1 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
3 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
4 Emit an error when the function has arguments.
6 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
8 * cfgloop.h (unswitch_loops): Remove.
9 * doc/passes.texi: Remove references to loop-unswitch.c
10 * timevar.def (TV_LOOP_UNSWITCH): Remove.
12 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
14 * tree-vect-data-refs.c (vect_grouped_load_supported): New
15 check for loads group of length 3.
16 (vect_permute_load_chain): New permutations for loads group of
18 * tree-vect-stmts.c (vect_model_load_cost): Change cost
19 of vec_perm_shuffle for the new permutations.
21 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
23 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
24 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
25 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
26 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
27 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
28 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
29 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
30 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
32 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
34 * loop-unswitch.c: Delete.
36 2014-05-07 Richard Biener <rguenther@suse.de>
38 * config.gcc: Always set need_64bit_hwint to yes.
40 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
42 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
43 of using optimize_size.
45 2014-05-06 Mike Stump <mikestump@comcast.net>
47 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
49 2014-05-06 Joseph Myers <joseph@codesourcery.com>
51 * config/i386/sse.md (*mov<mode>_internal)
52 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
53 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
54 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
55 (*<code><mode>3, *andnot<mode>3<mask_name>)
56 (<mask_codefor><code><mode>3<mask_name>): Only consider
57 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
59 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
62 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
64 * lra-constraints.c (valid_address_p): Move earlier in file.
65 Add a constraint argument to the address_info version.
66 (satisfies_memory_constraint_p): New function.
67 (satisfies_address_constraint_p): Likewise.
68 (process_alt_operands, curr_insn_transform): Use them.
69 (process_address): Pass the constraint to valid_address_p when
70 checking address operands.
72 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
74 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
75 to their respective blocks. Fix inadvertent use of "node".
77 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
79 * emit-rtl.c (init_derived_machine_modes): New functionm, split
81 (init_emit_once): ...here.
82 * rtl.h (init_derived_machine_modes): Declare.
83 * toplev.c (do_compile): Call it even if no_backend.
85 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
86 Mike Stump <mikestump@comcast.net>
87 Richard Sandiford <rdsandiford@googlemail.com>
88 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
90 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
91 (rtx_equal_for_memref_p): Update comment.
92 (adjust_offset_for_component_ref): Use wide-int interfaces.
93 * builtins.c (get_object_alignment_2): Likewise.
94 (c_readstr): Likewise.
95 (target_char_cast): Add comment.
96 (determine_block_size): Use wide-int interfaces.
97 (expand_builtin_signbit): Likewise.
98 (fold_builtin_int_roundingfn): Likewise.
99 (fold_builtin_bitop): Likewise.
100 (fold_builtin_bswap): Likewise.
101 (fold_builtin_logarithm): Use signop.
102 (fold_builtin_pow): Likewise.
103 (fold_builtin_memory_op): Use wide-int interfaces.
104 (fold_builtin_object_size): Likewise.
105 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
106 nb_iterations_estimate.
107 (record_niter_bound): Use wide-int interfaces.
108 (get_estimated_loop_iterations_int): Likewise.
109 (get_estimated_loop_iterations): Likewise.
110 (get_max_loop_iterations): Likewise.
111 * cfgloop.h: Include wide-int.h.
112 (struct nb_iter_bound): Change bound to widest_int.
113 (struct loop): Change nb_iterations_upper_bound and
114 nb_iterations_estimate to widest_int.
115 (record_niter_bound): Switch to use widest_int.
116 (get_estimated_loop_iterations): Likewise.
117 (get_max_loop_iterations): Likewise.
118 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
120 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
121 * combine.c (try_combine): Likewise.
122 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
123 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
125 (aarch64_float_const_representable_p): Likewise.
126 * config/arc/arc.c: Include wide-int.h.
127 (arc_can_use_doloop_p): Use wide-int interfaces.
128 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
129 (vfp3_const_double_index): Likewise.
130 * config/avr/avr.c (avr_out_round): Likewise.
131 (avr_fold_builtin): Likewise.
132 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
133 (bfin_can_use_doloop_p): Likewise.
134 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
135 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
136 * config/i386/i386.c: Include wide-int.h.
137 (ix86_data_alignment): Use wide-int interfaces.
138 (ix86_local_alignment): Likewise.
139 (ix86_emit_swsqrtsf): Update real_from_integer.
140 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
141 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
142 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
143 (zero_constant): Likewise.
144 (input_operand): Likewise.
145 (splat_input_operand): Likewise.
146 (non_logical_cint_operand): Change const_double to const_wide_int.
147 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
148 (easy_altivec_constant): Remove comment.
149 (paired_expand_vector_init): Use CONSTANT_P.
150 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
151 (rs6000_emit_move): Update checks.
152 (rs6000_aggregate_candidate): Use wide-int interfaces.
153 (rs6000_expand_ternop_builtin): Likewise.
154 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
155 (rs6000_assemble_integer): Likewise.
156 (rs6000_hash_constant): Likewise.
157 (output_toc): Likewise.
158 (rs6000_rtx_costs): Likewise.
159 (rs6000_emit_swrsqrt); Update call to real_from_integer.
160 * config/rs6000/rs6000-c.c: Include wide-int.h.
161 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
162 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
163 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
164 Handle CONST_WIDE_INT.
165 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
166 Use tree_fits_uhwi_p.
167 * config/sparc/sparc.c: Include wide-int.h.
168 (sparc_fold_builtin): Use wide-int interfaces.
169 * config/vax/vax.c: Include wide-int.h.
170 (vax_float_literal): Use real_from_integer.
171 * coretypes.h (struct hwivec_def): New.
174 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
175 (equiv_constant): Handle CONST_WIDE_INT.
176 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
177 (cselib_hash_rtx): Handle CONST_WIDE_INT.
178 * dbxout.c (stabstr_U): Use wide-int interfaces.
179 (dbxout_type): Update to use cst_fits_shwi_p.
180 * defaults.h (LOG2_BITS_PER_UNIT): Define.
181 (TARGET_SUPPORTS_WIDE_INT): Add default.
182 * dfp.c: Include wide-int.h.
183 (decimal_real_to_integer2): Use wide-int interfaces and rename to
184 decimal_real_to_integer.
185 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
186 decimal_real_to_integer.
187 * doc/generic.texi (Constant expressions): Update for wide_int.
188 * doc/rtl.texi (const_double): Likewise.
189 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
190 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
191 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
192 (REAL_VALUE_FROM_INT): Remove.
193 (TARGET_SUPPORTS_WIDE_INT): New.
194 * doc/tm.texi: Regenerate.
195 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
196 * double-int.h: Include wide-int.h.
197 (struct wi::int_traits): New.
198 * dwarf2out.c (get_full_len): New.
199 (dw_val_equal_p): Add case dw_val_class_wide_int.
200 (size_of_loc_descr): Likewise.
201 (output_loc_operands): Likewise.
202 (insert_double): Remove.
203 (insert_wide_int): New.
205 (print_die): Add case dw_val_class_wide_int.
206 (attr_checksum): Likewise.
207 (attr_checksum_ordered): Likewise.
208 (same_dw_val_p): Likewise.
209 (size_of_die): Likewise.
210 (value_format): Likewise.
211 (output_die): Likewise.
212 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
214 (clz_loc_descriptor): Use wide-int interfaces.
215 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
216 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
217 (round_up_to_align): Use wide-int interfaces.
218 (field_byte_offset): Likewise.
219 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
220 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
221 CONST_DOUBLE handling. Use wide-int interfaces.
222 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
223 (gen_enumeration_type_die): Use add_AT_wide.
224 (hash_loc_operands): Add case dw_val_class_wide_int.
225 (compare_loc_operands): Likewise.
226 * dwarf2out.h: Include wide-int.h.
228 (enum dw_val_class): Add dw_val_class_wide_int.
229 (struct dw_val_struct): Add val_wide.
230 * emit-rtl.c (const_wide_int_htab): New.
231 (const_wide_int_htab_hash): New.
232 (const_wide_int_htab_eq): New.
233 (lookup_const_wide_int): New.
234 (const_double_htab_hash): Use wide-int interfaces.
235 (const_double_htab_eq): Likewise.
236 (rtx_to_double_int): Conditionally compile for wide-int.
237 (immed_double_int_const): Rename to immed_wide_int_const and
239 (immed_double_const): Conditionally compile for wide-int.
240 (init_emit_once): Use wide-int interfaces.
241 * explow.c (plus_constant): Likewise.
242 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
243 (lshift_value): Use wide-int interfaces.
244 (expand_mult): Likewise.
245 (choose_multiplier): Likewise.
246 (expand_smod_pow2): Likewise.
247 (make_tree): Likewise.
248 * expr.c (convert_modes): Consolidate handling of constants.
249 Use wide-int interfaces.
250 (emit_group_load_1): Add note.
251 (store_expr): Update comment.
252 (get_inner_reference): Use wide-int interfaces.
253 (expand_constructor): Update comment.
254 (expand_expr_real_2): Use wide-int interfaces.
255 (expand_expr_real_1): Likewise.
256 (reduce_to_bit_field_precision): Likewise.
257 (const_vector_from_tree): Likewise.
258 * final.c: Include wide-int-print.h.
259 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
260 * fixed-value.c: Include wide-int.h.
261 (fixed_from_string): Use wide-int interfaces.
262 (fixed_to_decimal): Likewise.
263 (fixed_convert_from_real): Likewise.
264 (real_convert_from_fixed): Likewise.
265 * fold-const.h (mem_ref_offset): Return an offset_int.
266 (div_if_zero_remainder): Remove code parameter.
267 * fold-const.c (div_if_zero_remainder): Remove code parameter.
268 Use wide-int interfaces.
269 (may_negate_without_overflow_p): Use wide-int interfaces.
270 (negate_expr_p): Likewise.
271 (fold_negate_expr): Likewise.
272 (int_const_binop_1): Likewise.
273 (const_binop): Likewise.
274 (fold_convert_const_int_from_int): Likewise.
275 (fold_convert_const_int_from_real): Likewise.
276 (fold_convert_const_int_from_fixed): Likewise.
277 (fold_convert_const_fixed_from_int): Likewise.
278 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
279 (sign_bit_p): Use wide-int interfaces.
280 (make_range_step): Likewise.
281 (build_range_check): Likewise. Pass an integer of the correct type
282 instead of using integer_one_node.
283 (range_predecessor): Pass an integer of the correct type instead
284 of using integer_one_node.
285 (range_successor): Likewise.
286 (merge_ranges): Likewise.
287 (unextend): Use wide-int interfaces.
288 (extract_muldiv_1): Likewise.
289 (fold_div_compare): Likewise.
290 (fold_single_bit_test): Likewise.
291 (fold_sign_changed_comparison): Likewise.
292 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
293 (fold_plusminus_mult_expr): Use wide-int interfaces.
294 (native_encode_int): Likewise.
295 (native_interpret_int): Likewise.
296 (fold_unary_loc): Likewise.
297 (pointer_may_wrap_p): Likewise.
298 (size_low_cst): Likewise.
299 (mask_with_tz): Likewise.
300 (fold_binary_loc): Likewise.
301 (fold_ternary_loc): Likewise.
302 (multiple_of_p): Likewise.
303 (tree_call_nonnegative_warnv_p): Update calls to
304 tree_int_cst_min_precision and real_from_integer.
305 (fold_negate_const): Use wide-int interfaces.
306 (fold_abs_const): Likewise.
307 (fold_relational_const): Use tree_int_cst_lt.
308 (round_up_loc): Use wide-int interfaces.
309 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
310 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
311 * gengtype.c: Remove include of double-int.h.
312 (do_typedef): Use wide-int interfaces.
313 (open_base_files): Add wide-int.h.
314 (main): Add offset_int and widest_int typedefs.
315 * gengtype-lex.l: Handle "^".
316 (CXX_KEYWORD): Add "static".
317 * gengtype-parse.c (require3): New.
318 (require_template_declaration): Handle constant template arguments
319 and nested templates.
320 * gengtype-state.c: Don't include "double-int.h".
321 * genpreds.c (write_one_predicate_function): Update comment.
322 (write_tm_constrs_h): Add check for hval and lval use in
324 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
325 (add_to_sequence): Likewise.
326 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
327 and const_double_operand.
328 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
330 * gimple-fold.c (get_base_constructor): Likewise.
331 (fold_array_ctor_reference): Likewise.
332 (fold_nonarray_ctor_reference): Likewise.
333 (fold_const_aggregate_ref_1): Likewise.
334 (gimple_val_nonnegative_real_p): Likewise.
335 (gimple_fold_indirect_ref): Likewise.
336 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
337 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
338 (struct slsr_cand_d): Change index to be widest_int.
339 (struct incr_info_d): Change incr to be widest_int.
340 (alloc_cand_and_find_basis): Use wide-int interfaces.
341 (slsr_process_phi): Likewise.
342 (backtrace_base_for_ref): Likewise. Return a widest_int.
343 (restructure_reference): Take a widest_int instead of a double_int.
344 (slsr_process_ref): Use wide-int interfaces.
345 (create_mul_ssa_cand): Likewise.
346 (create_mul_imm_cand): Likewise.
347 (create_add_ssa_cand): Likewise.
348 (create_add_imm_cand): Take a widest_int instead of a double_int.
349 (slsr_process_add): Use wide-int interfaces.
350 (slsr_process_cast): Likewise.
351 (slsr_process_copy): Likewise.
352 (dump_candidate): Likewise.
353 (dump_incr_vec): Likewise.
354 (replace_ref): Likewise.
355 (cand_increment): Likewise. Return a widest_int.
356 (cand_abs_increment): Likewise.
357 (replace_mult_candidate): Take a widest_int instead of a double_int.
358 (replace_unconditional_candidate): Use wide-int interfaces.
359 (incr_vec_index): Take a widest_int instead of a double_int.
360 (create_add_on_incoming_edge): Likewise.
361 (create_phi_basis): Use wide-int interfaces.
362 (replace_conditional_candidate): Likewise.
363 (record_increment): Take a widest_int instead of a double_int.
364 (record_phi_increments): Use wide-int interfaces.
365 (phi_incr_cost): Take a widest_int instead of a double_int.
366 (lowest_cost_path): Likewise.
367 (total_savings): Likewise.
368 (analyze_increments): Use wide-int interfaces.
369 (ncd_with_phi): Take a widest_int instead of a double_int.
370 (ncd_of_cand_and_phis): Likewise.
371 (nearest_common_dominator_for_cands): Likewise.
372 (insert_initializers): Use wide-int interfaces.
373 (all_phi_incrs_profitable): Likewise.
374 (replace_one_candidate): Likewise.
375 (replace_profitable_candidates): Likewise.
376 * godump.c: Include wide-int-print.h.
377 (go_output_typedef): Use wide-int interfaces.
378 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
379 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
380 (build_loop_iteration_domains): Likewise.
381 * hooks.h: Include wide-int.h rather than double-int.h.
382 (hook_bool_dint_dint_uint_bool_true): Delete.
383 (hook_bool_wint_wint_uint_bool_true): Declare.
384 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
385 (hook_bool_wint_wint_uint_bool_true): New.
386 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
388 (ubsan_expand_si_overflow_mul_check): Likewise.
389 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
390 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
391 (get_ancestor_addr_info): Likewise.
392 (ipa_modify_call_arguments): Likewise.
393 * loop-doloop.c (doloop_modify): Likewise.
394 (doloop_optimize): Likewise.
395 * loop-iv.c (iv_number_of_iterations): Likewise.
396 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
397 (unroll_loop_constant_iterations): Likewise.
398 (decide_unroll_runtime_iterations): Likewise.
399 (unroll_loop_runtime_iterations): Likewise.
400 (decide_peel_simple): Likewise.
401 (decide_unroll_stupid): Likewise.
402 * lto-streamer-in.c (streamer_read_wi): Add.
403 (input_cfg): Use wide-int interfaces.
404 (lto_input_tree_1): Likewise.
405 * lto-streamer-out.c (streamer_write_wi): Add.
406 (hash_tree): Use wide-int interfaces.
407 (output_cfg): Likewise.
408 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
409 (GTFILES): Add wide-int.h and signop.h.
410 (TAGS): Look for .cc files too.
411 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
412 * optabs.c (expand_subword_shift): Likewise.
413 (expand_doubleword_shift): Likewise.
414 (expand_absneg_bit): Likewise.
415 (expand_copysign_absneg): Likewise.
416 (expand_copysign_bit): Likewise.
417 * postreload.c (reload_cse_simplify_set): Likewise.
418 * predict.c (predict_iv_comparison): Likewise.
419 * pretty-print.h: Include wide-int-print.h.
421 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
422 * print-tree.c: Include wide-int-print.h.
423 (print_node_brief): Use wide-int interfaces.
424 (print_node): Likewise.
425 * read-rtl.c (validate_const_wide_int): New.
426 (read_rtx_code): Add CONST_WIDE_INT case.
427 * real.c: Include wide-int.h.
428 (real_to_integer2): Delete.
429 (real_to_integer): New function, returning a wide_int.
430 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
431 (ten_to_ptwo): Update call to real_from_integer.
432 (real_digit): Likewise.
433 * real.h: Include signop.h, wide-int.h and insn-modes.h.
434 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
435 (REAL_VALUE_TO_INT): Delete.
436 (real_to_integer): Declare a wide-int form.
437 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
438 * recog.c (const_int_operand): Improve comment.
439 (const_scalar_int_operand): New.
440 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
441 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
442 (split_double): Likewise.
443 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
444 (rtx_size): Likewise.
445 (rtx_alloc_stat_v): New.
446 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
447 (cwi_output_hex): New.
448 (iterative_hash_rtx): Handle CONST_WIDE_INT.
449 (cwi_check_failed_bounds): New.
450 * rtl.def (CONST_WIDE_INT): New.
451 * rtl.h: Include <utility> and wide-int.h.
452 (struct hwivec_def): New.
453 (CWI_GET_NUM_ELEM): New.
454 (CWI_PUT_NUM_ELEM): New.
455 (struct rtx_def): Add num_elem and hwiv.
456 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
457 (CASE_CONST_UNIQUE): Likewise.
458 (CASE_CONST_ANY): Likewise.
459 (CONST_SCALAR_INT_P): Likewise.
460 (CONST_WIDE_INT_P): New.
463 (cwi_check_failed_bounds): New.
466 (CONST_WIDE_INT_VEC) New.
467 (CONST_WIDE_INT_NUNITS) New.
468 (CONST_WIDE_INT_ELT) New.
469 (rtx_mode_t): New type.
470 (wi::int_traits <rtx_mode_t>): New.
472 (wi::min_value): New.
473 (wi::max_value): New.
475 (const_wide_int_alloc): New.
476 (immed_wide_int_const): New.
477 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
478 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
479 * signop.h: New file.
480 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
481 (simplify_const_unary_operation): Use wide-int interfaces.
482 (simplify_binary_operation_1): Likewise.
483 (simplify_const_binary_operation): Likewise.
484 (simplify_const_relational_operation): Likewise.
485 (simplify_immed_subreg): Likewise.
486 * stmt.c (expand_case): Likewise.
487 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
488 signop rather than a bool.
489 * stor-layout.c (layout_type): Use wide-int interfaces.
490 (initialize_sizetypes): Update calls to
491 set_min_and_max_values_for_integral_type.
492 (set_min_and_max_values_for_integral_type): Take a signop rather
493 than a bool. Use wide-int interfaces.
494 (fixup_signed_type): Update accordingly. Remove
495 HOST_BITS_PER_DOUBLE_INT limit.
496 (fixup_unsigned_type): Likewise.
497 * system.h (STATIC_CONSTANT_P): New.
498 (STATIC_ASSERT): New.
499 * target.def (can_use_doloop_p): Take widest_ints rather than
501 * target.h: Include wide-int.h rather than double-int.h.
502 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
504 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
505 rather than INT_CST_LT_UNSIGNED.
506 (can_use_doloop_if_innermost): Take widest_ints rather than
508 * tree-affine.c: Include wide-int-print.h.
509 (double_int_ext_for_comb): Delete.
510 (wide_int_ext_for_comb): New.
511 (aff_combination_zero): Use wide-int interfaces.
512 (aff_combination_const): Take a widest_int instead of a double_int.
513 (aff_combination_elt): Use wide-int interfaces.
514 (aff_combination_scale): Take a widest_int instead of a double_int.
515 (aff_combination_add_elt): Likewise.
516 (aff_combination_add_cst): Likewise.
517 (aff_combination_add): Use wide-int interfaces.
518 (aff_combination_convert): Likewise.
519 (tree_to_aff_combination): Likewise.
520 (add_elt_to_tree): Take a widest_int instead of a double_int.
521 (aff_combination_to_tree): Use wide-int interfaces.
522 (aff_combination_remove_elt): Likewise.
523 (aff_combination_add_product): Take a widest_int instead of
525 (aff_combination_mult): Use wide-int interfaces.
526 (aff_combination_expand): Likewise.
527 (double_int_constant_multiple_p): Delete.
528 (wide_int_constant_multiple_p): New.
529 (aff_combination_constant_multiple_p): Take a widest_int pointer
530 instead of a double_int pointer.
531 (print_aff): Use wide-int interfaces.
532 (get_inner_reference_aff): Take a widest_int pointer
533 instead of a double_int pointer.
534 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
535 * tree-affine.h: Include wide-int.h.
536 (struct aff_comb_elt): Change type of coef to widest_int.
537 (struct affine_tree_combination): Change type of offset to widest_int.
538 (double_int_ext_for_comb): Delete.
539 (wide_int_ext_for_comb): New.
540 (aff_combination_const): Use widest_int instead of double_int.
541 (aff_combination_scale): Likewise.
542 (aff_combination_add_elt): Likewise.
543 (aff_combination_constant_multiple_p): Likewise.
544 (get_inner_reference_aff): Likewise.
545 (aff_comb_cannot_overlap_p): Likewise.
546 (aff_combination_zero_p): Use wide-int interfaces.
547 * tree.c: Include tree.h.
548 (init_ttree): Use make_int_cst.
549 (tree_code_size): Removed code for INTEGER_CST case.
550 (tree_size): Add INTEGER_CST case.
551 (make_node_stat): Update comment.
552 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
553 (build_int_cst_type): Use wide-int interfaces.
554 (double_int_to_tree): Likewise.
555 (double_int_fits_to_tree_p): Delete.
556 (force_fit_type_double): Delete.
557 (force_fit_type): New.
558 (int_cst_hash_hash): Use wide-int interfaces.
559 (int_cst_hash_eq): Likewise.
560 (build_int_cst_wide): Delete.
561 (wide_int_to_tree): New.
562 (cache_integer_cst): Use wide-int interfaces.
563 (build_low_bits_mask): Likewise.
564 (cst_and_fits_in_hwi): Likewise.
565 (real_value_from_int_cst): Likewise.
566 (make_int_cst_stat): New.
567 (integer_zerop): Use wide_int interfaces.
568 (integer_onep): Likewise.
569 (integer_all_onesp): Likewise.
570 (integer_pow2p): Likewise.
571 (integer_nonzerop): Likewise.
572 (tree_log2): Likewise.
573 (tree_floor_log2): Likewise.
574 (tree_ctz): Likewise.
575 (int_size_in_bytes): Likewise.
576 (mem_ref_offset): Return an offset_int rather than a double_int.
577 (build_type_attribute_qual_variant): Use wide_int interfaces.
578 (type_hash_eq): Likewise
579 (tree_int_cst_equal): Likewise.
580 (tree_int_cst_lt): Delete.
581 (tree_int_cst_compare): Likewise.
582 (tree_fits_shwi_p): Use wide_int interfaces.
583 (tree_fits_uhwi_p): Likewise.
584 (tree_int_cst_sign_bit): Likewise.
585 (tree_int_cst_sgn): Likewise.
586 (tree_int_cst_min_precision): Take a signop rather than a bool.
587 (simple_cst_equal): Use wide_int interfaces.
588 (compare_tree_int): Likewise.
589 (iterative_hash_expr): Likewise.
590 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
592 (get_type_static_bounds): Use wide_int interfaces.
593 (tree_int_cst_elt_check_failed): New.
594 (build_common_tree_nodes): Reordered to set prec before filling in
596 (int_cst_value): Check cst_and_fits_in_hwi.
597 (widest_int_cst_value): Use wide_int interfaces.
598 (upper_bound_in_type): Likewise.
599 (lower_bound_in_type): Likewise.
600 (num_ending_zeros): Likewise.
601 (drop_tree_overflow): Likewise.
602 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
603 (gen_conditions_for_pow_cst_base): Likewise.
604 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
605 (group_case_labels_stmt): Use wide-int interfaces.
606 (verify_gimple_assign_binary): Likewise.
607 (print_loop): Likewise.
608 * tree-chrec.c (tree_fold_binomial): Likewise.
609 * tree-core.h (struct tree_base): Add int_length.
610 (struct tree_int_cst): Change rep of value.
611 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
612 (dr_may_alias_p): Likewise.
613 (max_stmt_executions_tree): Likewise.
614 * tree.def (INTEGER_CST): Update comment.
615 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
616 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
617 * tree-dump.c: Include wide-int.h and wide-int-print.h.
618 (dequeue_and_dump): Use wide-int interfaces.
619 * tree.h: Include wide-int.h.
620 (NULL_TREE): Moved to earlier loc in file.
621 (TREE_INT_CST_ELT_CHECK): New.
622 (tree_int_cst_elt_check_failed): New.
624 (TREE_INT_CST): Delete.
625 (TREE_INT_CST_LOW): Use wide-int interfaces.
626 (TREE_INT_CST_HIGH): Delete.
627 (TREE_INT_CST_NUNITS): New.
628 (TREE_INT_CST_EXT_NUNITS): Likewise.
629 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
630 (TREE_INT_CST_ELT): Likewise.
631 (INT_CST_LT): Delete.
632 (tree_int_cst_elt_check): New (two forms).
633 (type_code_size): Update comment.
634 (make_int_cst_stat, make_int_cst): New.
635 (tree_to_double_int): Delete.
636 (double_int_fits_to_tree_p): Delete.
637 (force_fit_type_double): Delete.
638 (build_int_cstu): Replace with out-of-line function.
639 (build_int_cst_wide): Delete.
640 (tree_int_cst_lt): Define inline.
641 (tree_int_cst_le): New.
642 (tree_int_cst_compare): Define inline.
643 (tree_int_cst_min_precision): Take a signop rather than a bool.
644 (wi::int_traits <const_tree>): New.
645 (wi::int_traits <tree>): New.
646 (wi::extended_tree): New.
647 (wi::int_traits <wi::extended_tree>): New.
648 (wi::to_widest): New.
649 (wi::to_offset): New.
650 (wi::fits_to_tree_p): New.
651 (wi::min_value): New.
652 (wi::max_value): New.
653 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
654 (copy_tree_body_r): Likewise.
655 * tree-object-size.c (compute_object_offset): Likewise.
656 (addr_object_size): Likewise.
657 * tree-predcom.c: Include wide-int-print.h.
658 (struct dref_d): Change type of offset to widest_int.
659 (dump_dref): Call wide-int printer.
660 (aff_combination_dr_offset): Use wide-int interfaces.
661 (determine_offset): Take a widest_int pointer rather than a
663 (split_data_refs_to_components): Use wide-int interfaces.
664 (suitable_component_p): Likewise.
665 (order_drefs): Likewise.
666 (add_ref_to_chain): Likewise.
667 (valid_initializer_p): Likewise.
668 (determine_roots_comp): Likewise.
669 * tree-pretty-print.c: Include wide-int-print.h.
670 (dump_generic_node): Use wide-int interfaces.
671 * tree-sra.c (sra_ipa_modify_expr): Likewise.
672 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
673 (move_fixed_address_to_symbol): Likewise.
674 (move_hint_to_base): Likewise.
675 (move_pointer_to_base): Likewise.
676 (move_variant_to_index): Likewise.
677 (most_expensive_mult_to_index): Likewise.
678 (addr_to_parts): Likewise.
679 (copy_ref_info): Likewise.
680 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
681 (indirect_refs_may_alias_p): Likewise.
682 (stmt_kills_ref_p_1): Likewise.
683 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
684 * tree-ssa-ccp.c: Update comment at top of file. Include
686 (struct prop_value_d): Change type of mask to widest_int.
687 (extend_mask): New function.
688 (dump_lattice_value): Use wide-int interfaces.
689 (get_default_value): Likewise.
690 (set_constant_value): Likewise.
691 (set_value_varying): Likewise.
692 (valid_lattice_transition): Likewise.
693 (set_lattice_value): Likewise.
694 (value_to_double_int): Delete.
695 (value_to_wide_int): New.
696 (get_value_from_alignment): Use wide-int interfaces.
697 (get_value_for_expr): Likewise.
698 (do_dbg_cnt): Likewise.
699 (ccp_finalize): Likewise.
700 (ccp_lattice_meet): Likewise.
701 (bit_value_unop_1): Use widest_ints rather than double_ints.
702 (bit_value_binop_1): Likewise.
703 (bit_value_unop): Use wide-int interfaces.
704 (bit_value_binop): Likewise.
705 (bit_value_assume_aligned): Likewise.
706 (evaluate_stmt): Likewise.
707 (ccp_fold_stmt): Likewise.
708 (visit_cond_stmt): Likewise.
709 (ccp_visit_stmt): Likewise.
710 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
711 (constant_pointer_difference): Likewise.
712 (associate_pointerplus): Likewise.
713 (combine_conversions): Likewise.
714 * tree-ssa-loop.h: Include wide-int.h.
715 (struct tree_niter_desc): Change type of max to widest_int.
716 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
717 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
718 (remove_redundant_iv_tests): Likewise.
719 (canonicalize_loop_induction_variables): Likewise.
720 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
721 (constant_multiple_of): Take a widest_int pointer instead of
722 a double_int pointer.
723 (get_computation_aff): Use wide-int interfaces.
724 (ptr_difference_cost): Likewise.
725 (difference_cost): Likewise.
726 (get_loop_invariant_expr_id): Likewise.
727 (get_computation_cost_at): Likewise.
728 (iv_elimination_compare_lt): Likewise.
729 (may_eliminate_iv): Likewise.
730 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
731 instead of double_int.
732 (max_loop_iterations): Likewise.
733 (max_stmt_executions): Likewise.
734 (estimated_stmt_executions): Likewise.
735 * tree-ssa-loop-niter.c: Include wide-int-print.h.
736 (split_to_var_and_offset): Use wide-int interfaces.
737 (determine_value_range): Likewise.
738 (bound_difference_of_offsetted_base): Likewise.
739 (bounds_add): Take a widest_int instead of a double_int.
740 (number_of_iterations_ne_max): Use wide-int interfaces.
741 (number_of_iterations_ne): Likewise.
742 (number_of_iterations_lt_to_ne): Likewise.
743 (assert_loop_rolls_lt): Likewise.
744 (number_of_iterations_lt): Likewise.
745 (number_of_iterations_le): Likewise.
746 (number_of_iterations_cond): Likewise.
747 (number_of_iterations_exit): Likewise.
748 (finite_loop_p): Likewise.
749 (derive_constant_upper_bound_assign): Likewise.
750 (derive_constant_upper_bound): Return a widest_int.
751 (derive_constant_upper_bound_ops): Likewise.
752 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
753 (record_estimate): Take a widest_int rather than a double_int.
754 (record_nonwrapping_iv): Use wide-int interfaces.
755 (double_int_cmp): Delete.
757 (bound_index): Take a widest_int rather than a double_int.
758 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
759 (maybe_lower_iteration_bound): Likewise.
760 (estimate_numbers_of_iterations_loop): Likewise.
761 (estimated_loop_iterations): Take a widest_int pointer than than
762 a double_int pointer.
763 (estimated_loop_iterations_int): Use wide-int interfaces.
764 (max_loop_iterations): Take a widest_int pointer than than
765 a double_int pointer.
766 (max_loop_iterations_int): Use wide-int interfaces.
767 (max_stmt_executions): Take a widest_int pointer than than
768 a double_int pointer.
769 (estimated_stmt_executions): Likewise.
770 (n_of_executions_at_most): Use wide-int interfaces.
771 (scev_probably_wraps_p): Likewise.
772 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
774 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
776 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
777 double_ints. Adjust for trailing_wide_ints <3> representation.
778 (set_nonzero_bits): Likewise.
779 (get_range_info): Return wide_ints rather than double_ints.
780 Adjust for trailing_wide_ints <3> representation.
781 (get_nonzero_bits): Likewise.
782 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
784 * tree-ssanames.h (struct range_info_def): Replace min, max and
785 nonzero_bits with a trailing_wide_ints <3>.
786 (set_range_info): Use wide_int_refs rather than double_ints.
787 (set_nonzero_bits): Likewise.
788 (get_range_info): Return wide_ints rather than double_ints.
789 (get_nonzero_bits): Likewise.
790 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
791 * tree-ssa-pre.c (phi_translate_1): Likewise.
792 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
793 (acceptable_pow_call): Likewise.
794 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
796 (vn_reference_fold_indirect): Likewise.
797 (vn_reference_maybe_forwprop_address): Likewise.
798 (valueize_refs_1): Likewise.
799 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
800 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
801 tree_int_cst_lt and tree_int_cst_le.
802 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
804 (streamer_alloc_tree): Likewise.
805 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
806 (streamer_write_tree_header): Likewise.
807 (streamer_write_integer_cst): Likewise.
808 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
809 (build_constructors): Likewise.
810 (array_value_type): Likewise.
811 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
812 (vect_check_gather): Likewise.
813 * tree-vect-generic.c (build_replicated_const): Likewise.
814 (expand_vector_divmod): Likewise.
815 * tree-vect-loop.c (vect_transform_loop): Likewise.
816 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
817 (vect_do_peeling_for_alignment): Likewise.
818 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
819 * tree-vrp.c: Include wide-int.h.
820 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
821 (extract_range_from_assert): Use wide-int interfaces.
822 (vrp_int_const_binop): Likewise.
823 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
825 (ranges_from_anti_range): Use wide-int interfaces.
826 (quad_int_cmp): Delete.
827 (quad_int_pair_sort): Likewise.
828 (extract_range_from_binary_expr_1): Use wide-int interfaces.
829 (extract_range_from_unary_expr_1): Likewise.
830 (adjust_range_with_scev): Likewise.
831 (masked_increment): Take and return wide_ints rather than double_ints.
832 (register_edge_assert_for_2): Use wide-int interfaces.
833 (check_array_ref): Likewise.
834 (search_for_addr_array): Likewise.
835 (maybe_set_nonzero_bits): Likewise.
836 (union_ranges): Pass an integer of the correct type instead of
837 using integer_one_node.
838 (intersect_ranges): Likewise.
839 (simplify_truth_ops_using_ranges): Likewise.
840 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
841 (range_fits_type_p): Likewise.
842 (simplify_cond_using_ranges): Likewise. Take a signop rather than
844 (simplify_conversion_using_ranges): Use wide-int interfaces.
845 (simplify_float_conversion_using_ranges): Likewise.
846 (vrp_finalize): Likewise.
847 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
848 (gimple_stringops_transform): Likewise.
849 * varasm.c (decode_addr_const): Likewise.
850 (const_hash_1): Likewise.
851 (const_rtx_hash_1): Likewise
852 (output_constant): Likewise.
853 (array_size_for_constructor): Likewise.
854 (output_constructor_regular_field): Likewise.
855 (output_constructor_bitfield): Likewise.
856 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
857 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
859 * gencheck.c: Define BITS_PER_UNIT.
862 * wide-int-print.cc: New.
863 * wide-int-print.h: New.
865 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
867 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
869 2014-05-06 Richard Biener <rguenther@suse.de>
871 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
872 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
873 (TODO_verify_all): Adjust.
874 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
875 TODO_verify_stmts and TODO_verify_rtl_sharing.
876 * bb-reorder.c: Likewise.
877 * cfgexpand.c: Likewise.
880 * function.c: Likewise.
881 * fwprop.c: Likewise.
883 * gimple-ssa-isolate-paths.c: Likewise.
884 * gimple-ssa-strength-reduction.c: Likewise.
885 * ipa-split.c: Likewise.
886 * loop-init.c: Likewise.
887 * loop-unroll.c: Likewise.
888 * lower-subreg.c: Likewise.
889 * modulo-sched.c: Likewise.
890 * postreload-gcse.c: Likewise.
891 * predict.c: Likewise.
893 * sched-rgn.c: Likewise.
894 * store-motion.c: Likewise.
895 * tracer.c: Likewise.
896 * trans-mem.c: Likewise.
897 * tree-call-cdce.c: Likewise.
898 * tree-cfg.c: Likewise.
899 * tree-cfgcleanup.c: Likewise.
900 * tree-complex.c: Likewise.
901 * tree-eh.c: Likewise.
902 * tree-emutls.c: Likewise.
903 * tree-if-conv.c: Likewise.
904 * tree-into-ssa.c: Likewise.
905 * tree-loop-distribution.c: Likewise.
906 * tree-object-size.c: Likewise.
907 * tree-parloops.c: Likewise.
908 * tree-pass.h: Likewise.
909 * tree-sra.c: Likewise.
910 * tree-ssa-ccp.c: Likewise.
911 * tree-ssa-copy.c: Likewise.
912 * tree-ssa-copyrename.c: Likewise.
913 * tree-ssa-dce.c: Likewise.
914 * tree-ssa-dom.c: Likewise.
915 * tree-ssa-dse.c: Likewise.
916 * tree-ssa-forwprop.c: Likewise.
917 * tree-ssa-ifcombine.c: Likewise.
918 * tree-ssa-loop-ch.c: Likewise.
919 * tree-ssa-loop-ivcanon.c: Likewise.
920 * tree-ssa-loop.c: Likewise.
921 * tree-ssa-math-opts.c: Likewise.
922 * tree-ssa-phiopt.c: Likewise.
923 * tree-ssa-phiprop.c: Likewise.
924 * tree-ssa-pre.c: Likewise.
925 * tree-ssa-reassoc.c: Likewise.
926 * tree-ssa-sink.c: Likewise.
927 * tree-ssa-strlen.c: Likewise.
928 * tree-ssa-tail-merge.c: Likewise.
929 * tree-ssa-uncprop.c: Likewise.
930 * tree-switch-conversion.c: Likewise.
931 * tree-tailcall.c: Likewise.
932 * tree-vect-generic.c: Likewise.
933 * tree-vectorizer.c: Likewise.
934 * tree-vrp.c: Likewise.
936 * var-tracking.c: Likewise.
937 * bt-load.c: Likewise.
938 * cfgcleanup.c: Likewise.
939 * combine-stack-adj.c: Likewise.
940 * combine.c: Likewise.
941 * compare-elim.c: Likewise.
942 * config/epiphany/resolve-sw-modes.c: Likewise.
943 * config/i386/i386.c: Likewise.
944 * config/mips/mips.c: Likewise.
945 * config/s390/s390.c: Likewise.
946 * config/sh/sh_treg_combine.cc: Likewise.
947 * config/sparc/sparc.c: Likewise.
952 * mode-switching.c: Likewise.
953 * passes.c: Likewise.
954 * postreload.c: Likewise.
956 * reg-stack.c: Likewise.
957 * regcprop.c: Likewise.
958 * regrename.c: Likewise.
961 2014-05-06 Richard Biener <rguenther@suse.de>
964 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
965 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
967 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
970 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
972 2014-05-05 Radovan Obradovic <robradovic@mips.com>
973 Tom de Vries <tom@codesourcery.com>
975 * target.def (call_fusage_contains_non_callee_clobbers): New DEFHOOKPOD.
976 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
978 (@node Miscellaneous Register Hooks): New node.
979 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
980 * doc/tm.texi: Regenerate.
982 2014-05-05 Marek Polacek <polacek@redhat.com>
985 * opts.c (common_handle_option): Call error_at instead of warning_at.
987 2014-05-05 Richard Biener <rguenther@suse.de>
989 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
990 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
991 under the TODO_verify_il umbrella.
993 2014-05-05 Richard Biener <rguenther@suse.de>
995 * passes.c (execute_function_todo): Move TODO_verify_flow under
996 the TODO_verify_ul umbrella.
998 2014-05-05 Richard Biener <rguenther@suse.de>
1001 * fold-const.c (fold_binary_loc): Consistently avoid
1002 canonicalizing X & CST away from a CST that is the mask
1005 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1007 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
1008 int argument to enum machine_mode.
1009 (picochip_class_max_nregs): Ditto.
1010 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
1011 (picochip_class_max_nregs): Ditto.
1013 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1015 * target.def: Add new target hook.
1016 * doc/tm.texi: Regenerate.
1017 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
1018 * targhooks.c (default_keep_leaf_when_profiled): New function.
1020 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
1021 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
1023 2014-05-05 Bin Cheng <bin.cheng@arm.com>
1025 PR tree-optimization/60363
1026 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
1027 (copy_phi_args): New parameters. Call get_value_locus_in_path.
1028 (update_destination_phis): New parameter.
1029 (create_edge_and_update_destination_phis): Ditto.
1030 (ssa_fix_duplicate_block_edges): Pass new arguments.
1031 (thread_single_edge): Ditto.
1033 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
1035 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
1036 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
1037 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
1038 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
1039 Use RS6000_BTM_HARD_FLOAT.
1040 (BU_MISC_2): Likewise.
1041 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
1042 RS6000_BTM_HARD_FLOAT.
1043 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
1045 (rs6000_invalid_builtin): Add hard floating builtin support.
1046 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
1047 hard float builtins.
1048 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
1050 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1052 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
1053 Add missing function* argument.
1055 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1057 * lra-constraints.c (valid_address_p): Move earlier in file.
1058 Add a constraint argument to the address_info version.
1059 (satisfies_memory_constraint_p): New function.
1060 (satisfies_address_constraint_p): Likewise.
1061 (process_alt_operands, curr_insn_transform): Use them.
1062 (process_address): Pass the constraint to valid_address_p when
1063 checking address operands.
1065 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1067 * config/mips/mips.c (mips_isa_rev): New variable.
1068 (mips_set_architecture): Set it.
1069 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1071 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1072 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1073 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1074 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1075 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
1076 conditions in terms of mips_isa_rev.
1077 (mips_isa_rev): Declare.
1079 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1081 * config/sh/sh-mem.cc: Use tabs instead of spaces.
1082 (prob_unlikely, prob_likely): Make variables const.
1084 2014-05-03 Denis Chertykov <chertykov@gmail.com>
1086 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
1088 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1090 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
1092 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1094 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
1095 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
1096 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
1098 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
1099 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
1101 Replace usage of ROUND_REG with sh_round_reg.
1102 Use CEIL instead of ROUND_ADVANCE.
1104 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1107 * config/sh/sh.c: Include stdlib headers before everything else.
1109 2014-05-02 Jakub Jelinek <jakub@redhat.com>
1111 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
1112 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
1113 (gimplify_adjust_omp_clauses): Simd region is never
1114 directly nested in combined parallel. Instead, for linear
1115 with copyin/copyout, if in combined for simd loop, make decl
1116 firstprivate/lastprivate on OMP_FOR.
1117 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
1118 expand_omp_for_static_chunk): When setting endvar, also set
1119 fd->loop.v to the same value.
1121 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1123 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
1125 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
1127 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
1130 2014-05-02 Marek Polacek <polacek@redhat.com>
1132 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
1134 2014-05-02 Kito Cheng <kito@0xlab.org>
1136 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
1137 to a C expression marco.
1138 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
1139 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
1140 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
1141 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
1142 HONOR_REG_ALLOC_ORDER.
1143 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
1145 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1147 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
1149 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1151 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
1153 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
1155 * tree-if-conv.c (is_cond_scalar_reduction): New function.
1156 (convert_scalar_cond_reduction): Likewise.
1157 (predicate_scalar_phi): Add recognition and transformation
1158 of simple conditioanl reduction to be vectorizable.
1160 2014-05-01 Marek Polacek <polacek@redhat.com>
1163 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
1165 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
1167 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
1168 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
1169 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
1170 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
1171 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
1172 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
1173 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
1174 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
1176 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
1178 * config/arc/arc.opt (mlra): Move comment above option name
1179 to avoid mis-parsing as language options.
1181 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1183 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
1184 * config/sol2.h: ... here.
1185 * config/sol2-10.h: Remove.
1187 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
1188 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
1189 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
1190 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
1191 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
1192 * config/sol2.h: ... here.
1193 (SECTION_NAME_FORMAT): Don't redefine.
1194 (STARTFILE_ARCH32_SPEC): Rename to ...
1195 (STARTFILE_ARCH_SPEC): ... this.
1196 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
1197 * config/sparc/sol2.h: ... here.
1198 (SECTION_NAME_FORMAT): Don't undef.
1199 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
1200 (SUBTARGET_EXTRA_SPECS): Remove.
1201 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
1203 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
1204 (MD_STARTFILE_PREFIX): Remove.
1205 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
1206 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
1207 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
1208 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
1209 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
1210 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
1211 * config/i386/sol2.h: ... here.
1212 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
1213 * config/i386/sol2-bi.h: Remove.
1214 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
1215 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
1217 * config/i386/t-sol2-64: Rename to ...
1218 * config/i386/t-sol2: ... this.
1219 * config/sparc/t-sol2-64: Rename to ...
1220 * config/sparc/t-sol2: ... this.
1222 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
1223 sol2_tm_file_head, sol2_tm_file_tail.
1224 Include ${cpu_type}/sol2.h before sol2.h.
1226 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
1227 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
1228 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
1229 Reflect i386/t-sol2-64 renaming.
1230 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
1231 Reflect sparc/t-sol2-64 renaming.
1233 2014-04-30 Richard Biener <rguenther@suse.de>
1235 * passes.c (execute_function_todo): Move TODO_verify_stmts
1236 and TODO_verify_ssa under the TODO_verify_il umbrella.
1237 * tree-ssa.h (verify_ssa): Adjust prototype.
1238 * tree-ssa.c (verify_ssa): Add parameter to tell whether
1239 we should verify SSA operands.
1240 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
1241 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
1242 whether we should verify whether not throwing stmts have EH info.
1243 * graphite-scop-detection.c (create_sese_edges): Adjust.
1244 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
1245 * tree-eh.c (lower_try_finally_switch): Do not add the
1246 default case label twice.
1248 2014-04-30 Marek Polacek <polacek@redhat.com>
1250 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
1251 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
1252 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
1253 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
1255 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
1257 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
1258 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
1259 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
1260 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
1261 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
1262 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
1263 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
1264 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
1266 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1268 * tree-cfg.c (dump_function_to_file): Dump the return type of
1269 functions, in a line to itself before the function body, mimicking
1270 the layout of a C function.
1272 2014-04-29 Jakub Jelinek <jakub@redhat.com>
1274 PR tree-optimization/60971
1275 * tree-tailcall.c (process_assignment): Reject conversions which
1278 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
1280 * calls.c (initialize_argument_information): Always treat
1281 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
1282 (expand_call): Likewise.
1283 (emit_library_call_calue_1): Likewise.
1284 * expr.c (PUSH_ARGS_REVERSED): Do not define.
1285 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
1288 2014-04-29 Nick Clifton <nickc@redhat.com>
1290 * config/msp430/msp430.md (umulsidi): Fix typo.
1291 (mulhisi3): Enable even inside interrupt handlers.
1292 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
1293 bigger return address pushed in large mode.
1295 2014-04-29 Nick Clifton <nickc@redhat.com>
1297 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
1298 (arc_init_reg_tables): Use a machine_mode enum to iterate over
1300 * config/m32r/m32r.c (init_reg_tables): Likewise.
1301 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
1302 enum to hold the modes.
1304 2014-04-29 Richard Biener <rguenther@suse.de>
1306 * dominance.c (free_dominance_info): Add overload with
1308 (dom_info_state): Likewise.
1309 (dom_info_available_p): Likewise.
1310 * basic-block.h (free_dominance_info, dom_info_state,
1311 dom_info_available_p): Declare overloads.
1312 * passes.c (execute_function_todo): Verify that verifiers
1313 don't change dominator info state. Drop dominator info
1314 for IPA pass invocations.
1315 * cgraph.c (release_function_body): Restore asserts that
1316 dominator information is released.
1318 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
1320 * doc/invoke.texi: Fix typo.
1321 * tree-vrp.c: Fix typos.
1322 * gimple.c (infer_nonnull_range): Reorder operands of an &&
1325 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1327 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
1329 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1331 * config/aarch64/aarch64-builtins.c
1332 (aarch64_types_storestruct_lane_qualifiers): New.
1333 (TYPES_STORESTRUCT_LANE): Likewise.
1334 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
1335 (st3_lane): Likewise.
1336 (st4_lane): Likewise.
1337 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
1338 (vec_store_lanesci_lane<mode>): Likewise.
1339 (vec_store_lanesxi_lane<mode>): Likewise.
1340 (aarch64_st2_lane<VQ:mode>): Likewise.
1341 (aarch64_st3_lane<VQ:mode>): Likewise.
1342 (aarch64_st4_lane<VQ:mode>): Likewise.
1343 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
1344 * config/aarch64/arm_neon.h
1345 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
1346 use new macro arguments.
1347 (__ST3_LANE_FUNC): Likewise.
1348 (__ST4_LANE_FUNC): Likewise.
1349 * config/aarch64/iterators.md (V_TWO_ELEM): New.
1350 (V_THREE_ELEM): Likewise.
1351 (V_FOUR_ELEM): Likewise.
1353 2014-04-28 David Malcolm <dmalcolm@redhat.com>
1355 * doc/gimple.texi: Replace the description of the now-defunct
1356 union gimple_statement_d with a diagram showing the
1357 gimple_statement_base class hierarchy and its relationships to
1358 the GSS_ and GIMPLE_ enums.
1360 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1362 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
1363 * config/aarch64/aarch64.c
1364 (aarch64_cannot_change_mode_class): Weaken conditions.
1365 (aarch64_modes_tieable_p): New.
1366 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
1368 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
1370 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
1371 (loadsync_<mode>): Change mode.
1372 (load_quadpti, store_quadpti): New.
1373 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
1374 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
1376 2014-04-28 Martin Jambor <mjambor@suse.cz>
1378 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
1379 same alias type as the original statement.
1380 (subreplacement_assignment_data): New type.
1381 (handle_unscalarized_data_in_subtree): New type of parameter,
1382 generate new memory accesses with same alias type as the original
1384 (load_assign_lhs_subreplacements): Likewise.
1385 (sra_modify_constructor_assign): Generate new memory accesses with
1386 same alias type as the original statement.
1388 2014-04-28 Richard Biener <rguenther@suse.de>
1390 * tree-pass.h (TODO_verify_il): Define.
1391 (TODO_verify_all): Complete properly.
1392 * passes.c (execute_function_todo): Move existing loop-closed
1393 SSA verification under TODO_verify_il.
1394 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
1395 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
1396 Fix tree sharing issue.
1398 2014-04-28 Richard Biener <rguenther@suse.de>
1401 * builtins.def (DEF_C11_BUILTIN): Add.
1402 (BUILT_IN_ALIGNED_ALLOC): Likewise.
1403 * coretypes.h (enum function_class): Add function_c11_misc.
1404 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1405 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
1406 (call_may_clobber_ref_p_1): Likewise.
1407 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
1408 (mark_all_reaching_defs_necessary_1): Likewise.
1409 (propagate_necessity): Likewise.
1410 (eliminate_unnecessary_stmts): Likewise.
1411 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
1413 2014-04-28 Richard Biener <rguenther@suse.de>
1415 * tree-vrp.c (vrp_var_may_overflow): Remove.
1416 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
1417 with overflow immediately bump to one before that value and
1418 let iteration figure out overflow status.
1420 2014-04-28 Richard Biener <rguenther@suse.de>
1422 * configure.ac: Do valgrind header checks unconditionally.
1423 Add --enable-valgrind-annotations.
1424 * system.h: Guard valgrind header inclusion with
1425 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
1426 * alloc-pool.c (pool_alloc, pool_free): Use
1427 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
1428 to guard possibly dead code.
1429 * config.in: Regenerated.
1430 * configure: Likewise.
1432 2014-04-28 Jeff Law <law@redhat.com>
1434 PR tree-optimization/60902
1435 * tree-ssa-threadedge.c
1436 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
1437 over real defs when invalidating outputs from statements that do not
1438 produce useful outputs for threading.
1440 2014-04-28 Richard Biener <rguenther@suse.de>
1442 PR tree-optimization/60979
1443 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
1444 SCOPs that end in a block with a successor with abnormal
1447 2014-04-28 Richard Biener <rguenther@suse.de>
1449 * tree-pass.h (execute_pass_list): Adjust prototype.
1450 * passes.c (pass_manager::execute_early_local_passes):
1452 (do_per_function): Change callback signature, push all actual
1453 work to the callbals.
1454 (do_per_function_toporder): Likewise.
1455 (execute_function_dump): Adjust.
1456 (execute_function_todo): Likewise.
1457 (clear_last_verified): Likewise.
1458 (verify_curr_properties): Likewise.
1459 (update_properties_after_pass): Likewise.
1460 (execute_pass_list_1): Split out from ...
1461 (execute_pass_list): ... here. Adjust.
1462 (execute_ipa_pass_list): Likewise.
1463 * cgraphunit.c (cgraph_add_new_function): Adjust.
1464 (analyze_function): Likewise.
1465 (expand_function): Likewise.
1466 * cgraph.c (release_function_body): Free dominance info
1467 here instead of asserting it was magically freed elsewhere.
1469 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
1471 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
1472 * configure: Regenerate.
1473 * config/sparc/sparc.opt (muser-mode): New option.
1474 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
1476 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
1477 * doc/invoke.texi (SPARC options): Document -muser-mode.
1479 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
1481 * cselib.c (find_slot_memmode): Delete.
1482 (cselib_hasher): Change compare_type to a struct.
1483 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
1485 (preserve_constants_and_equivs): Adjust for new compare_type.
1486 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
1487 (wrap_constant): Delete.
1488 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
1490 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
1492 * doc/install.texi (Building with profile feedback): Remove
1495 2014-04-26 Tom de Vries <tom@codesourcery.com>
1497 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
1500 2014-04-25 Cary Coutant <ccoutant@google.com>
1503 * dwarf2out.c (should_move_die_to_comdat): A type definition
1504 can contain a subprogram definition, but don't move it to a
1506 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
1507 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
1509 (clone_tree_hash): Rename to...
1510 (clone_tree_partial): ...this; change callers. Copy
1511 DW_TAG_subprogram DIEs as declarations.
1512 (copy_decls_walk): Don't copy children of a declaration into a
1515 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
1518 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
1521 2014-04-25 Jiong Wang <jiong.wang@arm.com>
1523 * config/arm/predicates.md (call_insn_operand): Add long_call check.
1524 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
1526 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
1529 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1531 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
1534 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1536 PR tree-optimization/60930
1537 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
1538 creating a multiply candidate by folding two constant
1539 multiplicands when the result overflows.
1541 2014-04-25 Jakub Jelinek <jakub@redhat.com>
1543 PR tree-optimization/60960
1544 * tree-vect-generic.c (expand_vector_operation): Only call
1545 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
1547 2014-04-25 Tom de Vries <tom@codesourcery.com>
1549 * expr.c (clobber_reg_mode): New function.
1550 * expr.h (clobber_reg): New function.
1552 2014-04-25 Tom de Vries <tom@codesourcery.com>
1554 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
1557 2014-04-25 Radovan Obradovic <robradovic@mips.com>
1558 Tom de Vries <tom@codesourcery.com>
1560 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
1562 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
1563 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
1564 new argument to find_all_hard_reg_sets call.
1566 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1568 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
1569 Use HOST_WIDE_INT_C for mask literal.
1570 (aarch_rev16_shleft_mask_imm_p): Likewise.
1572 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
1575 * config/sparc/sparc.md (ashlsi3_extend): Delete.
1577 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
1579 PR preprocessor/56540
1580 * config/i386/i386-c.c (ix86_target_macros): Define
1581 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
1583 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1585 * configure.ac (tga_func): Remove.
1586 (LIB_TLS_SPEC): Remove.
1587 * configure: Regenerate.
1588 * config.in: Regenerate.
1589 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
1591 2014-04-25 Richard Biener <rguenther@suse.de>
1594 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
1595 call stmt use/clobber sets during stmt walk instead of
1596 walking the possibly incomplete set of caller edges.
1598 2014-04-25 Richard Biener <rguenther@suse.de>
1601 * passes.c (apply_ipa_transforms): Inline into only caller ...
1602 (execute_one_pass): ... here. Properly bring in function
1603 bodies for nodes we want to apply IPA transforms to.
1605 2014-04-24 Cong Hou <congh@google.com>
1607 PR tree-optimization/60896
1608 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
1609 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
1610 (vect_mark_pattern_stmts): Set the def type of all statements in
1611 PATTERN_DEF_SEQ as vect_internal_def.
1613 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
1615 * doc/extend.texi (PowerPC Built-in Functions): Document new
1616 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
1618 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
1620 * config/rs6000/predicates.md (const_0_to_3_operand): New
1621 predicate to match 0..3 integer constants.
1623 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
1624 to support adding miscellaneous builtin functions.
1625 (BU_DFP_MISC_2): Likewise.
1626 (BU_P7_MISC_1): Likewise.
1627 (BU_P7_MISC_2): Likewise.
1628 (BU_P8V_MISC_3): Likewise.
1629 (BU_MISC_1): Likewise.
1630 (BU_MISC_2): Likewise.
1631 (DIVWE): Add extended divide builtin functions.
1634 (DIVWEUO): Likewise.
1638 (DIVDEUO): Likewise.
1639 (DXEX): Add decimal floating-point builtin functions.
1642 (DDEDPDQ): Likewise.
1644 (DENBCDQ): Likewise.
1651 (CDTBCD): Add new BCD builtin functions.
1655 (BCDADD_LT): Likewise.
1656 (BCDADD_EQ): Likewise.
1657 (BCDADD_GT): Likewise.
1658 (BCDADD_OV): Likewise.
1660 (BCDSUB_LT): Likewise.
1661 (BCDSUB_EQ): Likewise.
1662 (BCDSUB_GT): Likewise.
1663 (BCDSUB_OV): Likewise.
1664 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
1665 (UNPACK_TD): Likewise.
1666 (PACK_TF): Likewise.
1667 (UNPACK_TF): Likewise.
1668 (UNPACK_TF_0): Likewise.
1669 (UNPACK_TF_1): Likewise.
1670 (PACK_V1TI): Likewise.
1671 (UNPACK_V1TI): Likewise.
1673 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
1674 support for decimal floating point builtin functions.
1675 (rs6000_expand_ternop_builtin): Add checks for the new builtin
1676 functions that take constant arguments.
1677 (rs6000_invalid_builtin): Add decimal floating point builtin
1679 (rs6000_init_builtins): Setup long double, _Decimal64, and
1680 _Decimal128 types for new builtin functions.
1681 (builtin_function_type): Set the unsigned flags appropriately for
1682 the new builtin functions.
1683 (rs6000_opt_masks): Add support for decimal floating point builtin
1686 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
1687 floating point builtin functions.
1688 (RS6000_BTM_COMMON): Likewise.
1689 (RS6000_BTI_long_double): Likewise.
1690 (RS6000_BTI_dfloat64): Likewise.
1691 (RS6000_BTI_dfloat128): Likewise.
1692 (long_double_type_internal_node): Likewise.
1693 (dfloat64_type_internal_node): Likewise.
1694 (dfloat128_type_internal_node): Likewise.
1696 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
1697 2.07 bcd arithmetic instructions.
1698 (UNSPEC_BCDSUB): Likewise.
1699 (UNSPEC_BCD_OVERFLOW): Likewise.
1700 (UNSPEC_BCD_ADD_SUB): Likewise.
1701 (bcd_add_sub): Likewise.
1702 (BCD_TEST): Likewise.
1703 (bcd<bcd_add_sub>): Likewise.
1704 (bcd<bcd_add_sub>_test): Likewise.
1705 (bcd<bcd_add_sub>_test2): Likewise.
1706 (bcd<bcd_add_sub>_<code>): Likewise.
1707 (peephole2 for combined bcd ops): Likewise.
1709 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
1710 decimal floating point builtin functions.
1711 (UNSPEC_DENBCD): Likewise.
1712 (UNSPEC_DXEX): Likewise.
1713 (UNSPEC_DIEX): Likewise.
1714 (UNSPEC_DSCLI): Likewise.
1715 (UNSPEC_DSCRI): Likewise.
1716 (D64_D128): Likewise.
1717 (dfp_suffix): Likewise.
1718 (dfp_ddedpd_<mode>): Likewise.
1719 (dfp_denbcd_<mode>): Likewise.
1720 (dfp_dxex_<mode>): Likewise.
1721 (dfp_diex_<mode>): Likewise.
1722 (dfp_dscli_<mode>): Likewise.
1723 (dfp_dscri_<mode>): Likewise.
1725 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
1727 (UNSPEC_CDTBCD): Likewise.
1728 (UNSPEC_CBCDTD): Likewise.
1729 (UNSPEC_DIVE): Add support for new extended divide builtin
1731 (UNSPEC_DIVEO): Likewise.
1732 (UNSPEC_DIVEU): Likewise.
1733 (UNSPEC_DIVEUO): Likewise.
1734 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
1735 pack/unpack 128-bit types.
1736 (UNSPEC_PACK_128BIT): Likewise.
1737 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
1738 (udiv<mode>3): Use idiv_ldiv mode attribute.
1739 (div<mode>3): Likewise.
1740 (addg6s): Add new BCD builtin functions.
1743 (UNSPEC_DIV_EXTEND): Add support for new extended divide
1745 (div_extend): Likewise.
1746 (div<div_extend>_<mode>"): Likewise.
1747 (FP128_64): Add support for new builtin functions to pack/unpack
1749 (unpack<mode>): Likewise.
1750 (unpacktf_0): Likewise.
1751 (unpacktf_1): Likewise.
1752 (unpack<mode>_dm): Likewise.
1753 (unpack<mode>_nodm): Likewise.
1754 (pack<mode>): Likewise.
1755 (unpackv1ti): Likewise.
1756 (packv1ti): Likewise.
1758 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
1760 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
1763 2014-04-24 Jakub Jelinek <jakub@redhat.com>
1765 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
1766 * gimplify.c (omp_is_private): Change last argument's type to int.
1767 Only diagnose lastprivate if the simd argument is 1, only diagnose
1768 linear if the simd argument is 2.
1769 (gimplify_omp_for): Adjust omp_is_private callers. When adding
1770 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
1771 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
1772 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
1773 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
1774 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
1775 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
1776 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
1777 * tree-nested.c (convert_nonlocal_omp_clauses,
1778 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
1780 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
1783 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
1786 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
1788 * flag-types.h (enum ivar_visibility): Add.
1790 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
1792 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
1793 function * argument.
1795 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
1797 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
1799 2014-04-24 Radovan Obradovic <robradovic@mips.com>
1800 Tom de Vries <tom@codesourcery.com>
1802 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
1803 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
1805 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
1806 * emit-rtl.c (try_split): Same.
1808 2014-04-24 Radovan Obradovic <robradovic@mips.com>
1809 Tom de Vries <tom@codesourcery.com>
1811 * common.opt (fuse-caller-save): New option.
1813 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
1815 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
1818 2014-04-24 Richard Biener <rguenther@suse.de>
1820 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
1821 during TER and instead use the sepops interface for expanding
1822 non-GIMPLE_SINGLE_RHS.
1824 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1826 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
1827 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
1829 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1831 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
1832 assembler 64-bit option.
1833 * configure: Regenerate.
1835 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1837 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
1838 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
1839 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
1840 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
1841 (TARGET_CRYPTO): Take TARGET_SIMD into account.
1843 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1845 * config/aarch64/aarch64-builtins.c
1846 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
1847 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
1848 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
1849 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
1851 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
1852 (Vrevsuff): New mode attribute.
1854 2014-04-24 Terry Guo <terry.guo@arm.com>
1856 * config/arm/arm.h (machine_function): Define variable
1857 after_arm_reorg here.
1858 * config/arm/arm.c (after_arm_reorg): Remove the definition.
1859 (arm_split_constant): Update the way to access variable
1862 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
1864 2014-04-23 Tom de Vries <tom@codesourcery.com>
1866 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
1868 2014-04-23 David Malcolm <dmalcolm@redhat.com>
1870 * is-a.h: Update comments to reflect the following changes to the
1871 "pointerness" of the API, making the template parameter match the
1872 return type, allowing use of is-a.h with typedefs of pointers.
1873 (is_a_helper::cast): Return a T rather then a pointer to a T, so
1874 that the return type matches the parameter to the is_a_helper.
1876 (dyn_cast): Likewise.
1878 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
1879 pointer from the is-a.h API.
1881 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
1882 (is_a_helper <cgraph_node *>::test): ...this, matching change to
1884 (is_a_helper <varpool_node>::test): Likewise, convert to...
1885 (is_a_helper <varpool_node *>::test): ...this.
1887 (varpool_first_variable): Update for removal of implicit pointer
1888 from the is-a.h API.
1889 (varpool_next_variable): Likewise.
1890 (varpool_first_static_initializer): Likewise.
1891 (varpool_next_static_initializer): Likewise.
1892 (varpool_first_defined_variable): Likewise.
1893 (varpool_next_defined_variable): Likewise.
1894 (cgraph_first_defined_function): Likewise.
1895 (cgraph_next_defined_function): Likewise.
1896 (cgraph_first_function): Likewise.
1897 (cgraph_next_function): Likewise.
1898 (cgraph_first_function_with_gimple_body): Likewise.
1899 (cgraph_next_function_with_gimple_body): Likewise.
1900 (cgraph_alias_target): Likewise.
1901 (varpool_alias_target): Likewise.
1902 (cgraph_function_or_thunk_node): Likewise.
1903 (varpool_variable_node): Likewise.
1904 (symtab_real_symbol_p): Likewise.
1905 * cgraphunit.c (referred_to_p): Likewise.
1906 (analyze_functions): Likewise.
1907 (handle_alias_pairs): Likewise.
1908 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1909 * gimple-ssa.h (gimple_vuse_op): Likewise.
1910 (gimple_vdef_op): Likewise.
1911 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
1912 * gimple.c (gimple_build_asm_1): Likewise.
1913 (gimple_build_try): Likewise.
1914 (gimple_build_resx): Likewise.
1915 (gimple_build_eh_dispatch): Likewise.
1916 (gimple_build_omp_for): Likewise.
1917 (gimple_omp_for_set_clauses): Likewise.
1919 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
1920 (is_a_helper <gimple_statement_asm *>::test): ...this.
1921 (is_a_helper <gimple_statement_bind>::test): Convert to...
1922 (is_a_helper <gimple_statement_bind *>::test): ...this.
1923 (is_a_helper <gimple_statement_call>::test): Convert to...
1924 (is_a_helper <gimple_statement_call *>::test): ...this.
1925 (is_a_helper <gimple_statement_catch>::test): Convert to...
1926 (is_a_helper <gimple_statement_catch *>::test): ...this.
1927 (is_a_helper <gimple_statement_resx>::test): Convert to...
1928 (is_a_helper <gimple_statement_resx *>::test): ...this.
1929 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
1930 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
1931 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
1932 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
1933 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
1934 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
1935 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
1936 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
1937 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
1938 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
1939 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
1940 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
1941 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
1942 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
1943 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
1944 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
1945 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
1946 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
1947 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
1948 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
1949 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
1950 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
1951 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
1952 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
1953 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
1954 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
1955 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
1956 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
1957 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
1958 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
1959 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
1960 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
1961 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
1962 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
1963 (is_a_helper <gimple_statement_phi>::test): Convert to...
1964 (is_a_helper <gimple_statement_phi *>::test): ...this.
1965 (is_a_helper <gimple_statement_transaction>::test): Convert to...
1966 (is_a_helper <gimple_statement_transaction *>::test): ...this.
1967 (is_a_helper <gimple_statement_try>::test): Convert to...
1968 (is_a_helper <gimple_statement_try *>::test): ...this.
1969 (is_a_helper <gimple_statement_wce>::test): Convert to...
1970 (is_a_helper <gimple_statement_wce *>::test): ...this.
1971 (is_a_helper <const gimple_statement_asm>::test): Convert to...
1972 (is_a_helper <const gimple_statement_asm *>::test): ...this.
1973 (is_a_helper <const gimple_statement_bind>::test): Convert to...
1974 (is_a_helper <const gimple_statement_bind *>::test): ...this.
1975 (is_a_helper <const gimple_statement_call>::test): Convert to...
1976 (is_a_helper <const gimple_statement_call *>::test): ...this.
1977 (is_a_helper <const gimple_statement_catch>::test): Convert to...
1978 (is_a_helper <const gimple_statement_catch *>::test): ...this.
1979 (is_a_helper <const gimple_statement_resx>::test): Convert to...
1980 (is_a_helper <const gimple_statement_resx *>::test): ...this.
1981 (is_a_helper <const gimple_statement_eh_dispatch>::test):
1983 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
1985 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
1987 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
1988 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
1990 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
1992 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
1994 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
1996 (is_a_helper <const gimple_statement_omp_return>::test): Convert
1998 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
1999 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
2001 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
2002 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
2004 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
2005 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
2006 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
2007 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
2009 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
2010 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
2012 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
2013 (is_a_helper <const gimple_statement_omp_target>::test): Convert
2015 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
2016 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
2018 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
2019 (is_a_helper <const gimple_statement_omp_single>::test): Convert
2021 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
2022 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
2024 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
2025 (is_a_helper <const gimple_statement_omp_task>::test): Convert
2027 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
2028 (is_a_helper <const gimple_statement_phi>::test): Convert to...
2029 (is_a_helper <const gimple_statement_phi *>::test): ...this.
2030 (is_a_helper <const gimple_statement_transaction>::test): Convert
2032 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
2033 (is_a_helper <const gimple_statement_with_ops>::test): Convert
2035 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
2036 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
2037 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
2038 (is_a_helper <const gimple_statement_with_memory_ops>::test):
2040 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
2042 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
2044 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
2046 (gimple_use_ops): Update for removal of implicit pointer from the
2048 (gimple_set_use_ops): Likewise.
2049 (gimple_vuse): Likewise.
2050 (gimple_vdef): Likewise.
2051 (gimple_vuse_ptr): Likewise.
2052 (gimple_vdef_ptr): Likewise.
2053 (gimple_set_vuse): Likewise.
2054 (gimple_set_vdef): Likewise.
2055 (gimple_omp_return_set_lhs): Likewise.
2056 (gimple_omp_return_lhs): Likewise.
2057 (gimple_omp_return_lhs_ptr): Likewise.
2058 (gimple_call_fntype): Likewise.
2059 (gimple_call_set_fntype): Likewise.
2060 (gimple_call_set_internal_fn): Likewise.
2061 (gimple_call_use_set): Likewise.
2062 (gimple_call_clobber_set): Likewise.
2063 (gimple_bind_vars): Likewise.
2064 (gimple_bind_set_vars): Likewise.
2065 (gimple_bind_body_ptr): Likewise.
2066 (gimple_bind_set_body): Likewise.
2067 (gimple_bind_add_stmt): Likewise.
2068 (gimple_bind_block): Likewise.
2069 (gimple_bind_set_block): Likewise.
2070 (gimple_asm_ninputs): Likewise.
2071 (gimple_asm_noutputs): Likewise.
2072 (gimple_asm_nclobbers): Likewise.
2073 (gimple_asm_nlabels): Likewise.
2074 (gimple_asm_input_op): Likewise.
2075 (gimple_asm_input_op_ptr): Likewise.
2076 (gimple_asm_output_op): Likewise.
2077 (gimple_asm_output_op_ptr): Likewise.
2078 (gimple_asm_set_output_op): Likewise.
2079 (gimple_asm_clobber_op): Likewise.
2080 (gimple_asm_set_clobber_op): Likewise.
2081 (gimple_asm_label_op): Likewise.
2082 (gimple_asm_set_label_op): Likewise.
2083 (gimple_asm_string): Likewise.
2084 (gimple_catch_types): Likewise.
2085 (gimple_catch_types_ptr): Likewise.
2086 (gimple_catch_handler_ptr): Likewise.
2087 (gimple_catch_set_types): Likewise.
2088 (gimple_catch_set_handler): Likewise.
2089 (gimple_eh_filter_types): Likewise.
2090 (gimple_eh_filter_types_ptr): Likewise.
2091 (gimple_eh_filter_failure_ptr): Likewise.
2092 (gimple_eh_filter_set_types): Likewise.
2093 (gimple_eh_filter_set_failure): Likewise.
2094 (gimple_eh_must_not_throw_fndecl): Likewise.
2095 (gimple_eh_must_not_throw_set_fndecl): Likewise.
2096 (gimple_eh_else_n_body_ptr): Likewise.
2097 (gimple_eh_else_e_body_ptr): Likewise.
2098 (gimple_eh_else_set_n_body): Likewise.
2099 (gimple_eh_else_set_e_body): Likewise.
2100 (gimple_try_eval_ptr): Likewise.
2101 (gimple_try_cleanup_ptr): Likewise.
2102 (gimple_try_set_eval): Likewise.
2103 (gimple_try_set_cleanup): Likewise.
2104 (gimple_wce_cleanup_ptr): Likewise.
2105 (gimple_wce_set_cleanup): Likewise.
2106 (gimple_phi_capacity): Likewise.
2107 (gimple_phi_num_args): Likewise.
2108 (gimple_phi_result): Likewise.
2109 (gimple_phi_result_ptr): Likewise.
2110 (gimple_phi_set_result): Likewise.
2111 (gimple_phi_arg): Likewise.
2112 (gimple_phi_set_arg): Likewise.
2113 (gimple_resx_region): Likewise.
2114 (gimple_resx_set_region): Likewise.
2115 (gimple_eh_dispatch_region): Likewise.
2116 (gimple_eh_dispatch_set_region): Likewise.
2117 (gimple_omp_critical_name): Likewise.
2118 (gimple_omp_critical_name_ptr): Likewise.
2119 (gimple_omp_critical_set_name): Likewise.
2120 (gimple_omp_for_clauses): Likewise.
2121 (gimple_omp_for_clauses_ptr): Likewise.
2122 (gimple_omp_for_set_clauses): Likewise.
2123 (gimple_omp_for_collapse): Likewise.
2124 (gimple_omp_for_index): Likewise.
2125 (gimple_omp_for_index_ptr): Likewise.
2126 (gimple_omp_for_set_index): Likewise.
2127 (gimple_omp_for_initial): Likewise.
2128 (gimple_omp_for_initial_ptr): Likewise.
2129 (gimple_omp_for_set_initial): Likewise.
2130 (gimple_omp_for_final): Likewise.
2131 (gimple_omp_for_final_ptr): Likewise.
2132 (gimple_omp_for_set_final): Likewise.
2133 (gimple_omp_for_incr): Likewise.
2134 (gimple_omp_for_incr_ptr): Likewise.
2135 (gimple_omp_for_set_incr): Likewise.
2136 (gimple_omp_for_pre_body_ptr): Likewise.
2137 (gimple_omp_for_set_pre_body): Likewise.
2138 (gimple_omp_parallel_clauses): Likewise.
2139 (gimple_omp_parallel_clauses_ptr): Likewise.
2140 (gimple_omp_parallel_set_clauses): Likewise.
2141 (gimple_omp_parallel_child_fn): Likewise.
2142 (gimple_omp_parallel_child_fn_ptr): Likewise.
2143 (gimple_omp_parallel_set_child_fn): Likewise.
2144 (gimple_omp_parallel_data_arg): Likewise.
2145 (gimple_omp_parallel_data_arg_ptr): Likewise.
2146 (gimple_omp_parallel_set_data_arg): Likewise.
2147 (gimple_omp_task_clauses): Likewise.
2148 (gimple_omp_task_clauses_ptr): Likewise.
2149 (gimple_omp_task_set_clauses): Likewise.
2150 (gimple_omp_task_child_fn): Likewise.
2151 (gimple_omp_task_child_fn_ptr): Likewise.
2152 (gimple_omp_task_set_child_fn): Likewise.
2153 (gimple_omp_task_data_arg): Likewise.
2154 (gimple_omp_task_data_arg_ptr): Likewise.
2155 (gimple_omp_task_set_data_arg): Likewise.
2156 (gimple_omp_taskreg_clauses): Likewise.
2157 (gimple_omp_taskreg_clauses_ptr): Likewise.
2158 (gimple_omp_taskreg_set_clauses): Likewise.
2159 (gimple_omp_taskreg_child_fn): Likewise.
2160 (gimple_omp_taskreg_child_fn_ptr): Likewise.
2161 (gimple_omp_taskreg_set_child_fn): Likewise.
2162 (gimple_omp_taskreg_data_arg): Likewise.
2163 (gimple_omp_taskreg_data_arg_ptr): Likewise.
2164 (gimple_omp_taskreg_set_data_arg): Likewise.
2165 (gimple_omp_task_copy_fn): Likewise.
2166 (gimple_omp_task_copy_fn_ptr): Likewise.
2167 (gimple_omp_task_set_copy_fn): Likewise.
2168 (gimple_omp_task_arg_size): Likewise.
2169 (gimple_omp_task_arg_size_ptr): Likewise.
2170 (gimple_omp_task_set_arg_size): Likewise.
2171 (gimple_omp_task_arg_align): Likewise.
2172 (gimple_omp_task_arg_align_ptr): Likewise.
2173 (gimple_omp_task_set_arg_align): Likewise.
2174 (gimple_omp_single_clauses): Likewise.
2175 (gimple_omp_single_clauses_ptr): Likewise.
2176 (gimple_omp_single_set_clauses): Likewise.
2177 (gimple_omp_target_clauses): Likewise.
2178 (gimple_omp_target_clauses_ptr): Likewise.
2179 (gimple_omp_target_set_clauses): Likewise.
2180 (gimple_omp_target_child_fn): Likewise.
2181 (gimple_omp_target_child_fn_ptr): Likewise.
2182 (gimple_omp_target_set_child_fn): Likewise.
2183 (gimple_omp_target_data_arg): Likewise.
2184 (gimple_omp_target_data_arg_ptr): Likewise.
2185 (gimple_omp_target_set_data_arg): Likewise.
2186 (gimple_omp_teams_clauses): Likewise.
2187 (gimple_omp_teams_clauses_ptr): Likewise.
2188 (gimple_omp_teams_set_clauses): Likewise.
2189 (gimple_omp_sections_clauses): Likewise.
2190 (gimple_omp_sections_clauses_ptr): Likewise.
2191 (gimple_omp_sections_set_clauses): Likewise.
2192 (gimple_omp_sections_control): Likewise.
2193 (gimple_omp_sections_control_ptr): Likewise.
2194 (gimple_omp_sections_set_control): Likewise.
2195 (gimple_omp_for_set_cond): Likewise.
2196 (gimple_omp_for_cond): Likewise.
2197 (gimple_omp_atomic_store_set_val): Likewise.
2198 (gimple_omp_atomic_store_val): Likewise.
2199 (gimple_omp_atomic_store_val_ptr): Likewise.
2200 (gimple_omp_atomic_load_set_lhs): Likewise.
2201 (gimple_omp_atomic_load_lhs): Likewise.
2202 (gimple_omp_atomic_load_lhs_ptr): Likewise.
2203 (gimple_omp_atomic_load_set_rhs): Likewise.
2204 (gimple_omp_atomic_load_rhs): Likewise.
2205 (gimple_omp_atomic_load_rhs_ptr): Likewise.
2206 (gimple_omp_continue_control_def): Likewise.
2207 (gimple_omp_continue_control_def_ptr): Likewise.
2208 (gimple_omp_continue_set_control_def): Likewise.
2209 (gimple_omp_continue_control_use): Likewise.
2210 (gimple_omp_continue_control_use_ptr): Likewise.
2211 (gimple_omp_continue_set_control_use): Likewise.
2212 (gimple_transaction_body_ptr): Likewise.
2213 (gimple_transaction_label): Likewise.
2214 (gimple_transaction_label_ptr): Likewise.
2215 (gimple_transaction_set_body): Likewise.
2216 (gimple_transaction_set_label): Likewise.
2218 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
2219 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2220 * ipa-ref.c (ipa_record_reference): Likewise.
2221 * ipa-reference.c (analyze_function): Likewise.
2222 (ipa_reference_write_optimization_summary): Likewise.
2223 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
2224 (address_taken_from_non_vtable_p): Likewise.
2225 (comdat_can_be_unshared_p_1): Likewise.
2226 * lto-cgraph.c (lto_output_ref): Likewise.
2227 (add_references): Likewise.
2228 (compute_ltrans_boundary): Likewise.
2229 (output_symtab): Likewise.
2230 (input_ref): Likewise.
2231 (input_cgraph_1): Likewise.
2232 (output_cgraph_opt_summary): Likewise.
2233 * lto-streamer-out.c (lto_output): Likewise.
2234 (output_symbol_p): Likewise.
2235 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
2236 (lsei_start_function_in_partition): Likewise.
2237 (lsei_next_variable_in_partition): Likewise.
2238 (lsei_start_variable_in_partition): Likewise.
2239 * symtab.c (insert_to_assembler_name_hash): Likewise.
2240 (unlink_from_assembler_name_hash): Likewise.
2241 (symtab_unregister_node): Likewise.
2242 (symtab_remove_node): Likewise.
2243 (dump_symtab_node): Likewise.
2244 (verify_symtab_base): Likewise.
2245 (verify_symtab_node): Likewise.
2246 (symtab_make_decl_local): Likewise.
2247 (symtab_alias_ultimate_target): Likewise.
2248 (symtab_resolve_alias): Likewise.
2249 (symtab_get_symbol_partitioning_class): Likewise.
2250 * tree-phinodes.c (allocate_phi_node): Likewise.
2251 (reserve_phi_args_for_new_edge): Likewise.
2252 (remove_phi_args): Likewise.
2253 * varpool.c (varpool_node_for_asm): Likewise.
2254 (varpool_remove_unreferenced_decls): Likewise.
2256 2014-04-23 Jeff Law <law@redhat.com>
2258 PR tree-optimization/60902
2259 * tree-ssa-threadedge.c
2260 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
2261 invalidate outputs from statements that do not produce useful
2262 outputs for threading.
2264 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2266 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
2267 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
2268 machine descriptions for Stack Smashing Protector.
2270 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
2272 * aarch64.md (<optab>_rol<mode>3): New pattern.
2273 (<optab>_rolsi3_uxtw): Likewise.
2274 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
2276 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
2278 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
2279 (arm_cortex_a12_tune): Likewise.
2281 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2283 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
2285 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2287 * config/arm/arm.md (arm_rev16si2): New pattern.
2288 (arm_rev16si2_alt): Likewise.
2289 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
2291 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2293 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
2294 (rev16<mode>2_alt): Likewise.
2295 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
2296 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
2297 (aarch_rev16_shleft_mask_imm_p): Likewise.
2298 (aarch_rev16_p_1): Likewise.
2299 (aarch_rev16_p): Likewise.
2300 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
2301 (aarch_rev16_shright_mask_imm_p): Likewise.
2302 (aarch_rev16_shleft_mask_imm_p): Likewise.
2304 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2306 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
2307 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
2309 (cortex_a53_extra_costs): Likewise.
2310 (cortex_a57_extra_costs): Likewise.
2311 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
2312 (cortexa7_extra_costs): Likewise.
2313 (cortexa8_extra_costs): Likewise.
2314 (cortexa12_extra_costs): Likewise.
2315 (cortexa15_extra_costs): Likewise.
2316 (v7m_extra_costs): Likewise.
2317 (arm_new_rtx_costs): Handle BSWAP.
2319 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2321 * config/arm/arm.c (cortexa8_extra_costs): New table.
2322 (arm_cortex_a8_tune): New tuning struct.
2323 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
2325 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2327 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
2329 2014-04-23 Richard Biener <rguenther@suse.de>
2331 * Makefile.in (OBJS): Remove loop-unswitch.o.
2332 * tree-pass.h (make_pass_rtl_unswitch): Remove.
2333 * passes.def (pass_rtl_unswitch): Likewise.
2334 * loop-init.c (gate_rtl_unswitch): Likewise.
2335 (rtl_unswitch): Likewise.
2336 (pass_data_rtl_unswitch): Likewise.
2337 (pass_rtl_unswitch): Likewise.
2338 (make_pass_rtl_unswitch): Likewise.
2339 * rtl.h (reversed_condition): Likewise.
2340 (compare_and_jump_seq): Likewise.
2341 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
2343 * loop-unroll.c (compare_and_jump_seq): Likewise.
2345 2014-04-23 Richard Biener <rguenther@suse.de>
2347 PR tree-optimization/60903
2348 * tree-ssa-loop-im.c (analyze_memory_references): Remove
2349 commented code block.
2350 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
2351 loop flags to newly created BBs and edges.
2353 2014-04-23 Nick Clifton <nickc@redhat.com>
2355 * config/msp430/msp430.c (msp430_handle_option): Move function
2357 (msp430_option_override): Simplify mcu and mcpu option handling.
2358 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
2359 support for -mhwmult command line option.
2360 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
2361 -mhwmult command line option.
2362 (msp430_hwmult_enabled): Delete.
2363 (msp43o_output_labelref): Add support for -mhwmult command line
2365 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
2366 (umulsidi3): Likewise.
2367 * config/msp430/msp430.opt (mmcu): Add Report attribute.
2368 (mcpu, mlarge, msmall): Likewise.
2369 (mhwmult): New option.
2370 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
2372 (msp430_is_f5_mcu): Remove prototype.
2373 (msp430_use_f5_series_hwmult): Add prototype.
2374 * config/msp430/msp430-opts.h: New file.
2375 * common/config/msp430: New directory.
2376 * common/config/msp430/msp430-common.c: New file.
2377 * config.gcc (msp430): Remove target_has_targetm_common.
2378 * doc/invoke.texi: Document -mhwmult command line option.
2380 2014-04-23 Nick Clifton <nickc@redhat.com>
2382 * config/i386/cygwin.h (ENDFILE_SPEC): Include
2383 default-manifest.o if it can be found in the search path.
2384 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
2386 2014-04-23 Terry Guo <terry.guo@arm.com>
2388 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
2390 2014-04-23 Richard Biener <rguenther@suse.de>
2393 * tree-inline.c (declare_return_variable): Use mark_addressable.
2395 2014-04-23 Richard Biener <rguenther@suse.de>
2398 * loop-init.c (loop_optimizer_init): Make sure to apply
2399 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
2401 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2404 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
2406 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
2407 if flag_sanitize_undefined_trap_on_error.
2408 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
2409 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
2410 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
2411 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
2412 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
2413 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
2414 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
2415 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
2416 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
2417 * ubsan.c (ubsan_instrument_unreachable): Return
2418 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
2419 (ubsan_expand_null_ifn): Emit __builtin_trap ()
2420 if flag_sanitize_undefined_trap_on_error and
2421 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
2422 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
2423 instrument_bool_enum_load): Emit __builtin_trap () if
2424 flag_sanitize_undefined_trap_on_error and
2425 __builtin_handle_*_abort () if !flag_sanitize_recover.
2426 * doc/invoke.texi (-fsanitize-recover,
2427 -fsanitize-undefined-trap-on-error): Document.
2429 2014-04-22 Christian Bruel <christian.bruel@st.com>
2431 * config/sh/sh.md (mov<mode>): Replace movQIHI.
2432 Force immediates to SImode.
2434 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
2436 * config/nios2/nios2.md (UNSPEC_ROUND): New.
2438 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
2439 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
2440 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
2441 (nios2_fpu_insn): Add entry for round.
2442 (N2FPU_NO_ERRNO_P): Define.
2443 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
2445 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
2447 2014-04-22 Richard Henderson <rth@redhat.com>
2449 * config/aarch64/aarch64 (addti3, subti3): New expanders.
2450 (add<GPI>3_compare0): Remove leading * from name.
2451 (add<GPI>3_carryin): Likewise.
2452 (sub<GPI>3_compare0): Likewise.
2453 (sub<GPI>3_carryin): Likewise.
2454 (<su_optab>mulditi3): New expander.
2455 (multi3): New expander.
2456 (madd<GPI>): Remove leading * from name.
2458 2014-04-22 Martin Jambor <mjambor@suse.cz>
2460 * cgraphclones.c (cgraph_function_versioning): Copy
2461 ipa_transforms_to_apply instead of asserting it is empty.
2463 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
2466 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
2467 on count_exp to get mode.
2469 2014-04-22 Andrew Pinski <apinski@cavium.com>
2471 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
2472 Handle TLS for ILP32.
2473 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
2474 (tlsie_small_<mode>): this and handle PTR.
2475 (tlsie_small_sidi): New pattern.
2476 (tlsle_small): Change to an expand to handle ILP32.
2477 (tlsle_small_<mode>): New pattern.
2478 (tlsdesc_small): Rename to ...
2479 (tlsdesc_small_<mode>): this and handle PTR.
2481 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2483 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
2485 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2487 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2488 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
2489 (aarch64_types_signed_poly_qualifiers): Likewise.
2490 (aarch64_types_unsigned_signed_qualifiers): Likewise.
2491 (aarch64_types_poly_signed_qualifiers): Likewise.
2492 (TYPES_REINTERP_SS): Type macro added.
2493 (TYPES_REINTERP_SU): Likewise.
2494 (TYPES_REINTERP_SP): Likewise.
2495 (TYPES_REINTERP_US): Likewise.
2496 (TYPES_REINTERP_PS): Likewise.
2497 (aarch64_fold_builtin): New expression folding added.
2498 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
2499 Declarations removed.
2500 (REINTERP_SS): Declarations added.
2501 (REINTERP_US): Likewise.
2502 (REINTERP_PS): Likewise.
2503 (REINTERP_SU): Likewise.
2504 (REINTERP_SP): Likewise.
2505 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
2506 (vreinterpretq_p8_f64): Likewise.
2507 (vreinterpret_p16_f64): Likewise.
2508 (vreinterpretq_p16_f64): Likewise.
2509 (vreinterpret_f32_f64): Likewise.
2510 (vreinterpretq_f32_f64): Likewise.
2511 (vreinterpret_f64_f32): Likewise.
2512 (vreinterpret_f64_p8): Likewise.
2513 (vreinterpret_f64_p16): Likewise.
2514 (vreinterpret_f64_s8): Likewise.
2515 (vreinterpret_f64_s16): Likewise.
2516 (vreinterpret_f64_s32): Likewise.
2517 (vreinterpret_f64_s64): Likewise.
2518 (vreinterpret_f64_u8): Likewise.
2519 (vreinterpret_f64_u16): Likewise.
2520 (vreinterpret_f64_u32): Likewise.
2521 (vreinterpret_f64_u64): Likewise.
2522 (vreinterpretq_f64_f32): Likewise.
2523 (vreinterpretq_f64_p8): Likewise.
2524 (vreinterpretq_f64_p16): Likewise.
2525 (vreinterpretq_f64_s8): Likewise.
2526 (vreinterpretq_f64_s16): Likewise.
2527 (vreinterpretq_f64_s32): Likewise.
2528 (vreinterpretq_f64_s64): Likewise.
2529 (vreinterpretq_f64_u8): Likewise.
2530 (vreinterpretq_f64_u16): Likewise.
2531 (vreinterpretq_f64_u32): Likewise.
2532 (vreinterpretq_f64_u64): Likewise.
2533 (vreinterpret_s64_f64): Likewise.
2534 (vreinterpretq_s64_f64): Likewise.
2535 (vreinterpret_u64_f64): Likewise.
2536 (vreinterpretq_u64_f64): Likewise.
2537 (vreinterpret_s8_f64): Likewise.
2538 (vreinterpretq_s8_f64): Likewise.
2539 (vreinterpret_s16_f64): Likewise.
2540 (vreinterpretq_s16_f64): Likewise.
2541 (vreinterpret_s32_f64): Likewise.
2542 (vreinterpretq_s32_f64): Likewise.
2543 (vreinterpret_u8_f64): Likewise.
2544 (vreinterpretq_u8_f64): Likewise.
2545 (vreinterpret_u16_f64): Likewise.
2546 (vreinterpretq_u16_f64): Likewise.
2547 (vreinterpret_u32_f64): Likewise.
2548 (vreinterpretq_u32_f64): Likewise.
2550 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2552 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2553 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
2554 (vreinterpret_p8_s8): Likewise.
2555 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
2556 (vreinterpret_p8_s16): Likewise.
2557 (vreinterpret_p8_s32): Likewise.
2558 (vreinterpret_p8_s64): Likewise.
2559 (vreinterpret_p8_f32): Likewise.
2560 (vreinterpret_p8_u8): Likewise.
2561 (vreinterpret_p8_u16): Likewise.
2562 (vreinterpret_p8_u32): Likewise.
2563 (vreinterpret_p8_u64): Likewise.
2564 (vreinterpret_p8_p16): Likewise.
2565 (vreinterpretq_p8_s8): Likewise.
2566 (vreinterpretq_p8_s16): Likewise.
2567 (vreinterpretq_p8_s32): Likewise.
2568 (vreinterpretq_p8_s64): Likewise.
2569 (vreinterpretq_p8_f32): Likewise.
2570 (vreinterpretq_p8_u8): Likewise.
2571 (vreinterpretq_p8_u16): Likewise.
2572 (vreinterpretq_p8_u32): Likewise.
2573 (vreinterpretq_p8_u64): Likewise.
2574 (vreinterpretq_p8_p16): Likewise.
2575 (vreinterpret_p16_s8): Likewise.
2576 (vreinterpret_p16_s16): Likewise.
2577 (vreinterpret_p16_s32): Likewise.
2578 (vreinterpret_p16_s64): Likewise.
2579 (vreinterpret_p16_f32): Likewise.
2580 (vreinterpret_p16_u8): Likewise.
2581 (vreinterpret_p16_u16): Likewise.
2582 (vreinterpret_p16_u32): Likewise.
2583 (vreinterpret_p16_u64): Likewise.
2584 (vreinterpret_p16_p8): Likewise.
2585 (vreinterpretq_p16_s8): Likewise.
2586 (vreinterpretq_p16_s16): Likewise.
2587 (vreinterpretq_p16_s32): Likewise.
2588 (vreinterpretq_p16_s64): Likewise.
2589 (vreinterpretq_p16_f32): Likewise.
2590 (vreinterpretq_p16_u8): Likewise.
2591 (vreinterpretq_p16_u16): Likewise.
2592 (vreinterpretq_p16_u32): Likewise.
2593 (vreinterpretq_p16_u64): Likewise.
2594 (vreinterpretq_p16_p8): Likewise.
2595 (vreinterpret_f32_s8): Likewise.
2596 (vreinterpret_f32_s16): Likewise.
2597 (vreinterpret_f32_s32): Likewise.
2598 (vreinterpret_f32_s64): Likewise.
2599 (vreinterpret_f32_u8): Likewise.
2600 (vreinterpret_f32_u16): Likewise.
2601 (vreinterpret_f32_u32): Likewise.
2602 (vreinterpret_f32_u64): Likewise.
2603 (vreinterpret_f32_p8): Likewise.
2604 (vreinterpret_f32_p16): Likewise.
2605 (vreinterpretq_f32_s8): Likewise.
2606 (vreinterpretq_f32_s16): Likewise.
2607 (vreinterpretq_f32_s32): Likewise.
2608 (vreinterpretq_f32_s64): Likewise.
2609 (vreinterpretq_f32_u8): Likewise.
2610 (vreinterpretq_f32_u16): Likewise.
2611 (vreinterpretq_f32_u32): Likewise.
2612 (vreinterpretq_f32_u64): Likewise.
2613 (vreinterpretq_f32_p8): Likewise.
2614 (vreinterpretq_f32_p16): Likewise.
2615 (vreinterpret_s64_s8): Likewise.
2616 (vreinterpret_s64_s16): Likewise.
2617 (vreinterpret_s64_s32): Likewise.
2618 (vreinterpret_s64_f32): Likewise.
2619 (vreinterpret_s64_u8): Likewise.
2620 (vreinterpret_s64_u16): Likewise.
2621 (vreinterpret_s64_u32): Likewise.
2622 (vreinterpret_s64_u64): Likewise.
2623 (vreinterpret_s64_p8): Likewise.
2624 (vreinterpret_s64_p16): Likewise.
2625 (vreinterpretq_s64_s8): Likewise.
2626 (vreinterpretq_s64_s16): Likewise.
2627 (vreinterpretq_s64_s32): Likewise.
2628 (vreinterpretq_s64_f32): Likewise.
2629 (vreinterpretq_s64_u8): Likewise.
2630 (vreinterpretq_s64_u16): Likewise.
2631 (vreinterpretq_s64_u32): Likewise.
2632 (vreinterpretq_s64_u64): Likewise.
2633 (vreinterpretq_s64_p8): Likewise.
2634 (vreinterpretq_s64_p16): Likewise.
2635 (vreinterpret_u64_s8): Likewise.
2636 (vreinterpret_u64_s16): Likewise.
2637 (vreinterpret_u64_s32): Likewise.
2638 (vreinterpret_u64_s64): Likewise.
2639 (vreinterpret_u64_f32): Likewise.
2640 (vreinterpret_u64_u8): Likewise.
2641 (vreinterpret_u64_u16): Likewise.
2642 (vreinterpret_u64_u32): Likewise.
2643 (vreinterpret_u64_p8): Likewise.
2644 (vreinterpret_u64_p16): Likewise.
2645 (vreinterpretq_u64_s8): Likewise.
2646 (vreinterpretq_u64_s16): Likewise.
2647 (vreinterpretq_u64_s32): Likewise.
2648 (vreinterpretq_u64_s64): Likewise.
2649 (vreinterpretq_u64_f32): Likewise.
2650 (vreinterpretq_u64_u8): Likewise.
2651 (vreinterpretq_u64_u16): Likewise.
2652 (vreinterpretq_u64_u32): Likewise.
2653 (vreinterpretq_u64_p8): Likewise.
2654 (vreinterpretq_u64_p16): Likewise.
2655 (vreinterpret_s8_s16): Likewise.
2656 (vreinterpret_s8_s32): Likewise.
2657 (vreinterpret_s8_s64): Likewise.
2658 (vreinterpret_s8_f32): Likewise.
2659 (vreinterpret_s8_u8): Likewise.
2660 (vreinterpret_s8_u16): Likewise.
2661 (vreinterpret_s8_u32): Likewise.
2662 (vreinterpret_s8_u64): Likewise.
2663 (vreinterpret_s8_p8): Likewise.
2664 (vreinterpret_s8_p16): Likewise.
2665 (vreinterpretq_s8_s16): Likewise.
2666 (vreinterpretq_s8_s32): Likewise.
2667 (vreinterpretq_s8_s64): Likewise.
2668 (vreinterpretq_s8_f32): Likewise.
2669 (vreinterpretq_s8_u8): Likewise.
2670 (vreinterpretq_s8_u16): Likewise.
2671 (vreinterpretq_s8_u32): Likewise.
2672 (vreinterpretq_s8_u64): Likewise.
2673 (vreinterpretq_s8_p8): Likewise.
2674 (vreinterpretq_s8_p16): Likewise.
2675 (vreinterpret_s16_s8): Likewise.
2676 (vreinterpret_s16_s32): Likewise.
2677 (vreinterpret_s16_s64): Likewise.
2678 (vreinterpret_s16_f32): Likewise.
2679 (vreinterpret_s16_u8): Likewise.
2680 (vreinterpret_s16_u16): Likewise.
2681 (vreinterpret_s16_u32): Likewise.
2682 (vreinterpret_s16_u64): Likewise.
2683 (vreinterpret_s16_p8): Likewise.
2684 (vreinterpret_s16_p16): Likewise.
2685 (vreinterpretq_s16_s8): Likewise.
2686 (vreinterpretq_s16_s32): Likewise.
2687 (vreinterpretq_s16_s64): Likewise.
2688 (vreinterpretq_s16_f32): Likewise.
2689 (vreinterpretq_s16_u8): Likewise.
2690 (vreinterpretq_s16_u16): Likewise.
2691 (vreinterpretq_s16_u32): Likewise.
2692 (vreinterpretq_s16_u64): Likewise.
2693 (vreinterpretq_s16_p8): Likewise.
2694 (vreinterpretq_s16_p16): Likewise.
2695 (vreinterpret_s32_s8): Likewise.
2696 (vreinterpret_s32_s16): Likewise.
2697 (vreinterpret_s32_s64): Likewise.
2698 (vreinterpret_s32_f32): Likewise.
2699 (vreinterpret_s32_u8): Likewise.
2700 (vreinterpret_s32_u16): Likewise.
2701 (vreinterpret_s32_u32): Likewise.
2702 (vreinterpret_s32_u64): Likewise.
2703 (vreinterpret_s32_p8): Likewise.
2704 (vreinterpret_s32_p16): Likewise.
2705 (vreinterpretq_s32_s8): Likewise.
2706 (vreinterpretq_s32_s16): Likewise.
2707 (vreinterpretq_s32_s64): Likewise.
2708 (vreinterpretq_s32_f32): Likewise.
2709 (vreinterpretq_s32_u8): Likewise.
2710 (vreinterpretq_s32_u16): Likewise.
2711 (vreinterpretq_s32_u32): Likewise.
2712 (vreinterpretq_s32_u64): Likewise.
2713 (vreinterpretq_s32_p8): Likewise.
2714 (vreinterpretq_s32_p16): Likewise.
2715 (vreinterpret_u8_s8): Likewise.
2716 (vreinterpret_u8_s16): Likewise.
2717 (vreinterpret_u8_s32): Likewise.
2718 (vreinterpret_u8_s64): Likewise.
2719 (vreinterpret_u8_f32): Likewise.
2720 (vreinterpret_u8_u16): Likewise.
2721 (vreinterpret_u8_u32): Likewise.
2722 (vreinterpret_u8_u64): Likewise.
2723 (vreinterpret_u8_p8): Likewise.
2724 (vreinterpret_u8_p16): Likewise.
2725 (vreinterpretq_u8_s8): Likewise.
2726 (vreinterpretq_u8_s16): Likewise.
2727 (vreinterpretq_u8_s32): Likewise.
2728 (vreinterpretq_u8_s64): Likewise.
2729 (vreinterpretq_u8_f32): Likewise.
2730 (vreinterpretq_u8_u16): Likewise.
2731 (vreinterpretq_u8_u32): Likewise.
2732 (vreinterpretq_u8_u64): Likewise.
2733 (vreinterpretq_u8_p8): Likewise.
2734 (vreinterpretq_u8_p16): Likewise.
2735 (vreinterpret_u16_s8): Likewise.
2736 (vreinterpret_u16_s16): Likewise.
2737 (vreinterpret_u16_s32): Likewise.
2738 (vreinterpret_u16_s64): Likewise.
2739 (vreinterpret_u16_f32): Likewise.
2740 (vreinterpret_u16_u8): Likewise.
2741 (vreinterpret_u16_u32): Likewise.
2742 (vreinterpret_u16_u64): Likewise.
2743 (vreinterpret_u16_p8): Likewise.
2744 (vreinterpret_u16_p16): Likewise.
2745 (vreinterpretq_u16_s8): Likewise.
2746 (vreinterpretq_u16_s16): Likewise.
2747 (vreinterpretq_u16_s32): Likewise.
2748 (vreinterpretq_u16_s64): Likewise.
2749 (vreinterpretq_u16_f32): Likewise.
2750 (vreinterpretq_u16_u8): Likewise.
2751 (vreinterpretq_u16_u32): Likewise.
2752 (vreinterpretq_u16_u64): Likewise.
2753 (vreinterpretq_u16_p8): Likewise.
2754 (vreinterpretq_u16_p16): Likewise.
2755 (vreinterpret_u32_s8): Likewise.
2756 (vreinterpret_u32_s16): Likewise.
2757 (vreinterpret_u32_s32): Likewise.
2758 (vreinterpret_u32_s64): Likewise.
2759 (vreinterpret_u32_f32): Likewise.
2760 (vreinterpret_u32_u8): Likewise.
2761 (vreinterpret_u32_u16): Likewise.
2762 (vreinterpret_u32_u64): Likewise.
2763 (vreinterpret_u32_p8): Likewise.
2764 (vreinterpret_u32_p16): Likewise.
2765 (vreinterpretq_u32_s8): Likewise.
2766 (vreinterpretq_u32_s16): Likewise.
2767 (vreinterpretq_u32_s32): Likewise.
2768 (vreinterpretq_u32_s64): Likewise.
2769 (vreinterpretq_u32_f32): Likewise.
2770 (vreinterpretq_u32_u8): Likewise.
2771 (vreinterpretq_u32_u16): Likewise.
2772 (vreinterpretq_u32_u64): Likewise.
2773 (vreinterpretq_u32_p8): Likewise.
2774 (vreinterpretq_u32_p16): Likewise.
2776 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2778 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
2780 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
2783 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
2784 (vqnegd_s64): Likewise.
2785 (vqabs_s64): Likewise.
2786 (vqabsd_s64): Likewise.
2788 2014-04-22 Richard Henderson <rth@redhat.com>
2790 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
2791 computation to the top of the loop.
2793 2014-04-22 Renlin <renlin.li@arm.com>
2794 Jiong Wang <jiong.wang@arm.com>
2796 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
2797 * config/aarch64/aarch64.c (aarch64_layout_frame)
2798 (aarch64_initial_elimination_offset): Likewise.
2800 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
2802 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
2805 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
2807 * machmode.h (bitwise_mode_for_mode): Declare.
2808 * stor-layout.h (bitwise_type_for_mode): Likewise.
2809 * stor-layout.c (bitwise_mode_for_mode): New function.
2810 (bitwise_type_for_mode): Likewise.
2811 * builtins.c (fold_builtin_memory_op): Use it instead of
2812 int_mode_for_mode and build_nonstandard_integer_type.
2814 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2816 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
2817 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
2818 (*-*-solaris2*): Simplify.
2819 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
2820 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
2821 *-*-solaris2.9* handling.
2823 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
2825 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
2826 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
2828 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
2829 * configure: Regenerate.
2831 * config/i386/sol2-9.h: Remove.
2833 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
2834 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
2835 Remove Solaris 9 references.
2837 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
2839 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
2840 (floatuns<GPI:mode><GPF:mode>2): Remove.
2841 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
2842 and floatuns conversions.
2843 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
2844 and floatuns conversions.
2845 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
2846 (w1,w2): New mode attributes for inequal width conversions.
2848 2014-04-22 Renlin Li <Renlin.Li@arm.com>
2850 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
2851 the output asm format.
2853 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
2855 * config/aarch64/aarch64-simd.md
2856 (aarch64_cm<optab>di): Always split.
2857 (*aarch64_cm<optab>di): New.
2858 (aarch64_cmtstdi): Always split.
2859 (*aarch64_cmtstdi): New.
2861 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2863 PR tree-optimization/60823
2864 * omp-low.c (ipa_simd_modify_function_body): Go through
2865 all SSA_NAMEs and for those refering to vector arguments
2866 which are going to be replaced adjust SSA_NAME_VAR and,
2867 if it is a default definition, change it into a non-default
2868 definition assigned at the beginning of function from new_decl.
2869 (ipa_simd_modify_stmt_ops): Rewritten.
2870 * tree-dfa.c (set_ssa_default_def): When removing default def,
2871 check for NULL loc instead of NULL *loc.
2873 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2875 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
2876 restrictions on core registers for DImode values in Thumb2.
2878 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2880 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
2881 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
2883 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2885 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
2886 (*iordi_notzesidi_di): Likewise.
2887 (*iordi_notsesidi_di): Likewise.
2889 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2891 * config/arm/arm-protos.h (tune_params): New struct members.
2892 * config/arm/arm.c: Initialise tune_params per processor.
2893 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
2894 for speed, based on new tune_params.
2896 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2898 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
2900 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
2902 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
2904 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
2905 * config/aarch64/arm_neon.h (vrnd_f64): Added.
2906 (vrnda_f64): Likewise.
2907 (vrndi_f64): Likewise.
2908 (vrndm_f64): Likewise.
2909 (vrndn_f64): Likewise.
2910 (vrndp_f64): Likewise.
2911 (vrndx_f64): Likewise.
2913 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2915 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
2916 GET_MODE_SIZE argument is enum machine_mode.
2918 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2921 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
2922 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
2924 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
2927 * asan.c (asan_emit_stack_protection): Force the base to align to
2928 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
2929 appropriate bits if STRICT_ALIGNMENT.
2930 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
2932 (expand_used_vars): Leave a space in the stack frame for alignment
2933 if STRICT_ALIGNMENT.
2935 2014-04-21 David Malcolm <dmalcolm@redhat.com>
2937 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
2939 (gimple_store_p): Likewise.
2940 (gimple_assign_load_p): Likewise.
2941 (gimple_assign_cast_p): Likewise.
2942 (gimple_clobber_p): Likewise.
2944 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
2945 rather than a gimple.
2946 (gimple_assign_cast_p): Likewise.
2948 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
2951 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
2952 If mode is DDmode and TARGET_E500_DOUBLE allow move.
2954 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
2955 more debug information for E500 if -mdebug=reg.
2957 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
2960 * config/i386/i386.c (ix86_expand_builtin)
2961 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
2962 register for target RTX.
2963 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
2965 2014-04-18 Cong Hou <congh@google.com>
2967 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
2968 the widen-mult pattern by handling two operands with different sizes,
2969 and operands whose size is smaller than half of the result type.
2971 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
2973 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
2974 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
2975 (do_estimate_edge_time): Compute it.
2976 * ipa-inline.c (want_inline_small_function_p): Bypass
2977 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
2979 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
2981 * ipa-inline.c (spec_rem): New static variable.
2982 (dump_overall_stats): New function.
2983 (dump_inline_stats): New function.
2985 2014-04-18 Richard Henderson <rth@redhat.com>
2987 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
2988 to GET_MODE_SIZE, not a reg_class_t.
2990 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2992 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
2993 (vsx_xxmrglw_<mode>): Likewise.
2995 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2998 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
2999 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
3000 (rs6000_init_hard_regno_mode_ok): Likewise.
3002 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
3004 * ipa-inline.c (inline_small_functions): Account only non-cold
3006 * doc/invoke.texi (inline-unit-growth): Update documentation.
3008 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
3010 * config/rs6000/rs6000.md (addti3, subti3): New.
3012 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
3015 * config/i386/i386.c (ix86_expand_clear): Remove outdated
3016 comment. Check optimize_insn_for_size_p instead of
3017 optimize_insn_for_speed_p.
3019 2014-04-17 Martin Jambor <mjambor@suse.cz>
3021 * gimple-iterator.c (gsi_start_edge): New function.
3022 * gimple-iterator.h (gsi_start_edge): Declare.
3023 * tree-sra.c (single_non_eh_succ): New function.
3024 (disqualify_ops_if_throwing_stmt): Renamed to
3025 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
3026 having one non-EH successor BB.
3027 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
3028 generate loads into replacements.
3029 (sra_modify_assign): Likewise and and also use the simple path for
3031 (sra_modify_function_body): Commit statements on edges.
3033 2014-04-17 Richard Biener <rguenther@suse.de>
3036 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
3037 comparison results and add clarifying comment.
3039 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3041 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
3042 (blank_mode): Initialize it.
3043 (emit_mode_size_inline, emit_mode_nunits_inline,
3044 emit_mode_inner_inline): New functions.
3045 (emit_insn_modes_h): Call them and surround their output with
3046 #if GCC_VERSION >= 4001 ... #endif.
3047 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
3048 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
3049 mode_* arrays if the argument is __builtin_constant_p.
3050 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
3051 is enum machine_mode.
3053 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3055 * passes.c (opt_pass::execute): Adjust.
3056 (pass_manager::execute_pass_mode_switching): Likewise.
3057 (early_local_passes::execute): Likewise.
3058 (execute_one_pass): Pass cfun to the pass's execute method.
3059 * tree-pass.h (opt_pass::execute): Add function * argument.
3060 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3061 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3062 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3063 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3064 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3065 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
3066 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
3067 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3068 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3069 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3070 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3071 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3072 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3073 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3074 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3075 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3076 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3077 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3078 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3079 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3080 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3081 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3082 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3083 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3084 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3085 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3086 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3087 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3088 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3091 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3093 * passes.c (opt_pass::gate): Take function * argument.
3094 (gate_all_early_local_passes): Merge into
3095 (early_local_passes::gate): this.
3096 (gate_all_early_optimizations): Merge into
3097 (all_early_optimizations::gate): this.
3098 (gate_all_optimizations): Mege into
3099 (all_optimizations::gate): this.
3100 (gate_all_optimizations_g): Merge into
3101 (all_optimizations_g::gate): this.
3102 (gate_rest_of_compilation): Mege into
3103 (rest_of_compilation::gate): this.
3104 (gate_postreload): Merge into
3105 (postreload::gate): this.
3106 (dump_one_pass): Pass cfun to the pass's gate method.
3107 (execute_ipa_summary_passes): Likewise.
3108 (execute_one_pass): Likewise.
3109 (ipa_write_summaries_2): Likewise.
3110 (ipa_write_optimization_summaries_1): Likewise.
3111 (ipa_read_summaries_1): Likewise.
3112 (ipa_read_optimization_summaries_1): Likewise.
3113 (execute_ipa_stmt_fixups): Likewise.
3114 * tree-pass.h (opt_pass::gate): Add function * argument.
3115 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
3116 combine-stack-adj.c, combine.c, compare-elim.c,
3117 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3118 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
3119 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
3120 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
3121 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3122 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3123 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3124 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
3125 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
3126 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
3127 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3128 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3129 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
3130 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3131 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3132 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3133 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3134 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3135 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3136 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3137 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3138 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3139 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3140 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
3141 var-tracking.c, vtable-verify.c, web.c: Adjust.
3143 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3145 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
3146 * configure: Regenerate.
3148 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3150 * passes.c (dump_one_pass): don't check pass->has_gate.
3151 (execute_ipa_summary_passes): Likewise.
3152 (execute_one_pass): Likewise.
3153 (ipa_write_summaries_2): Likewise.
3154 (ipa_write_optimization_summaries_1): Likewise.
3155 (ipa_read_optimization_summaries_1): Likewise.
3156 (execute_ipa_stmt_fixups): Likewise.
3157 * tree-pass.h (pass_data::has_gate): Remove.
3158 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3159 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3160 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3161 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3162 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3163 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
3164 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
3165 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
3166 gimple-low.c, gimple-ssa-isolate-paths.c,
3167 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
3168 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3169 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3170 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
3171 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
3172 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
3173 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
3174 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3175 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
3176 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3177 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3178 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3179 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3180 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3181 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3182 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3183 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3184 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3185 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3186 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3187 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3188 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3191 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3193 * pass_manager.h (pass_manager::register_dump_files_1): Remove
3195 * passes.c (pass_manager::register_dump_files_1): Merge into
3196 (pass_manager::register_dump_files): this, and remove its handling of
3197 properties since the pass always has the properties anyway.
3198 (pass_manager::pass_manager): Adjust.
3200 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3202 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
3203 * passes.c (pass_manager::register_dump_files_1): Remove dead code
3204 dealing with properties.
3205 (pass_manager::register_dump_files): Adjust.
3207 2014-03-20 Mark Wielaard <mjw@redhat.com>
3209 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
3210 then represent the bound as normal constant value.
3212 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3215 Forward port from 4.8 branch
3216 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
3218 * config/i386/bmiintrin.h (_blsi_u32): New.
3222 (_blsmsk_u32): Ditto.
3223 (_blsmsk_u64): Ditto.
3224 (_tzcnt_u32): Ditto.
3225 (_tzcnt_u64): Ditto.
3227 2014-04-17 Kito Cheng <kito@0xlab.org>
3229 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
3231 2014-04-17 Richard Biener <rguenther@suse.de>
3234 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
3235 boolean results for comparisons.
3237 2014-04-17 Richard Biener <rguenther@suse.de>
3239 PR tree-optimization/60836
3240 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
3241 initial PHI args to be gimple values.
3243 2014-04-17 Richard Biener <rguenther@suse.de>
3245 PR tree-optimization/60841
3246 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
3247 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
3248 of stmts to SLP build.
3249 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
3250 (vect_analyze_slp): Likewise.
3251 (vect_analyze_slp_instance): Likewise.
3252 (vect_build_slp_tree): Limit overall SLP tree growth.
3253 * tree-vectorizer.h (vect_analyze_data_refs,
3254 vect_analyze_slp): Adjust prototypes.
3256 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3258 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
3261 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3263 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
3264 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
3265 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
3266 for TARGET_SLOW_PSHUFB
3268 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3270 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
3271 * config/i386/i386.c (intel_cost): Ditto.
3273 2014-04-17 Joey Ye <joey.ye@arm.com>
3275 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
3277 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3279 * opts.c (common_handle_option): Disable -fipa-reference coorectly
3282 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3284 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
3285 (type_all_derivations_known_p): New predicate.
3286 (type_all_ctors_visible_p): New predicate.
3287 (type_possibly_instantiated_p): New predicate.
3288 (get_odr_type): Compute all_derivations_known.
3289 (dump_odr_type): Dump the flag.
3290 (maybe_record_type): Cleanup.
3291 (record_target_from_binfo): Add bases_to_consider array;
3292 record bases for types w/o instances and skip CXX destructor.
3293 (possible_polymorphic_call_targets_1): Add bases_to_consider
3294 and consider_construction parameters; check if type may have instance.
3295 (get_polymorphic_call_info): Set maybe_in_construction to true
3296 when we know nothing.
3297 (record_targets_from_bases): Skip CXX destructors; they are
3298 never called for types in construction.
3299 (possible_polymorphic_call_targets): Do not record target when
3300 type may not have instance.
3302 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3305 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
3306 external aliases alive, too.
3308 2014-04-16 Andrew Pinski <apinski@cavium.com>
3310 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
3313 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3315 * final.c (compute_alignments): Do not apply loop alignment to a block
3316 falling through to the exit.
3318 2014-04-16 Catherine Moore <clm@codesourcery.com>
3320 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
3321 Adjust constraints for microMIPS store patterns.
3323 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3325 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
3327 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3329 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
3330 (append_use): Run at -O0.
3331 (append_vdef): Likewise.
3332 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
3333 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
3335 2014-04-16 Jakub Jelinek <jakub@redhat.com>
3337 PR tree-optimization/60844
3338 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
3339 (propagate_op_to_single_use, remove_visited_stmt_chain,
3340 linearize_expr, repropagate_negates, reassociate_bb): Use it
3341 instead of gsi_remove.
3343 2014-04-16 Martin Jambor <mjambor@suse.cz>
3345 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
3346 ipa_transforms_to_apply.
3347 (cgraph_function_versioning): Assert that old_node has empty
3348 ipa_transforms_to_apply.
3349 * trans-mem.c (ipa_tm_create_version): Likewise.
3350 * tree-inline.c (tree_function_versioning): Do not duplicate
3351 ipa_transforms_to_apply.
3353 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3356 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
3358 Pass necessary as flags on 64-bit Solaris/x86.
3359 Use lowercase relocs for x86_64-*-*.
3360 * configure: Regenerate.
3362 2014-04-15 Jan Hubicka <jh@suse.cz>
3364 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
3365 (maybe_record_node, likely_target_p): Use it.
3367 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3370 Revert following patch
3372 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3375 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3376 software floating point or no floating point registers, do not
3377 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3378 in GPRs that occurs after we tested for GPRs that would never be
3381 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3382 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3383 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3384 specifically allow DDmode, since that does not use the SPE SIMD
3387 2014-03-21 Mark Wielaard <mjw@redhat.com>
3389 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
3390 as unsigned or int depending on type and value used.
3392 2014-04-15 Richard Biener <rguenther@suse.de>
3394 PR rtl-optimization/56965
3395 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
3396 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
3398 * alias.c (true_dependence_1): Do not call
3399 nonoverlapping_component_refs_p.
3400 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
3401 nonoverlapping_component_refs_p.
3402 (indirect_refs_may_alias_p): Likewise.
3404 2014-04-15 Teresa Johnson <tejohnson@google.com>
3406 * cfg.c (dump_bb_info): Fix flags check.
3407 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
3409 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3411 PR rtl-optimization/60663
3412 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
3415 2014-04-15 Richard Biener <rguenther@suse.de>
3417 * lto-streamer.h (LTO_major_version): Bump to 4.
3419 2014-04-15 Richard Biener <rguenther@suse.de>
3421 * common.opt (lto_partition_model): New enum.
3422 (flto-partition=): Merge separate options with a single with argument,
3423 add -flto-partition=one support.
3424 * flag-types.h (enum lto_partition_model): Declare.
3425 * opts.c (finish_options): Remove duplicate -flto-partition=
3427 * lto-wrapper.c (run_gcc): Adjust.
3429 2014-04-15 Richard Biener <rguenther@suse.de>
3431 * alias.c (ncr_compar): New function.
3432 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
3434 2014-04-15 Richard Biener <rguenther@suse.de>
3436 * alias.c (record_component_aliases): Do not walk BINFOs.
3438 2014-04-15 Richard Biener <rguenther@suse.de>
3440 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
3441 Add struct function argument and adjust.
3442 (find_func_aliases_for_call): Likewise.
3443 (find_func_aliases): Likewise.
3444 (find_func_clobbers): Likewise.
3445 (intra_create_variable_infos): Likewise.
3446 (compute_points_to_sets): Likewise.
3447 (ipa_pta_execute): Adjust. Do not push/pop cfun.
3449 2014-04-15 Richard Biener <rguenther@suse.de>
3451 * tree.c (iterative_hash_expr): Use enum tree_code_class
3452 to store TREE_CODE_CLASS.
3453 (tree_block): Likewise.
3454 (tree_set_block): Likewise.
3455 * tree.h (fold_build_pointer_plus_loc): Use
3456 convert_to_ptrofftype_loc.
3458 2014-04-15 Jakub Jelinek <jakub@redhat.com>
3461 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
3464 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
3466 * cfgloop.h (struct loop): Move force_vectorize down.
3467 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
3468 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
3469 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
3470 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
3471 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
3472 * tree-core.h (enum annot_expr_kind): Add new kind values.
3473 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
3474 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
3476 * tree.def (ANNOTATE_EXPR): Tweak comment.
3478 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3480 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
3483 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
3485 * tree.h (TYPE_IDENTIFIER): Declare.
3486 * tree.c (subrange_type_for_debug_p): Use it.
3487 * godump.c (go_format_type): Likewise.
3488 * dwarf2out.c (is_cxx_auto, modified_type_die,
3489 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
3490 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
3492 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3495 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
3497 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
3499 * config/i386/i386.c (examine_argument): Return bool. Return true if
3500 parameter should be passed in memory.
3501 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
3502 (construct_container): Update calls to examine_argument.
3503 (function_arg_advance_64): Ditto.
3504 (return_in_memory_32): Merge with ix86_return_in_memory.
3505 (return_in_memory_64): Ditto.
3506 (return_in_memory_ms_64): Ditto.
3508 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3510 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
3511 * coverage.c (coverage_compute_profile_id): Handle externally visible
3514 2014-04-14 Martin Jambor <mjambor@suse.cz>
3516 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
3517 DECL_DISREGARD_INLINE_LIMITS functions.
3519 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3522 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
3524 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3527 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
3528 optimize_insn_for_speed_p instead of
3529 optimize_function_for_speed_p.
3531 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
3533 * doc/invoke.texi (free): Document AArch64.
3535 2014-04-14 Richard Biener <rguenther@suse.de>
3537 PR tree-optimization/60042
3538 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
3539 (insert_into_preds_of_block): Do not prevent PHI insertion
3540 for REFERENCE exprs here ...
3541 (eliminate_dom_walker::before_dom_children): ... but prevent
3542 their use here under similar conditions when applied to the
3543 IL after PRE optimizations.
3545 2014-04-14 Richard Biener <rguenther@suse.de>
3547 * passes.def: Move early points-to after early SRA.
3549 2014-04-14 Richard Biener <rguenther@suse.de>
3551 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
3552 check for which sign-changes we allow when forwarding
3553 a converted value into a switch.
3555 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3557 * stor-layout.c (place_field): Finalize non-constant offset for the
3560 2014-04-14 Richard Biener <rguenther@suse.de>
3562 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
3564 (expand_switch_using_bit_tests_p): Likewise.
3565 (process_switch): Compute and pass on speed_p based on the
3567 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
3568 optimize_bb_for_speed_p.
3570 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3572 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
3573 * function.h (struct function): Rename has_force_vect_loops into
3574 has_force_vectorize_loops.
3575 * lto-streamer-in.c (input_cfg): Adjust for renaming.
3576 (input_struct_function_base): Likewise.
3577 * lto-streamer-out.c (output_cfg): Likewise.
3578 (output_struct_function_base): Likewise.
3579 * omp-low.c (expand_omp_simd): Likewise.
3580 * tree-cfg.c (move_sese_region_to_fn): Likewise.
3581 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
3582 (version_loop_for_if_conversion): Likewise.
3583 (tree_if_conversion): Likewise.
3584 (main_tree_if_conversion): Likewise.
3585 (gate_tree_if_conversion): Likewise.
3586 * tree-inline.c (copy_loops): Likewise.
3587 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
3588 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
3589 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
3590 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
3591 * tree-vectorizer.c (vectorize_loops): Likewise.
3592 * tree-vectorizer.h (unlimited_cost_model): Likewise.
3594 2014-04-14 Richard Biener <rguenther@suse.de>
3597 * lto-streamer-out.c (wrap_refs): New function.
3598 (lto_output): Wrap symbol references in global initializes in
3599 type-preserving MEM_REFs.
3601 2014-04-14 Christian Bruel <christian.bruel@st.com>
3603 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
3605 2014-04-14 Christian Bruel <christian.bruel@st.com>
3607 * config/sh/sh.md (setmemqi): New expand pattern.
3608 * config/sh/sh.h (CLEAR_RATIO): Define.
3609 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
3610 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
3612 2014-04-14 Richard Biener <rguenther@suse.de>
3615 * fold-const.c (negate_expr_p): Don't negate directional rounding
3617 (fold_negate_expr): Likewise.
3619 2014-04-14 Richard Biener <rguenther@suse.de>
3621 PR tree-optimization/59817
3622 PR tree-optimization/60453
3623 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
3624 recursion to catch all CHRECs in the scalar evolution and restrict
3625 the predicate for the remains appropriately.
3627 2014-04-12 Catherine Moore <clm@codesourcery.com>
3629 * config/mips/constraints.md: Add new register constraint "kb".
3630 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
3631 (*movhi_internal): Likewise.
3632 (*movqi_internal): Likewise.
3633 * config/mips/mips.h (M16_STORE_REGS): New register class.
3634 (REG_CLASS_NAMES): Add M16_STORE_REGS.
3635 (REG_CLASS_CONTENTS): Likewise.
3636 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
3638 2014-04-11 Tobias Burnus <burnus@net-b.de>
3641 * doc/invoke.texi (-Wformat-signedness): Document it.
3642 (Wformat=2): Mention that this enables -Wformat-signedness.
3644 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3646 * common/config/epiphany/epiphany-common.c
3647 (epiphany_option_optimization_table): Enable section anchors by
3648 default at -O1 or higher.
3649 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
3650 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
3651 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
3652 carries no extra cost.
3653 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
3654 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
3655 * config/epiphany/predicates.md (memclob_operand): New predicate.
3656 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
3657 Use memclob_operand predicate and X constraint for operand 3.
3659 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3661 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
3662 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
3665 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3667 PR rtl-optimization/60651
3668 * mode-switching.c (optimize_mode_switching): Make sure to emit
3669 sets of a lower numbered entity before sets of a higher numbered
3670 entity to a mode of the same or lower priority.
3671 When creating a seginfo for a basic block that starts with a code
3672 label, move the insertion point past the code label.
3673 (new_seginfo): Document and enforce requirement that
3674 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
3675 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
3676 * doc/tm.texi: Regenerate.
3678 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
3681 * config/arc/arc.c (arc_save_restore): Fix assert typo.
3683 2013-04-11 Jakub Jelinek <jakub@redhat.com>
3685 * BASE-VER: Set to 4.10.0.
3687 2014-04-11 Tobias Burnus <burnus@net-b.de>
3690 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
3691 * doc/gcc.texi (Service): Update description in the @menu
3692 * doc/invoke.texi (Option Summary): Remove misplaced and
3695 2014-04-11 Steve Ellcey <sellcey@mips.com>
3696 Jakub Jelinek <jakub@redhat.com>
3699 * expr.c (convert_move): Use emit_store_flag_force instead of
3700 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
3703 2014-04-11 Richard Biener <rguenther@suse.de>
3706 * varasm.c (assemble_alias): Avoid endless error reporting
3707 recursion by setting TREE_ASM_WRITTEN.
3709 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3711 * config/s390/s390.md: Add a splitter for NOT rtx.
3713 2014-04-11 Jakub Jelinek <jakub@redhat.com>
3715 PR rtl-optimization/60663
3716 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
3718 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
3719 Jakub Jelinek <jakub@redhat.com>
3722 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
3723 flag from decl_node to node.
3725 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3728 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
3729 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
3730 ameliorating the cases where it can be.
3732 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
3735 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
3737 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3738 (loadsync_<mode>): Change mode.
3739 (load_quadpti, store_quadpti): New.
3740 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3741 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3742 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
3744 2014-04-09 Cong Hou <congh@google.com>
3747 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
3750 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3752 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
3753 instead of vnor to exploit possible fusion opportunity in the
3755 (altivec_expand_vec_perm_const_le): Likewise.
3757 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
3759 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3760 (loadsync_<mode>): Change mode.
3761 (load_quadpti, store_quadpti): New.
3762 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3763 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3765 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
3768 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
3769 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
3770 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
3772 2014-04-08 Richard Biener <rguenther@suse.de>
3775 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
3776 a 64bit widest int print double-int similar to on HWI64 hosts.
3778 2014-04-08 Richard Biener <rguenther@suse.de>
3780 PR tree-optimization/60785
3781 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
3782 default defs properly.
3784 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
3786 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
3787 (Weffc++): Likewise.
3789 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
3791 * ipa-devirt.c (maybe_record_node): When node is not recorded,
3792 set completep to false rather than true.
3794 2014-04-07 Douglas B Rupp <rupp@adacore.com>
3797 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
3798 ARM_TARGET2_DWARF_FORMAT.
3800 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
3803 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
3804 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
3807 2014-04-07 Richard Biener <rguenther@suse.de>
3809 PR tree-optimization/60766
3810 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
3811 (may_eliminate_iv): Convert cand_value_at result to desired type.
3813 2014-04-07 Jason Merrill <jason@redhat.com>
3816 * common.opt (-fno-gnu-unique): Add.
3817 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
3819 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3821 * haifa-sched.c: Fix outdated function reference and minor
3822 grammar errors in introductory comment.
3824 2014-04-07 Richard Biener <rguenther@suse.de>
3827 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
3829 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
3831 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
3834 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
3835 size accounting for thunks.
3836 (pa_asm_output_mi_thunk): Use final_start_function() and
3837 final_end_function() to output function start and end directives.
3839 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3841 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
3842 device specific ISA/ feature information. Remove short_sp and
3843 errata_skip ds. Add avr_device_specific_features enum to have device
3845 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
3846 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
3847 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
3848 updated device specific info.
3849 * config/avr/avr-mcus.def: Merge device specific details to
3850 dev_attribute field.
3851 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
3853 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
3854 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
3855 assembler if RMW isa supported by current device.
3856 * config/avr/genmultilib.awk: Update as device info structure changed.
3857 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
3859 2014-04-04 Cong Hou <congh@google.com>
3861 PR tree-optimization/60656
3862 * tree-vect-stmts.c (supportable_widening_operation):
3863 Fix a bug that elements in a vector with vect_used_by_reduction
3864 property are incorrectly reordered when the operation on it is not
3865 consistant with the one in reduction operation.
3867 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
3869 PR rtl-optimization/60155
3870 * gcse.c (record_set_data): New function.
3871 (single_set_gcse): New function.
3872 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
3873 (hoist_code): Likewise.
3874 (get_pressure_class_and_nregs): Likewise.
3876 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
3878 * explow.c (probe_stack_range): Emit a final optimization blockage.
3880 2014-04-04 Anthony Green <green@moxielogic.com>
3882 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
3885 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
3888 * lto-cgraph.c (input_overwrite_node): Check that partitioning
3889 flags are set only during streaming.
3890 * ipa.c (process_references, walk_polymorphic_call_targets,
3891 symtab_remove_unreachable_nodes): Drop bodies of always inline
3892 after early inlining.
3893 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
3895 2014-04-04 Jakub Jelinek <jakub@redhat.com>
3896 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3899 * dwarf2out.c (const_ok_for_output_1): Reject expressions
3902 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3905 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
3907 (cortex_a53_fdivd): Likewise.
3909 2014-04-04 Martin Jambor <mjambor@suse.cz>
3912 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
3914 * cgraph.c (clone_of_p): Also return true if thunks match.
3915 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
3916 cgraph_function_or_thunk_node and an obsolete comment.
3917 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
3919 (build_function_decl_skip_args): Likewise.
3920 (set_new_clone_decl_and_node_flags): New function.
3921 (duplicate_thunk_for_node): Likewise.
3922 (redirect_edge_duplicating_thunks): Likewise.
3923 (cgraph_clone_node): New parameter args_to_skip, pass it to
3924 redirect_edge_duplicating_thunks which is called instead of
3925 cgraph_redirect_edge_callee.
3926 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
3927 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
3929 2014-04-04 Jeff Law <law@redhat.com>
3932 * config/arm/predicates.md (const_int_I_operand): New predicate.
3933 (const_int_M_operand): Similarly.
3934 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
3936 (insv_t2, extv_reg, extzv_t2): Likewise.
3937 (load_multiple_with_writeback): Similarly for const_int_I_operand.
3938 (pop_multiple_with_writeback_and_return): Likewise.
3939 (vfp_pop_multiple_with_writeback): Likewise
3941 2014-04-04 Richard Biener <rguenther@suse.de>
3944 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
3945 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
3947 * gimplify.h (gimple_add_tmp_var_fn): Declare.
3948 * gimplify.c (gimple_add_tmp_var_fn): New function.
3949 * gimple-expr.h (create_tmp_reg_fn): Declare.
3950 * gimple-expr.c (create_tmp_reg_fn): New function.
3951 * gimple-low.c (record_vars_into): Don't change cfun.
3952 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
3953 code generation without cfun.
3955 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
3958 * Makefile.in (install-driver): Fix shell scripting.
3960 2014-04-03 Cong Hou <congh@google.com>
3962 PR tree-optimization/60505
3963 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
3964 threshold of number of iterations below which no vectorization
3966 * tree-vect-loop.c (new_loop_vec_info):
3967 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
3968 * tree-vect-loop.c (vect_analyze_loop_operations):
3969 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
3970 * tree-vect-loop.c (vect_transform_loop):
3971 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
3972 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
3973 of iterations of the loop and see if we should build the epilogue.
3975 2014-04-03 Richard Biener <rguenther@suse.de>
3977 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
3978 (streamer_tree_cache_create): Adjust.
3979 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
3980 to allow optional nodes array.
3981 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
3982 (streamer_tree_cache_append): Likewise.
3983 (streamer_tree_cache_create): Create nodes array optionally
3984 as specified by parameter.
3985 * lto-streamer-out.c (create_output_block): Avoid maintaining
3986 the node array in the writer cache.
3987 (DFS_write_tree): Remove assertion.
3988 (produce_asm_for_decls): Free the out decl state hash table early.
3989 * lto-streamer-in.c (lto_data_in_create): Adjust for
3990 streamer_tree_cache_create prototype change.
3992 2014-04-03 Richard Biener <rguenther@suse.de>
3994 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
3995 set TREE_CHAIN to NULL_TREE.
3997 2014-04-03 Richard Biener <rguenther@suse.de>
3999 PR tree-optimization/60740
4000 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
4001 over all GIMPLE_COND operands.
4003 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
4005 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
4006 (Weffc++): Remove Scott's numbering, merge lists and reference
4009 2014-04-03 Nick Clifton <nickc@redhat.com>
4011 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
4014 2014-04-03 Martin Jambor <mjambor@suse.cz>
4016 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
4017 mention gcc_unreachable before failing.
4018 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
4021 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
4024 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
4025 inconsistent code and instead mark the context inconsistent.
4026 (possible_polymorphic_call_targets): For inconsistent contexts
4027 return empty complete list.
4029 2014-04-02 Anthony Green <green@moxielogic.com>
4031 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
4032 (extendqisi2, extendhisi2): Define.
4033 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
4034 (WCHAR_TYPE): Change to unsigned int.
4036 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4038 PR tree-optimization/60733
4039 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
4040 insertion point for PHI candidates to be the end of the feeding
4041 block for the PHI argument.
4043 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
4045 PR rtl-optimization/60650
4046 * lra-constraints.c (process_alt_operands): Decrease reject for
4047 earlyclobber matching.
4049 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4051 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
4053 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4055 * config/spu/spu.c (pad_bb): Do not crash when the last
4056 insn is CODE_FOR_blockage.
4058 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4060 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
4061 lies outside the target mode.
4063 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4066 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4067 software floating point or no floating point registers, do not
4068 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4069 in GPRs that occurs after we tested for GPRs that would never be
4072 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4073 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4074 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4075 specifically allow DDmode, since that does not use the SPE SIMD
4078 2014-04-02 Richard Biener <rguenther@suse.de>
4081 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4082 MODE_INTs. Properly use negv_optab.
4083 (expand_abs): Likewise.
4085 2014-04-02 Richard Biener <rguenther@suse.de>
4088 * Makefile.in (install-driver): Guard extra installs with special
4091 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4093 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4096 2014-04-01 Richard Henderson <rth@redhat.com>
4099 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
4100 alternative enabled before register allocation.
4102 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
4104 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
4105 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
4107 (nios2_large_got_address): Remove unneeded 'sym' parameter.
4108 (nios2_got_address): Update nios2_large_got_address call site.
4109 (nios2_delegitimize_address): New function.
4110 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
4111 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
4112 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
4114 2014-04-01 Martin Husemann <martin@duskware.de>
4116 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
4119 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
4121 PR rtl-optimization/60604
4122 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
4123 check from register_operand.
4124 (register_operand): Redefine in terms of general_operand.
4125 (nonmemory_operand): Use register_operand for the non-constant cases.
4127 2014-04-01 Richard Biener <rguenther@suse.de>
4129 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
4131 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
4133 * doc/invoke.texi (mapp-regs): Clarify.
4135 2014-03-31 Ulrich Drepper <drepper@gmail.com>
4137 * config/i386/avx512fintrin.h (__v32hi): Define type.
4138 (__v64qi): Likewise.
4139 (_mm512_set1_epi8): Define.
4140 (_mm512_set1_epi16): Define.
4141 (_mm512_set4_epi32): Define.
4142 (_mm512_set4_epi64): Define.
4143 (_mm512_set4_pd): Define.
4144 (_mm512_set4_ps): Define.
4145 (_mm512_setr4_epi64): Define.
4146 (_mm512_setr4_epi32): Define.
4147 (_mm512_setr4_pd): Define.
4148 (_mm512_setr4_ps): Define.
4149 (_mm512_setzero_epi32): Define.
4151 2014-03-31 Martin Jambor <mjambor@suse.cz>
4154 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
4155 callsite_arguments_match_p. Updated all callers. Also check types of
4156 corresponding formal parameters and actual arguments.
4157 (not_all_callers_have_enough_arguments_p) Renamed to
4158 some_callers_have_mismatched_arguments_p.
4160 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
4162 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
4164 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
4167 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
4170 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
4172 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
4173 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
4175 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
4176 Use FMAMODE_NOVF512 mode iterator.
4177 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
4178 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
4179 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
4180 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
4182 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
4183 Use VF_128_256 mode iterator.
4184 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
4187 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4189 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
4190 static chain if needed.
4192 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4195 * lra-constraints.c (index_part_to_reg): New.
4196 (process_address): Use it.
4198 2014-03-27 Jeff Law <law@redhat.com>
4199 Jakub Jelinek <jakub@redhat.com>
4202 * expr.c (do_tablejump): Use simplify_gen_binary rather than
4203 gen_rtx_{PLUS,MULT} to build up the address expression.
4205 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
4206 creating non-canonical RTL.
4208 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4211 * ipa-inline.c (want_inline_small_function_p): Short circuit large
4212 functions; reorganize to make cheap checks first.
4213 (inline_small_functions): Do not estimate growth when dumping;
4215 * ipa-inline.h (inline_summary): Add min_size.
4216 (growth_likely_positive): New function.
4217 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
4218 (set_cond_stmt_execution_predicate): Cleanup.
4219 (estimate_edge_size_and_time): Compute min_size.
4220 (estimate_calls_size_and_time): Likewise.
4221 (estimate_node_size_and_time): Likewise.
4222 (inline_update_overall_summary): Update min_size.
4223 (do_estimate_edge_time): Likewise.
4224 (do_estimate_edge_size): Update.
4225 (do_estimate_edge_hints): Update.
4226 (growth_likely_positive): New function.
4228 2014-03-28 Jakub Jelinek <jakub@redhat.com>
4231 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
4232 also if addr has VOIDmode.
4234 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4236 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
4237 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
4239 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
4240 instructions as well as AdvancedSIMD loads.
4242 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4244 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
4245 Use crypto_aese type.
4246 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
4247 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
4248 crypto_aese, crypto_aesmc. Move to types.md.
4249 * config/arm/types.md (crypto_aes): Split into crypto_aese,
4251 * config/arm/iterators.md (crypto_type): Likewise.
4253 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4255 * cgraph.c: Include expr.h and tree-dfa.h.
4256 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
4259 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4262 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
4263 regs from checking multi-reg pseudos.
4265 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4267 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
4269 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4271 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
4272 if it would clobber the stack pointer, even temporarily.
4274 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
4276 * mode-switching.c: Make small adjustments to the top comment.
4278 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4280 * config/rs6000/constraints.md (wD constraint): New constraint to
4281 match the constant integer to get the top DImode/DFmode out of a
4282 vector in a VSX register.
4284 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
4285 match the constant integer to get the top DImode/DFmode out of a
4286 vector in a VSX register.
4288 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
4291 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
4294 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
4295 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
4297 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
4298 Optimize vec_extract of 64-bit values, where the value being
4299 extracted is in the top word, where we can use scalar
4300 instructions. Add direct move and store support. Combine the big
4301 endian/little endian vector select load support into a single insn.
4302 (vsx_extract_<mode>_internal1): Likewise.
4303 (vsx_extract_<mode>_internal2): Likewise.
4304 (vsx_extract_<mode>_load): Likewise.
4305 (vsx_extract_<mode>_store): Likewise.
4306 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
4307 combined into vsx_extract_<mode>_load.
4308 (vsx_extract_<mode>_one_le): Likewise.
4310 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
4311 define the top 64-bit vector element.
4313 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
4316 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4317 Document vec_vbpermq builtin.
4320 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
4321 enable use of xxsldwi and xxpermdi builtin functions.
4322 (vec_xxpermdi): Likewise.
4324 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4325 Document use of vec_xxsldwi and vec_xxpermdi builtins.
4327 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
4329 PR rtl-optimization/60650
4330 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
4332 (find_spills_for): New.
4333 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
4334 Spill all pseudos on the second iteration.
4336 2014-03-27 Marek Polacek <polacek@redhat.com>
4339 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
4342 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4344 * config/s390/s390.c (s390_can_use_return_insn): Check for
4345 call-saved FPRs on 31 bit.
4347 2014-03-27 Jakub Jelinek <jakub@redhat.com>
4350 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
4351 if they need regimplification, just drop them instead of
4352 calling gimple_regimplify_operands on them.
4354 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
4357 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
4358 (aarch64_frame_pointer_required): Adjust logic.
4359 (aarch64_can_eliminate): Adjust logic.
4360 (aarch64_override_options_after_change): Adjust logic.
4362 2014-03-27 Dehao Chen <dehao@google.com>
4364 * ipa-inline.c (early_inliner): Update node's inline info.
4366 2014-03-26 Dehao Chen <dehao@google.com>
4368 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
4369 compiler inserted conditional jumps for NAN float check.
4371 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4373 * ubsan.h (ubsan_create_data): Change second argument's type
4374 to const location_t *.
4375 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
4377 (ubsan_create_data): Change second argument to const location_t *PLOC.
4378 Create Loc field whenever PLOC is non-NULL.
4379 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
4380 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
4384 * real.c (real_to_integer2): Change type of low to UHWI.
4386 2014-03-26 Tobias Burnus <burnus@net-b.de>
4388 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
4389 (CILK_SELF_SPECS): New define.
4390 (driver_self_specs): Use it.
4392 2014-03-26 Richard Biener <rguenther@suse.de>
4394 * tree-pretty-print.c (percent_K_format): Implement special
4395 case for LTO and its stripped down BLOCK tree.
4397 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4400 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
4402 * tree-vrp.c (simplify_internal_call_using_ranges): If only
4403 one range is range_int_cst_p, but not both, at least optimize
4404 addition/subtraction of 0 and multiplication by 0 or 1.
4405 * gimple-fold.c (gimple_fold_call): Fold
4406 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
4407 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
4408 INTEGER_CSTs, try to fold at least x * 0 and y - y.
4410 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
4412 PR rtl-optimization/60452
4413 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
4414 <case REG>: Return 1 for invalid offsets from the frame pointer.
4416 2014-03-26 Marek Polacek <polacek@redhat.com>
4419 * doc/extend.texi (C Extensions): Mention variable-length arrays in
4422 2014-03-26 Marek Polacek <polacek@redhat.com>
4425 * doc/extend.texi (Designated Inits): Describe what happens to omitted
4428 2014-03-26 Marek Polacek <polacek@redhat.com>
4431 * ira-color.c (update_conflict_hard_regno_costs): Perform the
4432 multiplication in unsigned type.
4434 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4436 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
4438 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4440 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
4442 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4445 * cif-code.def (UNREACHABLE) New code.
4446 * ipa-inline.c (inline_small_functions): Skip edges to
4447 __builtlin_unreachable.
4448 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
4449 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
4450 predicate to __bulitin_unreachable.
4451 (set_cond_stmt_execution_predicate): Fix issue when
4452 invert_tree_comparison returns ERROR_MARK.
4453 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
4454 propagate to inline clones.
4455 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
4457 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
4458 * cgraphclones.c (cgraph_clone_node): If call destination is already
4459 ureachable, do not redirect it back.
4460 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
4463 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4465 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
4466 Do not modify inline clones.
4468 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4470 * config/i386/i386.md (general_sext_operand): New mode attr.
4471 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
4472 don't generate (sign_extend (const_int)).
4473 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
4474 operands[2]. Use We constraint instead of <i> and
4475 <general_sext_operand> predicate instead of <general_operand>.
4476 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
4477 * config/i386/constraints.md (We): New constraint.
4478 * config/i386/predicates.md (x86_64_sext_operand,
4479 sext_operand): New predicates.
4481 2014-03-25 Martin Jambor <mjambor@suse.cz>
4484 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
4485 inconsistent devirtualizations to __builtin_unreachable.
4487 2014-03-25 Marek Polacek <polacek@redhat.com>
4490 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
4492 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
4494 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
4495 order of elements for big-endian.
4497 2014-03-25 Richard Biener <rguenther@suse.de>
4500 * gimplify-me.c (gimple_regimplify_operands): Update the
4503 2014-03-25 Martin Jambor <mjambor@suse.cz>
4506 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
4507 (lto_output_varpool_node): Likewise.
4508 (input_overwrite_node): Likewise.
4509 (input_varpool_node): Likewise.
4511 2014-03-25 Richard Biener <rguenther@suse.de>
4513 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
4514 (run_gcc): Likewise.
4516 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4518 * combine.c (simplify_compare_const): Add MODE argument.
4519 Handle mode_width 0 as very large mode_width.
4520 (try_combine, simplify_comparison): Adjust callers.
4522 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
4523 type to avoid signed integer overflow.
4524 * explow.c (plus_constant): Likewise.
4526 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
4528 * doc/generic.texi: Correct typos.
4530 2014-03-24 Tobias Burnus <burnus@net-b.de>
4532 * doc/invoke.texi (-flto): Expand section about
4533 using static libraries with LTO.
4535 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4537 PR rtl-optimization/60501
4538 * optabs.def (addptr3_optab): New optab.
4539 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
4540 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
4541 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
4543 * lra.c (emit_add3_insn): Use the addptr pattern if available.
4545 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
4547 2014-03-24 Ulrich Drepper <drepper@gmail.com>
4549 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
4552 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
4553 (_mm256_undefined_ps): Define.
4554 (_mm256_undefined_pd): Define.
4555 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
4556 (_mm_undefined_pd): Define.
4557 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
4558 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
4559 (_mm512_undefined_ps): Define.
4560 (_mm512_undefined_pd): Define.
4561 Use _mm*_undefined_*.
4562 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
4564 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
4566 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
4567 (lshr_simd): DI mode added.
4568 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
4569 (aarch64_ushr_simddi): Likewise.
4570 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
4571 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
4572 (vshrd_n_u64): Likewise.
4574 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4576 * Makefile.in (s-macro_list): Depend on cc1.
4578 2014-03-23 Teresa Johnson <tejohnson@google.com>
4580 * ipa-utils.c (ipa_print_order): Use specified dump file.
4582 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
4584 PR rtl-optimization/60601
4585 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
4587 * gcc.c (eval_spec_function): Initialize save_growing_value.
4589 2014-03-22 Jakub Jelinek <jakub@redhat.com>
4592 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
4593 code == MINUS_EXPR, never swap op0 with op1.
4595 * toplev.c (init_local_tick): Avoid signed integer multiplication
4597 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
4598 shift by first operand's bitsize.
4600 2014-03-21 Jakub Jelinek <jakub@redhat.com>
4603 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
4605 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
4606 TARGET_ISA_64BIT_P(x).
4608 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4610 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
4611 pattern for vector nor instead of subtract from splat(-1).
4612 (altivec_expand_vec_perm_const_le): Likewise.
4614 2014-03-21 Richard Henderson <rth@twiddle.net>
4617 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
4618 related insns after epilogue_completed.
4620 2014-03-21 Martin Jambor <mjambor@suse.cz>
4623 * cgraph.h (symtab_node): New flag body_removed.
4624 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
4625 when removing bodies.
4626 * symtab.c (dump_symtab_base): Dump body_removed flag.
4627 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
4628 had their bodies removed.
4630 2014-03-21 Martin Jambor <mjambor@suse.cz>
4633 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
4636 2014-03-21 Richard Biener <rguenther@suse.de>
4638 PR tree-optimization/60577
4639 * tree-core.h (struct tree_base): Document nothrow_flag use
4640 in VAR_DECL_NONALIASED.
4641 * tree.h (VAR_DECL_NONALIASED): New.
4642 (may_be_aliased): Adjust.
4643 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
4645 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4647 * expr.c (expand_expr_real_1): Remove outdated comment.
4649 2014-03-20 Jakub Jelinek <jakub@redhat.com>
4652 * ira.c (adjust_cleared_regs): Call copy_rtx on
4653 *reg_equiv[REGNO (loc)].src_p before passing it to
4654 simplify_replace_fn_rtx.
4657 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
4658 into CONST, put pic register as first operand of PLUS. Use
4659 gen_const_mem for both 32-bit and 64-bit PIC got loads.
4661 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4663 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
4665 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4667 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
4668 around for store forwarding issue in the FPU on the UT699.
4669 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
4670 loads and operations if -mfix-ut699 is specified.
4671 (divtf3_hq): Tweak attribute.
4672 (sqrttf2_hq): Likewise.
4674 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4676 * calls.c (store_one_arg): Remove incorrect const qualification on the
4677 type of the temporary.
4678 * cfgexpand.c (expand_return): Likewise.
4679 * expr.c (expand_constructor): Likewise.
4680 (expand_expr_real_1): Likewise.
4682 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4684 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
4687 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
4690 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
4692 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
4694 * config/arm/aarch-common-protos.h
4695 (alu_cost_table): Fix spelling of "extend".
4696 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
4698 2014-03-19 Richard Biener <rguenther@suse.de>
4701 * tree-core.h (tree_type_common): Re-order pointer members
4702 to reduce recursion depth during GC walks.
4704 2014-03-19 Marek Polacek <polacek@redhat.com>
4707 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
4708 before accessing it.
4710 2014-03-19 Richard Biener <rguenther@suse.de>
4713 * lto-streamer-in.c (input_function): In WPA stage do not drop
4716 2014-03-19 Jakub Jelinek <jakub@redhat.com>
4718 PR tree-optimization/60559
4719 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
4720 with build_zero_cst assignment.
4722 2014-03-18 Kai Tietz <ktietz@redhat.com>
4724 PR rtl-optimization/56356
4725 * sdbout.c (sdbout_parms): Verify that parms'
4726 incoming argument is valid.
4727 (sdbout_reg_parms): Likewise.
4729 2014-03-18 Richard Henderson <rth@redhat.com>
4732 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
4733 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
4734 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
4736 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
4738 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
4739 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
4740 Italicize plugin event names in description. Explain that
4741 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
4742 Remind that no GCC functions should be called after PLUGIN_FINISH.
4743 Explain what pragmas with expansion are.
4745 2014-03-18 Martin Liska <mliska@suse.cz>
4747 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
4748 gimple call statement is update.
4749 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
4750 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
4752 2014-03-18 Jakub Jelinek <jakub@redhat.com>
4755 * ubsan.c (ubsan_instrument_unreachable): Call
4756 initialize_sanitizer_builtins.
4757 (ubsan_pass): Likewise.
4760 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
4761 varpool_finalize_decl instead of rest_of_decl_compilation.
4763 2014-03-18 Richard Biener <rguenther@suse.de>
4765 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
4766 by using bitmap_and_compl instead of bitmap_and_compl_into.
4767 (df_rd_transfer_function): Likewise.
4769 2014-03-18 Richard Biener <rguenther@suse.de>
4771 * doc/lto.texi (fresolution): Fix typo.
4773 2014-03-18 Richard Biener <rguenther@suse.de>
4775 * doc/invoke.texi (flto): Update for changes in 4.9.
4777 2014-03-18 Richard Biener <rguenther@suse.de>
4779 * doc/loop.texi: Remove section on the removed lambda framework.
4780 Update loop docs with recent changes in preserving loop structure.
4782 2014-03-18 Richard Biener <rguenther@suse.de>
4784 * doc/lto.texi (-fresolution): Document.
4786 2014-03-18 Richard Biener <rguenther@suse.de>
4788 * doc/contrib.texi: Adjust my name.
4790 2014-03-18 Jakub Jelinek <jakub@redhat.com>
4793 * internal-fn.c: Include diagnostic-core.h.
4794 (expand_BUILTIN_EXPECT): New function.
4795 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
4796 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
4797 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
4798 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
4800 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
4801 Revert 3 argument __builtin_expect code.
4802 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
4803 * gimple-fold.c (gimple_fold_call): Likewise.
4804 * tree.h (fold_builtin_expect): New prototype.
4805 * builtins.c (build_builtin_expect_predicate): Add predictor
4806 argument, if non-NULL, create 3 argument __builtin_expect.
4807 (fold_builtin_expect): No longer static. Add ARG2 argument,
4808 pass it through to build_builtin_expect_predicate.
4809 (fold_builtin_2): Adjust caller.
4810 (fold_builtin_3): Handle BUILT_IN_EXPECT.
4811 * internal-fn.def (BUILTIN_EXPECT): New.
4813 2014-03-18 Tobias Burnus <burnus@net-b.de>
4816 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
4817 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
4818 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
4820 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
4823 * predict.c (combine_predictions_for_bb): Fix up formatting.
4824 (expr_expected_value_1, expr_expected_value): Add predictor argument,
4825 fill what it points to if non-NULL.
4826 (tree_predict_by_opcode): Adjust caller, use the predictor.
4827 * predict.def (PRED_COMPARE_AND_SWAP): Add.
4829 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
4831 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
4832 proper constant for the store mode.
4834 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
4836 * symtab.c (change_decl_assembler_name): Fix transparent alias
4839 2014-03-16 Renlin Li <Renlin.Li@arm.com>
4841 * config/aarch64/aarch64.c: Correct the comments about the
4842 aarch64 stack layout.
4844 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
4846 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
4847 check for GF_OMP_FOR_KIND_FOR.
4849 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
4851 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
4852 ymm and zmm register names.
4854 2014-03-17 Jakub Jelinek <jakub@redhat.com>
4857 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
4858 note creation for the 2010-08-31 changes.
4860 2014-03-17 Marek Polacek <polacek@redhat.com>
4863 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
4864 as -fno-tree-loop-vectorize.
4865 (expand_omp_simd): Likewise.
4867 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
4869 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
4870 (eligible_for_call_delay): New prototype.
4871 * config/sparc/sparc.c (tls_call_delay): Rename into...
4872 (eligible_for_call_delay): ...this. Return false if the instruction
4873 cannot be put in the delay slot of a branch.
4874 (eligible_for_restore_insn): Simplify.
4875 (eligible_for_return_delay): Return false if the instruction cannot be
4876 put in the delay slot of a branch and simplify.
4877 (eligible_for_sibcall_delay): Return false if the instruction cannot be
4878 put in the delay slot of a branch.
4879 * config/sparc/sparc.md (fix_ut699): New attribute.
4880 (tls_call_delay): Delete.
4881 (in_call_delay): Reimplement.
4882 (eligible_for_sibcall_delay): Rename into...
4883 (in_sibcall_delay): ...this.
4884 (eligible_for_return_delay): Rename into...
4885 (in_return_delay): ...this.
4886 (in_branch_delay): Reimplement.
4887 (in_uncond_branch_delay): Delete.
4888 (in_annul_branch_delay): Delete.
4890 2014-03-14 Richard Henderson <rth@redhat.com>
4893 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
4894 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
4895 (*floathi<X87MODEF>2_i387_with_temp): Remove.
4896 (floathi splitters): Remove.
4897 (float<SWI48x>xf2): New pattern.
4898 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
4899 code that tried to handle DImode for 32-bit, but which was excluded
4900 by the pattern's condition. Drop allocation of stack temporary.
4901 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
4902 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
4903 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
4904 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
4905 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
4906 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
4907 (*float<SWI48><MODEF>2_sse_interunit): Remove.
4908 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
4909 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
4910 (*float<SWI48x><X87MODEF>2_i387): Remove.
4911 (all float _with_temp splitters): Remove.
4912 (*float<SWI48x><MODEF>2_i387): New pattern.
4913 (*float<SWI48><MODEF>2_sse): New pattern.
4914 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
4915 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
4917 2014-03-14 Jakub Jelinek <jakub@redhat.com>
4918 Marek Polacek <polacek@redhat.com>
4921 * common.opt (dump_base_name_prefixed): New Variable.
4922 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
4923 if x_dump_base_name_prefixed is already set, set it at the end.
4925 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
4927 PR rtl-optimization/60508
4928 * lra-constraints.c (get_reload_reg): Add new parameter
4930 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
4931 Pass the new parameter values.
4933 2014-03-14 Richard Biener <rguenther@suse.de>
4935 * common.opt: Revert unintented changes from r205065.
4938 2014-03-14 Richard Biener <rguenther@suse.de>
4941 * cfghooks.c (split_block): Properly adjust all loops the
4942 block was a latch of.
4944 2014-03-14 Martin Jambor <mjambor@suse.cz>
4947 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
4950 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
4953 * config/avr/avr.c (avr_set_current_function): Pass function name
4954 through default_strip_name_encoding before sanity checking instead
4955 of skipping the first char of the assembler name.
4957 2014-03-13 Richard Henderson <rth@redhat.com>
4960 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
4961 (ix86_force_to_memory, ix86_free_from_memory): Remove.
4962 * config/i386/i386-protos.h: Likewise.
4963 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
4964 in the expander instead of a splitter.
4965 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
4966 any possibility of requiring a memory.
4967 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
4968 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
4969 (fp branch splitters): Update for ix86_split_fp_branch.
4970 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
4971 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
4972 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
4973 (*fop_<MODEF>_2_i387): Remove f/r alternative.
4974 (*fop_<MODEF>_3_i387): Likewise.
4975 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
4976 (splitters for the fop_* register patterns): Remove.
4977 (fscalexf4_i387): Rename from *fscalexf4_i387.
4978 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
4980 2014-03-13 Jakub Jelinek <jakub@redhat.com>
4982 PR tree-optimization/59779
4983 * tree-dfa.c (get_ref_base_and_extent): Use double_int
4984 type for bitsize and maxsize instead of HOST_WIDE_INT.
4986 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
4988 PR rtl-optimization/57320
4989 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
4990 the CFG after thread_prologue_and_epilogue_insns.
4992 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
4994 PR rtl-optimization/57189
4995 * lra-constraints.c (process_alt_operands): Disfavor spilling
4998 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
5000 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
5002 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5004 PR tree-optimization/59025
5006 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
5007 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
5009 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
5012 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
5013 calls of avr_out_plus_1.
5015 2014-03-13 Bin Cheng <bin.cheng@arm.com>
5017 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
5018 BB's single pred and update the father loop's latch info later.
5020 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5022 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
5026 (VEC_base): Likewise.
5027 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
5028 registers, we need to swap double words in little endian mode.
5030 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
5031 to be a container mode for 128-bit integer operations added in ISA
5032 2.07. Unlike TImode and PTImode, the preferred register set is
5033 the Altivec/VMX registers for the 128-bit operations.
5035 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
5037 (rs6000_split_128bit_ok_p): Likewise.
5039 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
5040 macros for creating ISA 2.07 normal and overloaded builtin
5041 functions with 3 arguments.
5042 (BU_P8V_OVERLOAD_3): Likewise.
5043 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
5044 for use as overloaded functions.
5045 (VPERM_1TI_UNS): Likewise.
5046 (VSEL_1TI): Likewise.
5047 (VSEL_1TI_UNS): Likewise.
5048 (ST_INTERNAL_1ti): Likewise.
5049 (LD_INTERNAL_1ti): Likewise.
5050 (XXSEL_1TI): Likewise.
5051 (XXSEL_1TI_UNS): Likewise.
5052 (VPERM_1TI): Likewise.
5053 (VPERM_1TI_UNS): Likewise.
5054 (XXPERMDI_1TI): Likewise.
5055 (SET_1TI): Likewise.
5056 (LXVD2X_V1TI): Likewise.
5057 (STXVD2X_V1TI): Likewise.
5058 (VEC_INIT_V1TI): Likewise.
5059 (VEC_SET_V1TI): Likewise.
5060 (VEC_EXT_V1TI): Likewise.
5061 (EQV_V1TI): Likewise.
5062 (NAND_V1TI): Likewise.
5063 (ORC_V1TI): Likewise.
5064 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
5065 added in ISA 2.07. Add both normal 'altivec' builtins, and the
5067 (VADDUQM): Likewise.
5068 (VSUBCUQ): Likewise.
5069 (VADDEUQM): Likewise.
5070 (VADDECUQ): Likewise.
5071 (VSUBEUQM): Likewise.
5072 (VSUBECUQ): Likewise.
5074 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5075 __int128_t and __uint128_t types.
5076 (__uint128_type): Likewise.
5077 (altivec_categorize_keyword): Add support for vector __int128_t,
5078 vector __uint128_t, vector __int128, and vector unsigned __int128
5079 as a container type for TImode operations that need to be done in
5080 VSX/Altivec registers.
5081 (rs6000_macro_to_expand): Likewise.
5082 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5083 to support 128-bit integer instructions vaddcuq, vadduqm,
5084 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5085 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5087 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5088 for V1TImode, and set up preferences to use VSX/Altivec registers.
5089 Setup VSX reload handlers.
5090 (rs6000_debug_reg_global): Likewise.
5091 (rs6000_init_hard_regno_mode_ok): Likewise.
5092 (rs6000_preferred_simd_mode): Likewise.
5093 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5094 (easy_altivec_constant): Likewise.
5095 (output_vec_const_move): Likewise.
5096 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5098 (rs6000_expand_vector_extract): Likewise.
5099 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
5101 (rs6000_const_vec): Add support for V1TImode.
5102 (rs6000_emit_le_vsx_load): Swap double words when loading or
5103 storing TImode/V1TImode.
5104 (rs6000_emit_le_vsx_store): Likewise.
5105 (rs6000_emit_le_vsx_move): Likewise.
5106 (rs6000_emit_move): Add support for V1TImode.
5107 (altivec_expand_ld_builtin): Likewise.
5108 (altivec_expand_st_builtin): Likewise.
5109 (altivec_expand_vec_init_builtin): Likewise.
5110 (altivec_expand_builtin): Likewise.
5111 (rs6000_init_builtins): Add support for V1TImode type. Add
5112 support for ISA 2.07 128-bit integer builtins. Define type names
5113 for the VSX/Altivec vector types.
5114 (altivec_init_builtins): Add support for overloaded vector
5115 functions with V1TImode type.
5116 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
5117 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
5119 (rs6000_split_128bit_ok_p): Likewise.
5120 (rs6000_handle_altivec_attribute): Create V1TImode from vector
5121 __int128_t and vector __uint128_t.
5123 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
5124 and mode attributes.
5131 (VS_scalar): Likewise.
5132 (VS_double): Likewise.
5133 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
5135 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
5136 we support the ISA 2.07 128-bit integer arithmetic instructions.
5137 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
5138 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
5139 and TImode types for use with the builtin functions.
5140 (V1TI_type_node): Likewise.
5141 (unsigned_V1TI_type_node): Likewise.
5142 (intTI_type_internal_node): Likewise.
5143 (uintTI_type_internal_node): Likewise.
5145 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
5146 128-bit builtin functions.
5147 (UNSPEC_VADDEUQM): Likewise.
5148 (UNSPEC_VADDECUQ): Likewise.
5149 (UNSPEC_VSUBCUQ): Likewise.
5150 (UNSPEC_VSUBEUQM): Likewise.
5151 (UNSPEC_VSUBECUQ): Likewise.
5152 (VM): Add V1TImode to vector mode iterators.
5154 (VI_unit): Likewise.
5155 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
5156 (altivec_vaddcuq): Likewise.
5157 (altivec_vsubuqm): Likewise.
5158 (altivec_vsubcuq): Likewise.
5159 (altivec_vaddeuqm): Likewise.
5160 (altivec_vaddecuq): Likewise.
5161 (altivec_vsubeuqm): Likewise.
5162 (altivec_vsubecuq): Likewise.
5164 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
5166 (BOOL_128): Likewise.
5167 (BOOL_REGS_OUTPUT): Likewise.
5168 (BOOL_REGS_OP1): Likewise.
5169 (BOOL_REGS_OP2): Likewise.
5170 (BOOL_REGS_UNARY): Likewise.
5171 (BOOL_REGS_AND_CR0): Likewise.
5173 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
5174 128-bit integer builtin support.
5175 (vec_vadduqm): Likewise.
5176 (vec_vaddecuq): Likewise.
5177 (vec_vaddeuqm): Likewise.
5178 (vec_vsubecuq): Likewise.
5179 (vec_vsubeuqm): Likewise.
5180 (vec_vsubcuq): Likewise.
5181 (vec_vsubuqm): Likewise.
5183 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5184 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
5185 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
5186 128-bit integer add/subtract to ISA 2.07.
5188 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
5190 * config/arc/arc.c (arc_predicate_delay_insns):
5191 Fix third argument passed to conditionalize_nonjump.
5193 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
5195 * config/aarch64/aarch64-builtins.c
5196 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
5197 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
5198 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
5199 instead of __builtin_lfloor.
5200 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
5202 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5204 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
5205 (tree_ssa_ifcombine_bb_1): New function.
5206 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
5207 is an empty forwarder block to then_bb or vice versa and then_bb
5208 and else_bb are effectively swapped.
5210 2014-03-12 Christian Bruel <christian.bruel@st.com>
5213 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
5214 REG_CFA_DEF_CFA note.
5215 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
5216 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
5218 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
5220 PR tree-optimization/60454
5221 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
5223 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5225 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
5226 Do not define target_cpu_default2 to generic.
5227 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
5228 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
5229 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
5231 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5232 Marc Glisse <marc.glisse@inria.fr>
5234 PR tree-optimization/60502
5235 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
5236 instead of build_low_bits_mask.
5238 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5241 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
5242 if there are multiple uses, but op doesn't live on E edge.
5243 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
5244 clobber stmts before __builtin_unreachable.
5246 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
5248 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
5249 hard_frame_pointer_rtx.
5250 * cse.c (cse_insn): Remove volatile check.
5251 * cselib.c (cselib_process_insn): Likewise.
5252 * dse.c (scan_insn): Likewise.
5254 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5256 * config/arc/arc.c (conditionalize_nonjump): New function,
5258 (arc_ifcvt): ... this.
5259 (arc_predicate_delay_insns): Use it.
5261 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5263 * config/arc/predicates.md (extend_operand): During/after reload,
5264 allow const_int_operand.
5265 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
5266 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
5267 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
5269 (umulsi3_highpart_i): Likewise.
5271 2014-03-11 Richard Biener <rguenther@suse.de>
5273 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
5274 Add asserts to guard possible wrong-code bugs.
5276 2014-03-11 Richard Biener <rguenther@suse.de>
5278 PR tree-optimization/60429
5279 PR tree-optimization/60485
5280 * tree-ssa-structalias.c (set_union_with_increment): Properly
5281 take into account all fields that overlap the shifted vars.
5282 (do_sd_constraint): Likewise.
5283 (do_ds_constraint): Likewise.
5284 (get_constraint_for_ptr_offset): Likewise.
5286 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
5288 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
5289 (nios2_compute_frame_layout):
5290 Add calculation of cfun->machine->fp_save_offset.
5291 (nios2_expand_prologue): Correct setting of frame pointer register
5293 (nios2_expand_epilogue): Update recovery of stack pointer from
5294 frame pointer accordingly.
5295 (nios2_initial_elimination_offset): Update calculation of offset
5296 for eliminating to HARD_FRAME_POINTER_REGNUM.
5298 2014-03-10 Jakub Jelinek <jakub@redhat.com>
5301 * ipa.c (symtab_remove_unreachable_nodes): Don't call
5302 cgraph_get_create_node on VAR_DECLs.
5304 2014-03-10 Richard Biener <rguenther@suse.de>
5307 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
5309 2014-03-08 Douglas B Rupp <rupp@gnat.com>
5311 * config/vms/vms.opt (vms_float_format): New variable.
5313 2014-03-08 Tobias Burnus <burnus@net-b.de>
5315 * doc/invoke.texi (-fcilkplus): Update implementation status.
5317 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
5318 Richard Biener <rguenther@suse.de>
5320 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
5321 consistently accross all TUs.
5322 (run_gcc): Enable -fshort-double automatically at link at link-time
5323 and disallow override.
5325 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
5328 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
5329 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
5330 if they can't be used.
5332 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5334 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
5335 for Solaris 11/x86 ld.
5336 * configure: Regenerate.
5338 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5340 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
5341 (LIB_TLS_SPEC): Save as ld_tls_libs.
5342 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
5343 (HAVE_AS_IX86_TLSLDM): New test.
5344 * configure, config.in: Regenerate.
5345 * config/i386/i386.c (legitimize_tls_address): Fall back to
5346 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
5347 cannot support TLS_MODEL_LOCAL_DYNAMIC.
5348 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
5349 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
5351 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
5353 * common.opt (fira-loop-pressure): Mark as optimization.
5355 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
5357 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
5358 an OpenMP mappable type.
5360 2014-03-06 Matthias Klose <doko@ubuntu.com>
5362 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
5363 MULTILIB_OSDIRNAMES is not defined.
5365 2014-03-06 Jakub Jelinek <jakub@redhat.com>
5366 Meador Inge <meadori@codesourcery.com>
5369 * config/arm/arm.c (arm_tls_symbol_p): Remove.
5370 (arm_legitimize_address): Call legitimize_tls_address for any
5371 arm_tls_referenced_p expression, handle constant addend. Call it
5372 before testing for !TARGET_ARM.
5373 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
5375 2014-03-06 Richard Biener <rguenther@suse.de>
5381 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5383 * tree-streamer.c (record_common_node): Assert we don't record
5384 nodes with type double.
5385 (preload_common_node): Skip type double, complex double and double
5386 pointer since it is now frontend dependent due to fshort-double option.
5388 2014-03-06 Richard Biener <rguenther@suse.de>
5390 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
5391 or -fno-lto is specified and the linker has full plugin support.
5392 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
5393 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
5394 * lto-wrapper.c (merge_and_complain): Merge compile-time
5395 optimization levels.
5396 (run_gcc): And pass it through to the link options.
5398 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
5402 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5404 * cselib.c (remove_useless_values): Skip to avoid quadratic
5405 behavior if the condition moved from...
5406 (cselib_process_insn): ... here holds.
5408 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5411 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
5412 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
5415 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
5416 (TM_H): Add x86-tune.def.
5418 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5420 * config/aarch64/aarch64.c (generic_tunings):
5421 Use cortexa57_extra_costs.
5423 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5426 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
5427 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
5428 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
5431 2014-03-04 Heiher <r@hev.cc>
5433 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
5434 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
5436 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
5438 * config/i386/predicates.md (const2356_operand): Change to ...
5439 (const2367_operand): ... this.
5440 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
5442 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5443 (*avx512pf_scatterpf<mode>sf): Ditto.
5444 (avx512pf_scatterpf<mode>df): Ditto.
5445 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5446 (*avx512pf_scatterpf<mode>df): Ditto.
5447 * config/i386/i386.c (ix86_expand_builtin): Update
5448 incorrect hint operand error message.
5450 2014-03-04 Richard Biener <rguenther@suse.de>
5452 * lto-section-in.c (lto_get_section_data): Fix const cast.
5454 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5456 * tree-streamer.c (record_common_node): Assert we don't record
5457 nodes with type double.
5458 (preload_common_node): Skip type double, complex double and double
5459 pointer since it is now frontend dependent due to fshort-double option.
5461 2014-03-04 Richard Biener <rguenther@suse.de>
5464 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
5465 (lto_input_toplevel_asms): Likewise.
5466 * lto-section-in.c (lto_get_section_data): Instead do it here
5469 2014-03-04 Richard Biener <rguenther@suse.de>
5471 PR tree-optimization/60382
5472 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
5473 dead PHIs a reduction.
5475 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
5477 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
5479 (_mm_prefetch): Move out of GCC target("sse") pragma.
5480 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
5481 GCC target("prfchw") pragma.
5482 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
5484 * config/i386/i386.c (ix86_option_override_internal): Enable
5485 -mprfchw with -mprefetchwt1.
5487 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5489 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
5492 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5494 * opts.h (CL_PCH_IGNORE): Define.
5495 * targhooks.c (option_affects_pch_p):
5496 Return false for options that have CL_PCH_IGNORE set.
5497 * opt-functions.awk: Process PchIgnore.
5498 * doc/options.texi: Document PchIgnore.
5500 * config/arc/arc.opt (misize): Add PchIgnore property.
5502 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5504 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
5505 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
5506 constraint on constants to permit them being loaded into
5507 GENERAL_REGS or BASE_REGS.
5509 2014-03-03 Nick Clifton <nickc@redhat.com>
5511 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
5512 anti-cacnonical alternatives.
5513 (negandhi3_real): New pattern.
5514 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
5516 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5518 * config/avr/avr-mcus.def: Remove atxmega16x1.
5519 * config/avr/avr-tables.opt: Regenerate.
5520 * config/avr/t-multilib: Regenerate.
5521 * doc/avr-mmcu.texi: Regenerate.
5523 2014-03-03 Tobias Grosser <tobias@grosser.es>
5524 Mircea Namolaru <mircea.namolaru@inria.fr>
5526 PR tree-optimization/58028
5527 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
5530 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5532 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
5533 not handled by recognizers.
5535 2014-03-03 Jakub Jelinek <jakub@redhat.com>
5538 * function.c (expand_function_end): Don't emit
5539 clobber_return_register sequence if clobber_after is a BARRIER.
5540 * cfgexpand.c (construct_exit_block): Append instructions before
5541 return_label to prev_bb.
5543 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5545 * config/rs6000/constraints.md: Document reserved use of "wc".
5547 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5550 * ipa.c (function_and_variable_visibility): When dissolving comdat
5551 group, also set all symbols to local.
5553 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5558 2013-12-14 Jan Hubicka <jh@suse.cz>
5560 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
5562 2014-03-02 Jon Beniston <jon@beniston.com>
5568 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
5569 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
5570 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
5571 (simple_return, *simple_return): New patterns
5572 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
5573 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
5575 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
5577 * dwarf2out.c (gen_subprogram_die): Tidy.
5579 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
5582 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
5583 (*mov_t_msb_neg_negc): ... this new insn.
5585 2014-02-28 Jason Merrill <jason@redhat.com>
5588 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
5591 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
5594 * dwarf2out.c (decltype_auto_die): New static.
5595 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
5596 (gen_type_die_with_usage): Handle 'decltype(auto)'.
5597 (is_cxx_auto): Likewise.
5599 2014-02-28 Ian Bolton <ian.bolton@arm.com>
5601 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
5602 we are not using general regs only.
5604 2014-02-28 Richard Biener <rguenther@suse.de>
5607 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
5608 previous fix and only allow to remove trivial pre-headers
5609 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
5610 (remove_forwarder_block): Properly update the latch of a loop.
5612 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5615 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
5616 (cselib_preserved_hash_table): New.
5617 (preserve_constants_and_equivs): Move preserved vals to it.
5618 (cselib_find_slot): Look it up first.
5619 (cselib_init): Initialize it.
5620 (cselib_finish): Release it.
5621 (dump_cselib_table): Dump it.
5623 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5626 * cselib.c (remove_useless_values): Skip to avoid quadratic
5627 behavior if the condition moved from...
5628 (cselib_process_insn): ... here holds.
5630 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5633 * var-tracking.c (vt_initialize): Apply the same condition to
5634 preserve the CFA base value.
5636 2014-02-28 Joey Ye <joey.ye@arm.com>
5639 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
5640 if reload in progress or completed.
5642 2014-02-28 Tobias Burnus <burnus@net-b.de>
5645 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
5648 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
5650 * doc/tm.texi.in (Condition Code Status): Update documention for
5651 relative locations of cc0-setter and cc0-user.
5653 2014-02-27 Jeff Law <law@redhat.com>
5655 PR rtl-optimization/52714
5656 * combine.c (try_combine): When splitting an unrecognized PARALLEL
5657 into two independent simple sets, if I3 is a jump, ensure the
5658 pattern we place into I3 is a (set (pc) ...).
5660 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
5661 Jeff Law <law@redhat.com>
5663 PR rtl-optimization/49847
5664 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
5665 are in different blocks.
5666 * doc/tm.texi (Condition Code Status): Update documention for
5667 relative locations of cc0-setter and cc0-user.
5669 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
5672 * lra.c (lra_emit_add): Check SUBREG too.
5674 2014-02-27 Andreas Schwab <schwab@suse.de>
5676 * config/m68k/m68k.c (m68k_option_override): Disable
5677 -flive-range-shrinkage for classic m68k.
5678 (m68k_override_options_after_change): Likewise.
5680 2014-02-27 Marek Polacek <polacek@redhat.com>
5683 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
5684 -Wmaybe-uninitialized.
5686 2014-02-27 Alan Modra <amodra@gmail.com>
5689 * reload1.c (emit_input_reload_insns): When reload_override_in,
5690 set old to rl->in_reg when rl->in_reg is a subreg.
5692 2014-02-26 Richard Biener <rguenther@suse.de>
5695 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
5697 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
5699 * common/config/i386/predicates.md (const1256_operand): Remove.
5700 (const2356_operand): New.
5701 (const_1_to_2_operand): Remove.
5702 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5703 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5704 (*avx512pf_gatherpf<mode>sf): Ditto.
5705 (avx512pf_gatherpf<mode>df): Ditto.
5706 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5707 (*avx512pf_gatherpf<mode>df): Ditto.
5708 (avx512pf_scatterpf<mode>sf): Ditto.
5709 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5710 (*avx512pf_scatterpf<mode>sf): Ditto.
5711 (avx512pf_scatterpf<mode>df): Ditto.
5712 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5713 (*avx512pf_scatterpf<mode>df): Ditto.
5714 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
5716 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
5718 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
5719 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
5720 (_mm512_mask_testn_epi64_mask): Move to ...
5721 * config/i386/avx512cdintrin.h: Here.
5722 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
5723 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
5724 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
5725 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
5726 TARGET_AVX512F from TARGET_AVX512CD.
5728 2014-02-26 Richard Biener <rguenther@suse.de>
5731 * ipa.c (walk_polymorphic_call_targets): Properly guard
5732 call to inline_update_overall_summary.
5734 2014-02-26 Bin Cheng <bin.cheng@arm.com>
5737 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
5738 and latches only if requested. Fix latch if it is removed.
5739 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
5740 LOOPS_HAVE_PREHEADERS.
5742 2014-02-25 Andrew Pinski <apinski@cavium.com>
5744 * builtins.c (expand_builtin_thread_pointer): Create a new target
5745 when the target is NULL.
5747 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
5749 PR rtl-optimization/60317
5750 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
5751 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
5752 * lra-assigns.c: Include params.h.
5753 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
5754 other reload pseudos considerations.
5756 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5758 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
5759 to use canonical form for nor<mode>3.
5761 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5764 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
5767 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
5769 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
5770 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
5771 (ix86_handle_option): Handle OPT_mprefetchwt1.
5772 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
5773 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
5775 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
5776 OPTION_MASK_ISA_PREFETCHWT1.
5777 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
5778 (PTA_PREFETCHWT1): New.
5779 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
5780 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
5781 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
5782 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
5783 (*prefetch_avx512pf_<mode>_: Change into ...
5784 (*prefetch_prefetchwt1_<mode>: This.
5785 * config/i386/i386.opt (mprefetchwt1): New.
5786 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
5787 (_mm_prefetch): Handle intent to write.
5788 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
5790 2014-02-25 Richard Biener <rguenther@suse.de>
5793 * emit-rtl.c (mem_attrs_htab): Remove.
5794 (mem_attrs_htab_hash): Likewise.
5795 (mem_attrs_htab_eq): Likewise.
5796 (set_mem_attrs): Always allocate new mem-attrs when something changed.
5797 (init_emit_once): Do not allocate mem_attrs_htab.
5799 2014-02-25 Richard Biener <rguenther@suse.de>
5802 * lto-opts.c (lto_write_options): Output non-explicit conservative
5803 -fwrapv, -fno-trapv and -fno-strict-overflow.
5804 * lto-wrapper.c (merge_and_complain): Handle merging those options.
5805 (run_gcc): And pass them through.
5807 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
5809 * sel-sched.c (calculate_new_fences): New parameter ptime.
5810 Calculate it as a maximum over all fence cycles.
5811 (sel_sched_region_2): Adjust the call to calculate_new_fences.
5812 Print the final schedule timing when sched_verbose.
5814 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
5816 PR rtl-optimization/60292
5817 * sel-sched.c (fill_vec_av_set): Do not reset target availability
5818 bit fot the fence instruction.
5820 2014-02-24 Alangi Derick <alangiderick@gmail.com>
5822 * calls.h: Fix typo in comment.
5824 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
5826 * config/pa/pa.c (pa_output_move_double): Don't valididate when
5827 adjusting offsetable addresses.
5829 2014-02-24 Guozhi Wei <carrot@google.com>
5831 * sparseset.h (sparseset_pop): Fix the wrong index.
5833 2014-02-24 Walter Lee <walt@tilera.com>
5835 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
5836 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
5838 * common/config/tilegx/tilegx-common.c
5839 (TARGET_DEFAULT_TARGET_FLAGS): Define.
5840 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
5842 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
5843 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
5844 (tilegx_gimplify_va_arg_expr): Handle big endian.
5845 (tilegx_expand_unaligned_load): Ditto.
5846 (tilegx_expand_unaligned_store): Ditto.
5847 (TARGET_RETURN_IN_MSB): New.
5848 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
5849 (TARGET_ENDIAN_DEFAULT): New.
5850 (TARGET_BIG_ENDIAN): Handle big endian.
5851 (BYTES_BIG_ENDIAN): Ditto.
5852 (WORDS_BIG_ENDIAN): Ditto.
5853 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
5856 * config/tilegx/tilegx.md (extv): Handle big endian.
5858 (insn_st<n>): Ditto.
5859 (insn_st<n>_add<bitsuffix>): Ditto.
5860 (insn_stnt<n>): Ditto.
5861 (insn_stnt<n>_add<bitsuffix>):Ditto.
5862 (vec_interleave_highv8qi): Handle big endian.
5863 (vec_interleave_highv8qi_be): New.
5864 (vec_interleave_highv8qi_le): New.
5865 (insn_v1int_h): Handle big endian.
5866 (vec_interleave_lowv8qi): Handle big endian.
5867 (vec_interleave_lowv8qi_be): New.
5868 (vec_interleave_lowv8qi_le): New.
5869 (insn_v1int_l): Handle big endian.
5870 (vec_interleave_highv4hi): Handle big endian.
5871 (vec_interleave_highv4hi_be): New.
5872 (vec_interleave_highv4hi_le): New.
5873 (insn_v2int_h): Handle big endian.
5874 (vec_interleave_lowv4hi): Handle big endian.
5875 (vec_interleave_lowv4hi_be): New.
5876 (vec_interleave_lowv4hi_le): New.
5877 (insn_v2int_l): Handle big endian.
5878 (vec_interleave_highv2si): Handle big endian.
5879 (vec_interleave_highv2si_be): New.
5880 (vec_interleave_highv2si_le): New.
5881 (insn_v4int_h): Handle big endian.
5882 (vec_interleave_lowv2si): Handle big endian.
5883 (vec_interleave_lowv2si_be): New.
5884 (vec_interleave_lowv2si_le): New.
5885 (insn_v4int_l): Handle big endian.
5886 * config/tilegx/tilegx.opt (mbig-endian): New option.
5887 (mlittle-endian): New option.
5888 * doc/install.texi: Document tilegxbe-linux.
5889 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
5891 2014-02-24 Martin Jambor <mjambor@suse.cz>
5894 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
5895 there are no parameter descriptors.
5897 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
5899 PR rtl-optimization/60268
5900 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
5901 initialization to ...
5902 (sched_rgn_init): ... here.
5903 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
5905 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5907 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
5910 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
5912 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
5915 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5917 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
5918 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
5920 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5922 * config/microblaze/predicates.md: Add cmp_op predicate.
5923 * config/microblaze/microblaze.md: Add branch_compare instruction
5924 which uses cmp_op predicate and emits cmp insn before branch.
5925 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
5926 to microblaze_expand_conditional_branch and consolidate logic.
5927 (microblaze_expand_conditional_branch): emit branch_compare
5928 insn instead of handling cmp op separate from branch insn.
5930 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5932 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
5935 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5937 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
5938 define_insn with define_expand and new define_insn
5939 *altivec_lve<VI_char>x_internal.
5940 (altivec_stve<VI_char>x): Replace define_insn with define_expand
5941 and new define_insn *altivec_stve<VI_char>x_internal.
5942 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
5944 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
5946 (altivec_expand_stvex_be): New function.
5948 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
5950 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
5951 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
5952 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
5953 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
5955 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
5958 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
5959 instead of emit_move_insn.
5961 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5963 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
5965 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
5966 gen_altivec_vsumsws.
5968 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5970 * config/rs6000/altivec.md (altivec_lvxl): Rename as
5971 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
5972 (altivec_lvxl_<mode>): New define_expand incorporating
5973 -maltivec=be semantics where needed.
5974 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
5975 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
5976 semantics where needed.
5977 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
5978 (altivec_stvx_<mode>): New define_expand incorporating
5979 -maltivec=be semantics where needed.
5980 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
5981 VM2 iterator instead of V4SI.
5982 (altivec_stvxl_<mode>): New define_expand incorporating
5983 -maltivec=be semantics where needed.
5984 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
5985 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
5986 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
5987 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
5988 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
5989 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
5990 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
5991 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
5992 ALTIVEC_BUILTIN_STVXL.
5993 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
5994 (altivec_expand_stvx_be): Likewise.
5995 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
5996 (altivec_expand_lvx_be): Likewise.
5997 (altivec_expand_stvx_be): Likewise.
5998 (altivec_expand_builtin): Add cases for
5999 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
6000 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
6001 (altivec_init_builtins): Add definitions for
6002 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
6003 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
6005 2014-02-21 Catherine Moore <clm@codesourcery.com>
6007 * doc/invoke.texi (mvirt, mno-virt): Document.
6008 * config/mips/mips.opt (mvirt): New option.
6009 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
6011 2014-02-21 Richard Biener <rguenther@suse.de>
6013 PR tree-optimization/60276
6014 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
6015 (STMT_VINFO_MIN_NEG_DIST): New macro.
6016 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
6017 STMT_VINFO_MIN_NEG_DIST.
6018 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
6019 made for negative dependence distances still hold.
6021 2014-02-21 Richard Biener <rguenther@suse.de>
6024 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
6025 DECL_INITIAL for globals not in the current function context.
6027 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6029 PR tree-optimization/56490
6030 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
6031 * tree-ssa-uninit.c: Include params.h.
6032 (compute_control_dep_chain): Add num_calls argument, return false
6033 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
6034 num_calls to recursive call.
6035 (find_predicates): Change dep_chain into normal array,
6036 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
6037 variable and adjust compute_control_dep_chain caller.
6038 (find_def_preds): Likewise.
6040 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
6042 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
6043 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
6045 2014-02-21 Nick Clifton <nickc@redhat.com>
6047 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
6048 (pushhi1): Likewise.
6049 (popqi1): Add mode to pre_dec.
6052 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6054 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
6055 mode for mask of V8SFmode permutation.
6057 2014-02-20 Richard Henderson <rth@redhat.com>
6060 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
6061 a new pseudo for OLDVAL.
6063 2014-02-20 Jakub Jelinek <jakub@redhat.com>
6066 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
6067 gen_reg_rtx if d->testing_p.
6068 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
6069 if d->testing_p and we will certainly return true.
6070 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6073 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6075 * emit-rtl.c (gen_reg_rtx): Assert that
6076 crtl->emit.regno_pointer_align_length is non-zero.
6078 2014-02-20 Richard Henderson <rth@redhat.com>
6081 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6082 on failure the store back into EXPECT.
6084 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6085 Sandra Loosemore <sandra@codesourcery.com>
6087 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6088 * config/nios2/nios2.c (nios2_function_profiler): Add
6089 -fPIC (flag_pic == 2) support.
6090 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6091 (nios2_large_offset_p): New function.
6092 (nios2_unspec_reloc_p): Move up position, update to use
6093 nios2_large_offset_p.
6094 (nios2_unspec_address): Remove function.
6095 (nios2_unspec_offset): New function.
6096 (nios2_large_got_address): New function.
6097 (nios2_got_address): Add large offset support.
6098 (nios2_legitimize_tls_address): Update usage of removed and new
6100 (nios2_symbol_binds_local_p): New function.
6101 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
6102 (nios2_legitimize_address): Update to use nios2_large_offset_p.
6103 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
6104 (nios2_print_operand): Merge H/L processing, add hiadj/lo
6105 processing for (const (unspec ...)).
6106 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
6108 2014-02-20 Richard Biener <rguenther@suse.de>
6110 * tree-cfg.c (replace_uses_by): Mark altered BBs before
6111 doing the substitution.
6112 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
6114 2014-02-20 Martin Jambor <mjambor@suse.cz>
6117 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
6118 info when checking whether lattices are bottom.
6120 2014-02-20 Richard Biener <rguenther@suse.de>
6123 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
6126 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
6129 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
6130 parameter specifying the scaling.
6131 (inline_call): Update.
6132 (want_inline_recursively): Guard division by zero.
6133 (recursive_inlining): Update.
6134 * ipa-inline.h (clone_inlined_nodes): Update.
6136 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6139 * config/i386/i386.c (classify_argument): Pass structures of size
6140 64 bytes or less in register.
6142 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6143 Kirill Yukhin <kirill.yukhin@intel.com>
6145 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
6146 (_mm_rcp28_round_ss): Ditto.
6147 (_mm_rsqrt28_round_sd): Ditto.
6148 (_mm_rsqrt28_round_ss): Ditto.
6149 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
6150 (_mm_rcp14_round_ss): Ditto.
6151 (_mm_rsqrt14_round_sd): Ditto.
6152 (_mm_rsqrt14_round_ss): Ditto.
6153 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
6154 the first input operand, get rid of match_dup.
6155 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
6157 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
6159 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
6160 operand as the first input operand, set type attribute.
6161 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
6163 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
6164 operand as the first input operand, set type attribute.
6166 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6168 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
6171 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
6174 * config/i386/i386.c (construct_container): Remove TFmode check
6175 for X86_64_INTEGER_CLASS.
6177 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
6180 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
6181 only when -Wpsabi is enabled.
6183 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
6186 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
6187 passing arrays in registers are the same as for structs, so remove the
6188 special case for them.
6190 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
6192 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
6193 destination type, extract only the valid bits if the source type is not
6194 integral and has a different mode.
6196 2014-02-19 Richard Biener <rguenther@suse.de>
6199 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
6202 2014-02-19 Richard Biener <rguenther@suse.de>
6205 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
6206 (ipa_modify_call_arguments): Emit an argument load explicitely and
6207 preserve virtual SSA form there and for the replacement call.
6208 Do not update SSA form nor free dominance info.
6210 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6212 * ipa.c (function_and_variable_visibility): Also clear WEAK
6213 flag when disolving COMDAT_GROUP.
6215 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6217 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
6218 * ipa-prop.c (ipa_set_jf_known_type): Return early when
6220 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
6221 do more sanity checks.
6222 (detect_type_change): Return true when giving up early.
6223 (compute_complex_assign_jump_func): Fix type parameter of
6224 ipa_set_ancestor_jf.
6225 (compute_complex_ancestor_jump_func): Likewise.
6226 (update_jump_functions_after_inlining): Fix updating of
6228 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
6230 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6232 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
6233 inline clones when edge disappears.
6235 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6238 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
6239 Split 64-bit moves into 2 patterns. Do not allow the use of
6240 direct move for TDmode in little endian, since the decimal value
6241 has little endian bytes within a word, but the 64-bit pieces are
6242 ordered in a big endian fashion, and normal subreg's of TDmode are
6244 (mov<mode>_64bit_dm): Likewise.
6245 (movtd_64bit_nodm): Likewise.
6247 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6249 PR tree-optimization/60174
6250 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
6251 statement of an SSA_NAME that occurs in an abnormal PHI node.
6253 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6256 * final.c (SEEN_BB): Remove.
6257 (SEEN_NOTE, SEEN_EMITTED): Renumber.
6258 (final_scan_insn): Don't force_source_line on second
6259 NOTE_INSN_BASIC_BLOCK.
6261 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
6264 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
6265 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
6266 (type_natural_mode): Warn ABI change when %zmm register is not
6267 available for AVX512F vector value passing.
6269 2014-02-18 Kai Tietz <ktietz@redhat.com>
6272 * config/i386/i386.c (ix86_expand_prologue): Use value in
6273 rax register as displacement when restoring %r10 or %rax.
6274 Fix wrong offset when restoring both registers.
6276 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6278 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
6279 assertion with conditional return.
6281 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6282 Uros Bizjak <ubizjak@gmail.com>
6285 * config/i386/driver-i386.c (host_detect_local_cpu): If
6286 YMM state is not saved by the OS, also clear has_f16c. Move
6287 CPUID 0x80000001 handling before YMM state saving checking.
6289 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
6291 PR rtl-optimization/58960
6292 * haifa-sched.c (alloc_global_sched_pressure_data): New,
6293 factored out from ...
6294 (sched_init): ... here.
6295 (free_global_sched_pressure_data): New, factored out from ...
6296 (sched_finish): ... here.
6297 * sched-int.h (free_global_sched_pressure_data): Declare.
6298 * sched-rgn.c (nr_regions_initial): New static global.
6299 (haifa_find_rgns): Initialize it.
6300 (schedule_region): Disable sched-pressure for the newly
6303 2014-02-17 Richard Biener <rguenther@suse.de>
6305 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
6306 release SSA defs of pattern stmts.
6308 2014-02-17 Richard Biener <rguenther@suse.de>
6310 * tree-inline.c (expand_call_inline): Release the virtual
6311 operand defined by the call we are about to inline.
6313 2014-02-17 Richard Biener <rguenther@suse.de>
6315 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
6317 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
6318 Ilya Tocar <ilya.tocar@intel.com>
6320 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
6321 arguments order in builtin.
6322 (_mm512_permutexvar_epi64): Ditto.
6323 (_mm512_mask_permutexvar_epi64): Ditto
6324 (_mm512_maskz_permutexvar_epi32): Ditto
6325 (_mm512_permutexvar_epi32): Ditto
6326 (_mm512_mask_permutexvar_epi32): Ditto
6328 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6330 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
6331 (p8_vmrgow): Likewise.
6333 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6335 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
6338 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
6341 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
6342 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
6343 into 64-bit and 32-bit moves. On 64-bit moves, add support for
6344 using direct move instructions on ISA 2.07. Also adjust
6345 instruction length for 64-bit.
6346 (mov<mode>_64bit, TFmode/TDmode): Likewise.
6347 (mov<mode>_32bit, TFmode/TDmode): Likewise.
6349 2014-02-15 Alan Modra <amodra@gmail.com>
6353 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
6354 find_replacement on parts of insn rtl that might be reloaded.
6356 2014-02-15 Richard Biener <rguenther@suse.de>
6358 PR tree-optimization/60183
6359 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
6360 (tree_ssa_phiprop): Calculate and free post-dominators.
6362 2014-02-14 Jeff Law <law@redhat.com>
6364 PR rtl-optimization/60131
6365 * ree.c (get_extended_src_reg): New function.
6366 (combine_reaching_defs): Use it rather than assuming location of REG.
6367 (find_and_remove_re): Verify first operand of extension is
6368 a REG before adding the insns to the copy list.
6370 2014-02-14 Roland McGrath <mcgrathr@google.com>
6372 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
6373 * configure: Regenerated.
6374 * config.in: Regenerated.
6375 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
6376 instead of ASM_SHORT.
6378 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
6379 Richard Earnshaw <rearnsha@arm.com>
6381 PR rtl-optimization/59535
6382 * lra-constraints.c (process_alt_operands): Encourage alternative
6383 when unassigned pseudo class is superset of the alternative class.
6384 (inherit_reload_reg): Don't inherit when optimizing for code size.
6385 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
6386 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
6387 modes not less than 4 for Thumb1.
6389 2014-02-14 Kyle McMartin <kyle@redhat.com>
6392 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
6394 2014-02-14 Richard Biener <rguenther@suse.de>
6396 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
6397 (get_frame_arg): Drop the assert with langhook types_compatible_p.
6398 Do not strip INDIRECT_REFs.
6400 2014-02-14 Richard Biener <rguenther@suse.de>
6403 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
6404 DECL_FUNCTION_SPECIFIC_TARGET.
6405 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
6406 * tree-streamer-out.c (pack_ts_target_option): Remove.
6407 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
6408 (write_ts_function_decl_tree_pointers): Do not stream
6409 DECL_FUNCTION_SPECIFIC_TARGET.
6410 * tree-streamer-in.c (unpack_ts_target_option): Remove.
6411 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
6412 (lto_input_ts_function_decl_tree_pointers): Do not stream
6413 DECL_FUNCTION_SPECIFIC_TARGET.
6415 2014-02-14 Jakub Jelinek <jakub@redhat.com>
6417 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
6418 (get_initial_def_for_induction, vectorizable_induction): Ignore
6419 debug stmts when looking for exit_phi.
6420 (vectorizable_live_operation): Fix up condition.
6422 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6424 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
6425 nreverse() because it changes the content of original tree list.
6427 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6429 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
6430 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
6432 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6434 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
6435 GNU coding standards.
6437 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6440 * dwarf2out.c (gen_subprogram_die): Don't call
6441 add_calling_convention_attribute if subr_die is old_die.
6443 2014-02-13 Sharad Singhai <singhai@google.com>
6445 * doc/optinfo.texi: Fix order of nodes.
6447 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
6449 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
6450 operands[2], not operands[3].
6452 2014-02-13 Richard Biener <rguenther@suse.de>
6455 * doc/install.texi (ISL): Update recommended version to 0.12.2,
6456 mention the possibility of an in-tree build.
6457 (CLooG): Update recommended version to 0.18.1, mention the
6458 possibility of an in-tree build and clarify that the ISL
6459 bundled with CLooG does not work.
6461 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6464 * expr.c (compress_float_constant): If x is a hard register,
6465 extend into a pseudo and then move to x.
6467 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
6469 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
6470 caused by bad second argument to warning_at() with -mhotpatch and
6471 nested functions (e.g. with gfortran).
6473 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
6475 * opts.c (option_name): Remove "enabled by default" rider.
6477 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
6479 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
6481 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
6482 Uros Bizjak <ubizjak@gmail.com>
6485 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
6486 * configure: Regenerated.
6488 2014-02-12 Richard Biener <rguenther@suse.de>
6490 * vec.c (vec_prefix::calculate_allocation): Move as
6491 inline variant to vec.h.
6492 (vec_prefix::calculate_allocation_1): New out-of-line version.
6493 * vec.h (vec_prefix::calculate_allocation_1): Declare.
6494 (vec_prefix::m_has_auto_buf): Rename to ...
6495 (vec_prefix::m_using_auto_storage): ... this.
6496 (vec_prefix::calculate_allocation): Inline the easy cases
6497 and dispatch to calculate_allocation_1 which doesn't need the
6499 (va_heap::reserve): Use gcc_checking_assert.
6500 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
6501 m_using_auto_storage.
6502 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
6504 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
6505 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
6506 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
6508 2014-02-12 Richard Biener <rguenther@suse.de>
6510 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
6511 when we found a dependence.
6513 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
6515 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
6517 (maybe_fold_stmt): ... into this new function.
6518 * omp-low.c (lower_omp): Update comment.
6520 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
6523 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
6526 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
6528 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
6529 identifiers in comments.
6530 (cortexa53_extra_costs): Likewise.
6531 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
6532 (cortexa7_extra_costs): Likewise.
6533 (cortexa12_extra_costs): Likewise.
6534 (cortexa15_extra_costs): Likewise.
6535 (v7m_extra_costs): Likewise.
6537 2014-02-12 Richard Biener <rguenther@suse.de>
6540 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
6541 of posix_memalign being successful.
6542 (lower_stmt): Restrict lowering of posix_memalign to when
6543 -ftree-bit-ccp is enabled.
6545 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6547 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
6549 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
6551 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
6553 PR rtl-optimization/60116
6554 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
6555 other_insn once the combination has been validated.
6557 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
6560 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
6562 * ipa-devirt.c: Include demangle.h
6563 (odr_violation_reported): New static variable.
6564 (add_type_duplicate): Update odr_violations.
6565 (maybe_record_node): Add completep parameter; update it.
6566 (record_target_from_binfo): Add COMPLETEP parameter;
6567 update it as needed.
6568 (possible_polymorphic_call_targets_1): Likewise.
6569 (struct polymorphic_call_target_d): Add nonconstruction_targets;
6570 rename FINAL to COMPLETE.
6571 (record_targets_from_bases): Sanity check we found the binfo;
6572 fix COMPLETEP updating.
6573 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
6574 parameter, fix computing of COMPLETEP.
6575 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
6576 at LTO time do demangling.
6577 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
6578 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
6580 (gimple_get_virt_method_for_binfo): Likewise.
6581 * gimple-fold.h (gimple_get_virt_method_for_binfo,
6582 gimple_get_virt_method_for_vtable): Update prototypes.
6584 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
6587 * genautomata.c (add_presence_absence): Fix typo with
6588 {final_}presence_list.
6590 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
6593 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
6594 for VSX/Altivec vectors that land in GPR registers.
6596 2014-02-11 Richard Henderson <rth@redhat.com>
6597 Jakub Jelinek <jakub@redhat.com>
6600 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
6601 around drhs if type conversion to lacc->type is not useless.
6603 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6605 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
6607 (cortex-a57.cortex-a53): Likewise.
6608 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
6610 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6612 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
6615 2014-02-11 Renlin Li <Renlin.Li@arm.com>
6617 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
6618 add_options_for_arm_vfp3.
6620 2014-02-11 Jeff Law <law@redhat.com>
6623 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
6624 object with an undesirable mode.
6626 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6629 * config/i386/sol2-9.h: New file.
6630 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
6631 *-*-solaris2.9*): Use it.
6633 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6635 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
6636 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
6638 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6640 * config/microblaze/microblaze.c: Extend mcpu version format
6642 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
6644 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
6646 2014-02-10 Richard Henderson <rth@redhat.com>
6649 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
6650 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
6651 ms-abi vs -mno-accumulate-outgoing-args.
6652 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
6653 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
6656 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
6659 * cfgexpand.c (expand_asm_operands): Attach source location to
6660 ASM_INPUT rtx objects.
6661 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
6663 2014-02-10 Nick Clifton <nickc@redhat.com>
6665 * config/mn10300/mn10300.c (popcount): New function.
6666 (mn10300_expand_prologue): Include saved registers in stack usage
6669 2014-02-10 Jeff Law <law@redhat.com>
6672 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
6673 when changing the SET_DEST of a prior insn to avoid an input reload.
6675 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6677 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
6678 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
6679 -mcall-openbsd, or -mcall-linux.
6680 (CC1_ENDIAN_BIG_SPEC): Remove.
6681 (CC1_ENDIAN_LITTLE_SPEC): Remove.
6682 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
6683 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
6684 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
6685 and %cc1_endian_default.
6686 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
6688 2014-02-10 Richard Biener <rguenther@suse.de>
6690 PR tree-optimization/60115
6691 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
6692 MEM_REF handling. Properly verify that the accesses are not
6693 out of the objects bound.
6695 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6697 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
6700 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
6702 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
6703 proper constants and fix formatting.
6704 (possible_polymorphic_call_targets): Fix formatting.
6706 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
6707 Ilya Tocar <ilya.tocar@intel.com>
6709 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
6710 (_mm512_loadu_epi32): Renamed into...
6711 (_mm512_loadu_si512): This.
6712 (_mm512_storeu_epi32): Renamed into...
6713 (_mm512_storeu_si512): This.
6714 (_mm512_maskz_ceil_ps): Removed.
6715 (_mm512_maskz_ceil_pd): Ditto.
6716 (_mm512_maskz_floor_ps): Ditto.
6717 (_mm512_maskz_floor_pd): Ditto.
6718 (_mm512_floor_round_ps): Ditto.
6719 (_mm512_floor_round_pd): Ditto.
6720 (_mm512_ceil_round_ps): Ditto.
6721 (_mm512_ceil_round_pd): Ditto.
6722 (_mm512_mask_floor_round_ps): Ditto.
6723 (_mm512_mask_floor_round_pd): Ditto.
6724 (_mm512_mask_ceil_round_ps): Ditto.
6725 (_mm512_mask_ceil_round_pd): Ditto.
6726 (_mm512_maskz_floor_round_ps): Ditto.
6727 (_mm512_maskz_floor_round_pd): Ditto.
6728 (_mm512_maskz_ceil_round_ps): Ditto.
6729 (_mm512_maskz_ceil_round_pd): Ditto.
6730 (_mm512_expand_pd): Ditto.
6731 (_mm512_expand_ps): Ditto.
6732 * config/i386/i386.c (ix86_builtins): Remove
6733 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
6734 (bdesc_args): Ditto.
6735 * config/i386/predicates.md (const1256_operand): New.
6736 (const_1_to_2_operand): Ditto.
6737 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6738 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6739 (*avx512pf_gatherpf<mode>sf): Ditto.
6740 (avx512pf_gatherpf<mode>df): Ditto.
6741 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6742 (*avx512pf_gatherpf<mode>df): Ditto.
6743 (avx512pf_scatterpf<mode>sf): Ditto.
6744 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6745 (*avx512pf_scatterpf<mode>sf): Ditto.
6746 (avx512pf_scatterpf<mode>df): Ditto.
6747 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6748 (*avx512pf_scatterpf<mode>df): Ditto.
6749 (avx512f_expand<mode>): Removed.
6750 (<shift_insn><mode>3<mask_name>): Change predicate type.
6752 2014-02-08 Jakub Jelinek <jakub@redhat.com>
6754 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
6755 not at the end of datarefs vector use ordered_remove to avoid
6756 reordering datarefs vector.
6759 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
6760 mark local addressable non-static vars as GOVD_PRIVATE
6761 instead of GOVD_LOCAL.
6762 * omp-low.c (lower_omp_for): Move gimple_bind_vars
6763 and BLOCK_VARS of gimple_bind_block to new_stmt rather
6767 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
6768 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
6769 assume_aligned or alloc_align attributes.
6770 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
6771 arguments. Handle also assume_aligned and alloc_align attributes.
6772 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
6773 calls to functions with assume_aligned or alloc_align attributes.
6774 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
6776 2014-02-08 Terry Guo <terry.guo@arm.com>
6778 * doc/invoke.texi: Document ARM -march=armv7e-m.
6780 2014-02-08 Jakub Jelinek <jakub@redhat.com>
6782 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
6783 flag on __cilkrts_rethrow builtin.
6786 * ipa-cp.c (determine_versionability): Fail at -O0
6787 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
6788 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
6791 2014-02-04 Jakub Jelinek <jakub@redhat.com>
6794 * tree-inline.c (copy_forbidden): Fail for
6795 __attribute__((optimize (0))) functions.
6797 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
6799 * varpool.c: Include pointer-set.h.
6800 (varpool_remove_unreferenced_decls): Variables in other partitions
6801 will not be output; be however careful to not lose information
6804 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
6806 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
6807 lookup in the vtable constructor.
6809 2014-02-07 Jeff Law <law@redhat.com>
6812 * config/m68k/m68k.md (ashldi_extsi): Turn into a
6813 define_insn_and_split.
6815 * ipa-inline.c (inline_small_functions): Fix typos.
6817 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6819 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
6820 (s390_can_use_return_insn): Declare.
6821 * config/s390/s390.h (EPILOGUE_USES): Define.
6822 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
6824 (s390_chunkify_start): Handle return JUMP_LABELs.
6825 (s390_early_mach): Emit a main_pool instruction on the entry edge.
6826 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
6827 (s390_can_use_return_insn): New functions.
6828 (s390_fix_long_loop_prediction): Handle conditional returns.
6829 (TARGET_SET_UP_BY_PROLOGUE): Define.
6830 * config/s390/s390.md (ANY_RETURN): New code iterator.
6831 (*creturn, *csimple_return, return, simple_return): New patterns.
6833 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6835 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
6836 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
6837 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
6838 REG_CFA_RESTORE list when deciding not to restore a register.
6840 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6842 * config/s390/s390.c: Include tree-pass.h and context.h.
6843 (s390_early_mach): New function, split out from...
6844 (s390_emit_prologue): ...here.
6845 (pass_data_s390_early_mach): New pass structure.
6846 (pass_s390_early_mach): New class.
6847 (s390_option_override): Create and register early_mach pass.
6848 Move to end of file.
6850 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6852 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
6853 to match for the exit block.
6855 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6857 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
6858 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
6859 Reject misaligned operands.
6861 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6863 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
6865 2014-02-07 Richard Biener <rguenther@suse.de>
6868 * gimple-low.c (lower_builtin_posix_memalign): New function.
6869 (lower_stmt): Call it to lower posix_memalign in a way
6870 to make alignment info accessible.
6872 2014-02-07 Jakub Jelinek <jakub@redhat.com>
6875 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
6876 __builtin_setjmp_receiver.
6878 2014-02-07 Richard Biener <rguenther@suse.de>
6881 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
6882 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
6883 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
6884 Handle BUILT_IN_POSIX_MEMALIGN.
6885 (find_func_clobbers): Likewise.
6886 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
6887 (call_may_clobber_ref_p_1): Likewise.
6889 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6892 * ipa-devirt.c (record_target_from_binfo): Remove overactive
6895 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6898 * lto-cgraph.c (lto_output_node): Use
6899 symtab_get_symbol_partitioning_class.
6900 (lto_output_varpool_node): likewise.
6901 (symtab_get_symbol_partitioning_class): Move here from
6903 * cgraph.h (symbol_partitioning_class): Likewise.
6904 (symtab_get_symbol_partitioning_class): Declare.
6906 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6908 * ggc.h (ggc_internal_cleared_alloc): New macro.
6909 * vec.h (vec_safe_copy): Handle memory stats.
6910 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
6911 * target-globals.c (save_target_globals): Likewise.
6913 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6916 * expr.c (emit_move_resolve_push): Export; be bit more selective
6917 on when to clear alias set.
6918 * expr.h (emit_move_resolve_push): Declare.
6919 * function.h (struct function): Add tail_call_marked.
6920 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
6921 * config/i386/i386-protos.h (ix86_expand_push): Remove.
6922 * config/i386/i386.md (TImode move expander): De not call
6924 (FP push expanders): Preserve memory attributes.
6925 * config/i386/sse.md (push<mode>1): Remove.
6926 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
6927 (ix86_expand_push): Remove.
6928 * config/i386/mmx.md (push<mode>1): Remove.
6930 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6932 PR rtl-optimization/60030
6933 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
6934 lopart with paradoxical subreg before shifting it up by hprec.
6936 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6938 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
6939 Remove extra newline at end of file.
6940 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
6941 (arm_issue_rate): Handle cortexa57.
6942 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
6943 (cortex-a57.cortex-a53): Likewise.
6945 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6948 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
6949 don't record in REG_FRAME_RELATED_EXPR registers not set in that
6951 (arm_expand_prologue): Adjust all callers.
6952 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
6953 info, registers also at the lowest numbered registers side. Use
6954 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
6958 * var-tracking.c (adjust_mems): Before adding a SET to
6959 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
6961 2014-02-06 Alan Modra <amodra@gmail.com>
6964 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
6965 change SDmode to DDmode when lra_in_progress.
6967 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6970 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
6971 free_data_ref on the dr first, and before goto again also set dr
6972 to the next dr. For simd_lane_access, free old datarefs[i] before
6973 overwriting it. For get_vectype_for_scalar_type failure, don't
6974 free_data_ref if simd_lane_access.
6976 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
6979 * tree.h (opts_for_fn): New inline function.
6980 (opt_for_fn): Define.
6981 * config/i386/i386.c (ix86_function_regparm): Use
6982 opt_for_fn (decl, optimize) instead of optimize.
6984 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
6986 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
6987 for SYMBOL_REF in large memory model.
6989 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6991 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
6993 (cortex-a57): Likewise.
6994 (cortex-a57.cortex-a53): Likewise.
6996 2014-02-06 Yury Gribov <y.gribov@samsung.com>
6997 Kugan Vivekanandarajah <kuganv@linaro.org>
6999 * config/arm/arm.c (arm_vector_alignment_reachable): Check
7001 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
7003 2014-02-06 Richard Biener <rguenther@suse.de>
7005 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
7006 set_loop_copy and initialize_original_copy_tables.
7008 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
7010 * config/aarch64/aarch64-simd.md
7011 (aarch64_ashr_simddi): Change QI to SI.
7013 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7014 Jakub Jelinek <jakub@redhat.com>
7017 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
7020 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7022 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
7023 CODE_FOR_altivec_vpku[hw]um to
7024 CODE_FOR_altivec_vpku[hw]um_direct.
7025 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
7026 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
7027 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
7028 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
7030 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7032 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
7033 generation for -maltivec=be.
7034 (altivec_vsumsws): Simplify redundant test.
7036 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7038 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
7039 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
7040 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
7041 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
7042 gen_altivec_vpkuwum.
7043 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
7045 (altivec_vpks<VI_char>ss): Likewise.
7046 (altivec_vpks<VI_char>us): Likewise.
7047 (altivec_vpku<VI_char>us): Likewise.
7048 (altivec_vpku<VI_char>um): Likewise.
7049 (altivec_vpku<VI_char>um_direct): New (copy of
7050 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
7052 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
7053 target is little endian and -maltivec=be is not specified.
7054 (*altivec_vupkhs<VU_char>_direct): New (copy of
7055 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
7056 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
7057 target is little endian and -maltivec=be is not specified.
7058 (*altivec_vupkls<VU_char>_direct): New (copy of
7059 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
7060 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
7061 little endian and -maltivec=be is not specified.
7062 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
7063 little endian and -maltivec=be is not specified.
7065 2014-02-05 Richard Henderson <rth@redhat.com>
7068 * combine-stack-adj.c: Revert r206943.
7069 * sched-int.h (struct deps_desc): Add last_args_size.
7070 * sched-deps.c (init_deps): Initialize it.
7071 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7072 contain REG_ARGS_SIZE notes.
7074 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7076 * lto-cgraph.c (asm_nodes_output): Make global.
7077 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7078 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7079 (driver_handle_option): Handle OPT_fwpa.
7081 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7084 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7085 a comment typo and formatting issue. If odr_hash hasn't been
7086 created, return vNULL and set *completep to false.
7089 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7090 bb with no successors.
7092 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7095 * doc/invoke.texi (-march): Clarify documentation for ARM.
7099 2014-02-05 Richard Biener <rguenther@suse.de>
7101 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
7102 when not vectorizing because of too many alias checks.
7103 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
7104 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
7106 2014-02-05 Nick Clifton <nickc@redhat.com>
7108 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
7109 accept extended registers in any mode when compiling for the MN10300.
7111 2014-02-05 Yury Gribov <y.gribov@samsung.com>
7113 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
7114 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
7115 sanitization attributes.
7116 (can_inline_edge_p): Likewise.
7117 (sanitize_attrs_match_for_inline_p): New function.
7119 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7121 * ipa-prop.c (detect_type_change): Shor circuit testing of
7122 type changes on THIS pointer.
7124 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
7127 * config/pa/pa.c (legitimize_tls_address): Return original address
7128 if not passed a SYMBOL_REF rtx.
7129 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
7131 (pa_emit_move_sequence): Simplify TLS source operands.
7132 (pa_legitimate_constant_p): Reject all TLS constants.
7133 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
7134 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
7136 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7138 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
7139 groups when we know they are controlled by LTO.
7140 * varasm.c (default_binds_local_p_1): If object is in other partition,
7141 it will be resolved locally.
7143 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7145 * config/host-linux.c (linux_gt_pch_use_address): Don't
7146 use SSIZE_MAX because it is not always defined.
7148 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
7151 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
7152 threshold for pseudo splitting.
7153 (update_ebb_live_info): Process call argument hard registers and
7154 hard registers from insn definition too.
7155 (max_small_class_regs_num): New constant.
7156 (inherit_in_ebb): Update live hard regs through EBBs. Update
7157 reloads_num only for small register classes. Don't split for
7160 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
7163 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
7166 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7168 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
7171 2014-02-04 Marek Polacek <polacek@redhat.com>
7173 * gdbinit.in (pel): Define.
7175 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7177 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
7180 2014-02-04 Richard Biener <rguenther@suse.de>
7183 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
7184 in function context local.
7185 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
7186 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
7187 similar to LTO_imported_decl_ref.
7189 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7191 PR tree-optimization/60002
7192 * cgraphclones.c (build_function_decl_skip_args): Clear
7195 PR tree-optimization/60023
7196 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
7197 false to gsi_replace.
7198 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
7199 has been in some EH region and vec_stmt could throw, add
7200 vec_stmt into the same EH region.
7201 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
7202 has no lhs, ignore it.
7203 * internal-fn.c (expand_MASK_LOAD): Likewise.
7206 * tree-inline.c (copy_forbidden): Fail for
7207 __attribute__((optimize (0))) functions.
7210 * omp-low.c (simd_clone_struct_copy): If from->inbranch
7211 is set, copy one less argument.
7212 (expand_simd_clones): Don't subtract clone_info->inbranch
7213 from simd_clone_struct_alloc argument.
7215 PR rtl-optimization/57915
7216 * recog.c (simplify_while_replacing): If all unary/binary/relational
7217 operation arguments are constant, attempt to simplify those.
7220 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
7221 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
7223 2014-02-04 Richard Biener <rguenther@suse.de>
7225 PR tree-optimization/60012
7226 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
7227 TBAA disambiguation to all DDRs.
7229 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7232 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
7233 (LINK_SPEC): Use it for -shared, -shared-libgcc.
7235 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7238 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
7240 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7242 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
7243 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
7245 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7248 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
7249 to figure out targets of polymorphic calls with known decl.
7250 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7251 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
7252 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
7253 (get_polymorphic_call_info): ... here.
7254 (get_polymorphic_call_info_from_invariant): New function.
7256 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7258 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
7259 lookup via vtable pointer; check for type consistency
7260 and turn inconsitent facts into UNREACHABLE.
7261 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7262 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
7263 type inconsistent querries; return UNREACHABLE instead.
7265 2014-02-03 Richard Henderson <rth@twiddle.net>
7268 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
7269 already processed this node.
7270 (normalize_one_pred_1): Pass along mark_set.
7271 (normalize_one_pred): Create and destroy a pointer_set_t.
7272 (normalize_one_pred_chain): Likewise.
7274 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
7276 PR gcov-profile/58602
7277 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
7279 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7282 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
7283 -fno-devirtualize; try to devirtualize by the knowledge of
7284 virtual table pointer given by aggregate propagation.
7285 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7286 (ipa_print_node_jump_functions): Dump also offset that
7287 is relevant for polymorphic calls.
7288 (determine_known_aggregate_parts): Add arg_type parameter; use it
7289 instead of determining the type from pointer type.
7290 (ipa_compute_jump_functions_for_edge): Update call of
7291 determine_known_aggregate_parts.
7292 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
7293 (gimple_get_virt_method_for_binfo): ... here; simplify using
7294 vtable_pointer_value_to_vtable.
7295 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
7296 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
7297 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
7298 (vtable_pointer_value_to_vtable): Break out from ...; handle also
7300 (vtable_pointer_value_to_binfo): ... here.
7301 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
7303 2014-02-03 Teresa Johnson <tejohnson@google.com>
7305 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
7306 redef of outer loop index variable.
7308 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
7312 * doc/extend.texi (Function Attributes): Typo.
7314 2014-02-03 Cong Hou <congh@google.com>
7316 PR tree-optimization/60000
7317 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
7318 if the vectorized statement is a store. A store statement can only
7319 appear at the end of pattern statements.
7321 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7323 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
7324 (ix86_option_override_internal): Default long double to 64-bit for
7325 32-bit Bionic and to 128-bit for 64-bit Bionic.
7327 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
7328 TARGET_LONG_DOUBLE_128 is true.
7329 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
7331 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
7332 (mlong-double-64): Negate -mlong-double-128.
7333 (mlong-double-128): New option.
7335 * config/i386/i386-c.c (ix86_target_macros): Define
7336 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
7338 * doc/invoke.texi: Document -mlong-double-128.
7340 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7342 PR rtl-optimization/60024
7343 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
7345 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
7347 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
7349 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7351 PR rtl-optimization/57662
7352 * sel-sched.c (code_motion_path_driver): Do not mark already not
7353 existing blocks in the visiting bitmap.
7355 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7357 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
7358 on the insn being emitted.
7360 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
7361 Will Deacon <will.deacon@arm.com>
7363 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
7365 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7367 * config/arm/arm-tables.opt: Regenerate.
7369 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7371 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
7372 for vector types other than V16QImode.
7373 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
7374 define_expand, and call altivec_expand_vec_perm_le when producing
7375 code with little endian element order.
7376 (*altivec_vperm_<mode>_internal): New insn having previous
7377 behavior of altivec_vperm_<mode>.
7378 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
7379 altivec_expand_vec_perm_le when producing code with little endian
7381 (*altivec_vperm_<mode>_uns_internal): New insn having previous
7382 behavior of altivec_vperm_<mode>_uns.
7384 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7386 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
7387 (altivec_vsumsws): Add handling for -maltivec=be with a little
7389 (altivec_vsumsws_direct): New.
7390 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
7391 gen_altivec_vsumsws.
7393 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7395 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
7396 vtable_pointer_value_to_binfo): New functions.
7397 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
7398 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
7400 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
7402 * config/nios2/nios2.md (load_got_register): Initialize GOT
7403 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
7404 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
7406 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7408 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
7409 preserverd by passthrough, do not propagate the type.
7411 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7413 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
7414 (mips_atomic_assign_expand_fenv): New function.
7415 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
7417 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7419 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
7420 (__builtin_mips_set_fcsr): Likewise.
7421 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
7422 MIPS_USI_FTYPE_VOID.
7423 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
7424 (mips16_expand_set_fcsr): Likewise.
7425 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
7426 (mips16_set_fcsr_stub): Likewise.
7427 (mips16_get_fcsr_one_only_stub): New class.
7428 (mips16_set_fcsr_one_only_stub): Likewise.
7429 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
7430 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
7431 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
7432 (hard_float): New availability predicate.
7433 (mips_builtins): Add get_fcsr and set_fcsr.
7434 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
7435 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
7436 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
7437 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
7438 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
7441 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7443 * config/mips/mips.c (mips_one_only_stub): New class.
7444 (mips_need_mips16_rdhwr_p): Replace with...
7445 (mips16_rdhwr_stub): ...this new variable.
7446 (mips16_stub_call_address): New function.
7447 (mips16_rdhwr_one_only_stub): New class.
7448 (mips_expand_thread_pointer): Use mips16_stub_call_address.
7449 (mips_output_mips16_rdhwr): Delete.
7450 (mips_finish_stub): New function.
7451 (mips_code_end): Use it to handle rdhwr stubs.
7453 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
7456 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
7457 when calculating size of integer atomic types.
7459 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
7461 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
7463 2014-02-01 Jakub Jelinek <jakub@redhat.com>
7465 PR tree-optimization/60003
7466 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
7467 * profile.c (branch_prob): Use gimple_call_builtin_p
7468 to check for BUILT_IN_SETJMP_RECEIVER.
7469 * tree-inline.c (copy_bb): Call notice_special_calls.
7471 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
7474 * lra-constraints.c (process_alt_operands): Update reload_sum only
7477 2014-01-31 Richard Henderson <rth@redhat.com>
7480 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
7481 until after else_eh is processed.
7483 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7485 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
7486 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
7487 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
7488 in smmintrin.h, remove them.
7489 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
7490 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
7491 * config/i386/i386.md (ROUND_SAE): Fix value.
7492 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
7493 (const48_operand): New.
7494 * config/i386/subst.md (round), (round_expand): Use
7495 const_4_or_8_to_11_operand.
7496 (round_saeonly), (round_saeonly_expand): Use const48_operand.
7498 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7500 * config/i386/constraints.md (Yk): Swap meaning with k.
7501 * config/i386/i386.md (movhi_internal): Change Yk to k.
7502 (movqi_internal): Ditto.
7503 (*k<logic><mode>): Ditto.
7506 (kandn<mode>): Ditto.
7507 (*<code>hi_1): Ditto.
7508 (*<code>qi_1): Ditto.
7509 (kxnor<mode>): Ditto.
7510 (kortestzhi): Ditto.
7511 (kortestchi): Ditto.
7513 (*one_cmplhi2_1): Ditto.
7514 (*one_cmplqi2_1): Ditto.
7515 * config/i386/sse.md (): Change k to Yk.
7516 (avx512f_load<mode>_mask): Ditto.
7517 (avx512f_blendm<mode>): Ditto.
7518 (avx512f_store<mode>_mask): Ditto.
7519 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
7520 (avx512f_storedqu<mode>_mask): Ditto.
7521 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
7523 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
7524 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
7525 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
7526 (avx512f_maskcmp<mode>3): Ditto.
7527 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
7528 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
7529 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
7530 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
7531 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
7532 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
7533 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
7534 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
7535 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
7536 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
7537 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
7538 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
7539 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
7540 (vec_extract_lo_<mode>_maskm): Ditto.
7541 (vec_extract_hi_<mode>_maskm): Ditto.
7542 (avx512f_vternlog<mode>_mask): Ditto.
7543 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
7544 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
7545 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
7546 (avx512f_<code>v8div16qi2_mask): Ditto.
7547 (avx512f_<code>v8div16qi2_mask_store): Ditto.
7548 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
7549 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
7550 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
7551 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
7552 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7553 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7554 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7555 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7556 (avx512cd_maskb_vec_dupv8di): Ditto.
7557 (avx512cd_maskw_vec_dupv16si): Ditto.
7558 (avx512f_vpermi2var<mode>3_maskz): Ditto.
7559 (avx512f_vpermi2var<mode>3_mask): Ditto.
7560 (avx512f_vpermi2var<mode>3_mask): Ditto.
7561 (avx512f_vpermt2var<mode>3_maskz): Ditto.
7562 (*avx512f_gathersi<mode>): Ditto.
7563 (*avx512f_gathersi<mode>_2): Ditto.
7564 (*avx512f_gatherdi<mode>): Ditto.
7565 (*avx512f_gatherdi<mode>_2): Ditto.
7566 (*avx512f_scattersi<mode>): Ditto.
7567 (*avx512f_scatterdi<mode>): Ditto.
7568 (avx512f_compress<mode>_mask): Ditto.
7569 (avx512f_compressstore<mode>_mask): Ditto.
7570 (avx512f_expand<mode>_mask): Ditto.
7571 * config/i386/subst.md (mask): Change k to Yk.
7572 (mask_scalar_merge): Ditto.
7575 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
7577 * doc/extend.texi (Vector Extensions): Document ?: in C++.
7579 2014-01-31 Richard Biener <rguenther@suse.de>
7582 * builtins.c (fold_builtin_memory_op): Make sure to not
7583 use a floating-point mode or a boolean or enumeral type for
7586 2014-01-30 DJ Delorie <dj@redhat.com>
7588 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
7589 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
7590 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
7591 whenever main() has an epilogue.
7593 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7595 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
7596 unused variable "field".
7597 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
7598 (vsx_mergeh_<mode>): Likewise.
7599 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
7600 (altivec_vmrghh): Likewise.
7601 (altivec_vmrghw): Likewise.
7602 (altivec_vmrglb): Likewise.
7603 (altivec_vmrglh): Likewise.
7604 (altivec_vmrglw): Likewise.
7605 (altivec_vspltb): Add missing uses.
7606 (altivec_vsplth): Likewise.
7607 (altivec_vspltw): Likewise.
7608 (altivec_vspltsf): Likewise.
7610 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7613 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
7614 frame related instructions.
7616 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
7618 PR rtl-optimization/59959
7619 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
7620 any reload of register whose subreg is invalid.
7622 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7624 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
7625 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
7626 Add missing return type - void.
7628 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7630 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
7631 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
7632 remove element index adjustment for endian (now handled in vsx.md
7634 (altivec_expand_vec_perm_const): Use
7635 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
7636 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
7637 (vsx_xxspltw_<mode>): Adjust element index for little endian.
7638 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
7639 define_expand and a new define_insn *altivec_vspltb_internal;
7640 adjust for -maltivec=be on a little endian target.
7641 (altivec_vspltb_direct): New.
7642 (altivec_vsplth): Divide into a define_expand and a new
7643 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
7644 little endian target.
7645 (altivec_vsplth_direct): New.
7646 (altivec_vspltw): Divide into a define_expand and a new
7647 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
7648 little endian target.
7649 (altivec_vspltw_direct): New.
7650 (altivec_vspltsf): Divide into a define_expand and a new
7651 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
7652 a little endian target.
7654 2014-01-30 Richard Biener <rguenther@suse.de>
7656 PR tree-optimization/59993
7657 * tree-ssa-forwprop.c (associate_pointerplus): Check we
7658 can propagate form the earlier stmt and avoid the transform
7659 when the intermediate result is needed.
7661 2014-01-30 Alangi Derick <alangiderick@gmail.com>
7663 * README.Portability: Fix typo.
7665 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
7667 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
7668 comparison_operator with ordered_comparison_operator.
7670 2014-01-30 Nick Clifton <nickc@redhat.com>
7672 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
7673 Rename to mn10300_store_multiple_regs.
7674 * config/mn10300/mn10300.c: Likewise.
7675 * config/mn10300/mn10300.md (store_movm): Fix typo: call
7676 store_multiple_regs.
7677 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
7678 Call mn10300_store_multiple_regs.
7680 2014-01-30 Nick Clifton <nickc@redhat.com>
7681 DJ Delorie <dj@redhat.com>
7683 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
7684 %fp 2 to keep registers after it properly word-aligned.
7685 (rl78_alloc_physical_registers_umul): Handle the case where both
7686 input operands are the same.
7688 2014-01-30 Richard Biener <rguenther@suse.de>
7690 PR tree-optimization/59903
7691 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
7694 2014-01-30 Jason Merrill <jason@redhat.com>
7697 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
7700 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
7702 2014-01-30 Richard Biener <rguenther@suse.de>
7704 PR tree-optimization/59951
7705 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
7707 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
7710 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
7711 SFmode to DFmode case.
7713 2014-01-29 DJ Delorie <dj@redhat.com>
7715 * config/msp430/msp430.opt (-minrt): New.
7716 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
7718 (ENDFILE_SPEC): Likewise.
7720 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
7722 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
7723 (estimate_function_body_sizes): Use it.
7725 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
7728 * dwarf2out.c (is_cxx_auto): New.
7729 (is_base_type): Use it.
7730 (gen_type_die_with_usage): Likewise.
7732 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7734 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
7735 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
7736 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
7737 -maltivec=be with LE targets.
7738 (vsx_mergeh_<mode>): Likewise.
7739 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
7740 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
7741 (altivec_vmrghb): Replace with define_expand and new
7742 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
7743 (altivec_vmrghb_direct): New define_insn.
7744 (altivec_vmrghh): Replace with define_expand and new
7745 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
7746 (altivec_vmrghh_direct): New define_insn.
7747 (altivec_vmrghw): Replace with define_expand and new
7748 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
7749 (altivec_vmrghw_direct): New define_insn.
7750 (*altivec_vmrghsf): Adjust for endianness.
7751 (altivec_vmrglb): Replace with define_expand and new
7752 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
7753 (altivec_vmrglb_direct): New define_insn.
7754 (altivec_vmrglh): Replace with define_expand and new
7755 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
7756 (altivec_vmrglh_direct): New define_insn.
7757 (altivec_vmrglw): Replace with define_expand and new
7758 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
7759 (altivec_vmrglw_direct): New define_insn.
7760 (*altivec_vmrglsf): Adjust for endianness.
7761 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
7762 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
7763 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
7764 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
7765 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
7766 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
7767 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
7768 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
7770 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
7772 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
7773 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
7776 2014-01-29 Richard Biener <rguenther@suse.de>
7778 PR tree-optimization/58742
7779 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
7780 associate_pointerplus_align.
7781 (associate_pointerplus_diff): New function.
7782 (associate_pointerplus): Likewise. Call associate_pointerplus_align
7783 and associate_pointerplus_diff.
7785 2014-01-29 Richard Biener <rguenther@suse.de>
7787 * lto-streamer.h (LTO_major_version): Bump to 3.
7788 (LTO_minor_version): Reset to 0.
7790 2014-01-29 Renlin Li <Renlin.Li@arm.com>
7792 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
7793 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
7794 (arm_file_start): Generate correct asm header for armv7ve.
7795 * config/arm/bpabi.h: Add multilib support for armv7ve.
7796 * config/arm/driver-arm.c: Change the architectures of cortex-a7
7797 and cortex-a15 to armv7ve.
7798 * config/arm/t-aprofile: Add multilib support for armv7ve.
7799 * doc/invoke.texi: Document -march=armv7ve.
7801 2014-01-29 Richard Biener <rguenther@suse.de>
7803 PR tree-optimization/58742
7804 * tree-ssa-forwprop.c (associate_plusminus): Return true
7805 if we changed sth, defer EH cleanup to ...
7806 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
7807 (simplify_mult): New function.
7809 2014-01-29 Jakub Jelinek <jakub@redhat.com>
7812 PR tree-optimization/59920
7813 * tree.c (build_common_builtin_nodes): Remove
7814 __builtin_setjmp_dispatcher initialization.
7815 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
7816 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
7817 instead of gsi_after_labels + manually skipping debug stmts.
7818 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
7819 ignore bbs with IFN_ABNORMAL_DISPATCHER.
7820 * tree-inline.c (copy_edges_for_bb): Remove
7821 can_make_abnormal_goto argument, instead add abnormal_goto_dest
7822 argument. Ignore computed_goto_p stmts. Don't call
7823 make_abnormal_goto_edges. If a call might need abnormal edges
7824 for non-local gotos, see if it already has an edge to
7825 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
7826 with true argument, don't do anything then, otherwise add
7827 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
7828 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
7830 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
7831 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
7832 (lower_stmt): Don't set data->calls_builtin_setjmp.
7833 (lower_builtin_setjmp): Adjust comment.
7834 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
7835 * tree-cfg.c (found_computed_goto): Remove.
7836 (factor_computed_gotos): Remove.
7837 (make_goto_expr_edges): Return bool, true for computed gotos.
7838 Don't call make_abnormal_goto_edges.
7839 (build_gimple_cfg): Don't set found_computed_goto, don't call
7840 factor_computed_gotos.
7841 (computed_goto_p): No longer static.
7842 (make_blocks): Don't set found_computed_goto.
7843 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
7844 (make_edges): If make_goto_expr_edges returns true, push bb
7845 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
7846 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
7847 vector. Record mapping between bbs and OpenMP regions if there
7848 are any, adjust make_gimple_omp_edges caller. Call
7849 handle_abnormal_edges.
7850 (make_abnormal_goto_edges): Remove.
7851 * tree-cfg.h (make_abnormal_goto_edges): Remove.
7852 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
7853 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
7854 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
7855 * internal-fn.def (ABNORMAL_DISPATCHER): New.
7856 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
7857 filling *region also set *region_idx to (*region)->entry->index.
7860 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
7861 For REGs set ORIGINAL_REGNO.
7863 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
7865 * doc/md.texi: Mention that a target shouldn't implement
7866 vec_widen_(s|u)mul_even/odd pair if it is less efficient
7869 2014-01-29 Jakub Jelinek <jakub@redhat.com>
7871 PR tree-optimization/59594
7872 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
7873 a copy of the datarefs vector rather than the vector itself.
7875 2014-01-28 Jason Merrill <jason@redhat.com>
7878 * dwarf2out.c (auto_die): New static.
7879 (gen_type_die_with_usage): Handle C++1y 'auto'.
7880 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
7883 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
7886 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
7887 (SPEC_X32): Likewise.
7888 (SPEC_64): Likewise.
7889 * config/i386/i386.c (ix86_option_override_internal): Turn off
7890 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
7892 (x86_file_start): Output .code16gcc for TARGET_16BIT.
7893 * config/i386/i386.h (TARGET_16BIT): New macro.
7894 (TARGET_16BIT_P): Likewise.
7895 * config/i386/i386.opt: Add m16.
7896 * doc/invoke.texi: Document -m16.
7898 2014-01-28 Jakub Jelinek <jakub@redhat.com>
7900 PR preprocessor/59935
7901 * input.c (location_get_source_line): Bail out on when line number
7902 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
7904 2014-01-28 Richard Biener <rguenther@suse.de>
7906 PR tree-optimization/58742
7907 * tree-ssa-forwprop.c (associate_plusminus): Handle
7908 pointer subtraction of the form (T)(P + A) - (T)P.
7910 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7912 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
7915 2014-01-28 Richard Biener <rguenther@suse.de>
7918 2014-01-28 Richard Biener <rguenther@suse.de>
7920 PR rtl-optimization/45364
7921 PR rtl-optimization/59890
7922 * var-tracking.c (local_get_addr_clear_given_value): Handle
7923 already cleared slot.
7924 (val_reset): Handle not allocated local_get_addr_cache.
7925 (vt_find_locations): Use post-order on the inverted CFG.
7927 2014-01-28 Richard Biener <rguenther@suse.de>
7929 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
7931 2014-01-28 Richard Biener <rguenther@suse.de>
7933 PR rtl-optimization/45364
7934 PR rtl-optimization/59890
7935 * var-tracking.c (local_get_addr_clear_given_value): Handle
7936 already cleared slot.
7937 (val_reset): Handle not allocated local_get_addr_cache.
7938 (vt_find_locations): Use post-order on the inverted CFG.
7940 2014-01-28 Alan Modra <amodra@gmail.com>
7942 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
7943 * configure.ac <recursive call for build != host>: Define
7944 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
7945 and LD_FOR_BUILD too.
7946 * configure: Regenerate.
7948 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
7950 * config/i386/i386.c (get_builtin_code_for_version): Separate
7951 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
7952 Broadwell from Haswell.
7954 2014-01-27 Steve Ellcey <sellcey@mips.com>
7956 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
7957 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
7958 * config/mips/mips.c (mips_option_override): Change setting
7960 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
7961 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
7962 Change from Mask to Var.
7964 2014-01-27 Jeff Law <law@redhat.com>
7966 * ipa-inline.c (inline_small_functions): Fix typo.
7968 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
7970 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
7971 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
7972 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
7973 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
7974 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
7975 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
7976 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
7977 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
7978 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
7979 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
7980 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
7981 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
7982 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
7983 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
7984 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
7985 (_mm512_storeu_epi64): Ditto.
7986 (_mm512_cmpge_epi32_mask): Ditto.
7987 (_mm512_cmpge_epu32_mask): Ditto.
7988 (_mm512_cmpge_epi64_mask): Ditto.
7989 (_mm512_cmpge_epu64_mask): Ditto.
7990 (_mm512_cmple_epi32_mask): Ditto.
7991 (_mm512_cmple_epu32_mask): Ditto.
7992 (_mm512_cmple_epi64_mask): Ditto.
7993 (_mm512_cmple_epu64_mask): Ditto.
7994 (_mm512_cmplt_epi32_mask): Ditto.
7995 (_mm512_cmplt_epu32_mask): Ditto.
7996 (_mm512_cmplt_epi64_mask): Ditto.
7997 (_mm512_cmplt_epu64_mask): Ditto.
7998 (_mm512_cmpneq_epi32_mask): Ditto.
7999 (_mm512_cmpneq_epu32_mask): Ditto.
8000 (_mm512_cmpneq_epi64_mask): Ditto.
8001 (_mm512_cmpneq_epu64_mask): Ditto.
8002 (_mm512_expand_pd): Ditto.
8003 (_mm512_expand_ps): Ditto.
8004 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
8005 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
8006 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
8007 * config/i386/i386.c (ix86_builtins): Add
8008 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
8009 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
8010 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
8011 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
8012 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
8013 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
8014 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
8015 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
8016 IX86_BUILTIN_PMOVUSQW512_MEM.
8017 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
8018 __builtin_ia32_pmovsqd512mem_mask,
8019 __builtin_ia32_pmovqd512mem_mask,
8020 __builtin_ia32_pmovusqw512mem_mask,
8021 __builtin_ia32_pmovsqw512mem_mask,
8022 __builtin_ia32_pmovqw512mem_mask,
8023 __builtin_ia32_pmovusdw512mem_mask,
8024 __builtin_ia32_pmovsdw512mem_mask,
8025 __builtin_ia32_pmovdw512mem_mask,
8026 __builtin_ia32_pmovqb512mem_mask,
8027 __builtin_ia32_pmovusqb512mem_mask,
8028 __builtin_ia32_pmovsqb512mem_mask,
8029 __builtin_ia32_pmovusdb512mem_mask,
8030 __builtin_ia32_pmovsdb512mem_mask,
8031 __builtin_ia32_pmovdb512mem_mask.
8032 (bdesc_args): Add __builtin_ia32_expanddf512,
8033 __builtin_ia32_expandsf512.
8034 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
8035 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
8036 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
8037 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
8038 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
8039 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
8040 (avx512f_<code>v8div16qi2_mask_store): This.
8041 (avx512f_expand<mode>): New.
8043 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
8045 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
8047 (_mm512_mask_prefetch_i64gather_pd): Ditto.
8048 (_mm512_prefetch_i32scatter_pd): Ditto.
8049 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
8050 (_mm512_prefetch_i64scatter_pd): Ditto.
8051 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
8052 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
8053 (_mm512_mask_prefetch_i64gather_ps): Ditto.
8054 (_mm512_prefetch_i32scatter_ps): Ditto.
8055 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
8056 (_mm512_prefetch_i64scatter_ps): Ditto.
8057 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
8058 * config/i386/i386-builtin-types.def: Define
8059 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
8060 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
8061 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
8062 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
8063 IX86_BUILTIN_SCATTERPFQPD.
8064 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
8065 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
8066 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
8067 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
8068 __builtin_ia32_scatterpfqps.
8069 (ix86_expand_builtin): Expand new built-ins.
8070 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8071 fix memory access data type.
8072 (*avx512pf_gatherpf<mode>_mask): Ditto.
8073 (*avx512pf_gatherpf<mode>): Ditto.
8074 (avx512pf_scatterpf<mode>): Ditto.
8075 (*avx512pf_scatterpf<mode>_mask): Ditto.
8076 (*avx512pf_scatterpf<mode>): Ditto.
8077 (GATHER_SCATTER_SF_MEM_MODE): New.
8078 (avx512pf_gatherpf<mode>df): Ditto.
8079 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8080 (*avx512pf_scatterpf<mode>df): Ditto.
8082 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8085 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8086 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8089 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8091 * common/config/arm/arm-common.c
8092 (arm_rewrite_mcpu): Handle multiple names.
8094 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8096 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8098 * gimple-builder.h (create_gimple_tmp): Delete.
8100 2014-01-27 Christian Bruel <christian.bruel@st.com>
8102 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
8105 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
8107 * config/pa/pa.md (call): Generate indirect long calls to non-local
8108 functions when outputing 32-bit code.
8109 (call_value): Likewise except for special call to buggy powf function.
8111 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
8112 portable runtime and PIC indirect calls.
8113 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
8114 and PIC call sequences. Use ldo instead of blr to set return register
8115 in PIC call sequence.
8117 2014-01-25 Walter Lee <walt@tilera.com>
8119 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
8120 avoid clobbering a live register.
8122 2014-01-25 Walter Lee <walt@tilera.com>
8124 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
8125 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
8126 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
8127 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
8129 2014-01-25 Walter Lee <walt@tilera.com>
8131 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
8132 arguments on even registers.
8133 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
8135 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
8136 (BIGGEST_ALIGNMENT): Ditto.
8137 (BIGGEST_FIELD_ALIGNMENT): Ditto.
8139 2014-01-25 Walter Lee <walt@tilera.com>
8141 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
8142 insns before bundling.
8143 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8145 2014-01-25 Walter Lee <walt@tilera.com>
8147 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
8148 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
8149 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
8151 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8153 * config/mips/constraints.md (kl): Delete.
8154 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
8155 define expands, using...
8156 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
8157 instructions for MIPS16.
8158 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
8159 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
8161 2014-01-25 Walter Lee <walt@tilera.com>
8163 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
8167 2014-01-25 Walter Lee <walt@tilera.com>
8169 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
8170 (TARGET_EXPAND_TO_RTL_HOOK): Define.
8172 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8174 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
8177 2014-01-25 Jakub Jelinek <jakub@redhat.com>
8179 * print-rtl.c (in_call_function_usage): New var.
8180 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
8181 EXPR_LIST mode as mode and not as reg note name.
8184 * cfgloopmanip.c (copy_loop_info): If
8185 loop->warned_aggressive_loop_optimizations, make sure
8186 the flag is set in target loop too.
8188 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
8190 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
8192 * builtins.def: Likewise.
8193 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
8194 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
8195 * ira.c (ira_setup_eliminable_regset): Likewise.
8196 * omp-low.c (gate_expand_omp): Likewise.
8197 (execute_lower_omp): Likewise.
8198 (diagnose_sb_0): Likewise.
8199 (gate_diagnose_omp_blocks): Likewise.
8200 (simd_clone_clauses_extract): Likewise.
8203 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8205 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
8206 correction for little endian...
8207 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
8210 2014-01-24 Jeff Law <law@redhat.com>
8212 PR tree-optimization/59919
8213 * tree-vrp.c (find_assert_locations_1): Do not register asserts
8214 for non-returning calls.
8216 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
8218 * common/config/aarch64/aarch64-common.c
8219 (aarch64_rewrite_mcpu): Handle multiple names.
8220 * config/aarch64/aarch64.h
8221 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8223 2014-01-24 Dodji Seketeli <dodji@redhat.com>
8225 * input.c (add_file_to_cache_tab): Handle the case where fopen
8228 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
8231 * config/i386/i386.md (pushsf splitter): Get stack adjustment
8232 from push operand if code of push isn't PRE_DEC.
8234 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8237 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
8238 -mquad-memory-atomic. Update -mquad-memory documentation to say
8239 it is only used for non-atomic loads/stores.
8241 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
8242 -mquad-memory or -mquad-memory-atomic switches.
8244 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
8245 -mquad-memory-atomic to ISA 2.07 support.
8247 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
8248 to separate support of normal quad word memory operations (ldq, stq)
8249 from the atomic quad word memory operations.
8251 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
8252 support to separate non-atomic quad word operations from atomic
8253 quad word operations. Disable non-atomic quad word operations in
8254 little endian mode so that we don't have to swap words after the
8255 load and before the store.
8256 (quad_load_store_p): Add comment about atomic quad word support.
8257 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
8258 options printed with -mdebug=reg.
8260 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
8261 -mquad-memory-atomic as the test for whether we have quad word
8262 atomic instructions.
8263 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
8264 or -mp8-vector are used, allow byte/half-word atomic operations.
8266 * config/rs6000/sync.md (load_lockedti): Insure that the address
8267 is a proper indexed or indirect address for the lqarx instruction.
8268 On little endian systems, swap the hi/lo registers after the lqarx
8270 (load_lockedpti): Use indexed_or_indirect_operand predicate to
8271 insure the address is valid for the lqarx instruction.
8272 (store_conditionalti): Insure that the address is a proper indexed
8273 or indirect address for the stqcrx. instruction. On little endian
8274 systems, swap the hi/lo registers before doing the stqcrx.
8276 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
8277 insure the address is valid for the stqcrx. instruction.
8279 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8280 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
8281 type of quad memory support is available.
8283 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
8286 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
8287 there is a danger of looping.
8289 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
8291 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8292 force flag_ira_loop_pressure if set via command line.
8294 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8296 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
8297 (ashr_simd): New builtin handling DI mode.
8298 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
8299 (aarch64_sshr_simddi): New match pattern.
8300 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
8301 (vshrd_n_s64): Likewise.
8302 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
8304 2014-01-23 Nick Clifton <nickc@redhat.com>
8306 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
8307 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
8308 favour of mcu specific scripts.
8309 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
8312 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8313 Alex Velenko <Alex.Velenko@arm.com>
8315 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
8316 (vaddv_s16): Likewise.
8317 (vaddv_s32): Likewise.
8318 (vaddv_u8): Likewise.
8319 (vaddv_u16): Likewise.
8320 (vaddv_u32): Likewise.
8321 (vaddvq_s8): Likewise.
8322 (vaddvq_s16): Likewise.
8323 (vaddvq_s32): Likewise.
8324 (vaddvq_s64): Likewise.
8325 (vaddvq_u8): Likewise.
8326 (vaddvq_u16): Likewise.
8327 (vaddvq_u32): Likewise.
8328 (vaddvq_u64): Likewise.
8329 (vaddv_f32): Likewise.
8330 (vaddvq_f32): Likewise.
8331 (vaddvq_f64): Likewise.
8332 (vmaxv_f32): Likewise.
8333 (vmaxv_s8): Likewise.
8334 (vmaxv_s16): Likewise.
8335 (vmaxv_s32): Likewise.
8336 (vmaxv_u8): Likewise.
8337 (vmaxv_u16): Likewise.
8338 (vmaxv_u32): Likewise.
8339 (vmaxvq_f32): Likewise.
8340 (vmaxvq_f64): Likewise.
8341 (vmaxvq_s8): Likewise.
8342 (vmaxvq_s16): Likewise.
8343 (vmaxvq_s32): Likewise.
8344 (vmaxvq_u8): Likewise.
8345 (vmaxvq_u16): Likewise.
8346 (vmaxvq_u32): Likewise.
8347 (vmaxnmv_f32): Likewise.
8348 (vmaxnmvq_f32): Likewise.
8349 (vmaxnmvq_f64): Likewise.
8350 (vminv_f32): Likewise.
8351 (vminv_s8): Likewise.
8352 (vminv_s16): Likewise.
8353 (vminv_s32): Likewise.
8354 (vminv_u8): Likewise.
8355 (vminv_u16): Likewise.
8356 (vminv_u32): Likewise.
8357 (vminvq_f32): Likewise.
8358 (vminvq_f64): Likewise.
8359 (vminvq_s8): Likewise.
8360 (vminvq_s16): Likewise.
8361 (vminvq_s32): Likewise.
8362 (vminvq_u8): Likewise.
8363 (vminvq_u16): Likewise.
8364 (vminvq_u32): Likewise.
8365 (vminnmv_f32): Likewise.
8366 (vminnmvq_f32): Likewise.
8367 (vminnmvq_f64): Likewise.
8369 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8371 * config/aarch64/aarch64-simd.md
8372 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
8373 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
8374 (*aarch64_mul3_elt<mode>): Likewise.
8375 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
8376 (*aarch64_mul3_elt_to_64v2df): Likewise.
8377 (*aarch64_mla_elt<mode>): Likewise.
8378 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8379 (*aarch64_mls_elt<mode>): Likewise.
8380 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8381 (*aarch64_fma4_elt<mode>): Likewise.
8382 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8383 (*aarch64_fma4_elt_to_64v2df): Likewise.
8384 (*aarch64_fnma4_elt<mode>): Likewise.
8385 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8386 (*aarch64_fnma4_elt_to_64v2df): Likewise.
8387 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
8388 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
8389 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8390 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8391 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
8392 (aarch64_sqdmull_lane<mode>_internal): Likewise.
8393 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
8395 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
8397 * config/aarch64/aarch64-simd.md
8398 (aarch64_be_checked_get_lane<mode>): New define_expand.
8399 * config/aarch64/aarch64-simd-builtins.def
8400 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
8401 New builtin definition.
8402 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
8403 Use new safe be builtin.
8405 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8407 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
8409 (aarch64_be_st1<mode>): Likewise.
8410 (aarch_ld1<VALL:mode>): Define_expand modified.
8411 (aarch_st1<VALL:mode>): Likewise.
8412 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
8413 (UNSPEC_ST1): Likewise.
8415 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
8417 * config/microblaze/microblaze.md: Add trap insn and attribute
8419 2014-01-23 Dodji Seketeli <dodji@redhat.com>
8421 PR preprocessor/58580
8422 * input.h (location_get_source_line): Take an additional line_size
8424 (void diagnostics_file_cache_fini): Declare new function.
8425 * input.c (struct fcache): New type.
8426 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
8427 New static constants.
8428 (diagnostic_file_cache_init, total_lines_num)
8429 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
8430 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
8431 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
8432 (get_next_line, read_next_line, goto_next_line, read_line_num):
8433 New static function definitions.
8434 (diagnostic_file_cache_fini): New function.
8435 (location_get_source_line): Take an additional output line_len
8436 parameter. Re-write using lookup_or_add_file_to_cache_tab and
8438 * diagnostic.c (diagnostic_finish): Call
8439 diagnostic_file_cache_fini.
8440 (adjust_line): Take an additional input parameter for the length
8441 of the line, rather than calculating it with strlen.
8442 (diagnostic_show_locus): Adjust the use of
8443 location_get_source_line and adjust_line with respect to their new
8444 signature. While displaying a line now, do not stop at the first
8445 null byte. Rather, display the zero byte as a space and keep
8446 going until we reach the size of the line.
8447 * Makefile.in: Add vec.o to OBJS-libcommon
8449 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8450 Ilya Tocar <ilya.tocar@intel.com>
8452 * config/i386/avx512fintrin.h (_mm512_kmov): New.
8453 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
8454 (__builtin_ia32_kmov16): Ditto.
8455 * config/i386/i386.md (UNSPEC_KMOV): New.
8458 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8460 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
8461 (_mm512_storeu_si512): Ditto.
8463 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
8466 * rtl.h (get_referenced_operands): Declare.
8467 * recog.c (get_referenced_operands): New function.
8468 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
8469 operands have been referenced when recording LO_SUM references.
8471 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
8473 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
8475 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8477 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
8478 Enable for generic and recent AMD targets.
8480 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8482 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
8483 ARG_SIZE note when adjustment was eliminated.
8485 2014-01-22 Jeff Law <law@redhat.com>
8487 PR tree-optimization/59597
8488 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
8489 in file. Accept new argument REGISTERING and use it to modify
8490 dump output appropriately.
8491 (register_jump_thread): Corresponding changes.
8492 (mark_threaded_blocks): Reinstate code to cancel unprofitable
8493 thread paths involving joiner blocks. Add code to dump cancelled
8494 jump threading paths.
8496 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
8498 PR rtl-optimization/59477
8499 * lra-constraints.c (inherit_in_ebb): Process call for living hard
8500 regs. Update reloads_num and potential_reload_hard_regs for all insns.
8502 2014-01-22 Tom Tromey <tromey@redhat.com>
8504 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
8506 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
8508 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8510 PR rtl-optimization/59896
8511 * lra-constraints.c (process_alt_operands): Check unused note for
8512 matched operands of insn with no output reloads.
8514 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
8516 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
8517 (mips_move_from_gpr_cost): Likewise.
8519 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8521 PR rtl-optimization/59858
8522 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
8523 ira_class_hard_regs_num.
8524 (process_alt_operands): Increase reject for dying matched operand.
8526 2014-01-21 Jakub Jelinek <jakub@redhat.com>
8529 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
8530 smaller than size, perform several stores or loads and stores
8531 at dst + count - size to store or copy all of size bytes, rather
8532 than just last modesize bytes.
8534 2014-01-20 DJ Delorie <dj@redhat.com>
8536 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
8537 that CLOBBERs are REGs before propogating their values.
8539 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
8542 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
8543 (cgraph_inline_failed_type): New function.
8544 * cgraph.h (DEFCIFCODE): Add type.
8545 (cgraph_inline_failed_type_t): New enum.
8546 (cgraph_inline_failed_type): New prototype.
8547 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
8548 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
8549 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
8550 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
8551 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
8552 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
8553 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
8554 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
8555 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
8556 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
8557 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
8558 OPTIMIZATION_MISMATCH.
8559 * tree-inline.c (expand_call_inline): Emit errors during
8560 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
8562 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8565 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
8566 mode attribute in insn output.
8568 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
8570 * output.h (output_constant): Delete.
8571 * varasm.c (output_constant): Make private.
8573 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
8575 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
8577 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8580 * tree.h (fold_builtin_strcat): New prototype.
8581 * builtins.c (fold_builtin_strcat): No longer static. Add len
8582 argument, if non-NULL, don't call c_strlen. Optimize
8583 directly into __builtin_memcpy instead of __builtin_strcpy.
8584 (fold_builtin_2): Adjust fold_builtin_strcat caller.
8585 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
8587 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8589 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8590 for SImode_address_operand operands, having only a REG argument.
8592 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
8594 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
8595 loader name using mbig-endian.
8596 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
8598 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8600 * doc/invoke.texi (-march): Clarify documentation for AArch64.
8604 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
8606 * config/aarch64/aarch64-protos.h
8607 (aarch64_cannot_change_mode_class_ptr): Declare.
8608 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
8609 aarch64_cannot_change_mode_class_ptr): New.
8610 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
8611 backend hook aarch64_cannot_change_mode_class.
8613 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8615 * common/config/aarch64/aarch64-common.c
8616 (aarch64_handle_option): Don't handle any option order logic here.
8617 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
8618 selected_cpu, warn on architecture version mismatch.
8619 (aarch64_override_options): Fix parsing order for option strings.
8621 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8622 Iain Sandoe <iain@codesourcery.com>
8625 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
8626 warning. Amend comment to reflect current functionality.
8628 2014-01-20 Richard Biener <rguenther@suse.de>
8631 * builtins.c (fold_builtin_strcat): Remove case better handled
8632 by tree-ssa-strlen.c.
8634 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
8636 * config/aarch64/aarch64.opt
8637 (mcpu, march, mtune): Make case-insensitive.
8639 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8642 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8643 if operands[1] is a REG or ZERO_EXTEND of a REG.
8645 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
8647 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
8649 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
8651 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
8652 long non-pic millicode calls.
8654 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8656 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
8658 2014-01-19 Kito Cheng <kito@0xlab.org>
8660 * builtins.c (expand_movstr): Check movstr expand done or fail.
8662 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8663 H.J. Lu <hongjiu.lu@intel.com>
8666 * config/i386/i386.md (*lea<mode>): Zero-extend return register
8667 to DImode for zero-extended addresses.
8669 2014-01-19 Jakub Jelinek <jakub@redhat.com>
8671 PR rtl-optimization/57763
8672 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
8673 on the new indirect jump_insn and increment LABEL_NUSES (label).
8675 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
8679 * config.gcc (x86_archs): New variable.
8680 (x86_64_archs): Likewise.
8681 (x86_cpus): Likewise.
8682 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
8683 --with-arch/--with-cpu= options.
8684 Support --with-arch=/--with-cpu={nehalem,westmere,
8685 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
8687 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8689 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
8690 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
8692 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8694 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
8696 2014-01-18 Jakub Jelinek <jakub@redhat.com>
8699 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
8700 clear cpp_get_options (parse_in)->warn_unused_macros for
8701 ix86_target_macros_internal with cpp_define.
8703 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
8705 * jump.c (delete_related_insns): Keep (use (insn))s.
8706 * reorg.c (redundant_insn): Check for barriers too.
8708 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8710 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
8712 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
8714 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
8715 call to $$dyncall when TARGET_LONG_CALLS is true.
8717 2014-01-17 Jeff Law <law@redhat.com>
8719 * ree.c (combine_set_extension): Temporarily disable test for
8720 changing number of hard registers.
8722 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
8725 * ipa-inline-analysis.c (inline_free_summary):
8726 Do not free summary of aliases.
8728 2014-01-17 Jakub Jelinek <jakub@redhat.com>
8731 * gimplify.c (gimplify_expr): Use create_tmp_var
8732 instead of create_tmp_var_raw. If cond doesn't have
8733 integral type, don't add the IFN_ANNOTATE builtin at all.
8735 2014-01-17 Martin Jambor <mjambor@suse.cz>
8738 * ipa-cp.c (prev_edge_clone): New variable.
8739 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
8740 Also resize prev_edge_clone vector.
8741 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
8742 (ipcp_edge_removal_hook): New function.
8743 (ipcp_driver): Register ipcp_edge_removal_hook.
8745 2014-01-17 Andrew Pinski <apinski@cavium.com>
8746 Steve Ellcey <sellcey@mips.com>
8749 * config/mips/mips.c (mips_print_operand): Check operand mode instead
8752 2014-01-17 Jeff Law <law@redhat.com>
8755 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
8756 so that pass_ccp runs first.
8758 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8760 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
8761 (ix86_adjust_cost): Use !TARGET_XXX.
8762 (do_reorder_for_imul): Likewise.
8763 (swap_top_of_ready_list): Likewise.
8764 (ix86_sched_reorder): Likewise.
8766 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8768 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
8769 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
8770 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
8771 (intel_memset): New. Duplicate slm_memset.
8772 (intel_cost): New. Duplicate slm_cost.
8773 (m_INTEL): New macro.
8774 (processor_target_table): Add "intel".
8775 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
8776 with PROCESSOR_INTEL for "intel".
8777 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
8778 PROCESSOR_SILVERMONT.
8779 (ix86_issue_rate): Likewise.
8780 (ix86_adjust_cost): Likewise.
8781 (ia32_multipass_dfa_lookahead): Likewise.
8782 (swap_top_of_ready_list): Likewise.
8783 (ix86_sched_reorder): Likewise.
8784 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
8785 instead of TARGET_OPT_AGU.
8786 * config/i386/i386.h (TARGET_INTEL): New.
8787 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
8788 (processor_type): Add PROCESSOR_INTEL.
8789 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
8790 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
8792 2014-01-17 Marek Polacek <polacek@redhat.com>
8795 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
8798 2014-01-17 Richard Biener <rguenther@suse.de>
8800 PR tree-optimization/46590
8801 * opts.c (default_options_table): Add entries for
8802 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
8803 all enabled at -O1 but not for -Og.
8804 * common.opt (fbranch-count-reg): Remove Init(1).
8805 (fmove-loop-invariants): Likewise.
8806 (ftree-pta): Likewise.
8808 2014-01-17 Jakub Jelinek <jakub@redhat.com>
8810 * config/i386/i386.c (ix86_data_alignment): For compatibility with
8811 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
8812 decls to at least the GCC 4.8 used alignments.
8815 * tree-nested.c (convert_nonlocal_reference_stmt,
8816 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
8817 of GIMPLE_BIND stmts, adjust associated decls.
8819 2014-01-17 Richard Biener <rguenther@suse.de>
8821 PR tree-optimization/46590
8822 * vec.h (vec<>::bseach): New member function implementing
8823 binary search according to C89 bsearch.
8824 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
8825 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
8826 bitmap pointer again. Make accesses_in_loop a flat array.
8827 (mem_ref_obstack): New global.
8828 (outermost_indep_loop): Adjust for mem_ref->stored changes.
8829 (mark_ref_stored): Likewise.
8830 (ref_indep_loop_p_2): Likewise.
8831 (set_ref_stored_in_loop): New helper function.
8832 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
8833 (memref_free): Adjust.
8834 (record_mem_ref_loc): Simplify.
8835 (gather_mem_refs_stmt): Adjust.
8836 (sort_locs_in_loop_postorder_cmp): New function.
8837 (analyze_memory_references): Sort accesses_in_loop after
8838 loop postorder number.
8839 (find_ref_loc_in_loop_cmp): New function.
8840 (for_all_locs_in_loop): Find relevant cluster of locs in
8841 accesses_in_loop and iterate without recursion.
8842 (execute_sm): Avoid uninit warning.
8843 (struct ref_always_accessed): Simplify.
8844 (ref_always_accessed::operator ()): Likewise.
8845 (ref_always_accessed_p): Likewise.
8846 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
8847 loop postorder numbers here.
8848 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
8851 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
8854 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
8855 on decls for which assemble_alias has been called.
8857 2014-01-17 Nick Clifton <nickc@redhat.com>
8859 * config/msp430/msp430.opt: (mcpu): New option.
8860 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
8861 (msp430_option_override): Parse target_cpu. If the MCU name
8862 matches a generic string, clear target_mcu.
8863 (msp430_attr): Allow numeric interrupt values up to 63.
8864 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
8865 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
8867 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
8869 * config/msp430/msp430.md (popm): Use %J rather than %I.
8870 (addsi3): Use msp430_nonimmediate_operand for operand 2.
8871 (addhi_cy_i): Use immediate_operand for operand 2.
8872 * doc/invoke.texi: Document -mcpu option.
8874 2014-01-17 Richard Biener <rguenther@suse.de>
8876 PR rtl-optimization/38518
8877 * df.h (df_analyze_loop): Declare.
8878 * df-core.c: Include cfgloop.h.
8879 (df_analyze_1): Split out main part of df_analyze.
8880 (df_analyze): Adjust.
8881 (loop_inverted_post_order_compute): New function.
8882 (loop_post_order_compute): Likewise.
8883 (df_analyze_loop): New function avoiding whole-function
8885 * loop-invariant.c (find_defs): Use df_analyze_loop.
8886 (find_invariants): Adjust.
8887 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
8889 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
8891 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
8892 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
8894 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
8896 * ipa-ref.c (ipa_remove_stmt_references): Fix references
8897 traversal when removing references.
8899 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
8902 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
8904 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
8907 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
8908 pushing a reload for an autoinc when we had previously reloaded an
8909 inner part of the address.
8911 2014-01-16 Jakub Jelinek <jakub@redhat.com>
8913 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
8915 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
8916 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
8917 when not giving up or versioning for alias only because of
8919 (vect_analyze_data_ref_dependences): Set to true.
8920 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
8922 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
8923 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
8927 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
8930 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
8931 operand 0 predicate for gathers, use a new pseudo as subtarget.
8933 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
8936 * lra-constraints.c (process_alt_operands): Add printing debug info.
8937 Check absence of input/output reloads for matched operands too.
8939 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
8941 PR rtl-optimization/59835
8942 * ira.c (ira_init_register_move_cost): Increase cost for
8945 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
8947 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
8949 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
8952 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
8953 non-register objects. Use gen_(high/low)part more consistently.
8956 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8959 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
8960 endian support, remove tests for WORDS_BIG_ENDIAN.
8961 (p8_mfvsrd_3_<mode>): Likewise.
8962 (reload_gpr_from_vsx<mode>): Likewise.
8963 (reload_gpr_from_vsxsf): Likewise.
8964 (p8_mfvsrd_4_disf): Likewise.
8966 2014-01-16 Richard Biener <rguenther@suse.de>
8968 PR rtl-optimization/46590
8969 * lcm.c (compute_antinout_edge): Use postorder iteration.
8970 (compute_laterin): Use inverted postorder iteration.
8972 2014-01-16 Nick Clifton <nickc@redhat.com>
8975 * varasm.c (output_constant): Return the number of bytes actually
8977 (output_constructor_array_range): Update the field size with the
8978 number of bytes emitted by output_constant.
8979 (output_constructor_regular_field): Likewise. Also do not
8980 complain if the total number of bytes emitted is now greater
8981 than the expected fieldpos.
8982 * output.h (output_constant): Update prototype and descriptive comment.
8984 2014-01-16 Marek Polacek <polacek@redhat.com>
8987 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
8988 it is error_mark_node.
8990 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
8992 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
8993 VALID_AVX256_REG_OR_OI_MODE.
8995 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
8997 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
8998 current procedure should be profiled.
9000 2014-01-15 Andrew Pinski <apinski@cavium.com>
9002 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
9003 of moving from/to the STACK_REG register class.
9005 2014-01-15 Richard Henderson <rth@redhat.com>
9008 * reginfo.c (global_regs_decl): Globalize.
9009 * rtl.h (global_regs_decl): Declare.
9010 * ira.c (do_reload): Diagnose frame_pointer_needed and it
9011 reserved via global_regs.
9013 2014-01-15 Teresa Johnson <tejohnson@google.com>
9015 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
9017 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9019 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
9020 and vmulosh rather than call gen_vec_widen_smult_*.
9021 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
9022 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
9023 (vec_widen_smult_even_v16qi): Likewise.
9024 (vec_widen_umult_even_v8hi): Likewise.
9025 (vec_widen_smult_even_v8hi): Likewise.
9026 (vec_widen_umult_odd_v16qi): Likewise.
9027 (vec_widen_smult_odd_v16qi): Likewise.
9028 (vec_widen_umult_odd_v8hi): Likewise.
9029 (vec_widen_smult_odd_v8hi): Likewise.
9030 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
9031 vmuloub rather than call gen_vec_widen_umult_*.
9032 (vec_widen_umult_lo_v16qi): Likewise.
9033 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
9034 vmulosb rather than call gen_vec_widen_smult_*.
9035 (vec_widen_smult_lo_v16qi): Likewise.
9036 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
9037 rather than call gen_vec_widen_umult_*.
9038 (vec_widen_umult_lo_v8hi): Likewise.
9039 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
9040 rather than call gen_vec_widen_smult_*.
9041 (vec_widen_smult_lo_v8hi): Likewise.
9043 2014-01-15 Jeff Law <law@redhat.com>
9045 PR tree-optimization/59747
9046 * ree.c (find_and_remove_re): Properly handle case where a second
9047 eliminated extension requires widening a copy created for elimination
9048 of a prior extension.
9049 (combine_set_extension): Ensure that the number of hard regs needed
9050 for a destination register does not change when we widen it.
9052 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
9054 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
9055 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
9056 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
9057 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
9058 (avr-*-rtems*): Likewise.
9059 (bfin*-rtems*): Likewise.
9060 (moxie-*-rtems*): Likewise.
9061 (h8300-*-rtems*): Likewise.
9062 (i[34567]86-*-rtems*): Likewise.
9063 (lm32-*-rtems*): Likewise.
9064 (m32r-*-rtems*): Likewise.
9065 (m68k-*-rtems*): Likewise.
9066 (microblaze*-*-rtems*): Likewise.
9067 (mips*-*-rtems*): Likewise.
9068 (powerpc-*-rtems*): Likewise.
9069 (sh-*-rtems*): Likewise.
9070 (sparc-*-rtems*): Likewise.
9071 (sparc64-*-rtems*): Likewise.
9072 (v850-*-rtems*): Likewise.
9073 (m32c-*-rtems*): Likewise.
9075 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9077 PR rtl-optimization/59511
9078 * ira.c (ira_init_register_move_cost): Use memory costs for some
9079 cases of register move cost calculations.
9080 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9081 instead of BB frequency.
9082 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9083 * lra-assigns.c (find_hard_regno_for): Ditto.
9085 2014-01-15 Richard Biener <rguenther@suse.de>
9087 PR tree-optimization/59822
9088 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9089 (vectorizable_load): Use it to hoist defs of uses of invariant
9090 loads out of the loop.
9092 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9093 Kugan Vivekanandarajah <kuganv@linaro.org>
9096 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9099 2014-01-15 Richard Biener <rguenther@suse.de>
9101 PR rtl-optimization/59802
9102 * lcm.c (compute_available): Use inverted postorder to seed
9103 the initial worklist.
9105 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9108 * config/s390/s390.c (s390_preferred_reload_class): Don't return
9109 ADDR_REGS for invalid symrefs in non-PIC code.
9111 2014-01-15 Jakub Jelinek <jakub@redhat.com>
9114 * builtins.c (determine_block_size): Initialize *probable_max_size
9115 even if len_rtx is CONST_INT.
9117 2014-01-14 Andrew Pinski <apinski@cavium.com>
9119 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
9120 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
9121 (cortexa53_tunings): Likewise.
9122 (aarch64_sched_issue_rate): New function.
9123 (TARGET_SCHED_ISSUE_RATE): Define.
9125 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9127 * ira-costs.c (find_costs_and_classes): Add missed
9128 ira_init_register_move_cost_if_necessary.
9130 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9133 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
9135 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
9138 * config/i386/i386.c (type_natural_mode): Add a bool parameter
9139 to indicate if type is used for function return value. Warn ABI
9140 change if the vector mode isn't available for function return value.
9141 (ix86_function_arg_advance): Pass false to type_natural_mode.
9142 (ix86_function_arg): Likewise.
9143 (ix86_gimplify_va_arg): Likewise.
9144 (function_arg_32): Don't warn ABI change.
9145 (ix86_function_value): Pass true to type_natural_mode.
9146 (ix86_return_in_memory): Likewise.
9147 (ix86_struct_value_rtx): Removed.
9148 (TARGET_STRUCT_VALUE_RTX): Likewise.
9150 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9152 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
9153 converting a conditional jump into a conditional return.
9155 2014-01-14 Richard Biener <rguenther@suse.de>
9157 PR tree-optimization/58921
9158 PR tree-optimization/59006
9159 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
9160 hoisting invariant stmts.
9161 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
9162 invariant loads on the preheader edge if possible.
9164 2014-01-14 Joey Ye <joey.ye@arm.com>
9166 * doc/plugin.texi (Building GCC plugins): Update to C++.
9168 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
9170 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
9171 (_mm_rcp28_round_ss): Ditto.
9172 (_mm_rsqrt28_round_sd): Ditto.
9173 (_mm_rsqrt28_round_ss): Ditto.
9174 (_mm_rcp28_sd): Ditto.
9175 (_mm_rcp28_ss): Ditto.
9176 (_mm_rsqrt28_sd): Ditto.
9177 (_mm_rsqrt28_ss): Ditto.
9178 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
9179 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
9180 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
9181 (IX86_BUILTIN_RCP28SD): Ditto.
9182 (IX86_BUILTIN_RCP28SS): Ditto.
9183 (IX86_BUILTIN_RSQRT28SD): Ditto.
9184 (IX86_BUILTIN_RSQRT28SS): Ditto.
9185 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
9186 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
9187 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
9188 (ix86_expand_special_args_builtin): Expand new FTYPE.
9189 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
9190 (srcp14<mode>): Make insn unary.
9191 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
9192 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
9193 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
9194 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
9195 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
9196 Fix rounding: make it SAE only.
9197 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9199 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9201 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
9202 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
9203 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
9204 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
9205 (round_saeonly_mask_scalar_operand4): Ditto.
9206 (round_saeonly_mask_scalar_op3): Ditto.
9207 (round_saeonly_mask_scalar_op4): Ditto.
9209 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9211 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9212 Implement -maltivec=be for vec_insert and vec_extract.
9214 2014-01-10 DJ Delorie <dj@redhat.com>
9216 * config/msp430/msp430.md (call_internal): Don't allow memory
9217 references with SP as the base register.
9218 (call_value_internal): Likewise.
9219 * config/msp430/constraints.md (Yc): New. For memory references
9220 that don't use SP as a base register.
9222 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
9223 "an integer without a # prefix"
9224 * config/msp430/msp430.md (epilogue_helper): Use it.
9226 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9229 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
9230 AVX512F gather builtins.
9231 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
9232 on gather decls with INTEGER_TYPE masktype.
9233 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
9234 directly into the builtin rather than hoisting it before loop.
9236 PR tree-optimization/59387
9237 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
9238 (scev_const_prop): If folded_casts and type has undefined overflow,
9239 use force_gimple_operand instead of force_gimple_operand_gsi and
9240 for each added stmt if it is assign with
9241 arith_code_with_undefined_signed_overflow, call
9242 rewrite_to_defined_overflow.
9243 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
9244 gimple-fold.h instead.
9245 (arith_code_with_undefined_signed_overflow,
9246 rewrite_to_defined_overflow): Moved to ...
9247 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
9248 rewrite_to_defined_overflow): ... here. No longer static.
9249 Include gimplify-me.h.
9250 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
9251 rewrite_to_defined_overflow): New prototypes.
9253 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9255 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
9257 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
9259 * builtins.c (get_object_alignment_2): Minor tweak.
9260 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
9262 2014-01-13 Christian Bruel <christian.bruel@st.com>
9264 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
9265 optimized non constant lengths.
9267 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9270 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
9271 load as __atomic_load_N if possible.
9273 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9275 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
9277 (rs6000_expand_builtin): Adjust call.
9279 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9282 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
9283 * config/rs6000/rs6000.c: Include target-globals.h.
9284 (rs6000_set_current_function): Instead of doing target_reinit
9285 unconditionally, use save_target_globals_default_opts and
9286 restore_target_globals.
9288 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
9290 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
9291 (rs6000_expand_builtin): Handle mffs and mtfsf.
9292 (rs6000_init_builtins): Define mffs and mtfsf.
9293 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
9294 (rs6000_mffs): New pattern.
9295 (rs6000_mtfsf): New pattern.
9297 2014-01-11 Bin Cheng <bin.cheng@arm.com>
9299 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
9300 Start narrowing with START. Apply candidate-use pair
9301 and check overall cost in narrowing.
9302 (iv_ca_prune): Pass new argument.
9304 2014-01-10 Jeff Law <law@redhat.com>
9307 * ree.c (combine_reaching_defs): Ensure the defining statement
9308 occurs before the extension when optimizing extensions with
9309 different source and destination hard registers.
9311 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9314 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
9315 vtables into the type inheritance graph.
9317 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9319 PR rtl-optimization/59754
9320 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
9321 modes in the REGNO != REGNO case.
9323 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9325 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
9327 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9329 PR tree-optimization/59745
9330 * tree-predcom.c (tree_predictive_commoning_loop): Call
9331 free_affine_expand_cache if giving up because components is NULL.
9333 * target-globals.c (save_target_globals): Allocate < 4KB structs using
9334 GC in payload of target_globals struct instead of allocating them on
9335 the heap and the larger structs separately using GC.
9336 * target-globals.h (struct target_globals): Make regs, hard_regs,
9337 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
9338 of GTY((skip)) and change type to void *.
9339 (reset_target_globals): Cast loads from those fields to corresponding
9342 2014-01-10 Steve Ellcey <sellcey@mips.com>
9345 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
9346 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
9347 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
9349 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
9352 * aarch64-modes.def (CC_Zmode): New flags mode.
9353 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
9354 represents an equality.
9355 (aarch64_get_condition_code): Handle CC_Zmode.
9356 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
9358 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9360 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
9361 extraction in good case.
9363 2014-01-10 Richard Biener <rguenther@suse.de>
9365 PR tree-optimization/59374
9366 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
9367 checking after SLP discovery. Mark stmts not participating
9368 in any SLP instance properly.
9370 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9372 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
9373 when handling a SET rtx.
9375 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9377 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
9378 (cortex-a57): Likewise.
9379 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
9381 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9383 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
9384 non-iwmmxt builtins.
9386 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9390 * ipa-devirt.c record_target_from_binfo): Take as argument
9391 stack of binfos and lookup matching one for virtual inheritance.
9392 (possible_polymorphic_call_targets_1): Update.
9394 2014-01-10 Huacai Chen <chenhc@lemote.com>
9396 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
9397 kernel strings for Loongson-2E/2F/3A.
9399 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9402 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
9403 is_gimple_call before calling gimple_call_internal_p.
9405 2014-01-09 Steve Ellcey <sellcey@mips.com>
9407 * Makefile.in (TREE_FLOW_H): Remove.
9408 (TREE_SSA_H): Add file names from tree-flow.h.
9409 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
9410 * tree.h: Remove tree-flow.h reference.
9411 * hash-table.h: Remove tree-flow.h reference.
9412 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
9413 reference with tree-ssa-loop.h.
9415 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9417 * doc/invoke.texi: Add -maltivec={be,le} options, and document
9418 default element-order behavior for -maltivec.
9419 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
9420 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
9421 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
9422 when targeting big endian, at least for now.
9423 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
9425 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9428 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
9429 var satisfies use_register_for_decl, just take into account type
9430 alignment, rather than decl alignment.
9432 PR tree-optimization/59622
9433 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
9434 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
9435 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
9436 Don't devirtualize for inplace at all. For targets.length () == 1,
9437 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
9439 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9441 * config/i386/i386.md (cpu): Remove the unused btver1.
9443 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9445 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
9447 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9450 * tree-core.h (struct target_globals): New forward declaration.
9451 (struct tree_target_option): Add globals field.
9452 * tree.h (TREE_TARGET_GLOBALS): Define.
9453 (prepare_target_option_nodes_for_pch): New prototype.
9454 * target-globals.h (struct target_globals): Define even if
9456 * tree.c (prepare_target_option_node_for_pch,
9457 prepare_target_option_nodes_for_pch): New functions.
9458 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
9459 * config/i386/i386.c: Include target-globals.h.
9460 (ix86_set_current_function): Instead of doing target_reinit
9461 unconditionally, use save_target_globals_default_opts and
9462 restore_target_globals.
9464 2014-01-09 Richard Biener <rguenther@suse.de>
9466 PR tree-optimization/59715
9467 * tree-cfg.h (split_critical_edges): Declare.
9468 * tree-cfg.c (split_critical_edges): Export.
9469 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
9471 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
9473 * cfgexpand.c (expand_stack_vars): Optionally disable
9474 asan stack protection.
9475 (expand_used_vars): Likewise.
9476 (partition_stack_vars): Likewise.
9477 * asan.c (asan_emit_stack_protection): Optionally disable
9478 after return stack usage.
9479 (instrument_derefs): Optionally disable memory access instrumentation.
9480 (instrument_builtin_call): Likewise.
9481 (instrument_strlen_call): Likewise.
9482 (asan_protect_global): Optionally disable global variables protection.
9483 * doc/invoke.texi: Added doc for new options.
9484 * params.def: Added new options.
9485 * params.h: Likewise.
9487 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9489 PR rtl-optimization/59724
9490 * ifcvt.c (cond_exec_process_if_block): Don't call
9491 flow_find_head_matching_sequence with 0 longest_match.
9492 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
9493 non-active insns if !stop_after.
9494 (try_head_merge_bb): Revert 2014-01-07 changes.
9496 2014-01-08 Jeff Law <law@redhat.com>
9498 * ree.c (get_sub_rtx): New function, extracted from...
9499 (merge_def_and_ext): Here.
9500 (combine_reaching_defs): Use get_sub_rtx.
9502 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
9504 * cgraph.h (varpool_variable_node): Do not choke on null node.
9506 2014-01-08 Catherine Moore <clm@codesourcery.com>
9508 * config/mips/mips.md (simple_return): Attempt to use JRC
9510 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
9512 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9514 PR rtl-optimization/59137
9515 * reorg.c (steal_delay_list_from_target): Call update_block for
9517 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
9519 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9521 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
9522 two duplicate entries.
9524 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9527 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
9529 * config/mips/mips.c (mips_truncated_op_cost): New function.
9530 (mips_rtx_costs): Adjust test for BADDU.
9531 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
9533 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
9535 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
9536 (*baddu_si): ...this new pattern.
9538 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9541 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
9543 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9546 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
9548 (expand_expr, expand_normal): Adjust.
9549 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
9550 inner_reference_p. Use inner_reference_p to expand inner references.
9551 (store_expr): Adjust.
9552 * cfgexpand.c (expand_call_stmt): Adjust.
9554 2014-01-08 Rong Xu <xur@google.com>
9556 * gcov-io.c (gcov_var): Move from gcov-io.h.
9557 (gcov_position): Ditto.
9558 (gcov_is_error): Ditto.
9559 (gcov_rewrite): Ditto.
9560 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
9561 only part to libgcc/libgcov.h.
9563 2014-01-08 Marek Polacek <polacek@redhat.com>
9566 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
9568 2014-01-08 Marek Polacek <polacek@redhat.com>
9571 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
9573 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9575 PR rtl-optimization/59649
9576 * stor-layout.c (get_mode_bounds): For BImode return
9577 0 and STORE_FLAG_VALUE.
9579 2014-01-08 Richard Biener <rguenther@suse.de>
9582 * gimple.h (is_gimple_builtin_call): Remove.
9583 (gimple_builtin_call_types_compatible_p): New.
9584 (gimple_call_builtin_p): New overload.
9585 * gimple.c (is_gimple_builtin_call): Remove.
9586 (validate_call): Rename to ...
9587 (gimple_builtin_call_types_compatible_p): ... this and export. Also
9589 (validate_type): New static function.
9590 (gimple_call_builtin_p): New overload and adjust.
9591 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
9592 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
9593 (gimple_fold_stmt_to_constant_1): Likewise.
9594 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
9596 2014-01-08 Richard Biener <rguenther@suse.de>
9599 * gimplify.c (gimplify_expr): Gimplify register-register type
9600 VIEW_CONVERT_EXPRs to separate stmts.
9602 2014-01-07 Jeff Law <law@redhat.com>
9605 * ree.c (combine_set_extension): Handle case where source
9606 and destination registers in an extension insn are different.
9607 (combine_reaching_defs): Allow source and destination registers
9608 in extension to be different under limited circumstances.
9609 (add_removable_extension): Remove restriction that the
9610 source and destination registers in the extension are the same.
9611 (find_and_remove_re): Emit a copy from the extension's
9612 destination to its source after the defining insn if
9613 the source and destination registers are different.
9616 * ifcvt.c (merge_if_block): If we are merging a block with more than
9617 one successor with a block with no successors, remove any BARRIER
9618 after the second block.
9620 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
9622 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
9624 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
9627 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
9628 for 14-bit register offsets when INT14_OK_STRICT is false.
9630 2014-01-07 Roland Stigge <stigge@antcom.de>
9631 Michael Meissner <meissner@linux.vnet.ibm.com>
9634 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
9635 Only check TFmode for SPE constants. Don't check TImode or TDmode.
9637 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
9639 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
9642 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
9644 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
9645 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
9646 rtx is const0_rtx or not.
9648 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
9651 * target-globals.c (save_target_globals): Remove this_fn_optab
9653 * toplev.c: Include optabs.h.
9654 (target_reinit): Temporarily restore the global options if another
9655 set of options are in force.
9657 2014-01-07 Jakub Jelinek <jakub@redhat.com>
9659 PR rtl-optimization/58668
9660 * cfgcleanup.c (flow_find_cross_jump): Don't count
9661 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
9662 to determine what is counted.
9663 (flow_find_head_matching_sequence): Use active_insn_p to determine
9665 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
9667 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
9668 determine what is counted.
9670 PR tree-optimization/59643
9671 * tree-predcom.c (split_data_refs_to_components): If one dr is
9672 read and one write, determine_offset fails and the write isn't
9673 in the bad component, just put the read into the bad component.
9675 2014-01-07 Mike Stump <mikestump@comcast.net>
9676 Jakub Jelinek <jakub@redhat.com>
9679 * tree-core.h (struct tree_optimization_option): Change optabs
9680 type from unsigned char * to void *.
9681 * optabs.c (init_tree_optimization_optabs): Adjust
9682 TREE_OPTIMIZATION_OPTABS initialization.
9684 2014-01-06 Jakub Jelinek <jakub@redhat.com>
9687 * config/i386/i386.h (struct machine_function): Add
9688 no_drap_save_restore field.
9689 * config/i386/i386.c (ix86_save_reg): Use
9690 !cfun->machine->no_drap_save_restore instead of
9691 crtl->stack_realign_needed.
9692 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
9693 this function clears frame_pointer_needed. Set
9694 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
9695 and DRAP reg is needed.
9697 2014-01-06 Marek Polacek <polacek@redhat.com>
9700 * doc/implement-c.texi: Mention that other integer types are
9701 permitted as bit-field types in strictly conforming mode.
9703 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
9705 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
9708 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
9710 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
9712 2014-01-06 Martin Jambor <mjambor@suse.cz>
9715 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
9717 * ipa-prop.c (ipa_print_node_params): Fix indentation.
9719 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
9723 * var-tracking.c (add_stores): Preserve the value of the source even if
9724 we don't record the store.
9726 2014-01-06 Terry Guo <terry.guo@arm.com>
9728 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
9730 2014-01-05 Iain Sandoe <iain@codesourcery.com>
9733 * config/darwin.c (darwin_function_section): Adjust return values to
9734 correspond to optimisation changes made in r206070.
9736 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
9738 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
9739 from prefetch_block tune setting.
9740 (nocona_cost): Correct size of prefetch block to 64.
9742 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
9744 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
9745 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
9746 used to save the static chain register in the computation of the offset
9747 from which the FP registers need to be restored.
9749 2014-01-04 Jakub Jelinek <jakub@redhat.com>
9751 PR tree-optimization/59519
9752 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
9753 ICE if get_current_def (current_new_name) is already non-NULL, as long
9754 as it is a phi result of some other phi in *new_exit_bb that has
9757 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
9758 or vmovdqu* for misaligned_operand.
9759 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
9760 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
9761 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
9762 aligned_mem for AVX512F masked aligned load and store builtins and for
9765 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
9767 PR tree-optimization/59651
9768 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
9769 Address range for negative step should be added by TYPE_SIZE_UNIT.
9771 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
9773 * config/m68k/m68k.c (handle_move_double): Handle pushes with
9774 overlapping registers also for registers other than the stack pointer.
9776 2014-01-03 Marek Polacek <polacek@redhat.com>
9779 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
9782 2014-01-03 Jakub Jelinek <jakub@redhat.com>
9785 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
9788 * config/i386/i386.md (MODE_SIZE): New mode attribute.
9789 (push splitter): Use <P:MODE_SIZE> instead of
9790 GET_MODE_SIZE (<P:MODE>mode).
9791 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
9792 (mov -1, reg peephole2): Likewise.
9793 * config/i386/sse.md (*mov<mode>_internal,
9794 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
9795 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
9796 *<code><mode>3, *andnot<mode>3<mask_name>,
9797 <mask_codefor><code><mode>3<mask_name>): Likewise.
9798 * config/i386/subst.md (mask_mode512bit_condition,
9799 sd_mask_mode512bit_condition): Likewise.
9801 2014-01-02 Xinliang David Li <davidxl@google.com>
9803 PR tree-optimization/59303
9804 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
9805 (dump_predicates): Better output format.
9806 (pred_equal_p): New function.
9807 (is_neq_relop_p): Ditto.
9808 (is_neq_zero_form_p): Ditto.
9809 (pred_expr_equal_p): Ditto.
9810 (pred_neg_p): Ditto.
9811 (simplify_pred): Ditto.
9812 (simplify_preds_2): Ditto.
9813 (simplify_preds_3): Ditto.
9814 (simplify_preds_4): Ditto.
9815 (simplify_preds): Ditto.
9817 (push_to_worklist): Ditto.
9818 (get_pred_info_from_cmp): Ditto.
9819 (is_degenerated_phi): Ditto.
9820 (normalize_one_pred_1): Ditto.
9821 (normalize_one_pred): Ditto.
9822 (normalize_one_pred_chain): Ditto.
9823 (normalize_preds): Ditto.
9824 (normalize_cond_1): Remove function.
9825 (normalize_cond): Ditto.
9826 (is_gcond_subset_of): Ditto.
9827 (is_subset_of_any): Ditto.
9828 (is_or_set_subset_of): Ditto.
9829 (is_and_set_subset_of): Ditto.
9830 (is_norm_cond_subset_of): Ditto.
9831 (pred_chain_length_cmp): Ditto.
9832 (convert_control_dep_chain_into_preds): Type change.
9833 (find_predicates): Ditto.
9834 (find_def_preds): Ditto.
9835 (destroy_predicates_vecs): Ditto.
9836 (find_matching_predicates_in_rest_chains): Ditto.
9837 (use_pred_not_overlap_with_undef_path_pred): Ditto.
9838 (is_pred_expr_subset): Ditto.
9839 (is_pred_chain_subset_of): Ditto.
9840 (is_included_in): Ditto.
9841 (is_superset_of): Ditto.
9843 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
9845 Update copyright years.
9847 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
9849 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
9850 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
9851 config/arc/arc.md, config/arc/arc.opt,
9852 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
9853 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
9854 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
9855 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
9856 config/linux-protos.h, config/linux.c, config/winnt-c.c,
9857 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
9858 vtable-verify.c, vtable-verify.h: Use the standard form for the
9861 2014-01-02 Tobias Burnus <burnus@net-b.de>
9863 * gcc.c (process_command): Update copyright notice dates.
9864 * gcov-dump.c: Ditto.
9866 * doc/cpp.texi: Bump @copying's copyright year.
9867 * doc/cppinternals.texi: Ditto.
9868 * doc/gcc.texi: Ditto.
9869 * doc/gccint.texi: Ditto.
9870 * doc/gcov.texi: Ditto.
9871 * doc/install.texi: Ditto.
9872 * doc/invoke.texi: Ditto.
9874 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9876 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
9878 2014-01-01 Jakub Jelinek <jakub@redhat.com>
9880 * config/i386/sse.md (*mov<mode>_internal): Guard
9881 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
9883 PR rtl-optimization/59647
9884 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
9885 new_rtx into UNSIGNED_FLOAT rtxes.
9887 Copyright (C) 2014 Free Software Foundation, Inc.
9889 Copying and distribution of this file, with or without modification,
9890 are permitted in any medium without royalty provided the copyright
9891 notice and this notice are preserved.