 (lower_bound_in_type): Likewise.
633 (num_ending_zeros): Likewise.
634 (drop_tree_overflow): Likewise.
635 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
636 (gen_conditions_for_pow_cst_base): Likewise.
637 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
638 (group_case_labels_stmt): Use wide-int interfaces.
639 (verify_gimple_assign_binary): Likewise.
640 (print_loop): Likewise.
641 * tree-chrec.c (tree_fold_binomial): Likewise.
642 * tree-core.h (struct tree_base): Add int_length.
643 (struct tree_int_cst): Change rep of value.
644 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
645 (dr_may_alias_p): Likewise.
646 (max_stmt_executions_tree): Likewise.
647 * tree.def (INTEGER_CST): Update comment.
648 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
649 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
650 * tree-dump.c: Include wide-int.h and wide-int-print.h.
651 (dequeue_and_dump): Use wide-int interfaces.
652 * tree.h: Include wide-int.h.
653 (NULL_TREE): Moved to earlier loc in file.
654 (TREE_INT_CST_ELT_CHECK): New.
655 (tree_int_cst_elt_check_failed): New.
657 (TREE_INT_CST): Delete.
658 (TREE_INT_CST_LOW): Use wide-int interfaces.
659 (TREE_INT_CST_HIGH): Delete.
660 (TREE_INT_CST_NUNITS): New.
661 (TREE_INT_CST_EXT_NUNITS): Likewise.
662 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
663 (TREE_INT_CST_ELT): Likewise.
664 (INT_CST_LT): Delete.
665 (tree_int_cst_elt_check): New (two forms).
666 (type_code_size): Update comment.
667 (make_int_cst_stat, make_int_cst): New.
668 (tree_to_double_int): Delete.
669 (double_int_fits_to_tree_p): Delete.
670 (force_fit_type_double): Delete.
671 (build_int_cstu): Replace with out-of-line function.
672 (build_int_cst_wide): Delete.
673 (tree_int_cst_lt): Define inline.
674 (tree_int_cst_le): New.
675 (tree_int_cst_compare): Define inline.
676 (tree_int_cst_min_precision): Take a signop rather than a bool.
677 (wi::int_traits <const_tree>): New.
678 (wi::int_traits <tree>): New.
679 (wi::extended_tree): New.
680 (wi::int_traits <wi::extended_tree>): New.
681 (wi::to_widest): New.
682 (wi::to_offset): New.
683 (wi::fits_to_tree_p): New.
684 (wi::min_value): New.
685 (wi::max_value): New.
686 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
687 (copy_tree_body_r): Likewise.
688 * tree-object-size.c (compute_object_offset): Likewise.
689 (addr_object_size): Likewise.
690 * tree-predcom.c: Include wide-int-print.h.
691 (struct dref_d): Change type of offset to widest_int.
692 (dump_dref): Call wide-int printer.
693 (aff_combination_dr_offset): Use wide-int interfaces.
694 (determine_offset): Take a widest_int pointer rather than a
696 (split_data_refs_to_components): Use wide-int interfaces.
697 (suitable_component_p): Likewise.
698 (order_drefs): Likewise.
699 (add_ref_to_chain): Likewise.
700 (valid_initializer_p): Likewise.
701 (determine_roots_comp): Likewise.
702 * tree-pretty-print.c: Include wide-int-print.h.
703 (dump_generic_node): Use wide-int interfaces.
704 * tree-sra.c (sra_ipa_modify_expr): Likewise.
705 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
706 (move_fixed_address_to_symbol): Likewise.
707 (move_hint_to_base): Likewise.
708 (move_pointer_to_base): Likewise.
709 (move_variant_to_index): Likewise.
710 (most_expensive_mult_to_index): Likewise.
711 (addr_to_parts): Likewise.
712 (copy_ref_info): Likewise.
713 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
714 (indirect_refs_may_alias_p): Likewise.
715 (stmt_kills_ref_p_1): Likewise.
716 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
717 * tree-ssa-ccp.c: Update comment at top of file. Include
719 (struct prop_value_d): Change type of mask to widest_int.
720 (extend_mask): New function.
721 (dump_lattice_value): Use wide-int interfaces.
722 (get_default_value): Likewise.
723 (set_constant_value): Likewise.
724 (set_value_varying): Likewise.
725 (valid_lattice_transition): Likewise.
726 (set_lattice_value): Likewise.
727 (value_to_double_int): Delete.
728 (value_to_wide_int): New.
729 (get_value_from_alignment): Use wide-int interfaces.
730 (get_value_for_expr): Likewise.
731 (do_dbg_cnt): Likewise.
732 (ccp_finalize): Likewise.
733 (ccp_lattice_meet): Likewise.
734 (bit_value_unop_1): Use widest_ints rather than double_ints.
735 (bit_value_binop_1): Likewise.
736 (bit_value_unop): Use wide-int interfaces.
737 (bit_value_binop): Likewise.
738 (bit_value_assume_aligned): Likewise.
739 (evaluate_stmt): Likewise.
740 (ccp_fold_stmt): Likewise.
741 (visit_cond_stmt): Likewise.
742 (ccp_visit_stmt): Likewise.
743 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
744 (constant_pointer_difference): Likewise.
745 (associate_pointerplus): Likewise.
746 (combine_conversions): Likewise.
747 * tree-ssa-loop.h: Include wide-int.h.
748 (struct tree_niter_desc): Change type of max to widest_int.
749 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
750 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
751 (remove_redundant_iv_tests): Likewise.
752 (canonicalize_loop_induction_variables): Likewise.
753 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
754 (constant_multiple_of): Take a widest_int pointer instead of
755 a double_int pointer.
756 (get_computation_aff): Use wide-int interfaces.
757 (ptr_difference_cost): Likewise.
758 (difference_cost): Likewise.
759 (get_loop_invariant_expr_id): Likewise.
760 (get_computation_cost_at): Likewise.
761 (iv_elimination_compare_lt): Likewise.
762 (may_eliminate_iv): Likewise.
763 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
764 instead of double_int.
765 (max_loop_iterations): Likewise.
766 (max_stmt_executions): Likewise.
767 (estimated_stmt_executions): Likewise.
768 * tree-ssa-loop-niter.c: Include wide-int-print.h.
769 (split_to_var_and_offset): Use wide-int interfaces.
770 (determine_value_range): Likewise.
771 (bound_difference_of_offsetted_base): Likewise.
772 (bounds_add): Take a widest_int instead of a double_int.
773 (number_of_iterations_ne_max): Use wide-int interfaces.
774 (number_of_iterations_ne): Likewise.
775 (number_of_iterations_lt_to_ne): Likewise.
776 (assert_loop_rolls_lt): Likewise.
777 (number_of_iterations_lt): Likewise.
778 (number_of_iterations_le): Likewise.
779 (number_of_iterations_cond): Likewise.
780 (number_of_iterations_exit): Likewise.
781 (finite_loop_p): Likewise.
782 (derive_constant_upper_bound_assign): Likewise.
783 (derive_constant_upper_bound): Return a widest_int.
784 (derive_constant_upper_bound_ops): Likewise.
785 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
786 (record_estimate): Take a widest_int rather than a double_int.
787 (record_nonwrapping_iv): Use wide-int interfaces.
788 (double_int_cmp): Delete.
790 (bound_index): Take a widest_int rather than a double_int.
791 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
792 (maybe_lower_iteration_bound): Likewise.
793 (estimate_numbers_of_iterations_loop): Likewise.
794 (estimated_loop_iterations): Take a widest_int pointer than than
795 a double_int pointer.
796 (estimated_loop_iterations_int): Use wide-int interfaces.
797 (max_loop_iterations): Take a widest_int pointer than than
798 a double_int pointer.
799 (max_loop_iterations_int): Use wide-int interfaces.
800 (max_stmt_executions): Take a widest_int pointer than than
801 a double_int pointer.
802 (estimated_stmt_executions): Likewise.
803 (n_of_executions_at_most): Use wide-int interfaces.
804 (scev_probably_wraps_p): Likewise.
805 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
807 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
809 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
810 double_ints. Adjust for trailing_wide_ints <3> representation.
811 (set_nonzero_bits): Likewise.
812 (get_range_info): Return wide_ints rather than double_ints.
813 Adjust for trailing_wide_ints <3> representation.
814 (get_nonzero_bits): Likewise.
815 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
817 * tree-ssanames.h (struct range_info_def): Replace min, max and
818 nonzero_bits with a trailing_wide_ints <3>.
819 (set_range_info): Use wide_int_refs rather than double_ints.
820 (set_nonzero_bits): Likewise.
821 (get_range_info): Return wide_ints rather than double_ints.
822 (get_nonzero_bits): Likewise.
823 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
824 * tree-ssa-pre.c (phi_translate_1): Likewise.
825 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
826 (acceptable_pow_call): Likewise.
827 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
829 (vn_reference_fold_indirect): Likewise.
830 (vn_reference_maybe_forwprop_address): Likewise.
831 (valueize_refs_1): Likewise.
832 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
833 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
834 tree_int_cst_lt and tree_int_cst_le.
835 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
837 (streamer_alloc_tree): Likewise.
838 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
839 (streamer_write_tree_header): Likewise.
840 (streamer_write_integer_cst): Likewise.
841 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
842 (build_constructors): Likewise.
843 (array_value_type): Likewise.
844 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
845 (vect_check_gather): Likewise.
846 * tree-vect-generic.c (build_replicated_const): Likewise.
847 (expand_vector_divmod): Likewise.
848 * tree-vect-loop.c (vect_transform_loop): Likewise.
849 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
850 (vect_do_peeling_for_alignment): Likewise.
851 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
852 * tree-vrp.c: Include wide-int.h.
853 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
854 (extract_range_from_assert): Use wide-int interfaces.
855 (vrp_int_const_binop): Likewise.
856 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
858 (ranges_from_anti_range): Use wide-int interfaces.
859 (quad_int_cmp): Delete.
860 (quad_int_pair_sort): Likewise.
861 (extract_range_from_binary_expr_1): Use wide-int interfaces.
862 (extract_range_from_unary_expr_1): Likewise.
863 (adjust_range_with_scev): Likewise.
864 (masked_increment): Take and return wide_ints rather than double_ints.
865 (register_edge_assert_for_2): Use wide-int interfaces.
866 (check_array_ref): Likewise.
867 (search_for_addr_array): Likewise.
868 (maybe_set_nonzero_bits): Likewise.
869 (union_ranges): Pass an integer of the correct type instead of
870 using integer_one_node.
871 (intersect_ranges): Likewise.
872 (simplify_truth_ops_using_ranges): Likewise.
873 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
874 (range_fits_type_p): Likewise.
875 (simplify_cond_using_ranges): Likewise. Take a signop rather than
877 (simplify_conversion_using_ranges): Use wide-int interfaces.
878 (simplify_float_conversion_using_ranges): Likewise.
879 (vrp_finalize): Likewise.
880 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
881 (gimple_stringops_transform): Likewise.
882 * varasm.c (decode_addr_const): Likewise.
883 (const_hash_1): Likewise.
884 (const_rtx_hash_1): Likewise
885 (output_constant): Likewise.
886 (array_size_for_constructor): Likewise.
887 (output_constructor_regular_field): Likewise.
888 (output_constructor_bitfield): Likewise.
889 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
890 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
892 * gencheck.c: Define BITS_PER_UNIT.
895 * wide-int-print.cc: New.
896 * wide-int-print.h: New.
898 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
900 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
902 2014-05-06 Richard Biener <rguenther@suse.de>
904 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
905 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
906 (TODO_verify_all): Adjust.
907 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
908 TODO_verify_stmts and TODO_verify_rtl_sharing.
909 * bb-reorder.c: Likewise.
910 * cfgexpand.c: Likewise.
913 * function.c: Likewise.
914 * fwprop.c: Likewise.
916 * gimple-ssa-isolate-paths.c: Likewise.
917 * gimple-ssa-strength-reduction.c: Likewise.
918 * ipa-split.c: Likewise.
919 * loop-init.c: Likewise.
920 * loop-unroll.c: Likewise.
921 * lower-subreg.c: Likewise.
922 * modulo-sched.c: Likewise.
923 * postreload-gcse.c: Likewise.
924 * predict.c: Likewise.
926 * sched-rgn.c: Likewise.
927 * store-motion.c: Likewise.
928 * tracer.c: Likewise.
929 * trans-mem.c: Likewise.
930 * tree-call-cdce.c: Likewise.
931 * tree-cfg.c: Likewise.
932 * tree-cfgcleanup.c: Likewise.
933 * tree-complex.c: Likewise.
934 * tree-eh.c: Likewise.
935 * tree-emutls.c: Likewise.
936 * tree-if-conv.c: Likewise.
937 * tree-into-ssa.c: Likewise.
938 * tree-loop-distribution.c: Likewise.
939 * tree-object-size.c: Likewise.
940 * tree-parloops.c: Likewise.
941 * tree-pass.h: Likewise.
942 * tree-sra.c: Likewise.
943 * tree-ssa-ccp.c: Likewise.
944 * tree-ssa-copy.c: Likewise.
945 * tree-ssa-copyrename.c: Likewise.
946 * tree-ssa-dce.c: Likewise.
947 * tree-ssa-dom.c: Likewise.
948 * tree-ssa-dse.c: Likewise.
949 * tree-ssa-forwprop.c: Likewise.
950 * tree-ssa-ifcombine.c: Likewise.
951 * tree-ssa-loop-ch.c: Likewise.
952 * tree-ssa-loop-ivcanon.c: Likewise.
953 * tree-ssa-loop.c: Likewise.
954 * tree-ssa-math-opts.c: Likewise.
955 * tree-ssa-phiopt.c: Likewise.
956 * tree-ssa-phiprop.c: Likewise.
957 * tree-ssa-pre.c: Likewise.
958 * tree-ssa-reassoc.c: Likewise.
959 * tree-ssa-sink.c: Likewise.
960 * tree-ssa-strlen.c: Likewise.
961 * tree-ssa-tail-merge.c: Likewise.
962 * tree-ssa-uncprop.c: Likewise.
963 * tree-switch-conversion.c: Likewise.
964 * tree-tailcall.c: Likewise.
965 * tree-vect-generic.c: Likewise.
966 * tree-vectorizer.c: Likewise.
967 * tree-vrp.c: Likewise.
969 * var-tracking.c: Likewise.
970 * bt-load.c: Likewise.
971 * cfgcleanup.c: Likewise.
972 * combine-stack-adj.c: Likewise.
973 * combine.c: Likewise.
974 * compare-elim.c: Likewise.
975 * config/epiphany/resolve-sw-modes.c: Likewise.
976 * config/i386/i386.c: Likewise.
977 * config/mips/mips.c: Likewise.
978 * config/s390/s390.c: Likewise.
979 * config/sh/sh_treg_combine.cc: Likewise.
980 * config/sparc/sparc.c: Likewise.
985 * mode-switching.c: Likewise.
986 * passes.c: Likewise.
987 * postreload.c: Likewise.
989 * reg-stack.c: Likewise.
990 * regcprop.c: Likewise.
991 * regrename.c: Likewise.
994 2014-05-06 Richard Biener <rguenther@suse.de>
997 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
998 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
1000 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
1003 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
1005 2014-05-05 Radovan Obradovic <robradovic@mips.com>
1006 Tom de Vries <tom@codesourcery.com>
1008 * target.def (call_fusage_contains_non_callee_clobbers): New DEFHOOKPOD.
1009 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
1011 (@node Miscellaneous Register Hooks): New node.
1012 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
1013 * doc/tm.texi: Regenerate.
1015 2014-05-05 Marek Polacek <polacek@redhat.com>
1018 * opts.c (common_handle_option): Call error_at instead of warning_at.
1020 2014-05-05 Richard Biener <rguenther@suse.de>
1022 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
1023 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
1024 under the TODO_verify_il umbrella.
1026 2014-05-05 Richard Biener <rguenther@suse.de>
1028 * passes.c (execute_function_todo): Move TODO_verify_flow under
1029 the TODO_verify_ul umbrella.
1031 2014-05-05 Richard Biener <rguenther@suse.de>
1034 * fold-const.c (fold_binary_loc): Consistently avoid
1035 canonicalizing X & CST away from a CST that is the mask
1038 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1040 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
1041 int argument to enum machine_mode.
1042 (picochip_class_max_nregs): Ditto.
1043 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
1044 (picochip_class_max_nregs): Ditto.
1046 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1048 * target.def: Add new target hook.
1049 * doc/tm.texi: Regenerate.
1050 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
1051 * targhooks.c (default_keep_leaf_when_profiled): New function.
1053 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
1054 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
1056 2014-05-05 Bin Cheng <bin.cheng@arm.com>
1058 PR tree-optimization/60363
1059 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
1060 (copy_phi_args): New parameters. Call get_value_locus_in_path.
1061 (update_destination_phis): New parameter.
1062 (create_edge_and_update_destination_phis): Ditto.
1063 (ssa_fix_duplicate_block_edges): Pass new arguments.
1064 (thread_single_edge): Ditto.
1066 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
1068 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
1069 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
1070 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
1071 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
1072 Use RS6000_BTM_HARD_FLOAT.
1073 (BU_MISC_2): Likewise.
1074 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
1075 RS6000_BTM_HARD_FLOAT.
1076 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
1078 (rs6000_invalid_builtin): Add hard floating builtin support.
1079 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
1080 hard float builtins.
1081 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
1083 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1085 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
1086 Add missing function* argument.
1088 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1090 * lra-constraints.c (valid_address_p): Move earlier in file.
1091 Add a constraint argument to the address_info version.
1092 (satisfies_memory_constraint_p): New function.
1093 (satisfies_address_constraint_p): Likewise.
1094 (process_alt_operands, curr_insn_transform): Use them.
1095 (process_address): Pass the constraint to valid_address_p when
1096 checking address operands.
1098 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1100 * config/mips/mips.c (mips_isa_rev): New variable.
1101 (mips_set_architecture): Set it.
1102 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1104 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1105 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1106 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1107 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1108 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
1109 conditions in terms of mips_isa_rev.
1110 (mips_isa_rev): Declare.
1112 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1114 * config/sh/sh-mem.cc: Use tabs instead of spaces.
1115 (prob_unlikely, prob_likely): Make variables const.
1117 2014-05-03 Denis Chertykov <chertykov@gmail.com>
1119 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
1121 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1123 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
1125 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1127 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
1128 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
1129 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
1131 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
1132 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
1134 Replace usage of ROUND_REG with sh_round_reg.
1135 Use CEIL instead of ROUND_ADVANCE.
1137 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1140 * config/sh/sh.c: Include stdlib headers before everything else.
1142 2014-05-02 Jakub Jelinek <jakub@redhat.com>
1144 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
1145 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
1146 (gimplify_adjust_omp_clauses): Simd region is never
1147 directly nested in combined parallel. Instead, for linear
1148 with copyin/copyout, if in combined for simd loop, make decl
1149 firstprivate/lastprivate on OMP_FOR.
1150 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
1151 expand_omp_for_static_chunk): When setting endvar, also set
1152 fd->loop.v to the same value.
1154 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1156 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
1158 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
1160 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
1163 2014-05-02 Marek Polacek <polacek@redhat.com>
1165 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
1167 2014-05-02 Kito Cheng <kito@0xlab.org>
1169 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
1170 to a C expression marco.
1171 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
1172 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
1173 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
1174 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
1175 HONOR_REG_ALLOC_ORDER.
1176 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
1178 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1180 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
1182 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1184 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
1186 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
1188 * tree-if-conv.c (is_cond_scalar_reduction): New function.
1189 (convert_scalar_cond_reduction): Likewise.
1190 (predicate_scalar_phi): Add recognition and transformation
1191 of simple conditioanl reduction to be vectorizable.
1193 2014-05-01 Marek Polacek <polacek@redhat.com>
1196 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
1198 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
1200 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
1201 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
1202 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
1203 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
1204 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
1205 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
1206 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
1207 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
1209 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
1211 * config/arc/arc.opt (mlra): Move comment above option name
1212 to avoid mis-parsing as language options.
1214 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1216 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
1217 * config/sol2.h: ... here.
1218 * config/sol2-10.h: Remove.
1220 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
1221 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
1222 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
1223 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
1224 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
1225 * config/sol2.h: ... here.
1226 (SECTION_NAME_FORMAT): Don't redefine.
1227 (STARTFILE_ARCH32_SPEC): Rename to ...
1228 (STARTFILE_ARCH_SPEC): ... this.
1229 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
1230 * config/sparc/sol2.h: ... here.
1231 (SECTION_NAME_FORMAT): Don't undef.
1232 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
1233 (SUBTARGET_EXTRA_SPECS): Remove.
1234 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
1236 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
1237 (MD_STARTFILE_PREFIX): Remove.
1238 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
1239 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
1240 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
1241 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
1242 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
1243 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
1244 * config/i386/sol2.h: ... here.
1245 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
1246 * config/i386/sol2-bi.h: Remove.
1247 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
1248 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
1250 * config/i386/t-sol2-64: Rename to ...
1251 * config/i386/t-sol2: ... this.
1252 * config/sparc/t-sol2-64: Rename to ...
1253 * config/sparc/t-sol2: ... this.
1255 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
1256 sol2_tm_file_head, sol2_tm_file_tail.
1257 Include ${cpu_type}/sol2.h before sol2.h.
1259 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
1260 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
1261 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
1262 Reflect i386/t-sol2-64 renaming.
1263 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
1264 Reflect sparc/t-sol2-64 renaming.
1266 2014-04-30 Richard Biener <rguenther@suse.de>
1268 * passes.c (execute_function_todo): Move TODO_verify_stmts
1269 and TODO_verify_ssa under the TODO_verify_il umbrella.
1270 * tree-ssa.h (verify_ssa): Adjust prototype.
1271 * tree-ssa.c (verify_ssa): Add parameter to tell whether
1272 we should verify SSA operands.
1273 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
1274 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
1275 whether we should verify whether not throwing stmts have EH info.
1276 * graphite-scop-detection.c (create_sese_edges): Adjust.
1277 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
1278 * tree-eh.c (lower_try_finally_switch): Do not add the
1279 default case label twice.
1281 2014-04-30 Marek Polacek <polacek@redhat.com>
1283 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
1284 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
1285 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
1286 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
1288 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
1290 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
1291 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
1292 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
1293 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
1294 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
1295 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
1296 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
1297 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
1299 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1301 * tree-cfg.c (dump_function_to_file): Dump the return type of
1302 functions, in a line to itself before the function body, mimicking
1303 the layout of a C function.
1305 2014-04-29 Jakub Jelinek <jakub@redhat.com>
1307 PR tree-optimization/60971
1308 * tree-tailcall.c (process_assignment): Reject conversions which
1311 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
1313 * calls.c (initialize_argument_information): Always treat
1314 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
1315 (expand_call): Likewise.
1316 (emit_library_call_calue_1): Likewise.
1317 * expr.c (PUSH_ARGS_REVERSED): Do not define.
1318 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
1321 2014-04-29 Nick Clifton <nickc@redhat.com>
1323 * config/msp430/msp430.md (umulsidi): Fix typo.
1324 (mulhisi3): Enable even inside interrupt handlers.
1325 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
1326 bigger return address pushed in large mode.
1328 2014-04-29 Nick Clifton <nickc@redhat.com>
1330 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
1331 (arc_init_reg_tables): Use a machine_mode enum to iterate over
1333 * config/m32r/m32r.c (init_reg_tables): Likewise.
1334 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
1335 enum to hold the modes.
1337 2014-04-29 Richard Biener <rguenther@suse.de>
1339 * dominance.c (free_dominance_info): Add overload with
1341 (dom_info_state): Likewise.
1342 (dom_info_available_p): Likewise.
1343 * basic-block.h (free_dominance_info, dom_info_state,
1344 dom_info_available_p): Declare overloads.
1345 * passes.c (execute_function_todo): Verify that verifiers
1346 don't change dominator info state. Drop dominator info
1347 for IPA pass invocations.
1348 * cgraph.c (release_function_body): Restore asserts that
1349 dominator information is released.
1351 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
1353 * doc/invoke.texi: Fix typo.
1354 * tree-vrp.c: Fix typos.
1355 * gimple.c (infer_nonnull_range): Reorder operands of an &&
1358 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1360 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
1362 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1364 * config/aarch64/aarch64-builtins.c
1365 (aarch64_types_storestruct_lane_qualifiers): New.
1366 (TYPES_STORESTRUCT_LANE): Likewise.
1367 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
1368 (st3_lane): Likewise.
1369 (st4_lane): Likewise.
1370 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
1371 (vec_store_lanesci_lane<mode>): Likewise.
1372 (vec_store_lanesxi_lane<mode>): Likewise.
1373 (aarch64_st2_lane<VQ:mode>): Likewise.
1374 (aarch64_st3_lane<VQ:mode>): Likewise.
1375 (aarch64_st4_lane<VQ:mode>): Likewise.
1376 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
1377 * config/aarch64/arm_neon.h
1378 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
1379 use new macro arguments.
1380 (__ST3_LANE_FUNC): Likewise.
1381 (__ST4_LANE_FUNC): Likewise.
1382 * config/aarch64/iterators.md (V_TWO_ELEM): New.
1383 (V_THREE_ELEM): Likewise.
1384 (V_FOUR_ELEM): Likewise.
1386 2014-04-28 David Malcolm <dmalcolm@redhat.com>
1388 * doc/gimple.texi: Replace the description of the now-defunct
1389 union gimple_statement_d with a diagram showing the
1390 gimple_statement_base class hierarchy and its relationships to
1391 the GSS_ and GIMPLE_ enums.
1393 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1395 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
1396 * config/aarch64/aarch64.c
1397 (aarch64_cannot_change_mode_class): Weaken conditions.
1398 (aarch64_modes_tieable_p): New.
1399 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
1401 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
1403 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
1404 (loadsync_<mode>): Change mode.
1405 (load_quadpti, store_quadpti): New.
1406 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
1407 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
1409 2014-04-28 Martin Jambor <mjambor@suse.cz>
1411 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
1412 same alias type as the original statement.
1413 (subreplacement_assignment_data): New type.
1414 (handle_unscalarized_data_in_subtree): New type of parameter,
1415 generate new memory accesses with same alias type as the original
1417 (load_assign_lhs_subreplacements): Likewise.
1418 (sra_modify_constructor_assign): Generate new memory accesses with
1419 same alias type as the original statement.
1421 2014-04-28 Richard Biener <rguenther@suse.de>
1423 * tree-pass.h (TODO_verify_il): Define.
1424 (TODO_verify_all): Complete properly.
1425 * passes.c (execute_function_todo): Move existing loop-closed
1426 SSA verification under TODO_verify_il.
1427 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
1428 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
1429 Fix tree sharing issue.
1431 2014-04-28 Richard Biener <rguenther@suse.de>
1434 * builtins.def (DEF_C11_BUILTIN): Add.
1435 (BUILT_IN_ALIGNED_ALLOC): Likewise.
1436 * coretypes.h (enum function_class): Add function_c11_misc.
1437 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1438 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
1439 (call_may_clobber_ref_p_1): Likewise.
1440 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
1441 (mark_all_reaching_defs_necessary_1): Likewise.
1442 (propagate_necessity): Likewise.
1443 (eliminate_unnecessary_stmts): Likewise.
1444 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
1446 2014-04-28 Richard Biener <rguenther@suse.de>
1448 * tree-vrp.c (vrp_var_may_overflow): Remove.
1449 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
1450 with overflow immediately bump to one before that value and
1451 let iteration figure out overflow status.
1453 2014-04-28 Richard Biener <rguenther@suse.de>
1455 * configure.ac: Do valgrind header checks unconditionally.
1456 Add --enable-valgrind-annotations.
1457 * system.h: Guard valgrind header inclusion with
1458 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
1459 * alloc-pool.c (pool_alloc, pool_free): Use
1460 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
1461 to guard possibly dead code.
1462 * config.in: Regenerated.
1463 * configure: Likewise.
1465 2014-04-28 Jeff Law <law@redhat.com>
1467 PR tree-optimization/60902
1468 * tree-ssa-threadedge.c
1469 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
1470 over real defs when invalidating outputs from statements that do not
1471 produce useful outputs for threading.
1473 2014-04-28 Richard Biener <rguenther@suse.de>
1475 PR tree-optimization/60979
1476 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
1477 SCOPs that end in a block with a successor with abnormal
1480 2014-04-28 Richard Biener <rguenther@suse.de>
1482 * tree-pass.h (execute_pass_list): Adjust prototype.
1483 * passes.c (pass_manager::execute_early_local_passes):
1485 (do_per_function): Change callback signature, push all actual
1486 work to the callbals.
1487 (do_per_function_toporder): Likewise.
1488 (execute_function_dump): Adjust.
1489 (execute_function_todo): Likewise.
1490 (clear_last_verified): Likewise.
1491 (verify_curr_properties): Likewise.
1492 (update_properties_after_pass): Likewise.
1493 (execute_pass_list_1): Split out from ...
1494 (execute_pass_list): ... here. Adjust.
1495 (execute_ipa_pass_list): Likewise.
1496 * cgraphunit.c (cgraph_add_new_function): Adjust.
1497 (analyze_function): Likewise.
1498 (expand_function): Likewise.
1499 * cgraph.c (release_function_body): Free dominance info
1500 here instead of asserting it was magically freed elsewhere.
1502 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
1504 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
1505 * configure: Regenerate.
1506 * config/sparc/sparc.opt (muser-mode): New option.
1507 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
1509 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
1510 * doc/invoke.texi (SPARC options): Document -muser-mode.
1512 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
1514 * cselib.c (find_slot_memmode): Delete.
1515 (cselib_hasher): Change compare_type to a struct.
1516 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
1518 (preserve_constants_and_equivs): Adjust for new compare_type.
1519 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
1520 (wrap_constant): Delete.
1521 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
1523 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
1525 * doc/install.texi (Building with profile feedback): Remove
1528 2014-04-26 Tom de Vries <tom@codesourcery.com>
1530 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
1533 2014-04-25 Cary Coutant <ccoutant@google.com>
1536 * dwarf2out.c (should_move_die_to_comdat): A type definition
1537 can contain a subprogram definition, but don't move it to a
1539 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
1540 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
1542 (clone_tree_hash): Rename to...
1543 (clone_tree_partial): ...this; change callers. Copy
1544 DW_TAG_subprogram DIEs as declarations.
1545 (copy_decls_walk): Don't copy children of a declaration into a
1548 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
1551 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
1554 2014-04-25 Jiong Wang <jiong.wang@arm.com>
1556 * config/arm/predicates.md (call_insn_operand): Add long_call check.
1557 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
1559 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
1562 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1564 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
1567 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1569 PR tree-optimization/60930
1570 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
1571 creating a multiply candidate by folding two constant
1572 multiplicands when the result overflows.
1574 2014-04-25 Jakub Jelinek <jakub@redhat.com>
1576 PR tree-optimization/60960
1577 * tree-vect-generic.c (expand_vector_operation): Only call
1578 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
1580 2014-04-25 Tom de Vries <tom@codesourcery.com>
1582 * expr.c (clobber_reg_mode): New function.
1583 * expr.h (clobber_reg): New function.
1585 2014-04-25 Tom de Vries <tom@codesourcery.com>
1587 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
1590 2014-04-25 Radovan Obradovic <robradovic@mips.com>
1591 Tom de Vries <tom@codesourcery.com>
1593 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
1595 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
1596 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
1597 new argument to find_all_hard_reg_sets call.
1599 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1601 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
1602 Use HOST_WIDE_INT_C for mask literal.
1603 (aarch_rev16_shleft_mask_imm_p): Likewise.
1605 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
1608 * config/sparc/sparc.md (ashlsi3_extend): Delete.
1610 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
1612 PR preprocessor/56540
1613 * config/i386/i386-c.c (ix86_target_macros): Define
1614 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
1616 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1618 * configure.ac (tga_func): Remove.
1619 (LIB_TLS_SPEC): Remove.
1620 * configure: Regenerate.
1621 * config.in: Regenerate.
1622 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
1624 2014-04-25 Richard Biener <rguenther@suse.de>
1627 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
1628 call stmt use/clobber sets during stmt walk instead of
1629 walking the possibly incomplete set of caller edges.
1631 2014-04-25 Richard Biener <rguenther@suse.de>
1634 * passes.c (apply_ipa_transforms): Inline into only caller ...
1635 (execute_one_pass): ... here. Properly bring in function
1636 bodies for nodes we want to apply IPA transforms to.
1638 2014-04-24 Cong Hou <congh@google.com>
1640 PR tree-optimization/60896
1641 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
1642 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
1643 (vect_mark_pattern_stmts): Set the def type of all statements in
1644 PATTERN_DEF_SEQ as vect_internal_def.
1646 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
1648 * doc/extend.texi (PowerPC Built-in Functions): Document new
1649 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
1651 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
1653 * config/rs6000/predicates.md (const_0_to_3_operand): New
1654 predicate to match 0..3 integer constants.
1656 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
1657 to support adding miscellaneous builtin functions.
1658 (BU_DFP_MISC_2): Likewise.
1659 (BU_P7_MISC_1): Likewise.
1660 (BU_P7_MISC_2): Likewise.
1661 (BU_P8V_MISC_3): Likewise.
1662 (BU_MISC_1): Likewise.
1663 (BU_MISC_2): Likewise.
1664 (DIVWE): Add extended divide builtin functions.
1667 (DIVWEUO): Likewise.
1671 (DIVDEUO): Likewise.
1672 (DXEX): Add decimal floating-point builtin functions.
1675 (DDEDPDQ): Likewise.
1677 (DENBCDQ): Likewise.
1684 (CDTBCD): Add new BCD builtin functions.
1688 (BCDADD_LT): Likewise.
1689 (BCDADD_EQ): Likewise.
1690 (BCDADD_GT): Likewise.
1691 (BCDADD_OV): Likewise.
1693 (BCDSUB_LT): Likewise.
1694 (BCDSUB_EQ): Likewise.
1695 (BCDSUB_GT): Likewise.
1696 (BCDSUB_OV): Likewise.
1697 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
1698 (UNPACK_TD): Likewise.
1699 (PACK_TF): Likewise.
1700 (UNPACK_TF): Likewise.
1701 (UNPACK_TF_0): Likewise.
1702 (UNPACK_TF_1): Likewise.
1703 (PACK_V1TI): Likewise.
1704 (UNPACK_V1TI): Likewise.
1706 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
1707 support for decimal floating point builtin functions.
1708 (rs6000_expand_ternop_builtin): Add checks for the new builtin
1709 functions that take constant arguments.
1710 (rs6000_invalid_builtin): Add decimal floating point builtin
1712 (rs6000_init_builtins): Setup long double, _Decimal64, and
1713 _Decimal128 types for new builtin functions.
1714 (builtin_function_type): Set the unsigned flags appropriately for
1715 the new builtin functions.
1716 (rs6000_opt_masks): Add support for decimal floating point builtin
1719 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
1720 floating point builtin functions.
1721 (RS6000_BTM_COMMON): Likewise.
1722 (RS6000_BTI_long_double): Likewise.
1723 (RS6000_BTI_dfloat64): Likewise.
1724 (RS6000_BTI_dfloat128): Likewise.
1725 (long_double_type_internal_node): Likewise.
1726 (dfloat64_type_internal_node): Likewise.
1727 (dfloat128_type_internal_node): Likewise.
1729 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
1730 2.07 bcd arithmetic instructions.
1731 (UNSPEC_BCDSUB): Likewise.
1732 (UNSPEC_BCD_OVERFLOW): Likewise.
1733 (UNSPEC_BCD_ADD_SUB): Likewise.
1734 (bcd_add_sub): Likewise.
1735 (BCD_TEST): Likewise.
1736 (bcd<bcd_add_sub>): Likewise.
1737 (bcd<bcd_add_sub>_test): Likewise.
1738 (bcd<bcd_add_sub>_test2): Likewise.
1739 (bcd<bcd_add_sub>_<code>): Likewise.
1740 (peephole2 for combined bcd ops): Likewise.
1742 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
1743 decimal floating point builtin functions.
1744 (UNSPEC_DENBCD): Likewise.
1745 (UNSPEC_DXEX): Likewise.
1746 (UNSPEC_DIEX): Likewise.
1747 (UNSPEC_DSCLI): Likewise.
1748 (UNSPEC_DSCRI): Likewise.
1749 (D64_D128): Likewise.
1750 (dfp_suffix): Likewise.
1751 (dfp_ddedpd_<mode>): Likewise.
1752 (dfp_denbcd_<mode>): Likewise.
1753 (dfp_dxex_<mode>): Likewise.
1754 (dfp_diex_<mode>): Likewise.
1755 (dfp_dscli_<mode>): Likewise.
1756 (dfp_dscri_<mode>): Likewise.
1758 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
1760 (UNSPEC_CDTBCD): Likewise.
1761 (UNSPEC_CBCDTD): Likewise.
1762 (UNSPEC_DIVE): Add support for new extended divide builtin
1764 (UNSPEC_DIVEO): Likewise.
1765 (UNSPEC_DIVEU): Likewise.
1766 (UNSPEC_DIVEUO): Likewise.
1767 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
1768 pack/unpack 128-bit types.
1769 (UNSPEC_PACK_128BIT): Likewise.
1770 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
1771 (udiv<mode>3): Use idiv_ldiv mode attribute.
1772 (div<mode>3): Likewise.
1773 (addg6s): Add new BCD builtin functions.
1776 (UNSPEC_DIV_EXTEND): Add support for new extended divide
1778 (div_extend): Likewise.
1779 (div<div_extend>_<mode>"): Likewise.
1780 (FP128_64): Add support for new builtin functions to pack/unpack
1782 (unpack<mode>): Likewise.
1783 (unpacktf_0): Likewise.
1784 (unpacktf_1): Likewise.
1785 (unpack<mode>_dm): Likewise.
1786 (unpack<mode>_nodm): Likewise.
1787 (pack<mode>): Likewise.
1788 (unpackv1ti): Likewise.
1789 (packv1ti): Likewise.
1791 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
1793 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
1796 2014-04-24 Jakub Jelinek <jakub@redhat.com>
1798 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
1799 * gimplify.c (omp_is_private): Change last argument's type to int.
1800 Only diagnose lastprivate if the simd argument is 1, only diagnose
1801 linear if the simd argument is 2.
1802 (gimplify_omp_for): Adjust omp_is_private callers. When adding
1803 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
1804 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
1805 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
1806 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
1807 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
1808 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
1809 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
1810 * tree-nested.c (convert_nonlocal_omp_clauses,
1811 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
1813 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
1816 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
1819 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
1821 * flag-types.h (enum ivar_visibility): Add.
1823 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
1825 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
1826 function * argument.
1828 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
1830 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
1832 2014-04-24 Radovan Obradovic <robradovic@mips.com>
1833 Tom de Vries <tom@codesourcery.com>
1835 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
1836 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
1838 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
1839 * emit-rtl.c (try_split): Same.
1841 2014-04-24 Radovan Obradovic <robradovic@mips.com>
1842 Tom de Vries <tom@codesourcery.com>
1844 * common.opt (fuse-caller-save): New option.
1846 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
1848 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
1851 2014-04-24 Richard Biener <rguenther@suse.de>
1853 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
1854 during TER and instead use the sepops interface for expanding
1855 non-GIMPLE_SINGLE_RHS.
1857 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1859 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
1860 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
1862 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1864 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
1865 assembler 64-bit option.
1866 * configure: Regenerate.
1868 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1870 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
1871 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
1872 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
1873 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
1874 (TARGET_CRYPTO): Take TARGET_SIMD into account.
1876 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1878 * config/aarch64/aarch64-builtins.c
1879 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
1880 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
1881 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
1882 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
1884 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
1885 (Vrevsuff): New mode attribute.
1887 2014-04-24 Terry Guo <terry.guo@arm.com>
1889 * config/arm/arm.h (machine_function): Define variable
1890 after_arm_reorg here.
1891 * config/arm/arm.c (after_arm_reorg): Remove the definition.
1892 (arm_split_constant): Update the way to access variable
1895 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
1897 2014-04-23 Tom de Vries <tom@codesourcery.com>
1899 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
1901 2014-04-23 David Malcolm <dmalcolm@redhat.com>
1903 * is-a.h: Update comments to reflect the following changes to the
1904 "pointerness" of the API, making the template parameter match the
1905 return type, allowing use of is-a.h with typedefs of pointers.
1906 (is_a_helper::cast): Return a T rather then a pointer to a T, so
1907 that the return type matches the parameter to the is_a_helper.
1909 (dyn_cast): Likewise.
1911 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
1912 pointer from the is-a.h API.
1914 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
1915 (is_a_helper <cgraph_node *>::test): ...this, matching change to
1917 (is_a_helper <varpool_node>::test): Likewise, convert to...
1918 (is_a_helper <varpool_node *>::test): ...this.
1920 (varpool_first_variable): Update for removal of implicit pointer
1921 from the is-a.h API.
1922 (varpool_next_variable): Likewise.
1923 (varpool_first_static_initializer): Likewise.
1924 (varpool_next_static_initializer): Likewise.
1925 (varpool_first_defined_variable): Likewise.
1926 (varpool_next_defined_variable): Likewise.
1927 (cgraph_first_defined_function): Likewise.
1928 (cgraph_next_defined_function): Likewise.
1929 (cgraph_first_function): Likewise.
1930 (cgraph_next_function): Likewise.
1931 (cgraph_first_function_with_gimple_body): Likewise.
1932 (cgraph_next_function_with_gimple_body): Likewise.
1933 (cgraph_alias_target): Likewise.
1934 (varpool_alias_target): Likewise.
1935 (cgraph_function_or_thunk_node): Likewise.
1936 (varpool_variable_node): Likewise.
1937 (symtab_real_symbol_p): Likewise.
1938 * cgraphunit.c (referred_to_p): Likewise.
1939 (analyze_functions): Likewise.
1940 (handle_alias_pairs): Likewise.
1941 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1942 * gimple-ssa.h (gimple_vuse_op): Likewise.
1943 (gimple_vdef_op): Likewise.
1944 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
1945 * gimple.c (gimple_build_asm_1): Likewise.
1946 (gimple_build_try): Likewise.
1947 (gimple_build_resx): Likewise.
1948 (gimple_build_eh_dispatch): Likewise.
1949 (gimple_build_omp_for): Likewise.
1950 (gimple_omp_for_set_clauses): Likewise.
1952 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
1953 (is_a_helper <gimple_statement_asm *>::test): ...this.
1954 (is_a_helper <gimple_statement_bind>::test): Convert to...
1955 (is_a_helper <gimple_statement_bind *>::test): ...this.
1956 (is_a_helper <gimple_statement_call>::test): Convert to...
1957 (is_a_helper <gimple_statement_call *>::test): ...this.
1958 (is_a_helper <gimple_statement_catch>::test): Convert to...
1959 (is_a_helper <gimple_statement_catch *>::test): ...this.
1960 (is_a_helper <gimple_statement_resx>::test): Convert to...
1961 (is_a_helper <gimple_statement_resx *>::test): ...this.
1962 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
1963 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
1964 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
1965 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
1966 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
1967 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
1968 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
1969 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
1970 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
1971 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
1972 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
1973 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
1974 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
1975 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
1976 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
1977 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
1978 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
1979 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
1980 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
1981 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
1982 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
1983 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
1984 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
1985 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
1986 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
1987 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
1988 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
1989 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
1990 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
1991 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
1992 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
1993 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
1994 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
1995 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
1996 (is_a_helper <gimple_statement_phi>::test): Convert to...
1997 (is_a_helper <gimple_statement_phi *>::test): ...this.
1998 (is_a_helper <gimple_statement_transaction>::test): Convert to...
1999 (is_a_helper <gimple_statement_transaction *>::test): ...this.
2000 (is_a_helper <gimple_statement_try>::test): Convert to...
2001 (is_a_helper <gimple_statement_try *>::test): ...this.
2002 (is_a_helper <gimple_statement_wce>::test): Convert to...
2003 (is_a_helper <gimple_statement_wce *>::test): ...this.
2004 (is_a_helper <const gimple_statement_asm>::test): Convert to...
2005 (is_a_helper <const gimple_statement_asm *>::test): ...this.
2006 (is_a_helper <const gimple_statement_bind>::test): Convert to...
2007 (is_a_helper <const gimple_statement_bind *>::test): ...this.
2008 (is_a_helper <const gimple_statement_call>::test): Convert to...
2009 (is_a_helper <const gimple_statement_call *>::test): ...this.
2010 (is_a_helper <const gimple_statement_catch>::test): Convert to...
2011 (is_a_helper <const gimple_statement_catch *>::test): ...this.
2012 (is_a_helper <const gimple_statement_resx>::test): Convert to...
2013 (is_a_helper <const gimple_statement_resx *>::test): ...this.
2014 (is_a_helper <const gimple_statement_eh_dispatch>::test):
2016 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
2018 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
2020 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
2021 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
2023 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
2025 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
2027 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
2029 (is_a_helper <const gimple_statement_omp_return>::test): Convert
2031 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
2032 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
2034 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
2035 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
2037 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
2038 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
2039 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
2040 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
2042 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
2043 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
2045 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
2046 (is_a_helper <const gimple_statement_omp_target>::test): Convert
2048 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
2049 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
2051 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
2052 (is_a_helper <const gimple_statement_omp_single>::test): Convert
2054 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
2055 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
2057 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
2058 (is_a_helper <const gimple_statement_omp_task>::test): Convert
2060 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
2061 (is_a_helper <const gimple_statement_phi>::test): Convert to...
2062 (is_a_helper <const gimple_statement_phi *>::test): ...this.
2063 (is_a_helper <const gimple_statement_transaction>::test): Convert
2065 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
2066 (is_a_helper <const gimple_statement_with_ops>::test): Convert
2068 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
2069 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
2070 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
2071 (is_a_helper <const gimple_statement_with_memory_ops>::test):
2073 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
2075 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
2077 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
2079 (gimple_use_ops): Update for removal of implicit pointer from the
2081 (gimple_set_use_ops): Likewise.
2082 (gimple_vuse): Likewise.
2083 (gimple_vdef): Likewise.
2084 (gimple_vuse_ptr): Likewise.
2085 (gimple_vdef_ptr): Likewise.
2086 (gimple_set_vuse): Likewise.
2087 (gimple_set_vdef): Likewise.
2088 (gimple_omp_return_set_lhs): Likewise.
2089 (gimple_omp_return_lhs): Likewise.
2090 (gimple_omp_return_lhs_ptr): Likewise.
2091 (gimple_call_fntype): Likewise.
2092 (gimple_call_set_fntype): Likewise.
2093 (gimple_call_set_internal_fn): Likewise.
2094 (gimple_call_use_set): Likewise.
2095 (gimple_call_clobber_set): Likewise.
2096 (gimple_bind_vars): Likewise.
2097 (gimple_bind_set_vars): Likewise.
2098 (gimple_bind_body_ptr): Likewise.
2099 (gimple_bind_set_body): Likewise.
2100 (gimple_bind_add_stmt): Likewise.
2101 (gimple_bind_block): Likewise.
2102 (gimple_bind_set_block): Likewise.
2103 (gimple_asm_ninputs): Likewise.
2104 (gimple_asm_noutputs): Likewise.
2105 (gimple_asm_nclobbers): Likewise.
2106 (gimple_asm_nlabels): Likewise.
2107 (gimple_asm_input_op): Likewise.
2108 (gimple_asm_input_op_ptr): Likewise.
2109 (gimple_asm_output_op): Likewise.
2110 (gimple_asm_output_op_ptr): Likewise.
2111 (gimple_asm_set_output_op): Likewise.
2112 (gimple_asm_clobber_op): Likewise.
2113 (gimple_asm_set_clobber_op): Likewise.
2114 (gimple_asm_label_op): Likewise.
2115 (gimple_asm_set_label_op): Likewise.
2116 (gimple_asm_string): Likewise.
2117 (gimple_catch_types): Likewise.
2118 (gimple_catch_types_ptr): Likewise.
2119 (gimple_catch_handler_ptr): Likewise.
2120 (gimple_catch_set_types): Likewise.
2121 (gimple_catch_set_handler): Likewise.
2122 (gimple_eh_filter_types): Likewise.
2123 (gimple_eh_filter_types_ptr): Likewise.
2124 (gimple_eh_filter_failure_ptr): Likewise.
2125 (gimple_eh_filter_set_types): Likewise.
2126 (gimple_eh_filter_set_failure): Likewise.
2127 (gimple_eh_must_not_throw_fndecl): Likewise.
2128 (gimple_eh_must_not_throw_set_fndecl): Likewise.
2129 (gimple_eh_else_n_body_ptr): Likewise.
2130 (gimple_eh_else_e_body_ptr): Likewise.
2131 (gimple_eh_else_set_n_body): Likewise.
2132 (gimple_eh_else_set_e_body): Likewise.
2133 (gimple_try_eval_ptr): Likewise.
2134 (gimple_try_cleanup_ptr): Likewise.
2135 (gimple_try_set_eval): Likewise.
2136 (gimple_try_set_cleanup): Likewise.
2137 (gimple_wce_cleanup_ptr): Likewise.
2138 (gimple_wce_set_cleanup): Likewise.
2139 (gimple_phi_capacity): Likewise.
2140 (gimple_phi_num_args): Likewise.
2141 (gimple_phi_result): Likewise.
2142 (gimple_phi_result_ptr): Likewise.
2143 (gimple_phi_set_result): Likewise.
2144 (gimple_phi_arg): Likewise.
2145 (gimple_phi_set_arg): Likewise.
2146 (gimple_resx_region): Likewise.
2147 (gimple_resx_set_region): Likewise.
2148 (gimple_eh_dispatch_region): Likewise.
2149 (gimple_eh_dispatch_set_region): Likewise.
2150 (gimple_omp_critical_name): Likewise.
2151 (gimple_omp_critical_name_ptr): Likewise.
2152 (gimple_omp_critical_set_name): Likewise.
2153 (gimple_omp_for_clauses): Likewise.
2154 (gimple_omp_for_clauses_ptr): Likewise.
2155 (gimple_omp_for_set_clauses): Likewise.
2156 (gimple_omp_for_collapse): Likewise.
2157 (gimple_omp_for_index): Likewise.
2158 (gimple_omp_for_index_ptr): Likewise.
2159 (gimple_omp_for_set_index): Likewise.
2160 (gimple_omp_for_initial): Likewise.
2161 (gimple_omp_for_initial_ptr): Likewise.
2162 (gimple_omp_for_set_initial): Likewise.
2163 (gimple_omp_for_final): Likewise.
2164 (gimple_omp_for_final_ptr): Likewise.
2165 (gimple_omp_for_set_final): Likewise.
2166 (gimple_omp_for_incr): Likewise.
2167 (gimple_omp_for_incr_ptr): Likewise.
2168 (gimple_omp_for_set_incr): Likewise.
2169 (gimple_omp_for_pre_body_ptr): Likewise.
2170 (gimple_omp_for_set_pre_body): Likewise.
2171 (gimple_omp_parallel_clauses): Likewise.
2172 (gimple_omp_parallel_clauses_ptr): Likewise.
2173 (gimple_omp_parallel_set_clauses): Likewise.
2174 (gimple_omp_parallel_child_fn): Likewise.
2175 (gimple_omp_parallel_child_fn_ptr): Likewise.
2176 (gimple_omp_parallel_set_child_fn): Likewise.
2177 (gimple_omp_parallel_data_arg): Likewise.
2178 (gimple_omp_parallel_data_arg_ptr): Likewise.
2179 (gimple_omp_parallel_set_data_arg): Likewise.
2180 (gimple_omp_task_clauses): Likewise.
2181 (gimple_omp_task_clauses_ptr): Likewise.
2182 (gimple_omp_task_set_clauses): Likewise.
2183 (gimple_omp_task_child_fn): Likewise.
2184 (gimple_omp_task_child_fn_ptr): Likewise.
2185 (gimple_omp_task_set_child_fn): Likewise.
2186 (gimple_omp_task_data_arg): Likewise.
2187 (gimple_omp_task_data_arg_ptr): Likewise.
2188 (gimple_omp_task_set_data_arg): Likewise.
2189 (gimple_omp_taskreg_clauses): Likewise.
2190 (gimple_omp_taskreg_clauses_ptr): Likewise.
2191 (gimple_omp_taskreg_set_clauses): Likewise.
2192 (gimple_omp_taskreg_child_fn): Likewise.
2193 (gimple_omp_taskreg_child_fn_ptr): Likewise.
2194 (gimple_omp_taskreg_set_child_fn): Likewise.
2195 (gimple_omp_taskreg_data_arg): Likewise.
2196 (gimple_omp_taskreg_data_arg_ptr): Likewise.
2197 (gimple_omp_taskreg_set_data_arg): Likewise.
2198 (gimple_omp_task_copy_fn): Likewise.
2199 (gimple_omp_task_copy_fn_ptr): Likewise.
2200 (gimple_omp_task_set_copy_fn): Likewise.
2201 (gimple_omp_task_arg_size): Likewise.
2202 (gimple_omp_task_arg_size_ptr): Likewise.
2203 (gimple_omp_task_set_arg_size): Likewise.
2204 (gimple_omp_task_arg_align): Likewise.
2205 (gimple_omp_task_arg_align_ptr): Likewise.
2206 (gimple_omp_task_set_arg_align): Likewise.
2207 (gimple_omp_single_clauses): Likewise.
2208 (gimple_omp_single_clauses_ptr): Likewise.
2209 (gimple_omp_single_set_clauses): Likewise.
2210 (gimple_omp_target_clauses): Likewise.
2211 (gimple_omp_target_clauses_ptr): Likewise.
2212 (gimple_omp_target_set_clauses): Likewise.
2213 (gimple_omp_target_child_fn): Likewise.
2214 (gimple_omp_target_child_fn_ptr): Likewise.
2215 (gimple_omp_target_set_child_fn): Likewise.
2216 (gimple_omp_target_data_arg): Likewise.
2217 (gimple_omp_target_data_arg_ptr): Likewise.
2218 (gimple_omp_target_set_data_arg): Likewise.
2219 (gimple_omp_teams_clauses): Likewise.
2220 (gimple_omp_teams_clauses_ptr): Likewise.
2221 (gimple_omp_teams_set_clauses): Likewise.
2222 (gimple_omp_sections_clauses): Likewise.
2223 (gimple_omp_sections_clauses_ptr): Likewise.
2224 (gimple_omp_sections_set_clauses): Likewise.
2225 (gimple_omp_sections_control): Likewise.
2226 (gimple_omp_sections_control_ptr): Likewise.
2227 (gimple_omp_sections_set_control): Likewise.
2228 (gimple_omp_for_set_cond): Likewise.
2229 (gimple_omp_for_cond): Likewise.
2230 (gimple_omp_atomic_store_set_val): Likewise.
2231 (gimple_omp_atomic_store_val): Likewise.
2232 (gimple_omp_atomic_store_val_ptr): Likewise.
2233 (gimple_omp_atomic_load_set_lhs): Likewise.
2234 (gimple_omp_atomic_load_lhs): Likewise.
2235 (gimple_omp_atomic_load_lhs_ptr): Likewise.
2236 (gimple_omp_atomic_load_set_rhs): Likewise.
2237 (gimple_omp_atomic_load_rhs): Likewise.
2238 (gimple_omp_atomic_load_rhs_ptr): Likewise.
2239 (gimple_omp_continue_control_def): Likewise.
2240 (gimple_omp_continue_control_def_ptr): Likewise.
2241 (gimple_omp_continue_set_control_def): Likewise.
2242 (gimple_omp_continue_control_use): Likewise.
2243 (gimple_omp_continue_control_use_ptr): Likewise.
2244 (gimple_omp_continue_set_control_use): Likewise.
2245 (gimple_transaction_body_ptr): Likewise.
2246 (gimple_transaction_label): Likewise.
2247 (gimple_transaction_label_ptr): Likewise.
2248 (gimple_transaction_set_body): Likewise.
2249 (gimple_transaction_set_label): Likewise.
2251 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
2252 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2253 * ipa-ref.c (ipa_record_reference): Likewise.
2254 * ipa-reference.c (analyze_function): Likewise.
2255 (ipa_reference_write_optimization_summary): Likewise.
2256 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
2257 (address_taken_from_non_vtable_p): Likewise.
2258 (comdat_can_be_unshared_p_1): Likewise.
2259 * lto-cgraph.c (lto_output_ref): Likewise.
2260 (add_references): Likewise.
2261 (compute_ltrans_boundary): Likewise.
2262 (output_symtab): Likewise.
2263 (input_ref): Likewise.
2264 (input_cgraph_1): Likewise.
2265 (output_cgraph_opt_summary): Likewise.
2266 * lto-streamer-out.c (lto_output): Likewise.
2267 (output_symbol_p): Likewise.
2268 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
2269 (lsei_start_function_in_partition): Likewise.
2270 (lsei_next_variable_in_partition): Likewise.
2271 (lsei_start_variable_in_partition): Likewise.
2272 * symtab.c (insert_to_assembler_name_hash): Likewise.
2273 (unlink_from_assembler_name_hash): Likewise.
2274 (symtab_unregister_node): Likewise.
2275 (symtab_remove_node): Likewise.
2276 (dump_symtab_node): Likewise.
2277 (verify_symtab_base): Likewise.
2278 (verify_symtab_node): Likewise.
2279 (symtab_make_decl_local): Likewise.
2280 (symtab_alias_ultimate_target): Likewise.
2281 (symtab_resolve_alias): Likewise.
2282 (symtab_get_symbol_partitioning_class): Likewise.
2283 * tree-phinodes.c (allocate_phi_node): Likewise.
2284 (reserve_phi_args_for_new_edge): Likewise.
2285 (remove_phi_args): Likewise.
2286 * varpool.c (varpool_node_for_asm): Likewise.
2287 (varpool_remove_unreferenced_decls): Likewise.
2289 2014-04-23 Jeff Law <law@redhat.com>
2291 PR tree-optimization/60902
2292 * tree-ssa-threadedge.c
2293 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
2294 invalidate outputs from statements that do not produce useful
2295 outputs for threading.
2297 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2299 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
2300 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
2301 machine descriptions for Stack Smashing Protector.
2303 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
2305 * aarch64.md (<optab>_rol<mode>3): New pattern.
2306 (<optab>_rolsi3_uxtw): Likewise.
2307 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
2309 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
2311 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
2312 (arm_cortex_a12_tune): Likewise.
2314 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2316 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
2318 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2320 * config/arm/arm.md (arm_rev16si2): New pattern.
2321 (arm_rev16si2_alt): Likewise.
2322 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
2324 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2326 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
2327 (rev16<mode>2_alt): Likewise.
2328 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
2329 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
2330 (aarch_rev16_shleft_mask_imm_p): Likewise.
2331 (aarch_rev16_p_1): Likewise.
2332 (aarch_rev16_p): Likewise.
2333 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
2334 (aarch_rev16_shright_mask_imm_p): Likewise.
2335 (aarch_rev16_shleft_mask_imm_p): Likewise.
2337 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2339 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
2340 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
2342 (cortex_a53_extra_costs): Likewise.
2343 (cortex_a57_extra_costs): Likewise.
2344 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
2345 (cortexa7_extra_costs): Likewise.
2346 (cortexa8_extra_costs): Likewise.
2347 (cortexa12_extra_costs): Likewise.
2348 (cortexa15_extra_costs): Likewise.
2349 (v7m_extra_costs): Likewise.
2350 (arm_new_rtx_costs): Handle BSWAP.
2352 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2354 * config/arm/arm.c (cortexa8_extra_costs): New table.
2355 (arm_cortex_a8_tune): New tuning struct.
2356 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
2358 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2360 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
2362 2014-04-23 Richard Biener <rguenther@suse.de>
2364 * Makefile.in (OBJS): Remove loop-unswitch.o.
2365 * tree-pass.h (make_pass_rtl_unswitch): Remove.
2366 * passes.def (pass_rtl_unswitch): Likewise.
2367 * loop-init.c (gate_rtl_unswitch): Likewise.
2368 (rtl_unswitch): Likewise.
2369 (pass_data_rtl_unswitch): Likewise.
2370 (pass_rtl_unswitch): Likewise.
2371 (make_pass_rtl_unswitch): Likewise.
2372 * rtl.h (reversed_condition): Likewise.
2373 (compare_and_jump_seq): Likewise.
2374 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
2376 * loop-unroll.c (compare_and_jump_seq): Likewise.
2378 2014-04-23 Richard Biener <rguenther@suse.de>
2380 PR tree-optimization/60903
2381 * tree-ssa-loop-im.c (analyze_memory_references): Remove
2382 commented code block.
2383 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
2384 loop flags to newly created BBs and edges.
2386 2014-04-23 Nick Clifton <nickc@redhat.com>
2388 * config/msp430/msp430.c (msp430_handle_option): Move function
2390 (msp430_option_override): Simplify mcu and mcpu option handling.
2391 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
2392 support for -mhwmult command line option.
2393 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
2394 -mhwmult command line option.
2395 (msp430_hwmult_enabled): Delete.
2396 (msp43o_output_labelref): Add support for -mhwmult command line
2398 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
2399 (umulsidi3): Likewise.
2400 * config/msp430/msp430.opt (mmcu): Add Report attribute.
2401 (mcpu, mlarge, msmall): Likewise.
2402 (mhwmult): New option.
2403 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
2405 (msp430_is_f5_mcu): Remove prototype.
2406 (msp430_use_f5_series_hwmult): Add prototype.
2407 * config/msp430/msp430-opts.h: New file.
2408 * common/config/msp430: New directory.
2409 * common/config/msp430/msp430-common.c: New file.
2410 * config.gcc (msp430): Remove target_has_targetm_common.
2411 * doc/invoke.texi: Document -mhwmult command line option.
2413 2014-04-23 Nick Clifton <nickc@redhat.com>
2415 * config/i386/cygwin.h (ENDFILE_SPEC): Include
2416 default-manifest.o if it can be found in the search path.
2417 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
2419 2014-04-23 Terry Guo <terry.guo@arm.com>
2421 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
2423 2014-04-23 Richard Biener <rguenther@suse.de>
2426 * tree-inline.c (declare_return_variable): Use mark_addressable.
2428 2014-04-23 Richard Biener <rguenther@suse.de>
2431 * loop-init.c (loop_optimizer_init): Make sure to apply
2432 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
2434 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2437 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
2439 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
2440 if flag_sanitize_undefined_trap_on_error.
2441 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
2442 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
2443 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
2444 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
2445 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
2446 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
2447 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
2448 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
2449 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
2450 * ubsan.c (ubsan_instrument_unreachable): Return
2451 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
2452 (ubsan_expand_null_ifn): Emit __builtin_trap ()
2453 if flag_sanitize_undefined_trap_on_error and
2454 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
2455 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
2456 instrument_bool_enum_load): Emit __builtin_trap () if
2457 flag_sanitize_undefined_trap_on_error and
2458 __builtin_handle_*_abort () if !flag_sanitize_recover.
2459 * doc/invoke.texi (-fsanitize-recover,
2460 -fsanitize-undefined-trap-on-error): Document.
2462 2014-04-22 Christian Bruel <christian.bruel@st.com>
2464 * config/sh/sh.md (mov<mode>): Replace movQIHI.
2465 Force immediates to SImode.
2467 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
2469 * config/nios2/nios2.md (UNSPEC_ROUND): New.
2471 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
2472 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
2473 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
2474 (nios2_fpu_insn): Add entry for round.
2475 (N2FPU_NO_ERRNO_P): Define.
2476 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
2478 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
2480 2014-04-22 Richard Henderson <rth@redhat.com>
2482 * config/aarch64/aarch64 (addti3, subti3): New expanders.
2483 (add<GPI>3_compare0): Remove leading * from name.
2484 (add<GPI>3_carryin): Likewise.
2485 (sub<GPI>3_compare0): Likewise.
2486 (sub<GPI>3_carryin): Likewise.
2487 (<su_optab>mulditi3): New expander.
2488 (multi3): New expander.
2489 (madd<GPI>): Remove leading * from name.
2491 2014-04-22 Martin Jambor <mjambor@suse.cz>
2493 * cgraphclones.c (cgraph_function_versioning): Copy
2494 ipa_transforms_to_apply instead of asserting it is empty.
2496 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
2499 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
2500 on count_exp to get mode.
2502 2014-04-22 Andrew Pinski <apinski@cavium.com>
2504 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
2505 Handle TLS for ILP32.
2506 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
2507 (tlsie_small_<mode>): this and handle PTR.
2508 (tlsie_small_sidi): New pattern.
2509 (tlsle_small): Change to an expand to handle ILP32.
2510 (tlsle_small_<mode>): New pattern.
2511 (tlsdesc_small): Rename to ...
2512 (tlsdesc_small_<mode>): this and handle PTR.
2514 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2516 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
2518 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2520 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2521 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
2522 (aarch64_types_signed_poly_qualifiers): Likewise.
2523 (aarch64_types_unsigned_signed_qualifiers): Likewise.
2524 (aarch64_types_poly_signed_qualifiers): Likewise.
2525 (TYPES_REINTERP_SS): Type macro added.
2526 (TYPES_REINTERP_SU): Likewise.
2527 (TYPES_REINTERP_SP): Likewise.
2528 (TYPES_REINTERP_US): Likewise.
2529 (TYPES_REINTERP_PS): Likewise.
2530 (aarch64_fold_builtin): New expression folding added.
2531 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
2532 Declarations removed.
2533 (REINTERP_SS): Declarations added.
2534 (REINTERP_US): Likewise.
2535 (REINTERP_PS): Likewise.
2536 (REINTERP_SU): Likewise.
2537 (REINTERP_SP): Likewise.
2538 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
2539 (vreinterpretq_p8_f64): Likewise.
2540 (vreinterpret_p16_f64): Likewise.
2541 (vreinterpretq_p16_f64): Likewise.
2542 (vreinterpret_f32_f64): Likewise.
2543 (vreinterpretq_f32_f64): Likewise.
2544 (vreinterpret_f64_f32): Likewise.
2545 (vreinterpret_f64_p8): Likewise.
2546 (vreinterpret_f64_p16): Likewise.
2547 (vreinterpret_f64_s8): Likewise.
2548 (vreinterpret_f64_s16): Likewise.
2549 (vreinterpret_f64_s32): Likewise.
2550 (vreinterpret_f64_s64): Likewise.
2551 (vreinterpret_f64_u8): Likewise.
2552 (vreinterpret_f64_u16): Likewise.
2553 (vreinterpret_f64_u32): Likewise.
2554 (vreinterpret_f64_u64): Likewise.
2555 (vreinterpretq_f64_f32): Likewise.
2556 (vreinterpretq_f64_p8): Likewise.
2557 (vreinterpretq_f64_p16): Likewise.
2558 (vreinterpretq_f64_s8): Likewise.
2559 (vreinterpretq_f64_s16): Likewise.
2560 (vreinterpretq_f64_s32): Likewise.
2561 (vreinterpretq_f64_s64): Likewise.
2562 (vreinterpretq_f64_u8): Likewise.
2563 (vreinterpretq_f64_u16): Likewise.
2564 (vreinterpretq_f64_u32): Likewise.
2565 (vreinterpretq_f64_u64): Likewise.
2566 (vreinterpret_s64_f64): Likewise.
2567 (vreinterpretq_s64_f64): Likewise.
2568 (vreinterpret_u64_f64): Likewise.
2569 (vreinterpretq_u64_f64): Likewise.
2570 (vreinterpret_s8_f64): Likewise.
2571 (vreinterpretq_s8_f64): Likewise.
2572 (vreinterpret_s16_f64): Likewise.
2573 (vreinterpretq_s16_f64): Likewise.
2574 (vreinterpret_s32_f64): Likewise.
2575 (vreinterpretq_s32_f64): Likewise.
2576 (vreinterpret_u8_f64): Likewise.
2577 (vreinterpretq_u8_f64): Likewise.
2578 (vreinterpret_u16_f64): Likewise.
2579 (vreinterpretq_u16_f64): Likewise.
2580 (vreinterpret_u32_f64): Likewise.
2581 (vreinterpretq_u32_f64): Likewise.
2583 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2585 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2586 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
2587 (vreinterpret_p8_s8): Likewise.
2588 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
2589 (vreinterpret_p8_s16): Likewise.
2590 (vreinterpret_p8_s32): Likewise.
2591 (vreinterpret_p8_s64): Likewise.
2592 (vreinterpret_p8_f32): Likewise.
2593 (vreinterpret_p8_u8): Likewise.
2594 (vreinterpret_p8_u16): Likewise.
2595 (vreinterpret_p8_u32): Likewise.
2596 (vreinterpret_p8_u64): Likewise.
2597 (vreinterpret_p8_p16): Likewise.
2598 (vreinterpretq_p8_s8): Likewise.
2599 (vreinterpretq_p8_s16): Likewise.
2600 (vreinterpretq_p8_s32): Likewise.
2601 (vreinterpretq_p8_s64): Likewise.
2602 (vreinterpretq_p8_f32): Likewise.
2603 (vreinterpretq_p8_u8): Likewise.
2604 (vreinterpretq_p8_u16): Likewise.
2605 (vreinterpretq_p8_u32): Likewise.
2606 (vreinterpretq_p8_u64): Likewise.
2607 (vreinterpretq_p8_p16): Likewise.
2608 (vreinterpret_p16_s8): Likewise.
2609 (vreinterpret_p16_s16): Likewise.
2610 (vreinterpret_p16_s32): Likewise.
2611 (vreinterpret_p16_s64): Likewise.
2612 (vreinterpret_p16_f32): Likewise.
2613 (vreinterpret_p16_u8): Likewise.
2614 (vreinterpret_p16_u16): Likewise.
2615 (vreinterpret_p16_u32): Likewise.
2616 (vreinterpret_p16_u64): Likewise.
2617 (vreinterpret_p16_p8): Likewise.
2618 (vreinterpretq_p16_s8): Likewise.
2619 (vreinterpretq_p16_s16): Likewise.
2620 (vreinterpretq_p16_s32): Likewise.
2621 (vreinterpretq_p16_s64): Likewise.
2622 (vreinterpretq_p16_f32): Likewise.
2623 (vreinterpretq_p16_u8): Likewise.
2624 (vreinterpretq_p16_u16): Likewise.
2625 (vreinterpretq_p16_u32): Likewise.
2626 (vreinterpretq_p16_u64): Likewise.
2627 (vreinterpretq_p16_p8): Likewise.
2628 (vreinterpret_f32_s8): Likewise.
2629 (vreinterpret_f32_s16): Likewise.
2630 (vreinterpret_f32_s32): Likewise.
2631 (vreinterpret_f32_s64): Likewise.
2632 (vreinterpret_f32_u8): Likewise.
2633 (vreinterpret_f32_u16): Likewise.
2634 (vreinterpret_f32_u32): Likewise.
2635 (vreinterpret_f32_u64): Likewise.
2636 (vreinterpret_f32_p8): Likewise.
2637 (vreinterpret_f32_p16): Likewise.
2638 (vreinterpretq_f32_s8): Likewise.
2639 (vreinterpretq_f32_s16): Likewise.
2640 (vreinterpretq_f32_s32): Likewise.
2641 (vreinterpretq_f32_s64): Likewise.
2642 (vreinterpretq_f32_u8): Likewise.
2643 (vreinterpretq_f32_u16): Likewise.
2644 (vreinterpretq_f32_u32): Likewise.
2645 (vreinterpretq_f32_u64): Likewise.
2646 (vreinterpretq_f32_p8): Likewise.
2647 (vreinterpretq_f32_p16): Likewise.
2648 (vreinterpret_s64_s8): Likewise.
2649 (vreinterpret_s64_s16): Likewise.
2650 (vreinterpret_s64_s32): Likewise.
2651 (vreinterpret_s64_f32): Likewise.
2652 (vreinterpret_s64_u8): Likewise.
2653 (vreinterpret_s64_u16): Likewise.
2654 (vreinterpret_s64_u32): Likewise.
2655 (vreinterpret_s64_u64): Likewise.
2656 (vreinterpret_s64_p8): Likewise.
2657 (vreinterpret_s64_p16): Likewise.
2658 (vreinterpretq_s64_s8): Likewise.
2659 (vreinterpretq_s64_s16): Likewise.
2660 (vreinterpretq_s64_s32): Likewise.
2661 (vreinterpretq_s64_f32): Likewise.
2662 (vreinterpretq_s64_u8): Likewise.
2663 (vreinterpretq_s64_u16): Likewise.
2664 (vreinterpretq_s64_u32): Likewise.
2665 (vreinterpretq_s64_u64): Likewise.
2666 (vreinterpretq_s64_p8): Likewise.
2667 (vreinterpretq_s64_p16): Likewise.
2668 (vreinterpret_u64_s8): Likewise.
2669 (vreinterpret_u64_s16): Likewise.
2670 (vreinterpret_u64_s32): Likewise.
2671 (vreinterpret_u64_s64): Likewise.
2672 (vreinterpret_u64_f32): Likewise.
2673 (vreinterpret_u64_u8): Likewise.
2674 (vreinterpret_u64_u16): Likewise.
2675 (vreinterpret_u64_u32): Likewise.
2676 (vreinterpret_u64_p8): Likewise.
2677 (vreinterpret_u64_p16): Likewise.
2678 (vreinterpretq_u64_s8): Likewise.
2679 (vreinterpretq_u64_s16): Likewise.
2680 (vreinterpretq_u64_s32): Likewise.
2681 (vreinterpretq_u64_s64): Likewise.
2682 (vreinterpretq_u64_f32): Likewise.
2683 (vreinterpretq_u64_u8): Likewise.
2684 (vreinterpretq_u64_u16): Likewise.
2685 (vreinterpretq_u64_u32): Likewise.
2686 (vreinterpretq_u64_p8): Likewise.
2687 (vreinterpretq_u64_p16): Likewise.
2688 (vreinterpret_s8_s16): Likewise.
2689 (vreinterpret_s8_s32): Likewise.
2690 (vreinterpret_s8_s64): Likewise.
2691 (vreinterpret_s8_f32): Likewise.
2692 (vreinterpret_s8_u8): Likewise.
2693 (vreinterpret_s8_u16): Likewise.
2694 (vreinterpret_s8_u32): Likewise.
2695 (vreinterpret_s8_u64): Likewise.
2696 (vreinterpret_s8_p8): Likewise.
2697 (vreinterpret_s8_p16): Likewise.
2698 (vreinterpretq_s8_s16): Likewise.
2699 (vreinterpretq_s8_s32): Likewise.
2700 (vreinterpretq_s8_s64): Likewise.
2701 (vreinterpretq_s8_f32): Likewise.
2702 (vreinterpretq_s8_u8): Likewise.
2703 (vreinterpretq_s8_u16): Likewise.
2704 (vreinterpretq_s8_u32): Likewise.
2705 (vreinterpretq_s8_u64): Likewise.
2706 (vreinterpretq_s8_p8): Likewise.
2707 (vreinterpretq_s8_p16): Likewise.
2708 (vreinterpret_s16_s8): Likewise.
2709 (vreinterpret_s16_s32): Likewise.
2710 (vreinterpret_s16_s64): Likewise.
2711 (vreinterpret_s16_f32): Likewise.
2712 (vreinterpret_s16_u8): Likewise.
2713 (vreinterpret_s16_u16): Likewise.
2714 (vreinterpret_s16_u32): Likewise.
2715 (vreinterpret_s16_u64): Likewise.
2716 (vreinterpret_s16_p8): Likewise.
2717 (vreinterpret_s16_p16): Likewise.
2718 (vreinterpretq_s16_s8): Likewise.
2719 (vreinterpretq_s16_s32): Likewise.
2720 (vreinterpretq_s16_s64): Likewise.
2721 (vreinterpretq_s16_f32): Likewise.
2722 (vreinterpretq_s16_u8): Likewise.
2723 (vreinterpretq_s16_u16): Likewise.
2724 (vreinterpretq_s16_u32): Likewise.
2725 (vreinterpretq_s16_u64): Likewise.
2726 (vreinterpretq_s16_p8): Likewise.
2727 (vreinterpretq_s16_p16): Likewise.
2728 (vreinterpret_s32_s8): Likewise.
2729 (vreinterpret_s32_s16): Likewise.
2730 (vreinterpret_s32_s64): Likewise.
2731 (vreinterpret_s32_f32): Likewise.
2732 (vreinterpret_s32_u8): Likewise.
2733 (vreinterpret_s32_u16): Likewise.
2734 (vreinterpret_s32_u32): Likewise.
2735 (vreinterpret_s32_u64): Likewise.
2736 (vreinterpret_s32_p8): Likewise.
2737 (vreinterpret_s32_p16): Likewise.
2738 (vreinterpretq_s32_s8): Likewise.
2739 (vreinterpretq_s32_s16): Likewise.
2740 (vreinterpretq_s32_s64): Likewise.
2741 (vreinterpretq_s32_f32): Likewise.
2742 (vreinterpretq_s32_u8): Likewise.
2743 (vreinterpretq_s32_u16): Likewise.
2744 (vreinterpretq_s32_u32): Likewise.
2745 (vreinterpretq_s32_u64): Likewise.
2746 (vreinterpretq_s32_p8): Likewise.
2747 (vreinterpretq_s32_p16): Likewise.
2748 (vreinterpret_u8_s8): Likewise.
2749 (vreinterpret_u8_s16): Likewise.
2750 (vreinterpret_u8_s32): Likewise.
2751 (vreinterpret_u8_s64): Likewise.
2752 (vreinterpret_u8_f32): Likewise.
2753 (vreinterpret_u8_u16): Likewise.
2754 (vreinterpret_u8_u32): Likewise.
2755 (vreinterpret_u8_u64): Likewise.
2756 (vreinterpret_u8_p8): Likewise.
2757 (vreinterpret_u8_p16): Likewise.
2758 (vreinterpretq_u8_s8): Likewise.
2759 (vreinterpretq_u8_s16): Likewise.
2760 (vreinterpretq_u8_s32): Likewise.
2761 (vreinterpretq_u8_s64): Likewise.
2762 (vreinterpretq_u8_f32): Likewise.
2763 (vreinterpretq_u8_u16): Likewise.
2764 (vreinterpretq_u8_u32): Likewise.
2765 (vreinterpretq_u8_u64): Likewise.
2766 (vreinterpretq_u8_p8): Likewise.
2767 (vreinterpretq_u8_p16): Likewise.
2768 (vreinterpret_u16_s8): Likewise.
2769 (vreinterpret_u16_s16): Likewise.
2770 (vreinterpret_u16_s32): Likewise.
2771 (vreinterpret_u16_s64): Likewise.
2772 (vreinterpret_u16_f32): Likewise.
2773 (vreinterpret_u16_u8): Likewise.
2774 (vreinterpret_u16_u32): Likewise.
2775 (vreinterpret_u16_u64): Likewise.
2776 (vreinterpret_u16_p8): Likewise.
2777 (vreinterpret_u16_p16): Likewise.
2778 (vreinterpretq_u16_s8): Likewise.
2779 (vreinterpretq_u16_s16): Likewise.
2780 (vreinterpretq_u16_s32): Likewise.
2781 (vreinterpretq_u16_s64): Likewise.
2782 (vreinterpretq_u16_f32): Likewise.
2783 (vreinterpretq_u16_u8): Likewise.
2784 (vreinterpretq_u16_u32): Likewise.
2785 (vreinterpretq_u16_u64): Likewise.
2786 (vreinterpretq_u16_p8): Likewise.
2787 (vreinterpretq_u16_p16): Likewise.
2788 (vreinterpret_u32_s8): Likewise.
2789 (vreinterpret_u32_s16): Likewise.
2790 (vreinterpret_u32_s32): Likewise.
2791 (vreinterpret_u32_s64): Likewise.
2792 (vreinterpret_u32_f32): Likewise.
2793 (vreinterpret_u32_u8): Likewise.
2794 (vreinterpret_u32_u16): Likewise.
2795 (vreinterpret_u32_u64): Likewise.
2796 (vreinterpret_u32_p8): Likewise.
2797 (vreinterpret_u32_p16): Likewise.
2798 (vreinterpretq_u32_s8): Likewise.
2799 (vreinterpretq_u32_s16): Likewise.
2800 (vreinterpretq_u32_s32): Likewise.
2801 (vreinterpretq_u32_s64): Likewise.
2802 (vreinterpretq_u32_f32): Likewise.
2803 (vreinterpretq_u32_u8): Likewise.
2804 (vreinterpretq_u32_u16): Likewise.
2805 (vreinterpretq_u32_u64): Likewise.
2806 (vreinterpretq_u32_p8): Likewise.
2807 (vreinterpretq_u32_p16): Likewise.
2809 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2811 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
2813 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
2816 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
2817 (vqnegd_s64): Likewise.
2818 (vqabs_s64): Likewise.
2819 (vqabsd_s64): Likewise.
2821 2014-04-22 Richard Henderson <rth@redhat.com>
2823 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
2824 computation to the top of the loop.
2826 2014-04-22 Renlin <renlin.li@arm.com>
2827 Jiong Wang <jiong.wang@arm.com>
2829 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
2830 * config/aarch64/aarch64.c (aarch64_layout_frame)
2831 (aarch64_initial_elimination_offset): Likewise.
2833 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
2835 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
2838 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
2840 * machmode.h (bitwise_mode_for_mode): Declare.
2841 * stor-layout.h (bitwise_type_for_mode): Likewise.
2842 * stor-layout.c (bitwise_mode_for_mode): New function.
2843 (bitwise_type_for_mode): Likewise.
2844 * builtins.c (fold_builtin_memory_op): Use it instead of
2845 int_mode_for_mode and build_nonstandard_integer_type.
2847 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2849 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
2850 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
2851 (*-*-solaris2*): Simplify.
2852 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
2853 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
2854 *-*-solaris2.9* handling.
2856 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
2858 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
2859 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
2861 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
2862 * configure: Regenerate.
2864 * config/i386/sol2-9.h: Remove.
2866 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
2867 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
2868 Remove Solaris 9 references.
2870 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
2872 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
2873 (floatuns<GPI:mode><GPF:mode>2): Remove.
2874 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
2875 and floatuns conversions.
2876 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
2877 and floatuns conversions.
2878 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
2879 (w1,w2): New mode attributes for inequal width conversions.
2881 2014-04-22 Renlin Li <Renlin.Li@arm.com>
2883 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
2884 the output asm format.
2886 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
2888 * config/aarch64/aarch64-simd.md
2889 (aarch64_cm<optab>di): Always split.
2890 (*aarch64_cm<optab>di): New.
2891 (aarch64_cmtstdi): Always split.
2892 (*aarch64_cmtstdi): New.
2894 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2896 PR tree-optimization/60823
2897 * omp-low.c (ipa_simd_modify_function_body): Go through
2898 all SSA_NAMEs and for those refering to vector arguments
2899 which are going to be replaced adjust SSA_NAME_VAR and,
2900 if it is a default definition, change it into a non-default
2901 definition assigned at the beginning of function from new_decl.
2902 (ipa_simd_modify_stmt_ops): Rewritten.
2903 * tree-dfa.c (set_ssa_default_def): When removing default def,
2904 check for NULL loc instead of NULL *loc.
2906 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2908 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
2909 restrictions on core registers for DImode values in Thumb2.
2911 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2913 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
2914 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
2916 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2918 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
2919 (*iordi_notzesidi_di): Likewise.
2920 (*iordi_notsesidi_di): Likewise.
2922 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2924 * config/arm/arm-protos.h (tune_params): New struct members.
2925 * config/arm/arm.c: Initialise tune_params per processor.
2926 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
2927 for speed, based on new tune_params.
2929 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2931 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
2933 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
2935 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
2937 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
2938 * config/aarch64/arm_neon.h (vrnd_f64): Added.
2939 (vrnda_f64): Likewise.
2940 (vrndi_f64): Likewise.
2941 (vrndm_f64): Likewise.
2942 (vrndn_f64): Likewise.
2943 (vrndp_f64): Likewise.
2944 (vrndx_f64): Likewise.
2946 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2948 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
2949 GET_MODE_SIZE argument is enum machine_mode.
2951 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2954 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
2955 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
2957 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
2960 * asan.c (asan_emit_stack_protection): Force the base to align to
2961 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
2962 appropriate bits if STRICT_ALIGNMENT.
2963 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
2965 (expand_used_vars): Leave a space in the stack frame for alignment
2966 if STRICT_ALIGNMENT.
2968 2014-04-21 David Malcolm <dmalcolm@redhat.com>
2970 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
2972 (gimple_store_p): Likewise.
2973 (gimple_assign_load_p): Likewise.
2974 (gimple_assign_cast_p): Likewise.
2975 (gimple_clobber_p): Likewise.
2977 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
2978 rather than a gimple.
2979 (gimple_assign_cast_p): Likewise.
2981 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
2984 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
2985 If mode is DDmode and TARGET_E500_DOUBLE allow move.
2987 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
2988 more debug information for E500 if -mdebug=reg.
2990 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
2993 * config/i386/i386.c (ix86_expand_builtin)
2994 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
2995 register for target RTX.
2996 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
2998 2014-04-18 Cong Hou <congh@google.com>
3000 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
3001 the widen-mult pattern by handling two operands with different sizes,
3002 and operands whose size is smaller than half of the result type.
3004 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3006 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
3007 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
3008 (do_estimate_edge_time): Compute it.
3009 * ipa-inline.c (want_inline_small_function_p): Bypass
3010 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
3012 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3014 * ipa-inline.c (spec_rem): New static variable.
3015 (dump_overall_stats): New function.
3016 (dump_inline_stats): New function.
3018 2014-04-18 Richard Henderson <rth@redhat.com>
3020 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
3021 to GET_MODE_SIZE, not a reg_class_t.
3023 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3025 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
3026 (vsx_xxmrglw_<mode>): Likewise.
3028 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3031 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
3032 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
3033 (rs6000_init_hard_regno_mode_ok): Likewise.
3035 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
3037 * ipa-inline.c (inline_small_functions): Account only non-cold
3039 * doc/invoke.texi (inline-unit-growth): Update documentation.
3041 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
3043 * config/rs6000/rs6000.md (addti3, subti3): New.
3045 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
3048 * config/i386/i386.c (ix86_expand_clear): Remove outdated
3049 comment. Check optimize_insn_for_size_p instead of
3050 optimize_insn_for_speed_p.
3052 2014-04-17 Martin Jambor <mjambor@suse.cz>
3054 * gimple-iterator.c (gsi_start_edge): New function.
3055 * gimple-iterator.h (gsi_start_edge): Declare.
3056 * tree-sra.c (single_non_eh_succ): New function.
3057 (disqualify_ops_if_throwing_stmt): Renamed to
3058 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
3059 having one non-EH successor BB.
3060 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
3061 generate loads into replacements.
3062 (sra_modify_assign): Likewise and and also use the simple path for
3064 (sra_modify_function_body): Commit statements on edges.
3066 2014-04-17 Richard Biener <rguenther@suse.de>
3069 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
3070 comparison results and add clarifying comment.
3072 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3074 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
3075 (blank_mode): Initialize it.
3076 (emit_mode_size_inline, emit_mode_nunits_inline,
3077 emit_mode_inner_inline): New functions.
3078 (emit_insn_modes_h): Call them and surround their output with
3079 #if GCC_VERSION >= 4001 ... #endif.
3080 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
3081 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
3082 mode_* arrays if the argument is __builtin_constant_p.
3083 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
3084 is enum machine_mode.
3086 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3088 * passes.c (opt_pass::execute): Adjust.
3089 (pass_manager::execute_pass_mode_switching): Likewise.
3090 (early_local_passes::execute): Likewise.
3091 (execute_one_pass): Pass cfun to the pass's execute method.
3092 * tree-pass.h (opt_pass::execute): Add function * argument.
3093 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3094 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3095 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3096 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3097 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3098 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
3099 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
3100 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3101 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3102 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3103 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3104 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3105 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3106 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3107 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3108 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3109 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3110 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3111 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3112 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3113 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3114 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3115 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3116 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3117 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3118 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3119 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3120 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3121 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3124 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3126 * passes.c (opt_pass::gate): Take function * argument.
3127 (gate_all_early_local_passes): Merge into
3128 (early_local_passes::gate): this.
3129 (gate_all_early_optimizations): Merge into
3130 (all_early_optimizations::gate): this.
3131 (gate_all_optimizations): Mege into
3132 (all_optimizations::gate): this.
3133 (gate_all_optimizations_g): Merge into
3134 (all_optimizations_g::gate): this.
3135 (gate_rest_of_compilation): Mege into
3136 (rest_of_compilation::gate): this.
3137 (gate_postreload): Merge into
3138 (postreload::gate): this.
3139 (dump_one_pass): Pass cfun to the pass's gate method.
3140 (execute_ipa_summary_passes): Likewise.
3141 (execute_one_pass): Likewise.
3142 (ipa_write_summaries_2): Likewise.
3143 (ipa_write_optimization_summaries_1): Likewise.
3144 (ipa_read_summaries_1): Likewise.
3145 (ipa_read_optimization_summaries_1): Likewise.
3146 (execute_ipa_stmt_fixups): Likewise.
3147 * tree-pass.h (opt_pass::gate): Add function * argument.
3148 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
3149 combine-stack-adj.c, combine.c, compare-elim.c,
3150 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3151 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
3152 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
3153 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
3154 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3155 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3156 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3157 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
3158 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
3159 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
3160 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3161 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3162 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
3163 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3164 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3165 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3166 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3167 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3168 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3169 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3170 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3171 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3172 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3173 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
3174 var-tracking.c, vtable-verify.c, web.c: Adjust.
3176 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3178 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
3179 * configure: Regenerate.
3181 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3183 * passes.c (dump_one_pass): don't check pass->has_gate.
3184 (execute_ipa_summary_passes): Likewise.
3185 (execute_one_pass): Likewise.
3186 (ipa_write_summaries_2): Likewise.
3187 (ipa_write_optimization_summaries_1): Likewise.
3188 (ipa_read_optimization_summaries_1): Likewise.
3189 (execute_ipa_stmt_fixups): Likewise.
3190 * tree-pass.h (pass_data::has_gate): Remove.
3191 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3192 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3193 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3194 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3195 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3196 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
3197 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
3198 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
3199 gimple-low.c, gimple-ssa-isolate-paths.c,
3200 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
3201 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3202 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3203 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
3204 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
3205 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
3206 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
3207 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3208 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
3209 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3210 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3211 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3212 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3213 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3214 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3215 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3216 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3217 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3218 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3219 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3220 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3221 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3224 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3226 * pass_manager.h (pass_manager::register_dump_files_1): Remove
3228 * passes.c (pass_manager::register_dump_files_1): Merge into
3229 (pass_manager::register_dump_files): this, and remove its handling of
3230 properties since the pass always has the properties anyway.
3231 (pass_manager::pass_manager): Adjust.
3233 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3235 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
3236 * passes.c (pass_manager::register_dump_files_1): Remove dead code
3237 dealing with properties.
3238 (pass_manager::register_dump_files): Adjust.
3240 2014-03-20 Mark Wielaard <mjw@redhat.com>
3242 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
3243 then represent the bound as normal constant value.
3245 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3248 Forward port from 4.8 branch
3249 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
3251 * config/i386/bmiintrin.h (_blsi_u32): New.
3255 (_blsmsk_u32): Ditto.
3256 (_blsmsk_u64): Ditto.
3257 (_tzcnt_u32): Ditto.
3258 (_tzcnt_u64): Ditto.
3260 2014-04-17 Kito Cheng <kito@0xlab.org>
3262 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
3264 2014-04-17 Richard Biener <rguenther@suse.de>
3267 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
3268 boolean results for comparisons.
3270 2014-04-17 Richard Biener <rguenther@suse.de>
3272 PR tree-optimization/60836
3273 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
3274 initial PHI args to be gimple values.
3276 2014-04-17 Richard Biener <rguenther@suse.de>
3278 PR tree-optimization/60841
3279 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
3280 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
3281 of stmts to SLP build.
3282 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
3283 (vect_analyze_slp): Likewise.
3284 (vect_analyze_slp_instance): Likewise.
3285 (vect_build_slp_tree): Limit overall SLP tree growth.
3286 * tree-vectorizer.h (vect_analyze_data_refs,
3287 vect_analyze_slp): Adjust prototypes.
3289 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3291 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
3294 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3296 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
3297 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
3298 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
3299 for TARGET_SLOW_PSHUFB
3301 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3303 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
3304 * config/i386/i386.c (intel_cost): Ditto.
3306 2014-04-17 Joey Ye <joey.ye@arm.com>
3308 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
3310 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3312 * opts.c (common_handle_option): Disable -fipa-reference coorectly
3315 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3317 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
3318 (type_all_derivations_known_p): New predicate.
3319 (type_all_ctors_visible_p): New predicate.
3320 (type_possibly_instantiated_p): New predicate.
3321 (get_odr_type): Compute all_derivations_known.
3322 (dump_odr_type): Dump the flag.
3323 (maybe_record_type): Cleanup.
3324 (record_target_from_binfo): Add bases_to_consider array;
3325 record bases for types w/o instances and skip CXX destructor.
3326 (possible_polymorphic_call_targets_1): Add bases_to_consider
3327 and consider_construction parameters; check if type may have instance.
3328 (get_polymorphic_call_info): Set maybe_in_construction to true
3329 when we know nothing.
3330 (record_targets_from_bases): Skip CXX destructors; they are
3331 never called for types in construction.
3332 (possible_polymorphic_call_targets): Do not record target when
3333 type may not have instance.
3335 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3338 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
3339 external aliases alive, too.
3341 2014-04-16 Andrew Pinski <apinski@cavium.com>
3343 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
3346 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3348 * final.c (compute_alignments): Do not apply loop alignment to a block
3349 falling through to the exit.
3351 2014-04-16 Catherine Moore <clm@codesourcery.com>
3353 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
3354 Adjust constraints for microMIPS store patterns.
3356 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3358 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
3360 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3362 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
3363 (append_use): Run at -O0.
3364 (append_vdef): Likewise.
3365 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
3366 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
3368 2014-04-16 Jakub Jelinek <jakub@redhat.com>
3370 PR tree-optimization/60844
3371 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
3372 (propagate_op_to_single_use, remove_visited_stmt_chain,
3373 linearize_expr, repropagate_negates, reassociate_bb): Use it
3374 instead of gsi_remove.
3376 2014-04-16 Martin Jambor <mjambor@suse.cz>
3378 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
3379 ipa_transforms_to_apply.
3380 (cgraph_function_versioning): Assert that old_node has empty
3381 ipa_transforms_to_apply.
3382 * trans-mem.c (ipa_tm_create_version): Likewise.
3383 * tree-inline.c (tree_function_versioning): Do not duplicate
3384 ipa_transforms_to_apply.
3386 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3389 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
3391 Pass necessary as flags on 64-bit Solaris/x86.
3392 Use lowercase relocs for x86_64-*-*.
3393 * configure: Regenerate.
3395 2014-04-15 Jan Hubicka <jh@suse.cz>
3397 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
3398 (maybe_record_node, likely_target_p): Use it.
3400 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3403 Revert following patch
3405 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3408 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3409 software floating point or no floating point registers, do not
3410 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3411 in GPRs that occurs after we tested for GPRs that would never be
3414 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3415 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3416 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3417 specifically allow DDmode, since that does not use the SPE SIMD
3420 2014-03-21 Mark Wielaard <mjw@redhat.com>
3422 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
3423 as unsigned or int depending on type and value used.
3425 2014-04-15 Richard Biener <rguenther@suse.de>
3427 PR rtl-optimization/56965
3428 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
3429 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
3431 * alias.c (true_dependence_1): Do not call
3432 nonoverlapping_component_refs_p.
3433 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
3434 nonoverlapping_component_refs_p.
3435 (indirect_refs_may_alias_p): Likewise.
3437 2014-04-15 Teresa Johnson <tejohnson@google.com>
3439 * cfg.c (dump_bb_info): Fix flags check.
3440 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
3442 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3444 PR rtl-optimization/60663
3445 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
3448 2014-04-15 Richard Biener <rguenther@suse.de>
3450 * lto-streamer.h (LTO_major_version): Bump to 4.
3452 2014-04-15 Richard Biener <rguenther@suse.de>
3454 * common.opt (lto_partition_model): New enum.
3455 (flto-partition=): Merge separate options with a single with argument,
3456 add -flto-partition=one support.
3457 * flag-types.h (enum lto_partition_model): Declare.
3458 * opts.c (finish_options): Remove duplicate -flto-partition=
3460 * lto-wrapper.c (run_gcc): Adjust.
3462 2014-04-15 Richard Biener <rguenther@suse.de>
3464 * alias.c (ncr_compar): New function.
3465 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
3467 2014-04-15 Richard Biener <rguenther@suse.de>
3469 * alias.c (record_component_aliases): Do not walk BINFOs.
3471 2014-04-15 Richard Biener <rguenther@suse.de>
3473 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
3474 Add struct function argument and adjust.
3475 (find_func_aliases_for_call): Likewise.
3476 (find_func_aliases): Likewise.
3477 (find_func_clobbers): Likewise.
3478 (intra_create_variable_infos): Likewise.
3479 (compute_points_to_sets): Likewise.
3480 (ipa_pta_execute): Adjust. Do not push/pop cfun.
3482 2014-04-15 Richard Biener <rguenther@suse.de>
3484 * tree.c (iterative_hash_expr): Use enum tree_code_class
3485 to store TREE_CODE_CLASS.
3486 (tree_block): Likewise.
3487 (tree_set_block): Likewise.
3488 * tree.h (fold_build_pointer_plus_loc): Use
3489 convert_to_ptrofftype_loc.
3491 2014-04-15 Jakub Jelinek <jakub@redhat.com>
3494 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
3497 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
3499 * cfgloop.h (struct loop): Move force_vectorize down.
3500 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
3501 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
3502 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
3503 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
3504 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
3505 * tree-core.h (enum annot_expr_kind): Add new kind values.
3506 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
3507 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
3509 * tree.def (ANNOTATE_EXPR): Tweak comment.
3511 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3513 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
3516 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
3518 * tree.h (TYPE_IDENTIFIER): Declare.
3519 * tree.c (subrange_type_for_debug_p): Use it.
3520 * godump.c (go_format_type): Likewise.
3521 * dwarf2out.c (is_cxx_auto, modified_type_die,
3522 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
3523 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
3525 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3528 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
3530 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
3532 * config/i386/i386.c (examine_argument): Return bool. Return true if
3533 parameter should be passed in memory.
3534 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
3535 (construct_container): Update calls to examine_argument.
3536 (function_arg_advance_64): Ditto.
3537 (return_in_memory_32): Merge with ix86_return_in_memory.
3538 (return_in_memory_64): Ditto.
3539 (return_in_memory_ms_64): Ditto.
3541 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3543 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
3544 * coverage.c (coverage_compute_profile_id): Handle externally visible
3547 2014-04-14 Martin Jambor <mjambor@suse.cz>
3549 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
3550 DECL_DISREGARD_INLINE_LIMITS functions.
3552 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3555 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
3557 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3560 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
3561 optimize_insn_for_speed_p instead of
3562 optimize_function_for_speed_p.
3564 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
3566 * doc/invoke.texi (free): Document AArch64.
3568 2014-04-14 Richard Biener <rguenther@suse.de>
3570 PR tree-optimization/60042
3571 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
3572 (insert_into_preds_of_block): Do not prevent PHI insertion
3573 for REFERENCE exprs here ...
3574 (eliminate_dom_walker::before_dom_children): ... but prevent
3575 their use here under similar conditions when applied to the
3576 IL after PRE optimizations.
3578 2014-04-14 Richard Biener <rguenther@suse.de>
3580 * passes.def: Move early points-to after early SRA.
3582 2014-04-14 Richard Biener <rguenther@suse.de>
3584 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
3585 check for which sign-changes we allow when forwarding
3586 a converted value into a switch.
3588 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3590 * stor-layout.c (place_field): Finalize non-constant offset for the
3593 2014-04-14 Richard Biener <rguenther@suse.de>
3595 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
3597 (expand_switch_using_bit_tests_p): Likewise.
3598 (process_switch): Compute and pass on speed_p based on the
3600 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
3601 optimize_bb_for_speed_p.
3603 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3605 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
3606 * function.h (struct function): Rename has_force_vect_loops into
3607 has_force_vectorize_loops.
3608 * lto-streamer-in.c (input_cfg): Adjust for renaming.
3609 (input_struct_function_base): Likewise.
3610 * lto-streamer-out.c (output_cfg): Likewise.
3611 (output_struct_function_base): Likewise.
3612 * omp-low.c (expand_omp_simd): Likewise.
3613 * tree-cfg.c (move_sese_region_to_fn): Likewise.
3614 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
3615 (version_loop_for_if_conversion): Likewise.
3616 (tree_if_conversion): Likewise.
3617 (main_tree_if_conversion): Likewise.
3618 (gate_tree_if_conversion): Likewise.
3619 * tree-inline.c (copy_loops): Likewise.
3620 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
3621 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
3622 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
3623 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
3624 * tree-vectorizer.c (vectorize_loops): Likewise.
3625 * tree-vectorizer.h (unlimited_cost_model): Likewise.
3627 2014-04-14 Richard Biener <rguenther@suse.de>
3630 * lto-streamer-out.c (wrap_refs): New function.
3631 (lto_output): Wrap symbol references in global initializes in
3632 type-preserving MEM_REFs.
3634 2014-04-14 Christian Bruel <christian.bruel@st.com>
3636 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
3638 2014-04-14 Christian Bruel <christian.bruel@st.com>
3640 * config/sh/sh.md (setmemqi): New expand pattern.
3641 * config/sh/sh.h (CLEAR_RATIO): Define.
3642 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
3643 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
3645 2014-04-14 Richard Biener <rguenther@suse.de>
3648 * fold-const.c (negate_expr_p): Don't negate directional rounding
3650 (fold_negate_expr): Likewise.
3652 2014-04-14 Richard Biener <rguenther@suse.de>
3654 PR tree-optimization/59817
3655 PR tree-optimization/60453
3656 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
3657 recursion to catch all CHRECs in the scalar evolution and restrict
3658 the predicate for the remains appropriately.
3660 2014-04-12 Catherine Moore <clm@codesourcery.com>
3662 * config/mips/constraints.md: Add new register constraint "kb".
3663 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
3664 (*movhi_internal): Likewise.
3665 (*movqi_internal): Likewise.
3666 * config/mips/mips.h (M16_STORE_REGS): New register class.
3667 (REG_CLASS_NAMES): Add M16_STORE_REGS.
3668 (REG_CLASS_CONTENTS): Likewise.
3669 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
3671 2014-04-11 Tobias Burnus <burnus@net-b.de>
3674 * doc/invoke.texi (-Wformat-signedness): Document it.
3675 (Wformat=2): Mention that this enables -Wformat-signedness.
3677 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3679 * common/config/epiphany/epiphany-common.c
3680 (epiphany_option_optimization_table): Enable section anchors by
3681 default at -O1 or higher.
3682 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
3683 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
3684 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
3685 carries no extra cost.
3686 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
3687 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
3688 * config/epiphany/predicates.md (memclob_operand): New predicate.
3689 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
3690 Use memclob_operand predicate and X constraint for operand 3.
3692 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3694 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
3695 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
3698 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3700 PR rtl-optimization/60651
3701 * mode-switching.c (optimize_mode_switching): Make sure to emit
3702 sets of a lower numbered entity before sets of a higher numbered
3703 entity to a mode of the same or lower priority.
3704 When creating a seginfo for a basic block that starts with a code
3705 label, move the insertion point past the code label.
3706 (new_seginfo): Document and enforce requirement that
3707 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
3708 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
3709 * doc/tm.texi: Regenerate.
3711 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
3714 * config/arc/arc.c (arc_save_restore): Fix assert typo.
3716 2013-04-11 Jakub Jelinek <jakub@redhat.com>
3718 * BASE-VER: Set to 4.10.0.
3720 2014-04-11 Tobias Burnus <burnus@net-b.de>
3723 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
3724 * doc/gcc.texi (Service): Update description in the @menu
3725 * doc/invoke.texi (Option Summary): Remove misplaced and
3728 2014-04-11 Steve Ellcey <sellcey@mips.com>
3729 Jakub Jelinek <jakub@redhat.com>
3732 * expr.c (convert_move): Use emit_store_flag_force instead of
3733 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
3736 2014-04-11 Richard Biener <rguenther@suse.de>
3739 * varasm.c (assemble_alias): Avoid endless error reporting
3740 recursion by setting TREE_ASM_WRITTEN.
3742 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3744 * config/s390/s390.md: Add a splitter for NOT rtx.
3746 2014-04-11 Jakub Jelinek <jakub@redhat.com>
3748 PR rtl-optimization/60663
3749 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
3751 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
3752 Jakub Jelinek <jakub@redhat.com>
3755 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
3756 flag from decl_node to node.
3758 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3761 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
3762 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
3763 ameliorating the cases where it can be.
3765 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
3768 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
3770 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3771 (loadsync_<mode>): Change mode.
3772 (load_quadpti, store_quadpti): New.
3773 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3774 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3775 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
3777 2014-04-09 Cong Hou <congh@google.com>
3780 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
3783 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3785 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
3786 instead of vnor to exploit possible fusion opportunity in the
3788 (altivec_expand_vec_perm_const_le): Likewise.
3790 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
3792 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3793 (loadsync_<mode>): Change mode.
3794 (load_quadpti, store_quadpti): New.
3795 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3796 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3798 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
3801 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
3802 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
3803 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
3805 2014-04-08 Richard Biener <rguenther@suse.de>
3808 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
3809 a 64bit widest int print double-int similar to on HWI64 hosts.
3811 2014-04-08 Richard Biener <rguenther@suse.de>
3813 PR tree-optimization/60785
3814 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
3815 default defs properly.
3817 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
3819 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
3820 (Weffc++): Likewise.
3822 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
3824 * ipa-devirt.c (maybe_record_node): When node is not recorded,
3825 set completep to false rather than true.
3827 2014-04-07 Douglas B Rupp <rupp@adacore.com>
3830 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
3831 ARM_TARGET2_DWARF_FORMAT.
3833 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
3836 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
3837 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
3840 2014-04-07 Richard Biener <rguenther@suse.de>
3842 PR tree-optimization/60766
3843 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
3844 (may_eliminate_iv): Convert cand_value_at result to desired type.
3846 2014-04-07 Jason Merrill <jason@redhat.com>
3849 * common.opt (-fno-gnu-unique): Add.
3850 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
3852 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3854 * haifa-sched.c: Fix outdated function reference and minor
3855 grammar errors in introductory comment.
3857 2014-04-07 Richard Biener <rguenther@suse.de>
3860 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
3862 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
3864 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
3867 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
3868 size accounting for thunks.
3869 (pa_asm_output_mi_thunk): Use final_start_function() and
3870 final_end_function() to output function start and end directives.
3872 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3874 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
3875 device specific ISA/ feature information. Remove short_sp and
3876 errata_skip ds. Add avr_device_specific_features enum to have device
3878 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
3879 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
3880 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
3881 updated device specific info.
3882 * config/avr/avr-mcus.def: Merge device specific details to
3883 dev_attribute field.
3884 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
3886 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
3887 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
3888 assembler if RMW isa supported by current device.
3889 * config/avr/genmultilib.awk: Update as device info structure changed.
3890 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
3892 2014-04-04 Cong Hou <congh@google.com>
3894 PR tree-optimization/60656
3895 * tree-vect-stmts.c (supportable_widening_operation):
3896 Fix a bug that elements in a vector with vect_used_by_reduction
3897 property are incorrectly reordered when the operation on it is not
3898 consistant with the one in reduction operation.
3900 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
3902 PR rtl-optimization/60155
3903 * gcse.c (record_set_data): New function.
3904 (single_set_gcse): New function.
3905 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
3906 (hoist_code): Likewise.
3907 (get_pressure_class_and_nregs): Likewise.
3909 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
3911 * explow.c (probe_stack_range): Emit a final optimization blockage.
3913 2014-04-04 Anthony Green <green@moxielogic.com>
3915 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
3918 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
3921 * lto-cgraph.c (input_overwrite_node): Check that partitioning
3922 flags are set only during streaming.
3923 * ipa.c (process_references, walk_polymorphic_call_targets,
3924 symtab_remove_unreachable_nodes): Drop bodies of always inline
3925 after early inlining.
3926 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
3928 2014-04-04 Jakub Jelinek <jakub@redhat.com>
3929 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3932 * dwarf2out.c (const_ok_for_output_1): Reject expressions
3935 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3938 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
3940 (cortex_a53_fdivd): Likewise.
3942 2014-04-04 Martin Jambor <mjambor@suse.cz>
3945 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
3947 * cgraph.c (clone_of_p): Also return true if thunks match.
3948 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
3949 cgraph_function_or_thunk_node and an obsolete comment.
3950 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
3952 (build_function_decl_skip_args): Likewise.
3953 (set_new_clone_decl_and_node_flags): New function.
3954 (duplicate_thunk_for_node): Likewise.
3955 (redirect_edge_duplicating_thunks): Likewise.
3956 (cgraph_clone_node): New parameter args_to_skip, pass it to
3957 redirect_edge_duplicating_thunks which is called instead of
3958 cgraph_redirect_edge_callee.
3959 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
3960 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
3962 2014-04-04 Jeff Law <law@redhat.com>
3965 * config/arm/predicates.md (const_int_I_operand): New predicate.
3966 (const_int_M_operand): Similarly.
3967 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
3969 (insv_t2, extv_reg, extzv_t2): Likewise.
3970 (load_multiple_with_writeback): Similarly for const_int_I_operand.
3971 (pop_multiple_with_writeback_and_return): Likewise.
3972 (vfp_pop_multiple_with_writeback): Likewise
3974 2014-04-04 Richard Biener <rguenther@suse.de>
3977 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
3978 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
3980 * gimplify.h (gimple_add_tmp_var_fn): Declare.
3981 * gimplify.c (gimple_add_tmp_var_fn): New function.
3982 * gimple-expr.h (create_tmp_reg_fn): Declare.
3983 * gimple-expr.c (create_tmp_reg_fn): New function.
3984 * gimple-low.c (record_vars_into): Don't change cfun.
3985 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
3986 code generation without cfun.
3988 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
3991 * Makefile.in (install-driver): Fix shell scripting.
3993 2014-04-03 Cong Hou <congh@google.com>
3995 PR tree-optimization/60505
3996 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
3997 threshold of number of iterations below which no vectorization
3999 * tree-vect-loop.c (new_loop_vec_info):
4000 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
4001 * tree-vect-loop.c (vect_analyze_loop_operations):
4002 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
4003 * tree-vect-loop.c (vect_transform_loop):
4004 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
4005 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
4006 of iterations of the loop and see if we should build the epilogue.
4008 2014-04-03 Richard Biener <rguenther@suse.de>
4010 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
4011 (streamer_tree_cache_create): Adjust.
4012 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
4013 to allow optional nodes array.
4014 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
4015 (streamer_tree_cache_append): Likewise.
4016 (streamer_tree_cache_create): Create nodes array optionally
4017 as specified by parameter.
4018 * lto-streamer-out.c (create_output_block): Avoid maintaining
4019 the node array in the writer cache.
4020 (DFS_write_tree): Remove assertion.
4021 (produce_asm_for_decls): Free the out decl state hash table early.
4022 * lto-streamer-in.c (lto_data_in_create): Adjust for
4023 streamer_tree_cache_create prototype change.
4025 2014-04-03 Richard Biener <rguenther@suse.de>
4027 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
4028 set TREE_CHAIN to NULL_TREE.
4030 2014-04-03 Richard Biener <rguenther@suse.de>
4032 PR tree-optimization/60740
4033 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
4034 over all GIMPLE_COND operands.
4036 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
4038 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
4039 (Weffc++): Remove Scott's numbering, merge lists and reference
4042 2014-04-03 Nick Clifton <nickc@redhat.com>
4044 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
4047 2014-04-03 Martin Jambor <mjambor@suse.cz>
4049 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
4050 mention gcc_unreachable before failing.
4051 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
4054 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
4057 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
4058 inconsistent code and instead mark the context inconsistent.
4059 (possible_polymorphic_call_targets): For inconsistent contexts
4060 return empty complete list.
4062 2014-04-02 Anthony Green <green@moxielogic.com>
4064 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
4065 (extendqisi2, extendhisi2): Define.
4066 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
4067 (WCHAR_TYPE): Change to unsigned int.
4069 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4071 PR tree-optimization/60733
4072 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
4073 insertion point for PHI candidates to be the end of the feeding
4074 block for the PHI argument.
4076 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
4078 PR rtl-optimization/60650
4079 * lra-constraints.c (process_alt_operands): Decrease reject for
4080 earlyclobber matching.
4082 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4084 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
4086 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4088 * config/spu/spu.c (pad_bb): Do not crash when the last
4089 insn is CODE_FOR_blockage.
4091 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4093 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
4094 lies outside the target mode.
4096 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4099 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4100 software floating point or no floating point registers, do not
4101 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4102 in GPRs that occurs after we tested for GPRs that would never be
4105 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4106 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4107 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4108 specifically allow DDmode, since that does not use the SPE SIMD
4111 2014-04-02 Richard Biener <rguenther@suse.de>
4114 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4115 MODE_INTs. Properly use negv_optab.
4116 (expand_abs): Likewise.
4118 2014-04-02 Richard Biener <rguenther@suse.de>
4121 * Makefile.in (install-driver): Guard extra installs with special
4124 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4126 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4129 2014-04-01 Richard Henderson <rth@redhat.com>
4132 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
4133 alternative enabled before register allocation.
4135 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
4137 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
4138 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
4140 (nios2_large_got_address): Remove unneeded 'sym' parameter.
4141 (nios2_got_address): Update nios2_large_got_address call site.
4142 (nios2_delegitimize_address): New function.
4143 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
4144 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
4145 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
4147 2014-04-01 Martin Husemann <martin@duskware.de>
4149 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
4152 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
4154 PR rtl-optimization/60604
4155 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
4156 check from register_operand.
4157 (register_operand): Redefine in terms of general_operand.
4158 (nonmemory_operand): Use register_operand for the non-constant cases.
4160 2014-04-01 Richard Biener <rguenther@suse.de>
4162 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
4164 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
4166 * doc/invoke.texi (mapp-regs): Clarify.
4168 2014-03-31 Ulrich Drepper <drepper@gmail.com>
4170 * config/i386/avx512fintrin.h (__v32hi): Define type.
4171 (__v64qi): Likewise.
4172 (_mm512_set1_epi8): Define.
4173 (_mm512_set1_epi16): Define.
4174 (_mm512_set4_epi32): Define.
4175 (_mm512_set4_epi64): Define.
4176 (_mm512_set4_pd): Define.
4177 (_mm512_set4_ps): Define.
4178 (_mm512_setr4_epi64): Define.
4179 (_mm512_setr4_epi32): Define.
4180 (_mm512_setr4_pd): Define.
4181 (_mm512_setr4_ps): Define.
4182 (_mm512_setzero_epi32): Define.
4184 2014-03-31 Martin Jambor <mjambor@suse.cz>
4187 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
4188 callsite_arguments_match_p. Updated all callers. Also check types of
4189 corresponding formal parameters and actual arguments.
4190 (not_all_callers_have_enough_arguments_p) Renamed to
4191 some_callers_have_mismatched_arguments_p.
4193 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
4195 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
4197 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
4200 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
4203 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
4205 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
4206 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
4208 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
4209 Use FMAMODE_NOVF512 mode iterator.
4210 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
4211 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
4212 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
4213 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
4215 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
4216 Use VF_128_256 mode iterator.
4217 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
4220 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4222 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
4223 static chain if needed.
4225 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4228 * lra-constraints.c (index_part_to_reg): New.
4229 (process_address): Use it.
4231 2014-03-27 Jeff Law <law@redhat.com>
4232 Jakub Jelinek <jakub@redhat.com>
4235 * expr.c (do_tablejump): Use simplify_gen_binary rather than
4236 gen_rtx_{PLUS,MULT} to build up the address expression.
4238 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
4239 creating non-canonical RTL.
4241 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4244 * ipa-inline.c (want_inline_small_function_p): Short circuit large
4245 functions; reorganize to make cheap checks first.
4246 (inline_small_functions): Do not estimate growth when dumping;
4248 * ipa-inline.h (inline_summary): Add min_size.
4249 (growth_likely_positive): New function.
4250 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
4251 (set_cond_stmt_execution_predicate): Cleanup.
4252 (estimate_edge_size_and_time): Compute min_size.
4253 (estimate_calls_size_and_time): Likewise.
4254 (estimate_node_size_and_time): Likewise.
4255 (inline_update_overall_summary): Update min_size.
4256 (do_estimate_edge_time): Likewise.
4257 (do_estimate_edge_size): Update.
4258 (do_estimate_edge_hints): Update.
4259 (growth_likely_positive): New function.
4261 2014-03-28 Jakub Jelinek <jakub@redhat.com>
4264 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
4265 also if addr has VOIDmode.
4267 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4269 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
4270 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
4272 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
4273 instructions as well as AdvancedSIMD loads.
4275 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4277 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
4278 Use crypto_aese type.
4279 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
4280 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
4281 crypto_aese, crypto_aesmc. Move to types.md.
4282 * config/arm/types.md (crypto_aes): Split into crypto_aese,
4284 * config/arm/iterators.md (crypto_type): Likewise.
4286 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4288 * cgraph.c: Include expr.h and tree-dfa.h.
4289 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
4292 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4295 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
4296 regs from checking multi-reg pseudos.
4298 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4300 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
4302 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4304 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
4305 if it would clobber the stack pointer, even temporarily.
4307 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
4309 * mode-switching.c: Make small adjustments to the top comment.
4311 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4313 * config/rs6000/constraints.md (wD constraint): New constraint to
4314 match the constant integer to get the top DImode/DFmode out of a
4315 vector in a VSX register.
4317 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
4318 match the constant integer to get the top DImode/DFmode out of a
4319 vector in a VSX register.
4321 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
4324 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
4327 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
4328 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
4330 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
4331 Optimize vec_extract of 64-bit values, where the value being
4332 extracted is in the top word, where we can use scalar
4333 instructions. Add direct move and store support. Combine the big
4334 endian/little endian vector select load support into a single insn.
4335 (vsx_extract_<mode>_internal1): Likewise.
4336 (vsx_extract_<mode>_internal2): Likewise.
4337 (vsx_extract_<mode>_load): Likewise.
4338 (vsx_extract_<mode>_store): Likewise.
4339 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
4340 combined into vsx_extract_<mode>_load.
4341 (vsx_extract_<mode>_one_le): Likewise.
4343 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
4344 define the top 64-bit vector element.
4346 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
4349 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4350 Document vec_vbpermq builtin.
4353 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
4354 enable use of xxsldwi and xxpermdi builtin functions.
4355 (vec_xxpermdi): Likewise.
4357 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4358 Document use of vec_xxsldwi and vec_xxpermdi builtins.
4360 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
4362 PR rtl-optimization/60650
4363 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
4365 (find_spills_for): New.
4366 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
4367 Spill all pseudos on the second iteration.
4369 2014-03-27 Marek Polacek <polacek@redhat.com>
4372 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
4375 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4377 * config/s390/s390.c (s390_can_use_return_insn): Check for
4378 call-saved FPRs on 31 bit.
4380 2014-03-27 Jakub Jelinek <jakub@redhat.com>
4383 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
4384 if they need regimplification, just drop them instead of
4385 calling gimple_regimplify_operands on them.
4387 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
4390 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
4391 (aarch64_frame_pointer_required): Adjust logic.
4392 (aarch64_can_eliminate): Adjust logic.
4393 (aarch64_override_options_after_change): Adjust logic.
4395 2014-03-27 Dehao Chen <dehao@google.com>
4397 * ipa-inline.c (early_inliner): Update node's inline info.
4399 2014-03-26 Dehao Chen <dehao@google.com>
4401 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
4402 compiler inserted conditional jumps for NAN float check.
4404 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4406 * ubsan.h (ubsan_create_data): Change second argument's type
4407 to const location_t *.
4408 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
4410 (ubsan_create_data): Change second argument to const location_t *PLOC.
4411 Create Loc field whenever PLOC is non-NULL.
4412 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
4413 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
4417 * real.c (real_to_integer2): Change type of low to UHWI.
4419 2014-03-26 Tobias Burnus <burnus@net-b.de>
4421 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
4422 (CILK_SELF_SPECS): New define.
4423 (driver_self_specs): Use it.
4425 2014-03-26 Richard Biener <rguenther@suse.de>
4427 * tree-pretty-print.c (percent_K_format): Implement special
4428 case for LTO and its stripped down BLOCK tree.
4430 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4433 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
4435 * tree-vrp.c (simplify_internal_call_using_ranges): If only
4436 one range is range_int_cst_p, but not both, at least optimize
4437 addition/subtraction of 0 and multiplication by 0 or 1.
4438 * gimple-fold.c (gimple_fold_call): Fold
4439 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
4440 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
4441 INTEGER_CSTs, try to fold at least x * 0 and y - y.
4443 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
4445 PR rtl-optimization/60452
4446 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
4447 <case REG>: Return 1 for invalid offsets from the frame pointer.
4449 2014-03-26 Marek Polacek <polacek@redhat.com>
4452 * doc/extend.texi (C Extensions): Mention variable-length arrays in
4455 2014-03-26 Marek Polacek <polacek@redhat.com>
4458 * doc/extend.texi (Designated Inits): Describe what happens to omitted
4461 2014-03-26 Marek Polacek <polacek@redhat.com>
4464 * ira-color.c (update_conflict_hard_regno_costs): Perform the
4465 multiplication in unsigned type.
4467 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4469 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
4471 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4473 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
4475 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4478 * cif-code.def (UNREACHABLE) New code.
4479 * ipa-inline.c (inline_small_functions): Skip edges to
4480 __builtlin_unreachable.
4481 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
4482 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
4483 predicate to __bulitin_unreachable.
4484 (set_cond_stmt_execution_predicate): Fix issue when
4485 invert_tree_comparison returns ERROR_MARK.
4486 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
4487 propagate to inline clones.
4488 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
4490 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
4491 * cgraphclones.c (cgraph_clone_node): If call destination is already
4492 ureachable, do not redirect it back.
4493 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
4496 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4498 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
4499 Do not modify inline clones.
4501 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4503 * config/i386/i386.md (general_sext_operand): New mode attr.
4504 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
4505 don't generate (sign_extend (const_int)).
4506 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
4507 operands[2]. Use We constraint instead of <i> and
4508 <general_sext_operand> predicate instead of <general_operand>.
4509 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
4510 * config/i386/constraints.md (We): New constraint.
4511 * config/i386/predicates.md (x86_64_sext_operand,
4512 sext_operand): New predicates.
4514 2014-03-25 Martin Jambor <mjambor@suse.cz>
4517 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
4518 inconsistent devirtualizations to __builtin_unreachable.
4520 2014-03-25 Marek Polacek <polacek@redhat.com>
4523 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
4525 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
4527 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
4528 order of elements for big-endian.
4530 2014-03-25 Richard Biener <rguenther@suse.de>
4533 * gimplify-me.c (gimple_regimplify_operands): Update the
4536 2014-03-25 Martin Jambor <mjambor@suse.cz>
4539 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
4540 (lto_output_varpool_node): Likewise.
4541 (input_overwrite_node): Likewise.
4542 (input_varpool_node): Likewise.
4544 2014-03-25 Richard Biener <rguenther@suse.de>
4546 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
4547 (run_gcc): Likewise.
4549 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4551 * combine.c (simplify_compare_const): Add MODE argument.
4552 Handle mode_width 0 as very large mode_width.
4553 (try_combine, simplify_comparison): Adjust callers.
4555 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
4556 type to avoid signed integer overflow.
4557 * explow.c (plus_constant): Likewise.
4559 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
4561 * doc/generic.texi: Correct typos.
4563 2014-03-24 Tobias Burnus <burnus@net-b.de>
4565 * doc/invoke.texi (-flto): Expand section about
4566 using static libraries with LTO.
4568 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4570 PR rtl-optimization/60501
4571 * optabs.def (addptr3_optab): New optab.
4572 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
4573 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
4574 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
4576 * lra.c (emit_add3_insn): Use the addptr pattern if available.
4578 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
4580 2014-03-24 Ulrich Drepper <drepper@gmail.com>
4582 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
4585 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
4586 (_mm256_undefined_ps): Define.
4587 (_mm256_undefined_pd): Define.
4588 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
4589 (_mm_undefined_pd): Define.
4590 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
4591 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
4592 (_mm512_undefined_ps): Define.
4593 (_mm512_undefined_pd): Define.
4594 Use _mm*_undefined_*.
4595 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
4597 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
4599 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
4600 (lshr_simd): DI mode added.
4601 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
4602 (aarch64_ushr_simddi): Likewise.
4603 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
4604 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
4605 (vshrd_n_u64): Likewise.
4607 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4609 * Makefile.in (s-macro_list): Depend on cc1.
4611 2014-03-23 Teresa Johnson <tejohnson@google.com>
4613 * ipa-utils.c (ipa_print_order): Use specified dump file.
4615 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
4617 PR rtl-optimization/60601
4618 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
4620 * gcc.c (eval_spec_function): Initialize save_growing_value.
4622 2014-03-22 Jakub Jelinek <jakub@redhat.com>
4625 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
4626 code == MINUS_EXPR, never swap op0 with op1.
4628 * toplev.c (init_local_tick): Avoid signed integer multiplication
4630 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
4631 shift by first operand's bitsize.
4633 2014-03-21 Jakub Jelinek <jakub@redhat.com>
4636 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
4638 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
4639 TARGET_ISA_64BIT_P(x).
4641 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4643 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
4644 pattern for vector nor instead of subtract from splat(-1).
4645 (altivec_expand_vec_perm_const_le): Likewise.
4647 2014-03-21 Richard Henderson <rth@twiddle.net>
4650 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
4651 related insns after epilogue_completed.
4653 2014-03-21 Martin Jambor <mjambor@suse.cz>
4656 * cgraph.h (symtab_node): New flag body_removed.
4657 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
4658 when removing bodies.
4659 * symtab.c (dump_symtab_base): Dump body_removed flag.
4660 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
4661 had their bodies removed.
4663 2014-03-21 Martin Jambor <mjambor@suse.cz>
4666 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
4669 2014-03-21 Richard Biener <rguenther@suse.de>
4671 PR tree-optimization/60577
4672 * tree-core.h (struct tree_base): Document nothrow_flag use
4673 in VAR_DECL_NONALIASED.
4674 * tree.h (VAR_DECL_NONALIASED): New.
4675 (may_be_aliased): Adjust.
4676 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
4678 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4680 * expr.c (expand_expr_real_1): Remove outdated comment.
4682 2014-03-20 Jakub Jelinek <jakub@redhat.com>
4685 * ira.c (adjust_cleared_regs): Call copy_rtx on
4686 *reg_equiv[REGNO (loc)].src_p before passing it to
4687 simplify_replace_fn_rtx.
4690 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
4691 into CONST, put pic register as first operand of PLUS. Use
4692 gen_const_mem for both 32-bit and 64-bit PIC got loads.
4694 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4696 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
4698 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4700 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
4701 around for store forwarding issue in the FPU on the UT699.
4702 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
4703 loads and operations if -mfix-ut699 is specified.
4704 (divtf3_hq): Tweak attribute.
4705 (sqrttf2_hq): Likewise.
4707 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4709 * calls.c (store_one_arg): Remove incorrect const qualification on the
4710 type of the temporary.
4711 * cfgexpand.c (expand_return): Likewise.
4712 * expr.c (expand_constructor): Likewise.
4713 (expand_expr_real_1): Likewise.
4715 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4717 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
4720 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
4723 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
4725 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
4727 * config/arm/aarch-common-protos.h
4728 (alu_cost_table): Fix spelling of "extend".
4729 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
4731 2014-03-19 Richard Biener <rguenther@suse.de>
4734 * tree-core.h (tree_type_common): Re-order pointer members
4735 to reduce recursion depth during GC walks.
4737 2014-03-19 Marek Polacek <polacek@redhat.com>
4740 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
4741 before accessing it.
4743 2014-03-19 Richard Biener <rguenther@suse.de>
4746 * lto-streamer-in.c (input_function): In WPA stage do not drop
4749 2014-03-19 Jakub Jelinek <jakub@redhat.com>
4751 PR tree-optimization/60559
4752 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
4753 with build_zero_cst assignment.
4755 2014-03-18 Kai Tietz <ktietz@redhat.com>
4757 PR rtl-optimization/56356
4758 * sdbout.c (sdbout_parms): Verify that parms'
4759 incoming argument is valid.
4760 (sdbout_reg_parms): Likewise.
4762 2014-03-18 Richard Henderson <rth@redhat.com>
4765 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
4766 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
4767 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
4769 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
4771 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
4772 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
4773 Italicize plugin event names in description. Explain that
4774 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
4775 Remind that no GCC functions should be called after PLUGIN_FINISH.
4776 Explain what pragmas with expansion are.
4778 2014-03-18 Martin Liska <mliska@suse.cz>
4780 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
4781 gimple call statement is update.
4782 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
4783 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
4785 2014-03-18 Jakub Jelinek <jakub@redhat.com>
4788 * ubsan.c (ubsan_instrument_unreachable): Call
4789 initialize_sanitizer_builtins.
4790 (ubsan_pass): Likewise.
4793 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
4794 varpool_finalize_decl instead of rest_of_decl_compilation.
4796 2014-03-18 Richard Biener <rguenther@suse.de>
4798 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
4799 by using bitmap_and_compl instead of bitmap_and_compl_into.
4800 (df_rd_transfer_function): Likewise.
4802 2014-03-18 Richard Biener <rguenther@suse.de>
4804 * doc/lto.texi (fresolution): Fix typo.
4806 2014-03-18 Richard Biener <rguenther@suse.de>
4808 * doc/invoke.texi (flto): Update for changes in 4.9.
4810 2014-03-18 Richard Biener <rguenther@suse.de>
4812 * doc/loop.texi: Remove section on the removed lambda framework.
4813 Update loop docs with recent changes in preserving loop structure.
4815 2014-03-18 Richard Biener <rguenther@suse.de>
4817 * doc/lto.texi (-fresolution): Document.
4819 2014-03-18 Richard Biener <rguenther@suse.de>
4821 * doc/contrib.texi: Adjust my name.
4823 2014-03-18 Jakub Jelinek <jakub@redhat.com>
4826 * internal-fn.c: Include diagnostic-core.h.
4827 (expand_BUILTIN_EXPECT): New function.
4828 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
4829 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
4830 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
4831 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
4833 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
4834 Revert 3 argument __builtin_expect code.
4835 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
4836 * gimple-fold.c (gimple_fold_call): Likewise.
4837 * tree.h (fold_builtin_expect): New prototype.
4838 * builtins.c (build_builtin_expect_predicate): Add predictor
4839 argument, if non-NULL, create 3 argument __builtin_expect.
4840 (fold_builtin_expect): No longer static. Add ARG2 argument,
4841 pass it through to build_builtin_expect_predicate.
4842 (fold_builtin_2): Adjust caller.
4843 (fold_builtin_3): Handle BUILT_IN_EXPECT.
4844 * internal-fn.def (BUILTIN_EXPECT): New.
4846 2014-03-18 Tobias Burnus <burnus@net-b.de>
4849 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
4850 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
4851 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
4853 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
4856 * predict.c (combine_predictions_for_bb): Fix up formatting.
4857 (expr_expected_value_1, expr_expected_value): Add predictor argument,
4858 fill what it points to if non-NULL.
4859 (tree_predict_by_opcode): Adjust caller, use the predictor.
4860 * predict.def (PRED_COMPARE_AND_SWAP): Add.
4862 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
4864 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
4865 proper constant for the store mode.
4867 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
4869 * symtab.c (change_decl_assembler_name): Fix transparent alias
4872 2014-03-16 Renlin Li <Renlin.Li@arm.com>
4874 * config/aarch64/aarch64.c: Correct the comments about the
4875 aarch64 stack layout.
4877 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
4879 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
4880 check for GF_OMP_FOR_KIND_FOR.
4882 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
4884 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
4885 ymm and zmm register names.
4887 2014-03-17 Jakub Jelinek <jakub@redhat.com>
4890 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
4891 note creation for the 2010-08-31 changes.
4893 2014-03-17 Marek Polacek <polacek@redhat.com>
4896 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
4897 as -fno-tree-loop-vectorize.
4898 (expand_omp_simd): Likewise.
4900 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
4902 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
4903 (eligible_for_call_delay): New prototype.
4904 * config/sparc/sparc.c (tls_call_delay): Rename into...
4905 (eligible_for_call_delay): ...this. Return false if the instruction
4906 cannot be put in the delay slot of a branch.
4907 (eligible_for_restore_insn): Simplify.
4908 (eligible_for_return_delay): Return false if the instruction cannot be
4909 put in the delay slot of a branch and simplify.
4910 (eligible_for_sibcall_delay): Return false if the instruction cannot be
4911 put in the delay slot of a branch.
4912 * config/sparc/sparc.md (fix_ut699): New attribute.
4913 (tls_call_delay): Delete.
4914 (in_call_delay): Reimplement.
4915 (eligible_for_sibcall_delay): Rename into...
4916 (in_sibcall_delay): ...this.
4917 (eligible_for_return_delay): Rename into...
4918 (in_return_delay): ...this.
4919 (in_branch_delay): Reimplement.
4920 (in_uncond_branch_delay): Delete.
4921 (in_annul_branch_delay): Delete.
4923 2014-03-14 Richard Henderson <rth@redhat.com>
4926 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
4927 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
4928 (*floathi<X87MODEF>2_i387_with_temp): Remove.
4929 (floathi splitters): Remove.
4930 (float<SWI48x>xf2): New pattern.
4931 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
4932 code that tried to handle DImode for 32-bit, but which was excluded
4933 by the pattern's condition. Drop allocation of stack temporary.
4934 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
4935 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
4936 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
4937 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
4938 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
4939 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
4940 (*float<SWI48><MODEF>2_sse_interunit): Remove.
4941 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
4942 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
4943 (*float<SWI48x><X87MODEF>2_i387): Remove.
4944 (all float _with_temp splitters): Remove.
4945 (*float<SWI48x><MODEF>2_i387): New pattern.
4946 (*float<SWI48><MODEF>2_sse): New pattern.
4947 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
4948 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
4950 2014-03-14 Jakub Jelinek <jakub@redhat.com>
4951 Marek Polacek <polacek@redhat.com>
4954 * common.opt (dump_base_name_prefixed): New Variable.
4955 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
4956 if x_dump_base_name_prefixed is already set, set it at the end.
4958 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
4960 PR rtl-optimization/60508
4961 * lra-constraints.c (get_reload_reg): Add new parameter
4963 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
4964 Pass the new parameter values.
4966 2014-03-14 Richard Biener <rguenther@suse.de>
4968 * common.opt: Revert unintented changes from r205065.
4971 2014-03-14 Richard Biener <rguenther@suse.de>
4974 * cfghooks.c (split_block): Properly adjust all loops the
4975 block was a latch of.
4977 2014-03-14 Martin Jambor <mjambor@suse.cz>
4980 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
4983 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
4986 * config/avr/avr.c (avr_set_current_function): Pass function name
4987 through default_strip_name_encoding before sanity checking instead
4988 of skipping the first char of the assembler name.
4990 2014-03-13 Richard Henderson <rth@redhat.com>
4993 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
4994 (ix86_force_to_memory, ix86_free_from_memory): Remove.
4995 * config/i386/i386-protos.h: Likewise.
4996 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
4997 in the expander instead of a splitter.
4998 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
4999 any possibility of requiring a memory.
5000 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
5001 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
5002 (fp branch splitters): Update for ix86_split_fp_branch.
5003 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
5004 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
5005 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
5006 (*fop_<MODEF>_2_i387): Remove f/r alternative.
5007 (*fop_<MODEF>_3_i387): Likewise.
5008 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
5009 (splitters for the fop_* register patterns): Remove.
5010 (fscalexf4_i387): Rename from *fscalexf4_i387.
5011 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
5013 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5015 PR tree-optimization/59779
5016 * tree-dfa.c (get_ref_base_and_extent): Use double_int
5017 type for bitsize and maxsize instead of HOST_WIDE_INT.
5019 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
5021 PR rtl-optimization/57320
5022 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
5023 the CFG after thread_prologue_and_epilogue_insns.
5025 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
5027 PR rtl-optimization/57189
5028 * lra-constraints.c (process_alt_operands): Disfavor spilling
5031 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
5033 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
5035 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5037 PR tree-optimization/59025
5039 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
5040 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
5042 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
5045 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
5046 calls of avr_out_plus_1.
5048 2014-03-13 Bin Cheng <bin.cheng@arm.com>
5050 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
5051 BB's single pred and update the father loop's latch info later.
5053 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5055 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
5059 (VEC_base): Likewise.
5060 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
5061 registers, we need to swap double words in little endian mode.
5063 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
5064 to be a container mode for 128-bit integer operations added in ISA
5065 2.07. Unlike TImode and PTImode, the preferred register set is
5066 the Altivec/VMX registers for the 128-bit operations.
5068 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
5070 (rs6000_split_128bit_ok_p): Likewise.
5072 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
5073 macros for creating ISA 2.07 normal and overloaded builtin
5074 functions with 3 arguments.
5075 (BU_P8V_OVERLOAD_3): Likewise.
5076 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
5077 for use as overloaded functions.
5078 (VPERM_1TI_UNS): Likewise.
5079 (VSEL_1TI): Likewise.
5080 (VSEL_1TI_UNS): Likewise.
5081 (ST_INTERNAL_1ti): Likewise.
5082 (LD_INTERNAL_1ti): Likewise.
5083 (XXSEL_1TI): Likewise.
5084 (XXSEL_1TI_UNS): Likewise.
5085 (VPERM_1TI): Likewise.
5086 (VPERM_1TI_UNS): Likewise.
5087 (XXPERMDI_1TI): Likewise.
5088 (SET_1TI): Likewise.
5089 (LXVD2X_V1TI): Likewise.
5090 (STXVD2X_V1TI): Likewise.
5091 (VEC_INIT_V1TI): Likewise.
5092 (VEC_SET_V1TI): Likewise.
5093 (VEC_EXT_V1TI): Likewise.
5094 (EQV_V1TI): Likewise.
5095 (NAND_V1TI): Likewise.
5096 (ORC_V1TI): Likewise.
5097 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
5098 added in ISA 2.07. Add both normal 'altivec' builtins, and the
5100 (VADDUQM): Likewise.
5101 (VSUBCUQ): Likewise.
5102 (VADDEUQM): Likewise.
5103 (VADDECUQ): Likewise.
5104 (VSUBEUQM): Likewise.
5105 (VSUBECUQ): Likewise.
5107 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5108 __int128_t and __uint128_t types.
5109 (__uint128_type): Likewise.
5110 (altivec_categorize_keyword): Add support for vector __int128_t,
5111 vector __uint128_t, vector __int128, and vector unsigned __int128
5112 as a container type for TImode operations that need to be done in
5113 VSX/Altivec registers.
5114 (rs6000_macro_to_expand): Likewise.
5115 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5116 to support 128-bit integer instructions vaddcuq, vadduqm,
5117 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5118 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5120 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5121 for V1TImode, and set up preferences to use VSX/Altivec registers.
5122 Setup VSX reload handlers.
5123 (rs6000_debug_reg_global): Likewise.
5124 (rs6000_init_hard_regno_mode_ok): Likewise.
5125 (rs6000_preferred_simd_mode): Likewise.
5126 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5127 (easy_altivec_constant): Likewise.
5128 (output_vec_const_move): Likewise.
5129 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5131 (rs6000_expand_vector_extract): Likewise.
5132 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
5134 (rs6000_const_vec): Add support for V1TImode.
5135 (rs6000_emit_le_vsx_load): Swap double words when loading or
5136 storing TImode/V1TImode.
5137 (rs6000_emit_le_vsx_store): Likewise.
5138 (rs6000_emit_le_vsx_move): Likewise.
5139 (rs6000_emit_move): Add support for V1TImode.
5140 (altivec_expand_ld_builtin): Likewise.
5141 (altivec_expand_st_builtin): Likewise.
5142 (altivec_expand_vec_init_builtin): Likewise.
5143 (altivec_expand_builtin): Likewise.
5144 (rs6000_init_builtins): Add support for V1TImode type. Add
5145 support for ISA 2.07 128-bit integer builtins. Define type names
5146 for the VSX/Altivec vector types.
5147 (altivec_init_builtins): Add support for overloaded vector
5148 functions with V1TImode type.
5149 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
5150 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
5152 (rs6000_split_128bit_ok_p): Likewise.
5153 (rs6000_handle_altivec_attribute): Create V1TImode from vector
5154 __int128_t and vector __uint128_t.
5156 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
5157 and mode attributes.
5164 (VS_scalar): Likewise.
5165 (VS_double): Likewise.
5166 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
5168 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
5169 we support the ISA 2.07 128-bit integer arithmetic instructions.
5170 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
5171 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
5172 and TImode types for use with the builtin functions.
5173 (V1TI_type_node): Likewise.
5174 (unsigned_V1TI_type_node): Likewise.
5175 (intTI_type_internal_node): Likewise.
5176 (uintTI_type_internal_node): Likewise.
5178 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
5179 128-bit builtin functions.
5180 (UNSPEC_VADDEUQM): Likewise.
5181 (UNSPEC_VADDECUQ): Likewise.
5182 (UNSPEC_VSUBCUQ): Likewise.
5183 (UNSPEC_VSUBEUQM): Likewise.
5184 (UNSPEC_VSUBECUQ): Likewise.
5185 (VM): Add V1TImode to vector mode iterators.
5187 (VI_unit): Likewise.
5188 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
5189 (altivec_vaddcuq): Likewise.
5190 (altivec_vsubuqm): Likewise.
5191 (altivec_vsubcuq): Likewise.
5192 (altivec_vaddeuqm): Likewise.
5193 (altivec_vaddecuq): Likewise.
5194 (altivec_vsubeuqm): Likewise.
5195 (altivec_vsubecuq): Likewise.
5197 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
5199 (BOOL_128): Likewise.
5200 (BOOL_REGS_OUTPUT): Likewise.
5201 (BOOL_REGS_OP1): Likewise.
5202 (BOOL_REGS_OP2): Likewise.
5203 (BOOL_REGS_UNARY): Likewise.
5204 (BOOL_REGS_AND_CR0): Likewise.
5206 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
5207 128-bit integer builtin support.
5208 (vec_vadduqm): Likewise.
5209 (vec_vaddecuq): Likewise.
5210 (vec_vaddeuqm): Likewise.
5211 (vec_vsubecuq): Likewise.
5212 (vec_vsubeuqm): Likewise.
5213 (vec_vsubcuq): Likewise.
5214 (vec_vsubuqm): Likewise.
5216 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5217 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
5218 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
5219 128-bit integer add/subtract to ISA 2.07.
5221 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
5223 * config/arc/arc.c (arc_predicate_delay_insns):
5224 Fix third argument passed to conditionalize_nonjump.
5226 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
5228 * config/aarch64/aarch64-builtins.c
5229 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
5230 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
5231 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
5232 instead of __builtin_lfloor.
5233 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
5235 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5237 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
5238 (tree_ssa_ifcombine_bb_1): New function.
5239 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
5240 is an empty forwarder block to then_bb or vice versa and then_bb
5241 and else_bb are effectively swapped.
5243 2014-03-12 Christian Bruel <christian.bruel@st.com>
5246 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
5247 REG_CFA_DEF_CFA note.
5248 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
5249 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
5251 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
5253 PR tree-optimization/60454
5254 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
5256 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5258 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
5259 Do not define target_cpu_default2 to generic.
5260 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
5261 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
5262 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
5264 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5265 Marc Glisse <marc.glisse@inria.fr>
5267 PR tree-optimization/60502
5268 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
5269 instead of build_low_bits_mask.
5271 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5274 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
5275 if there are multiple uses, but op doesn't live on E edge.
5276 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
5277 clobber stmts before __builtin_unreachable.
5279 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
5281 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
5282 hard_frame_pointer_rtx.
5283 * cse.c (cse_insn): Remove volatile check.
5284 * cselib.c (cselib_process_insn): Likewise.
5285 * dse.c (scan_insn): Likewise.
5287 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5289 * config/arc/arc.c (conditionalize_nonjump): New function,
5291 (arc_ifcvt): ... this.
5292 (arc_predicate_delay_insns): Use it.
5294 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5296 * config/arc/predicates.md (extend_operand): During/after reload,
5297 allow const_int_operand.
5298 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
5299 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
5300 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
5302 (umulsi3_highpart_i): Likewise.
5304 2014-03-11 Richard Biener <rguenther@suse.de>
5306 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
5307 Add asserts to guard possible wrong-code bugs.
5309 2014-03-11 Richard Biener <rguenther@suse.de>
5311 PR tree-optimization/60429
5312 PR tree-optimization/60485
5313 * tree-ssa-structalias.c (set_union_with_increment): Properly
5314 take into account all fields that overlap the shifted vars.
5315 (do_sd_constraint): Likewise.
5316 (do_ds_constraint): Likewise.
5317 (get_constraint_for_ptr_offset): Likewise.
5319 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
5321 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
5322 (nios2_compute_frame_layout):
5323 Add calculation of cfun->machine->fp_save_offset.
5324 (nios2_expand_prologue): Correct setting of frame pointer register
5326 (nios2_expand_epilogue): Update recovery of stack pointer from
5327 frame pointer accordingly.
5328 (nios2_initial_elimination_offset): Update calculation of offset
5329 for eliminating to HARD_FRAME_POINTER_REGNUM.
5331 2014-03-10 Jakub Jelinek <jakub@redhat.com>
5334 * ipa.c (symtab_remove_unreachable_nodes): Don't call
5335 cgraph_get_create_node on VAR_DECLs.
5337 2014-03-10 Richard Biener <rguenther@suse.de>
5340 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
5342 2014-03-08 Douglas B Rupp <rupp@gnat.com>
5344 * config/vms/vms.opt (vms_float_format): New variable.
5346 2014-03-08 Tobias Burnus <burnus@net-b.de>
5348 * doc/invoke.texi (-fcilkplus): Update implementation status.
5350 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
5351 Richard Biener <rguenther@suse.de>
5353 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
5354 consistently accross all TUs.
5355 (run_gcc): Enable -fshort-double automatically at link at link-time
5356 and disallow override.
5358 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
5361 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
5362 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
5363 if they can't be used.
5365 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5367 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
5368 for Solaris 11/x86 ld.
5369 * configure: Regenerate.
5371 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5373 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
5374 (LIB_TLS_SPEC): Save as ld_tls_libs.
5375 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
5376 (HAVE_AS_IX86_TLSLDM): New test.
5377 * configure, config.in: Regenerate.
5378 * config/i386/i386.c (legitimize_tls_address): Fall back to
5379 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
5380 cannot support TLS_MODEL_LOCAL_DYNAMIC.
5381 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
5382 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
5384 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
5386 * common.opt (fira-loop-pressure): Mark as optimization.
5388 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
5390 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
5391 an OpenMP mappable type.
5393 2014-03-06 Matthias Klose <doko@ubuntu.com>
5395 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
5396 MULTILIB_OSDIRNAMES is not defined.
5398 2014-03-06 Jakub Jelinek <jakub@redhat.com>
5399 Meador Inge <meadori@codesourcery.com>
5402 * config/arm/arm.c (arm_tls_symbol_p): Remove.
5403 (arm_legitimize_address): Call legitimize_tls_address for any
5404 arm_tls_referenced_p expression, handle constant addend. Call it
5405 before testing for !TARGET_ARM.
5406 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
5408 2014-03-06 Richard Biener <rguenther@suse.de>
5414 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5416 * tree-streamer.c (record_common_node): Assert we don't record
5417 nodes with type double.
5418 (preload_common_node): Skip type double, complex double and double
5419 pointer since it is now frontend dependent due to fshort-double option.
5421 2014-03-06 Richard Biener <rguenther@suse.de>
5423 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
5424 or -fno-lto is specified and the linker has full plugin support.
5425 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
5426 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
5427 * lto-wrapper.c (merge_and_complain): Merge compile-time
5428 optimization levels.
5429 (run_gcc): And pass it through to the link options.
5431 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
5435 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5437 * cselib.c (remove_useless_values): Skip to avoid quadratic
5438 behavior if the condition moved from...
5439 (cselib_process_insn): ... here holds.
5441 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5444 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
5445 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
5448 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
5449 (TM_H): Add x86-tune.def.
5451 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5453 * config/aarch64/aarch64.c (generic_tunings):
5454 Use cortexa57_extra_costs.
5456 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5459 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
5460 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
5461 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
5464 2014-03-04 Heiher <r@hev.cc>
5466 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
5467 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
5469 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
5471 * config/i386/predicates.md (const2356_operand): Change to ...
5472 (const2367_operand): ... this.
5473 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
5475 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5476 (*avx512pf_scatterpf<mode>sf): Ditto.
5477 (avx512pf_scatterpf<mode>df): Ditto.
5478 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5479 (*avx512pf_scatterpf<mode>df): Ditto.
5480 * config/i386/i386.c (ix86_expand_builtin): Update
5481 incorrect hint operand error message.
5483 2014-03-04 Richard Biener <rguenther@suse.de>
5485 * lto-section-in.c (lto_get_section_data): Fix const cast.
5487 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5489 * tree-streamer.c (record_common_node): Assert we don't record
5490 nodes with type double.
5491 (preload_common_node): Skip type double, complex double and double
5492 pointer since it is now frontend dependent due to fshort-double option.
5494 2014-03-04 Richard Biener <rguenther@suse.de>
5497 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
5498 (lto_input_toplevel_asms): Likewise.
5499 * lto-section-in.c (lto_get_section_data): Instead do it here
5502 2014-03-04 Richard Biener <rguenther@suse.de>
5504 PR tree-optimization/60382
5505 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
5506 dead PHIs a reduction.
5508 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
5510 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
5512 (_mm_prefetch): Move out of GCC target("sse") pragma.
5513 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
5514 GCC target("prfchw") pragma.
5515 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
5517 * config/i386/i386.c (ix86_option_override_internal): Enable
5518 -mprfchw with -mprefetchwt1.
5520 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5522 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
5525 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5527 * opts.h (CL_PCH_IGNORE): Define.
5528 * targhooks.c (option_affects_pch_p):
5529 Return false for options that have CL_PCH_IGNORE set.
5530 * opt-functions.awk: Process PchIgnore.
5531 * doc/options.texi: Document PchIgnore.
5533 * config/arc/arc.opt (misize): Add PchIgnore property.
5535 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5537 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
5538 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
5539 constraint on constants to permit them being loaded into
5540 GENERAL_REGS or BASE_REGS.
5542 2014-03-03 Nick Clifton <nickc@redhat.com>
5544 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
5545 anti-cacnonical alternatives.
5546 (negandhi3_real): New pattern.
5547 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
5549 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5551 * config/avr/avr-mcus.def: Remove atxmega16x1.
5552 * config/avr/avr-tables.opt: Regenerate.
5553 * config/avr/t-multilib: Regenerate.
5554 * doc/avr-mmcu.texi: Regenerate.
5556 2014-03-03 Tobias Grosser <tobias@grosser.es>
5557 Mircea Namolaru <mircea.namolaru@inria.fr>
5559 PR tree-optimization/58028
5560 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
5563 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5565 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
5566 not handled by recognizers.
5568 2014-03-03 Jakub Jelinek <jakub@redhat.com>
5571 * function.c (expand_function_end): Don't emit
5572 clobber_return_register sequence if clobber_after is a BARRIER.
5573 * cfgexpand.c (construct_exit_block): Append instructions before
5574 return_label to prev_bb.
5576 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5578 * config/rs6000/constraints.md: Document reserved use of "wc".
5580 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5583 * ipa.c (function_and_variable_visibility): When dissolving comdat
5584 group, also set all symbols to local.
5586 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5591 2013-12-14 Jan Hubicka <jh@suse.cz>
5593 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
5595 2014-03-02 Jon Beniston <jon@beniston.com>
5601 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
5602 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
5603 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
5604 (simple_return, *simple_return): New patterns
5605 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
5606 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
5608 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
5610 * dwarf2out.c (gen_subprogram_die): Tidy.
5612 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
5615 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
5616 (*mov_t_msb_neg_negc): ... this new insn.
5618 2014-02-28 Jason Merrill <jason@redhat.com>
5621 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
5624 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
5627 * dwarf2out.c (decltype_auto_die): New static.
5628 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
5629 (gen_type_die_with_usage): Handle 'decltype(auto)'.
5630 (is_cxx_auto): Likewise.
5632 2014-02-28 Ian Bolton <ian.bolton@arm.com>
5634 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
5635 we are not using general regs only.
5637 2014-02-28 Richard Biener <rguenther@suse.de>
5640 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
5641 previous fix and only allow to remove trivial pre-headers
5642 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
5643 (remove_forwarder_block): Properly update the latch of a loop.
5645 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5648 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
5649 (cselib_preserved_hash_table): New.
5650 (preserve_constants_and_equivs): Move preserved vals to it.
5651 (cselib_find_slot): Look it up first.
5652 (cselib_init): Initialize it.
5653 (cselib_finish): Release it.
5654 (dump_cselib_table): Dump it.
5656 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5659 * cselib.c (remove_useless_values): Skip to avoid quadratic
5660 behavior if the condition moved from...
5661 (cselib_process_insn): ... here holds.
5663 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5666 * var-tracking.c (vt_initialize): Apply the same condition to
5667 preserve the CFA base value.
5669 2014-02-28 Joey Ye <joey.ye@arm.com>
5672 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
5673 if reload in progress or completed.
5675 2014-02-28 Tobias Burnus <burnus@net-b.de>
5678 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
5681 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
5683 * doc/tm.texi.in (Condition Code Status): Update documention for
5684 relative locations of cc0-setter and cc0-user.
5686 2014-02-27 Jeff Law <law@redhat.com>
5688 PR rtl-optimization/52714
5689 * combine.c (try_combine): When splitting an unrecognized PARALLEL
5690 into two independent simple sets, if I3 is a jump, ensure the
5691 pattern we place into I3 is a (set (pc) ...).
5693 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
5694 Jeff Law <law@redhat.com>
5696 PR rtl-optimization/49847
5697 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
5698 are in different blocks.
5699 * doc/tm.texi (Condition Code Status): Update documention for
5700 relative locations of cc0-setter and cc0-user.
5702 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
5705 * lra.c (lra_emit_add): Check SUBREG too.
5707 2014-02-27 Andreas Schwab <schwab@suse.de>
5709 * config/m68k/m68k.c (m68k_option_override): Disable
5710 -flive-range-shrinkage for classic m68k.
5711 (m68k_override_options_after_change): Likewise.
5713 2014-02-27 Marek Polacek <polacek@redhat.com>
5716 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
5717 -Wmaybe-uninitialized.
5719 2014-02-27 Alan Modra <amodra@gmail.com>
5722 * reload1.c (emit_input_reload_insns): When reload_override_in,
5723 set old to rl->in_reg when rl->in_reg is a subreg.
5725 2014-02-26 Richard Biener <rguenther@suse.de>
5728 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
5730 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
5732 * common/config/i386/predicates.md (const1256_operand): Remove.
5733 (const2356_operand): New.
5734 (const_1_to_2_operand): Remove.
5735 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5736 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5737 (*avx512pf_gatherpf<mode>sf): Ditto.
5738 (avx512pf_gatherpf<mode>df): Ditto.
5739 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5740 (*avx512pf_gatherpf<mode>df): Ditto.
5741 (avx512pf_scatterpf<mode>sf): Ditto.
5742 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5743 (*avx512pf_scatterpf<mode>sf): Ditto.
5744 (avx512pf_scatterpf<mode>df): Ditto.
5745 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5746 (*avx512pf_scatterpf<mode>df): Ditto.
5747 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
5749 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
5751 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
5752 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
5753 (_mm512_mask_testn_epi64_mask): Move to ...
5754 * config/i386/avx512cdintrin.h: Here.
5755 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
5756 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
5757 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
5758 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
5759 TARGET_AVX512F from TARGET_AVX512CD.
5761 2014-02-26 Richard Biener <rguenther@suse.de>
5764 * ipa.c (walk_polymorphic_call_targets): Properly guard
5765 call to inline_update_overall_summary.
5767 2014-02-26 Bin Cheng <bin.cheng@arm.com>
5770 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
5771 and latches only if requested. Fix latch if it is removed.
5772 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
5773 LOOPS_HAVE_PREHEADERS.
5775 2014-02-25 Andrew Pinski <apinski@cavium.com>
5777 * builtins.c (expand_builtin_thread_pointer): Create a new target
5778 when the target is NULL.
5780 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
5782 PR rtl-optimization/60317
5783 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
5784 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
5785 * lra-assigns.c: Include params.h.
5786 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
5787 other reload pseudos considerations.
5789 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5791 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
5792 to use canonical form for nor<mode>3.
5794 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5797 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
5800 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
5802 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
5803 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
5804 (ix86_handle_option): Handle OPT_mprefetchwt1.
5805 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
5806 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
5808 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
5809 OPTION_MASK_ISA_PREFETCHWT1.
5810 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
5811 (PTA_PREFETCHWT1): New.
5812 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
5813 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
5814 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
5815 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
5816 (*prefetch_avx512pf_<mode>_: Change into ...
5817 (*prefetch_prefetchwt1_<mode>: This.
5818 * config/i386/i386.opt (mprefetchwt1): New.
5819 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
5820 (_mm_prefetch): Handle intent to write.
5821 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
5823 2014-02-25 Richard Biener <rguenther@suse.de>
5826 * emit-rtl.c (mem_attrs_htab): Remove.
5827 (mem_attrs_htab_hash): Likewise.
5828 (mem_attrs_htab_eq): Likewise.
5829 (set_mem_attrs): Always allocate new mem-attrs when something changed.
5830 (init_emit_once): Do not allocate mem_attrs_htab.
5832 2014-02-25 Richard Biener <rguenther@suse.de>
5835 * lto-opts.c (lto_write_options): Output non-explicit conservative
5836 -fwrapv, -fno-trapv and -fno-strict-overflow.
5837 * lto-wrapper.c (merge_and_complain): Handle merging those options.
5838 (run_gcc): And pass them through.
5840 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
5842 * sel-sched.c (calculate_new_fences): New parameter ptime.
5843 Calculate it as a maximum over all fence cycles.
5844 (sel_sched_region_2): Adjust the call to calculate_new_fences.
5845 Print the final schedule timing when sched_verbose.
5847 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
5849 PR rtl-optimization/60292
5850 * sel-sched.c (fill_vec_av_set): Do not reset target availability
5851 bit fot the fence instruction.
5853 2014-02-24 Alangi Derick <alangiderick@gmail.com>
5855 * calls.h: Fix typo in comment.
5857 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
5859 * config/pa/pa.c (pa_output_move_double): Don't valididate when
5860 adjusting offsetable addresses.
5862 2014-02-24 Guozhi Wei <carrot@google.com>
5864 * sparseset.h (sparseset_pop): Fix the wrong index.
5866 2014-02-24 Walter Lee <walt@tilera.com>
5868 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
5869 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
5871 * common/config/tilegx/tilegx-common.c
5872 (TARGET_DEFAULT_TARGET_FLAGS): Define.
5873 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
5875 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
5876 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
5877 (tilegx_gimplify_va_arg_expr): Handle big endian.
5878 (tilegx_expand_unaligned_load): Ditto.
5879 (tilegx_expand_unaligned_store): Ditto.
5880 (TARGET_RETURN_IN_MSB): New.
5881 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
5882 (TARGET_ENDIAN_DEFAULT): New.
5883 (TARGET_BIG_ENDIAN): Handle big endian.
5884 (BYTES_BIG_ENDIAN): Ditto.
5885 (WORDS_BIG_ENDIAN): Ditto.
5886 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
5889 * config/tilegx/tilegx.md (extv): Handle big endian.
5891 (insn_st<n>): Ditto.
5892 (insn_st<n>_add<bitsuffix>): Ditto.
5893 (insn_stnt<n>): Ditto.
5894 (insn_stnt<n>_add<bitsuffix>):Ditto.
5895 (vec_interleave_highv8qi): Handle big endian.
5896 (vec_interleave_highv8qi_be): New.
5897 (vec_interleave_highv8qi_le): New.
5898 (insn_v1int_h): Handle big endian.
5899 (vec_interleave_lowv8qi): Handle big endian.
5900 (vec_interleave_lowv8qi_be): New.
5901 (vec_interleave_lowv8qi_le): New.
5902 (insn_v1int_l): Handle big endian.
5903 (vec_interleave_highv4hi): Handle big endian.
5904 (vec_interleave_highv4hi_be): New.
5905 (vec_interleave_highv4hi_le): New.
5906 (insn_v2int_h): Handle big endian.
5907 (vec_interleave_lowv4hi): Handle big endian.
5908 (vec_interleave_lowv4hi_be): New.
5909 (vec_interleave_lowv4hi_le): New.
5910 (insn_v2int_l): Handle big endian.
5911 (vec_interleave_highv2si): Handle big endian.
5912 (vec_interleave_highv2si_be): New.
5913 (vec_interleave_highv2si_le): New.
5914 (insn_v4int_h): Handle big endian.
5915 (vec_interleave_lowv2si): Handle big endian.
5916 (vec_interleave_lowv2si_be): New.
5917 (vec_interleave_lowv2si_le): New.
5918 (insn_v4int_l): Handle big endian.
5919 * config/tilegx/tilegx.opt (mbig-endian): New option.
5920 (mlittle-endian): New option.
5921 * doc/install.texi: Document tilegxbe-linux.
5922 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
5924 2014-02-24 Martin Jambor <mjambor@suse.cz>
5927 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
5928 there are no parameter descriptors.
5930 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
5932 PR rtl-optimization/60268
5933 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
5934 initialization to ...
5935 (sched_rgn_init): ... here.
5936 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
5938 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5940 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
5943 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
5945 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
5948 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5950 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
5951 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
5953 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5955 * config/microblaze/predicates.md: Add cmp_op predicate.
5956 * config/microblaze/microblaze.md: Add branch_compare instruction
5957 which uses cmp_op predicate and emits cmp insn before branch.
5958 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
5959 to microblaze_expand_conditional_branch and consolidate logic.
5960 (microblaze_expand_conditional_branch): emit branch_compare
5961 insn instead of handling cmp op separate from branch insn.
5963 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5965 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
5968 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5970 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
5971 define_insn with define_expand and new define_insn
5972 *altivec_lve<VI_char>x_internal.
5973 (altivec_stve<VI_char>x): Replace define_insn with define_expand
5974 and new define_insn *altivec_stve<VI_char>x_internal.
5975 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
5977 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
5979 (altivec_expand_stvex_be): New function.
5981 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
5983 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
5984 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
5985 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
5986 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
5988 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
5991 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
5992 instead of emit_move_insn.
5994 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5996 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
5998 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
5999 gen_altivec_vsumsws.
6001 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6003 * config/rs6000/altivec.md (altivec_lvxl): Rename as
6004 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
6005 (altivec_lvxl_<mode>): New define_expand incorporating
6006 -maltivec=be semantics where needed.
6007 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
6008 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
6009 semantics where needed.
6010 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
6011 (altivec_stvx_<mode>): New define_expand incorporating
6012 -maltivec=be semantics where needed.
6013 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
6014 VM2 iterator instead of V4SI.
6015 (altivec_stvxl_<mode>): New define_expand incorporating
6016 -maltivec=be semantics where needed.
6017 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
6018 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
6019 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
6020 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
6021 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
6022 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
6023 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
6024 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
6025 ALTIVEC_BUILTIN_STVXL.
6026 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
6027 (altivec_expand_stvx_be): Likewise.
6028 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
6029 (altivec_expand_lvx_be): Likewise.
6030 (altivec_expand_stvx_be): Likewise.
6031 (altivec_expand_builtin): Add cases for
6032 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
6033 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
6034 (altivec_init_builtins): Add definitions for
6035 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
6036 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
6038 2014-02-21 Catherine Moore <clm@codesourcery.com>
6040 * doc/invoke.texi (mvirt, mno-virt): Document.
6041 * config/mips/mips.opt (mvirt): New option.
6042 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
6044 2014-02-21 Richard Biener <rguenther@suse.de>
6046 PR tree-optimization/60276
6047 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
6048 (STMT_VINFO_MIN_NEG_DIST): New macro.
6049 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
6050 STMT_VINFO_MIN_NEG_DIST.
6051 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
6052 made for negative dependence distances still hold.
6054 2014-02-21 Richard Biener <rguenther@suse.de>
6057 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
6058 DECL_INITIAL for globals not in the current function context.
6060 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6062 PR tree-optimization/56490
6063 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
6064 * tree-ssa-uninit.c: Include params.h.
6065 (compute_control_dep_chain): Add num_calls argument, return false
6066 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
6067 num_calls to recursive call.
6068 (find_predicates): Change dep_chain into normal array,
6069 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
6070 variable and adjust compute_control_dep_chain caller.
6071 (find_def_preds): Likewise.
6073 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
6075 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
6076 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
6078 2014-02-21 Nick Clifton <nickc@redhat.com>
6080 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
6081 (pushhi1): Likewise.
6082 (popqi1): Add mode to pre_dec.
6085 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6087 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
6088 mode for mask of V8SFmode permutation.
6090 2014-02-20 Richard Henderson <rth@redhat.com>
6093 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
6094 a new pseudo for OLDVAL.
6096 2014-02-20 Jakub Jelinek <jakub@redhat.com>
6099 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
6100 gen_reg_rtx if d->testing_p.
6101 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
6102 if d->testing_p and we will certainly return true.
6103 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6106 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6108 * emit-rtl.c (gen_reg_rtx): Assert that
6109 crtl->emit.regno_pointer_align_length is non-zero.
6111 2014-02-20 Richard Henderson <rth@redhat.com>
6114 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6115 on failure the store back into EXPECT.
6117 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6118 Sandra Loosemore <sandra@codesourcery.com>
6120 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6121 * config/nios2/nios2.c (nios2_function_profiler): Add
6122 -fPIC (flag_pic == 2) support.
6123 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6124 (nios2_large_offset_p): New function.
6125 (nios2_unspec_reloc_p): Move up position, update to use
6126 nios2_large_offset_p.
6127 (nios2_unspec_address): Remove function.
6128 (nios2_unspec_offset): New function.
6129 (nios2_large_got_address): New function.
6130 (nios2_got_address): Add large offset support.
6131 (nios2_legitimize_tls_address): Update usage of removed and new
6133 (nios2_symbol_binds_local_p): New function.
6134 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
6135 (nios2_legitimize_address): Update to use nios2_large_offset_p.
6136 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
6137 (nios2_print_operand): Merge H/L processing, add hiadj/lo
6138 processing for (const (unspec ...)).
6139 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
6141 2014-02-20 Richard Biener <rguenther@suse.de>
6143 * tree-cfg.c (replace_uses_by): Mark altered BBs before
6144 doing the substitution.
6145 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
6147 2014-02-20 Martin Jambor <mjambor@suse.cz>
6150 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
6151 info when checking whether lattices are bottom.
6153 2014-02-20 Richard Biener <rguenther@suse.de>
6156 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
6159 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
6162 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
6163 parameter specifying the scaling.
6164 (inline_call): Update.
6165 (want_inline_recursively): Guard division by zero.
6166 (recursive_inlining): Update.
6167 * ipa-inline.h (clone_inlined_nodes): Update.
6169 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6172 * config/i386/i386.c (classify_argument): Pass structures of size
6173 64 bytes or less in register.
6175 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6176 Kirill Yukhin <kirill.yukhin@intel.com>
6178 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
6179 (_mm_rcp28_round_ss): Ditto.
6180 (_mm_rsqrt28_round_sd): Ditto.
6181 (_mm_rsqrt28_round_ss): Ditto.
6182 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
6183 (_mm_rcp14_round_ss): Ditto.
6184 (_mm_rsqrt14_round_sd): Ditto.
6185 (_mm_rsqrt14_round_ss): Ditto.
6186 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
6187 the first input operand, get rid of match_dup.
6188 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
6190 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
6192 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
6193 operand as the first input operand, set type attribute.
6194 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
6196 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
6197 operand as the first input operand, set type attribute.
6199 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6201 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
6204 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
6207 * config/i386/i386.c (construct_container): Remove TFmode check
6208 for X86_64_INTEGER_CLASS.
6210 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
6213 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
6214 only when -Wpsabi is enabled.
6216 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
6219 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
6220 passing arrays in registers are the same as for structs, so remove the
6221 special case for them.
6223 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
6225 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
6226 destination type, extract only the valid bits if the source type is not
6227 integral and has a different mode.
6229 2014-02-19 Richard Biener <rguenther@suse.de>
6232 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
6235 2014-02-19 Richard Biener <rguenther@suse.de>
6238 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
6239 (ipa_modify_call_arguments): Emit an argument load explicitely and
6240 preserve virtual SSA form there and for the replacement call.
6241 Do not update SSA form nor free dominance info.
6243 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6245 * ipa.c (function_and_variable_visibility): Also clear WEAK
6246 flag when disolving COMDAT_GROUP.
6248 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6250 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
6251 * ipa-prop.c (ipa_set_jf_known_type): Return early when
6253 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
6254 do more sanity checks.
6255 (detect_type_change): Return true when giving up early.
6256 (compute_complex_assign_jump_func): Fix type parameter of
6257 ipa_set_ancestor_jf.
6258 (compute_complex_ancestor_jump_func): Likewise.
6259 (update_jump_functions_after_inlining): Fix updating of
6261 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
6263 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6265 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
6266 inline clones when edge disappears.
6268 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6271 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
6272 Split 64-bit moves into 2 patterns. Do not allow the use of
6273 direct move for TDmode in little endian, since the decimal value
6274 has little endian bytes within a word, but the 64-bit pieces are
6275 ordered in a big endian fashion, and normal subreg's of TDmode are
6277 (mov<mode>_64bit_dm): Likewise.
6278 (movtd_64bit_nodm): Likewise.
6280 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6282 PR tree-optimization/60174
6283 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
6284 statement of an SSA_NAME that occurs in an abnormal PHI node.
6286 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6289 * final.c (SEEN_BB): Remove.
6290 (SEEN_NOTE, SEEN_EMITTED): Renumber.
6291 (final_scan_insn): Don't force_source_line on second
6292 NOTE_INSN_BASIC_BLOCK.
6294 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
6297 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
6298 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
6299 (type_natural_mode): Warn ABI change when %zmm register is not
6300 available for AVX512F vector value passing.
6302 2014-02-18 Kai Tietz <ktietz@redhat.com>
6305 * config/i386/i386.c (ix86_expand_prologue): Use value in
6306 rax register as displacement when restoring %r10 or %rax.
6307 Fix wrong offset when restoring both registers.
6309 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6311 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
6312 assertion with conditional return.
6314 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6315 Uros Bizjak <ubizjak@gmail.com>
6318 * config/i386/driver-i386.c (host_detect_local_cpu): If
6319 YMM state is not saved by the OS, also clear has_f16c. Move
6320 CPUID 0x80000001 handling before YMM state saving checking.
6322 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
6324 PR rtl-optimization/58960
6325 * haifa-sched.c (alloc_global_sched_pressure_data): New,
6326 factored out from ...
6327 (sched_init): ... here.
6328 (free_global_sched_pressure_data): New, factored out from ...
6329 (sched_finish): ... here.
6330 * sched-int.h (free_global_sched_pressure_data): Declare.
6331 * sched-rgn.c (nr_regions_initial): New static global.
6332 (haifa_find_rgns): Initialize it.
6333 (schedule_region): Disable sched-pressure for the newly
6336 2014-02-17 Richard Biener <rguenther@suse.de>
6338 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
6339 release SSA defs of pattern stmts.
6341 2014-02-17 Richard Biener <rguenther@suse.de>
6343 * tree-inline.c (expand_call_inline): Release the virtual
6344 operand defined by the call we are about to inline.
6346 2014-02-17 Richard Biener <rguenther@suse.de>
6348 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
6350 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
6351 Ilya Tocar <ilya.tocar@intel.com>
6353 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
6354 arguments order in builtin.
6355 (_mm512_permutexvar_epi64): Ditto.
6356 (_mm512_mask_permutexvar_epi64): Ditto
6357 (_mm512_maskz_permutexvar_epi32): Ditto
6358 (_mm512_permutexvar_epi32): Ditto
6359 (_mm512_mask_permutexvar_epi32): Ditto
6361 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6363 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
6364 (p8_vmrgow): Likewise.
6366 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6368 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
6371 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
6374 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
6375 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
6376 into 64-bit and 32-bit moves. On 64-bit moves, add support for
6377 using direct move instructions on ISA 2.07. Also adjust
6378 instruction length for 64-bit.
6379 (mov<mode>_64bit, TFmode/TDmode): Likewise.
6380 (mov<mode>_32bit, TFmode/TDmode): Likewise.
6382 2014-02-15 Alan Modra <amodra@gmail.com>
6386 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
6387 find_replacement on parts of insn rtl that might be reloaded.
6389 2014-02-15 Richard Biener <rguenther@suse.de>
6391 PR tree-optimization/60183
6392 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
6393 (tree_ssa_phiprop): Calculate and free post-dominators.
6395 2014-02-14 Jeff Law <law@redhat.com>
6397 PR rtl-optimization/60131
6398 * ree.c (get_extended_src_reg): New function.
6399 (combine_reaching_defs): Use it rather than assuming location of REG.
6400 (find_and_remove_re): Verify first operand of extension is
6401 a REG before adding the insns to the copy list.
6403 2014-02-14 Roland McGrath <mcgrathr@google.com>
6405 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
6406 * configure: Regenerated.
6407 * config.in: Regenerated.
6408 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
6409 instead of ASM_SHORT.
6411 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
6412 Richard Earnshaw <rearnsha@arm.com>
6414 PR rtl-optimization/59535
6415 * lra-constraints.c (process_alt_operands): Encourage alternative
6416 when unassigned pseudo class is superset of the alternative class.
6417 (inherit_reload_reg): Don't inherit when optimizing for code size.
6418 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
6419 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
6420 modes not less than 4 for Thumb1.
6422 2014-02-14 Kyle McMartin <kyle@redhat.com>
6425 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
6427 2014-02-14 Richard Biener <rguenther@suse.de>
6429 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
6430 (get_frame_arg): Drop the assert with langhook types_compatible_p.
6431 Do not strip INDIRECT_REFs.
6433 2014-02-14 Richard Biener <rguenther@suse.de>
6436 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
6437 DECL_FUNCTION_SPECIFIC_TARGET.
6438 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
6439 * tree-streamer-out.c (pack_ts_target_option): Remove.
6440 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
6441 (write_ts_function_decl_tree_pointers): Do not stream
6442 DECL_FUNCTION_SPECIFIC_TARGET.
6443 * tree-streamer-in.c (unpack_ts_target_option): Remove.
6444 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
6445 (lto_input_ts_function_decl_tree_pointers): Do not stream
6446 DECL_FUNCTION_SPECIFIC_TARGET.
6448 2014-02-14 Jakub Jelinek <jakub@redhat.com>
6450 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
6451 (get_initial_def_for_induction, vectorizable_induction): Ignore
6452 debug stmts when looking for exit_phi.
6453 (vectorizable_live_operation): Fix up condition.
6455 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6457 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
6458 nreverse() because it changes the content of original tree list.
6460 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6462 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
6463 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
6465 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6467 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
6468 GNU coding standards.
6470 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6473 * dwarf2out.c (gen_subprogram_die): Don't call
6474 add_calling_convention_attribute if subr_die is old_die.
6476 2014-02-13 Sharad Singhai <singhai@google.com>
6478 * doc/optinfo.texi: Fix order of nodes.
6480 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
6482 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
6483 operands[2], not operands[3].
6485 2014-02-13 Richard Biener <rguenther@suse.de>
6488 * doc/install.texi (ISL): Update recommended version to 0.12.2,
6489 mention the possibility of an in-tree build.
6490 (CLooG): Update recommended version to 0.18.1, mention the
6491 possibility of an in-tree build and clarify that the ISL
6492 bundled with CLooG does not work.
6494 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6497 * expr.c (compress_float_constant): If x is a hard register,
6498 extend into a pseudo and then move to x.
6500 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
6502 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
6503 caused by bad second argument to warning_at() with -mhotpatch and
6504 nested functions (e.g. with gfortran).
6506 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
6508 * opts.c (option_name): Remove "enabled by default" rider.
6510 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
6512 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
6514 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
6515 Uros Bizjak <ubizjak@gmail.com>
6518 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
6519 * configure: Regenerated.
6521 2014-02-12 Richard Biener <rguenther@suse.de>
6523 * vec.c (vec_prefix::calculate_allocation): Move as
6524 inline variant to vec.h.
6525 (vec_prefix::calculate_allocation_1): New out-of-line version.
6526 * vec.h (vec_prefix::calculate_allocation_1): Declare.
6527 (vec_prefix::m_has_auto_buf): Rename to ...
6528 (vec_prefix::m_using_auto_storage): ... this.
6529 (vec_prefix::calculate_allocation): Inline the easy cases
6530 and dispatch to calculate_allocation_1 which doesn't need the
6532 (va_heap::reserve): Use gcc_checking_assert.
6533 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
6534 m_using_auto_storage.
6535 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
6537 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
6538 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
6539 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
6541 2014-02-12 Richard Biener <rguenther@suse.de>
6543 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
6544 when we found a dependence.
6546 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
6548 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
6550 (maybe_fold_stmt): ... into this new function.
6551 * omp-low.c (lower_omp): Update comment.
6553 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
6556 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
6559 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
6561 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
6562 identifiers in comments.
6563 (cortexa53_extra_costs): Likewise.
6564 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
6565 (cortexa7_extra_costs): Likewise.
6566 (cortexa12_extra_costs): Likewise.
6567 (cortexa15_extra_costs): Likewise.
6568 (v7m_extra_costs): Likewise.
6570 2014-02-12 Richard Biener <rguenther@suse.de>
6573 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
6574 of posix_memalign being successful.
6575 (lower_stmt): Restrict lowering of posix_memalign to when
6576 -ftree-bit-ccp is enabled.
6578 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6580 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
6582 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
6584 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
6586 PR rtl-optimization/60116
6587 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
6588 other_insn once the combination has been validated.
6590 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
6593 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
6595 * ipa-devirt.c: Include demangle.h
6596 (odr_violation_reported): New static variable.
6597 (add_type_duplicate): Update odr_violations.
6598 (maybe_record_node): Add completep parameter; update it.
6599 (record_target_from_binfo): Add COMPLETEP parameter;
6600 update it as needed.
6601 (possible_polymorphic_call_targets_1): Likewise.
6602 (struct polymorphic_call_target_d): Add nonconstruction_targets;
6603 rename FINAL to COMPLETE.
6604 (record_targets_from_bases): Sanity check we found the binfo;
6605 fix COMPLETEP updating.
6606 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
6607 parameter, fix computing of COMPLETEP.
6608 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
6609 at LTO time do demangling.
6610 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
6611 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
6613 (gimple_get_virt_method_for_binfo): Likewise.
6614 * gimple-fold.h (gimple_get_virt_method_for_binfo,
6615 gimple_get_virt_method_for_vtable): Update prototypes.
6617 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
6620 * genautomata.c (add_presence_absence): Fix typo with
6621 {final_}presence_list.
6623 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
6626 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
6627 for VSX/Altivec vectors that land in GPR registers.
6629 2014-02-11 Richard Henderson <rth@redhat.com>
6630 Jakub Jelinek <jakub@redhat.com>
6633 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
6634 around drhs if type conversion to lacc->type is not useless.
6636 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6638 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
6640 (cortex-a57.cortex-a53): Likewise.
6641 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
6643 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6645 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
6648 2014-02-11 Renlin Li <Renlin.Li@arm.com>
6650 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
6651 add_options_for_arm_vfp3.
6653 2014-02-11 Jeff Law <law@redhat.com>
6656 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
6657 object with an undesirable mode.
6659 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6662 * config/i386/sol2-9.h: New file.
6663 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
6664 *-*-solaris2.9*): Use it.
6666 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6668 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
6669 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
6671 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6673 * config/microblaze/microblaze.c: Extend mcpu version format
6675 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
6677 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
6679 2014-02-10 Richard Henderson <rth@redhat.com>
6682 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
6683 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
6684 ms-abi vs -mno-accumulate-outgoing-args.
6685 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
6686 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
6689 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
6692 * cfgexpand.c (expand_asm_operands): Attach source location to
6693 ASM_INPUT rtx objects.
6694 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
6696 2014-02-10 Nick Clifton <nickc@redhat.com>
6698 * config/mn10300/mn10300.c (popcount): New function.
6699 (mn10300_expand_prologue): Include saved registers in stack usage
6702 2014-02-10 Jeff Law <law@redhat.com>
6705 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
6706 when changing the SET_DEST of a prior insn to avoid an input reload.
6708 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6710 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
6711 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
6712 -mcall-openbsd, or -mcall-linux.
6713 (CC1_ENDIAN_BIG_SPEC): Remove.
6714 (CC1_ENDIAN_LITTLE_SPEC): Remove.
6715 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
6716 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
6717 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
6718 and %cc1_endian_default.
6719 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
6721 2014-02-10 Richard Biener <rguenther@suse.de>
6723 PR tree-optimization/60115
6724 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
6725 MEM_REF handling. Properly verify that the accesses are not
6726 out of the objects bound.
6728 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6730 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
6733 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
6735 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
6736 proper constants and fix formatting.
6737 (possible_polymorphic_call_targets): Fix formatting.
6739 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
6740 Ilya Tocar <ilya.tocar@intel.com>
6742 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
6743 (_mm512_loadu_epi32): Renamed into...
6744 (_mm512_loadu_si512): This.
6745 (_mm512_storeu_epi32): Renamed into...
6746 (_mm512_storeu_si512): This.
6747 (_mm512_maskz_ceil_ps): Removed.
6748 (_mm512_maskz_ceil_pd): Ditto.
6749 (_mm512_maskz_floor_ps): Ditto.
6750 (_mm512_maskz_floor_pd): Ditto.
6751 (_mm512_floor_round_ps): Ditto.
6752 (_mm512_floor_round_pd): Ditto.
6753 (_mm512_ceil_round_ps): Ditto.
6754 (_mm512_ceil_round_pd): Ditto.
6755 (_mm512_mask_floor_round_ps): Ditto.
6756 (_mm512_mask_floor_round_pd): Ditto.
6757 (_mm512_mask_ceil_round_ps): Ditto.
6758 (_mm512_mask_ceil_round_pd): Ditto.
6759 (_mm512_maskz_floor_round_ps): Ditto.
6760 (_mm512_maskz_floor_round_pd): Ditto.
6761 (_mm512_maskz_ceil_round_ps): Ditto.
6762 (_mm512_maskz_ceil_round_pd): Ditto.
6763 (_mm512_expand_pd): Ditto.
6764 (_mm512_expand_ps): Ditto.
6765 * config/i386/i386.c (ix86_builtins): Remove
6766 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
6767 (bdesc_args): Ditto.
6768 * config/i386/predicates.md (const1256_operand): New.
6769 (const_1_to_2_operand): Ditto.
6770 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6771 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6772 (*avx512pf_gatherpf<mode>sf): Ditto.
6773 (avx512pf_gatherpf<mode>df): Ditto.
6774 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6775 (*avx512pf_gatherpf<mode>df): Ditto.
6776 (avx512pf_scatterpf<mode>sf): Ditto.
6777 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6778 (*avx512pf_scatterpf<mode>sf): Ditto.
6779 (avx512pf_scatterpf<mode>df): Ditto.
6780 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6781 (*avx512pf_scatterpf<mode>df): Ditto.
6782 (avx512f_expand<mode>): Removed.
6783 (<shift_insn><mode>3<mask_name>): Change predicate type.
6785 2014-02-08 Jakub Jelinek <jakub@redhat.com>
6787 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
6788 not at the end of datarefs vector use ordered_remove to avoid
6789 reordering datarefs vector.
6792 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
6793 mark local addressable non-static vars as GOVD_PRIVATE
6794 instead of GOVD_LOCAL.
6795 * omp-low.c (lower_omp_for): Move gimple_bind_vars
6796 and BLOCK_VARS of gimple_bind_block to new_stmt rather
6800 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
6801 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
6802 assume_aligned or alloc_align attributes.
6803 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
6804 arguments. Handle also assume_aligned and alloc_align attributes.
6805 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
6806 calls to functions with assume_aligned or alloc_align attributes.
6807 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
6809 2014-02-08 Terry Guo <terry.guo@arm.com>
6811 * doc/invoke.texi: Document ARM -march=armv7e-m.
6813 2014-02-08 Jakub Jelinek <jakub@redhat.com>
6815 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
6816 flag on __cilkrts_rethrow builtin.
6819 * ipa-cp.c (determine_versionability): Fail at -O0
6820 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
6821 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
6824 2014-02-04 Jakub Jelinek <jakub@redhat.com>
6827 * tree-inline.c (copy_forbidden): Fail for
6828 __attribute__((optimize (0))) functions.
6830 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
6832 * varpool.c: Include pointer-set.h.
6833 (varpool_remove_unreferenced_decls): Variables in other partitions
6834 will not be output; be however careful to not lose information
6837 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
6839 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
6840 lookup in the vtable constructor.
6842 2014-02-07 Jeff Law <law@redhat.com>
6845 * config/m68k/m68k.md (ashldi_extsi): Turn into a
6846 define_insn_and_split.
6848 * ipa-inline.c (inline_small_functions): Fix typos.
6850 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6852 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
6853 (s390_can_use_return_insn): Declare.
6854 * config/s390/s390.h (EPILOGUE_USES): Define.
6855 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
6857 (s390_chunkify_start): Handle return JUMP_LABELs.
6858 (s390_early_mach): Emit a main_pool instruction on the entry edge.
6859 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
6860 (s390_can_use_return_insn): New functions.
6861 (s390_fix_long_loop_prediction): Handle conditional returns.
6862 (TARGET_SET_UP_BY_PROLOGUE): Define.
6863 * config/s390/s390.md (ANY_RETURN): New code iterator.
6864 (*creturn, *csimple_return, return, simple_return): New patterns.
6866 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6868 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
6869 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
6870 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
6871 REG_CFA_RESTORE list when deciding not to restore a register.
6873 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6875 * config/s390/s390.c: Include tree-pass.h and context.h.
6876 (s390_early_mach): New function, split out from...
6877 (s390_emit_prologue): ...here.
6878 (pass_data_s390_early_mach): New pass structure.
6879 (pass_s390_early_mach): New class.
6880 (s390_option_override): Create and register early_mach pass.
6881 Move to end of file.
6883 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6885 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
6886 to match for the exit block.
6888 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6890 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
6891 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
6892 Reject misaligned operands.
6894 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6896 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
6898 2014-02-07 Richard Biener <rguenther@suse.de>
6901 * gimple-low.c (lower_builtin_posix_memalign): New function.
6902 (lower_stmt): Call it to lower posix_memalign in a way
6903 to make alignment info accessible.
6905 2014-02-07 Jakub Jelinek <jakub@redhat.com>
6908 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
6909 __builtin_setjmp_receiver.
6911 2014-02-07 Richard Biener <rguenther@suse.de>
6914 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
6915 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
6916 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
6917 Handle BUILT_IN_POSIX_MEMALIGN.
6918 (find_func_clobbers): Likewise.
6919 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
6920 (call_may_clobber_ref_p_1): Likewise.
6922 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6925 * ipa-devirt.c (record_target_from_binfo): Remove overactive
6928 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6931 * lto-cgraph.c (lto_output_node): Use
6932 symtab_get_symbol_partitioning_class.
6933 (lto_output_varpool_node): likewise.
6934 (symtab_get_symbol_partitioning_class): Move here from
6936 * cgraph.h (symbol_partitioning_class): Likewise.
6937 (symtab_get_symbol_partitioning_class): Declare.
6939 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6941 * ggc.h (ggc_internal_cleared_alloc): New macro.
6942 * vec.h (vec_safe_copy): Handle memory stats.
6943 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
6944 * target-globals.c (save_target_globals): Likewise.
6946 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6949 * expr.c (emit_move_resolve_push): Export; be bit more selective
6950 on when to clear alias set.
6951 * expr.h (emit_move_resolve_push): Declare.
6952 * function.h (struct function): Add tail_call_marked.
6953 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
6954 * config/i386/i386-protos.h (ix86_expand_push): Remove.
6955 * config/i386/i386.md (TImode move expander): De not call
6957 (FP push expanders): Preserve memory attributes.
6958 * config/i386/sse.md (push<mode>1): Remove.
6959 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
6960 (ix86_expand_push): Remove.
6961 * config/i386/mmx.md (push<mode>1): Remove.
6963 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6965 PR rtl-optimization/60030
6966 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
6967 lopart with paradoxical subreg before shifting it up by hprec.
6969 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6971 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
6972 Remove extra newline at end of file.
6973 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
6974 (arm_issue_rate): Handle cortexa57.
6975 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
6976 (cortex-a57.cortex-a53): Likewise.
6978 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6981 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
6982 don't record in REG_FRAME_RELATED_EXPR registers not set in that
6984 (arm_expand_prologue): Adjust all callers.
6985 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
6986 info, registers also at the lowest numbered registers side. Use
6987 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
6991 * var-tracking.c (adjust_mems): Before adding a SET to
6992 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
6994 2014-02-06 Alan Modra <amodra@gmail.com>
6997 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
6998 change SDmode to DDmode when lra_in_progress.
7000 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7003 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
7004 free_data_ref on the dr first, and before goto again also set dr
7005 to the next dr. For simd_lane_access, free old datarefs[i] before
7006 overwriting it. For get_vectype_for_scalar_type failure, don't
7007 free_data_ref if simd_lane_access.
7009 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
7012 * tree.h (opts_for_fn): New inline function.
7013 (opt_for_fn): Define.
7014 * config/i386/i386.c (ix86_function_regparm): Use
7015 opt_for_fn (decl, optimize) instead of optimize.
7017 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
7019 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
7020 for SYMBOL_REF in large memory model.
7022 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7024 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
7026 (cortex-a57): Likewise.
7027 (cortex-a57.cortex-a53): Likewise.
7029 2014-02-06 Yury Gribov <y.gribov@samsung.com>
7030 Kugan Vivekanandarajah <kuganv@linaro.org>
7032 * config/arm/arm.c (arm_vector_alignment_reachable): Check
7034 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
7036 2014-02-06 Richard Biener <rguenther@suse.de>
7038 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
7039 set_loop_copy and initialize_original_copy_tables.
7041 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
7043 * config/aarch64/aarch64-simd.md
7044 (aarch64_ashr_simddi): Change QI to SI.
7046 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7047 Jakub Jelinek <jakub@redhat.com>
7050 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
7053 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7055 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
7056 CODE_FOR_altivec_vpku[hw]um to
7057 CODE_FOR_altivec_vpku[hw]um_direct.
7058 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
7059 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
7060 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
7061 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
7063 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7065 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
7066 generation for -maltivec=be.
7067 (altivec_vsumsws): Simplify redundant test.
7069 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7071 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
7072 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
7073 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
7074 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
7075 gen_altivec_vpkuwum.
7076 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
7078 (altivec_vpks<VI_char>ss): Likewise.
7079 (altivec_vpks<VI_char>us): Likewise.
7080 (altivec_vpku<VI_char>us): Likewise.
7081 (altivec_vpku<VI_char>um): Likewise.
7082 (altivec_vpku<VI_char>um_direct): New (copy of
7083 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
7085 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
7086 target is little endian and -maltivec=be is not specified.
7087 (*altivec_vupkhs<VU_char>_direct): New (copy of
7088 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
7089 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
7090 target is little endian and -maltivec=be is not specified.
7091 (*altivec_vupkls<VU_char>_direct): New (copy of
7092 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
7093 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
7094 little endian and -maltivec=be is not specified.
7095 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
7096 little endian and -maltivec=be is not specified.
7098 2014-02-05 Richard Henderson <rth@redhat.com>
7101 * combine-stack-adj.c: Revert r206943.
7102 * sched-int.h (struct deps_desc): Add last_args_size.
7103 * sched-deps.c (init_deps): Initialize it.
7104 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7105 contain REG_ARGS_SIZE notes.
7107 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7109 * lto-cgraph.c (asm_nodes_output): Make global.
7110 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7111 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7112 (driver_handle_option): Handle OPT_fwpa.
7114 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7117 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7118 a comment typo and formatting issue. If odr_hash hasn't been
7119 created, return vNULL and set *completep to false.
7122 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7123 bb with no successors.
7125 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7128 * doc/invoke.texi (-march): Clarify documentation for ARM.
7132 2014-02-05 Richard Biener <rguenther@suse.de>
7134 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
7135 when not vectorizing because of too many alias checks.
7136 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
7137 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
7139 2014-02-05 Nick Clifton <nickc@redhat.com>
7141 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
7142 accept extended registers in any mode when compiling for the MN10300.
7144 2014-02-05 Yury Gribov <y.gribov@samsung.com>
7146 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
7147 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
7148 sanitization attributes.
7149 (can_inline_edge_p): Likewise.
7150 (sanitize_attrs_match_for_inline_p): New function.
7152 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7154 * ipa-prop.c (detect_type_change): Shor circuit testing of
7155 type changes on THIS pointer.
7157 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
7160 * config/pa/pa.c (legitimize_tls_address): Return original address
7161 if not passed a SYMBOL_REF rtx.
7162 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
7164 (pa_emit_move_sequence): Simplify TLS source operands.
7165 (pa_legitimate_constant_p): Reject all TLS constants.
7166 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
7167 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
7169 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7171 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
7172 groups when we know they are controlled by LTO.
7173 * varasm.c (default_binds_local_p_1): If object is in other partition,
7174 it will be resolved locally.
7176 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7178 * config/host-linux.c (linux_gt_pch_use_address): Don't
7179 use SSIZE_MAX because it is not always defined.
7181 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
7184 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
7185 threshold for pseudo splitting.
7186 (update_ebb_live_info): Process call argument hard registers and
7187 hard registers from insn definition too.
7188 (max_small_class_regs_num): New constant.
7189 (inherit_in_ebb): Update live hard regs through EBBs. Update
7190 reloads_num only for small register classes. Don't split for
7193 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
7196 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
7199 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7201 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
7204 2014-02-04 Marek Polacek <polacek@redhat.com>
7206 * gdbinit.in (pel): Define.
7208 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7210 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
7213 2014-02-04 Richard Biener <rguenther@suse.de>
7216 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
7217 in function context local.
7218 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
7219 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
7220 similar to LTO_imported_decl_ref.
7222 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7224 PR tree-optimization/60002
7225 * cgraphclones.c (build_function_decl_skip_args): Clear
7228 PR tree-optimization/60023
7229 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
7230 false to gsi_replace.
7231 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
7232 has been in some EH region and vec_stmt could throw, add
7233 vec_stmt into the same EH region.
7234 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
7235 has no lhs, ignore it.
7236 * internal-fn.c (expand_MASK_LOAD): Likewise.
7239 * tree-inline.c (copy_forbidden): Fail for
7240 __attribute__((optimize (0))) functions.
7243 * omp-low.c (simd_clone_struct_copy): If from->inbranch
7244 is set, copy one less argument.
7245 (expand_simd_clones): Don't subtract clone_info->inbranch
7246 from simd_clone_struct_alloc argument.
7248 PR rtl-optimization/57915
7249 * recog.c (simplify_while_replacing): If all unary/binary/relational
7250 operation arguments are constant, attempt to simplify those.
7253 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
7254 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
7256 2014-02-04 Richard Biener <rguenther@suse.de>
7258 PR tree-optimization/60012
7259 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
7260 TBAA disambiguation to all DDRs.
7262 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7265 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
7266 (LINK_SPEC): Use it for -shared, -shared-libgcc.
7268 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7271 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
7273 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7275 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
7276 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
7278 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7281 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
7282 to figure out targets of polymorphic calls with known decl.
7283 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7284 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
7285 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
7286 (get_polymorphic_call_info): ... here.
7287 (get_polymorphic_call_info_from_invariant): New function.
7289 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7291 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
7292 lookup via vtable pointer; check for type consistency
7293 and turn inconsitent facts into UNREACHABLE.
7294 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7295 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
7296 type inconsistent querries; return UNREACHABLE instead.
7298 2014-02-03 Richard Henderson <rth@twiddle.net>
7301 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
7302 already processed this node.
7303 (normalize_one_pred_1): Pass along mark_set.
7304 (normalize_one_pred): Create and destroy a pointer_set_t.
7305 (normalize_one_pred_chain): Likewise.
7307 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
7309 PR gcov-profile/58602
7310 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
7312 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7315 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
7316 -fno-devirtualize; try to devirtualize by the knowledge of
7317 virtual table pointer given by aggregate propagation.
7318 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7319 (ipa_print_node_jump_functions): Dump also offset that
7320 is relevant for polymorphic calls.
7321 (determine_known_aggregate_parts): Add arg_type parameter; use it
7322 instead of determining the type from pointer type.
7323 (ipa_compute_jump_functions_for_edge): Update call of
7324 determine_known_aggregate_parts.
7325 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
7326 (gimple_get_virt_method_for_binfo): ... here; simplify using
7327 vtable_pointer_value_to_vtable.
7328 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
7329 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
7330 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
7331 (vtable_pointer_value_to_vtable): Break out from ...; handle also
7333 (vtable_pointer_value_to_binfo): ... here.
7334 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
7336 2014-02-03 Teresa Johnson <tejohnson@google.com>
7338 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
7339 redef of outer loop index variable.
7341 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
7345 * doc/extend.texi (Function Attributes): Typo.
7347 2014-02-03 Cong Hou <congh@google.com>
7349 PR tree-optimization/60000
7350 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
7351 if the vectorized statement is a store. A store statement can only
7352 appear at the end of pattern statements.
7354 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7356 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
7357 (ix86_option_override_internal): Default long double to 64-bit for
7358 32-bit Bionic and to 128-bit for 64-bit Bionic.
7360 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
7361 TARGET_LONG_DOUBLE_128 is true.
7362 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
7364 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
7365 (mlong-double-64): Negate -mlong-double-128.
7366 (mlong-double-128): New option.
7368 * config/i386/i386-c.c (ix86_target_macros): Define
7369 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
7371 * doc/invoke.texi: Document -mlong-double-128.
7373 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7375 PR rtl-optimization/60024
7376 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
7378 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
7380 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
7382 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7384 PR rtl-optimization/57662
7385 * sel-sched.c (code_motion_path_driver): Do not mark already not
7386 existing blocks in the visiting bitmap.
7388 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7390 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
7391 on the insn being emitted.
7393 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
7394 Will Deacon <will.deacon@arm.com>
7396 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
7398 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7400 * config/arm/arm-tables.opt: Regenerate.
7402 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7404 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
7405 for vector types other than V16QImode.
7406 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
7407 define_expand, and call altivec_expand_vec_perm_le when producing
7408 code with little endian element order.
7409 (*altivec_vperm_<mode>_internal): New insn having previous
7410 behavior of altivec_vperm_<mode>.
7411 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
7412 altivec_expand_vec_perm_le when producing code with little endian
7414 (*altivec_vperm_<mode>_uns_internal): New insn having previous
7415 behavior of altivec_vperm_<mode>_uns.
7417 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7419 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
7420 (altivec_vsumsws): Add handling for -maltivec=be with a little
7422 (altivec_vsumsws_direct): New.
7423 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
7424 gen_altivec_vsumsws.
7426 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7428 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
7429 vtable_pointer_value_to_binfo): New functions.
7430 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
7431 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
7433 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
7435 * config/nios2/nios2.md (load_got_register): Initialize GOT
7436 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
7437 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
7439 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7441 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
7442 preserverd by passthrough, do not propagate the type.
7444 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7446 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
7447 (mips_atomic_assign_expand_fenv): New function.
7448 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
7450 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7452 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
7453 (__builtin_mips_set_fcsr): Likewise.
7454 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
7455 MIPS_USI_FTYPE_VOID.
7456 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
7457 (mips16_expand_set_fcsr): Likewise.
7458 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
7459 (mips16_set_fcsr_stub): Likewise.
7460 (mips16_get_fcsr_one_only_stub): New class.
7461 (mips16_set_fcsr_one_only_stub): Likewise.
7462 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
7463 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
7464 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
7465 (hard_float): New availability predicate.
7466 (mips_builtins): Add get_fcsr and set_fcsr.
7467 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
7468 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
7469 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
7470 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
7471 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
7474 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7476 * config/mips/mips.c (mips_one_only_stub): New class.
7477 (mips_need_mips16_rdhwr_p): Replace with...
7478 (mips16_rdhwr_stub): ...this new variable.
7479 (mips16_stub_call_address): New function.
7480 (mips16_rdhwr_one_only_stub): New class.
7481 (mips_expand_thread_pointer): Use mips16_stub_call_address.
7482 (mips_output_mips16_rdhwr): Delete.
7483 (mips_finish_stub): New function.
7484 (mips_code_end): Use it to handle rdhwr stubs.
7486 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
7489 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
7490 when calculating size of integer atomic types.
7492 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
7494 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
7496 2014-02-01 Jakub Jelinek <jakub@redhat.com>
7498 PR tree-optimization/60003
7499 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
7500 * profile.c (branch_prob): Use gimple_call_builtin_p
7501 to check for BUILT_IN_SETJMP_RECEIVER.
7502 * tree-inline.c (copy_bb): Call notice_special_calls.
7504 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
7507 * lra-constraints.c (process_alt_operands): Update reload_sum only
7510 2014-01-31 Richard Henderson <rth@redhat.com>
7513 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
7514 until after else_eh is processed.
7516 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7518 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
7519 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
7520 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
7521 in smmintrin.h, remove them.
7522 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
7523 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
7524 * config/i386/i386.md (ROUND_SAE): Fix value.
7525 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
7526 (const48_operand): New.
7527 * config/i386/subst.md (round), (round_expand): Use
7528 const_4_or_8_to_11_operand.
7529 (round_saeonly), (round_saeonly_expand): Use const48_operand.
7531 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7533 * config/i386/constraints.md (Yk): Swap meaning with k.
7534 * config/i386/i386.md (movhi_internal): Change Yk to k.
7535 (movqi_internal): Ditto.
7536 (*k<logic><mode>): Ditto.
7539 (kandn<mode>): Ditto.
7540 (*<code>hi_1): Ditto.
7541 (*<code>qi_1): Ditto.
7542 (kxnor<mode>): Ditto.
7543 (kortestzhi): Ditto.
7544 (kortestchi): Ditto.
7546 (*one_cmplhi2_1): Ditto.
7547 (*one_cmplqi2_1): Ditto.
7548 * config/i386/sse.md (): Change k to Yk.
7549 (avx512f_load<mode>_mask): Ditto.
7550 (avx512f_blendm<mode>): Ditto.
7551 (avx512f_store<mode>_mask): Ditto.
7552 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
7553 (avx512f_storedqu<mode>_mask): Ditto.
7554 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
7556 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
7557 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
7558 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
7559 (avx512f_maskcmp<mode>3): Ditto.
7560 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
7561 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
7562 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
7563 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
7564 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
7565 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
7566 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
7567 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
7568 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
7569 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
7570 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
7571 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
7572 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
7573 (vec_extract_lo_<mode>_maskm): Ditto.
7574 (vec_extract_hi_<mode>_maskm): Ditto.
7575 (avx512f_vternlog<mode>_mask): Ditto.
7576 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
7577 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
7578 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
7579 (avx512f_<code>v8div16qi2_mask): Ditto.
7580 (avx512f_<code>v8div16qi2_mask_store): Ditto.
7581 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
7582 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
7583 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
7584 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
7585 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7586 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7587 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7588 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7589 (avx512cd_maskb_vec_dupv8di): Ditto.
7590 (avx512cd_maskw_vec_dupv16si): Ditto.
7591 (avx512f_vpermi2var<mode>3_maskz): Ditto.
7592 (avx512f_vpermi2var<mode>3_mask): Ditto.
7593 (avx512f_vpermi2var<mode>3_mask): Ditto.
7594 (avx512f_vpermt2var<mode>3_maskz): Ditto.
7595 (*avx512f_gathersi<mode>): Ditto.
7596 (*avx512f_gathersi<mode>_2): Ditto.
7597 (*avx512f_gatherdi<mode>): Ditto.
7598 (*avx512f_gatherdi<mode>_2): Ditto.
7599 (*avx512f_scattersi<mode>): Ditto.
7600 (*avx512f_scatterdi<mode>): Ditto.
7601 (avx512f_compress<mode>_mask): Ditto.
7602 (avx512f_compressstore<mode>_mask): Ditto.
7603 (avx512f_expand<mode>_mask): Ditto.
7604 * config/i386/subst.md (mask): Change k to Yk.
7605 (mask_scalar_merge): Ditto.
7608 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
7610 * doc/extend.texi (Vector Extensions): Document ?: in C++.
7612 2014-01-31 Richard Biener <rguenther@suse.de>
7615 * builtins.c (fold_builtin_memory_op): Make sure to not
7616 use a floating-point mode or a boolean or enumeral type for
7619 2014-01-30 DJ Delorie <dj@redhat.com>
7621 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
7622 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
7623 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
7624 whenever main() has an epilogue.
7626 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7628 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
7629 unused variable "field".
7630 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
7631 (vsx_mergeh_<mode>): Likewise.
7632 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
7633 (altivec_vmrghh): Likewise.
7634 (altivec_vmrghw): Likewise.
7635 (altivec_vmrglb): Likewise.
7636 (altivec_vmrglh): Likewise.
7637 (altivec_vmrglw): Likewise.
7638 (altivec_vspltb): Add missing uses.
7639 (altivec_vsplth): Likewise.
7640 (altivec_vspltw): Likewise.
7641 (altivec_vspltsf): Likewise.
7643 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7646 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
7647 frame related instructions.
7649 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
7651 PR rtl-optimization/59959
7652 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
7653 any reload of register whose subreg is invalid.
7655 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7657 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
7658 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
7659 Add missing return type - void.
7661 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7663 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
7664 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
7665 remove element index adjustment for endian (now handled in vsx.md
7667 (altivec_expand_vec_perm_const): Use
7668 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
7669 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
7670 (vsx_xxspltw_<mode>): Adjust element index for little endian.
7671 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
7672 define_expand and a new define_insn *altivec_vspltb_internal;
7673 adjust for -maltivec=be on a little endian target.
7674 (altivec_vspltb_direct): New.
7675 (altivec_vsplth): Divide into a define_expand and a new
7676 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
7677 little endian target.
7678 (altivec_vsplth_direct): New.
7679 (altivec_vspltw): Divide into a define_expand and a new
7680 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
7681 little endian target.
7682 (altivec_vspltw_direct): New.
7683 (altivec_vspltsf): Divide into a define_expand and a new
7684 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
7685 a little endian target.
7687 2014-01-30 Richard Biener <rguenther@suse.de>
7689 PR tree-optimization/59993
7690 * tree-ssa-forwprop.c (associate_pointerplus): Check we
7691 can propagate form the earlier stmt and avoid the transform
7692 when the intermediate result is needed.
7694 2014-01-30 Alangi Derick <alangiderick@gmail.com>
7696 * README.Portability: Fix typo.
7698 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
7700 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
7701 comparison_operator with ordered_comparison_operator.
7703 2014-01-30 Nick Clifton <nickc@redhat.com>
7705 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
7706 Rename to mn10300_store_multiple_regs.
7707 * config/mn10300/mn10300.c: Likewise.
7708 * config/mn10300/mn10300.md (store_movm): Fix typo: call
7709 store_multiple_regs.
7710 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
7711 Call mn10300_store_multiple_regs.
7713 2014-01-30 Nick Clifton <nickc@redhat.com>
7714 DJ Delorie <dj@redhat.com>
7716 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
7717 %fp 2 to keep registers after it properly word-aligned.
7718 (rl78_alloc_physical_registers_umul): Handle the case where both
7719 input operands are the same.
7721 2014-01-30 Richard Biener <rguenther@suse.de>
7723 PR tree-optimization/59903
7724 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
7727 2014-01-30 Jason Merrill <jason@redhat.com>
7730 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
7733 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
7735 2014-01-30 Richard Biener <rguenther@suse.de>
7737 PR tree-optimization/59951
7738 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
7740 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
7743 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
7744 SFmode to DFmode case.
7746 2014-01-29 DJ Delorie <dj@redhat.com>
7748 * config/msp430/msp430.opt (-minrt): New.
7749 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
7751 (ENDFILE_SPEC): Likewise.
7753 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
7755 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
7756 (estimate_function_body_sizes): Use it.
7758 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
7761 * dwarf2out.c (is_cxx_auto): New.
7762 (is_base_type): Use it.
7763 (gen_type_die_with_usage): Likewise.
7765 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7767 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
7768 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
7769 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
7770 -maltivec=be with LE targets.
7771 (vsx_mergeh_<mode>): Likewise.
7772 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
7773 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
7774 (altivec_vmrghb): Replace with define_expand and new
7775 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
7776 (altivec_vmrghb_direct): New define_insn.
7777 (altivec_vmrghh): Replace with define_expand and new
7778 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
7779 (altivec_vmrghh_direct): New define_insn.
7780 (altivec_vmrghw): Replace with define_expand and new
7781 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
7782 (altivec_vmrghw_direct): New define_insn.
7783 (*altivec_vmrghsf): Adjust for endianness.
7784 (altivec_vmrglb): Replace with define_expand and new
7785 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
7786 (altivec_vmrglb_direct): New define_insn.
7787 (altivec_vmrglh): Replace with define_expand and new
7788 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
7789 (altivec_vmrglh_direct): New define_insn.
7790 (altivec_vmrglw): Replace with define_expand and new
7791 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
7792 (altivec_vmrglw_direct): New define_insn.
7793 (*altivec_vmrglsf): Adjust for endianness.
7794 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
7795 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
7796 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
7797 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
7798 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
7799 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
7800 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
7801 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
7803 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
7805 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
7806 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
7809 2014-01-29 Richard Biener <rguenther@suse.de>
7811 PR tree-optimization/58742
7812 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
7813 associate_pointerplus_align.
7814 (associate_pointerplus_diff): New function.
7815 (associate_pointerplus): Likewise. Call associate_pointerplus_align
7816 and associate_pointerplus_diff.
7818 2014-01-29 Richard Biener <rguenther@suse.de>
7820 * lto-streamer.h (LTO_major_version): Bump to 3.
7821 (LTO_minor_version): Reset to 0.
7823 2014-01-29 Renlin Li <Renlin.Li@arm.com>
7825 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
7826 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
7827 (arm_file_start): Generate correct asm header for armv7ve.
7828 * config/arm/bpabi.h: Add multilib support for armv7ve.
7829 * config/arm/driver-arm.c: Change the architectures of cortex-a7
7830 and cortex-a15 to armv7ve.
7831 * config/arm/t-aprofile: Add multilib support for armv7ve.
7832 * doc/invoke.texi: Document -march=armv7ve.
7834 2014-01-29 Richard Biener <rguenther@suse.de>
7836 PR tree-optimization/58742
7837 * tree-ssa-forwprop.c (associate_plusminus): Return true
7838 if we changed sth, defer EH cleanup to ...
7839 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
7840 (simplify_mult): New function.
7842 2014-01-29 Jakub Jelinek <jakub@redhat.com>
7845 PR tree-optimization/59920
7846 * tree.c (build_common_builtin_nodes): Remove
7847 __builtin_setjmp_dispatcher initialization.
7848 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
7849 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
7850 instead of gsi_after_labels + manually skipping debug stmts.
7851 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
7852 ignore bbs with IFN_ABNORMAL_DISPATCHER.
7853 * tree-inline.c (copy_edges_for_bb): Remove
7854 can_make_abnormal_goto argument, instead add abnormal_goto_dest
7855 argument. Ignore computed_goto_p stmts. Don't call
7856 make_abnormal_goto_edges. If a call might need abnormal edges
7857 for non-local gotos, see if it already has an edge to
7858 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
7859 with true argument, don't do anything then, otherwise add
7860 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
7861 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
7863 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
7864 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
7865 (lower_stmt): Don't set data->calls_builtin_setjmp.
7866 (lower_builtin_setjmp): Adjust comment.
7867 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
7868 * tree-cfg.c (found_computed_goto): Remove.
7869 (factor_computed_gotos): Remove.
7870 (make_goto_expr_edges): Return bool, true for computed gotos.
7871 Don't call make_abnormal_goto_edges.
7872 (build_gimple_cfg): Don't set found_computed_goto, don't call
7873 factor_computed_gotos.
7874 (computed_goto_p): No longer static.
7875 (make_blocks): Don't set found_computed_goto.
7876 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
7877 (make_edges): If make_goto_expr_edges returns true, push bb
7878 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
7879 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
7880 vector. Record mapping between bbs and OpenMP regions if there
7881 are any, adjust make_gimple_omp_edges caller. Call
7882 handle_abnormal_edges.
7883 (make_abnormal_goto_edges): Remove.
7884 * tree-cfg.h (make_abnormal_goto_edges): Remove.
7885 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
7886 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
7887 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
7888 * internal-fn.def (ABNORMAL_DISPATCHER): New.
7889 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
7890 filling *region also set *region_idx to (*region)->entry->index.
7893 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
7894 For REGs set ORIGINAL_REGNO.
7896 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
7898 * doc/md.texi: Mention that a target shouldn't implement
7899 vec_widen_(s|u)mul_even/odd pair if it is less efficient
7902 2014-01-29 Jakub Jelinek <jakub@redhat.com>
7904 PR tree-optimization/59594
7905 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
7906 a copy of the datarefs vector rather than the vector itself.
7908 2014-01-28 Jason Merrill <jason@redhat.com>
7911 * dwarf2out.c (auto_die): New static.
7912 (gen_type_die_with_usage): Handle C++1y 'auto'.
7913 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
7916 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
7919 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
7920 (SPEC_X32): Likewise.
7921 (SPEC_64): Likewise.
7922 * config/i386/i386.c (ix86_option_override_internal): Turn off
7923 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
7925 (x86_file_start): Output .code16gcc for TARGET_16BIT.
7926 * config/i386/i386.h (TARGET_16BIT): New macro.
7927 (TARGET_16BIT_P): Likewise.
7928 * config/i386/i386.opt: Add m16.
7929 * doc/invoke.texi: Document -m16.
7931 2014-01-28 Jakub Jelinek <jakub@redhat.com>
7933 PR preprocessor/59935
7934 * input.c (location_get_source_line): Bail out on when line number
7935 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
7937 2014-01-28 Richard Biener <rguenther@suse.de>
7939 PR tree-optimization/58742
7940 * tree-ssa-forwprop.c (associate_plusminus): Handle
7941 pointer subtraction of the form (T)(P + A) - (T)P.
7943 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7945 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
7948 2014-01-28 Richard Biener <rguenther@suse.de>
7951 2014-01-28 Richard Biener <rguenther@suse.de>
7953 PR rtl-optimization/45364
7954 PR rtl-optimization/59890
7955 * var-tracking.c (local_get_addr_clear_given_value): Handle
7956 already cleared slot.
7957 (val_reset): Handle not allocated local_get_addr_cache.
7958 (vt_find_locations): Use post-order on the inverted CFG.
7960 2014-01-28 Richard Biener <rguenther@suse.de>
7962 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
7964 2014-01-28 Richard Biener <rguenther@suse.de>
7966 PR rtl-optimization/45364
7967 PR rtl-optimization/59890
7968 * var-tracking.c (local_get_addr_clear_given_value): Handle
7969 already cleared slot.
7970 (val_reset): Handle not allocated local_get_addr_cache.
7971 (vt_find_locations): Use post-order on the inverted CFG.
7973 2014-01-28 Alan Modra <amodra@gmail.com>
7975 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
7976 * configure.ac <recursive call for build != host>: Define
7977 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
7978 and LD_FOR_BUILD too.
7979 * configure: Regenerate.
7981 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
7983 * config/i386/i386.c (get_builtin_code_for_version): Separate
7984 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
7985 Broadwell from Haswell.
7987 2014-01-27 Steve Ellcey <sellcey@mips.com>
7989 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
7990 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
7991 * config/mips/mips.c (mips_option_override): Change setting
7993 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
7994 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
7995 Change from Mask to Var.
7997 2014-01-27 Jeff Law <law@redhat.com>
7999 * ipa-inline.c (inline_small_functions): Fix typo.
8001 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
8003 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
8004 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
8005 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
8006 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
8007 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
8008 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
8009 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
8010 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
8011 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
8012 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
8013 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
8014 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
8015 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
8016 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
8017 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
8018 (_mm512_storeu_epi64): Ditto.
8019 (_mm512_cmpge_epi32_mask): Ditto.
8020 (_mm512_cmpge_epu32_mask): Ditto.
8021 (_mm512_cmpge_epi64_mask): Ditto.
8022 (_mm512_cmpge_epu64_mask): Ditto.
8023 (_mm512_cmple_epi32_mask): Ditto.
8024 (_mm512_cmple_epu32_mask): Ditto.
8025 (_mm512_cmple_epi64_mask): Ditto.
8026 (_mm512_cmple_epu64_mask): Ditto.
8027 (_mm512_cmplt_epi32_mask): Ditto.
8028 (_mm512_cmplt_epu32_mask): Ditto.
8029 (_mm512_cmplt_epi64_mask): Ditto.
8030 (_mm512_cmplt_epu64_mask): Ditto.
8031 (_mm512_cmpneq_epi32_mask): Ditto.
8032 (_mm512_cmpneq_epu32_mask): Ditto.
8033 (_mm512_cmpneq_epi64_mask): Ditto.
8034 (_mm512_cmpneq_epu64_mask): Ditto.
8035 (_mm512_expand_pd): Ditto.
8036 (_mm512_expand_ps): Ditto.
8037 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
8038 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
8039 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
8040 * config/i386/i386.c (ix86_builtins): Add
8041 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
8042 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
8043 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
8044 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
8045 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
8046 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
8047 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
8048 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
8049 IX86_BUILTIN_PMOVUSQW512_MEM.
8050 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
8051 __builtin_ia32_pmovsqd512mem_mask,
8052 __builtin_ia32_pmovqd512mem_mask,
8053 __builtin_ia32_pmovusqw512mem_mask,
8054 __builtin_ia32_pmovsqw512mem_mask,
8055 __builtin_ia32_pmovqw512mem_mask,
8056 __builtin_ia32_pmovusdw512mem_mask,
8057 __builtin_ia32_pmovsdw512mem_mask,
8058 __builtin_ia32_pmovdw512mem_mask,
8059 __builtin_ia32_pmovqb512mem_mask,
8060 __builtin_ia32_pmovusqb512mem_mask,
8061 __builtin_ia32_pmovsqb512mem_mask,
8062 __builtin_ia32_pmovusdb512mem_mask,
8063 __builtin_ia32_pmovsdb512mem_mask,
8064 __builtin_ia32_pmovdb512mem_mask.
8065 (bdesc_args): Add __builtin_ia32_expanddf512,
8066 __builtin_ia32_expandsf512.
8067 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
8068 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
8069 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
8070 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
8071 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
8072 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
8073 (avx512f_<code>v8div16qi2_mask_store): This.
8074 (avx512f_expand<mode>): New.
8076 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
8078 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
8080 (_mm512_mask_prefetch_i64gather_pd): Ditto.
8081 (_mm512_prefetch_i32scatter_pd): Ditto.
8082 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
8083 (_mm512_prefetch_i64scatter_pd): Ditto.
8084 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
8085 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
8086 (_mm512_mask_prefetch_i64gather_ps): Ditto.
8087 (_mm512_prefetch_i32scatter_ps): Ditto.
8088 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
8089 (_mm512_prefetch_i64scatter_ps): Ditto.
8090 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
8091 * config/i386/i386-builtin-types.def: Define
8092 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
8093 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
8094 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
8095 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
8096 IX86_BUILTIN_SCATTERPFQPD.
8097 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
8098 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
8099 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
8100 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
8101 __builtin_ia32_scatterpfqps.
8102 (ix86_expand_builtin): Expand new built-ins.
8103 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8104 fix memory access data type.
8105 (*avx512pf_gatherpf<mode>_mask): Ditto.
8106 (*avx512pf_gatherpf<mode>): Ditto.
8107 (avx512pf_scatterpf<mode>): Ditto.
8108 (*avx512pf_scatterpf<mode>_mask): Ditto.
8109 (*avx512pf_scatterpf<mode>): Ditto.
8110 (GATHER_SCATTER_SF_MEM_MODE): New.
8111 (avx512pf_gatherpf<mode>df): Ditto.
8112 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8113 (*avx512pf_scatterpf<mode>df): Ditto.
8115 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8118 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8119 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8122 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8124 * common/config/arm/arm-common.c
8125 (arm_rewrite_mcpu): Handle multiple names.
8127 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8129 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8131 * gimple-builder.h (create_gimple_tmp): Delete.
8133 2014-01-27 Christian Bruel <christian.bruel@st.com>
8135 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
8138 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
8140 * config/pa/pa.md (call): Generate indirect long calls to non-local
8141 functions when outputing 32-bit code.
8142 (call_value): Likewise except for special call to buggy powf function.
8144 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
8145 portable runtime and PIC indirect calls.
8146 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
8147 and PIC call sequences. Use ldo instead of blr to set return register
8148 in PIC call sequence.
8150 2014-01-25 Walter Lee <walt@tilera.com>
8152 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
8153 avoid clobbering a live register.
8155 2014-01-25 Walter Lee <walt@tilera.com>
8157 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
8158 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
8159 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
8160 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
8162 2014-01-25 Walter Lee <walt@tilera.com>
8164 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
8165 arguments on even registers.
8166 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
8168 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
8169 (BIGGEST_ALIGNMENT): Ditto.
8170 (BIGGEST_FIELD_ALIGNMENT): Ditto.
8172 2014-01-25 Walter Lee <walt@tilera.com>
8174 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
8175 insns before bundling.
8176 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8178 2014-01-25 Walter Lee <walt@tilera.com>
8180 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
8181 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
8182 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
8184 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8186 * config/mips/constraints.md (kl): Delete.
8187 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
8188 define expands, using...
8189 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
8190 instructions for MIPS16.
8191 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
8192 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
8194 2014-01-25 Walter Lee <walt@tilera.com>
8196 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
8200 2014-01-25 Walter Lee <walt@tilera.com>
8202 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
8203 (TARGET_EXPAND_TO_RTL_HOOK): Define.
8205 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8207 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
8210 2014-01-25 Jakub Jelinek <jakub@redhat.com>
8212 * print-rtl.c (in_call_function_usage): New var.
8213 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
8214 EXPR_LIST mode as mode and not as reg note name.
8217 * cfgloopmanip.c (copy_loop_info): If
8218 loop->warned_aggressive_loop_optimizations, make sure
8219 the flag is set in target loop too.
8221 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
8223 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
8225 * builtins.def: Likewise.
8226 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
8227 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
8228 * ira.c (ira_setup_eliminable_regset): Likewise.
8229 * omp-low.c (gate_expand_omp): Likewise.
8230 (execute_lower_omp): Likewise.
8231 (diagnose_sb_0): Likewise.
8232 (gate_diagnose_omp_blocks): Likewise.
8233 (simd_clone_clauses_extract): Likewise.
8236 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8238 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
8239 correction for little endian...
8240 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
8243 2014-01-24 Jeff Law <law@redhat.com>
8245 PR tree-optimization/59919
8246 * tree-vrp.c (find_assert_locations_1): Do not register asserts
8247 for non-returning calls.
8249 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
8251 * common/config/aarch64/aarch64-common.c
8252 (aarch64_rewrite_mcpu): Handle multiple names.
8253 * config/aarch64/aarch64.h
8254 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8256 2014-01-24 Dodji Seketeli <dodji@redhat.com>
8258 * input.c (add_file_to_cache_tab): Handle the case where fopen
8261 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
8264 * config/i386/i386.md (pushsf splitter): Get stack adjustment
8265 from push operand if code of push isn't PRE_DEC.
8267 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8270 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
8271 -mquad-memory-atomic. Update -mquad-memory documentation to say
8272 it is only used for non-atomic loads/stores.
8274 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
8275 -mquad-memory or -mquad-memory-atomic switches.
8277 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
8278 -mquad-memory-atomic to ISA 2.07 support.
8280 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
8281 to separate support of normal quad word memory operations (ldq, stq)
8282 from the atomic quad word memory operations.
8284 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
8285 support to separate non-atomic quad word operations from atomic
8286 quad word operations. Disable non-atomic quad word operations in
8287 little endian mode so that we don't have to swap words after the
8288 load and before the store.
8289 (quad_load_store_p): Add comment about atomic quad word support.
8290 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
8291 options printed with -mdebug=reg.
8293 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
8294 -mquad-memory-atomic as the test for whether we have quad word
8295 atomic instructions.
8296 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
8297 or -mp8-vector are used, allow byte/half-word atomic operations.
8299 * config/rs6000/sync.md (load_lockedti): Insure that the address
8300 is a proper indexed or indirect address for the lqarx instruction.
8301 On little endian systems, swap the hi/lo registers after the lqarx
8303 (load_lockedpti): Use indexed_or_indirect_operand predicate to
8304 insure the address is valid for the lqarx instruction.
8305 (store_conditionalti): Insure that the address is a proper indexed
8306 or indirect address for the stqcrx. instruction. On little endian
8307 systems, swap the hi/lo registers before doing the stqcrx.
8309 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
8310 insure the address is valid for the stqcrx. instruction.
8312 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8313 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
8314 type of quad memory support is available.
8316 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
8319 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
8320 there is a danger of looping.
8322 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
8324 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8325 force flag_ira_loop_pressure if set via command line.
8327 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8329 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
8330 (ashr_simd): New builtin handling DI mode.
8331 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
8332 (aarch64_sshr_simddi): New match pattern.
8333 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
8334 (vshrd_n_s64): Likewise.
8335 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
8337 2014-01-23 Nick Clifton <nickc@redhat.com>
8339 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
8340 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
8341 favour of mcu specific scripts.
8342 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
8345 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8346 Alex Velenko <Alex.Velenko@arm.com>
8348 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
8349 (vaddv_s16): Likewise.
8350 (vaddv_s32): Likewise.
8351 (vaddv_u8): Likewise.
8352 (vaddv_u16): Likewise.
8353 (vaddv_u32): Likewise.
8354 (vaddvq_s8): Likewise.
8355 (vaddvq_s16): Likewise.
8356 (vaddvq_s32): Likewise.
8357 (vaddvq_s64): Likewise.
8358 (vaddvq_u8): Likewise.
8359 (vaddvq_u16): Likewise.
8360 (vaddvq_u32): Likewise.
8361 (vaddvq_u64): Likewise.
8362 (vaddv_f32): Likewise.
8363 (vaddvq_f32): Likewise.
8364 (vaddvq_f64): Likewise.
8365 (vmaxv_f32): Likewise.
8366 (vmaxv_s8): Likewise.
8367 (vmaxv_s16): Likewise.
8368 (vmaxv_s32): Likewise.
8369 (vmaxv_u8): Likewise.
8370 (vmaxv_u16): Likewise.
8371 (vmaxv_u32): Likewise.
8372 (vmaxvq_f32): Likewise.
8373 (vmaxvq_f64): Likewise.
8374 (vmaxvq_s8): Likewise.
8375 (vmaxvq_s16): Likewise.
8376 (vmaxvq_s32): Likewise.
8377 (vmaxvq_u8): Likewise.
8378 (vmaxvq_u16): Likewise.
8379 (vmaxvq_u32): Likewise.
8380 (vmaxnmv_f32): Likewise.
8381 (vmaxnmvq_f32): Likewise.
8382 (vmaxnmvq_f64): Likewise.
8383 (vminv_f32): Likewise.
8384 (vminv_s8): Likewise.
8385 (vminv_s16): Likewise.
8386 (vminv_s32): Likewise.
8387 (vminv_u8): Likewise.
8388 (vminv_u16): Likewise.
8389 (vminv_u32): Likewise.
8390 (vminvq_f32): Likewise.
8391 (vminvq_f64): Likewise.
8392 (vminvq_s8): Likewise.
8393 (vminvq_s16): Likewise.
8394 (vminvq_s32): Likewise.
8395 (vminvq_u8): Likewise.
8396 (vminvq_u16): Likewise.
8397 (vminvq_u32): Likewise.
8398 (vminnmv_f32): Likewise.
8399 (vminnmvq_f32): Likewise.
8400 (vminnmvq_f64): Likewise.
8402 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8404 * config/aarch64/aarch64-simd.md
8405 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
8406 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
8407 (*aarch64_mul3_elt<mode>): Likewise.
8408 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
8409 (*aarch64_mul3_elt_to_64v2df): Likewise.
8410 (*aarch64_mla_elt<mode>): Likewise.
8411 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8412 (*aarch64_mls_elt<mode>): Likewise.
8413 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8414 (*aarch64_fma4_elt<mode>): Likewise.
8415 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8416 (*aarch64_fma4_elt_to_64v2df): Likewise.
8417 (*aarch64_fnma4_elt<mode>): Likewise.
8418 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8419 (*aarch64_fnma4_elt_to_64v2df): Likewise.
8420 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
8421 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
8422 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8423 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8424 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
8425 (aarch64_sqdmull_lane<mode>_internal): Likewise.
8426 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
8428 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
8430 * config/aarch64/aarch64-simd.md
8431 (aarch64_be_checked_get_lane<mode>): New define_expand.
8432 * config/aarch64/aarch64-simd-builtins.def
8433 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
8434 New builtin definition.
8435 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
8436 Use new safe be builtin.
8438 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8440 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
8442 (aarch64_be_st1<mode>): Likewise.
8443 (aarch_ld1<VALL:mode>): Define_expand modified.
8444 (aarch_st1<VALL:mode>): Likewise.
8445 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
8446 (UNSPEC_ST1): Likewise.
8448 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
8450 * config/microblaze/microblaze.md: Add trap insn and attribute
8452 2014-01-23 Dodji Seketeli <dodji@redhat.com>
8454 PR preprocessor/58580
8455 * input.h (location_get_source_line): Take an additional line_size
8457 (void diagnostics_file_cache_fini): Declare new function.
8458 * input.c (struct fcache): New type.
8459 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
8460 New static constants.
8461 (diagnostic_file_cache_init, total_lines_num)
8462 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
8463 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
8464 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
8465 (get_next_line, read_next_line, goto_next_line, read_line_num):
8466 New static function definitions.
8467 (diagnostic_file_cache_fini): New function.
8468 (location_get_source_line): Take an additional output line_len
8469 parameter. Re-write using lookup_or_add_file_to_cache_tab and
8471 * diagnostic.c (diagnostic_finish): Call
8472 diagnostic_file_cache_fini.
8473 (adjust_line): Take an additional input parameter for the length
8474 of the line, rather than calculating it with strlen.
8475 (diagnostic_show_locus): Adjust the use of
8476 location_get_source_line and adjust_line with respect to their new
8477 signature. While displaying a line now, do not stop at the first
8478 null byte. Rather, display the zero byte as a space and keep
8479 going until we reach the size of the line.
8480 * Makefile.in: Add vec.o to OBJS-libcommon
8482 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8483 Ilya Tocar <ilya.tocar@intel.com>
8485 * config/i386/avx512fintrin.h (_mm512_kmov): New.
8486 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
8487 (__builtin_ia32_kmov16): Ditto.
8488 * config/i386/i386.md (UNSPEC_KMOV): New.
8491 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8493 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
8494 (_mm512_storeu_si512): Ditto.
8496 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
8499 * rtl.h (get_referenced_operands): Declare.
8500 * recog.c (get_referenced_operands): New function.
8501 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
8502 operands have been referenced when recording LO_SUM references.
8504 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
8506 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
8508 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8510 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
8511 Enable for generic and recent AMD targets.
8513 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8515 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
8516 ARG_SIZE note when adjustment was eliminated.
8518 2014-01-22 Jeff Law <law@redhat.com>
8520 PR tree-optimization/59597
8521 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
8522 in file. Accept new argument REGISTERING and use it to modify
8523 dump output appropriately.
8524 (register_jump_thread): Corresponding changes.
8525 (mark_threaded_blocks): Reinstate code to cancel unprofitable
8526 thread paths involving joiner blocks. Add code to dump cancelled
8527 jump threading paths.
8529 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
8531 PR rtl-optimization/59477
8532 * lra-constraints.c (inherit_in_ebb): Process call for living hard
8533 regs. Update reloads_num and potential_reload_hard_regs for all insns.
8535 2014-01-22 Tom Tromey <tromey@redhat.com>
8537 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
8539 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
8541 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8543 PR rtl-optimization/59896
8544 * lra-constraints.c (process_alt_operands): Check unused note for
8545 matched operands of insn with no output reloads.
8547 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
8549 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
8550 (mips_move_from_gpr_cost): Likewise.
8552 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8554 PR rtl-optimization/59858
8555 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
8556 ira_class_hard_regs_num.
8557 (process_alt_operands): Increase reject for dying matched operand.
8559 2014-01-21 Jakub Jelinek <jakub@redhat.com>
8562 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
8563 smaller than size, perform several stores or loads and stores
8564 at dst + count - size to store or copy all of size bytes, rather
8565 than just last modesize bytes.
8567 2014-01-20 DJ Delorie <dj@redhat.com>
8569 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
8570 that CLOBBERs are REGs before propogating their values.
8572 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
8575 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
8576 (cgraph_inline_failed_type): New function.
8577 * cgraph.h (DEFCIFCODE): Add type.
8578 (cgraph_inline_failed_type_t): New enum.
8579 (cgraph_inline_failed_type): New prototype.
8580 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
8581 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
8582 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
8583 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
8584 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
8585 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
8586 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
8587 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
8588 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
8589 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
8590 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
8591 OPTIMIZATION_MISMATCH.
8592 * tree-inline.c (expand_call_inline): Emit errors during
8593 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
8595 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8598 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
8599 mode attribute in insn output.
8601 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
8603 * output.h (output_constant): Delete.
8604 * varasm.c (output_constant): Make private.
8606 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
8608 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
8610 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8613 * tree.h (fold_builtin_strcat): New prototype.
8614 * builtins.c (fold_builtin_strcat): No longer static. Add len
8615 argument, if non-NULL, don't call c_strlen. Optimize
8616 directly into __builtin_memcpy instead of __builtin_strcpy.
8617 (fold_builtin_2): Adjust fold_builtin_strcat caller.
8618 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
8620 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8622 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8623 for SImode_address_operand operands, having only a REG argument.
8625 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
8627 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
8628 loader name using mbig-endian.
8629 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
8631 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8633 * doc/invoke.texi (-march): Clarify documentation for AArch64.
8637 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
8639 * config/aarch64/aarch64-protos.h
8640 (aarch64_cannot_change_mode_class_ptr): Declare.
8641 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
8642 aarch64_cannot_change_mode_class_ptr): New.
8643 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
8644 backend hook aarch64_cannot_change_mode_class.
8646 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8648 * common/config/aarch64/aarch64-common.c
8649 (aarch64_handle_option): Don't handle any option order logic here.
8650 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
8651 selected_cpu, warn on architecture version mismatch.
8652 (aarch64_override_options): Fix parsing order for option strings.
8654 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8655 Iain Sandoe <iain@codesourcery.com>
8658 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
8659 warning. Amend comment to reflect current functionality.
8661 2014-01-20 Richard Biener <rguenther@suse.de>
8664 * builtins.c (fold_builtin_strcat): Remove case better handled
8665 by tree-ssa-strlen.c.
8667 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
8669 * config/aarch64/aarch64.opt
8670 (mcpu, march, mtune): Make case-insensitive.
8672 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8675 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8676 if operands[1] is a REG or ZERO_EXTEND of a REG.
8678 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
8680 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
8682 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
8684 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
8685 long non-pic millicode calls.
8687 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8689 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
8691 2014-01-19 Kito Cheng <kito@0xlab.org>
8693 * builtins.c (expand_movstr): Check movstr expand done or fail.
8695 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8696 H.J. Lu <hongjiu.lu@intel.com>
8699 * config/i386/i386.md (*lea<mode>): Zero-extend return register
8700 to DImode for zero-extended addresses.
8702 2014-01-19 Jakub Jelinek <jakub@redhat.com>
8704 PR rtl-optimization/57763
8705 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
8706 on the new indirect jump_insn and increment LABEL_NUSES (label).
8708 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
8712 * config.gcc (x86_archs): New variable.
8713 (x86_64_archs): Likewise.
8714 (x86_cpus): Likewise.
8715 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
8716 --with-arch/--with-cpu= options.
8717 Support --with-arch=/--with-cpu={nehalem,westmere,
8718 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
8720 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8722 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
8723 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
8725 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8727 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
8729 2014-01-18 Jakub Jelinek <jakub@redhat.com>
8732 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
8733 clear cpp_get_options (parse_in)->warn_unused_macros for
8734 ix86_target_macros_internal with cpp_define.
8736 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
8738 * jump.c (delete_related_insns): Keep (use (insn))s.
8739 * reorg.c (redundant_insn): Check for barriers too.
8741 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8743 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
8745 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
8747 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
8748 call to $$dyncall when TARGET_LONG_CALLS is true.
8750 2014-01-17 Jeff Law <law@redhat.com>
8752 * ree.c (combine_set_extension): Temporarily disable test for
8753 changing number of hard registers.
8755 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
8758 * ipa-inline-analysis.c (inline_free_summary):
8759 Do not free summary of aliases.
8761 2014-01-17 Jakub Jelinek <jakub@redhat.com>
8764 * gimplify.c (gimplify_expr): Use create_tmp_var
8765 instead of create_tmp_var_raw. If cond doesn't have
8766 integral type, don't add the IFN_ANNOTATE builtin at all.
8768 2014-01-17 Martin Jambor <mjambor@suse.cz>
8771 * ipa-cp.c (prev_edge_clone): New variable.
8772 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
8773 Also resize prev_edge_clone vector.
8774 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
8775 (ipcp_edge_removal_hook): New function.
8776 (ipcp_driver): Register ipcp_edge_removal_hook.
8778 2014-01-17 Andrew Pinski <apinski@cavium.com>
8779 Steve Ellcey <sellcey@mips.com>
8782 * config/mips/mips.c (mips_print_operand): Check operand mode instead
8785 2014-01-17 Jeff Law <law@redhat.com>
8788 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
8789 so that pass_ccp runs first.
8791 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8793 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
8794 (ix86_adjust_cost): Use !TARGET_XXX.
8795 (do_reorder_for_imul): Likewise.
8796 (swap_top_of_ready_list): Likewise.
8797 (ix86_sched_reorder): Likewise.
8799 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8801 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
8802 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
8803 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
8804 (intel_memset): New. Duplicate slm_memset.
8805 (intel_cost): New. Duplicate slm_cost.
8806 (m_INTEL): New macro.
8807 (processor_target_table): Add "intel".
8808 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
8809 with PROCESSOR_INTEL for "intel".
8810 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
8811 PROCESSOR_SILVERMONT.
8812 (ix86_issue_rate): Likewise.
8813 (ix86_adjust_cost): Likewise.
8814 (ia32_multipass_dfa_lookahead): Likewise.
8815 (swap_top_of_ready_list): Likewise.
8816 (ix86_sched_reorder): Likewise.
8817 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
8818 instead of TARGET_OPT_AGU.
8819 * config/i386/i386.h (TARGET_INTEL): New.
8820 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
8821 (processor_type): Add PROCESSOR_INTEL.
8822 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
8823 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
8825 2014-01-17 Marek Polacek <polacek@redhat.com>
8828 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
8831 2014-01-17 Richard Biener <rguenther@suse.de>
8833 PR tree-optimization/46590
8834 * opts.c (default_options_table): Add entries for
8835 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
8836 all enabled at -O1 but not for -Og.
8837 * common.opt (fbranch-count-reg): Remove Init(1).
8838 (fmove-loop-invariants): Likewise.
8839 (ftree-pta): Likewise.
8841 2014-01-17 Jakub Jelinek <jakub@redhat.com>
8843 * config/i386/i386.c (ix86_data_alignment): For compatibility with
8844 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
8845 decls to at least the GCC 4.8 used alignments.
8848 * tree-nested.c (convert_nonlocal_reference_stmt,
8849 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
8850 of GIMPLE_BIND stmts, adjust associated decls.
8852 2014-01-17 Richard Biener <rguenther@suse.de>
8854 PR tree-optimization/46590
8855 * vec.h (vec<>::bseach): New member function implementing
8856 binary search according to C89 bsearch.
8857 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
8858 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
8859 bitmap pointer again. Make accesses_in_loop a flat array.
8860 (mem_ref_obstack): New global.
8861 (outermost_indep_loop): Adjust for mem_ref->stored changes.
8862 (mark_ref_stored): Likewise.
8863 (ref_indep_loop_p_2): Likewise.
8864 (set_ref_stored_in_loop): New helper function.
8865 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
8866 (memref_free): Adjust.
8867 (record_mem_ref_loc): Simplify.
8868 (gather_mem_refs_stmt): Adjust.
8869 (sort_locs_in_loop_postorder_cmp): New function.
8870 (analyze_memory_references): Sort accesses_in_loop after
8871 loop postorder number.
8872 (find_ref_loc_in_loop_cmp): New function.
8873 (for_all_locs_in_loop): Find relevant cluster of locs in
8874 accesses_in_loop and iterate without recursion.
8875 (execute_sm): Avoid uninit warning.
8876 (struct ref_always_accessed): Simplify.
8877 (ref_always_accessed::operator ()): Likewise.
8878 (ref_always_accessed_p): Likewise.
8879 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
8880 loop postorder numbers here.
8881 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
8884 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
8887 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
8888 on decls for which assemble_alias has been called.
8890 2014-01-17 Nick Clifton <nickc@redhat.com>
8892 * config/msp430/msp430.opt: (mcpu): New option.
8893 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
8894 (msp430_option_override): Parse target_cpu. If the MCU name
8895 matches a generic string, clear target_mcu.
8896 (msp430_attr): Allow numeric interrupt values up to 63.
8897 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
8898 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
8900 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
8902 * config/msp430/msp430.md (popm): Use %J rather than %I.
8903 (addsi3): Use msp430_nonimmediate_operand for operand 2.
8904 (addhi_cy_i): Use immediate_operand for operand 2.
8905 * doc/invoke.texi: Document -mcpu option.
8907 2014-01-17 Richard Biener <rguenther@suse.de>
8909 PR rtl-optimization/38518
8910 * df.h (df_analyze_loop): Declare.
8911 * df-core.c: Include cfgloop.h.
8912 (df_analyze_1): Split out main part of df_analyze.
8913 (df_analyze): Adjust.
8914 (loop_inverted_post_order_compute): New function.
8915 (loop_post_order_compute): Likewise.
8916 (df_analyze_loop): New function avoiding whole-function
8918 * loop-invariant.c (find_defs): Use df_analyze_loop.
8919 (find_invariants): Adjust.
8920 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
8922 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
8924 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
8925 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
8927 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
8929 * ipa-ref.c (ipa_remove_stmt_references): Fix references
8930 traversal when removing references.
8932 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
8935 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
8937 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
8940 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
8941 pushing a reload for an autoinc when we had previously reloaded an
8942 inner part of the address.
8944 2014-01-16 Jakub Jelinek <jakub@redhat.com>
8946 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
8948 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
8949 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
8950 when not giving up or versioning for alias only because of
8952 (vect_analyze_data_ref_dependences): Set to true.
8953 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
8955 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
8956 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
8960 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
8963 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
8964 operand 0 predicate for gathers, use a new pseudo as subtarget.
8966 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
8969 * lra-constraints.c (process_alt_operands): Add printing debug info.
8970 Check absence of input/output reloads for matched operands too.
8972 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
8974 PR rtl-optimization/59835
8975 * ira.c (ira_init_register_move_cost): Increase cost for
8978 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
8980 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
8982 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
8985 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
8986 non-register objects. Use gen_(high/low)part more consistently.
8989 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8992 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
8993 endian support, remove tests for WORDS_BIG_ENDIAN.
8994 (p8_mfvsrd_3_<mode>): Likewise.
8995 (reload_gpr_from_vsx<mode>): Likewise.
8996 (reload_gpr_from_vsxsf): Likewise.
8997 (p8_mfvsrd_4_disf): Likewise.
8999 2014-01-16 Richard Biener <rguenther@suse.de>
9001 PR rtl-optimization/46590
9002 * lcm.c (compute_antinout_edge): Use postorder iteration.
9003 (compute_laterin): Use inverted postorder iteration.
9005 2014-01-16 Nick Clifton <nickc@redhat.com>
9008 * varasm.c (output_constant): Return the number of bytes actually
9010 (output_constructor_array_range): Update the field size with the
9011 number of bytes emitted by output_constant.
9012 (output_constructor_regular_field): Likewise. Also do not
9013 complain if the total number of bytes emitted is now greater
9014 than the expected fieldpos.
9015 * output.h (output_constant): Update prototype and descriptive comment.
9017 2014-01-16 Marek Polacek <polacek@redhat.com>
9020 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
9021 it is error_mark_node.
9023 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
9025 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
9026 VALID_AVX256_REG_OR_OI_MODE.
9028 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
9030 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
9031 current procedure should be profiled.
9033 2014-01-15 Andrew Pinski <apinski@cavium.com>
9035 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
9036 of moving from/to the STACK_REG register class.
9038 2014-01-15 Richard Henderson <rth@redhat.com>
9041 * reginfo.c (global_regs_decl): Globalize.
9042 * rtl.h (global_regs_decl): Declare.
9043 * ira.c (do_reload): Diagnose frame_pointer_needed and it
9044 reserved via global_regs.
9046 2014-01-15 Teresa Johnson <tejohnson@google.com>
9048 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
9050 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9052 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
9053 and vmulosh rather than call gen_vec_widen_smult_*.
9054 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
9055 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
9056 (vec_widen_smult_even_v16qi): Likewise.
9057 (vec_widen_umult_even_v8hi): Likewise.
9058 (vec_widen_smult_even_v8hi): Likewise.
9059 (vec_widen_umult_odd_v16qi): Likewise.
9060 (vec_widen_smult_odd_v16qi): Likewise.
9061 (vec_widen_umult_odd_v8hi): Likewise.
9062 (vec_widen_smult_odd_v8hi): Likewise.
9063 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
9064 vmuloub rather than call gen_vec_widen_umult_*.
9065 (vec_widen_umult_lo_v16qi): Likewise.
9066 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
9067 vmulosb rather than call gen_vec_widen_smult_*.
9068 (vec_widen_smult_lo_v16qi): Likewise.
9069 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
9070 rather than call gen_vec_widen_umult_*.
9071 (vec_widen_umult_lo_v8hi): Likewise.
9072 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
9073 rather than call gen_vec_widen_smult_*.
9074 (vec_widen_smult_lo_v8hi): Likewise.
9076 2014-01-15 Jeff Law <law@redhat.com>
9078 PR tree-optimization/59747
9079 * ree.c (find_and_remove_re): Properly handle case where a second
9080 eliminated extension requires widening a copy created for elimination
9081 of a prior extension.
9082 (combine_set_extension): Ensure that the number of hard regs needed
9083 for a destination register does not change when we widen it.
9085 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
9087 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
9088 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
9089 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
9090 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
9091 (avr-*-rtems*): Likewise.
9092 (bfin*-rtems*): Likewise.
9093 (moxie-*-rtems*): Likewise.
9094 (h8300-*-rtems*): Likewise.
9095 (i[34567]86-*-rtems*): Likewise.
9096 (lm32-*-rtems*): Likewise.
9097 (m32r-*-rtems*): Likewise.
9098 (m68k-*-rtems*): Likewise.
9099 (microblaze*-*-rtems*): Likewise.
9100 (mips*-*-rtems*): Likewise.
9101 (powerpc-*-rtems*): Likewise.
9102 (sh-*-rtems*): Likewise.
9103 (sparc-*-rtems*): Likewise.
9104 (sparc64-*-rtems*): Likewise.
9105 (v850-*-rtems*): Likewise.
9106 (m32c-*-rtems*): Likewise.
9108 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9110 PR rtl-optimization/59511
9111 * ira.c (ira_init_register_move_cost): Use memory costs for some
9112 cases of register move cost calculations.
9113 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9114 instead of BB frequency.
9115 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9116 * lra-assigns.c (find_hard_regno_for): Ditto.
9118 2014-01-15 Richard Biener <rguenther@suse.de>
9120 PR tree-optimization/59822
9121 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9122 (vectorizable_load): Use it to hoist defs of uses of invariant
9123 loads out of the loop.
9125 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9126 Kugan Vivekanandarajah <kuganv@linaro.org>
9129 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9132 2014-01-15 Richard Biener <rguenther@suse.de>
9134 PR rtl-optimization/59802
9135 * lcm.c (compute_available): Use inverted postorder to seed
9136 the initial worklist.
9138 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9141 * config/s390/s390.c (s390_preferred_reload_class): Don't return
9142 ADDR_REGS for invalid symrefs in non-PIC code.
9144 2014-01-15 Jakub Jelinek <jakub@redhat.com>
9147 * builtins.c (determine_block_size): Initialize *probable_max_size
9148 even if len_rtx is CONST_INT.
9150 2014-01-14 Andrew Pinski <apinski@cavium.com>
9152 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
9153 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
9154 (cortexa53_tunings): Likewise.
9155 (aarch64_sched_issue_rate): New function.
9156 (TARGET_SCHED_ISSUE_RATE): Define.
9158 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9160 * ira-costs.c (find_costs_and_classes): Add missed
9161 ira_init_register_move_cost_if_necessary.
9163 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9166 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
9168 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
9171 * config/i386/i386.c (type_natural_mode): Add a bool parameter
9172 to indicate if type is used for function return value. Warn ABI
9173 change if the vector mode isn't available for function return value.
9174 (ix86_function_arg_advance): Pass false to type_natural_mode.
9175 (ix86_function_arg): Likewise.
9176 (ix86_gimplify_va_arg): Likewise.
9177 (function_arg_32): Don't warn ABI change.
9178 (ix86_function_value): Pass true to type_natural_mode.
9179 (ix86_return_in_memory): Likewise.
9180 (ix86_struct_value_rtx): Removed.
9181 (TARGET_STRUCT_VALUE_RTX): Likewise.
9183 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9185 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
9186 converting a conditional jump into a conditional return.
9188 2014-01-14 Richard Biener <rguenther@suse.de>
9190 PR tree-optimization/58921
9191 PR tree-optimization/59006
9192 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
9193 hoisting invariant stmts.
9194 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
9195 invariant loads on the preheader edge if possible.
9197 2014-01-14 Joey Ye <joey.ye@arm.com>
9199 * doc/plugin.texi (Building GCC plugins): Update to C++.
9201 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
9203 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
9204 (_mm_rcp28_round_ss): Ditto.
9205 (_mm_rsqrt28_round_sd): Ditto.
9206 (_mm_rsqrt28_round_ss): Ditto.
9207 (_mm_rcp28_sd): Ditto.
9208 (_mm_rcp28_ss): Ditto.
9209 (_mm_rsqrt28_sd): Ditto.
9210 (_mm_rsqrt28_ss): Ditto.
9211 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
9212 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
9213 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
9214 (IX86_BUILTIN_RCP28SD): Ditto.
9215 (IX86_BUILTIN_RCP28SS): Ditto.
9216 (IX86_BUILTIN_RSQRT28SD): Ditto.
9217 (IX86_BUILTIN_RSQRT28SS): Ditto.
9218 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
9219 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
9220 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
9221 (ix86_expand_special_args_builtin): Expand new FTYPE.
9222 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
9223 (srcp14<mode>): Make insn unary.
9224 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
9225 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
9226 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
9227 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
9228 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
9229 Fix rounding: make it SAE only.
9230 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9232 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9234 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
9235 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
9236 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
9237 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
9238 (round_saeonly_mask_scalar_operand4): Ditto.
9239 (round_saeonly_mask_scalar_op3): Ditto.
9240 (round_saeonly_mask_scalar_op4): Ditto.
9242 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9244 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9245 Implement -maltivec=be for vec_insert and vec_extract.
9247 2014-01-10 DJ Delorie <dj@redhat.com>
9249 * config/msp430/msp430.md (call_internal): Don't allow memory
9250 references with SP as the base register.
9251 (call_value_internal): Likewise.
9252 * config/msp430/constraints.md (Yc): New. For memory references
9253 that don't use SP as a base register.
9255 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
9256 "an integer without a # prefix"
9257 * config/msp430/msp430.md (epilogue_helper): Use it.
9259 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9262 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
9263 AVX512F gather builtins.
9264 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
9265 on gather decls with INTEGER_TYPE masktype.
9266 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
9267 directly into the builtin rather than hoisting it before loop.
9269 PR tree-optimization/59387
9270 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
9271 (scev_const_prop): If folded_casts and type has undefined overflow,
9272 use force_gimple_operand instead of force_gimple_operand_gsi and
9273 for each added stmt if it is assign with
9274 arith_code_with_undefined_signed_overflow, call
9275 rewrite_to_defined_overflow.
9276 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
9277 gimple-fold.h instead.
9278 (arith_code_with_undefined_signed_overflow,
9279 rewrite_to_defined_overflow): Moved to ...
9280 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
9281 rewrite_to_defined_overflow): ... here. No longer static.
9282 Include gimplify-me.h.
9283 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
9284 rewrite_to_defined_overflow): New prototypes.
9286 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9288 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
9290 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
9292 * builtins.c (get_object_alignment_2): Minor tweak.
9293 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
9295 2014-01-13 Christian Bruel <christian.bruel@st.com>
9297 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
9298 optimized non constant lengths.
9300 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9303 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
9304 load as __atomic_load_N if possible.
9306 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9308 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
9310 (rs6000_expand_builtin): Adjust call.
9312 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9315 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
9316 * config/rs6000/rs6000.c: Include target-globals.h.
9317 (rs6000_set_current_function): Instead of doing target_reinit
9318 unconditionally, use save_target_globals_default_opts and
9319 restore_target_globals.
9321 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
9323 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
9324 (rs6000_expand_builtin): Handle mffs and mtfsf.
9325 (rs6000_init_builtins): Define mffs and mtfsf.
9326 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
9327 (rs6000_mffs): New pattern.
9328 (rs6000_mtfsf): New pattern.
9330 2014-01-11 Bin Cheng <bin.cheng@arm.com>
9332 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
9333 Start narrowing with START. Apply candidate-use pair
9334 and check overall cost in narrowing.
9335 (iv_ca_prune): Pass new argument.
9337 2014-01-10 Jeff Law <law@redhat.com>
9340 * ree.c (combine_reaching_defs): Ensure the defining statement
9341 occurs before the extension when optimizing extensions with
9342 different source and destination hard registers.
9344 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9347 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
9348 vtables into the type inheritance graph.
9350 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9352 PR rtl-optimization/59754
9353 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
9354 modes in the REGNO != REGNO case.
9356 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9358 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
9360 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9362 PR tree-optimization/59745
9363 * tree-predcom.c (tree_predictive_commoning_loop): Call
9364 free_affine_expand_cache if giving up because components is NULL.
9366 * target-globals.c (save_target_globals): Allocate < 4KB structs using
9367 GC in payload of target_globals struct instead of allocating them on
9368 the heap and the larger structs separately using GC.
9369 * target-globals.h (struct target_globals): Make regs, hard_regs,
9370 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
9371 of GTY((skip)) and change type to void *.
9372 (reset_target_globals): Cast loads from those fields to corresponding
9375 2014-01-10 Steve Ellcey <sellcey@mips.com>
9378 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
9379 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
9380 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
9382 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
9385 * aarch64-modes.def (CC_Zmode): New flags mode.
9386 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
9387 represents an equality.
9388 (aarch64_get_condition_code): Handle CC_Zmode.
9389 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
9391 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9393 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
9394 extraction in good case.
9396 2014-01-10 Richard Biener <rguenther@suse.de>
9398 PR tree-optimization/59374
9399 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
9400 checking after SLP discovery. Mark stmts not participating
9401 in any SLP instance properly.
9403 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9405 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
9406 when handling a SET rtx.
9408 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9410 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
9411 (cortex-a57): Likewise.
9412 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
9414 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9416 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
9417 non-iwmmxt builtins.
9419 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9423 * ipa-devirt.c record_target_from_binfo): Take as argument
9424 stack of binfos and lookup matching one for virtual inheritance.
9425 (possible_polymorphic_call_targets_1): Update.
9427 2014-01-10 Huacai Chen <chenhc@lemote.com>
9429 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
9430 kernel strings for Loongson-2E/2F/3A.
9432 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9435 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
9436 is_gimple_call before calling gimple_call_internal_p.
9438 2014-01-09 Steve Ellcey <sellcey@mips.com>
9440 * Makefile.in (TREE_FLOW_H): Remove.
9441 (TREE_SSA_H): Add file names from tree-flow.h.
9442 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
9443 * tree.h: Remove tree-flow.h reference.
9444 * hash-table.h: Remove tree-flow.h reference.
9445 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
9446 reference with tree-ssa-loop.h.
9448 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9450 * doc/invoke.texi: Add -maltivec={be,le} options, and document
9451 default element-order behavior for -maltivec.
9452 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
9453 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
9454 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
9455 when targeting big endian, at least for now.
9456 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
9458 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9461 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
9462 var satisfies use_register_for_decl, just take into account type
9463 alignment, rather than decl alignment.
9465 PR tree-optimization/59622
9466 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
9467 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
9468 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
9469 Don't devirtualize for inplace at all. For targets.length () == 1,
9470 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
9472 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9474 * config/i386/i386.md (cpu): Remove the unused btver1.
9476 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9478 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
9480 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9483 * tree-core.h (struct target_globals): New forward declaration.
9484 (struct tree_target_option): Add globals field.
9485 * tree.h (TREE_TARGET_GLOBALS): Define.
9486 (prepare_target_option_nodes_for_pch): New prototype.
9487 * target-globals.h (struct target_globals): Define even if
9489 * tree.c (prepare_target_option_node_for_pch,
9490 prepare_target_option_nodes_for_pch): New functions.
9491 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
9492 * config/i386/i386.c: Include target-globals.h.
9493 (ix86_set_current_function): Instead of doing target_reinit
9494 unconditionally, use save_target_globals_default_opts and
9495 restore_target_globals.
9497 2014-01-09 Richard Biener <rguenther@suse.de>
9499 PR tree-optimization/59715
9500 * tree-cfg.h (split_critical_edges): Declare.
9501 * tree-cfg.c (split_critical_edges): Export.
9502 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
9504 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
9506 * cfgexpand.c (expand_stack_vars): Optionally disable
9507 asan stack protection.
9508 (expand_used_vars): Likewise.
9509 (partition_stack_vars): Likewise.
9510 * asan.c (asan_emit_stack_protection): Optionally disable
9511 after return stack usage.
9512 (instrument_derefs): Optionally disable memory access instrumentation.
9513 (instrument_builtin_call): Likewise.
9514 (instrument_strlen_call): Likewise.
9515 (asan_protect_global): Optionally disable global variables protection.
9516 * doc/invoke.texi: Added doc for new options.
9517 * params.def: Added new options.
9518 * params.h: Likewise.
9520 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9522 PR rtl-optimization/59724
9523 * ifcvt.c (cond_exec_process_if_block): Don't call
9524 flow_find_head_matching_sequence with 0 longest_match.
9525 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
9526 non-active insns if !stop_after.
9527 (try_head_merge_bb): Revert 2014-01-07 changes.
9529 2014-01-08 Jeff Law <law@redhat.com>
9531 * ree.c (get_sub_rtx): New function, extracted from...
9532 (merge_def_and_ext): Here.
9533 (combine_reaching_defs): Use get_sub_rtx.
9535 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
9537 * cgraph.h (varpool_variable_node): Do not choke on null node.
9539 2014-01-08 Catherine Moore <clm@codesourcery.com>
9541 * config/mips/mips.md (simple_return): Attempt to use JRC
9543 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
9545 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9547 PR rtl-optimization/59137
9548 * reorg.c (steal_delay_list_from_target): Call update_block for
9550 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
9552 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9554 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
9555 two duplicate entries.
9557 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9560 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
9562 * config/mips/mips.c (mips_truncated_op_cost): New function.
9563 (mips_rtx_costs): Adjust test for BADDU.
9564 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
9566 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
9568 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
9569 (*baddu_si): ...this new pattern.
9571 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9574 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
9576 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9579 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
9581 (expand_expr, expand_normal): Adjust.
9582 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
9583 inner_reference_p. Use inner_reference_p to expand inner references.
9584 (store_expr): Adjust.
9585 * cfgexpand.c (expand_call_stmt): Adjust.
9587 2014-01-08 Rong Xu <xur@google.com>
9589 * gcov-io.c (gcov_var): Move from gcov-io.h.
9590 (gcov_position): Ditto.
9591 (gcov_is_error): Ditto.
9592 (gcov_rewrite): Ditto.
9593 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
9594 only part to libgcc/libgcov.h.
9596 2014-01-08 Marek Polacek <polacek@redhat.com>
9599 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
9601 2014-01-08 Marek Polacek <polacek@redhat.com>
9604 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
9606 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9608 PR rtl-optimization/59649
9609 * stor-layout.c (get_mode_bounds): For BImode return
9610 0 and STORE_FLAG_VALUE.
9612 2014-01-08 Richard Biener <rguenther@suse.de>
9615 * gimple.h (is_gimple_builtin_call): Remove.
9616 (gimple_builtin_call_types_compatible_p): New.
9617 (gimple_call_builtin_p): New overload.
9618 * gimple.c (is_gimple_builtin_call): Remove.
9619 (validate_call): Rename to ...
9620 (gimple_builtin_call_types_compatible_p): ... this and export. Also
9622 (validate_type): New static function.
9623 (gimple_call_builtin_p): New overload and adjust.
9624 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
9625 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
9626 (gimple_fold_stmt_to_constant_1): Likewise.
9627 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
9629 2014-01-08 Richard Biener <rguenther@suse.de>
9632 * gimplify.c (gimplify_expr): Gimplify register-register type
9633 VIEW_CONVERT_EXPRs to separate stmts.
9635 2014-01-07 Jeff Law <law@redhat.com>
9638 * ree.c (combine_set_extension): Handle case where source
9639 and destination registers in an extension insn are different.
9640 (combine_reaching_defs): Allow source and destination registers
9641 in extension to be different under limited circumstances.
9642 (add_removable_extension): Remove restriction that the
9643 source and destination registers in the extension are the same.
9644 (find_and_remove_re): Emit a copy from the extension's
9645 destination to its source after the defining insn if
9646 the source and destination registers are different.
9649 * ifcvt.c (merge_if_block): If we are merging a block with more than
9650 one successor with a block with no successors, remove any BARRIER
9651 after the second block.
9653 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
9655 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
9657 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
9660 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
9661 for 14-bit register offsets when INT14_OK_STRICT is false.
9663 2014-01-07 Roland Stigge <stigge@antcom.de>
9664 Michael Meissner <meissner@linux.vnet.ibm.com>
9667 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
9668 Only check TFmode for SPE constants. Don't check TImode or TDmode.
9670 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
9672 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
9675 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
9677 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
9678 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
9679 rtx is const0_rtx or not.
9681 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
9684 * target-globals.c (save_target_globals): Remove this_fn_optab
9686 * toplev.c: Include optabs.h.
9687 (target_reinit): Temporarily restore the global options if another
9688 set of options are in force.
9690 2014-01-07 Jakub Jelinek <jakub@redhat.com>
9692 PR rtl-optimization/58668
9693 * cfgcleanup.c (flow_find_cross_jump): Don't count
9694 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
9695 to determine what is counted.
9696 (flow_find_head_matching_sequence): Use active_insn_p to determine
9698 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
9700 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
9701 determine what is counted.
9703 PR tree-optimization/59643
9704 * tree-predcom.c (split_data_refs_to_components): If one dr is
9705 read and one write, determine_offset fails and the write isn't
9706 in the bad component, just put the read into the bad component.
9708 2014-01-07 Mike Stump <mikestump@comcast.net>
9709 Jakub Jelinek <jakub@redhat.com>
9712 * tree-core.h (struct tree_optimization_option): Change optabs
9713 type from unsigned char * to void *.
9714 * optabs.c (init_tree_optimization_optabs): Adjust
9715 TREE_OPTIMIZATION_OPTABS initialization.
9717 2014-01-06 Jakub Jelinek <jakub@redhat.com>
9720 * config/i386/i386.h (struct machine_function): Add
9721 no_drap_save_restore field.
9722 * config/i386/i386.c (ix86_save_reg): Use
9723 !cfun->machine->no_drap_save_restore instead of
9724 crtl->stack_realign_needed.
9725 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
9726 this function clears frame_pointer_needed. Set
9727 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
9728 and DRAP reg is needed.
9730 2014-01-06 Marek Polacek <polacek@redhat.com>
9733 * doc/implement-c.texi: Mention that other integer types are
9734 permitted as bit-field types in strictly conforming mode.
9736 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
9738 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
9741 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
9743 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
9745 2014-01-06 Martin Jambor <mjambor@suse.cz>
9748 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
9750 * ipa-prop.c (ipa_print_node_params): Fix indentation.
9752 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
9756 * var-tracking.c (add_stores): Preserve the value of the source even if
9757 we don't record the store.
9759 2014-01-06 Terry Guo <terry.guo@arm.com>
9761 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
9763 2014-01-05 Iain Sandoe <iain@codesourcery.com>
9766 * config/darwin.c (darwin_function_section): Adjust return values to
9767 correspond to optimisation changes made in r206070.
9769 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
9771 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
9772 from prefetch_block tune setting.
9773 (nocona_cost): Correct size of prefetch block to 64.
9775 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
9777 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
9778 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
9779 used to save the static chain register in the computation of the offset
9780 from which the FP registers need to be restored.
9782 2014-01-04 Jakub Jelinek <jakub@redhat.com>
9784 PR tree-optimization/59519
9785 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
9786 ICE if get_current_def (current_new_name) is already non-NULL, as long
9787 as it is a phi result of some other phi in *new_exit_bb that has
9790 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
9791 or vmovdqu* for misaligned_operand.
9792 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
9793 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
9794 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
9795 aligned_mem for AVX512F masked aligned load and store builtins and for
9798 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
9800 PR tree-optimization/59651
9801 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
9802 Address range for negative step should be added by TYPE_SIZE_UNIT.
9804 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
9806 * config/m68k/m68k.c (handle_move_double): Handle pushes with
9807 overlapping registers also for registers other than the stack pointer.
9809 2014-01-03 Marek Polacek <polacek@redhat.com>
9812 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
9815 2014-01-03 Jakub Jelinek <jakub@redhat.com>
9818 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
9821 * config/i386/i386.md (MODE_SIZE): New mode attribute.
9822 (push splitter): Use <P:MODE_SIZE> instead of
9823 GET_MODE_SIZE (<P:MODE>mode).
9824 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
9825 (mov -1, reg peephole2): Likewise.
9826 * config/i386/sse.md (*mov<mode>_internal,
9827 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
9828 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
9829 *<code><mode>3, *andnot<mode>3<mask_name>,
9830 <mask_codefor><code><mode>3<mask_name>): Likewise.
9831 * config/i386/subst.md (mask_mode512bit_condition,
9832 sd_mask_mode512bit_condition): Likewise.
9834 2014-01-02 Xinliang David Li <davidxl@google.com>
9836 PR tree-optimization/59303
9837 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
9838 (dump_predicates): Better output format.
9839 (pred_equal_p): New function.
9840 (is_neq_relop_p): Ditto.
9841 (is_neq_zero_form_p): Ditto.
9842 (pred_expr_equal_p): Ditto.
9843 (pred_neg_p): Ditto.
9844 (simplify_pred): Ditto.
9845 (simplify_preds_2): Ditto.
9846 (simplify_preds_3): Ditto.
9847 (simplify_preds_4): Ditto.
9848 (simplify_preds): Ditto.
9850 (push_to_worklist): Ditto.
9851 (get_pred_info_from_cmp): Ditto.
9852 (is_degenerated_phi): Ditto.
9853 (normalize_one_pred_1): Ditto.
9854 (normalize_one_pred): Ditto.
9855 (normalize_one_pred_chain): Ditto.
9856 (normalize_preds): Ditto.
9857 (normalize_cond_1): Remove function.
9858 (normalize_cond): Ditto.
9859 (is_gcond_subset_of): Ditto.
9860 (is_subset_of_any): Ditto.
9861 (is_or_set_subset_of): Ditto.
9862 (is_and_set_subset_of): Ditto.
9863 (is_norm_cond_subset_of): Ditto.
9864 (pred_chain_length_cmp): Ditto.
9865 (convert_control_dep_chain_into_preds): Type change.
9866 (find_predicates): Ditto.
9867 (find_def_preds): Ditto.
9868 (destroy_predicates_vecs): Ditto.
9869 (find_matching_predicates_in_rest_chains): Ditto.
9870 (use_pred_not_overlap_with_undef_path_pred): Ditto.
9871 (is_pred_expr_subset): Ditto.
9872 (is_pred_chain_subset_of): Ditto.
9873 (is_included_in): Ditto.
9874 (is_superset_of): Ditto.
9876 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
9878 Update copyright years.
9880 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
9882 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
9883 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
9884 config/arc/arc.md, config/arc/arc.opt,
9885 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
9886 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
9887 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
9888 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
9889 config/linux-protos.h, config/linux.c, config/winnt-c.c,
9890 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
9891 vtable-verify.c, vtable-verify.h: Use the standard form for the
9894 2014-01-02 Tobias Burnus <burnus@net-b.de>
9896 * gcc.c (process_command): Update copyright notice dates.
9897 * gcov-dump.c: Ditto.
9899 * doc/cpp.texi: Bump @copying's copyright year.
9900 * doc/cppinternals.texi: Ditto.
9901 * doc/gcc.texi: Ditto.
9902 * doc/gccint.texi: Ditto.
9903 * doc/gcov.texi: Ditto.
9904 * doc/install.texi: Ditto.
9905 * doc/invoke.texi: Ditto.
9907 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9909 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
9911 2014-01-01 Jakub Jelinek <jakub@redhat.com>
9913 * config/i386/sse.md (*mov<mode>_internal): Guard
9914 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
9916 PR rtl-optimization/59647
9917 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
9918 new_rtx into UNSIGNED_FLOAT rtxes.
9920 Copyright (C) 2014 Free Software Foundation, Inc.
9922 Copying and distribution of this file, with or without modification,
9923 are permitted in any medium without royalty provided the copyright
9924 notice and this notice are preserved.