1 2014-05-08 Bin Cheng <bin.cheng@arm.com>
3 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
4 core part of address expressions.
6 2014-05-08 Alan Modra <amodra@gmail.com>
9 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
10 loads and stores when -mno-strict-align at any alignment.
11 (expand_block_clear): Similarly. Also correct calculation of
14 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
17 * tree-complex.c (expand_complex_move): Keep line info when expanding
19 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
20 of complex expression. Use new argument to display correct location
21 for values coming from phi statement.
22 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
23 (warn_uninitialized_phi): Pass location of phi argument to
25 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
26 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
28 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
30 * config/rs6000/predicates.md (indexed_address_mem): New.
31 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
32 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
33 fpstore_ux, fpstore_u.
34 (sign_extend, indexed, update): New.
36 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
37 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
38 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
39 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
40 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
41 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
42 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
43 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
44 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
45 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
46 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
47 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
48 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
49 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
50 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
52 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
53 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
54 *vsx_extract_<mode>_store): Adjust.
55 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
56 is_cracked_insn, insn_must_be_first_in_group,
57 insn_must_be_last_in_group): Adjust.
59 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
61 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
62 ppc440-fpstore): Adjust.
63 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
64 ppc476-fpstore): Adjust.
65 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
66 ppc601-fpstore): Adjust.
67 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
69 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
71 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
72 ppc7450-fpstore): Adjust.
73 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
74 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
75 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
77 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
78 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
79 cell-fpstore, cell-fpstore-update): Adjust.
80 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
81 ppce300c3_store, ppce300c3_fpstore): Adjust.
82 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
83 e500mc_fpstore): Adjust.
84 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
85 e500mc64_store, e500mc64_fpstore): Adjust.
86 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
87 e5500_fpstore): Adjust.
88 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
89 e6500_fpstore): Adjust.
90 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
92 * config/rs6000/power4.md (power4-load, power4-load-ext,
93 power4-load-ext-update, power4-load-ext-update-indexed,
94 power4-load-update-indexed, power4-load-update, power4-fpload,
95 power4-fpload-update, power4-store, power4-store-update,
96 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
98 * config/rs6000/power5.md (power5-load, power5-load-ext,
99 power5-load-ext-update, power5-load-ext-update-indexed,
100 power5-load-update-indexed, power5-load-update, power5-fpload,
101 power5-fpload-update, power5-store, power5-store-update,
102 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
104 * config/rs6000/power6.md (power6-load, power6-load-ext,
105 power6-load-update, power6-load-update-indexed,
106 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
107 power6-fpload-update, power6-store, power6-store-update,
108 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
110 * config/rs6000/power7.md (power7-load, power7-load-ext,
111 power7-load-update, power7-load-update-indexed,
112 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
113 power7-fpload-update, power7-store, power7-store-update,
114 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
116 * config/rs6000/power8.md (power8-load, power8-load-update,
117 power8-load-ext, power8-load-ext-update, power8-fpload,
118 power8-fpload-update, power8-store, power8-store-update-indexed,
119 power8-fpstore, power8-fpstore-update): Adjust.
120 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
122 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
123 titan_lsu_store, titan_lsu_fpstore): Adjust.
124 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
126 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
129 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
130 unrolled byte insns. Emit address increments after move insns.
132 2014-05-07 David Malcolm <dmalcolm@redhat.com>
134 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
135 const_gimple, rather than a gimple.
136 (gimple_call_builtin_p): Likewise, for the three variants.
138 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
139 (gimple_call_builtin_p): Likewise, for the three variants.
141 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
143 PR tree-optimization/61095
144 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
146 2014-05-07 Richard Biener <rguenther@suse.de>
148 PR tree-optimization/61034
149 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
150 (maybe_skip_until): Use translate to take into account
151 lattices when trying to do disambiguations.
152 (get_continuation_for_phi_1): Likewise.
153 (get_continuation_for_phi): Adjust for added translate
155 (walk_non_aliased_vuses): Likewise.
156 * tree-ssa-alias.h (get_continuation_for_phi): Adjust
158 (walk_non_aliased_vuses): Likewise.
159 (call_may_clobber_ref_p_1): Declare.
160 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also
161 disambiguate against calls. Stop early if we are
162 only supposed to disambiguate.
163 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
165 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
167 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
168 Emit an error when the function has arguments.
170 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
172 * cfgloop.h (unswitch_loops): Remove.
173 * doc/passes.texi: Remove references to loop-unswitch.c
174 * timevar.def (TV_LOOP_UNSWITCH): Remove.
176 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
178 * tree-vect-data-refs.c (vect_grouped_load_supported): New
179 check for loads group of length 3.
180 (vect_permute_load_chain): New permutations for loads group of
182 * tree-vect-stmts.c (vect_model_load_cost): Change cost
183 of vec_perm_shuffle for the new permutations.
185 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
187 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
188 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
189 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
190 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
191 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
192 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
193 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
194 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
196 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
198 * loop-unswitch.c: Delete.
200 2014-05-07 Richard Biener <rguenther@suse.de>
202 * config.gcc: Always set need_64bit_hwint to yes.
204 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
206 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
207 of using optimize_size.
209 2014-05-06 Mike Stump <mikestump@comcast.net>
211 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
213 2014-05-06 Joseph Myers <joseph@codesourcery.com>
215 * config/i386/sse.md (*mov<mode>_internal)
216 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
217 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
218 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
219 (*<code><mode>3, *andnot<mode>3<mask_name>)
220 (<mask_codefor><code><mode>3<mask_name>): Only consider
221 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
223 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
226 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
228 * lra-constraints.c (valid_address_p): Move earlier in file.
229 Add a constraint argument to the address_info version.
230 (satisfies_memory_constraint_p): New function.
231 (satisfies_address_constraint_p): Likewise.
232 (process_alt_operands, curr_insn_transform): Use them.
233 (process_address): Pass the constraint to valid_address_p when
234 checking address operands.
236 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
238 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
239 to their respective blocks. Fix inadvertent use of "node".
241 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
243 * emit-rtl.c (init_derived_machine_modes): New functionm, split
245 (init_emit_once): ...here.
246 * rtl.h (init_derived_machine_modes): Declare.
247 * toplev.c (do_compile): Call it even if no_backend.
249 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
250 Mike Stump <mikestump@comcast.net>
251 Richard Sandiford <rdsandiford@googlemail.com>
252 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
254 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
255 (rtx_equal_for_memref_p): Update comment.
256 (adjust_offset_for_component_ref): Use wide-int interfaces.
257 * builtins.c (get_object_alignment_2): Likewise.
258 (c_readstr): Likewise.
259 (target_char_cast): Add comment.
260 (determine_block_size): Use wide-int interfaces.
261 (expand_builtin_signbit): Likewise.
262 (fold_builtin_int_roundingfn): Likewise.
263 (fold_builtin_bitop): Likewise.
264 (fold_builtin_bswap): Likewise.
265 (fold_builtin_logarithm): Use signop.
266 (fold_builtin_pow): Likewise.
267 (fold_builtin_memory_op): Use wide-int interfaces.
268 (fold_builtin_object_size): Likewise.
269 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
270 nb_iterations_estimate.
271 (record_niter_bound): Use wide-int interfaces.
272 (get_estimated_loop_iterations_int): Likewise.
273 (get_estimated_loop_iterations): Likewise.
274 (get_max_loop_iterations): Likewise.
275 * cfgloop.h: Include wide-int.h.
276 (struct nb_iter_bound): Change bound to widest_int.
277 (struct loop): Change nb_iterations_upper_bound and
278 nb_iterations_estimate to widest_int.
279 (record_niter_bound): Switch to use widest_int.
280 (get_estimated_loop_iterations): Likewise.
281 (get_max_loop_iterations): Likewise.
282 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
284 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
285 * combine.c (try_combine): Likewise.
286 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
287 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
289 (aarch64_float_const_representable_p): Likewise.
290 * config/arc/arc.c: Include wide-int.h.
291 (arc_can_use_doloop_p): Use wide-int interfaces.
292 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
293 (vfp3_const_double_index): Likewise.
294 * config/avr/avr.c (avr_out_round): Likewise.
295 (avr_fold_builtin): Likewise.
296 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
297 (bfin_can_use_doloop_p): Likewise.
298 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
299 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
300 * config/i386/i386.c: Include wide-int.h.
301 (ix86_data_alignment): Use wide-int interfaces.
302 (ix86_local_alignment): Likewise.
303 (ix86_emit_swsqrtsf): Update real_from_integer.
304 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
305 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
306 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
307 (zero_constant): Likewise.
308 (input_operand): Likewise.
309 (splat_input_operand): Likewise.
310 (non_logical_cint_operand): Change const_double to const_wide_int.
311 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
312 (easy_altivec_constant): Remove comment.
313 (paired_expand_vector_init): Use CONSTANT_P.
314 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
315 (rs6000_emit_move): Update checks.
316 (rs6000_aggregate_candidate): Use wide-int interfaces.
317 (rs6000_expand_ternop_builtin): Likewise.
318 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
319 (rs6000_assemble_integer): Likewise.
320 (rs6000_hash_constant): Likewise.
321 (output_toc): Likewise.
322 (rs6000_rtx_costs): Likewise.
323 (rs6000_emit_swrsqrt); Update call to real_from_integer.
324 * config/rs6000/rs6000-c.c: Include wide-int.h.
325 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
326 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
327 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
328 Handle CONST_WIDE_INT.
329 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
330 Use tree_fits_uhwi_p.
331 * config/sparc/sparc.c: Include wide-int.h.
332 (sparc_fold_builtin): Use wide-int interfaces.
333 * config/vax/vax.c: Include wide-int.h.
334 (vax_float_literal): Use real_from_integer.
335 * coretypes.h (struct hwivec_def): New.
338 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
339 (equiv_constant): Handle CONST_WIDE_INT.
340 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
341 (cselib_hash_rtx): Handle CONST_WIDE_INT.
342 * dbxout.c (stabstr_U): Use wide-int interfaces.
343 (dbxout_type): Update to use cst_fits_shwi_p.
344 * defaults.h (LOG2_BITS_PER_UNIT): Define.
345 (TARGET_SUPPORTS_WIDE_INT): Add default.
346 * dfp.c: Include wide-int.h.
347 (decimal_real_to_integer2): Use wide-int interfaces and rename to
348 decimal_real_to_integer.
349 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
350 decimal_real_to_integer.
351 * doc/generic.texi (Constant expressions): Update for wide_int.
352 * doc/rtl.texi (const_double): Likewise.
353 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
354 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
355 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
356 (REAL_VALUE_FROM_INT): Remove.
357 (TARGET_SUPPORTS_WIDE_INT): New.
358 * doc/tm.texi: Regenerate.
359 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
360 * double-int.h: Include wide-int.h.
361 (struct wi::int_traits): New.
362 * dwarf2out.c (get_full_len): New.
363 (dw_val_equal_p): Add case dw_val_class_wide_int.
364 (size_of_loc_descr): Likewise.
365 (output_loc_operands): Likewise.
366 (insert_double): Remove.
367 (insert_wide_int): New.
369 (print_die): Add case dw_val_class_wide_int.
370 (attr_checksum): Likewise.
371 (attr_checksum_ordered): Likewise.
372 (same_dw_val_p): Likewise.
373 (size_of_die): Likewise.
374 (value_format): Likewise.
375 (output_die): Likewise.
376 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
378 (clz_loc_descriptor): Use wide-int interfaces.
379 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
380 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
381 (round_up_to_align): Use wide-int interfaces.
382 (field_byte_offset): Likewise.
383 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
384 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
385 CONST_DOUBLE handling. Use wide-int interfaces.
386 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
387 (gen_enumeration_type_die): Use add_AT_wide.
388 (hash_loc_operands): Add case dw_val_class_wide_int.
389 (compare_loc_operands): Likewise.
390 * dwarf2out.h: Include wide-int.h.
392 (enum dw_val_class): Add dw_val_class_wide_int.
393 (struct dw_val_struct): Add val_wide.
394 * emit-rtl.c (const_wide_int_htab): New.
395 (const_wide_int_htab_hash): New.
396 (const_wide_int_htab_eq): New.
397 (lookup_const_wide_int): New.
398 (const_double_htab_hash): Use wide-int interfaces.
399 (const_double_htab_eq): Likewise.
400 (rtx_to_double_int): Conditionally compile for wide-int.
401 (immed_double_int_const): Rename to immed_wide_int_const and
403 (immed_double_const): Conditionally compile for wide-int.
404 (init_emit_once): Use wide-int interfaces.
405 * explow.c (plus_constant): Likewise.
406 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
407 (lshift_value): Use wide-int interfaces.
408 (expand_mult): Likewise.
409 (choose_multiplier): Likewise.
410 (expand_smod_pow2): Likewise.
411 (make_tree): Likewise.
412 * expr.c (convert_modes): Consolidate handling of constants.
413 Use wide-int interfaces.
414 (emit_group_load_1): Add note.
415 (store_expr): Update comment.
416 (get_inner_reference): Use wide-int interfaces.
417 (expand_constructor): Update comment.
418 (expand_expr_real_2): Use wide-int interfaces.
419 (expand_expr_real_1): Likewise.
420 (reduce_to_bit_field_precision): Likewise.
421 (const_vector_from_tree): Likewise.
422 * final.c: Include wide-int-print.h.
423 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
424 * fixed-value.c: Include wide-int.h.
425 (fixed_from_string): Use wide-int interfaces.
426 (fixed_to_decimal): Likewise.
427 (fixed_convert_from_real): Likewise.
428 (real_convert_from_fixed): Likewise.
429 * fold-const.h (mem_ref_offset): Return an offset_int.
430 (div_if_zero_remainder): Remove code parameter.
431 * fold-const.c (div_if_zero_remainder): Remove code parameter.
432 Use wide-int interfaces.
433 (may_negate_without_overflow_p): Use wide-int interfaces.
434 (negate_expr_p): Likewise.
435 (fold_negate_expr): Likewise.
436 (int_const_binop_1): Likewise.
437 (const_binop): Likewise.
438 (fold_convert_const_int_from_int): Likewise.
439 (fold_convert_const_int_from_real): Likewise.
440 (fold_convert_const_int_from_fixed): Likewise.
441 (fold_convert_const_fixed_from_int): Likewise.
442 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
443 (sign_bit_p): Use wide-int interfaces.
444 (make_range_step): Likewise.
445 (build_range_check): Likewise. Pass an integer of the correct type
446 instead of using integer_one_node.
447 (range_predecessor): Pass an integer of the correct type instead
448 of using integer_one_node.
449 (range_successor): Likewise.
450 (merge_ranges): Likewise.
451 (unextend): Use wide-int interfaces.
452 (extract_muldiv_1): Likewise.
453 (fold_div_compare): Likewise.
454 (fold_single_bit_test): Likewise.
455 (fold_sign_changed_comparison): Likewise.
456 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
457 (fold_plusminus_mult_expr): Use wide-int interfaces.
458 (native_encode_int): Likewise.
459 (native_interpret_int): Likewise.
460 (fold_unary_loc): Likewise.
461 (pointer_may_wrap_p): Likewise.
462 (size_low_cst): Likewise.
463 (mask_with_tz): Likewise.
464 (fold_binary_loc): Likewise.
465 (fold_ternary_loc): Likewise.
466 (multiple_of_p): Likewise.
467 (tree_call_nonnegative_warnv_p): Update calls to
468 tree_int_cst_min_precision and real_from_integer.
469 (fold_negate_const): Use wide-int interfaces.
470 (fold_abs_const): Likewise.
471 (fold_relational_const): Use tree_int_cst_lt.
472 (round_up_loc): Use wide-int interfaces.
473 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
474 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
475 * gengtype.c: Remove include of double-int.h.
476 (do_typedef): Use wide-int interfaces.
477 (open_base_files): Add wide-int.h.
478 (main): Add offset_int and widest_int typedefs.
479 * gengtype-lex.l: Handle "^".
480 (CXX_KEYWORD): Add "static".
481 * gengtype-parse.c (require3): New.
482 (require_template_declaration): Handle constant template arguments
483 and nested templates.
484 * gengtype-state.c: Don't include "double-int.h".
485 * genpreds.c (write_one_predicate_function): Update comment.
486 (write_tm_constrs_h): Add check for hval and lval use in
488 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
489 (add_to_sequence): Likewise.
490 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
491 and const_double_operand.
492 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
494 * gimple-fold.c (get_base_constructor): Likewise.
495 (fold_array_ctor_reference): Likewise.
496 (fold_nonarray_ctor_reference): Likewise.
497 (fold_const_aggregate_ref_1): Likewise.
498 (gimple_val_nonnegative_real_p): Likewise.
499 (gimple_fold_indirect_ref): Likewise.
500 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
501 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
502 (struct slsr_cand_d): Change index to be widest_int.
503 (struct incr_info_d): Change incr to be widest_int.
504 (alloc_cand_and_find_basis): Use wide-int interfaces.
505 (slsr_process_phi): Likewise.
506 (backtrace_base_for_ref): Likewise. Return a widest_int.
507 (restructure_reference): Take a widest_int instead of a double_int.
508 (slsr_process_ref): Use wide-int interfaces.
509 (create_mul_ssa_cand): Likewise.
510 (create_mul_imm_cand): Likewise.
511 (create_add_ssa_cand): Likewise.
512 (create_add_imm_cand): Take a widest_int instead of a double_int.
513 (slsr_process_add): Use wide-int interfaces.
514 (slsr_process_cast): Likewise.
515 (slsr_process_copy): Likewise.
516 (dump_candidate): Likewise.
517 (dump_incr_vec): Likewise.
518 (replace_ref): Likewise.
519 (cand_increment): Likewise. Return a widest_int.
520 (cand_abs_increment): Likewise.
521 (replace_mult_candidate): Take a widest_int instead of a double_int.
522 (replace_unconditional_candidate): Use wide-int interfaces.
523 (incr_vec_index): Take a widest_int instead of a double_int.
524 (create_add_on_incoming_edge): Likewise.
525 (create_phi_basis): Use wide-int interfaces.
526 (replace_conditional_candidate): Likewise.
527 (record_increment): Take a widest_int instead of a double_int.
528 (record_phi_increments): Use wide-int interfaces.
529 (phi_incr_cost): Take a widest_int instead of a double_int.
530 (lowest_cost_path): Likewise.
531 (total_savings): Likewise.
532 (analyze_increments): Use wide-int interfaces.
533 (ncd_with_phi): Take a widest_int instead of a double_int.
534 (ncd_of_cand_and_phis): Likewise.
535 (nearest_common_dominator_for_cands): Likewise.
536 (insert_initializers): Use wide-int interfaces.
537 (all_phi_incrs_profitable): Likewise.
538 (replace_one_candidate): Likewise.
539 (replace_profitable_candidates): Likewise.
540 * godump.c: Include wide-int-print.h.
541 (go_output_typedef): Use wide-int interfaces.
542 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
543 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
544 (build_loop_iteration_domains): Likewise.
545 * hooks.h: Include wide-int.h rather than double-int.h.
546 (hook_bool_dint_dint_uint_bool_true): Delete.
547 (hook_bool_wint_wint_uint_bool_true): Declare.
548 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
549 (hook_bool_wint_wint_uint_bool_true): New.
550 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
552 (ubsan_expand_si_overflow_mul_check): Likewise.
553 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
554 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
555 (get_ancestor_addr_info): Likewise.
556 (ipa_modify_call_arguments): Likewise.
557 * loop-doloop.c (doloop_modify): Likewise.
558 (doloop_optimize): Likewise.
559 * loop-iv.c (iv_number_of_iterations): Likewise.
560 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
561 (unroll_loop_constant_iterations): Likewise.
562 (decide_unroll_runtime_iterations): Likewise.
563 (unroll_loop_runtime_iterations): Likewise.
564 (decide_peel_simple): Likewise.
565 (decide_unroll_stupid): Likewise.
566 * lto-streamer-in.c (streamer_read_wi): Add.
567 (input_cfg): Use wide-int interfaces.
568 (lto_input_tree_1): Likewise.
569 * lto-streamer-out.c (streamer_write_wi): Add.
570 (hash_tree): Use wide-int interfaces.
571 (output_cfg): Likewise.
572 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
573 (GTFILES): Add wide-int.h and signop.h.
574 (TAGS): Look for .cc files too.
575 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
576 * optabs.c (expand_subword_shift): Likewise.
577 (expand_doubleword_shift): Likewise.
578 (expand_absneg_bit): Likewise.
579 (expand_copysign_absneg): Likewise.
580 (expand_copysign_bit): Likewise.
581 * postreload.c (reload_cse_simplify_set): Likewise.
582 * predict.c (predict_iv_comparison): Likewise.
583 * pretty-print.h: Include wide-int-print.h.
585 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
586 * print-tree.c: Include wide-int-print.h.
587 (print_node_brief): Use wide-int interfaces.
588 (print_node): Likewise.
589 * read-rtl.c (validate_const_wide_int): New.
590 (read_rtx_code): Add CONST_WIDE_INT case.
591 * real.c: Include wide-int.h.
592 (real_to_integer2): Delete.
593 (real_to_integer): New function, returning a wide_int.
594 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
595 (ten_to_ptwo): Update call to real_from_integer.
596 (real_digit): Likewise.
597 * real.h: Include signop.h, wide-int.h and insn-modes.h.
598 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
599 (REAL_VALUE_TO_INT): Delete.
600 (real_to_integer): Declare a wide-int form.
601 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
602 * recog.c (const_int_operand): Improve comment.
603 (const_scalar_int_operand): New.
604 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
605 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
606 (split_double): Likewise.
607 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
608 (rtx_size): Likewise.
609 (rtx_alloc_stat_v): New.
610 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
611 (cwi_output_hex): New.
612 (iterative_hash_rtx): Handle CONST_WIDE_INT.
613 (cwi_check_failed_bounds): New.
614 * rtl.def (CONST_WIDE_INT): New.
615 * rtl.h: Include <utility> and wide-int.h.
616 (struct hwivec_def): New.
617 (CWI_GET_NUM_ELEM): New.
618 (CWI_PUT_NUM_ELEM): New.
619 (struct rtx_def): Add num_elem and hwiv.
620 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
621 (CASE_CONST_UNIQUE): Likewise.
622 (CASE_CONST_ANY): Likewise.
623 (CONST_SCALAR_INT_P): Likewise.
624 (CONST_WIDE_INT_P): New.
627 (cwi_check_failed_bounds): New.
630 (CONST_WIDE_INT_VEC) New.
631 (CONST_WIDE_INT_NUNITS) New.
632 (CONST_WIDE_INT_ELT) New.
633 (rtx_mode_t): New type.
634 (wi::int_traits <rtx_mode_t>): New.
636 (wi::min_value): New.
637 (wi::max_value): New.
639 (const_wide_int_alloc): New.
640 (immed_wide_int_const): New.
641 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
642 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
643 * signop.h: New file.
644 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
645 (simplify_const_unary_operation): Use wide-int interfaces.
646 (simplify_binary_operation_1): Likewise.
647 (simplify_const_binary_operation): Likewise.
648 (simplify_const_relational_operation): Likewise.
649 (simplify_immed_subreg): Likewise.
650 * stmt.c (expand_case): Likewise.
651 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
652 signop rather than a bool.
653 * stor-layout.c (layout_type): Use wide-int interfaces.
654 (initialize_sizetypes): Update calls to
655 set_min_and_max_values_for_integral_type.
656 (set_min_and_max_values_for_integral_type): Take a signop rather
657 than a bool. Use wide-int interfaces.
658 (fixup_signed_type): Update accordingly. Remove
659 HOST_BITS_PER_DOUBLE_INT limit.
660 (fixup_unsigned_type): Likewise.
661 * system.h (STATIC_CONSTANT_P): New.
662 (STATIC_ASSERT): New.
663 * target.def (can_use_doloop_p): Take widest_ints rather than
665 * target.h: Include wide-int.h rather than double-int.h.
666 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
668 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
669 rather than INT_CST_LT_UNSIGNED.
670 (can_use_doloop_if_innermost): Take widest_ints rather than
672 * tree-affine.c: Include wide-int-print.h.
673 (double_int_ext_for_comb): Delete.
674 (wide_int_ext_for_comb): New.
675 (aff_combination_zero): Use wide-int interfaces.
676 (aff_combination_const): Take a widest_int instead of a double_int.
677 (aff_combination_elt): Use wide-int interfaces.
678 (aff_combination_scale): Take a widest_int instead of a double_int.
679 (aff_combination_add_elt): Likewise.
680 (aff_combination_add_cst): Likewise.
681 (aff_combination_add): Use wide-int interfaces.
682 (aff_combination_convert): Likewise.
683 (tree_to_aff_combination): Likewise.
684 (add_elt_to_tree): Take a widest_int instead of a double_int.
685 (aff_combination_to_tree): Use wide-int interfaces.
686 (aff_combination_remove_elt): Likewise.
687 (aff_combination_add_product): Take a widest_int instead of
689 (aff_combination_mult): Use wide-int interfaces.
690 (aff_combination_expand): Likewise.
691 (double_int_constant_multiple_p): Delete.
692 (wide_int_constant_multiple_p): New.
693 (aff_combination_constant_multiple_p): Take a widest_int pointer
694 instead of a double_int pointer.
695 (print_aff): Use wide-int interfaces.
696 (get_inner_reference_aff): Take a widest_int pointer
697 instead of a double_int pointer.
698 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
699 * tree-affine.h: Include wide-int.h.
700 (struct aff_comb_elt): Change type of coef to widest_int.
701 (struct affine_tree_combination): Change type of offset to widest_int.
702 (double_int_ext_for_comb): Delete.
703 (wide_int_ext_for_comb): New.
704 (aff_combination_const): Use widest_int instead of double_int.
705 (aff_combination_scale): Likewise.
706 (aff_combination_add_elt): Likewise.
707 (aff_combination_constant_multiple_p): Likewise.
708 (get_inner_reference_aff): Likewise.
709 (aff_comb_cannot_overlap_p): Likewise.
710 (aff_combination_zero_p): Use wide-int interfaces.
711 * tree.c: Include tree.h.
712 (init_ttree): Use make_int_cst.
713 (tree_code_size): Removed code for INTEGER_CST case.
714 (tree_size): Add INTEGER_CST case.
715 (make_node_stat): Update comment.
716 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
717 (build_int_cst_type): Use wide-int interfaces.
718 (double_int_to_tree): Likewise.
719 (double_int_fits_to_tree_p): Delete.
720 (force_fit_type_double): Delete.
721 (force_fit_type): New.
722 (int_cst_hash_hash): Use wide-int interfaces.
723 (int_cst_hash_eq): Likewise.
724 (build_int_cst_wide): Delete.
725 (wide_int_to_tree): New.
726 (cache_integer_cst): Use wide-int interfaces.
727 (build_low_bits_mask): Likewise.
728 (cst_and_fits_in_hwi): Likewise.
729 (real_value_from_int_cst): Likewise.
730 (make_int_cst_stat): New.
731 (integer_zerop): Use wide_int interfaces.
732 (integer_onep): Likewise.
733 (integer_all_onesp): Likewise.
734 (integer_pow2p): Likewise.
735 (integer_nonzerop): Likewise.
736 (tree_log2): Likewise.
737 (tree_floor_log2): Likewise.
738 (tree_ctz): Likewise.
739 (int_size_in_bytes): Likewise.
740 (mem_ref_offset): Return an offset_int rather than a double_int.
741 (build_type_attribute_qual_variant): Use wide_int interfaces.
742 (type_hash_eq): Likewise
743 (tree_int_cst_equal): Likewise.
744 (tree_int_cst_lt): Delete.
745 (tree_int_cst_compare): Likewise.
746 (tree_fits_shwi_p): Use wide_int interfaces.
747 (tree_fits_uhwi_p): Likewise.
748 (tree_int_cst_sign_bit): Likewise.
749 (tree_int_cst_sgn): Likewise.
750 (tree_int_cst_min_precision): Take a signop rather than a bool.
751 (simple_cst_equal): Use wide_int interfaces.
752 (compare_tree_int): Likewise.
753 (iterative_hash_expr): Likewise.
754 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
756 (get_type_static_bounds): Use wide_int interfaces.
757 (tree_int_cst_elt_check_failed): New.
758 (build_common_tree_nodes): Reordered to set prec before filling in
760 (int_cst_value): Check cst_and_fits_in_hwi.
761 (widest_int_cst_value): Use wide_int interfaces.
762 (upper_bound_in_type): Likewise.
763 (lower_bound_in_type): Likewise.
764 (num_ending_zeros): Likewise.
765 (drop_tree_overflow): Likewise.
766 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
767 (gen_conditions_for_pow_cst_base): Likewise.
768 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
769 (group_case_labels_stmt): Use wide-int interfaces.
770 (verify_gimple_assign_binary): Likewise.
771 (print_loop): Likewise.
772 * tree-chrec.c (tree_fold_binomial): Likewise.
773 * tree-core.h (struct tree_base): Add int_length.
774 (struct tree_int_cst): Change rep of value.
775 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
776 (dr_may_alias_p): Likewise.
777 (max_stmt_executions_tree): Likewise.
778 * tree.def (INTEGER_CST): Update comment.
779 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
780 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
781 * tree-dump.c: Include wide-int.h and wide-int-print.h.
782 (dequeue_and_dump): Use wide-int interfaces.
783 * tree.h: Include wide-int.h.
784 (NULL_TREE): Moved to earlier loc in file.
785 (TREE_INT_CST_ELT_CHECK): New.
786 (tree_int_cst_elt_check_failed): New.
788 (TREE_INT_CST): Delete.
789 (TREE_INT_CST_LOW): Use wide-int interfaces.
790 (TREE_INT_CST_HIGH): Delete.
791 (TREE_INT_CST_NUNITS): New.
792 (TREE_INT_CST_EXT_NUNITS): Likewise.
793 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
794 (TREE_INT_CST_ELT): Likewise.
795 (INT_CST_LT): Delete.
796 (tree_int_cst_elt_check): New (two forms).
797 (type_code_size): Update comment.
798 (make_int_cst_stat, make_int_cst): New.
799 (tree_to_double_int): Delete.
800 (double_int_fits_to_tree_p): Delete.
801 (force_fit_type_double): Delete.
802 (build_int_cstu): Replace with out-of-line function.
803 (build_int_cst_wide): Delete.
804 (tree_int_cst_lt): Define inline.
805 (tree_int_cst_le): New.
806 (tree_int_cst_compare): Define inline.
807 (tree_int_cst_min_precision): Take a signop rather than a bool.
808 (wi::int_traits <const_tree>): New.
809 (wi::int_traits <tree>): New.
810 (wi::extended_tree): New.
811 (wi::int_traits <wi::extended_tree>): New.
812 (wi::to_widest): New.
813 (wi::to_offset): New.
814 (wi::fits_to_tree_p): New.
815 (wi::min_value): New.
816 (wi::max_value): New.
817 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
818 (copy_tree_body_r): Likewise.
819 * tree-object-size.c (compute_object_offset): Likewise.
820 (addr_object_size): Likewise.
821 * tree-predcom.c: Include wide-int-print.h.
822 (struct dref_d): Change type of offset to widest_int.
823 (dump_dref): Call wide-int printer.
824 (aff_combination_dr_offset): Use wide-int interfaces.
825 (determine_offset): Take a widest_int pointer rather than a
827 (split_data_refs_to_components): Use wide-int interfaces.
828 (suitable_component_p): Likewise.
829 (order_drefs): Likewise.
830 (add_ref_to_chain): Likewise.
831 (valid_initializer_p): Likewise.
832 (determine_roots_comp): Likewise.
833 * tree-pretty-print.c: Include wide-int-print.h.
834 (dump_generic_node): Use wide-int interfaces.
835 * tree-sra.c (sra_ipa_modify_expr): Likewise.
836 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
837 (move_fixed_address_to_symbol): Likewise.
838 (move_hint_to_base): Likewise.
839 (move_pointer_to_base): Likewise.
840 (move_variant_to_index): Likewise.
841 (most_expensive_mult_to_index): Likewise.
842 (addr_to_parts): Likewise.
843 (copy_ref_info): Likewise.
844 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
845 (indirect_refs_may_alias_p): Likewise.
846 (stmt_kills_ref_p_1): Likewise.
847 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
848 * tree-ssa-ccp.c: Update comment at top of file. Include
850 (struct prop_value_d): Change type of mask to widest_int.
851 (extend_mask): New function.
852 (dump_lattice_value): Use wide-int interfaces.
853 (get_default_value): Likewise.
854 (set_constant_value): Likewise.
855 (set_value_varying): Likewise.
856 (valid_lattice_transition): Likewise.
857 (set_lattice_value): Likewise.
858 (value_to_double_int): Delete.
859 (value_to_wide_int): New.
860 (get_value_from_alignment): Use wide-int interfaces.
861 (get_value_for_expr): Likewise.
862 (do_dbg_cnt): Likewise.
863 (ccp_finalize): Likewise.
864 (ccp_lattice_meet): Likewise.
865 (bit_value_unop_1): Use widest_ints rather than double_ints.
866 (bit_value_binop_1): Likewise.
867 (bit_value_unop): Use wide-int interfaces.
868 (bit_value_binop): Likewise.
869 (bit_value_assume_aligned): Likewise.
870 (evaluate_stmt): Likewise.
871 (ccp_fold_stmt): Likewise.
872 (visit_cond_stmt): Likewise.
873 (ccp_visit_stmt): Likewise.
874 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
875 (constant_pointer_difference): Likewise.
876 (associate_pointerplus): Likewise.
877 (combine_conversions): Likewise.
878 * tree-ssa-loop.h: Include wide-int.h.
879 (struct tree_niter_desc): Change type of max to widest_int.
880 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
881 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
882 (remove_redundant_iv_tests): Likewise.
883 (canonicalize_loop_induction_variables): Likewise.
884 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
885 (constant_multiple_of): Take a widest_int pointer instead of
886 a double_int pointer.
887 (get_computation_aff): Use wide-int interfaces.
888 (ptr_difference_cost): Likewise.
889 (difference_cost): Likewise.
890 (get_loop_invariant_expr_id): Likewise.
891 (get_computation_cost_at): Likewise.
892 (iv_elimination_compare_lt): Likewise.
893 (may_eliminate_iv): Likewise.
894 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
895 instead of double_int.
896 (max_loop_iterations): Likewise.
897 (max_stmt_executions): Likewise.
898 (estimated_stmt_executions): Likewise.
899 * tree-ssa-loop-niter.c: Include wide-int-print.h.
900 (split_to_var_and_offset): Use wide-int interfaces.
901 (determine_value_range): Likewise.
902 (bound_difference_of_offsetted_base): Likewise.
903 (bounds_add): Take a widest_int instead of a double_int.
904 (number_of_iterations_ne_max): Use wide-int interfaces.
905 (number_of_iterations_ne): Likewise.
906 (number_of_iterations_lt_to_ne): Likewise.
907 (assert_loop_rolls_lt): Likewise.
908 (number_of_iterations_lt): Likewise.
909 (number_of_iterations_le): Likewise.
910 (number_of_iterations_cond): Likewise.
911 (number_of_iterations_exit): Likewise.
912 (finite_loop_p): Likewise.
913 (derive_constant_upper_bound_assign): Likewise.
914 (derive_constant_upper_bound): Return a widest_int.
915 (derive_constant_upper_bound_ops): Likewise.
916 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
917 (record_estimate): Take a widest_int rather than a double_int.
918 (record_nonwrapping_iv): Use wide-int interfaces.
919 (double_int_cmp): Delete.
921 (bound_index): Take a widest_int rather than a double_int.
922 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
923 (maybe_lower_iteration_bound): Likewise.
924 (estimate_numbers_of_iterations_loop): Likewise.
925 (estimated_loop_iterations): Take a widest_int pointer than than
926 a double_int pointer.
927 (estimated_loop_iterations_int): Use wide-int interfaces.
928 (max_loop_iterations): Take a widest_int pointer than than
929 a double_int pointer.
930 (max_loop_iterations_int): Use wide-int interfaces.
931 (max_stmt_executions): Take a widest_int pointer than than
932 a double_int pointer.
933 (estimated_stmt_executions): Likewise.
934 (n_of_executions_at_most): Use wide-int interfaces.
935 (scev_probably_wraps_p): Likewise.
936 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
938 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
940 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
941 double_ints. Adjust for trailing_wide_ints <3> representation.
942 (set_nonzero_bits): Likewise.
943 (get_range_info): Return wide_ints rather than double_ints.
944 Adjust for trailing_wide_ints <3> representation.
945 (get_nonzero_bits): Likewise.
946 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
948 * tree-ssanames.h (struct range_info_def): Replace min, max and
949 nonzero_bits with a trailing_wide_ints <3>.
950 (set_range_info): Use wide_int_refs rather than double_ints.
951 (set_nonzero_bits): Likewise.
952 (get_range_info): Return wide_ints rather than double_ints.
953 (get_nonzero_bits): Likewise.
954 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
955 * tree-ssa-pre.c (phi_translate_1): Likewise.
956 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
957 (acceptable_pow_call): Likewise.
958 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
960 (vn_reference_fold_indirect): Likewise.
961 (vn_reference_maybe_forwprop_address): Likewise.
962 (valueize_refs_1): Likewise.
963 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
964 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
965 tree_int_cst_lt and tree_int_cst_le.
966 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
968 (streamer_alloc_tree): Likewise.
969 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
970 (streamer_write_tree_header): Likewise.
971 (streamer_write_integer_cst): Likewise.
972 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
973 (build_constructors): Likewise.
974 (array_value_type): Likewise.
975 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
976 (vect_check_gather): Likewise.
977 * tree-vect-generic.c (build_replicated_const): Likewise.
978 (expand_vector_divmod): Likewise.
979 * tree-vect-loop.c (vect_transform_loop): Likewise.
980 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
981 (vect_do_peeling_for_alignment): Likewise.
982 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
983 * tree-vrp.c: Include wide-int.h.
984 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
985 (extract_range_from_assert): Use wide-int interfaces.
986 (vrp_int_const_binop): Likewise.
987 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
989 (ranges_from_anti_range): Use wide-int interfaces.
990 (quad_int_cmp): Delete.
991 (quad_int_pair_sort): Likewise.
992 (extract_range_from_binary_expr_1): Use wide-int interfaces.
993 (extract_range_from_unary_expr_1): Likewise.
994 (adjust_range_with_scev): Likewise.
995 (masked_increment): Take and return wide_ints rather than double_ints.
996 (register_edge_assert_for_2): Use wide-int interfaces.
997 (check_array_ref): Likewise.
998 (search_for_addr_array): Likewise.
999 (maybe_set_nonzero_bits): Likewise.
1000 (union_ranges): Pass an integer of the correct type instead of
1001 using integer_one_node.
1002 (intersect_ranges): Likewise.
1003 (simplify_truth_ops_using_ranges): Likewise.
1004 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
1005 (range_fits_type_p): Likewise.
1006 (simplify_cond_using_ranges): Likewise. Take a signop rather than
1008 (simplify_conversion_using_ranges): Use wide-int interfaces.
1009 (simplify_float_conversion_using_ranges): Likewise.
1010 (vrp_finalize): Likewise.
1011 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
1012 (gimple_stringops_transform): Likewise.
1013 * varasm.c (decode_addr_const): Likewise.
1014 (const_hash_1): Likewise.
1015 (const_rtx_hash_1): Likewise
1016 (output_constant): Likewise.
1017 (array_size_for_constructor): Likewise.
1018 (output_constructor_regular_field): Likewise.
1019 (output_constructor_bitfield): Likewise.
1020 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
1021 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
1023 * gencheck.c: Define BITS_PER_UNIT.
1026 * wide-int-print.cc: New.
1027 * wide-int-print.h: New.
1029 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1031 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
1033 2014-05-06 Richard Biener <rguenther@suse.de>
1035 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
1036 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
1037 (TODO_verify_all): Adjust.
1038 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
1039 TODO_verify_stmts and TODO_verify_rtl_sharing.
1040 * bb-reorder.c: Likewise.
1041 * cfgexpand.c: Likewise.
1042 * cprop.c: Likewise.
1044 * function.c: Likewise.
1045 * fwprop.c: Likewise.
1047 * gimple-ssa-isolate-paths.c: Likewise.
1048 * gimple-ssa-strength-reduction.c: Likewise.
1049 * ipa-split.c: Likewise.
1050 * loop-init.c: Likewise.
1051 * loop-unroll.c: Likewise.
1052 * lower-subreg.c: Likewise.
1053 * modulo-sched.c: Likewise.
1054 * postreload-gcse.c: Likewise.
1055 * predict.c: Likewise.
1056 * recog.c: Likewise.
1057 * sched-rgn.c: Likewise.
1058 * store-motion.c: Likewise.
1059 * tracer.c: Likewise.
1060 * trans-mem.c: Likewise.
1061 * tree-call-cdce.c: Likewise.
1062 * tree-cfg.c: Likewise.
1063 * tree-cfgcleanup.c: Likewise.
1064 * tree-complex.c: Likewise.
1065 * tree-eh.c: Likewise.
1066 * tree-emutls.c: Likewise.
1067 * tree-if-conv.c: Likewise.
1068 * tree-into-ssa.c: Likewise.
1069 * tree-loop-distribution.c: Likewise.
1070 * tree-object-size.c: Likewise.
1071 * tree-parloops.c: Likewise.
1072 * tree-pass.h: Likewise.
1073 * tree-sra.c: Likewise.
1074 * tree-ssa-ccp.c: Likewise.
1075 * tree-ssa-copy.c: Likewise.
1076 * tree-ssa-copyrename.c: Likewise.
1077 * tree-ssa-dce.c: Likewise.
1078 * tree-ssa-dom.c: Likewise.
1079 * tree-ssa-dse.c: Likewise.
1080 * tree-ssa-forwprop.c: Likewise.
1081 * tree-ssa-ifcombine.c: Likewise.
1082 * tree-ssa-loop-ch.c: Likewise.
1083 * tree-ssa-loop-ivcanon.c: Likewise.
1084 * tree-ssa-loop.c: Likewise.
1085 * tree-ssa-math-opts.c: Likewise.
1086 * tree-ssa-phiopt.c: Likewise.
1087 * tree-ssa-phiprop.c: Likewise.
1088 * tree-ssa-pre.c: Likewise.
1089 * tree-ssa-reassoc.c: Likewise.
1090 * tree-ssa-sink.c: Likewise.
1091 * tree-ssa-strlen.c: Likewise.
1092 * tree-ssa-tail-merge.c: Likewise.
1093 * tree-ssa-uncprop.c: Likewise.
1094 * tree-switch-conversion.c: Likewise.
1095 * tree-tailcall.c: Likewise.
1096 * tree-vect-generic.c: Likewise.
1097 * tree-vectorizer.c: Likewise.
1098 * tree-vrp.c: Likewise.
1100 * var-tracking.c: Likewise.
1101 * bt-load.c: Likewise.
1102 * cfgcleanup.c: Likewise.
1103 * combine-stack-adj.c: Likewise.
1104 * combine.c: Likewise.
1105 * compare-elim.c: Likewise.
1106 * config/epiphany/resolve-sw-modes.c: Likewise.
1107 * config/i386/i386.c: Likewise.
1108 * config/mips/mips.c: Likewise.
1109 * config/s390/s390.c: Likewise.
1110 * config/sh/sh_treg_combine.cc: Likewise.
1111 * config/sparc/sparc.c: Likewise.
1114 * final.c: Likewise.
1115 * ifcvt.c: Likewise.
1116 * mode-switching.c: Likewise.
1117 * passes.c: Likewise.
1118 * postreload.c: Likewise.
1120 * reg-stack.c: Likewise.
1121 * regcprop.c: Likewise.
1122 * regrename.c: Likewise.
1125 2014-05-06 Richard Biener <rguenther@suse.de>
1128 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
1129 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
1131 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
1134 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
1136 2014-05-05 Radovan Obradovic <robradovic@mips.com>
1137 Tom de Vries <tom@codesourcery.com>
1139 * target.def (call_fusage_contains_non_callee_clobbers): New DEFHOOKPOD.
1140 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
1142 (@node Miscellaneous Register Hooks): New node.
1143 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
1144 * doc/tm.texi: Regenerate.
1146 2014-05-05 Marek Polacek <polacek@redhat.com>
1149 * opts.c (common_handle_option): Call error_at instead of warning_at.
1151 2014-05-05 Richard Biener <rguenther@suse.de>
1153 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
1154 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
1155 under the TODO_verify_il umbrella.
1157 2014-05-05 Richard Biener <rguenther@suse.de>
1159 * passes.c (execute_function_todo): Move TODO_verify_flow under
1160 the TODO_verify_ul umbrella.
1162 2014-05-05 Richard Biener <rguenther@suse.de>
1165 * fold-const.c (fold_binary_loc): Consistently avoid
1166 canonicalizing X & CST away from a CST that is the mask
1169 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1171 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
1172 int argument to enum machine_mode.
1173 (picochip_class_max_nregs): Ditto.
1174 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
1175 (picochip_class_max_nregs): Ditto.
1177 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1179 * target.def: Add new target hook.
1180 * doc/tm.texi: Regenerate.
1181 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
1182 * targhooks.c (default_keep_leaf_when_profiled): New function.
1184 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
1185 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
1187 2014-05-05 Bin Cheng <bin.cheng@arm.com>
1189 PR tree-optimization/60363
1190 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
1191 (copy_phi_args): New parameters. Call get_value_locus_in_path.
1192 (update_destination_phis): New parameter.
1193 (create_edge_and_update_destination_phis): Ditto.
1194 (ssa_fix_duplicate_block_edges): Pass new arguments.
1195 (thread_single_edge): Ditto.
1197 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
1199 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
1200 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
1201 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
1202 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
1203 Use RS6000_BTM_HARD_FLOAT.
1204 (BU_MISC_2): Likewise.
1205 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
1206 RS6000_BTM_HARD_FLOAT.
1207 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
1209 (rs6000_invalid_builtin): Add hard floating builtin support.
1210 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
1211 hard float builtins.
1212 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
1214 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1216 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
1217 Add missing function* argument.
1219 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1221 * lra-constraints.c (valid_address_p): Move earlier in file.
1222 Add a constraint argument to the address_info version.
1223 (satisfies_memory_constraint_p): New function.
1224 (satisfies_address_constraint_p): Likewise.
1225 (process_alt_operands, curr_insn_transform): Use them.
1226 (process_address): Pass the constraint to valid_address_p when
1227 checking address operands.
1229 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1231 * config/mips/mips.c (mips_isa_rev): New variable.
1232 (mips_set_architecture): Set it.
1233 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1235 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1236 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1237 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1238 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1239 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
1240 conditions in terms of mips_isa_rev.
1241 (mips_isa_rev): Declare.
1243 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1245 * config/sh/sh-mem.cc: Use tabs instead of spaces.
1246 (prob_unlikely, prob_likely): Make variables const.
1248 2014-05-03 Denis Chertykov <chertykov@gmail.com>
1250 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
1252 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1254 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
1256 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1258 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
1259 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
1260 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
1262 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
1263 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
1265 Replace usage of ROUND_REG with sh_round_reg.
1266 Use CEIL instead of ROUND_ADVANCE.
1268 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1271 * config/sh/sh.c: Include stdlib headers before everything else.
1273 2014-05-02 Jakub Jelinek <jakub@redhat.com>
1275 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
1276 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
1277 (gimplify_adjust_omp_clauses): Simd region is never
1278 directly nested in combined parallel. Instead, for linear
1279 with copyin/copyout, if in combined for simd loop, make decl
1280 firstprivate/lastprivate on OMP_FOR.
1281 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
1282 expand_omp_for_static_chunk): When setting endvar, also set
1283 fd->loop.v to the same value.
1285 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1287 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
1289 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
1291 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
1294 2014-05-02 Marek Polacek <polacek@redhat.com>
1296 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
1298 2014-05-02 Kito Cheng <kito@0xlab.org>
1300 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
1301 to a C expression marco.
1302 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
1303 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
1304 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
1305 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
1306 HONOR_REG_ALLOC_ORDER.
1307 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
1309 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1311 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
1313 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1315 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
1317 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
1319 * tree-if-conv.c (is_cond_scalar_reduction): New function.
1320 (convert_scalar_cond_reduction): Likewise.
1321 (predicate_scalar_phi): Add recognition and transformation
1322 of simple conditioanl reduction to be vectorizable.
1324 2014-05-01 Marek Polacek <polacek@redhat.com>
1327 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
1329 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
1331 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
1332 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
1333 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
1334 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
1335 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
1336 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
1337 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
1338 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
1340 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
1342 * config/arc/arc.opt (mlra): Move comment above option name
1343 to avoid mis-parsing as language options.
1345 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1347 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
1348 * config/sol2.h: ... here.
1349 * config/sol2-10.h: Remove.
1351 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
1352 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
1353 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
1354 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
1355 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
1356 * config/sol2.h: ... here.
1357 (SECTION_NAME_FORMAT): Don't redefine.
1358 (STARTFILE_ARCH32_SPEC): Rename to ...
1359 (STARTFILE_ARCH_SPEC): ... this.
1360 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
1361 * config/sparc/sol2.h: ... here.
1362 (SECTION_NAME_FORMAT): Don't undef.
1363 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
1364 (SUBTARGET_EXTRA_SPECS): Remove.
1365 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
1367 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
1368 (MD_STARTFILE_PREFIX): Remove.
1369 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
1370 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
1371 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
1372 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
1373 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
1374 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
1375 * config/i386/sol2.h: ... here.
1376 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
1377 * config/i386/sol2-bi.h: Remove.
1378 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
1379 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
1381 * config/i386/t-sol2-64: Rename to ...
1382 * config/i386/t-sol2: ... this.
1383 * config/sparc/t-sol2-64: Rename to ...
1384 * config/sparc/t-sol2: ... this.
1386 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
1387 sol2_tm_file_head, sol2_tm_file_tail.
1388 Include ${cpu_type}/sol2.h before sol2.h.
1390 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
1391 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
1392 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
1393 Reflect i386/t-sol2-64 renaming.
1394 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
1395 Reflect sparc/t-sol2-64 renaming.
1397 2014-04-30 Richard Biener <rguenther@suse.de>
1399 * passes.c (execute_function_todo): Move TODO_verify_stmts
1400 and TODO_verify_ssa under the TODO_verify_il umbrella.
1401 * tree-ssa.h (verify_ssa): Adjust prototype.
1402 * tree-ssa.c (verify_ssa): Add parameter to tell whether
1403 we should verify SSA operands.
1404 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
1405 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
1406 whether we should verify whether not throwing stmts have EH info.
1407 * graphite-scop-detection.c (create_sese_edges): Adjust.
1408 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
1409 * tree-eh.c (lower_try_finally_switch): Do not add the
1410 default case label twice.
1412 2014-04-30 Marek Polacek <polacek@redhat.com>
1414 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
1415 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
1416 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
1417 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
1419 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
1421 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
1422 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
1423 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
1424 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
1425 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
1426 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
1427 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
1428 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
1430 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1432 * tree-cfg.c (dump_function_to_file): Dump the return type of
1433 functions, in a line to itself before the function body, mimicking
1434 the layout of a C function.
1436 2014-04-29 Jakub Jelinek <jakub@redhat.com>
1438 PR tree-optimization/60971
1439 * tree-tailcall.c (process_assignment): Reject conversions which
1442 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
1444 * calls.c (initialize_argument_information): Always treat
1445 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
1446 (expand_call): Likewise.
1447 (emit_library_call_calue_1): Likewise.
1448 * expr.c (PUSH_ARGS_REVERSED): Do not define.
1449 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
1452 2014-04-29 Nick Clifton <nickc@redhat.com>
1454 * config/msp430/msp430.md (umulsidi): Fix typo.
1455 (mulhisi3): Enable even inside interrupt handlers.
1456 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
1457 bigger return address pushed in large mode.
1459 2014-04-29 Nick Clifton <nickc@redhat.com>
1461 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
1462 (arc_init_reg_tables): Use a machine_mode enum to iterate over
1464 * config/m32r/m32r.c (init_reg_tables): Likewise.
1465 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
1466 enum to hold the modes.
1468 2014-04-29 Richard Biener <rguenther@suse.de>
1470 * dominance.c (free_dominance_info): Add overload with
1472 (dom_info_state): Likewise.
1473 (dom_info_available_p): Likewise.
1474 * basic-block.h (free_dominance_info, dom_info_state,
1475 dom_info_available_p): Declare overloads.
1476 * passes.c (execute_function_todo): Verify that verifiers
1477 don't change dominator info state. Drop dominator info
1478 for IPA pass invocations.
1479 * cgraph.c (release_function_body): Restore asserts that
1480 dominator information is released.
1482 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
1484 * doc/invoke.texi: Fix typo.
1485 * tree-vrp.c: Fix typos.
1486 * gimple.c (infer_nonnull_range): Reorder operands of an &&
1489 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1491 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
1493 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1495 * config/aarch64/aarch64-builtins.c
1496 (aarch64_types_storestruct_lane_qualifiers): New.
1497 (TYPES_STORESTRUCT_LANE): Likewise.
1498 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
1499 (st3_lane): Likewise.
1500 (st4_lane): Likewise.
1501 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
1502 (vec_store_lanesci_lane<mode>): Likewise.
1503 (vec_store_lanesxi_lane<mode>): Likewise.
1504 (aarch64_st2_lane<VQ:mode>): Likewise.
1505 (aarch64_st3_lane<VQ:mode>): Likewise.
1506 (aarch64_st4_lane<VQ:mode>): Likewise.
1507 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
1508 * config/aarch64/arm_neon.h
1509 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
1510 use new macro arguments.
1511 (__ST3_LANE_FUNC): Likewise.
1512 (__ST4_LANE_FUNC): Likewise.
1513 * config/aarch64/iterators.md (V_TWO_ELEM): New.
1514 (V_THREE_ELEM): Likewise.
1515 (V_FOUR_ELEM): Likewise.
1517 2014-04-28 David Malcolm <dmalcolm@redhat.com>
1519 * doc/gimple.texi: Replace the description of the now-defunct
1520 union gimple_statement_d with a diagram showing the
1521 gimple_statement_base class hierarchy and its relationships to
1522 the GSS_ and GIMPLE_ enums.
1524 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1526 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
1527 * config/aarch64/aarch64.c
1528 (aarch64_cannot_change_mode_class): Weaken conditions.
1529 (aarch64_modes_tieable_p): New.
1530 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
1532 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
1534 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
1535 (loadsync_<mode>): Change mode.
1536 (load_quadpti, store_quadpti): New.
1537 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
1538 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
1540 2014-04-28 Martin Jambor <mjambor@suse.cz>
1542 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
1543 same alias type as the original statement.
1544 (subreplacement_assignment_data): New type.
1545 (handle_unscalarized_data_in_subtree): New type of parameter,
1546 generate new memory accesses with same alias type as the original
1548 (load_assign_lhs_subreplacements): Likewise.
1549 (sra_modify_constructor_assign): Generate new memory accesses with
1550 same alias type as the original statement.
1552 2014-04-28 Richard Biener <rguenther@suse.de>
1554 * tree-pass.h (TODO_verify_il): Define.
1555 (TODO_verify_all): Complete properly.
1556 * passes.c (execute_function_todo): Move existing loop-closed
1557 SSA verification under TODO_verify_il.
1558 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
1559 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
1560 Fix tree sharing issue.
1562 2014-04-28 Richard Biener <rguenther@suse.de>
1565 * builtins.def (DEF_C11_BUILTIN): Add.
1566 (BUILT_IN_ALIGNED_ALLOC): Likewise.
1567 * coretypes.h (enum function_class): Add function_c11_misc.
1568 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1569 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
1570 (call_may_clobber_ref_p_1): Likewise.
1571 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
1572 (mark_all_reaching_defs_necessary_1): Likewise.
1573 (propagate_necessity): Likewise.
1574 (eliminate_unnecessary_stmts): Likewise.
1575 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
1577 2014-04-28 Richard Biener <rguenther@suse.de>
1579 * tree-vrp.c (vrp_var_may_overflow): Remove.
1580 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
1581 with overflow immediately bump to one before that value and
1582 let iteration figure out overflow status.
1584 2014-04-28 Richard Biener <rguenther@suse.de>
1586 * configure.ac: Do valgrind header checks unconditionally.
1587 Add --enable-valgrind-annotations.
1588 * system.h: Guard valgrind header inclusion with
1589 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
1590 * alloc-pool.c (pool_alloc, pool_free): Use
1591 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
1592 to guard possibly dead code.
1593 * config.in: Regenerated.
1594 * configure: Likewise.
1596 2014-04-28 Jeff Law <law@redhat.com>
1598 PR tree-optimization/60902
1599 * tree-ssa-threadedge.c
1600 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
1601 over real defs when invalidating outputs from statements that do not
1602 produce useful outputs for threading.
1604 2014-04-28 Richard Biener <rguenther@suse.de>
1606 PR tree-optimization/60979
1607 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
1608 SCOPs that end in a block with a successor with abnormal
1611 2014-04-28 Richard Biener <rguenther@suse.de>
1613 * tree-pass.h (execute_pass_list): Adjust prototype.
1614 * passes.c (pass_manager::execute_early_local_passes):
1616 (do_per_function): Change callback signature, push all actual
1617 work to the callbals.
1618 (do_per_function_toporder): Likewise.
1619 (execute_function_dump): Adjust.
1620 (execute_function_todo): Likewise.
1621 (clear_last_verified): Likewise.
1622 (verify_curr_properties): Likewise.
1623 (update_properties_after_pass): Likewise.
1624 (execute_pass_list_1): Split out from ...
1625 (execute_pass_list): ... here. Adjust.
1626 (execute_ipa_pass_list): Likewise.
1627 * cgraphunit.c (cgraph_add_new_function): Adjust.
1628 (analyze_function): Likewise.
1629 (expand_function): Likewise.
1630 * cgraph.c (release_function_body): Free dominance info
1631 here instead of asserting it was magically freed elsewhere.
1633 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
1635 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
1636 * configure: Regenerate.
1637 * config/sparc/sparc.opt (muser-mode): New option.
1638 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
1640 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
1641 * doc/invoke.texi (SPARC options): Document -muser-mode.
1643 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
1645 * cselib.c (find_slot_memmode): Delete.
1646 (cselib_hasher): Change compare_type to a struct.
1647 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
1649 (preserve_constants_and_equivs): Adjust for new compare_type.
1650 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
1651 (wrap_constant): Delete.
1652 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
1654 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
1656 * doc/install.texi (Building with profile feedback): Remove
1659 2014-04-26 Tom de Vries <tom@codesourcery.com>
1661 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
1664 2014-04-25 Cary Coutant <ccoutant@google.com>
1667 * dwarf2out.c (should_move_die_to_comdat): A type definition
1668 can contain a subprogram definition, but don't move it to a
1670 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
1671 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
1673 (clone_tree_hash): Rename to...
1674 (clone_tree_partial): ...this; change callers. Copy
1675 DW_TAG_subprogram DIEs as declarations.
1676 (copy_decls_walk): Don't copy children of a declaration into a
1679 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
1682 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
1685 2014-04-25 Jiong Wang <jiong.wang@arm.com>
1687 * config/arm/predicates.md (call_insn_operand): Add long_call check.
1688 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
1690 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
1693 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1695 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
1698 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1700 PR tree-optimization/60930
1701 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
1702 creating a multiply candidate by folding two constant
1703 multiplicands when the result overflows.
1705 2014-04-25 Jakub Jelinek <jakub@redhat.com>
1707 PR tree-optimization/60960
1708 * tree-vect-generic.c (expand_vector_operation): Only call
1709 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
1711 2014-04-25 Tom de Vries <tom@codesourcery.com>
1713 * expr.c (clobber_reg_mode): New function.
1714 * expr.h (clobber_reg): New function.
1716 2014-04-25 Tom de Vries <tom@codesourcery.com>
1718 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
1721 2014-04-25 Radovan Obradovic <robradovic@mips.com>
1722 Tom de Vries <tom@codesourcery.com>
1724 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
1726 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
1727 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
1728 new argument to find_all_hard_reg_sets call.
1730 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1732 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
1733 Use HOST_WIDE_INT_C for mask literal.
1734 (aarch_rev16_shleft_mask_imm_p): Likewise.
1736 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
1739 * config/sparc/sparc.md (ashlsi3_extend): Delete.
1741 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
1743 PR preprocessor/56540
1744 * config/i386/i386-c.c (ix86_target_macros): Define
1745 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
1747 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1749 * configure.ac (tga_func): Remove.
1750 (LIB_TLS_SPEC): Remove.
1751 * configure: Regenerate.
1752 * config.in: Regenerate.
1753 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
1755 2014-04-25 Richard Biener <rguenther@suse.de>
1758 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
1759 call stmt use/clobber sets during stmt walk instead of
1760 walking the possibly incomplete set of caller edges.
1762 2014-04-25 Richard Biener <rguenther@suse.de>
1765 * passes.c (apply_ipa_transforms): Inline into only caller ...
1766 (execute_one_pass): ... here. Properly bring in function
1767 bodies for nodes we want to apply IPA transforms to.
1769 2014-04-24 Cong Hou <congh@google.com>
1771 PR tree-optimization/60896
1772 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
1773 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
1774 (vect_mark_pattern_stmts): Set the def type of all statements in
1775 PATTERN_DEF_SEQ as vect_internal_def.
1777 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
1779 * doc/extend.texi (PowerPC Built-in Functions): Document new
1780 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
1782 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
1784 * config/rs6000/predicates.md (const_0_to_3_operand): New
1785 predicate to match 0..3 integer constants.
1787 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
1788 to support adding miscellaneous builtin functions.
1789 (BU_DFP_MISC_2): Likewise.
1790 (BU_P7_MISC_1): Likewise.
1791 (BU_P7_MISC_2): Likewise.
1792 (BU_P8V_MISC_3): Likewise.
1793 (BU_MISC_1): Likewise.
1794 (BU_MISC_2): Likewise.
1795 (DIVWE): Add extended divide builtin functions.
1798 (DIVWEUO): Likewise.
1802 (DIVDEUO): Likewise.
1803 (DXEX): Add decimal floating-point builtin functions.
1806 (DDEDPDQ): Likewise.
1808 (DENBCDQ): Likewise.
1815 (CDTBCD): Add new BCD builtin functions.
1819 (BCDADD_LT): Likewise.
1820 (BCDADD_EQ): Likewise.
1821 (BCDADD_GT): Likewise.
1822 (BCDADD_OV): Likewise.
1824 (BCDSUB_LT): Likewise.
1825 (BCDSUB_EQ): Likewise.
1826 (BCDSUB_GT): Likewise.
1827 (BCDSUB_OV): Likewise.
1828 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
1829 (UNPACK_TD): Likewise.
1830 (PACK_TF): Likewise.
1831 (UNPACK_TF): Likewise.
1832 (UNPACK_TF_0): Likewise.
1833 (UNPACK_TF_1): Likewise.
1834 (PACK_V1TI): Likewise.
1835 (UNPACK_V1TI): Likewise.
1837 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
1838 support for decimal floating point builtin functions.
1839 (rs6000_expand_ternop_builtin): Add checks for the new builtin
1840 functions that take constant arguments.
1841 (rs6000_invalid_builtin): Add decimal floating point builtin
1843 (rs6000_init_builtins): Setup long double, _Decimal64, and
1844 _Decimal128 types for new builtin functions.
1845 (builtin_function_type): Set the unsigned flags appropriately for
1846 the new builtin functions.
1847 (rs6000_opt_masks): Add support for decimal floating point builtin
1850 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
1851 floating point builtin functions.
1852 (RS6000_BTM_COMMON): Likewise.
1853 (RS6000_BTI_long_double): Likewise.
1854 (RS6000_BTI_dfloat64): Likewise.
1855 (RS6000_BTI_dfloat128): Likewise.
1856 (long_double_type_internal_node): Likewise.
1857 (dfloat64_type_internal_node): Likewise.
1858 (dfloat128_type_internal_node): Likewise.
1860 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
1861 2.07 bcd arithmetic instructions.
1862 (UNSPEC_BCDSUB): Likewise.
1863 (UNSPEC_BCD_OVERFLOW): Likewise.
1864 (UNSPEC_BCD_ADD_SUB): Likewise.
1865 (bcd_add_sub): Likewise.
1866 (BCD_TEST): Likewise.
1867 (bcd<bcd_add_sub>): Likewise.
1868 (bcd<bcd_add_sub>_test): Likewise.
1869 (bcd<bcd_add_sub>_test2): Likewise.
1870 (bcd<bcd_add_sub>_<code>): Likewise.
1871 (peephole2 for combined bcd ops): Likewise.
1873 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
1874 decimal floating point builtin functions.
1875 (UNSPEC_DENBCD): Likewise.
1876 (UNSPEC_DXEX): Likewise.
1877 (UNSPEC_DIEX): Likewise.
1878 (UNSPEC_DSCLI): Likewise.
1879 (UNSPEC_DSCRI): Likewise.
1880 (D64_D128): Likewise.
1881 (dfp_suffix): Likewise.
1882 (dfp_ddedpd_<mode>): Likewise.
1883 (dfp_denbcd_<mode>): Likewise.
1884 (dfp_dxex_<mode>): Likewise.
1885 (dfp_diex_<mode>): Likewise.
1886 (dfp_dscli_<mode>): Likewise.
1887 (dfp_dscri_<mode>): Likewise.
1889 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
1891 (UNSPEC_CDTBCD): Likewise.
1892 (UNSPEC_CBCDTD): Likewise.
1893 (UNSPEC_DIVE): Add support for new extended divide builtin
1895 (UNSPEC_DIVEO): Likewise.
1896 (UNSPEC_DIVEU): Likewise.
1897 (UNSPEC_DIVEUO): Likewise.
1898 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
1899 pack/unpack 128-bit types.
1900 (UNSPEC_PACK_128BIT): Likewise.
1901 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
1902 (udiv<mode>3): Use idiv_ldiv mode attribute.
1903 (div<mode>3): Likewise.
1904 (addg6s): Add new BCD builtin functions.
1907 (UNSPEC_DIV_EXTEND): Add support for new extended divide
1909 (div_extend): Likewise.
1910 (div<div_extend>_<mode>"): Likewise.
1911 (FP128_64): Add support for new builtin functions to pack/unpack
1913 (unpack<mode>): Likewise.
1914 (unpacktf_0): Likewise.
1915 (unpacktf_1): Likewise.
1916 (unpack<mode>_dm): Likewise.
1917 (unpack<mode>_nodm): Likewise.
1918 (pack<mode>): Likewise.
1919 (unpackv1ti): Likewise.
1920 (packv1ti): Likewise.
1922 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
1924 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
1927 2014-04-24 Jakub Jelinek <jakub@redhat.com>
1929 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
1930 * gimplify.c (omp_is_private): Change last argument's type to int.
1931 Only diagnose lastprivate if the simd argument is 1, only diagnose
1932 linear if the simd argument is 2.
1933 (gimplify_omp_for): Adjust omp_is_private callers. When adding
1934 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
1935 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
1936 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
1937 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
1938 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
1939 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
1940 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
1941 * tree-nested.c (convert_nonlocal_omp_clauses,
1942 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
1944 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
1947 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
1950 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
1952 * flag-types.h (enum ivar_visibility): Add.
1954 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
1956 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
1957 function * argument.
1959 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
1961 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
1963 2014-04-24 Radovan Obradovic <robradovic@mips.com>
1964 Tom de Vries <tom@codesourcery.com>
1966 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
1967 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
1969 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
1970 * emit-rtl.c (try_split): Same.
1972 2014-04-24 Radovan Obradovic <robradovic@mips.com>
1973 Tom de Vries <tom@codesourcery.com>
1975 * common.opt (fuse-caller-save): New option.
1977 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
1979 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
1982 2014-04-24 Richard Biener <rguenther@suse.de>
1984 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
1985 during TER and instead use the sepops interface for expanding
1986 non-GIMPLE_SINGLE_RHS.
1988 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1990 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
1991 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
1993 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1995 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
1996 assembler 64-bit option.
1997 * configure: Regenerate.
1999 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2001 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
2002 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
2003 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
2004 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
2005 (TARGET_CRYPTO): Take TARGET_SIMD into account.
2007 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2009 * config/aarch64/aarch64-builtins.c
2010 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
2011 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
2012 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
2013 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
2015 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
2016 (Vrevsuff): New mode attribute.
2018 2014-04-24 Terry Guo <terry.guo@arm.com>
2020 * config/arm/arm.h (machine_function): Define variable
2021 after_arm_reorg here.
2022 * config/arm/arm.c (after_arm_reorg): Remove the definition.
2023 (arm_split_constant): Update the way to access variable
2026 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
2028 2014-04-23 Tom de Vries <tom@codesourcery.com>
2030 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
2032 2014-04-23 David Malcolm <dmalcolm@redhat.com>
2034 * is-a.h: Update comments to reflect the following changes to the
2035 "pointerness" of the API, making the template parameter match the
2036 return type, allowing use of is-a.h with typedefs of pointers.
2037 (is_a_helper::cast): Return a T rather then a pointer to a T, so
2038 that the return type matches the parameter to the is_a_helper.
2040 (dyn_cast): Likewise.
2042 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
2043 pointer from the is-a.h API.
2045 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
2046 (is_a_helper <cgraph_node *>::test): ...this, matching change to
2048 (is_a_helper <varpool_node>::test): Likewise, convert to...
2049 (is_a_helper <varpool_node *>::test): ...this.
2051 (varpool_first_variable): Update for removal of implicit pointer
2052 from the is-a.h API.
2053 (varpool_next_variable): Likewise.
2054 (varpool_first_static_initializer): Likewise.
2055 (varpool_next_static_initializer): Likewise.
2056 (varpool_first_defined_variable): Likewise.
2057 (varpool_next_defined_variable): Likewise.
2058 (cgraph_first_defined_function): Likewise.
2059 (cgraph_next_defined_function): Likewise.
2060 (cgraph_first_function): Likewise.
2061 (cgraph_next_function): Likewise.
2062 (cgraph_first_function_with_gimple_body): Likewise.
2063 (cgraph_next_function_with_gimple_body): Likewise.
2064 (cgraph_alias_target): Likewise.
2065 (varpool_alias_target): Likewise.
2066 (cgraph_function_or_thunk_node): Likewise.
2067 (varpool_variable_node): Likewise.
2068 (symtab_real_symbol_p): Likewise.
2069 * cgraphunit.c (referred_to_p): Likewise.
2070 (analyze_functions): Likewise.
2071 (handle_alias_pairs): Likewise.
2072 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2073 * gimple-ssa.h (gimple_vuse_op): Likewise.
2074 (gimple_vdef_op): Likewise.
2075 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
2076 * gimple.c (gimple_build_asm_1): Likewise.
2077 (gimple_build_try): Likewise.
2078 (gimple_build_resx): Likewise.
2079 (gimple_build_eh_dispatch): Likewise.
2080 (gimple_build_omp_for): Likewise.
2081 (gimple_omp_for_set_clauses): Likewise.
2083 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
2084 (is_a_helper <gimple_statement_asm *>::test): ...this.
2085 (is_a_helper <gimple_statement_bind>::test): Convert to...
2086 (is_a_helper <gimple_statement_bind *>::test): ...this.
2087 (is_a_helper <gimple_statement_call>::test): Convert to...
2088 (is_a_helper <gimple_statement_call *>::test): ...this.
2089 (is_a_helper <gimple_statement_catch>::test): Convert to...
2090 (is_a_helper <gimple_statement_catch *>::test): ...this.
2091 (is_a_helper <gimple_statement_resx>::test): Convert to...
2092 (is_a_helper <gimple_statement_resx *>::test): ...this.
2093 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
2094 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
2095 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
2096 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
2097 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
2098 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
2099 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
2100 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
2101 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
2102 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
2103 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
2104 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
2105 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
2106 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
2107 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
2108 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
2109 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
2110 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
2111 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
2112 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
2113 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
2114 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
2115 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
2116 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
2117 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
2118 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
2119 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
2120 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
2121 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
2122 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
2123 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
2124 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
2125 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
2126 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
2127 (is_a_helper <gimple_statement_phi>::test): Convert to...
2128 (is_a_helper <gimple_statement_phi *>::test): ...this.
2129 (is_a_helper <gimple_statement_transaction>::test): Convert to...
2130 (is_a_helper <gimple_statement_transaction *>::test): ...this.
2131 (is_a_helper <gimple_statement_try>::test): Convert to...
2132 (is_a_helper <gimple_statement_try *>::test): ...this.
2133 (is_a_helper <gimple_statement_wce>::test): Convert to...
2134 (is_a_helper <gimple_statement_wce *>::test): ...this.
2135 (is_a_helper <const gimple_statement_asm>::test): Convert to...
2136 (is_a_helper <const gimple_statement_asm *>::test): ...this.
2137 (is_a_helper <const gimple_statement_bind>::test): Convert to...
2138 (is_a_helper <const gimple_statement_bind *>::test): ...this.
2139 (is_a_helper <const gimple_statement_call>::test): Convert to...
2140 (is_a_helper <const gimple_statement_call *>::test): ...this.
2141 (is_a_helper <const gimple_statement_catch>::test): Convert to...
2142 (is_a_helper <const gimple_statement_catch *>::test): ...this.
2143 (is_a_helper <const gimple_statement_resx>::test): Convert to...
2144 (is_a_helper <const gimple_statement_resx *>::test): ...this.
2145 (is_a_helper <const gimple_statement_eh_dispatch>::test):
2147 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
2149 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
2151 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
2152 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
2154 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
2156 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
2158 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
2160 (is_a_helper <const gimple_statement_omp_return>::test): Convert
2162 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
2163 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
2165 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
2166 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
2168 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
2169 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
2170 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
2171 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
2173 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
2174 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
2176 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
2177 (is_a_helper <const gimple_statement_omp_target>::test): Convert
2179 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
2180 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
2182 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
2183 (is_a_helper <const gimple_statement_omp_single>::test): Convert
2185 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
2186 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
2188 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
2189 (is_a_helper <const gimple_statement_omp_task>::test): Convert
2191 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
2192 (is_a_helper <const gimple_statement_phi>::test): Convert to...
2193 (is_a_helper <const gimple_statement_phi *>::test): ...this.
2194 (is_a_helper <const gimple_statement_transaction>::test): Convert
2196 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
2197 (is_a_helper <const gimple_statement_with_ops>::test): Convert
2199 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
2200 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
2201 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
2202 (is_a_helper <const gimple_statement_with_memory_ops>::test):
2204 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
2206 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
2208 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
2210 (gimple_use_ops): Update for removal of implicit pointer from the
2212 (gimple_set_use_ops): Likewise.
2213 (gimple_vuse): Likewise.
2214 (gimple_vdef): Likewise.
2215 (gimple_vuse_ptr): Likewise.
2216 (gimple_vdef_ptr): Likewise.
2217 (gimple_set_vuse): Likewise.
2218 (gimple_set_vdef): Likewise.
2219 (gimple_omp_return_set_lhs): Likewise.
2220 (gimple_omp_return_lhs): Likewise.
2221 (gimple_omp_return_lhs_ptr): Likewise.
2222 (gimple_call_fntype): Likewise.
2223 (gimple_call_set_fntype): Likewise.
2224 (gimple_call_set_internal_fn): Likewise.
2225 (gimple_call_use_set): Likewise.
2226 (gimple_call_clobber_set): Likewise.
2227 (gimple_bind_vars): Likewise.
2228 (gimple_bind_set_vars): Likewise.
2229 (gimple_bind_body_ptr): Likewise.
2230 (gimple_bind_set_body): Likewise.
2231 (gimple_bind_add_stmt): Likewise.
2232 (gimple_bind_block): Likewise.
2233 (gimple_bind_set_block): Likewise.
2234 (gimple_asm_ninputs): Likewise.
2235 (gimple_asm_noutputs): Likewise.
2236 (gimple_asm_nclobbers): Likewise.
2237 (gimple_asm_nlabels): Likewise.
2238 (gimple_asm_input_op): Likewise.
2239 (gimple_asm_input_op_ptr): Likewise.
2240 (gimple_asm_output_op): Likewise.
2241 (gimple_asm_output_op_ptr): Likewise.
2242 (gimple_asm_set_output_op): Likewise.
2243 (gimple_asm_clobber_op): Likewise.
2244 (gimple_asm_set_clobber_op): Likewise.
2245 (gimple_asm_label_op): Likewise.
2246 (gimple_asm_set_label_op): Likewise.
2247 (gimple_asm_string): Likewise.
2248 (gimple_catch_types): Likewise.
2249 (gimple_catch_types_ptr): Likewise.
2250 (gimple_catch_handler_ptr): Likewise.
2251 (gimple_catch_set_types): Likewise.
2252 (gimple_catch_set_handler): Likewise.
2253 (gimple_eh_filter_types): Likewise.
2254 (gimple_eh_filter_types_ptr): Likewise.
2255 (gimple_eh_filter_failure_ptr): Likewise.
2256 (gimple_eh_filter_set_types): Likewise.
2257 (gimple_eh_filter_set_failure): Likewise.
2258 (gimple_eh_must_not_throw_fndecl): Likewise.
2259 (gimple_eh_must_not_throw_set_fndecl): Likewise.
2260 (gimple_eh_else_n_body_ptr): Likewise.
2261 (gimple_eh_else_e_body_ptr): Likewise.
2262 (gimple_eh_else_set_n_body): Likewise.
2263 (gimple_eh_else_set_e_body): Likewise.
2264 (gimple_try_eval_ptr): Likewise.
2265 (gimple_try_cleanup_ptr): Likewise.
2266 (gimple_try_set_eval): Likewise.
2267 (gimple_try_set_cleanup): Likewise.
2268 (gimple_wce_cleanup_ptr): Likewise.
2269 (gimple_wce_set_cleanup): Likewise.
2270 (gimple_phi_capacity): Likewise.
2271 (gimple_phi_num_args): Likewise.
2272 (gimple_phi_result): Likewise.
2273 (gimple_phi_result_ptr): Likewise.
2274 (gimple_phi_set_result): Likewise.
2275 (gimple_phi_arg): Likewise.
2276 (gimple_phi_set_arg): Likewise.
2277 (gimple_resx_region): Likewise.
2278 (gimple_resx_set_region): Likewise.
2279 (gimple_eh_dispatch_region): Likewise.
2280 (gimple_eh_dispatch_set_region): Likewise.
2281 (gimple_omp_critical_name): Likewise.
2282 (gimple_omp_critical_name_ptr): Likewise.
2283 (gimple_omp_critical_set_name): Likewise.
2284 (gimple_omp_for_clauses): Likewise.
2285 (gimple_omp_for_clauses_ptr): Likewise.
2286 (gimple_omp_for_set_clauses): Likewise.
2287 (gimple_omp_for_collapse): Likewise.
2288 (gimple_omp_for_index): Likewise.
2289 (gimple_omp_for_index_ptr): Likewise.
2290 (gimple_omp_for_set_index): Likewise.
2291 (gimple_omp_for_initial): Likewise.
2292 (gimple_omp_for_initial_ptr): Likewise.
2293 (gimple_omp_for_set_initial): Likewise.
2294 (gimple_omp_for_final): Likewise.
2295 (gimple_omp_for_final_ptr): Likewise.
2296 (gimple_omp_for_set_final): Likewise.
2297 (gimple_omp_for_incr): Likewise.
2298 (gimple_omp_for_incr_ptr): Likewise.
2299 (gimple_omp_for_set_incr): Likewise.
2300 (gimple_omp_for_pre_body_ptr): Likewise.
2301 (gimple_omp_for_set_pre_body): Likewise.
2302 (gimple_omp_parallel_clauses): Likewise.
2303 (gimple_omp_parallel_clauses_ptr): Likewise.
2304 (gimple_omp_parallel_set_clauses): Likewise.
2305 (gimple_omp_parallel_child_fn): Likewise.
2306 (gimple_omp_parallel_child_fn_ptr): Likewise.
2307 (gimple_omp_parallel_set_child_fn): Likewise.
2308 (gimple_omp_parallel_data_arg): Likewise.
2309 (gimple_omp_parallel_data_arg_ptr): Likewise.
2310 (gimple_omp_parallel_set_data_arg): Likewise.
2311 (gimple_omp_task_clauses): Likewise.
2312 (gimple_omp_task_clauses_ptr): Likewise.
2313 (gimple_omp_task_set_clauses): Likewise.
2314 (gimple_omp_task_child_fn): Likewise.
2315 (gimple_omp_task_child_fn_ptr): Likewise.
2316 (gimple_omp_task_set_child_fn): Likewise.
2317 (gimple_omp_task_data_arg): Likewise.
2318 (gimple_omp_task_data_arg_ptr): Likewise.
2319 (gimple_omp_task_set_data_arg): Likewise.
2320 (gimple_omp_taskreg_clauses): Likewise.
2321 (gimple_omp_taskreg_clauses_ptr): Likewise.
2322 (gimple_omp_taskreg_set_clauses): Likewise.
2323 (gimple_omp_taskreg_child_fn): Likewise.
2324 (gimple_omp_taskreg_child_fn_ptr): Likewise.
2325 (gimple_omp_taskreg_set_child_fn): Likewise.
2326 (gimple_omp_taskreg_data_arg): Likewise.
2327 (gimple_omp_taskreg_data_arg_ptr): Likewise.
2328 (gimple_omp_taskreg_set_data_arg): Likewise.
2329 (gimple_omp_task_copy_fn): Likewise.
2330 (gimple_omp_task_copy_fn_ptr): Likewise.
2331 (gimple_omp_task_set_copy_fn): Likewise.
2332 (gimple_omp_task_arg_size): Likewise.
2333 (gimple_omp_task_arg_size_ptr): Likewise.
2334 (gimple_omp_task_set_arg_size): Likewise.
2335 (gimple_omp_task_arg_align): Likewise.
2336 (gimple_omp_task_arg_align_ptr): Likewise.
2337 (gimple_omp_task_set_arg_align): Likewise.
2338 (gimple_omp_single_clauses): Likewise.
2339 (gimple_omp_single_clauses_ptr): Likewise.
2340 (gimple_omp_single_set_clauses): Likewise.
2341 (gimple_omp_target_clauses): Likewise.
2342 (gimple_omp_target_clauses_ptr): Likewise.
2343 (gimple_omp_target_set_clauses): Likewise.
2344 (gimple_omp_target_child_fn): Likewise.
2345 (gimple_omp_target_child_fn_ptr): Likewise.
2346 (gimple_omp_target_set_child_fn): Likewise.
2347 (gimple_omp_target_data_arg): Likewise.
2348 (gimple_omp_target_data_arg_ptr): Likewise.
2349 (gimple_omp_target_set_data_arg): Likewise.
2350 (gimple_omp_teams_clauses): Likewise.
2351 (gimple_omp_teams_clauses_ptr): Likewise.
2352 (gimple_omp_teams_set_clauses): Likewise.
2353 (gimple_omp_sections_clauses): Likewise.
2354 (gimple_omp_sections_clauses_ptr): Likewise.
2355 (gimple_omp_sections_set_clauses): Likewise.
2356 (gimple_omp_sections_control): Likewise.
2357 (gimple_omp_sections_control_ptr): Likewise.
2358 (gimple_omp_sections_set_control): Likewise.
2359 (gimple_omp_for_set_cond): Likewise.
2360 (gimple_omp_for_cond): Likewise.
2361 (gimple_omp_atomic_store_set_val): Likewise.
2362 (gimple_omp_atomic_store_val): Likewise.
2363 (gimple_omp_atomic_store_val_ptr): Likewise.
2364 (gimple_omp_atomic_load_set_lhs): Likewise.
2365 (gimple_omp_atomic_load_lhs): Likewise.
2366 (gimple_omp_atomic_load_lhs_ptr): Likewise.
2367 (gimple_omp_atomic_load_set_rhs): Likewise.
2368 (gimple_omp_atomic_load_rhs): Likewise.
2369 (gimple_omp_atomic_load_rhs_ptr): Likewise.
2370 (gimple_omp_continue_control_def): Likewise.
2371 (gimple_omp_continue_control_def_ptr): Likewise.
2372 (gimple_omp_continue_set_control_def): Likewise.
2373 (gimple_omp_continue_control_use): Likewise.
2374 (gimple_omp_continue_control_use_ptr): Likewise.
2375 (gimple_omp_continue_set_control_use): Likewise.
2376 (gimple_transaction_body_ptr): Likewise.
2377 (gimple_transaction_label): Likewise.
2378 (gimple_transaction_label_ptr): Likewise.
2379 (gimple_transaction_set_body): Likewise.
2380 (gimple_transaction_set_label): Likewise.
2382 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
2383 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2384 * ipa-ref.c (ipa_record_reference): Likewise.
2385 * ipa-reference.c (analyze_function): Likewise.
2386 (ipa_reference_write_optimization_summary): Likewise.
2387 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
2388 (address_taken_from_non_vtable_p): Likewise.
2389 (comdat_can_be_unshared_p_1): Likewise.
2390 * lto-cgraph.c (lto_output_ref): Likewise.
2391 (add_references): Likewise.
2392 (compute_ltrans_boundary): Likewise.
2393 (output_symtab): Likewise.
2394 (input_ref): Likewise.
2395 (input_cgraph_1): Likewise.
2396 (output_cgraph_opt_summary): Likewise.
2397 * lto-streamer-out.c (lto_output): Likewise.
2398 (output_symbol_p): Likewise.
2399 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
2400 (lsei_start_function_in_partition): Likewise.
2401 (lsei_next_variable_in_partition): Likewise.
2402 (lsei_start_variable_in_partition): Likewise.
2403 * symtab.c (insert_to_assembler_name_hash): Likewise.
2404 (unlink_from_assembler_name_hash): Likewise.
2405 (symtab_unregister_node): Likewise.
2406 (symtab_remove_node): Likewise.
2407 (dump_symtab_node): Likewise.
2408 (verify_symtab_base): Likewise.
2409 (verify_symtab_node): Likewise.
2410 (symtab_make_decl_local): Likewise.
2411 (symtab_alias_ultimate_target): Likewise.
2412 (symtab_resolve_alias): Likewise.
2413 (symtab_get_symbol_partitioning_class): Likewise.
2414 * tree-phinodes.c (allocate_phi_node): Likewise.
2415 (reserve_phi_args_for_new_edge): Likewise.
2416 (remove_phi_args): Likewise.
2417 * varpool.c (varpool_node_for_asm): Likewise.
2418 (varpool_remove_unreferenced_decls): Likewise.
2420 2014-04-23 Jeff Law <law@redhat.com>
2422 PR tree-optimization/60902
2423 * tree-ssa-threadedge.c
2424 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
2425 invalidate outputs from statements that do not produce useful
2426 outputs for threading.
2428 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2430 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
2431 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
2432 machine descriptions for Stack Smashing Protector.
2434 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
2436 * aarch64.md (<optab>_rol<mode>3): New pattern.
2437 (<optab>_rolsi3_uxtw): Likewise.
2438 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
2440 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
2442 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
2443 (arm_cortex_a12_tune): Likewise.
2445 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2447 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
2449 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2451 * config/arm/arm.md (arm_rev16si2): New pattern.
2452 (arm_rev16si2_alt): Likewise.
2453 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
2455 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2457 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
2458 (rev16<mode>2_alt): Likewise.
2459 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
2460 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
2461 (aarch_rev16_shleft_mask_imm_p): Likewise.
2462 (aarch_rev16_p_1): Likewise.
2463 (aarch_rev16_p): Likewise.
2464 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
2465 (aarch_rev16_shright_mask_imm_p): Likewise.
2466 (aarch_rev16_shleft_mask_imm_p): Likewise.
2468 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2470 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
2471 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
2473 (cortex_a53_extra_costs): Likewise.
2474 (cortex_a57_extra_costs): Likewise.
2475 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
2476 (cortexa7_extra_costs): Likewise.
2477 (cortexa8_extra_costs): Likewise.
2478 (cortexa12_extra_costs): Likewise.
2479 (cortexa15_extra_costs): Likewise.
2480 (v7m_extra_costs): Likewise.
2481 (arm_new_rtx_costs): Handle BSWAP.
2483 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2485 * config/arm/arm.c (cortexa8_extra_costs): New table.
2486 (arm_cortex_a8_tune): New tuning struct.
2487 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
2489 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2491 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
2493 2014-04-23 Richard Biener <rguenther@suse.de>
2495 * Makefile.in (OBJS): Remove loop-unswitch.o.
2496 * tree-pass.h (make_pass_rtl_unswitch): Remove.
2497 * passes.def (pass_rtl_unswitch): Likewise.
2498 * loop-init.c (gate_rtl_unswitch): Likewise.
2499 (rtl_unswitch): Likewise.
2500 (pass_data_rtl_unswitch): Likewise.
2501 (pass_rtl_unswitch): Likewise.
2502 (make_pass_rtl_unswitch): Likewise.
2503 * rtl.h (reversed_condition): Likewise.
2504 (compare_and_jump_seq): Likewise.
2505 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
2507 * loop-unroll.c (compare_and_jump_seq): Likewise.
2509 2014-04-23 Richard Biener <rguenther@suse.de>
2511 PR tree-optimization/60903
2512 * tree-ssa-loop-im.c (analyze_memory_references): Remove
2513 commented code block.
2514 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
2515 loop flags to newly created BBs and edges.
2517 2014-04-23 Nick Clifton <nickc@redhat.com>
2519 * config/msp430/msp430.c (msp430_handle_option): Move function
2521 (msp430_option_override): Simplify mcu and mcpu option handling.
2522 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
2523 support for -mhwmult command line option.
2524 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
2525 -mhwmult command line option.
2526 (msp430_hwmult_enabled): Delete.
2527 (msp43o_output_labelref): Add support for -mhwmult command line
2529 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
2530 (umulsidi3): Likewise.
2531 * config/msp430/msp430.opt (mmcu): Add Report attribute.
2532 (mcpu, mlarge, msmall): Likewise.
2533 (mhwmult): New option.
2534 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
2536 (msp430_is_f5_mcu): Remove prototype.
2537 (msp430_use_f5_series_hwmult): Add prototype.
2538 * config/msp430/msp430-opts.h: New file.
2539 * common/config/msp430: New directory.
2540 * common/config/msp430/msp430-common.c: New file.
2541 * config.gcc (msp430): Remove target_has_targetm_common.
2542 * doc/invoke.texi: Document -mhwmult command line option.
2544 2014-04-23 Nick Clifton <nickc@redhat.com>
2546 * config/i386/cygwin.h (ENDFILE_SPEC): Include
2547 default-manifest.o if it can be found in the search path.
2548 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
2550 2014-04-23 Terry Guo <terry.guo@arm.com>
2552 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
2554 2014-04-23 Richard Biener <rguenther@suse.de>
2557 * tree-inline.c (declare_return_variable): Use mark_addressable.
2559 2014-04-23 Richard Biener <rguenther@suse.de>
2562 * loop-init.c (loop_optimizer_init): Make sure to apply
2563 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
2565 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2568 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
2570 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
2571 if flag_sanitize_undefined_trap_on_error.
2572 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
2573 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
2574 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
2575 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
2576 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
2577 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
2578 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
2579 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
2580 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
2581 * ubsan.c (ubsan_instrument_unreachable): Return
2582 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
2583 (ubsan_expand_null_ifn): Emit __builtin_trap ()
2584 if flag_sanitize_undefined_trap_on_error and
2585 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
2586 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
2587 instrument_bool_enum_load): Emit __builtin_trap () if
2588 flag_sanitize_undefined_trap_on_error and
2589 __builtin_handle_*_abort () if !flag_sanitize_recover.
2590 * doc/invoke.texi (-fsanitize-recover,
2591 -fsanitize-undefined-trap-on-error): Document.
2593 2014-04-22 Christian Bruel <christian.bruel@st.com>
2595 * config/sh/sh.md (mov<mode>): Replace movQIHI.
2596 Force immediates to SImode.
2598 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
2600 * config/nios2/nios2.md (UNSPEC_ROUND): New.
2602 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
2603 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
2604 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
2605 (nios2_fpu_insn): Add entry for round.
2606 (N2FPU_NO_ERRNO_P): Define.
2607 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
2609 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
2611 2014-04-22 Richard Henderson <rth@redhat.com>
2613 * config/aarch64/aarch64 (addti3, subti3): New expanders.
2614 (add<GPI>3_compare0): Remove leading * from name.
2615 (add<GPI>3_carryin): Likewise.
2616 (sub<GPI>3_compare0): Likewise.
2617 (sub<GPI>3_carryin): Likewise.
2618 (<su_optab>mulditi3): New expander.
2619 (multi3): New expander.
2620 (madd<GPI>): Remove leading * from name.
2622 2014-04-22 Martin Jambor <mjambor@suse.cz>
2624 * cgraphclones.c (cgraph_function_versioning): Copy
2625 ipa_transforms_to_apply instead of asserting it is empty.
2627 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
2630 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
2631 on count_exp to get mode.
2633 2014-04-22 Andrew Pinski <apinski@cavium.com>
2635 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
2636 Handle TLS for ILP32.
2637 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
2638 (tlsie_small_<mode>): this and handle PTR.
2639 (tlsie_small_sidi): New pattern.
2640 (tlsle_small): Change to an expand to handle ILP32.
2641 (tlsle_small_<mode>): New pattern.
2642 (tlsdesc_small): Rename to ...
2643 (tlsdesc_small_<mode>): this and handle PTR.
2645 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2647 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
2649 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2651 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2652 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
2653 (aarch64_types_signed_poly_qualifiers): Likewise.
2654 (aarch64_types_unsigned_signed_qualifiers): Likewise.
2655 (aarch64_types_poly_signed_qualifiers): Likewise.
2656 (TYPES_REINTERP_SS): Type macro added.
2657 (TYPES_REINTERP_SU): Likewise.
2658 (TYPES_REINTERP_SP): Likewise.
2659 (TYPES_REINTERP_US): Likewise.
2660 (TYPES_REINTERP_PS): Likewise.
2661 (aarch64_fold_builtin): New expression folding added.
2662 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
2663 Declarations removed.
2664 (REINTERP_SS): Declarations added.
2665 (REINTERP_US): Likewise.
2666 (REINTERP_PS): Likewise.
2667 (REINTERP_SU): Likewise.
2668 (REINTERP_SP): Likewise.
2669 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
2670 (vreinterpretq_p8_f64): Likewise.
2671 (vreinterpret_p16_f64): Likewise.
2672 (vreinterpretq_p16_f64): Likewise.
2673 (vreinterpret_f32_f64): Likewise.
2674 (vreinterpretq_f32_f64): Likewise.
2675 (vreinterpret_f64_f32): Likewise.
2676 (vreinterpret_f64_p8): Likewise.
2677 (vreinterpret_f64_p16): Likewise.
2678 (vreinterpret_f64_s8): Likewise.
2679 (vreinterpret_f64_s16): Likewise.
2680 (vreinterpret_f64_s32): Likewise.
2681 (vreinterpret_f64_s64): Likewise.
2682 (vreinterpret_f64_u8): Likewise.
2683 (vreinterpret_f64_u16): Likewise.
2684 (vreinterpret_f64_u32): Likewise.
2685 (vreinterpret_f64_u64): Likewise.
2686 (vreinterpretq_f64_f32): Likewise.
2687 (vreinterpretq_f64_p8): Likewise.
2688 (vreinterpretq_f64_p16): Likewise.
2689 (vreinterpretq_f64_s8): Likewise.
2690 (vreinterpretq_f64_s16): Likewise.
2691 (vreinterpretq_f64_s32): Likewise.
2692 (vreinterpretq_f64_s64): Likewise.
2693 (vreinterpretq_f64_u8): Likewise.
2694 (vreinterpretq_f64_u16): Likewise.
2695 (vreinterpretq_f64_u32): Likewise.
2696 (vreinterpretq_f64_u64): Likewise.
2697 (vreinterpret_s64_f64): Likewise.
2698 (vreinterpretq_s64_f64): Likewise.
2699 (vreinterpret_u64_f64): Likewise.
2700 (vreinterpretq_u64_f64): Likewise.
2701 (vreinterpret_s8_f64): Likewise.
2702 (vreinterpretq_s8_f64): Likewise.
2703 (vreinterpret_s16_f64): Likewise.
2704 (vreinterpretq_s16_f64): Likewise.
2705 (vreinterpret_s32_f64): Likewise.
2706 (vreinterpretq_s32_f64): Likewise.
2707 (vreinterpret_u8_f64): Likewise.
2708 (vreinterpretq_u8_f64): Likewise.
2709 (vreinterpret_u16_f64): Likewise.
2710 (vreinterpretq_u16_f64): Likewise.
2711 (vreinterpret_u32_f64): Likewise.
2712 (vreinterpretq_u32_f64): Likewise.
2714 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2716 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2717 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
2718 (vreinterpret_p8_s8): Likewise.
2719 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
2720 (vreinterpret_p8_s16): Likewise.
2721 (vreinterpret_p8_s32): Likewise.
2722 (vreinterpret_p8_s64): Likewise.
2723 (vreinterpret_p8_f32): Likewise.
2724 (vreinterpret_p8_u8): Likewise.
2725 (vreinterpret_p8_u16): Likewise.
2726 (vreinterpret_p8_u32): Likewise.
2727 (vreinterpret_p8_u64): Likewise.
2728 (vreinterpret_p8_p16): Likewise.
2729 (vreinterpretq_p8_s8): Likewise.
2730 (vreinterpretq_p8_s16): Likewise.
2731 (vreinterpretq_p8_s32): Likewise.
2732 (vreinterpretq_p8_s64): Likewise.
2733 (vreinterpretq_p8_f32): Likewise.
2734 (vreinterpretq_p8_u8): Likewise.
2735 (vreinterpretq_p8_u16): Likewise.
2736 (vreinterpretq_p8_u32): Likewise.
2737 (vreinterpretq_p8_u64): Likewise.
2738 (vreinterpretq_p8_p16): Likewise.
2739 (vreinterpret_p16_s8): Likewise.
2740 (vreinterpret_p16_s16): Likewise.
2741 (vreinterpret_p16_s32): Likewise.
2742 (vreinterpret_p16_s64): Likewise.
2743 (vreinterpret_p16_f32): Likewise.
2744 (vreinterpret_p16_u8): Likewise.
2745 (vreinterpret_p16_u16): Likewise.
2746 (vreinterpret_p16_u32): Likewise.
2747 (vreinterpret_p16_u64): Likewise.
2748 (vreinterpret_p16_p8): Likewise.
2749 (vreinterpretq_p16_s8): Likewise.
2750 (vreinterpretq_p16_s16): Likewise.
2751 (vreinterpretq_p16_s32): Likewise.
2752 (vreinterpretq_p16_s64): Likewise.
2753 (vreinterpretq_p16_f32): Likewise.
2754 (vreinterpretq_p16_u8): Likewise.
2755 (vreinterpretq_p16_u16): Likewise.
2756 (vreinterpretq_p16_u32): Likewise.
2757 (vreinterpretq_p16_u64): Likewise.
2758 (vreinterpretq_p16_p8): Likewise.
2759 (vreinterpret_f32_s8): Likewise.
2760 (vreinterpret_f32_s16): Likewise.
2761 (vreinterpret_f32_s32): Likewise.
2762 (vreinterpret_f32_s64): Likewise.
2763 (vreinterpret_f32_u8): Likewise.
2764 (vreinterpret_f32_u16): Likewise.
2765 (vreinterpret_f32_u32): Likewise.
2766 (vreinterpret_f32_u64): Likewise.
2767 (vreinterpret_f32_p8): Likewise.
2768 (vreinterpret_f32_p16): Likewise.
2769 (vreinterpretq_f32_s8): Likewise.
2770 (vreinterpretq_f32_s16): Likewise.
2771 (vreinterpretq_f32_s32): Likewise.
2772 (vreinterpretq_f32_s64): Likewise.
2773 (vreinterpretq_f32_u8): Likewise.
2774 (vreinterpretq_f32_u16): Likewise.
2775 (vreinterpretq_f32_u32): Likewise.
2776 (vreinterpretq_f32_u64): Likewise.
2777 (vreinterpretq_f32_p8): Likewise.
2778 (vreinterpretq_f32_p16): Likewise.
2779 (vreinterpret_s64_s8): Likewise.
2780 (vreinterpret_s64_s16): Likewise.
2781 (vreinterpret_s64_s32): Likewise.
2782 (vreinterpret_s64_f32): Likewise.
2783 (vreinterpret_s64_u8): Likewise.
2784 (vreinterpret_s64_u16): Likewise.
2785 (vreinterpret_s64_u32): Likewise.
2786 (vreinterpret_s64_u64): Likewise.
2787 (vreinterpret_s64_p8): Likewise.
2788 (vreinterpret_s64_p16): Likewise.
2789 (vreinterpretq_s64_s8): Likewise.
2790 (vreinterpretq_s64_s16): Likewise.
2791 (vreinterpretq_s64_s32): Likewise.
2792 (vreinterpretq_s64_f32): Likewise.
2793 (vreinterpretq_s64_u8): Likewise.
2794 (vreinterpretq_s64_u16): Likewise.
2795 (vreinterpretq_s64_u32): Likewise.
2796 (vreinterpretq_s64_u64): Likewise.
2797 (vreinterpretq_s64_p8): Likewise.
2798 (vreinterpretq_s64_p16): Likewise.
2799 (vreinterpret_u64_s8): Likewise.
2800 (vreinterpret_u64_s16): Likewise.
2801 (vreinterpret_u64_s32): Likewise.
2802 (vreinterpret_u64_s64): Likewise.
2803 (vreinterpret_u64_f32): Likewise.
2804 (vreinterpret_u64_u8): Likewise.
2805 (vreinterpret_u64_u16): Likewise.
2806 (vreinterpret_u64_u32): Likewise.
2807 (vreinterpret_u64_p8): Likewise.
2808 (vreinterpret_u64_p16): Likewise.
2809 (vreinterpretq_u64_s8): Likewise.
2810 (vreinterpretq_u64_s16): Likewise.
2811 (vreinterpretq_u64_s32): Likewise.
2812 (vreinterpretq_u64_s64): Likewise.
2813 (vreinterpretq_u64_f32): Likewise.
2814 (vreinterpretq_u64_u8): Likewise.
2815 (vreinterpretq_u64_u16): Likewise.
2816 (vreinterpretq_u64_u32): Likewise.
2817 (vreinterpretq_u64_p8): Likewise.
2818 (vreinterpretq_u64_p16): Likewise.
2819 (vreinterpret_s8_s16): Likewise.
2820 (vreinterpret_s8_s32): Likewise.
2821 (vreinterpret_s8_s64): Likewise.
2822 (vreinterpret_s8_f32): Likewise.
2823 (vreinterpret_s8_u8): Likewise.
2824 (vreinterpret_s8_u16): Likewise.
2825 (vreinterpret_s8_u32): Likewise.
2826 (vreinterpret_s8_u64): Likewise.
2827 (vreinterpret_s8_p8): Likewise.
2828 (vreinterpret_s8_p16): Likewise.
2829 (vreinterpretq_s8_s16): Likewise.
2830 (vreinterpretq_s8_s32): Likewise.
2831 (vreinterpretq_s8_s64): Likewise.
2832 (vreinterpretq_s8_f32): Likewise.
2833 (vreinterpretq_s8_u8): Likewise.
2834 (vreinterpretq_s8_u16): Likewise.
2835 (vreinterpretq_s8_u32): Likewise.
2836 (vreinterpretq_s8_u64): Likewise.
2837 (vreinterpretq_s8_p8): Likewise.
2838 (vreinterpretq_s8_p16): Likewise.
2839 (vreinterpret_s16_s8): Likewise.
2840 (vreinterpret_s16_s32): Likewise.
2841 (vreinterpret_s16_s64): Likewise.
2842 (vreinterpret_s16_f32): Likewise.
2843 (vreinterpret_s16_u8): Likewise.
2844 (vreinterpret_s16_u16): Likewise.
2845 (vreinterpret_s16_u32): Likewise.
2846 (vreinterpret_s16_u64): Likewise.
2847 (vreinterpret_s16_p8): Likewise.
2848 (vreinterpret_s16_p16): Likewise.
2849 (vreinterpretq_s16_s8): Likewise.
2850 (vreinterpretq_s16_s32): Likewise.
2851 (vreinterpretq_s16_s64): Likewise.
2852 (vreinterpretq_s16_f32): Likewise.
2853 (vreinterpretq_s16_u8): Likewise.
2854 (vreinterpretq_s16_u16): Likewise.
2855 (vreinterpretq_s16_u32): Likewise.
2856 (vreinterpretq_s16_u64): Likewise.
2857 (vreinterpretq_s16_p8): Likewise.
2858 (vreinterpretq_s16_p16): Likewise.
2859 (vreinterpret_s32_s8): Likewise.
2860 (vreinterpret_s32_s16): Likewise.
2861 (vreinterpret_s32_s64): Likewise.
2862 (vreinterpret_s32_f32): Likewise.
2863 (vreinterpret_s32_u8): Likewise.
2864 (vreinterpret_s32_u16): Likewise.
2865 (vreinterpret_s32_u32): Likewise.
2866 (vreinterpret_s32_u64): Likewise.
2867 (vreinterpret_s32_p8): Likewise.
2868 (vreinterpret_s32_p16): Likewise.
2869 (vreinterpretq_s32_s8): Likewise.
2870 (vreinterpretq_s32_s16): Likewise.
2871 (vreinterpretq_s32_s64): Likewise.
2872 (vreinterpretq_s32_f32): Likewise.
2873 (vreinterpretq_s32_u8): Likewise.
2874 (vreinterpretq_s32_u16): Likewise.
2875 (vreinterpretq_s32_u32): Likewise.
2876 (vreinterpretq_s32_u64): Likewise.
2877 (vreinterpretq_s32_p8): Likewise.
2878 (vreinterpretq_s32_p16): Likewise.
2879 (vreinterpret_u8_s8): Likewise.
2880 (vreinterpret_u8_s16): Likewise.
2881 (vreinterpret_u8_s32): Likewise.
2882 (vreinterpret_u8_s64): Likewise.
2883 (vreinterpret_u8_f32): Likewise.
2884 (vreinterpret_u8_u16): Likewise.
2885 (vreinterpret_u8_u32): Likewise.
2886 (vreinterpret_u8_u64): Likewise.
2887 (vreinterpret_u8_p8): Likewise.
2888 (vreinterpret_u8_p16): Likewise.
2889 (vreinterpretq_u8_s8): Likewise.
2890 (vreinterpretq_u8_s16): Likewise.
2891 (vreinterpretq_u8_s32): Likewise.
2892 (vreinterpretq_u8_s64): Likewise.
2893 (vreinterpretq_u8_f32): Likewise.
2894 (vreinterpretq_u8_u16): Likewise.
2895 (vreinterpretq_u8_u32): Likewise.
2896 (vreinterpretq_u8_u64): Likewise.
2897 (vreinterpretq_u8_p8): Likewise.
2898 (vreinterpretq_u8_p16): Likewise.
2899 (vreinterpret_u16_s8): Likewise.
2900 (vreinterpret_u16_s16): Likewise.
2901 (vreinterpret_u16_s32): Likewise.
2902 (vreinterpret_u16_s64): Likewise.
2903 (vreinterpret_u16_f32): Likewise.
2904 (vreinterpret_u16_u8): Likewise.
2905 (vreinterpret_u16_u32): Likewise.
2906 (vreinterpret_u16_u64): Likewise.
2907 (vreinterpret_u16_p8): Likewise.
2908 (vreinterpret_u16_p16): Likewise.
2909 (vreinterpretq_u16_s8): Likewise.
2910 (vreinterpretq_u16_s16): Likewise.
2911 (vreinterpretq_u16_s32): Likewise.
2912 (vreinterpretq_u16_s64): Likewise.
2913 (vreinterpretq_u16_f32): Likewise.
2914 (vreinterpretq_u16_u8): Likewise.
2915 (vreinterpretq_u16_u32): Likewise.
2916 (vreinterpretq_u16_u64): Likewise.
2917 (vreinterpretq_u16_p8): Likewise.
2918 (vreinterpretq_u16_p16): Likewise.
2919 (vreinterpret_u32_s8): Likewise.
2920 (vreinterpret_u32_s16): Likewise.
2921 (vreinterpret_u32_s32): Likewise.
2922 (vreinterpret_u32_s64): Likewise.
2923 (vreinterpret_u32_f32): Likewise.
2924 (vreinterpret_u32_u8): Likewise.
2925 (vreinterpret_u32_u16): Likewise.
2926 (vreinterpret_u32_u64): Likewise.
2927 (vreinterpret_u32_p8): Likewise.
2928 (vreinterpret_u32_p16): Likewise.
2929 (vreinterpretq_u32_s8): Likewise.
2930 (vreinterpretq_u32_s16): Likewise.
2931 (vreinterpretq_u32_s32): Likewise.
2932 (vreinterpretq_u32_s64): Likewise.
2933 (vreinterpretq_u32_f32): Likewise.
2934 (vreinterpretq_u32_u8): Likewise.
2935 (vreinterpretq_u32_u16): Likewise.
2936 (vreinterpretq_u32_u64): Likewise.
2937 (vreinterpretq_u32_p8): Likewise.
2938 (vreinterpretq_u32_p16): Likewise.
2940 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2942 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
2944 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
2947 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
2948 (vqnegd_s64): Likewise.
2949 (vqabs_s64): Likewise.
2950 (vqabsd_s64): Likewise.
2952 2014-04-22 Richard Henderson <rth@redhat.com>
2954 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
2955 computation to the top of the loop.
2957 2014-04-22 Renlin <renlin.li@arm.com>
2958 Jiong Wang <jiong.wang@arm.com>
2960 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
2961 * config/aarch64/aarch64.c (aarch64_layout_frame)
2962 (aarch64_initial_elimination_offset): Likewise.
2964 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
2966 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
2969 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
2971 * machmode.h (bitwise_mode_for_mode): Declare.
2972 * stor-layout.h (bitwise_type_for_mode): Likewise.
2973 * stor-layout.c (bitwise_mode_for_mode): New function.
2974 (bitwise_type_for_mode): Likewise.
2975 * builtins.c (fold_builtin_memory_op): Use it instead of
2976 int_mode_for_mode and build_nonstandard_integer_type.
2978 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2980 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
2981 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
2982 (*-*-solaris2*): Simplify.
2983 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
2984 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
2985 *-*-solaris2.9* handling.
2987 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
2989 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
2990 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
2992 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
2993 * configure: Regenerate.
2995 * config/i386/sol2-9.h: Remove.
2997 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
2998 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
2999 Remove Solaris 9 references.
3001 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
3003 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
3004 (floatuns<GPI:mode><GPF:mode>2): Remove.
3005 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
3006 and floatuns conversions.
3007 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
3008 and floatuns conversions.
3009 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
3010 (w1,w2): New mode attributes for inequal width conversions.
3012 2014-04-22 Renlin Li <Renlin.Li@arm.com>
3014 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
3015 the output asm format.
3017 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
3019 * config/aarch64/aarch64-simd.md
3020 (aarch64_cm<optab>di): Always split.
3021 (*aarch64_cm<optab>di): New.
3022 (aarch64_cmtstdi): Always split.
3023 (*aarch64_cmtstdi): New.
3025 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3027 PR tree-optimization/60823
3028 * omp-low.c (ipa_simd_modify_function_body): Go through
3029 all SSA_NAMEs and for those refering to vector arguments
3030 which are going to be replaced adjust SSA_NAME_VAR and,
3031 if it is a default definition, change it into a non-default
3032 definition assigned at the beginning of function from new_decl.
3033 (ipa_simd_modify_stmt_ops): Rewritten.
3034 * tree-dfa.c (set_ssa_default_def): When removing default def,
3035 check for NULL loc instead of NULL *loc.
3037 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3039 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
3040 restrictions on core registers for DImode values in Thumb2.
3042 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3044 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
3045 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
3047 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3049 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
3050 (*iordi_notzesidi_di): Likewise.
3051 (*iordi_notsesidi_di): Likewise.
3053 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3055 * config/arm/arm-protos.h (tune_params): New struct members.
3056 * config/arm/arm.c: Initialise tune_params per processor.
3057 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
3058 for speed, based on new tune_params.
3060 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3062 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
3064 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
3066 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
3068 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
3069 * config/aarch64/arm_neon.h (vrnd_f64): Added.
3070 (vrnda_f64): Likewise.
3071 (vrndi_f64): Likewise.
3072 (vrndm_f64): Likewise.
3073 (vrndn_f64): Likewise.
3074 (vrndp_f64): Likewise.
3075 (vrndx_f64): Likewise.
3077 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3079 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
3080 GET_MODE_SIZE argument is enum machine_mode.
3082 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3085 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
3086 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
3088 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
3091 * asan.c (asan_emit_stack_protection): Force the base to align to
3092 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
3093 appropriate bits if STRICT_ALIGNMENT.
3094 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
3096 (expand_used_vars): Leave a space in the stack frame for alignment
3097 if STRICT_ALIGNMENT.
3099 2014-04-21 David Malcolm <dmalcolm@redhat.com>
3101 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
3103 (gimple_store_p): Likewise.
3104 (gimple_assign_load_p): Likewise.
3105 (gimple_assign_cast_p): Likewise.
3106 (gimple_clobber_p): Likewise.
3108 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
3109 rather than a gimple.
3110 (gimple_assign_cast_p): Likewise.
3112 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
3115 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
3116 If mode is DDmode and TARGET_E500_DOUBLE allow move.
3118 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
3119 more debug information for E500 if -mdebug=reg.
3121 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
3124 * config/i386/i386.c (ix86_expand_builtin)
3125 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
3126 register for target RTX.
3127 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
3129 2014-04-18 Cong Hou <congh@google.com>
3131 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
3132 the widen-mult pattern by handling two operands with different sizes,
3133 and operands whose size is smaller than half of the result type.
3135 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3137 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
3138 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
3139 (do_estimate_edge_time): Compute it.
3140 * ipa-inline.c (want_inline_small_function_p): Bypass
3141 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
3143 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3145 * ipa-inline.c (spec_rem): New static variable.
3146 (dump_overall_stats): New function.
3147 (dump_inline_stats): New function.
3149 2014-04-18 Richard Henderson <rth@redhat.com>
3151 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
3152 to GET_MODE_SIZE, not a reg_class_t.
3154 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3156 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
3157 (vsx_xxmrglw_<mode>): Likewise.
3159 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3162 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
3163 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
3164 (rs6000_init_hard_regno_mode_ok): Likewise.
3166 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
3168 * ipa-inline.c (inline_small_functions): Account only non-cold
3170 * doc/invoke.texi (inline-unit-growth): Update documentation.
3172 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
3174 * config/rs6000/rs6000.md (addti3, subti3): New.
3176 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
3179 * config/i386/i386.c (ix86_expand_clear): Remove outdated
3180 comment. Check optimize_insn_for_size_p instead of
3181 optimize_insn_for_speed_p.
3183 2014-04-17 Martin Jambor <mjambor@suse.cz>
3185 * gimple-iterator.c (gsi_start_edge): New function.
3186 * gimple-iterator.h (gsi_start_edge): Declare.
3187 * tree-sra.c (single_non_eh_succ): New function.
3188 (disqualify_ops_if_throwing_stmt): Renamed to
3189 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
3190 having one non-EH successor BB.
3191 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
3192 generate loads into replacements.
3193 (sra_modify_assign): Likewise and and also use the simple path for
3195 (sra_modify_function_body): Commit statements on edges.
3197 2014-04-17 Richard Biener <rguenther@suse.de>
3200 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
3201 comparison results and add clarifying comment.
3203 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3205 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
3206 (blank_mode): Initialize it.
3207 (emit_mode_size_inline, emit_mode_nunits_inline,
3208 emit_mode_inner_inline): New functions.
3209 (emit_insn_modes_h): Call them and surround their output with
3210 #if GCC_VERSION >= 4001 ... #endif.
3211 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
3212 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
3213 mode_* arrays if the argument is __builtin_constant_p.
3214 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
3215 is enum machine_mode.
3217 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3219 * passes.c (opt_pass::execute): Adjust.
3220 (pass_manager::execute_pass_mode_switching): Likewise.
3221 (early_local_passes::execute): Likewise.
3222 (execute_one_pass): Pass cfun to the pass's execute method.
3223 * tree-pass.h (opt_pass::execute): Add function * argument.
3224 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3225 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3226 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3227 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3228 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3229 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
3230 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
3231 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3232 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3233 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3234 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3235 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3236 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3237 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3238 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3239 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3240 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3241 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3242 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3243 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3244 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3245 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3246 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3247 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3248 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3249 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3250 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3251 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3252 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3255 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3257 * passes.c (opt_pass::gate): Take function * argument.
3258 (gate_all_early_local_passes): Merge into
3259 (early_local_passes::gate): this.
3260 (gate_all_early_optimizations): Merge into
3261 (all_early_optimizations::gate): this.
3262 (gate_all_optimizations): Mege into
3263 (all_optimizations::gate): this.
3264 (gate_all_optimizations_g): Merge into
3265 (all_optimizations_g::gate): this.
3266 (gate_rest_of_compilation): Mege into
3267 (rest_of_compilation::gate): this.
3268 (gate_postreload): Merge into
3269 (postreload::gate): this.
3270 (dump_one_pass): Pass cfun to the pass's gate method.
3271 (execute_ipa_summary_passes): Likewise.
3272 (execute_one_pass): Likewise.
3273 (ipa_write_summaries_2): Likewise.
3274 (ipa_write_optimization_summaries_1): Likewise.
3275 (ipa_read_summaries_1): Likewise.
3276 (ipa_read_optimization_summaries_1): Likewise.
3277 (execute_ipa_stmt_fixups): Likewise.
3278 * tree-pass.h (opt_pass::gate): Add function * argument.
3279 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
3280 combine-stack-adj.c, combine.c, compare-elim.c,
3281 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3282 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
3283 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
3284 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
3285 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3286 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3287 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3288 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
3289 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
3290 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
3291 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3292 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3293 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
3294 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3295 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3296 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3297 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3298 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3299 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3300 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3301 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3302 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3303 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3304 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
3305 var-tracking.c, vtable-verify.c, web.c: Adjust.
3307 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3309 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
3310 * configure: Regenerate.
3312 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3314 * passes.c (dump_one_pass): don't check pass->has_gate.
3315 (execute_ipa_summary_passes): Likewise.
3316 (execute_one_pass): Likewise.
3317 (ipa_write_summaries_2): Likewise.
3318 (ipa_write_optimization_summaries_1): Likewise.
3319 (ipa_read_optimization_summaries_1): Likewise.
3320 (execute_ipa_stmt_fixups): Likewise.
3321 * tree-pass.h (pass_data::has_gate): Remove.
3322 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3323 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3324 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3325 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3326 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3327 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
3328 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
3329 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
3330 gimple-low.c, gimple-ssa-isolate-paths.c,
3331 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
3332 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3333 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3334 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
3335 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
3336 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
3337 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
3338 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3339 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
3340 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3341 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3342 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3343 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3344 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3345 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3346 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3347 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3348 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3349 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3350 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3351 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3352 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3355 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3357 * pass_manager.h (pass_manager::register_dump_files_1): Remove
3359 * passes.c (pass_manager::register_dump_files_1): Merge into
3360 (pass_manager::register_dump_files): this, and remove its handling of
3361 properties since the pass always has the properties anyway.
3362 (pass_manager::pass_manager): Adjust.
3364 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3366 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
3367 * passes.c (pass_manager::register_dump_files_1): Remove dead code
3368 dealing with properties.
3369 (pass_manager::register_dump_files): Adjust.
3371 2014-03-20 Mark Wielaard <mjw@redhat.com>
3373 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
3374 then represent the bound as normal constant value.
3376 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3379 Forward port from 4.8 branch
3380 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
3382 * config/i386/bmiintrin.h (_blsi_u32): New.
3386 (_blsmsk_u32): Ditto.
3387 (_blsmsk_u64): Ditto.
3388 (_tzcnt_u32): Ditto.
3389 (_tzcnt_u64): Ditto.
3391 2014-04-17 Kito Cheng <kito@0xlab.org>
3393 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
3395 2014-04-17 Richard Biener <rguenther@suse.de>
3398 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
3399 boolean results for comparisons.
3401 2014-04-17 Richard Biener <rguenther@suse.de>
3403 PR tree-optimization/60836
3404 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
3405 initial PHI args to be gimple values.
3407 2014-04-17 Richard Biener <rguenther@suse.de>
3409 PR tree-optimization/60841
3410 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
3411 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
3412 of stmts to SLP build.
3413 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
3414 (vect_analyze_slp): Likewise.
3415 (vect_analyze_slp_instance): Likewise.
3416 (vect_build_slp_tree): Limit overall SLP tree growth.
3417 * tree-vectorizer.h (vect_analyze_data_refs,
3418 vect_analyze_slp): Adjust prototypes.
3420 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3422 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
3425 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3427 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
3428 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
3429 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
3430 for TARGET_SLOW_PSHUFB
3432 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3434 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
3435 * config/i386/i386.c (intel_cost): Ditto.
3437 2014-04-17 Joey Ye <joey.ye@arm.com>
3439 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
3441 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3443 * opts.c (common_handle_option): Disable -fipa-reference coorectly
3446 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3448 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
3449 (type_all_derivations_known_p): New predicate.
3450 (type_all_ctors_visible_p): New predicate.
3451 (type_possibly_instantiated_p): New predicate.
3452 (get_odr_type): Compute all_derivations_known.
3453 (dump_odr_type): Dump the flag.
3454 (maybe_record_type): Cleanup.
3455 (record_target_from_binfo): Add bases_to_consider array;
3456 record bases for types w/o instances and skip CXX destructor.
3457 (possible_polymorphic_call_targets_1): Add bases_to_consider
3458 and consider_construction parameters; check if type may have instance.
3459 (get_polymorphic_call_info): Set maybe_in_construction to true
3460 when we know nothing.
3461 (record_targets_from_bases): Skip CXX destructors; they are
3462 never called for types in construction.
3463 (possible_polymorphic_call_targets): Do not record target when
3464 type may not have instance.
3466 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3469 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
3470 external aliases alive, too.
3472 2014-04-16 Andrew Pinski <apinski@cavium.com>
3474 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
3477 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3479 * final.c (compute_alignments): Do not apply loop alignment to a block
3480 falling through to the exit.
3482 2014-04-16 Catherine Moore <clm@codesourcery.com>
3484 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
3485 Adjust constraints for microMIPS store patterns.
3487 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3489 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
3491 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3493 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
3494 (append_use): Run at -O0.
3495 (append_vdef): Likewise.
3496 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
3497 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
3499 2014-04-16 Jakub Jelinek <jakub@redhat.com>
3501 PR tree-optimization/60844
3502 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
3503 (propagate_op_to_single_use, remove_visited_stmt_chain,
3504 linearize_expr, repropagate_negates, reassociate_bb): Use it
3505 instead of gsi_remove.
3507 2014-04-16 Martin Jambor <mjambor@suse.cz>
3509 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
3510 ipa_transforms_to_apply.
3511 (cgraph_function_versioning): Assert that old_node has empty
3512 ipa_transforms_to_apply.
3513 * trans-mem.c (ipa_tm_create_version): Likewise.
3514 * tree-inline.c (tree_function_versioning): Do not duplicate
3515 ipa_transforms_to_apply.
3517 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3520 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
3522 Pass necessary as flags on 64-bit Solaris/x86.
3523 Use lowercase relocs for x86_64-*-*.
3524 * configure: Regenerate.
3526 2014-04-15 Jan Hubicka <jh@suse.cz>
3528 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
3529 (maybe_record_node, likely_target_p): Use it.
3531 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3534 Revert following patch
3536 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3539 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3540 software floating point or no floating point registers, do not
3541 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3542 in GPRs that occurs after we tested for GPRs that would never be
3545 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3546 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3547 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3548 specifically allow DDmode, since that does not use the SPE SIMD
3551 2014-03-21 Mark Wielaard <mjw@redhat.com>
3553 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
3554 as unsigned or int depending on type and value used.
3556 2014-04-15 Richard Biener <rguenther@suse.de>
3558 PR rtl-optimization/56965
3559 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
3560 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
3562 * alias.c (true_dependence_1): Do not call
3563 nonoverlapping_component_refs_p.
3564 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
3565 nonoverlapping_component_refs_p.
3566 (indirect_refs_may_alias_p): Likewise.
3568 2014-04-15 Teresa Johnson <tejohnson@google.com>
3570 * cfg.c (dump_bb_info): Fix flags check.
3571 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
3573 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3575 PR rtl-optimization/60663
3576 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
3579 2014-04-15 Richard Biener <rguenther@suse.de>
3581 * lto-streamer.h (LTO_major_version): Bump to 4.
3583 2014-04-15 Richard Biener <rguenther@suse.de>
3585 * common.opt (lto_partition_model): New enum.
3586 (flto-partition=): Merge separate options with a single with argument,
3587 add -flto-partition=one support.
3588 * flag-types.h (enum lto_partition_model): Declare.
3589 * opts.c (finish_options): Remove duplicate -flto-partition=
3591 * lto-wrapper.c (run_gcc): Adjust.
3593 2014-04-15 Richard Biener <rguenther@suse.de>
3595 * alias.c (ncr_compar): New function.
3596 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
3598 2014-04-15 Richard Biener <rguenther@suse.de>
3600 * alias.c (record_component_aliases): Do not walk BINFOs.
3602 2014-04-15 Richard Biener <rguenther@suse.de>
3604 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
3605 Add struct function argument and adjust.
3606 (find_func_aliases_for_call): Likewise.
3607 (find_func_aliases): Likewise.
3608 (find_func_clobbers): Likewise.
3609 (intra_create_variable_infos): Likewise.
3610 (compute_points_to_sets): Likewise.
3611 (ipa_pta_execute): Adjust. Do not push/pop cfun.
3613 2014-04-15 Richard Biener <rguenther@suse.de>
3615 * tree.c (iterative_hash_expr): Use enum tree_code_class
3616 to store TREE_CODE_CLASS.
3617 (tree_block): Likewise.
3618 (tree_set_block): Likewise.
3619 * tree.h (fold_build_pointer_plus_loc): Use
3620 convert_to_ptrofftype_loc.
3622 2014-04-15 Jakub Jelinek <jakub@redhat.com>
3625 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
3628 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
3630 * cfgloop.h (struct loop): Move force_vectorize down.
3631 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
3632 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
3633 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
3634 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
3635 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
3636 * tree-core.h (enum annot_expr_kind): Add new kind values.
3637 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
3638 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
3640 * tree.def (ANNOTATE_EXPR): Tweak comment.
3642 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3644 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
3647 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
3649 * tree.h (TYPE_IDENTIFIER): Declare.
3650 * tree.c (subrange_type_for_debug_p): Use it.
3651 * godump.c (go_format_type): Likewise.
3652 * dwarf2out.c (is_cxx_auto, modified_type_die,
3653 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
3654 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
3656 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3659 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
3661 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
3663 * config/i386/i386.c (examine_argument): Return bool. Return true if
3664 parameter should be passed in memory.
3665 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
3666 (construct_container): Update calls to examine_argument.
3667 (function_arg_advance_64): Ditto.
3668 (return_in_memory_32): Merge with ix86_return_in_memory.
3669 (return_in_memory_64): Ditto.
3670 (return_in_memory_ms_64): Ditto.
3672 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3674 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
3675 * coverage.c (coverage_compute_profile_id): Handle externally visible
3678 2014-04-14 Martin Jambor <mjambor@suse.cz>
3680 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
3681 DECL_DISREGARD_INLINE_LIMITS functions.
3683 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3686 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
3688 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3691 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
3692 optimize_insn_for_speed_p instead of
3693 optimize_function_for_speed_p.
3695 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
3697 * doc/invoke.texi (free): Document AArch64.
3699 2014-04-14 Richard Biener <rguenther@suse.de>
3701 PR tree-optimization/60042
3702 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
3703 (insert_into_preds_of_block): Do not prevent PHI insertion
3704 for REFERENCE exprs here ...
3705 (eliminate_dom_walker::before_dom_children): ... but prevent
3706 their use here under similar conditions when applied to the
3707 IL after PRE optimizations.
3709 2014-04-14 Richard Biener <rguenther@suse.de>
3711 * passes.def: Move early points-to after early SRA.
3713 2014-04-14 Richard Biener <rguenther@suse.de>
3715 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
3716 check for which sign-changes we allow when forwarding
3717 a converted value into a switch.
3719 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3721 * stor-layout.c (place_field): Finalize non-constant offset for the
3724 2014-04-14 Richard Biener <rguenther@suse.de>
3726 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
3728 (expand_switch_using_bit_tests_p): Likewise.
3729 (process_switch): Compute and pass on speed_p based on the
3731 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
3732 optimize_bb_for_speed_p.
3734 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3736 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
3737 * function.h (struct function): Rename has_force_vect_loops into
3738 has_force_vectorize_loops.
3739 * lto-streamer-in.c (input_cfg): Adjust for renaming.
3740 (input_struct_function_base): Likewise.
3741 * lto-streamer-out.c (output_cfg): Likewise.
3742 (output_struct_function_base): Likewise.
3743 * omp-low.c (expand_omp_simd): Likewise.
3744 * tree-cfg.c (move_sese_region_to_fn): Likewise.
3745 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
3746 (version_loop_for_if_conversion): Likewise.
3747 (tree_if_conversion): Likewise.
3748 (main_tree_if_conversion): Likewise.
3749 (gate_tree_if_conversion): Likewise.
3750 * tree-inline.c (copy_loops): Likewise.
3751 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
3752 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
3753 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
3754 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
3755 * tree-vectorizer.c (vectorize_loops): Likewise.
3756 * tree-vectorizer.h (unlimited_cost_model): Likewise.
3758 2014-04-14 Richard Biener <rguenther@suse.de>
3761 * lto-streamer-out.c (wrap_refs): New function.
3762 (lto_output): Wrap symbol references in global initializes in
3763 type-preserving MEM_REFs.
3765 2014-04-14 Christian Bruel <christian.bruel@st.com>
3767 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
3769 2014-04-14 Christian Bruel <christian.bruel@st.com>
3771 * config/sh/sh.md (setmemqi): New expand pattern.
3772 * config/sh/sh.h (CLEAR_RATIO): Define.
3773 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
3774 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
3776 2014-04-14 Richard Biener <rguenther@suse.de>
3779 * fold-const.c (negate_expr_p): Don't negate directional rounding
3781 (fold_negate_expr): Likewise.
3783 2014-04-14 Richard Biener <rguenther@suse.de>
3785 PR tree-optimization/59817
3786 PR tree-optimization/60453
3787 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
3788 recursion to catch all CHRECs in the scalar evolution and restrict
3789 the predicate for the remains appropriately.
3791 2014-04-12 Catherine Moore <clm@codesourcery.com>
3793 * config/mips/constraints.md: Add new register constraint "kb".
3794 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
3795 (*movhi_internal): Likewise.
3796 (*movqi_internal): Likewise.
3797 * config/mips/mips.h (M16_STORE_REGS): New register class.
3798 (REG_CLASS_NAMES): Add M16_STORE_REGS.
3799 (REG_CLASS_CONTENTS): Likewise.
3800 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
3802 2014-04-11 Tobias Burnus <burnus@net-b.de>
3805 * doc/invoke.texi (-Wformat-signedness): Document it.
3806 (Wformat=2): Mention that this enables -Wformat-signedness.
3808 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3810 * common/config/epiphany/epiphany-common.c
3811 (epiphany_option_optimization_table): Enable section anchors by
3812 default at -O1 or higher.
3813 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
3814 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
3815 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
3816 carries no extra cost.
3817 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
3818 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
3819 * config/epiphany/predicates.md (memclob_operand): New predicate.
3820 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
3821 Use memclob_operand predicate and X constraint for operand 3.
3823 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3825 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
3826 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
3829 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3831 PR rtl-optimization/60651
3832 * mode-switching.c (optimize_mode_switching): Make sure to emit
3833 sets of a lower numbered entity before sets of a higher numbered
3834 entity to a mode of the same or lower priority.
3835 When creating a seginfo for a basic block that starts with a code
3836 label, move the insertion point past the code label.
3837 (new_seginfo): Document and enforce requirement that
3838 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
3839 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
3840 * doc/tm.texi: Regenerate.
3842 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
3845 * config/arc/arc.c (arc_save_restore): Fix assert typo.
3847 2013-04-11 Jakub Jelinek <jakub@redhat.com>
3849 * BASE-VER: Set to 4.10.0.
3851 2014-04-11 Tobias Burnus <burnus@net-b.de>
3854 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
3855 * doc/gcc.texi (Service): Update description in the @menu
3856 * doc/invoke.texi (Option Summary): Remove misplaced and
3859 2014-04-11 Steve Ellcey <sellcey@mips.com>
3860 Jakub Jelinek <jakub@redhat.com>
3863 * expr.c (convert_move): Use emit_store_flag_force instead of
3864 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
3867 2014-04-11 Richard Biener <rguenther@suse.de>
3870 * varasm.c (assemble_alias): Avoid endless error reporting
3871 recursion by setting TREE_ASM_WRITTEN.
3873 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3875 * config/s390/s390.md: Add a splitter for NOT rtx.
3877 2014-04-11 Jakub Jelinek <jakub@redhat.com>
3879 PR rtl-optimization/60663
3880 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
3882 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
3883 Jakub Jelinek <jakub@redhat.com>
3886 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
3887 flag from decl_node to node.
3889 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3892 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
3893 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
3894 ameliorating the cases where it can be.
3896 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
3899 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
3901 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3902 (loadsync_<mode>): Change mode.
3903 (load_quadpti, store_quadpti): New.
3904 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3905 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3906 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
3908 2014-04-09 Cong Hou <congh@google.com>
3911 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
3914 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3916 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
3917 instead of vnor to exploit possible fusion opportunity in the
3919 (altivec_expand_vec_perm_const_le): Likewise.
3921 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
3923 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3924 (loadsync_<mode>): Change mode.
3925 (load_quadpti, store_quadpti): New.
3926 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3927 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3929 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
3932 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
3933 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
3934 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
3936 2014-04-08 Richard Biener <rguenther@suse.de>
3939 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
3940 a 64bit widest int print double-int similar to on HWI64 hosts.
3942 2014-04-08 Richard Biener <rguenther@suse.de>
3944 PR tree-optimization/60785
3945 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
3946 default defs properly.
3948 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
3950 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
3951 (Weffc++): Likewise.
3953 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
3955 * ipa-devirt.c (maybe_record_node): When node is not recorded,
3956 set completep to false rather than true.
3958 2014-04-07 Douglas B Rupp <rupp@adacore.com>
3961 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
3962 ARM_TARGET2_DWARF_FORMAT.
3964 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
3967 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
3968 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
3971 2014-04-07 Richard Biener <rguenther@suse.de>
3973 PR tree-optimization/60766
3974 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
3975 (may_eliminate_iv): Convert cand_value_at result to desired type.
3977 2014-04-07 Jason Merrill <jason@redhat.com>
3980 * common.opt (-fno-gnu-unique): Add.
3981 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
3983 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3985 * haifa-sched.c: Fix outdated function reference and minor
3986 grammar errors in introductory comment.
3988 2014-04-07 Richard Biener <rguenther@suse.de>
3991 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
3993 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
3995 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
3998 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
3999 size accounting for thunks.
4000 (pa_asm_output_mi_thunk): Use final_start_function() and
4001 final_end_function() to output function start and end directives.
4003 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4005 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
4006 device specific ISA/ feature information. Remove short_sp and
4007 errata_skip ds. Add avr_device_specific_features enum to have device
4009 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
4010 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
4011 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
4012 updated device specific info.
4013 * config/avr/avr-mcus.def: Merge device specific details to
4014 dev_attribute field.
4015 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
4017 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
4018 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
4019 assembler if RMW isa supported by current device.
4020 * config/avr/genmultilib.awk: Update as device info structure changed.
4021 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
4023 2014-04-04 Cong Hou <congh@google.com>
4025 PR tree-optimization/60656
4026 * tree-vect-stmts.c (supportable_widening_operation):
4027 Fix a bug that elements in a vector with vect_used_by_reduction
4028 property are incorrectly reordered when the operation on it is not
4029 consistant with the one in reduction operation.
4031 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
4033 PR rtl-optimization/60155
4034 * gcse.c (record_set_data): New function.
4035 (single_set_gcse): New function.
4036 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
4037 (hoist_code): Likewise.
4038 (get_pressure_class_and_nregs): Likewise.
4040 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
4042 * explow.c (probe_stack_range): Emit a final optimization blockage.
4044 2014-04-04 Anthony Green <green@moxielogic.com>
4046 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
4049 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
4052 * lto-cgraph.c (input_overwrite_node): Check that partitioning
4053 flags are set only during streaming.
4054 * ipa.c (process_references, walk_polymorphic_call_targets,
4055 symtab_remove_unreachable_nodes): Drop bodies of always inline
4056 after early inlining.
4057 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
4059 2014-04-04 Jakub Jelinek <jakub@redhat.com>
4060 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4063 * dwarf2out.c (const_ok_for_output_1): Reject expressions
4066 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4069 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
4071 (cortex_a53_fdivd): Likewise.
4073 2014-04-04 Martin Jambor <mjambor@suse.cz>
4076 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
4078 * cgraph.c (clone_of_p): Also return true if thunks match.
4079 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
4080 cgraph_function_or_thunk_node and an obsolete comment.
4081 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
4083 (build_function_decl_skip_args): Likewise.
4084 (set_new_clone_decl_and_node_flags): New function.
4085 (duplicate_thunk_for_node): Likewise.
4086 (redirect_edge_duplicating_thunks): Likewise.
4087 (cgraph_clone_node): New parameter args_to_skip, pass it to
4088 redirect_edge_duplicating_thunks which is called instead of
4089 cgraph_redirect_edge_callee.
4090 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
4091 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
4093 2014-04-04 Jeff Law <law@redhat.com>
4096 * config/arm/predicates.md (const_int_I_operand): New predicate.
4097 (const_int_M_operand): Similarly.
4098 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
4100 (insv_t2, extv_reg, extzv_t2): Likewise.
4101 (load_multiple_with_writeback): Similarly for const_int_I_operand.
4102 (pop_multiple_with_writeback_and_return): Likewise.
4103 (vfp_pop_multiple_with_writeback): Likewise
4105 2014-04-04 Richard Biener <rguenther@suse.de>
4108 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
4109 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
4111 * gimplify.h (gimple_add_tmp_var_fn): Declare.
4112 * gimplify.c (gimple_add_tmp_var_fn): New function.
4113 * gimple-expr.h (create_tmp_reg_fn): Declare.
4114 * gimple-expr.c (create_tmp_reg_fn): New function.
4115 * gimple-low.c (record_vars_into): Don't change cfun.
4116 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
4117 code generation without cfun.
4119 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
4122 * Makefile.in (install-driver): Fix shell scripting.
4124 2014-04-03 Cong Hou <congh@google.com>
4126 PR tree-optimization/60505
4127 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
4128 threshold of number of iterations below which no vectorization
4130 * tree-vect-loop.c (new_loop_vec_info):
4131 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
4132 * tree-vect-loop.c (vect_analyze_loop_operations):
4133 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
4134 * tree-vect-loop.c (vect_transform_loop):
4135 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
4136 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
4137 of iterations of the loop and see if we should build the epilogue.
4139 2014-04-03 Richard Biener <rguenther@suse.de>
4141 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
4142 (streamer_tree_cache_create): Adjust.
4143 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
4144 to allow optional nodes array.
4145 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
4146 (streamer_tree_cache_append): Likewise.
4147 (streamer_tree_cache_create): Create nodes array optionally
4148 as specified by parameter.
4149 * lto-streamer-out.c (create_output_block): Avoid maintaining
4150 the node array in the writer cache.
4151 (DFS_write_tree): Remove assertion.
4152 (produce_asm_for_decls): Free the out decl state hash table early.
4153 * lto-streamer-in.c (lto_data_in_create): Adjust for
4154 streamer_tree_cache_create prototype change.
4156 2014-04-03 Richard Biener <rguenther@suse.de>
4158 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
4159 set TREE_CHAIN to NULL_TREE.
4161 2014-04-03 Richard Biener <rguenther@suse.de>
4163 PR tree-optimization/60740
4164 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
4165 over all GIMPLE_COND operands.
4167 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
4169 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
4170 (Weffc++): Remove Scott's numbering, merge lists and reference
4173 2014-04-03 Nick Clifton <nickc@redhat.com>
4175 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
4178 2014-04-03 Martin Jambor <mjambor@suse.cz>
4180 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
4181 mention gcc_unreachable before failing.
4182 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
4185 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
4188 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
4189 inconsistent code and instead mark the context inconsistent.
4190 (possible_polymorphic_call_targets): For inconsistent contexts
4191 return empty complete list.
4193 2014-04-02 Anthony Green <green@moxielogic.com>
4195 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
4196 (extendqisi2, extendhisi2): Define.
4197 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
4198 (WCHAR_TYPE): Change to unsigned int.
4200 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4202 PR tree-optimization/60733
4203 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
4204 insertion point for PHI candidates to be the end of the feeding
4205 block for the PHI argument.
4207 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
4209 PR rtl-optimization/60650
4210 * lra-constraints.c (process_alt_operands): Decrease reject for
4211 earlyclobber matching.
4213 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4215 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
4217 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4219 * config/spu/spu.c (pad_bb): Do not crash when the last
4220 insn is CODE_FOR_blockage.
4222 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4224 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
4225 lies outside the target mode.
4227 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4230 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4231 software floating point or no floating point registers, do not
4232 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4233 in GPRs that occurs after we tested for GPRs that would never be
4236 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4237 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4238 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4239 specifically allow DDmode, since that does not use the SPE SIMD
4242 2014-04-02 Richard Biener <rguenther@suse.de>
4245 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4246 MODE_INTs. Properly use negv_optab.
4247 (expand_abs): Likewise.
4249 2014-04-02 Richard Biener <rguenther@suse.de>
4252 * Makefile.in (install-driver): Guard extra installs with special
4255 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4257 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4260 2014-04-01 Richard Henderson <rth@redhat.com>
4263 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
4264 alternative enabled before register allocation.
4266 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
4268 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
4269 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
4271 (nios2_large_got_address): Remove unneeded 'sym' parameter.
4272 (nios2_got_address): Update nios2_large_got_address call site.
4273 (nios2_delegitimize_address): New function.
4274 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
4275 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
4276 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
4278 2014-04-01 Martin Husemann <martin@duskware.de>
4280 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
4283 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
4285 PR rtl-optimization/60604
4286 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
4287 check from register_operand.
4288 (register_operand): Redefine in terms of general_operand.
4289 (nonmemory_operand): Use register_operand for the non-constant cases.
4291 2014-04-01 Richard Biener <rguenther@suse.de>
4293 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
4295 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
4297 * doc/invoke.texi (mapp-regs): Clarify.
4299 2014-03-31 Ulrich Drepper <drepper@gmail.com>
4301 * config/i386/avx512fintrin.h (__v32hi): Define type.
4302 (__v64qi): Likewise.
4303 (_mm512_set1_epi8): Define.
4304 (_mm512_set1_epi16): Define.
4305 (_mm512_set4_epi32): Define.
4306 (_mm512_set4_epi64): Define.
4307 (_mm512_set4_pd): Define.
4308 (_mm512_set4_ps): Define.
4309 (_mm512_setr4_epi64): Define.
4310 (_mm512_setr4_epi32): Define.
4311 (_mm512_setr4_pd): Define.
4312 (_mm512_setr4_ps): Define.
4313 (_mm512_setzero_epi32): Define.
4315 2014-03-31 Martin Jambor <mjambor@suse.cz>
4318 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
4319 callsite_arguments_match_p. Updated all callers. Also check types of
4320 corresponding formal parameters and actual arguments.
4321 (not_all_callers_have_enough_arguments_p) Renamed to
4322 some_callers_have_mismatched_arguments_p.
4324 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
4326 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
4328 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
4331 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
4334 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
4336 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
4337 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
4339 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
4340 Use FMAMODE_NOVF512 mode iterator.
4341 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
4342 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
4343 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
4344 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
4346 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
4347 Use VF_128_256 mode iterator.
4348 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
4351 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4353 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
4354 static chain if needed.
4356 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4359 * lra-constraints.c (index_part_to_reg): New.
4360 (process_address): Use it.
4362 2014-03-27 Jeff Law <law@redhat.com>
4363 Jakub Jelinek <jakub@redhat.com>
4366 * expr.c (do_tablejump): Use simplify_gen_binary rather than
4367 gen_rtx_{PLUS,MULT} to build up the address expression.
4369 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
4370 creating non-canonical RTL.
4372 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4375 * ipa-inline.c (want_inline_small_function_p): Short circuit large
4376 functions; reorganize to make cheap checks first.
4377 (inline_small_functions): Do not estimate growth when dumping;
4379 * ipa-inline.h (inline_summary): Add min_size.
4380 (growth_likely_positive): New function.
4381 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
4382 (set_cond_stmt_execution_predicate): Cleanup.
4383 (estimate_edge_size_and_time): Compute min_size.
4384 (estimate_calls_size_and_time): Likewise.
4385 (estimate_node_size_and_time): Likewise.
4386 (inline_update_overall_summary): Update min_size.
4387 (do_estimate_edge_time): Likewise.
4388 (do_estimate_edge_size): Update.
4389 (do_estimate_edge_hints): Update.
4390 (growth_likely_positive): New function.
4392 2014-03-28 Jakub Jelinek <jakub@redhat.com>
4395 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
4396 also if addr has VOIDmode.
4398 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4400 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
4401 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
4403 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
4404 instructions as well as AdvancedSIMD loads.
4406 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4408 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
4409 Use crypto_aese type.
4410 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
4411 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
4412 crypto_aese, crypto_aesmc. Move to types.md.
4413 * config/arm/types.md (crypto_aes): Split into crypto_aese,
4415 * config/arm/iterators.md (crypto_type): Likewise.
4417 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4419 * cgraph.c: Include expr.h and tree-dfa.h.
4420 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
4423 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4426 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
4427 regs from checking multi-reg pseudos.
4429 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4431 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
4433 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4435 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
4436 if it would clobber the stack pointer, even temporarily.
4438 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
4440 * mode-switching.c: Make small adjustments to the top comment.
4442 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4444 * config/rs6000/constraints.md (wD constraint): New constraint to
4445 match the constant integer to get the top DImode/DFmode out of a
4446 vector in a VSX register.
4448 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
4449 match the constant integer to get the top DImode/DFmode out of a
4450 vector in a VSX register.
4452 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
4455 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
4458 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
4459 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
4461 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
4462 Optimize vec_extract of 64-bit values, where the value being
4463 extracted is in the top word, where we can use scalar
4464 instructions. Add direct move and store support. Combine the big
4465 endian/little endian vector select load support into a single insn.
4466 (vsx_extract_<mode>_internal1): Likewise.
4467 (vsx_extract_<mode>_internal2): Likewise.
4468 (vsx_extract_<mode>_load): Likewise.
4469 (vsx_extract_<mode>_store): Likewise.
4470 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
4471 combined into vsx_extract_<mode>_load.
4472 (vsx_extract_<mode>_one_le): Likewise.
4474 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
4475 define the top 64-bit vector element.
4477 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
4480 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4481 Document vec_vbpermq builtin.
4484 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
4485 enable use of xxsldwi and xxpermdi builtin functions.
4486 (vec_xxpermdi): Likewise.
4488 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4489 Document use of vec_xxsldwi and vec_xxpermdi builtins.
4491 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
4493 PR rtl-optimization/60650
4494 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
4496 (find_spills_for): New.
4497 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
4498 Spill all pseudos on the second iteration.
4500 2014-03-27 Marek Polacek <polacek@redhat.com>
4503 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
4506 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4508 * config/s390/s390.c (s390_can_use_return_insn): Check for
4509 call-saved FPRs on 31 bit.
4511 2014-03-27 Jakub Jelinek <jakub@redhat.com>
4514 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
4515 if they need regimplification, just drop them instead of
4516 calling gimple_regimplify_operands on them.
4518 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
4521 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
4522 (aarch64_frame_pointer_required): Adjust logic.
4523 (aarch64_can_eliminate): Adjust logic.
4524 (aarch64_override_options_after_change): Adjust logic.
4526 2014-03-27 Dehao Chen <dehao@google.com>
4528 * ipa-inline.c (early_inliner): Update node's inline info.
4530 2014-03-26 Dehao Chen <dehao@google.com>
4532 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
4533 compiler inserted conditional jumps for NAN float check.
4535 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4537 * ubsan.h (ubsan_create_data): Change second argument's type
4538 to const location_t *.
4539 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
4541 (ubsan_create_data): Change second argument to const location_t *PLOC.
4542 Create Loc field whenever PLOC is non-NULL.
4543 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
4544 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
4548 * real.c (real_to_integer2): Change type of low to UHWI.
4550 2014-03-26 Tobias Burnus <burnus@net-b.de>
4552 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
4553 (CILK_SELF_SPECS): New define.
4554 (driver_self_specs): Use it.
4556 2014-03-26 Richard Biener <rguenther@suse.de>
4558 * tree-pretty-print.c (percent_K_format): Implement special
4559 case for LTO and its stripped down BLOCK tree.
4561 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4564 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
4566 * tree-vrp.c (simplify_internal_call_using_ranges): If only
4567 one range is range_int_cst_p, but not both, at least optimize
4568 addition/subtraction of 0 and multiplication by 0 or 1.
4569 * gimple-fold.c (gimple_fold_call): Fold
4570 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
4571 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
4572 INTEGER_CSTs, try to fold at least x * 0 and y - y.
4574 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
4576 PR rtl-optimization/60452
4577 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
4578 <case REG>: Return 1 for invalid offsets from the frame pointer.
4580 2014-03-26 Marek Polacek <polacek@redhat.com>
4583 * doc/extend.texi (C Extensions): Mention variable-length arrays in
4586 2014-03-26 Marek Polacek <polacek@redhat.com>
4589 * doc/extend.texi (Designated Inits): Describe what happens to omitted
4592 2014-03-26 Marek Polacek <polacek@redhat.com>
4595 * ira-color.c (update_conflict_hard_regno_costs): Perform the
4596 multiplication in unsigned type.
4598 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4600 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
4602 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4604 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
4606 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4609 * cif-code.def (UNREACHABLE) New code.
4610 * ipa-inline.c (inline_small_functions): Skip edges to
4611 __builtlin_unreachable.
4612 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
4613 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
4614 predicate to __bulitin_unreachable.
4615 (set_cond_stmt_execution_predicate): Fix issue when
4616 invert_tree_comparison returns ERROR_MARK.
4617 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
4618 propagate to inline clones.
4619 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
4621 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
4622 * cgraphclones.c (cgraph_clone_node): If call destination is already
4623 ureachable, do not redirect it back.
4624 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
4627 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4629 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
4630 Do not modify inline clones.
4632 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4634 * config/i386/i386.md (general_sext_operand): New mode attr.
4635 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
4636 don't generate (sign_extend (const_int)).
4637 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
4638 operands[2]. Use We constraint instead of <i> and
4639 <general_sext_operand> predicate instead of <general_operand>.
4640 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
4641 * config/i386/constraints.md (We): New constraint.
4642 * config/i386/predicates.md (x86_64_sext_operand,
4643 sext_operand): New predicates.
4645 2014-03-25 Martin Jambor <mjambor@suse.cz>
4648 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
4649 inconsistent devirtualizations to __builtin_unreachable.
4651 2014-03-25 Marek Polacek <polacek@redhat.com>
4654 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
4656 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
4658 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
4659 order of elements for big-endian.
4661 2014-03-25 Richard Biener <rguenther@suse.de>
4664 * gimplify-me.c (gimple_regimplify_operands): Update the
4667 2014-03-25 Martin Jambor <mjambor@suse.cz>
4670 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
4671 (lto_output_varpool_node): Likewise.
4672 (input_overwrite_node): Likewise.
4673 (input_varpool_node): Likewise.
4675 2014-03-25 Richard Biener <rguenther@suse.de>
4677 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
4678 (run_gcc): Likewise.
4680 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4682 * combine.c (simplify_compare_const): Add MODE argument.
4683 Handle mode_width 0 as very large mode_width.
4684 (try_combine, simplify_comparison): Adjust callers.
4686 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
4687 type to avoid signed integer overflow.
4688 * explow.c (plus_constant): Likewise.
4690 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
4692 * doc/generic.texi: Correct typos.
4694 2014-03-24 Tobias Burnus <burnus@net-b.de>
4696 * doc/invoke.texi (-flto): Expand section about
4697 using static libraries with LTO.
4699 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4701 PR rtl-optimization/60501
4702 * optabs.def (addptr3_optab): New optab.
4703 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
4704 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
4705 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
4707 * lra.c (emit_add3_insn): Use the addptr pattern if available.
4709 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
4711 2014-03-24 Ulrich Drepper <drepper@gmail.com>
4713 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
4716 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
4717 (_mm256_undefined_ps): Define.
4718 (_mm256_undefined_pd): Define.
4719 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
4720 (_mm_undefined_pd): Define.
4721 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
4722 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
4723 (_mm512_undefined_ps): Define.
4724 (_mm512_undefined_pd): Define.
4725 Use _mm*_undefined_*.
4726 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
4728 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
4730 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
4731 (lshr_simd): DI mode added.
4732 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
4733 (aarch64_ushr_simddi): Likewise.
4734 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
4735 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
4736 (vshrd_n_u64): Likewise.
4738 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4740 * Makefile.in (s-macro_list): Depend on cc1.
4742 2014-03-23 Teresa Johnson <tejohnson@google.com>
4744 * ipa-utils.c (ipa_print_order): Use specified dump file.
4746 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
4748 PR rtl-optimization/60601
4749 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
4751 * gcc.c (eval_spec_function): Initialize save_growing_value.
4753 2014-03-22 Jakub Jelinek <jakub@redhat.com>
4756 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
4757 code == MINUS_EXPR, never swap op0 with op1.
4759 * toplev.c (init_local_tick): Avoid signed integer multiplication
4761 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
4762 shift by first operand's bitsize.
4764 2014-03-21 Jakub Jelinek <jakub@redhat.com>
4767 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
4769 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
4770 TARGET_ISA_64BIT_P(x).
4772 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4774 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
4775 pattern for vector nor instead of subtract from splat(-1).
4776 (altivec_expand_vec_perm_const_le): Likewise.
4778 2014-03-21 Richard Henderson <rth@twiddle.net>
4781 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
4782 related insns after epilogue_completed.
4784 2014-03-21 Martin Jambor <mjambor@suse.cz>
4787 * cgraph.h (symtab_node): New flag body_removed.
4788 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
4789 when removing bodies.
4790 * symtab.c (dump_symtab_base): Dump body_removed flag.
4791 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
4792 had their bodies removed.
4794 2014-03-21 Martin Jambor <mjambor@suse.cz>
4797 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
4800 2014-03-21 Richard Biener <rguenther@suse.de>
4802 PR tree-optimization/60577
4803 * tree-core.h (struct tree_base): Document nothrow_flag use
4804 in VAR_DECL_NONALIASED.
4805 * tree.h (VAR_DECL_NONALIASED): New.
4806 (may_be_aliased): Adjust.
4807 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
4809 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4811 * expr.c (expand_expr_real_1): Remove outdated comment.
4813 2014-03-20 Jakub Jelinek <jakub@redhat.com>
4816 * ira.c (adjust_cleared_regs): Call copy_rtx on
4817 *reg_equiv[REGNO (loc)].src_p before passing it to
4818 simplify_replace_fn_rtx.
4821 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
4822 into CONST, put pic register as first operand of PLUS. Use
4823 gen_const_mem for both 32-bit and 64-bit PIC got loads.
4825 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4827 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
4829 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4831 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
4832 around for store forwarding issue in the FPU on the UT699.
4833 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
4834 loads and operations if -mfix-ut699 is specified.
4835 (divtf3_hq): Tweak attribute.
4836 (sqrttf2_hq): Likewise.
4838 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4840 * calls.c (store_one_arg): Remove incorrect const qualification on the
4841 type of the temporary.
4842 * cfgexpand.c (expand_return): Likewise.
4843 * expr.c (expand_constructor): Likewise.
4844 (expand_expr_real_1): Likewise.
4846 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4848 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
4851 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
4854 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
4856 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
4858 * config/arm/aarch-common-protos.h
4859 (alu_cost_table): Fix spelling of "extend".
4860 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
4862 2014-03-19 Richard Biener <rguenther@suse.de>
4865 * tree-core.h (tree_type_common): Re-order pointer members
4866 to reduce recursion depth during GC walks.
4868 2014-03-19 Marek Polacek <polacek@redhat.com>
4871 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
4872 before accessing it.
4874 2014-03-19 Richard Biener <rguenther@suse.de>
4877 * lto-streamer-in.c (input_function): In WPA stage do not drop
4880 2014-03-19 Jakub Jelinek <jakub@redhat.com>
4882 PR tree-optimization/60559
4883 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
4884 with build_zero_cst assignment.
4886 2014-03-18 Kai Tietz <ktietz@redhat.com>
4888 PR rtl-optimization/56356
4889 * sdbout.c (sdbout_parms): Verify that parms'
4890 incoming argument is valid.
4891 (sdbout_reg_parms): Likewise.
4893 2014-03-18 Richard Henderson <rth@redhat.com>
4896 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
4897 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
4898 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
4900 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
4902 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
4903 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
4904 Italicize plugin event names in description. Explain that
4905 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
4906 Remind that no GCC functions should be called after PLUGIN_FINISH.
4907 Explain what pragmas with expansion are.
4909 2014-03-18 Martin Liska <mliska@suse.cz>
4911 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
4912 gimple call statement is update.
4913 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
4914 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
4916 2014-03-18 Jakub Jelinek <jakub@redhat.com>
4919 * ubsan.c (ubsan_instrument_unreachable): Call
4920 initialize_sanitizer_builtins.
4921 (ubsan_pass): Likewise.
4924 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
4925 varpool_finalize_decl instead of rest_of_decl_compilation.
4927 2014-03-18 Richard Biener <rguenther@suse.de>
4929 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
4930 by using bitmap_and_compl instead of bitmap_and_compl_into.
4931 (df_rd_transfer_function): Likewise.
4933 2014-03-18 Richard Biener <rguenther@suse.de>
4935 * doc/lto.texi (fresolution): Fix typo.
4937 2014-03-18 Richard Biener <rguenther@suse.de>
4939 * doc/invoke.texi (flto): Update for changes in 4.9.
4941 2014-03-18 Richard Biener <rguenther@suse.de>
4943 * doc/loop.texi: Remove section on the removed lambda framework.
4944 Update loop docs with recent changes in preserving loop structure.
4946 2014-03-18 Richard Biener <rguenther@suse.de>
4948 * doc/lto.texi (-fresolution): Document.
4950 2014-03-18 Richard Biener <rguenther@suse.de>
4952 * doc/contrib.texi: Adjust my name.
4954 2014-03-18 Jakub Jelinek <jakub@redhat.com>
4957 * internal-fn.c: Include diagnostic-core.h.
4958 (expand_BUILTIN_EXPECT): New function.
4959 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
4960 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
4961 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
4962 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
4964 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
4965 Revert 3 argument __builtin_expect code.
4966 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
4967 * gimple-fold.c (gimple_fold_call): Likewise.
4968 * tree.h (fold_builtin_expect): New prototype.
4969 * builtins.c (build_builtin_expect_predicate): Add predictor
4970 argument, if non-NULL, create 3 argument __builtin_expect.
4971 (fold_builtin_expect): No longer static. Add ARG2 argument,
4972 pass it through to build_builtin_expect_predicate.
4973 (fold_builtin_2): Adjust caller.
4974 (fold_builtin_3): Handle BUILT_IN_EXPECT.
4975 * internal-fn.def (BUILTIN_EXPECT): New.
4977 2014-03-18 Tobias Burnus <burnus@net-b.de>
4980 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
4981 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
4982 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
4984 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
4987 * predict.c (combine_predictions_for_bb): Fix up formatting.
4988 (expr_expected_value_1, expr_expected_value): Add predictor argument,
4989 fill what it points to if non-NULL.
4990 (tree_predict_by_opcode): Adjust caller, use the predictor.
4991 * predict.def (PRED_COMPARE_AND_SWAP): Add.
4993 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
4995 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
4996 proper constant for the store mode.
4998 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
5000 * symtab.c (change_decl_assembler_name): Fix transparent alias
5003 2014-03-16 Renlin Li <Renlin.Li@arm.com>
5005 * config/aarch64/aarch64.c: Correct the comments about the
5006 aarch64 stack layout.
5008 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
5010 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
5011 check for GF_OMP_FOR_KIND_FOR.
5013 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
5015 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
5016 ymm and zmm register names.
5018 2014-03-17 Jakub Jelinek <jakub@redhat.com>
5021 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
5022 note creation for the 2010-08-31 changes.
5024 2014-03-17 Marek Polacek <polacek@redhat.com>
5027 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
5028 as -fno-tree-loop-vectorize.
5029 (expand_omp_simd): Likewise.
5031 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
5033 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
5034 (eligible_for_call_delay): New prototype.
5035 * config/sparc/sparc.c (tls_call_delay): Rename into...
5036 (eligible_for_call_delay): ...this. Return false if the instruction
5037 cannot be put in the delay slot of a branch.
5038 (eligible_for_restore_insn): Simplify.
5039 (eligible_for_return_delay): Return false if the instruction cannot be
5040 put in the delay slot of a branch and simplify.
5041 (eligible_for_sibcall_delay): Return false if the instruction cannot be
5042 put in the delay slot of a branch.
5043 * config/sparc/sparc.md (fix_ut699): New attribute.
5044 (tls_call_delay): Delete.
5045 (in_call_delay): Reimplement.
5046 (eligible_for_sibcall_delay): Rename into...
5047 (in_sibcall_delay): ...this.
5048 (eligible_for_return_delay): Rename into...
5049 (in_return_delay): ...this.
5050 (in_branch_delay): Reimplement.
5051 (in_uncond_branch_delay): Delete.
5052 (in_annul_branch_delay): Delete.
5054 2014-03-14 Richard Henderson <rth@redhat.com>
5057 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
5058 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
5059 (*floathi<X87MODEF>2_i387_with_temp): Remove.
5060 (floathi splitters): Remove.
5061 (float<SWI48x>xf2): New pattern.
5062 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
5063 code that tried to handle DImode for 32-bit, but which was excluded
5064 by the pattern's condition. Drop allocation of stack temporary.
5065 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
5066 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
5067 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
5068 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
5069 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
5070 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
5071 (*float<SWI48><MODEF>2_sse_interunit): Remove.
5072 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
5073 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
5074 (*float<SWI48x><X87MODEF>2_i387): Remove.
5075 (all float _with_temp splitters): Remove.
5076 (*float<SWI48x><MODEF>2_i387): New pattern.
5077 (*float<SWI48><MODEF>2_sse): New pattern.
5078 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
5079 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
5081 2014-03-14 Jakub Jelinek <jakub@redhat.com>
5082 Marek Polacek <polacek@redhat.com>
5085 * common.opt (dump_base_name_prefixed): New Variable.
5086 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
5087 if x_dump_base_name_prefixed is already set, set it at the end.
5089 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
5091 PR rtl-optimization/60508
5092 * lra-constraints.c (get_reload_reg): Add new parameter
5094 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
5095 Pass the new parameter values.
5097 2014-03-14 Richard Biener <rguenther@suse.de>
5099 * common.opt: Revert unintented changes from r205065.
5102 2014-03-14 Richard Biener <rguenther@suse.de>
5105 * cfghooks.c (split_block): Properly adjust all loops the
5106 block was a latch of.
5108 2014-03-14 Martin Jambor <mjambor@suse.cz>
5111 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
5114 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
5117 * config/avr/avr.c (avr_set_current_function): Pass function name
5118 through default_strip_name_encoding before sanity checking instead
5119 of skipping the first char of the assembler name.
5121 2014-03-13 Richard Henderson <rth@redhat.com>
5124 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
5125 (ix86_force_to_memory, ix86_free_from_memory): Remove.
5126 * config/i386/i386-protos.h: Likewise.
5127 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
5128 in the expander instead of a splitter.
5129 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
5130 any possibility of requiring a memory.
5131 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
5132 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
5133 (fp branch splitters): Update for ix86_split_fp_branch.
5134 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
5135 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
5136 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
5137 (*fop_<MODEF>_2_i387): Remove f/r alternative.
5138 (*fop_<MODEF>_3_i387): Likewise.
5139 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
5140 (splitters for the fop_* register patterns): Remove.
5141 (fscalexf4_i387): Rename from *fscalexf4_i387.
5142 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
5144 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5146 PR tree-optimization/59779
5147 * tree-dfa.c (get_ref_base_and_extent): Use double_int
5148 type for bitsize and maxsize instead of HOST_WIDE_INT.
5150 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
5152 PR rtl-optimization/57320
5153 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
5154 the CFG after thread_prologue_and_epilogue_insns.
5156 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
5158 PR rtl-optimization/57189
5159 * lra-constraints.c (process_alt_operands): Disfavor spilling
5162 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
5164 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
5166 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5168 PR tree-optimization/59025
5170 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
5171 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
5173 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
5176 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
5177 calls of avr_out_plus_1.
5179 2014-03-13 Bin Cheng <bin.cheng@arm.com>
5181 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
5182 BB's single pred and update the father loop's latch info later.
5184 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5186 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
5190 (VEC_base): Likewise.
5191 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
5192 registers, we need to swap double words in little endian mode.
5194 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
5195 to be a container mode for 128-bit integer operations added in ISA
5196 2.07. Unlike TImode and PTImode, the preferred register set is
5197 the Altivec/VMX registers for the 128-bit operations.
5199 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
5201 (rs6000_split_128bit_ok_p): Likewise.
5203 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
5204 macros for creating ISA 2.07 normal and overloaded builtin
5205 functions with 3 arguments.
5206 (BU_P8V_OVERLOAD_3): Likewise.
5207 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
5208 for use as overloaded functions.
5209 (VPERM_1TI_UNS): Likewise.
5210 (VSEL_1TI): Likewise.
5211 (VSEL_1TI_UNS): Likewise.
5212 (ST_INTERNAL_1ti): Likewise.
5213 (LD_INTERNAL_1ti): Likewise.
5214 (XXSEL_1TI): Likewise.
5215 (XXSEL_1TI_UNS): Likewise.
5216 (VPERM_1TI): Likewise.
5217 (VPERM_1TI_UNS): Likewise.
5218 (XXPERMDI_1TI): Likewise.
5219 (SET_1TI): Likewise.
5220 (LXVD2X_V1TI): Likewise.
5221 (STXVD2X_V1TI): Likewise.
5222 (VEC_INIT_V1TI): Likewise.
5223 (VEC_SET_V1TI): Likewise.
5224 (VEC_EXT_V1TI): Likewise.
5225 (EQV_V1TI): Likewise.
5226 (NAND_V1TI): Likewise.
5227 (ORC_V1TI): Likewise.
5228 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
5229 added in ISA 2.07. Add both normal 'altivec' builtins, and the
5231 (VADDUQM): Likewise.
5232 (VSUBCUQ): Likewise.
5233 (VADDEUQM): Likewise.
5234 (VADDECUQ): Likewise.
5235 (VSUBEUQM): Likewise.
5236 (VSUBECUQ): Likewise.
5238 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5239 __int128_t and __uint128_t types.
5240 (__uint128_type): Likewise.
5241 (altivec_categorize_keyword): Add support for vector __int128_t,
5242 vector __uint128_t, vector __int128, and vector unsigned __int128
5243 as a container type for TImode operations that need to be done in
5244 VSX/Altivec registers.
5245 (rs6000_macro_to_expand): Likewise.
5246 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5247 to support 128-bit integer instructions vaddcuq, vadduqm,
5248 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5249 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5251 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5252 for V1TImode, and set up preferences to use VSX/Altivec registers.
5253 Setup VSX reload handlers.
5254 (rs6000_debug_reg_global): Likewise.
5255 (rs6000_init_hard_regno_mode_ok): Likewise.
5256 (rs6000_preferred_simd_mode): Likewise.
5257 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5258 (easy_altivec_constant): Likewise.
5259 (output_vec_const_move): Likewise.
5260 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5262 (rs6000_expand_vector_extract): Likewise.
5263 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
5265 (rs6000_const_vec): Add support for V1TImode.
5266 (rs6000_emit_le_vsx_load): Swap double words when loading or
5267 storing TImode/V1TImode.
5268 (rs6000_emit_le_vsx_store): Likewise.
5269 (rs6000_emit_le_vsx_move): Likewise.
5270 (rs6000_emit_move): Add support for V1TImode.
5271 (altivec_expand_ld_builtin): Likewise.
5272 (altivec_expand_st_builtin): Likewise.
5273 (altivec_expand_vec_init_builtin): Likewise.
5274 (altivec_expand_builtin): Likewise.
5275 (rs6000_init_builtins): Add support for V1TImode type. Add
5276 support for ISA 2.07 128-bit integer builtins. Define type names
5277 for the VSX/Altivec vector types.
5278 (altivec_init_builtins): Add support for overloaded vector
5279 functions with V1TImode type.
5280 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
5281 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
5283 (rs6000_split_128bit_ok_p): Likewise.
5284 (rs6000_handle_altivec_attribute): Create V1TImode from vector
5285 __int128_t and vector __uint128_t.
5287 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
5288 and mode attributes.
5295 (VS_scalar): Likewise.
5296 (VS_double): Likewise.
5297 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
5299 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
5300 we support the ISA 2.07 128-bit integer arithmetic instructions.
5301 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
5302 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
5303 and TImode types for use with the builtin functions.
5304 (V1TI_type_node): Likewise.
5305 (unsigned_V1TI_type_node): Likewise.
5306 (intTI_type_internal_node): Likewise.
5307 (uintTI_type_internal_node): Likewise.
5309 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
5310 128-bit builtin functions.
5311 (UNSPEC_VADDEUQM): Likewise.
5312 (UNSPEC_VADDECUQ): Likewise.
5313 (UNSPEC_VSUBCUQ): Likewise.
5314 (UNSPEC_VSUBEUQM): Likewise.
5315 (UNSPEC_VSUBECUQ): Likewise.
5316 (VM): Add V1TImode to vector mode iterators.
5318 (VI_unit): Likewise.
5319 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
5320 (altivec_vaddcuq): Likewise.
5321 (altivec_vsubuqm): Likewise.
5322 (altivec_vsubcuq): Likewise.
5323 (altivec_vaddeuqm): Likewise.
5324 (altivec_vaddecuq): Likewise.
5325 (altivec_vsubeuqm): Likewise.
5326 (altivec_vsubecuq): Likewise.
5328 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
5330 (BOOL_128): Likewise.
5331 (BOOL_REGS_OUTPUT): Likewise.
5332 (BOOL_REGS_OP1): Likewise.
5333 (BOOL_REGS_OP2): Likewise.
5334 (BOOL_REGS_UNARY): Likewise.
5335 (BOOL_REGS_AND_CR0): Likewise.
5337 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
5338 128-bit integer builtin support.
5339 (vec_vadduqm): Likewise.
5340 (vec_vaddecuq): Likewise.
5341 (vec_vaddeuqm): Likewise.
5342 (vec_vsubecuq): Likewise.
5343 (vec_vsubeuqm): Likewise.
5344 (vec_vsubcuq): Likewise.
5345 (vec_vsubuqm): Likewise.
5347 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5348 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
5349 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
5350 128-bit integer add/subtract to ISA 2.07.
5352 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
5354 * config/arc/arc.c (arc_predicate_delay_insns):
5355 Fix third argument passed to conditionalize_nonjump.
5357 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
5359 * config/aarch64/aarch64-builtins.c
5360 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
5361 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
5362 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
5363 instead of __builtin_lfloor.
5364 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
5366 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5368 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
5369 (tree_ssa_ifcombine_bb_1): New function.
5370 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
5371 is an empty forwarder block to then_bb or vice versa and then_bb
5372 and else_bb are effectively swapped.
5374 2014-03-12 Christian Bruel <christian.bruel@st.com>
5377 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
5378 REG_CFA_DEF_CFA note.
5379 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
5380 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
5382 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
5384 PR tree-optimization/60454
5385 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
5387 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5389 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
5390 Do not define target_cpu_default2 to generic.
5391 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
5392 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
5393 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
5395 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5396 Marc Glisse <marc.glisse@inria.fr>
5398 PR tree-optimization/60502
5399 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
5400 instead of build_low_bits_mask.
5402 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5405 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
5406 if there are multiple uses, but op doesn't live on E edge.
5407 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
5408 clobber stmts before __builtin_unreachable.
5410 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
5412 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
5413 hard_frame_pointer_rtx.
5414 * cse.c (cse_insn): Remove volatile check.
5415 * cselib.c (cselib_process_insn): Likewise.
5416 * dse.c (scan_insn): Likewise.
5418 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5420 * config/arc/arc.c (conditionalize_nonjump): New function,
5422 (arc_ifcvt): ... this.
5423 (arc_predicate_delay_insns): Use it.
5425 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5427 * config/arc/predicates.md (extend_operand): During/after reload,
5428 allow const_int_operand.
5429 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
5430 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
5431 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
5433 (umulsi3_highpart_i): Likewise.
5435 2014-03-11 Richard Biener <rguenther@suse.de>
5437 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
5438 Add asserts to guard possible wrong-code bugs.
5440 2014-03-11 Richard Biener <rguenther@suse.de>
5442 PR tree-optimization/60429
5443 PR tree-optimization/60485
5444 * tree-ssa-structalias.c (set_union_with_increment): Properly
5445 take into account all fields that overlap the shifted vars.
5446 (do_sd_constraint): Likewise.
5447 (do_ds_constraint): Likewise.
5448 (get_constraint_for_ptr_offset): Likewise.
5450 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
5452 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
5453 (nios2_compute_frame_layout):
5454 Add calculation of cfun->machine->fp_save_offset.
5455 (nios2_expand_prologue): Correct setting of frame pointer register
5457 (nios2_expand_epilogue): Update recovery of stack pointer from
5458 frame pointer accordingly.
5459 (nios2_initial_elimination_offset): Update calculation of offset
5460 for eliminating to HARD_FRAME_POINTER_REGNUM.
5462 2014-03-10 Jakub Jelinek <jakub@redhat.com>
5465 * ipa.c (symtab_remove_unreachable_nodes): Don't call
5466 cgraph_get_create_node on VAR_DECLs.
5468 2014-03-10 Richard Biener <rguenther@suse.de>
5471 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
5473 2014-03-08 Douglas B Rupp <rupp@gnat.com>
5475 * config/vms/vms.opt (vms_float_format): New variable.
5477 2014-03-08 Tobias Burnus <burnus@net-b.de>
5479 * doc/invoke.texi (-fcilkplus): Update implementation status.
5481 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
5482 Richard Biener <rguenther@suse.de>
5484 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
5485 consistently accross all TUs.
5486 (run_gcc): Enable -fshort-double automatically at link at link-time
5487 and disallow override.
5489 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
5492 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
5493 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
5494 if they can't be used.
5496 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5498 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
5499 for Solaris 11/x86 ld.
5500 * configure: Regenerate.
5502 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5504 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
5505 (LIB_TLS_SPEC): Save as ld_tls_libs.
5506 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
5507 (HAVE_AS_IX86_TLSLDM): New test.
5508 * configure, config.in: Regenerate.
5509 * config/i386/i386.c (legitimize_tls_address): Fall back to
5510 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
5511 cannot support TLS_MODEL_LOCAL_DYNAMIC.
5512 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
5513 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
5515 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
5517 * common.opt (fira-loop-pressure): Mark as optimization.
5519 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
5521 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
5522 an OpenMP mappable type.
5524 2014-03-06 Matthias Klose <doko@ubuntu.com>
5526 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
5527 MULTILIB_OSDIRNAMES is not defined.
5529 2014-03-06 Jakub Jelinek <jakub@redhat.com>
5530 Meador Inge <meadori@codesourcery.com>
5533 * config/arm/arm.c (arm_tls_symbol_p): Remove.
5534 (arm_legitimize_address): Call legitimize_tls_address for any
5535 arm_tls_referenced_p expression, handle constant addend. Call it
5536 before testing for !TARGET_ARM.
5537 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
5539 2014-03-06 Richard Biener <rguenther@suse.de>
5545 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5547 * tree-streamer.c (record_common_node): Assert we don't record
5548 nodes with type double.
5549 (preload_common_node): Skip type double, complex double and double
5550 pointer since it is now frontend dependent due to fshort-double option.
5552 2014-03-06 Richard Biener <rguenther@suse.de>
5554 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
5555 or -fno-lto is specified and the linker has full plugin support.
5556 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
5557 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
5558 * lto-wrapper.c (merge_and_complain): Merge compile-time
5559 optimization levels.
5560 (run_gcc): And pass it through to the link options.
5562 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
5566 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5568 * cselib.c (remove_useless_values): Skip to avoid quadratic
5569 behavior if the condition moved from...
5570 (cselib_process_insn): ... here holds.
5572 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5575 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
5576 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
5579 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
5580 (TM_H): Add x86-tune.def.
5582 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5584 * config/aarch64/aarch64.c (generic_tunings):
5585 Use cortexa57_extra_costs.
5587 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5590 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
5591 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
5592 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
5595 2014-03-04 Heiher <r@hev.cc>
5597 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
5598 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
5600 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
5602 * config/i386/predicates.md (const2356_operand): Change to ...
5603 (const2367_operand): ... this.
5604 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
5606 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5607 (*avx512pf_scatterpf<mode>sf): Ditto.
5608 (avx512pf_scatterpf<mode>df): Ditto.
5609 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5610 (*avx512pf_scatterpf<mode>df): Ditto.
5611 * config/i386/i386.c (ix86_expand_builtin): Update
5612 incorrect hint operand error message.
5614 2014-03-04 Richard Biener <rguenther@suse.de>
5616 * lto-section-in.c (lto_get_section_data): Fix const cast.
5618 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5620 * tree-streamer.c (record_common_node): Assert we don't record
5621 nodes with type double.
5622 (preload_common_node): Skip type double, complex double and double
5623 pointer since it is now frontend dependent due to fshort-double option.
5625 2014-03-04 Richard Biener <rguenther@suse.de>
5628 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
5629 (lto_input_toplevel_asms): Likewise.
5630 * lto-section-in.c (lto_get_section_data): Instead do it here
5633 2014-03-04 Richard Biener <rguenther@suse.de>
5635 PR tree-optimization/60382
5636 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
5637 dead PHIs a reduction.
5639 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
5641 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
5643 (_mm_prefetch): Move out of GCC target("sse") pragma.
5644 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
5645 GCC target("prfchw") pragma.
5646 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
5648 * config/i386/i386.c (ix86_option_override_internal): Enable
5649 -mprfchw with -mprefetchwt1.
5651 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5653 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
5656 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5658 * opts.h (CL_PCH_IGNORE): Define.
5659 * targhooks.c (option_affects_pch_p):
5660 Return false for options that have CL_PCH_IGNORE set.
5661 * opt-functions.awk: Process PchIgnore.
5662 * doc/options.texi: Document PchIgnore.
5664 * config/arc/arc.opt (misize): Add PchIgnore property.
5666 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5668 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
5669 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
5670 constraint on constants to permit them being loaded into
5671 GENERAL_REGS or BASE_REGS.
5673 2014-03-03 Nick Clifton <nickc@redhat.com>
5675 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
5676 anti-cacnonical alternatives.
5677 (negandhi3_real): New pattern.
5678 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
5680 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5682 * config/avr/avr-mcus.def: Remove atxmega16x1.
5683 * config/avr/avr-tables.opt: Regenerate.
5684 * config/avr/t-multilib: Regenerate.
5685 * doc/avr-mmcu.texi: Regenerate.
5687 2014-03-03 Tobias Grosser <tobias@grosser.es>
5688 Mircea Namolaru <mircea.namolaru@inria.fr>
5690 PR tree-optimization/58028
5691 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
5694 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5696 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
5697 not handled by recognizers.
5699 2014-03-03 Jakub Jelinek <jakub@redhat.com>
5702 * function.c (expand_function_end): Don't emit
5703 clobber_return_register sequence if clobber_after is a BARRIER.
5704 * cfgexpand.c (construct_exit_block): Append instructions before
5705 return_label to prev_bb.
5707 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5709 * config/rs6000/constraints.md: Document reserved use of "wc".
5711 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5714 * ipa.c (function_and_variable_visibility): When dissolving comdat
5715 group, also set all symbols to local.
5717 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5722 2013-12-14 Jan Hubicka <jh@suse.cz>
5724 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
5726 2014-03-02 Jon Beniston <jon@beniston.com>
5732 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
5733 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
5734 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
5735 (simple_return, *simple_return): New patterns
5736 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
5737 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
5739 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
5741 * dwarf2out.c (gen_subprogram_die): Tidy.
5743 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
5746 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
5747 (*mov_t_msb_neg_negc): ... this new insn.
5749 2014-02-28 Jason Merrill <jason@redhat.com>
5752 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
5755 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
5758 * dwarf2out.c (decltype_auto_die): New static.
5759 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
5760 (gen_type_die_with_usage): Handle 'decltype(auto)'.
5761 (is_cxx_auto): Likewise.
5763 2014-02-28 Ian Bolton <ian.bolton@arm.com>
5765 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
5766 we are not using general regs only.
5768 2014-02-28 Richard Biener <rguenther@suse.de>
5771 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
5772 previous fix and only allow to remove trivial pre-headers
5773 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
5774 (remove_forwarder_block): Properly update the latch of a loop.
5776 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5779 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
5780 (cselib_preserved_hash_table): New.
5781 (preserve_constants_and_equivs): Move preserved vals to it.
5782 (cselib_find_slot): Look it up first.
5783 (cselib_init): Initialize it.
5784 (cselib_finish): Release it.
5785 (dump_cselib_table): Dump it.
5787 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5790 * cselib.c (remove_useless_values): Skip to avoid quadratic
5791 behavior if the condition moved from...
5792 (cselib_process_insn): ... here holds.
5794 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5797 * var-tracking.c (vt_initialize): Apply the same condition to
5798 preserve the CFA base value.
5800 2014-02-28 Joey Ye <joey.ye@arm.com>
5803 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
5804 if reload in progress or completed.
5806 2014-02-28 Tobias Burnus <burnus@net-b.de>
5809 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
5812 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
5814 * doc/tm.texi.in (Condition Code Status): Update documention for
5815 relative locations of cc0-setter and cc0-user.
5817 2014-02-27 Jeff Law <law@redhat.com>
5819 PR rtl-optimization/52714
5820 * combine.c (try_combine): When splitting an unrecognized PARALLEL
5821 into two independent simple sets, if I3 is a jump, ensure the
5822 pattern we place into I3 is a (set (pc) ...).
5824 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
5825 Jeff Law <law@redhat.com>
5827 PR rtl-optimization/49847
5828 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
5829 are in different blocks.
5830 * doc/tm.texi (Condition Code Status): Update documention for
5831 relative locations of cc0-setter and cc0-user.
5833 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
5836 * lra.c (lra_emit_add): Check SUBREG too.
5838 2014-02-27 Andreas Schwab <schwab@suse.de>
5840 * config/m68k/m68k.c (m68k_option_override): Disable
5841 -flive-range-shrinkage for classic m68k.
5842 (m68k_override_options_after_change): Likewise.
5844 2014-02-27 Marek Polacek <polacek@redhat.com>
5847 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
5848 -Wmaybe-uninitialized.
5850 2014-02-27 Alan Modra <amodra@gmail.com>
5853 * reload1.c (emit_input_reload_insns): When reload_override_in,
5854 set old to rl->in_reg when rl->in_reg is a subreg.
5856 2014-02-26 Richard Biener <rguenther@suse.de>
5859 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
5861 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
5863 * common/config/i386/predicates.md (const1256_operand): Remove.
5864 (const2356_operand): New.
5865 (const_1_to_2_operand): Remove.
5866 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5867 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5868 (*avx512pf_gatherpf<mode>sf): Ditto.
5869 (avx512pf_gatherpf<mode>df): Ditto.
5870 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5871 (*avx512pf_gatherpf<mode>df): Ditto.
5872 (avx512pf_scatterpf<mode>sf): Ditto.
5873 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5874 (*avx512pf_scatterpf<mode>sf): Ditto.
5875 (avx512pf_scatterpf<mode>df): Ditto.
5876 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5877 (*avx512pf_scatterpf<mode>df): Ditto.
5878 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
5880 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
5882 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
5883 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
5884 (_mm512_mask_testn_epi64_mask): Move to ...
5885 * config/i386/avx512cdintrin.h: Here.
5886 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
5887 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
5888 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
5889 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
5890 TARGET_AVX512F from TARGET_AVX512CD.
5892 2014-02-26 Richard Biener <rguenther@suse.de>
5895 * ipa.c (walk_polymorphic_call_targets): Properly guard
5896 call to inline_update_overall_summary.
5898 2014-02-26 Bin Cheng <bin.cheng@arm.com>
5901 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
5902 and latches only if requested. Fix latch if it is removed.
5903 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
5904 LOOPS_HAVE_PREHEADERS.
5906 2014-02-25 Andrew Pinski <apinski@cavium.com>
5908 * builtins.c (expand_builtin_thread_pointer): Create a new target
5909 when the target is NULL.
5911 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
5913 PR rtl-optimization/60317
5914 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
5915 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
5916 * lra-assigns.c: Include params.h.
5917 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
5918 other reload pseudos considerations.
5920 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5922 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
5923 to use canonical form for nor<mode>3.
5925 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5928 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
5931 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
5933 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
5934 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
5935 (ix86_handle_option): Handle OPT_mprefetchwt1.
5936 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
5937 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
5939 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
5940 OPTION_MASK_ISA_PREFETCHWT1.
5941 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
5942 (PTA_PREFETCHWT1): New.
5943 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
5944 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
5945 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
5946 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
5947 (*prefetch_avx512pf_<mode>_: Change into ...
5948 (*prefetch_prefetchwt1_<mode>: This.
5949 * config/i386/i386.opt (mprefetchwt1): New.
5950 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
5951 (_mm_prefetch): Handle intent to write.
5952 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
5954 2014-02-25 Richard Biener <rguenther@suse.de>
5957 * emit-rtl.c (mem_attrs_htab): Remove.
5958 (mem_attrs_htab_hash): Likewise.
5959 (mem_attrs_htab_eq): Likewise.
5960 (set_mem_attrs): Always allocate new mem-attrs when something changed.
5961 (init_emit_once): Do not allocate mem_attrs_htab.
5963 2014-02-25 Richard Biener <rguenther@suse.de>
5966 * lto-opts.c (lto_write_options): Output non-explicit conservative
5967 -fwrapv, -fno-trapv and -fno-strict-overflow.
5968 * lto-wrapper.c (merge_and_complain): Handle merging those options.
5969 (run_gcc): And pass them through.
5971 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
5973 * sel-sched.c (calculate_new_fences): New parameter ptime.
5974 Calculate it as a maximum over all fence cycles.
5975 (sel_sched_region_2): Adjust the call to calculate_new_fences.
5976 Print the final schedule timing when sched_verbose.
5978 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
5980 PR rtl-optimization/60292
5981 * sel-sched.c (fill_vec_av_set): Do not reset target availability
5982 bit fot the fence instruction.
5984 2014-02-24 Alangi Derick <alangiderick@gmail.com>
5986 * calls.h: Fix typo in comment.
5988 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
5990 * config/pa/pa.c (pa_output_move_double): Don't valididate when
5991 adjusting offsetable addresses.
5993 2014-02-24 Guozhi Wei <carrot@google.com>
5995 * sparseset.h (sparseset_pop): Fix the wrong index.
5997 2014-02-24 Walter Lee <walt@tilera.com>
5999 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
6000 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
6002 * common/config/tilegx/tilegx-common.c
6003 (TARGET_DEFAULT_TARGET_FLAGS): Define.
6004 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
6006 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
6007 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
6008 (tilegx_gimplify_va_arg_expr): Handle big endian.
6009 (tilegx_expand_unaligned_load): Ditto.
6010 (tilegx_expand_unaligned_store): Ditto.
6011 (TARGET_RETURN_IN_MSB): New.
6012 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
6013 (TARGET_ENDIAN_DEFAULT): New.
6014 (TARGET_BIG_ENDIAN): Handle big endian.
6015 (BYTES_BIG_ENDIAN): Ditto.
6016 (WORDS_BIG_ENDIAN): Ditto.
6017 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
6020 * config/tilegx/tilegx.md (extv): Handle big endian.
6022 (insn_st<n>): Ditto.
6023 (insn_st<n>_add<bitsuffix>): Ditto.
6024 (insn_stnt<n>): Ditto.
6025 (insn_stnt<n>_add<bitsuffix>):Ditto.
6026 (vec_interleave_highv8qi): Handle big endian.
6027 (vec_interleave_highv8qi_be): New.
6028 (vec_interleave_highv8qi_le): New.
6029 (insn_v1int_h): Handle big endian.
6030 (vec_interleave_lowv8qi): Handle big endian.
6031 (vec_interleave_lowv8qi_be): New.
6032 (vec_interleave_lowv8qi_le): New.
6033 (insn_v1int_l): Handle big endian.
6034 (vec_interleave_highv4hi): Handle big endian.
6035 (vec_interleave_highv4hi_be): New.
6036 (vec_interleave_highv4hi_le): New.
6037 (insn_v2int_h): Handle big endian.
6038 (vec_interleave_lowv4hi): Handle big endian.
6039 (vec_interleave_lowv4hi_be): New.
6040 (vec_interleave_lowv4hi_le): New.
6041 (insn_v2int_l): Handle big endian.
6042 (vec_interleave_highv2si): Handle big endian.
6043 (vec_interleave_highv2si_be): New.
6044 (vec_interleave_highv2si_le): New.
6045 (insn_v4int_h): Handle big endian.
6046 (vec_interleave_lowv2si): Handle big endian.
6047 (vec_interleave_lowv2si_be): New.
6048 (vec_interleave_lowv2si_le): New.
6049 (insn_v4int_l): Handle big endian.
6050 * config/tilegx/tilegx.opt (mbig-endian): New option.
6051 (mlittle-endian): New option.
6052 * doc/install.texi: Document tilegxbe-linux.
6053 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
6055 2014-02-24 Martin Jambor <mjambor@suse.cz>
6058 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
6059 there are no parameter descriptors.
6061 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
6063 PR rtl-optimization/60268
6064 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
6065 initialization to ...
6066 (sched_rgn_init): ... here.
6067 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
6069 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6071 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
6074 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
6076 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
6079 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6081 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
6082 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
6084 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6086 * config/microblaze/predicates.md: Add cmp_op predicate.
6087 * config/microblaze/microblaze.md: Add branch_compare instruction
6088 which uses cmp_op predicate and emits cmp insn before branch.
6089 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
6090 to microblaze_expand_conditional_branch and consolidate logic.
6091 (microblaze_expand_conditional_branch): emit branch_compare
6092 insn instead of handling cmp op separate from branch insn.
6094 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6096 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
6099 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6101 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
6102 define_insn with define_expand and new define_insn
6103 *altivec_lve<VI_char>x_internal.
6104 (altivec_stve<VI_char>x): Replace define_insn with define_expand
6105 and new define_insn *altivec_stve<VI_char>x_internal.
6106 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
6108 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
6110 (altivec_expand_stvex_be): New function.
6112 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
6114 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
6115 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
6116 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
6117 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
6119 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
6122 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
6123 instead of emit_move_insn.
6125 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6127 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
6129 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
6130 gen_altivec_vsumsws.
6132 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6134 * config/rs6000/altivec.md (altivec_lvxl): Rename as
6135 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
6136 (altivec_lvxl_<mode>): New define_expand incorporating
6137 -maltivec=be semantics where needed.
6138 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
6139 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
6140 semantics where needed.
6141 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
6142 (altivec_stvx_<mode>): New define_expand incorporating
6143 -maltivec=be semantics where needed.
6144 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
6145 VM2 iterator instead of V4SI.
6146 (altivec_stvxl_<mode>): New define_expand incorporating
6147 -maltivec=be semantics where needed.
6148 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
6149 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
6150 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
6151 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
6152 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
6153 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
6154 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
6155 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
6156 ALTIVEC_BUILTIN_STVXL.
6157 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
6158 (altivec_expand_stvx_be): Likewise.
6159 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
6160 (altivec_expand_lvx_be): Likewise.
6161 (altivec_expand_stvx_be): Likewise.
6162 (altivec_expand_builtin): Add cases for
6163 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
6164 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
6165 (altivec_init_builtins): Add definitions for
6166 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
6167 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
6169 2014-02-21 Catherine Moore <clm@codesourcery.com>
6171 * doc/invoke.texi (mvirt, mno-virt): Document.
6172 * config/mips/mips.opt (mvirt): New option.
6173 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
6175 2014-02-21 Richard Biener <rguenther@suse.de>
6177 PR tree-optimization/60276
6178 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
6179 (STMT_VINFO_MIN_NEG_DIST): New macro.
6180 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
6181 STMT_VINFO_MIN_NEG_DIST.
6182 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
6183 made for negative dependence distances still hold.
6185 2014-02-21 Richard Biener <rguenther@suse.de>
6188 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
6189 DECL_INITIAL for globals not in the current function context.
6191 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6193 PR tree-optimization/56490
6194 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
6195 * tree-ssa-uninit.c: Include params.h.
6196 (compute_control_dep_chain): Add num_calls argument, return false
6197 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
6198 num_calls to recursive call.
6199 (find_predicates): Change dep_chain into normal array,
6200 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
6201 variable and adjust compute_control_dep_chain caller.
6202 (find_def_preds): Likewise.
6204 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
6206 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
6207 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
6209 2014-02-21 Nick Clifton <nickc@redhat.com>
6211 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
6212 (pushhi1): Likewise.
6213 (popqi1): Add mode to pre_dec.
6216 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6218 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
6219 mode for mask of V8SFmode permutation.
6221 2014-02-20 Richard Henderson <rth@redhat.com>
6224 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
6225 a new pseudo for OLDVAL.
6227 2014-02-20 Jakub Jelinek <jakub@redhat.com>
6230 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
6231 gen_reg_rtx if d->testing_p.
6232 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
6233 if d->testing_p and we will certainly return true.
6234 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6237 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6239 * emit-rtl.c (gen_reg_rtx): Assert that
6240 crtl->emit.regno_pointer_align_length is non-zero.
6242 2014-02-20 Richard Henderson <rth@redhat.com>
6245 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6246 on failure the store back into EXPECT.
6248 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6249 Sandra Loosemore <sandra@codesourcery.com>
6251 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6252 * config/nios2/nios2.c (nios2_function_profiler): Add
6253 -fPIC (flag_pic == 2) support.
6254 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6255 (nios2_large_offset_p): New function.
6256 (nios2_unspec_reloc_p): Move up position, update to use
6257 nios2_large_offset_p.
6258 (nios2_unspec_address): Remove function.
6259 (nios2_unspec_offset): New function.
6260 (nios2_large_got_address): New function.
6261 (nios2_got_address): Add large offset support.
6262 (nios2_legitimize_tls_address): Update usage of removed and new
6264 (nios2_symbol_binds_local_p): New function.
6265 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
6266 (nios2_legitimize_address): Update to use nios2_large_offset_p.
6267 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
6268 (nios2_print_operand): Merge H/L processing, add hiadj/lo
6269 processing for (const (unspec ...)).
6270 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
6272 2014-02-20 Richard Biener <rguenther@suse.de>
6274 * tree-cfg.c (replace_uses_by): Mark altered BBs before
6275 doing the substitution.
6276 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
6278 2014-02-20 Martin Jambor <mjambor@suse.cz>
6281 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
6282 info when checking whether lattices are bottom.
6284 2014-02-20 Richard Biener <rguenther@suse.de>
6287 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
6290 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
6293 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
6294 parameter specifying the scaling.
6295 (inline_call): Update.
6296 (want_inline_recursively): Guard division by zero.
6297 (recursive_inlining): Update.
6298 * ipa-inline.h (clone_inlined_nodes): Update.
6300 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6303 * config/i386/i386.c (classify_argument): Pass structures of size
6304 64 bytes or less in register.
6306 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6307 Kirill Yukhin <kirill.yukhin@intel.com>
6309 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
6310 (_mm_rcp28_round_ss): Ditto.
6311 (_mm_rsqrt28_round_sd): Ditto.
6312 (_mm_rsqrt28_round_ss): Ditto.
6313 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
6314 (_mm_rcp14_round_ss): Ditto.
6315 (_mm_rsqrt14_round_sd): Ditto.
6316 (_mm_rsqrt14_round_ss): Ditto.
6317 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
6318 the first input operand, get rid of match_dup.
6319 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
6321 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
6323 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
6324 operand as the first input operand, set type attribute.
6325 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
6327 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
6328 operand as the first input operand, set type attribute.
6330 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6332 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
6335 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
6338 * config/i386/i386.c (construct_container): Remove TFmode check
6339 for X86_64_INTEGER_CLASS.
6341 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
6344 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
6345 only when -Wpsabi is enabled.
6347 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
6350 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
6351 passing arrays in registers are the same as for structs, so remove the
6352 special case for them.
6354 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
6356 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
6357 destination type, extract only the valid bits if the source type is not
6358 integral and has a different mode.
6360 2014-02-19 Richard Biener <rguenther@suse.de>
6363 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
6366 2014-02-19 Richard Biener <rguenther@suse.de>
6369 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
6370 (ipa_modify_call_arguments): Emit an argument load explicitely and
6371 preserve virtual SSA form there and for the replacement call.
6372 Do not update SSA form nor free dominance info.
6374 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6376 * ipa.c (function_and_variable_visibility): Also clear WEAK
6377 flag when disolving COMDAT_GROUP.
6379 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6381 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
6382 * ipa-prop.c (ipa_set_jf_known_type): Return early when
6384 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
6385 do more sanity checks.
6386 (detect_type_change): Return true when giving up early.
6387 (compute_complex_assign_jump_func): Fix type parameter of
6388 ipa_set_ancestor_jf.
6389 (compute_complex_ancestor_jump_func): Likewise.
6390 (update_jump_functions_after_inlining): Fix updating of
6392 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
6394 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6396 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
6397 inline clones when edge disappears.
6399 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6402 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
6403 Split 64-bit moves into 2 patterns. Do not allow the use of
6404 direct move for TDmode in little endian, since the decimal value
6405 has little endian bytes within a word, but the 64-bit pieces are
6406 ordered in a big endian fashion, and normal subreg's of TDmode are
6408 (mov<mode>_64bit_dm): Likewise.
6409 (movtd_64bit_nodm): Likewise.
6411 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6413 PR tree-optimization/60174
6414 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
6415 statement of an SSA_NAME that occurs in an abnormal PHI node.
6417 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6420 * final.c (SEEN_BB): Remove.
6421 (SEEN_NOTE, SEEN_EMITTED): Renumber.
6422 (final_scan_insn): Don't force_source_line on second
6423 NOTE_INSN_BASIC_BLOCK.
6425 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
6428 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
6429 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
6430 (type_natural_mode): Warn ABI change when %zmm register is not
6431 available for AVX512F vector value passing.
6433 2014-02-18 Kai Tietz <ktietz@redhat.com>
6436 * config/i386/i386.c (ix86_expand_prologue): Use value in
6437 rax register as displacement when restoring %r10 or %rax.
6438 Fix wrong offset when restoring both registers.
6440 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6442 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
6443 assertion with conditional return.
6445 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6446 Uros Bizjak <ubizjak@gmail.com>
6449 * config/i386/driver-i386.c (host_detect_local_cpu): If
6450 YMM state is not saved by the OS, also clear has_f16c. Move
6451 CPUID 0x80000001 handling before YMM state saving checking.
6453 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
6455 PR rtl-optimization/58960
6456 * haifa-sched.c (alloc_global_sched_pressure_data): New,
6457 factored out from ...
6458 (sched_init): ... here.
6459 (free_global_sched_pressure_data): New, factored out from ...
6460 (sched_finish): ... here.
6461 * sched-int.h (free_global_sched_pressure_data): Declare.
6462 * sched-rgn.c (nr_regions_initial): New static global.
6463 (haifa_find_rgns): Initialize it.
6464 (schedule_region): Disable sched-pressure for the newly
6467 2014-02-17 Richard Biener <rguenther@suse.de>
6469 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
6470 release SSA defs of pattern stmts.
6472 2014-02-17 Richard Biener <rguenther@suse.de>
6474 * tree-inline.c (expand_call_inline): Release the virtual
6475 operand defined by the call we are about to inline.
6477 2014-02-17 Richard Biener <rguenther@suse.de>
6479 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
6481 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
6482 Ilya Tocar <ilya.tocar@intel.com>
6484 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
6485 arguments order in builtin.
6486 (_mm512_permutexvar_epi64): Ditto.
6487 (_mm512_mask_permutexvar_epi64): Ditto
6488 (_mm512_maskz_permutexvar_epi32): Ditto
6489 (_mm512_permutexvar_epi32): Ditto
6490 (_mm512_mask_permutexvar_epi32): Ditto
6492 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6494 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
6495 (p8_vmrgow): Likewise.
6497 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6499 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
6502 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
6505 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
6506 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
6507 into 64-bit and 32-bit moves. On 64-bit moves, add support for
6508 using direct move instructions on ISA 2.07. Also adjust
6509 instruction length for 64-bit.
6510 (mov<mode>_64bit, TFmode/TDmode): Likewise.
6511 (mov<mode>_32bit, TFmode/TDmode): Likewise.
6513 2014-02-15 Alan Modra <amodra@gmail.com>
6517 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
6518 find_replacement on parts of insn rtl that might be reloaded.
6520 2014-02-15 Richard Biener <rguenther@suse.de>
6522 PR tree-optimization/60183
6523 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
6524 (tree_ssa_phiprop): Calculate and free post-dominators.
6526 2014-02-14 Jeff Law <law@redhat.com>
6528 PR rtl-optimization/60131
6529 * ree.c (get_extended_src_reg): New function.
6530 (combine_reaching_defs): Use it rather than assuming location of REG.
6531 (find_and_remove_re): Verify first operand of extension is
6532 a REG before adding the insns to the copy list.
6534 2014-02-14 Roland McGrath <mcgrathr@google.com>
6536 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
6537 * configure: Regenerated.
6538 * config.in: Regenerated.
6539 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
6540 instead of ASM_SHORT.
6542 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
6543 Richard Earnshaw <rearnsha@arm.com>
6545 PR rtl-optimization/59535
6546 * lra-constraints.c (process_alt_operands): Encourage alternative
6547 when unassigned pseudo class is superset of the alternative class.
6548 (inherit_reload_reg): Don't inherit when optimizing for code size.
6549 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
6550 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
6551 modes not less than 4 for Thumb1.
6553 2014-02-14 Kyle McMartin <kyle@redhat.com>
6556 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
6558 2014-02-14 Richard Biener <rguenther@suse.de>
6560 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
6561 (get_frame_arg): Drop the assert with langhook types_compatible_p.
6562 Do not strip INDIRECT_REFs.
6564 2014-02-14 Richard Biener <rguenther@suse.de>
6567 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
6568 DECL_FUNCTION_SPECIFIC_TARGET.
6569 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
6570 * tree-streamer-out.c (pack_ts_target_option): Remove.
6571 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
6572 (write_ts_function_decl_tree_pointers): Do not stream
6573 DECL_FUNCTION_SPECIFIC_TARGET.
6574 * tree-streamer-in.c (unpack_ts_target_option): Remove.
6575 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
6576 (lto_input_ts_function_decl_tree_pointers): Do not stream
6577 DECL_FUNCTION_SPECIFIC_TARGET.
6579 2014-02-14 Jakub Jelinek <jakub@redhat.com>
6581 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
6582 (get_initial_def_for_induction, vectorizable_induction): Ignore
6583 debug stmts when looking for exit_phi.
6584 (vectorizable_live_operation): Fix up condition.
6586 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6588 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
6589 nreverse() because it changes the content of original tree list.
6591 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6593 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
6594 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
6596 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6598 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
6599 GNU coding standards.
6601 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6604 * dwarf2out.c (gen_subprogram_die): Don't call
6605 add_calling_convention_attribute if subr_die is old_die.
6607 2014-02-13 Sharad Singhai <singhai@google.com>
6609 * doc/optinfo.texi: Fix order of nodes.
6611 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
6613 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
6614 operands[2], not operands[3].
6616 2014-02-13 Richard Biener <rguenther@suse.de>
6619 * doc/install.texi (ISL): Update recommended version to 0.12.2,
6620 mention the possibility of an in-tree build.
6621 (CLooG): Update recommended version to 0.18.1, mention the
6622 possibility of an in-tree build and clarify that the ISL
6623 bundled with CLooG does not work.
6625 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6628 * expr.c (compress_float_constant): If x is a hard register,
6629 extend into a pseudo and then move to x.
6631 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
6633 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
6634 caused by bad second argument to warning_at() with -mhotpatch and
6635 nested functions (e.g. with gfortran).
6637 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
6639 * opts.c (option_name): Remove "enabled by default" rider.
6641 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
6643 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
6645 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
6646 Uros Bizjak <ubizjak@gmail.com>
6649 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
6650 * configure: Regenerated.
6652 2014-02-12 Richard Biener <rguenther@suse.de>
6654 * vec.c (vec_prefix::calculate_allocation): Move as
6655 inline variant to vec.h.
6656 (vec_prefix::calculate_allocation_1): New out-of-line version.
6657 * vec.h (vec_prefix::calculate_allocation_1): Declare.
6658 (vec_prefix::m_has_auto_buf): Rename to ...
6659 (vec_prefix::m_using_auto_storage): ... this.
6660 (vec_prefix::calculate_allocation): Inline the easy cases
6661 and dispatch to calculate_allocation_1 which doesn't need the
6663 (va_heap::reserve): Use gcc_checking_assert.
6664 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
6665 m_using_auto_storage.
6666 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
6668 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
6669 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
6670 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
6672 2014-02-12 Richard Biener <rguenther@suse.de>
6674 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
6675 when we found a dependence.
6677 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
6679 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
6681 (maybe_fold_stmt): ... into this new function.
6682 * omp-low.c (lower_omp): Update comment.
6684 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
6687 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
6690 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
6692 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
6693 identifiers in comments.
6694 (cortexa53_extra_costs): Likewise.
6695 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
6696 (cortexa7_extra_costs): Likewise.
6697 (cortexa12_extra_costs): Likewise.
6698 (cortexa15_extra_costs): Likewise.
6699 (v7m_extra_costs): Likewise.
6701 2014-02-12 Richard Biener <rguenther@suse.de>
6704 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
6705 of posix_memalign being successful.
6706 (lower_stmt): Restrict lowering of posix_memalign to when
6707 -ftree-bit-ccp is enabled.
6709 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6711 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
6713 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
6715 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
6717 PR rtl-optimization/60116
6718 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
6719 other_insn once the combination has been validated.
6721 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
6724 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
6726 * ipa-devirt.c: Include demangle.h
6727 (odr_violation_reported): New static variable.
6728 (add_type_duplicate): Update odr_violations.
6729 (maybe_record_node): Add completep parameter; update it.
6730 (record_target_from_binfo): Add COMPLETEP parameter;
6731 update it as needed.
6732 (possible_polymorphic_call_targets_1): Likewise.
6733 (struct polymorphic_call_target_d): Add nonconstruction_targets;
6734 rename FINAL to COMPLETE.
6735 (record_targets_from_bases): Sanity check we found the binfo;
6736 fix COMPLETEP updating.
6737 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
6738 parameter, fix computing of COMPLETEP.
6739 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
6740 at LTO time do demangling.
6741 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
6742 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
6744 (gimple_get_virt_method_for_binfo): Likewise.
6745 * gimple-fold.h (gimple_get_virt_method_for_binfo,
6746 gimple_get_virt_method_for_vtable): Update prototypes.
6748 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
6751 * genautomata.c (add_presence_absence): Fix typo with
6752 {final_}presence_list.
6754 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
6757 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
6758 for VSX/Altivec vectors that land in GPR registers.
6760 2014-02-11 Richard Henderson <rth@redhat.com>
6761 Jakub Jelinek <jakub@redhat.com>
6764 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
6765 around drhs if type conversion to lacc->type is not useless.
6767 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6769 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
6771 (cortex-a57.cortex-a53): Likewise.
6772 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
6774 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6776 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
6779 2014-02-11 Renlin Li <Renlin.Li@arm.com>
6781 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
6782 add_options_for_arm_vfp3.
6784 2014-02-11 Jeff Law <law@redhat.com>
6787 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
6788 object with an undesirable mode.
6790 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6793 * config/i386/sol2-9.h: New file.
6794 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
6795 *-*-solaris2.9*): Use it.
6797 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6799 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
6800 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
6802 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6804 * config/microblaze/microblaze.c: Extend mcpu version format
6806 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
6808 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
6810 2014-02-10 Richard Henderson <rth@redhat.com>
6813 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
6814 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
6815 ms-abi vs -mno-accumulate-outgoing-args.
6816 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
6817 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
6820 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
6823 * cfgexpand.c (expand_asm_operands): Attach source location to
6824 ASM_INPUT rtx objects.
6825 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
6827 2014-02-10 Nick Clifton <nickc@redhat.com>
6829 * config/mn10300/mn10300.c (popcount): New function.
6830 (mn10300_expand_prologue): Include saved registers in stack usage
6833 2014-02-10 Jeff Law <law@redhat.com>
6836 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
6837 when changing the SET_DEST of a prior insn to avoid an input reload.
6839 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6841 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
6842 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
6843 -mcall-openbsd, or -mcall-linux.
6844 (CC1_ENDIAN_BIG_SPEC): Remove.
6845 (CC1_ENDIAN_LITTLE_SPEC): Remove.
6846 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
6847 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
6848 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
6849 and %cc1_endian_default.
6850 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
6852 2014-02-10 Richard Biener <rguenther@suse.de>
6854 PR tree-optimization/60115
6855 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
6856 MEM_REF handling. Properly verify that the accesses are not
6857 out of the objects bound.
6859 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6861 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
6864 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
6866 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
6867 proper constants and fix formatting.
6868 (possible_polymorphic_call_targets): Fix formatting.
6870 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
6871 Ilya Tocar <ilya.tocar@intel.com>
6873 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
6874 (_mm512_loadu_epi32): Renamed into...
6875 (_mm512_loadu_si512): This.
6876 (_mm512_storeu_epi32): Renamed into...
6877 (_mm512_storeu_si512): This.
6878 (_mm512_maskz_ceil_ps): Removed.
6879 (_mm512_maskz_ceil_pd): Ditto.
6880 (_mm512_maskz_floor_ps): Ditto.
6881 (_mm512_maskz_floor_pd): Ditto.
6882 (_mm512_floor_round_ps): Ditto.
6883 (_mm512_floor_round_pd): Ditto.
6884 (_mm512_ceil_round_ps): Ditto.
6885 (_mm512_ceil_round_pd): Ditto.
6886 (_mm512_mask_floor_round_ps): Ditto.
6887 (_mm512_mask_floor_round_pd): Ditto.
6888 (_mm512_mask_ceil_round_ps): Ditto.
6889 (_mm512_mask_ceil_round_pd): Ditto.
6890 (_mm512_maskz_floor_round_ps): Ditto.
6891 (_mm512_maskz_floor_round_pd): Ditto.
6892 (_mm512_maskz_ceil_round_ps): Ditto.
6893 (_mm512_maskz_ceil_round_pd): Ditto.
6894 (_mm512_expand_pd): Ditto.
6895 (_mm512_expand_ps): Ditto.
6896 * config/i386/i386.c (ix86_builtins): Remove
6897 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
6898 (bdesc_args): Ditto.
6899 * config/i386/predicates.md (const1256_operand): New.
6900 (const_1_to_2_operand): Ditto.
6901 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6902 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6903 (*avx512pf_gatherpf<mode>sf): Ditto.
6904 (avx512pf_gatherpf<mode>df): Ditto.
6905 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6906 (*avx512pf_gatherpf<mode>df): Ditto.
6907 (avx512pf_scatterpf<mode>sf): Ditto.
6908 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6909 (*avx512pf_scatterpf<mode>sf): Ditto.
6910 (avx512pf_scatterpf<mode>df): Ditto.
6911 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6912 (*avx512pf_scatterpf<mode>df): Ditto.
6913 (avx512f_expand<mode>): Removed.
6914 (<shift_insn><mode>3<mask_name>): Change predicate type.
6916 2014-02-08 Jakub Jelinek <jakub@redhat.com>
6918 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
6919 not at the end of datarefs vector use ordered_remove to avoid
6920 reordering datarefs vector.
6923 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
6924 mark local addressable non-static vars as GOVD_PRIVATE
6925 instead of GOVD_LOCAL.
6926 * omp-low.c (lower_omp_for): Move gimple_bind_vars
6927 and BLOCK_VARS of gimple_bind_block to new_stmt rather
6931 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
6932 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
6933 assume_aligned or alloc_align attributes.
6934 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
6935 arguments. Handle also assume_aligned and alloc_align attributes.
6936 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
6937 calls to functions with assume_aligned or alloc_align attributes.
6938 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
6940 2014-02-08 Terry Guo <terry.guo@arm.com>
6942 * doc/invoke.texi: Document ARM -march=armv7e-m.
6944 2014-02-08 Jakub Jelinek <jakub@redhat.com>
6946 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
6947 flag on __cilkrts_rethrow builtin.
6950 * ipa-cp.c (determine_versionability): Fail at -O0
6951 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
6952 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
6955 2014-02-04 Jakub Jelinek <jakub@redhat.com>
6958 * tree-inline.c (copy_forbidden): Fail for
6959 __attribute__((optimize (0))) functions.
6961 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
6963 * varpool.c: Include pointer-set.h.
6964 (varpool_remove_unreferenced_decls): Variables in other partitions
6965 will not be output; be however careful to not lose information
6968 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
6970 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
6971 lookup in the vtable constructor.
6973 2014-02-07 Jeff Law <law@redhat.com>
6976 * config/m68k/m68k.md (ashldi_extsi): Turn into a
6977 define_insn_and_split.
6979 * ipa-inline.c (inline_small_functions): Fix typos.
6981 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6983 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
6984 (s390_can_use_return_insn): Declare.
6985 * config/s390/s390.h (EPILOGUE_USES): Define.
6986 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
6988 (s390_chunkify_start): Handle return JUMP_LABELs.
6989 (s390_early_mach): Emit a main_pool instruction on the entry edge.
6990 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
6991 (s390_can_use_return_insn): New functions.
6992 (s390_fix_long_loop_prediction): Handle conditional returns.
6993 (TARGET_SET_UP_BY_PROLOGUE): Define.
6994 * config/s390/s390.md (ANY_RETURN): New code iterator.
6995 (*creturn, *csimple_return, return, simple_return): New patterns.
6997 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6999 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
7000 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
7001 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
7002 REG_CFA_RESTORE list when deciding not to restore a register.
7004 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7006 * config/s390/s390.c: Include tree-pass.h and context.h.
7007 (s390_early_mach): New function, split out from...
7008 (s390_emit_prologue): ...here.
7009 (pass_data_s390_early_mach): New pass structure.
7010 (pass_s390_early_mach): New class.
7011 (s390_option_override): Create and register early_mach pass.
7012 Move to end of file.
7014 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7016 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
7017 to match for the exit block.
7019 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7021 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
7022 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
7023 Reject misaligned operands.
7025 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7027 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
7029 2014-02-07 Richard Biener <rguenther@suse.de>
7032 * gimple-low.c (lower_builtin_posix_memalign): New function.
7033 (lower_stmt): Call it to lower posix_memalign in a way
7034 to make alignment info accessible.
7036 2014-02-07 Jakub Jelinek <jakub@redhat.com>
7039 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
7040 __builtin_setjmp_receiver.
7042 2014-02-07 Richard Biener <rguenther@suse.de>
7045 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
7046 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
7047 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7048 Handle BUILT_IN_POSIX_MEMALIGN.
7049 (find_func_clobbers): Likewise.
7050 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
7051 (call_may_clobber_ref_p_1): Likewise.
7053 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7056 * ipa-devirt.c (record_target_from_binfo): Remove overactive
7059 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7062 * lto-cgraph.c (lto_output_node): Use
7063 symtab_get_symbol_partitioning_class.
7064 (lto_output_varpool_node): likewise.
7065 (symtab_get_symbol_partitioning_class): Move here from
7067 * cgraph.h (symbol_partitioning_class): Likewise.
7068 (symtab_get_symbol_partitioning_class): Declare.
7070 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7072 * ggc.h (ggc_internal_cleared_alloc): New macro.
7073 * vec.h (vec_safe_copy): Handle memory stats.
7074 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
7075 * target-globals.c (save_target_globals): Likewise.
7077 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7080 * expr.c (emit_move_resolve_push): Export; be bit more selective
7081 on when to clear alias set.
7082 * expr.h (emit_move_resolve_push): Declare.
7083 * function.h (struct function): Add tail_call_marked.
7084 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
7085 * config/i386/i386-protos.h (ix86_expand_push): Remove.
7086 * config/i386/i386.md (TImode move expander): De not call
7088 (FP push expanders): Preserve memory attributes.
7089 * config/i386/sse.md (push<mode>1): Remove.
7090 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
7091 (ix86_expand_push): Remove.
7092 * config/i386/mmx.md (push<mode>1): Remove.
7094 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7096 PR rtl-optimization/60030
7097 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
7098 lopart with paradoxical subreg before shifting it up by hprec.
7100 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7102 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
7103 Remove extra newline at end of file.
7104 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
7105 (arm_issue_rate): Handle cortexa57.
7106 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
7107 (cortex-a57.cortex-a53): Likewise.
7109 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7112 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
7113 don't record in REG_FRAME_RELATED_EXPR registers not set in that
7115 (arm_expand_prologue): Adjust all callers.
7116 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
7117 info, registers also at the lowest numbered registers side. Use
7118 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
7122 * var-tracking.c (adjust_mems): Before adding a SET to
7123 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
7125 2014-02-06 Alan Modra <amodra@gmail.com>
7128 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
7129 change SDmode to DDmode when lra_in_progress.
7131 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7134 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
7135 free_data_ref on the dr first, and before goto again also set dr
7136 to the next dr. For simd_lane_access, free old datarefs[i] before
7137 overwriting it. For get_vectype_for_scalar_type failure, don't
7138 free_data_ref if simd_lane_access.
7140 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
7143 * tree.h (opts_for_fn): New inline function.
7144 (opt_for_fn): Define.
7145 * config/i386/i386.c (ix86_function_regparm): Use
7146 opt_for_fn (decl, optimize) instead of optimize.
7148 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
7150 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
7151 for SYMBOL_REF in large memory model.
7153 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7155 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
7157 (cortex-a57): Likewise.
7158 (cortex-a57.cortex-a53): Likewise.
7160 2014-02-06 Yury Gribov <y.gribov@samsung.com>
7161 Kugan Vivekanandarajah <kuganv@linaro.org>
7163 * config/arm/arm.c (arm_vector_alignment_reachable): Check
7165 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
7167 2014-02-06 Richard Biener <rguenther@suse.de>
7169 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
7170 set_loop_copy and initialize_original_copy_tables.
7172 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
7174 * config/aarch64/aarch64-simd.md
7175 (aarch64_ashr_simddi): Change QI to SI.
7177 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7178 Jakub Jelinek <jakub@redhat.com>
7181 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
7184 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7186 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
7187 CODE_FOR_altivec_vpku[hw]um to
7188 CODE_FOR_altivec_vpku[hw]um_direct.
7189 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
7190 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
7191 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
7192 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
7194 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7196 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
7197 generation for -maltivec=be.
7198 (altivec_vsumsws): Simplify redundant test.
7200 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7202 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
7203 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
7204 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
7205 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
7206 gen_altivec_vpkuwum.
7207 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
7209 (altivec_vpks<VI_char>ss): Likewise.
7210 (altivec_vpks<VI_char>us): Likewise.
7211 (altivec_vpku<VI_char>us): Likewise.
7212 (altivec_vpku<VI_char>um): Likewise.
7213 (altivec_vpku<VI_char>um_direct): New (copy of
7214 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
7216 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
7217 target is little endian and -maltivec=be is not specified.
7218 (*altivec_vupkhs<VU_char>_direct): New (copy of
7219 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
7220 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
7221 target is little endian and -maltivec=be is not specified.
7222 (*altivec_vupkls<VU_char>_direct): New (copy of
7223 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
7224 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
7225 little endian and -maltivec=be is not specified.
7226 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
7227 little endian and -maltivec=be is not specified.
7229 2014-02-05 Richard Henderson <rth@redhat.com>
7232 * combine-stack-adj.c: Revert r206943.
7233 * sched-int.h (struct deps_desc): Add last_args_size.
7234 * sched-deps.c (init_deps): Initialize it.
7235 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7236 contain REG_ARGS_SIZE notes.
7238 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7240 * lto-cgraph.c (asm_nodes_output): Make global.
7241 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7242 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7243 (driver_handle_option): Handle OPT_fwpa.
7245 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7248 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7249 a comment typo and formatting issue. If odr_hash hasn't been
7250 created, return vNULL and set *completep to false.
7253 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7254 bb with no successors.
7256 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7259 * doc/invoke.texi (-march): Clarify documentation for ARM.
7263 2014-02-05 Richard Biener <rguenther@suse.de>
7265 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
7266 when not vectorizing because of too many alias checks.
7267 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
7268 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
7270 2014-02-05 Nick Clifton <nickc@redhat.com>
7272 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
7273 accept extended registers in any mode when compiling for the MN10300.
7275 2014-02-05 Yury Gribov <y.gribov@samsung.com>
7277 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
7278 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
7279 sanitization attributes.
7280 (can_inline_edge_p): Likewise.
7281 (sanitize_attrs_match_for_inline_p): New function.
7283 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7285 * ipa-prop.c (detect_type_change): Shor circuit testing of
7286 type changes on THIS pointer.
7288 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
7291 * config/pa/pa.c (legitimize_tls_address): Return original address
7292 if not passed a SYMBOL_REF rtx.
7293 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
7295 (pa_emit_move_sequence): Simplify TLS source operands.
7296 (pa_legitimate_constant_p): Reject all TLS constants.
7297 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
7298 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
7300 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7302 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
7303 groups when we know they are controlled by LTO.
7304 * varasm.c (default_binds_local_p_1): If object is in other partition,
7305 it will be resolved locally.
7307 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7309 * config/host-linux.c (linux_gt_pch_use_address): Don't
7310 use SSIZE_MAX because it is not always defined.
7312 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
7315 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
7316 threshold for pseudo splitting.
7317 (update_ebb_live_info): Process call argument hard registers and
7318 hard registers from insn definition too.
7319 (max_small_class_regs_num): New constant.
7320 (inherit_in_ebb): Update live hard regs through EBBs. Update
7321 reloads_num only for small register classes. Don't split for
7324 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
7327 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
7330 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7332 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
7335 2014-02-04 Marek Polacek <polacek@redhat.com>
7337 * gdbinit.in (pel): Define.
7339 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7341 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
7344 2014-02-04 Richard Biener <rguenther@suse.de>
7347 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
7348 in function context local.
7349 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
7350 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
7351 similar to LTO_imported_decl_ref.
7353 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7355 PR tree-optimization/60002
7356 * cgraphclones.c (build_function_decl_skip_args): Clear
7359 PR tree-optimization/60023
7360 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
7361 false to gsi_replace.
7362 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
7363 has been in some EH region and vec_stmt could throw, add
7364 vec_stmt into the same EH region.
7365 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
7366 has no lhs, ignore it.
7367 * internal-fn.c (expand_MASK_LOAD): Likewise.
7370 * tree-inline.c (copy_forbidden): Fail for
7371 __attribute__((optimize (0))) functions.
7374 * omp-low.c (simd_clone_struct_copy): If from->inbranch
7375 is set, copy one less argument.
7376 (expand_simd_clones): Don't subtract clone_info->inbranch
7377 from simd_clone_struct_alloc argument.
7379 PR rtl-optimization/57915
7380 * recog.c (simplify_while_replacing): If all unary/binary/relational
7381 operation arguments are constant, attempt to simplify those.
7384 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
7385 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
7387 2014-02-04 Richard Biener <rguenther@suse.de>
7389 PR tree-optimization/60012
7390 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
7391 TBAA disambiguation to all DDRs.
7393 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7396 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
7397 (LINK_SPEC): Use it for -shared, -shared-libgcc.
7399 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7402 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
7404 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7406 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
7407 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
7409 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7412 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
7413 to figure out targets of polymorphic calls with known decl.
7414 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7415 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
7416 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
7417 (get_polymorphic_call_info): ... here.
7418 (get_polymorphic_call_info_from_invariant): New function.
7420 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7422 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
7423 lookup via vtable pointer; check for type consistency
7424 and turn inconsitent facts into UNREACHABLE.
7425 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7426 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
7427 type inconsistent querries; return UNREACHABLE instead.
7429 2014-02-03 Richard Henderson <rth@twiddle.net>
7432 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
7433 already processed this node.
7434 (normalize_one_pred_1): Pass along mark_set.
7435 (normalize_one_pred): Create and destroy a pointer_set_t.
7436 (normalize_one_pred_chain): Likewise.
7438 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
7440 PR gcov-profile/58602
7441 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
7443 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7446 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
7447 -fno-devirtualize; try to devirtualize by the knowledge of
7448 virtual table pointer given by aggregate propagation.
7449 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7450 (ipa_print_node_jump_functions): Dump also offset that
7451 is relevant for polymorphic calls.
7452 (determine_known_aggregate_parts): Add arg_type parameter; use it
7453 instead of determining the type from pointer type.
7454 (ipa_compute_jump_functions_for_edge): Update call of
7455 determine_known_aggregate_parts.
7456 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
7457 (gimple_get_virt_method_for_binfo): ... here; simplify using
7458 vtable_pointer_value_to_vtable.
7459 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
7460 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
7461 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
7462 (vtable_pointer_value_to_vtable): Break out from ...; handle also
7464 (vtable_pointer_value_to_binfo): ... here.
7465 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
7467 2014-02-03 Teresa Johnson <tejohnson@google.com>
7469 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
7470 redef of outer loop index variable.
7472 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
7476 * doc/extend.texi (Function Attributes): Typo.
7478 2014-02-03 Cong Hou <congh@google.com>
7480 PR tree-optimization/60000
7481 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
7482 if the vectorized statement is a store. A store statement can only
7483 appear at the end of pattern statements.
7485 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7487 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
7488 (ix86_option_override_internal): Default long double to 64-bit for
7489 32-bit Bionic and to 128-bit for 64-bit Bionic.
7491 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
7492 TARGET_LONG_DOUBLE_128 is true.
7493 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
7495 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
7496 (mlong-double-64): Negate -mlong-double-128.
7497 (mlong-double-128): New option.
7499 * config/i386/i386-c.c (ix86_target_macros): Define
7500 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
7502 * doc/invoke.texi: Document -mlong-double-128.
7504 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7506 PR rtl-optimization/60024
7507 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
7509 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
7511 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
7513 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7515 PR rtl-optimization/57662
7516 * sel-sched.c (code_motion_path_driver): Do not mark already not
7517 existing blocks in the visiting bitmap.
7519 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7521 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
7522 on the insn being emitted.
7524 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
7525 Will Deacon <will.deacon@arm.com>
7527 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
7529 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7531 * config/arm/arm-tables.opt: Regenerate.
7533 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7535 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
7536 for vector types other than V16QImode.
7537 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
7538 define_expand, and call altivec_expand_vec_perm_le when producing
7539 code with little endian element order.
7540 (*altivec_vperm_<mode>_internal): New insn having previous
7541 behavior of altivec_vperm_<mode>.
7542 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
7543 altivec_expand_vec_perm_le when producing code with little endian
7545 (*altivec_vperm_<mode>_uns_internal): New insn having previous
7546 behavior of altivec_vperm_<mode>_uns.
7548 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7550 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
7551 (altivec_vsumsws): Add handling for -maltivec=be with a little
7553 (altivec_vsumsws_direct): New.
7554 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
7555 gen_altivec_vsumsws.
7557 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7559 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
7560 vtable_pointer_value_to_binfo): New functions.
7561 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
7562 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
7564 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
7566 * config/nios2/nios2.md (load_got_register): Initialize GOT
7567 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
7568 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
7570 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7572 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
7573 preserverd by passthrough, do not propagate the type.
7575 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7577 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
7578 (mips_atomic_assign_expand_fenv): New function.
7579 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
7581 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7583 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
7584 (__builtin_mips_set_fcsr): Likewise.
7585 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
7586 MIPS_USI_FTYPE_VOID.
7587 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
7588 (mips16_expand_set_fcsr): Likewise.
7589 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
7590 (mips16_set_fcsr_stub): Likewise.
7591 (mips16_get_fcsr_one_only_stub): New class.
7592 (mips16_set_fcsr_one_only_stub): Likewise.
7593 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
7594 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
7595 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
7596 (hard_float): New availability predicate.
7597 (mips_builtins): Add get_fcsr and set_fcsr.
7598 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
7599 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
7600 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
7601 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
7602 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
7605 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7607 * config/mips/mips.c (mips_one_only_stub): New class.
7608 (mips_need_mips16_rdhwr_p): Replace with...
7609 (mips16_rdhwr_stub): ...this new variable.
7610 (mips16_stub_call_address): New function.
7611 (mips16_rdhwr_one_only_stub): New class.
7612 (mips_expand_thread_pointer): Use mips16_stub_call_address.
7613 (mips_output_mips16_rdhwr): Delete.
7614 (mips_finish_stub): New function.
7615 (mips_code_end): Use it to handle rdhwr stubs.
7617 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
7620 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
7621 when calculating size of integer atomic types.
7623 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
7625 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
7627 2014-02-01 Jakub Jelinek <jakub@redhat.com>
7629 PR tree-optimization/60003
7630 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
7631 * profile.c (branch_prob): Use gimple_call_builtin_p
7632 to check for BUILT_IN_SETJMP_RECEIVER.
7633 * tree-inline.c (copy_bb): Call notice_special_calls.
7635 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
7638 * lra-constraints.c (process_alt_operands): Update reload_sum only
7641 2014-01-31 Richard Henderson <rth@redhat.com>
7644 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
7645 until after else_eh is processed.
7647 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7649 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
7650 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
7651 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
7652 in smmintrin.h, remove them.
7653 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
7654 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
7655 * config/i386/i386.md (ROUND_SAE): Fix value.
7656 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
7657 (const48_operand): New.
7658 * config/i386/subst.md (round), (round_expand): Use
7659 const_4_or_8_to_11_operand.
7660 (round_saeonly), (round_saeonly_expand): Use const48_operand.
7662 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7664 * config/i386/constraints.md (Yk): Swap meaning with k.
7665 * config/i386/i386.md (movhi_internal): Change Yk to k.
7666 (movqi_internal): Ditto.
7667 (*k<logic><mode>): Ditto.
7670 (kandn<mode>): Ditto.
7671 (*<code>hi_1): Ditto.
7672 (*<code>qi_1): Ditto.
7673 (kxnor<mode>): Ditto.
7674 (kortestzhi): Ditto.
7675 (kortestchi): Ditto.
7677 (*one_cmplhi2_1): Ditto.
7678 (*one_cmplqi2_1): Ditto.
7679 * config/i386/sse.md (): Change k to Yk.
7680 (avx512f_load<mode>_mask): Ditto.
7681 (avx512f_blendm<mode>): Ditto.
7682 (avx512f_store<mode>_mask): Ditto.
7683 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
7684 (avx512f_storedqu<mode>_mask): Ditto.
7685 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
7687 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
7688 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
7689 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
7690 (avx512f_maskcmp<mode>3): Ditto.
7691 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
7692 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
7693 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
7694 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
7695 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
7696 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
7697 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
7698 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
7699 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
7700 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
7701 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
7702 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
7703 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
7704 (vec_extract_lo_<mode>_maskm): Ditto.
7705 (vec_extract_hi_<mode>_maskm): Ditto.
7706 (avx512f_vternlog<mode>_mask): Ditto.
7707 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
7708 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
7709 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
7710 (avx512f_<code>v8div16qi2_mask): Ditto.
7711 (avx512f_<code>v8div16qi2_mask_store): Ditto.
7712 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
7713 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
7714 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
7715 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
7716 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7717 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7718 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7719 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7720 (avx512cd_maskb_vec_dupv8di): Ditto.
7721 (avx512cd_maskw_vec_dupv16si): Ditto.
7722 (avx512f_vpermi2var<mode>3_maskz): Ditto.
7723 (avx512f_vpermi2var<mode>3_mask): Ditto.
7724 (avx512f_vpermi2var<mode>3_mask): Ditto.
7725 (avx512f_vpermt2var<mode>3_maskz): Ditto.
7726 (*avx512f_gathersi<mode>): Ditto.
7727 (*avx512f_gathersi<mode>_2): Ditto.
7728 (*avx512f_gatherdi<mode>): Ditto.
7729 (*avx512f_gatherdi<mode>_2): Ditto.
7730 (*avx512f_scattersi<mode>): Ditto.
7731 (*avx512f_scatterdi<mode>): Ditto.
7732 (avx512f_compress<mode>_mask): Ditto.
7733 (avx512f_compressstore<mode>_mask): Ditto.
7734 (avx512f_expand<mode>_mask): Ditto.
7735 * config/i386/subst.md (mask): Change k to Yk.
7736 (mask_scalar_merge): Ditto.
7739 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
7741 * doc/extend.texi (Vector Extensions): Document ?: in C++.
7743 2014-01-31 Richard Biener <rguenther@suse.de>
7746 * builtins.c (fold_builtin_memory_op): Make sure to not
7747 use a floating-point mode or a boolean or enumeral type for
7750 2014-01-30 DJ Delorie <dj@redhat.com>
7752 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
7753 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
7754 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
7755 whenever main() has an epilogue.
7757 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7759 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
7760 unused variable "field".
7761 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
7762 (vsx_mergeh_<mode>): Likewise.
7763 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
7764 (altivec_vmrghh): Likewise.
7765 (altivec_vmrghw): Likewise.
7766 (altivec_vmrglb): Likewise.
7767 (altivec_vmrglh): Likewise.
7768 (altivec_vmrglw): Likewise.
7769 (altivec_vspltb): Add missing uses.
7770 (altivec_vsplth): Likewise.
7771 (altivec_vspltw): Likewise.
7772 (altivec_vspltsf): Likewise.
7774 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7777 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
7778 frame related instructions.
7780 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
7782 PR rtl-optimization/59959
7783 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
7784 any reload of register whose subreg is invalid.
7786 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7788 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
7789 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
7790 Add missing return type - void.
7792 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7794 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
7795 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
7796 remove element index adjustment for endian (now handled in vsx.md
7798 (altivec_expand_vec_perm_const): Use
7799 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
7800 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
7801 (vsx_xxspltw_<mode>): Adjust element index for little endian.
7802 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
7803 define_expand and a new define_insn *altivec_vspltb_internal;
7804 adjust for -maltivec=be on a little endian target.
7805 (altivec_vspltb_direct): New.
7806 (altivec_vsplth): Divide into a define_expand and a new
7807 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
7808 little endian target.
7809 (altivec_vsplth_direct): New.
7810 (altivec_vspltw): Divide into a define_expand and a new
7811 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
7812 little endian target.
7813 (altivec_vspltw_direct): New.
7814 (altivec_vspltsf): Divide into a define_expand and a new
7815 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
7816 a little endian target.
7818 2014-01-30 Richard Biener <rguenther@suse.de>
7820 PR tree-optimization/59993
7821 * tree-ssa-forwprop.c (associate_pointerplus): Check we
7822 can propagate form the earlier stmt and avoid the transform
7823 when the intermediate result is needed.
7825 2014-01-30 Alangi Derick <alangiderick@gmail.com>
7827 * README.Portability: Fix typo.
7829 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
7831 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
7832 comparison_operator with ordered_comparison_operator.
7834 2014-01-30 Nick Clifton <nickc@redhat.com>
7836 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
7837 Rename to mn10300_store_multiple_regs.
7838 * config/mn10300/mn10300.c: Likewise.
7839 * config/mn10300/mn10300.md (store_movm): Fix typo: call
7840 store_multiple_regs.
7841 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
7842 Call mn10300_store_multiple_regs.
7844 2014-01-30 Nick Clifton <nickc@redhat.com>
7845 DJ Delorie <dj@redhat.com>
7847 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
7848 %fp 2 to keep registers after it properly word-aligned.
7849 (rl78_alloc_physical_registers_umul): Handle the case where both
7850 input operands are the same.
7852 2014-01-30 Richard Biener <rguenther@suse.de>
7854 PR tree-optimization/59903
7855 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
7858 2014-01-30 Jason Merrill <jason@redhat.com>
7861 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
7864 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
7866 2014-01-30 Richard Biener <rguenther@suse.de>
7868 PR tree-optimization/59951
7869 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
7871 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
7874 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
7875 SFmode to DFmode case.
7877 2014-01-29 DJ Delorie <dj@redhat.com>
7879 * config/msp430/msp430.opt (-minrt): New.
7880 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
7882 (ENDFILE_SPEC): Likewise.
7884 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
7886 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
7887 (estimate_function_body_sizes): Use it.
7889 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
7892 * dwarf2out.c (is_cxx_auto): New.
7893 (is_base_type): Use it.
7894 (gen_type_die_with_usage): Likewise.
7896 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7898 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
7899 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
7900 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
7901 -maltivec=be with LE targets.
7902 (vsx_mergeh_<mode>): Likewise.
7903 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
7904 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
7905 (altivec_vmrghb): Replace with define_expand and new
7906 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
7907 (altivec_vmrghb_direct): New define_insn.
7908 (altivec_vmrghh): Replace with define_expand and new
7909 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
7910 (altivec_vmrghh_direct): New define_insn.
7911 (altivec_vmrghw): Replace with define_expand and new
7912 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
7913 (altivec_vmrghw_direct): New define_insn.
7914 (*altivec_vmrghsf): Adjust for endianness.
7915 (altivec_vmrglb): Replace with define_expand and new
7916 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
7917 (altivec_vmrglb_direct): New define_insn.
7918 (altivec_vmrglh): Replace with define_expand and new
7919 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
7920 (altivec_vmrglh_direct): New define_insn.
7921 (altivec_vmrglw): Replace with define_expand and new
7922 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
7923 (altivec_vmrglw_direct): New define_insn.
7924 (*altivec_vmrglsf): Adjust for endianness.
7925 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
7926 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
7927 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
7928 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
7929 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
7930 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
7931 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
7932 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
7934 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
7936 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
7937 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
7940 2014-01-29 Richard Biener <rguenther@suse.de>
7942 PR tree-optimization/58742
7943 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
7944 associate_pointerplus_align.
7945 (associate_pointerplus_diff): New function.
7946 (associate_pointerplus): Likewise. Call associate_pointerplus_align
7947 and associate_pointerplus_diff.
7949 2014-01-29 Richard Biener <rguenther@suse.de>
7951 * lto-streamer.h (LTO_major_version): Bump to 3.
7952 (LTO_minor_version): Reset to 0.
7954 2014-01-29 Renlin Li <Renlin.Li@arm.com>
7956 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
7957 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
7958 (arm_file_start): Generate correct asm header for armv7ve.
7959 * config/arm/bpabi.h: Add multilib support for armv7ve.
7960 * config/arm/driver-arm.c: Change the architectures of cortex-a7
7961 and cortex-a15 to armv7ve.
7962 * config/arm/t-aprofile: Add multilib support for armv7ve.
7963 * doc/invoke.texi: Document -march=armv7ve.
7965 2014-01-29 Richard Biener <rguenther@suse.de>
7967 PR tree-optimization/58742
7968 * tree-ssa-forwprop.c (associate_plusminus): Return true
7969 if we changed sth, defer EH cleanup to ...
7970 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
7971 (simplify_mult): New function.
7973 2014-01-29 Jakub Jelinek <jakub@redhat.com>
7976 PR tree-optimization/59920
7977 * tree.c (build_common_builtin_nodes): Remove
7978 __builtin_setjmp_dispatcher initialization.
7979 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
7980 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
7981 instead of gsi_after_labels + manually skipping debug stmts.
7982 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
7983 ignore bbs with IFN_ABNORMAL_DISPATCHER.
7984 * tree-inline.c (copy_edges_for_bb): Remove
7985 can_make_abnormal_goto argument, instead add abnormal_goto_dest
7986 argument. Ignore computed_goto_p stmts. Don't call
7987 make_abnormal_goto_edges. If a call might need abnormal edges
7988 for non-local gotos, see if it already has an edge to
7989 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
7990 with true argument, don't do anything then, otherwise add
7991 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
7992 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
7994 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
7995 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
7996 (lower_stmt): Don't set data->calls_builtin_setjmp.
7997 (lower_builtin_setjmp): Adjust comment.
7998 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
7999 * tree-cfg.c (found_computed_goto): Remove.
8000 (factor_computed_gotos): Remove.
8001 (make_goto_expr_edges): Return bool, true for computed gotos.
8002 Don't call make_abnormal_goto_edges.
8003 (build_gimple_cfg): Don't set found_computed_goto, don't call
8004 factor_computed_gotos.
8005 (computed_goto_p): No longer static.
8006 (make_blocks): Don't set found_computed_goto.
8007 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
8008 (make_edges): If make_goto_expr_edges returns true, push bb
8009 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
8010 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
8011 vector. Record mapping between bbs and OpenMP regions if there
8012 are any, adjust make_gimple_omp_edges caller. Call
8013 handle_abnormal_edges.
8014 (make_abnormal_goto_edges): Remove.
8015 * tree-cfg.h (make_abnormal_goto_edges): Remove.
8016 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
8017 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
8018 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
8019 * internal-fn.def (ABNORMAL_DISPATCHER): New.
8020 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
8021 filling *region also set *region_idx to (*region)->entry->index.
8024 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
8025 For REGs set ORIGINAL_REGNO.
8027 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
8029 * doc/md.texi: Mention that a target shouldn't implement
8030 vec_widen_(s|u)mul_even/odd pair if it is less efficient
8033 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8035 PR tree-optimization/59594
8036 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
8037 a copy of the datarefs vector rather than the vector itself.
8039 2014-01-28 Jason Merrill <jason@redhat.com>
8042 * dwarf2out.c (auto_die): New static.
8043 (gen_type_die_with_usage): Handle C++1y 'auto'.
8044 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
8047 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
8050 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
8051 (SPEC_X32): Likewise.
8052 (SPEC_64): Likewise.
8053 * config/i386/i386.c (ix86_option_override_internal): Turn off
8054 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
8056 (x86_file_start): Output .code16gcc for TARGET_16BIT.
8057 * config/i386/i386.h (TARGET_16BIT): New macro.
8058 (TARGET_16BIT_P): Likewise.
8059 * config/i386/i386.opt: Add m16.
8060 * doc/invoke.texi: Document -m16.
8062 2014-01-28 Jakub Jelinek <jakub@redhat.com>
8064 PR preprocessor/59935
8065 * input.c (location_get_source_line): Bail out on when line number
8066 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
8068 2014-01-28 Richard Biener <rguenther@suse.de>
8070 PR tree-optimization/58742
8071 * tree-ssa-forwprop.c (associate_plusminus): Handle
8072 pointer subtraction of the form (T)(P + A) - (T)P.
8074 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8076 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
8079 2014-01-28 Richard Biener <rguenther@suse.de>
8082 2014-01-28 Richard Biener <rguenther@suse.de>
8084 PR rtl-optimization/45364
8085 PR rtl-optimization/59890
8086 * var-tracking.c (local_get_addr_clear_given_value): Handle
8087 already cleared slot.
8088 (val_reset): Handle not allocated local_get_addr_cache.
8089 (vt_find_locations): Use post-order on the inverted CFG.
8091 2014-01-28 Richard Biener <rguenther@suse.de>
8093 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
8095 2014-01-28 Richard Biener <rguenther@suse.de>
8097 PR rtl-optimization/45364
8098 PR rtl-optimization/59890
8099 * var-tracking.c (local_get_addr_clear_given_value): Handle
8100 already cleared slot.
8101 (val_reset): Handle not allocated local_get_addr_cache.
8102 (vt_find_locations): Use post-order on the inverted CFG.
8104 2014-01-28 Alan Modra <amodra@gmail.com>
8106 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
8107 * configure.ac <recursive call for build != host>: Define
8108 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
8109 and LD_FOR_BUILD too.
8110 * configure: Regenerate.
8112 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
8114 * config/i386/i386.c (get_builtin_code_for_version): Separate
8115 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
8116 Broadwell from Haswell.
8118 2014-01-27 Steve Ellcey <sellcey@mips.com>
8120 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
8121 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
8122 * config/mips/mips.c (mips_option_override): Change setting
8124 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
8125 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
8126 Change from Mask to Var.
8128 2014-01-27 Jeff Law <law@redhat.com>
8130 * ipa-inline.c (inline_small_functions): Fix typo.
8132 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
8134 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
8135 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
8136 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
8137 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
8138 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
8139 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
8140 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
8141 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
8142 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
8143 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
8144 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
8145 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
8146 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
8147 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
8148 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
8149 (_mm512_storeu_epi64): Ditto.
8150 (_mm512_cmpge_epi32_mask): Ditto.
8151 (_mm512_cmpge_epu32_mask): Ditto.
8152 (_mm512_cmpge_epi64_mask): Ditto.
8153 (_mm512_cmpge_epu64_mask): Ditto.
8154 (_mm512_cmple_epi32_mask): Ditto.
8155 (_mm512_cmple_epu32_mask): Ditto.
8156 (_mm512_cmple_epi64_mask): Ditto.
8157 (_mm512_cmple_epu64_mask): Ditto.
8158 (_mm512_cmplt_epi32_mask): Ditto.
8159 (_mm512_cmplt_epu32_mask): Ditto.
8160 (_mm512_cmplt_epi64_mask): Ditto.
8161 (_mm512_cmplt_epu64_mask): Ditto.
8162 (_mm512_cmpneq_epi32_mask): Ditto.
8163 (_mm512_cmpneq_epu32_mask): Ditto.
8164 (_mm512_cmpneq_epi64_mask): Ditto.
8165 (_mm512_cmpneq_epu64_mask): Ditto.
8166 (_mm512_expand_pd): Ditto.
8167 (_mm512_expand_ps): Ditto.
8168 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
8169 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
8170 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
8171 * config/i386/i386.c (ix86_builtins): Add
8172 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
8173 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
8174 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
8175 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
8176 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
8177 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
8178 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
8179 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
8180 IX86_BUILTIN_PMOVUSQW512_MEM.
8181 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
8182 __builtin_ia32_pmovsqd512mem_mask,
8183 __builtin_ia32_pmovqd512mem_mask,
8184 __builtin_ia32_pmovusqw512mem_mask,
8185 __builtin_ia32_pmovsqw512mem_mask,
8186 __builtin_ia32_pmovqw512mem_mask,
8187 __builtin_ia32_pmovusdw512mem_mask,
8188 __builtin_ia32_pmovsdw512mem_mask,
8189 __builtin_ia32_pmovdw512mem_mask,
8190 __builtin_ia32_pmovqb512mem_mask,
8191 __builtin_ia32_pmovusqb512mem_mask,
8192 __builtin_ia32_pmovsqb512mem_mask,
8193 __builtin_ia32_pmovusdb512mem_mask,
8194 __builtin_ia32_pmovsdb512mem_mask,
8195 __builtin_ia32_pmovdb512mem_mask.
8196 (bdesc_args): Add __builtin_ia32_expanddf512,
8197 __builtin_ia32_expandsf512.
8198 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
8199 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
8200 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
8201 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
8202 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
8203 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
8204 (avx512f_<code>v8div16qi2_mask_store): This.
8205 (avx512f_expand<mode>): New.
8207 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
8209 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
8211 (_mm512_mask_prefetch_i64gather_pd): Ditto.
8212 (_mm512_prefetch_i32scatter_pd): Ditto.
8213 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
8214 (_mm512_prefetch_i64scatter_pd): Ditto.
8215 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
8216 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
8217 (_mm512_mask_prefetch_i64gather_ps): Ditto.
8218 (_mm512_prefetch_i32scatter_ps): Ditto.
8219 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
8220 (_mm512_prefetch_i64scatter_ps): Ditto.
8221 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
8222 * config/i386/i386-builtin-types.def: Define
8223 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
8224 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
8225 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
8226 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
8227 IX86_BUILTIN_SCATTERPFQPD.
8228 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
8229 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
8230 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
8231 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
8232 __builtin_ia32_scatterpfqps.
8233 (ix86_expand_builtin): Expand new built-ins.
8234 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8235 fix memory access data type.
8236 (*avx512pf_gatherpf<mode>_mask): Ditto.
8237 (*avx512pf_gatherpf<mode>): Ditto.
8238 (avx512pf_scatterpf<mode>): Ditto.
8239 (*avx512pf_scatterpf<mode>_mask): Ditto.
8240 (*avx512pf_scatterpf<mode>): Ditto.
8241 (GATHER_SCATTER_SF_MEM_MODE): New.
8242 (avx512pf_gatherpf<mode>df): Ditto.
8243 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8244 (*avx512pf_scatterpf<mode>df): Ditto.
8246 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8249 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8250 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8253 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8255 * common/config/arm/arm-common.c
8256 (arm_rewrite_mcpu): Handle multiple names.
8258 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8260 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8262 * gimple-builder.h (create_gimple_tmp): Delete.
8264 2014-01-27 Christian Bruel <christian.bruel@st.com>
8266 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
8269 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
8271 * config/pa/pa.md (call): Generate indirect long calls to non-local
8272 functions when outputing 32-bit code.
8273 (call_value): Likewise except for special call to buggy powf function.
8275 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
8276 portable runtime and PIC indirect calls.
8277 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
8278 and PIC call sequences. Use ldo instead of blr to set return register
8279 in PIC call sequence.
8281 2014-01-25 Walter Lee <walt@tilera.com>
8283 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
8284 avoid clobbering a live register.
8286 2014-01-25 Walter Lee <walt@tilera.com>
8288 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
8289 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
8290 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
8291 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
8293 2014-01-25 Walter Lee <walt@tilera.com>
8295 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
8296 arguments on even registers.
8297 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
8299 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
8300 (BIGGEST_ALIGNMENT): Ditto.
8301 (BIGGEST_FIELD_ALIGNMENT): Ditto.
8303 2014-01-25 Walter Lee <walt@tilera.com>
8305 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
8306 insns before bundling.
8307 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8309 2014-01-25 Walter Lee <walt@tilera.com>
8311 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
8312 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
8313 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
8315 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8317 * config/mips/constraints.md (kl): Delete.
8318 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
8319 define expands, using...
8320 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
8321 instructions for MIPS16.
8322 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
8323 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
8325 2014-01-25 Walter Lee <walt@tilera.com>
8327 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
8331 2014-01-25 Walter Lee <walt@tilera.com>
8333 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
8334 (TARGET_EXPAND_TO_RTL_HOOK): Define.
8336 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8338 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
8341 2014-01-25 Jakub Jelinek <jakub@redhat.com>
8343 * print-rtl.c (in_call_function_usage): New var.
8344 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
8345 EXPR_LIST mode as mode and not as reg note name.
8348 * cfgloopmanip.c (copy_loop_info): If
8349 loop->warned_aggressive_loop_optimizations, make sure
8350 the flag is set in target loop too.
8352 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
8354 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
8356 * builtins.def: Likewise.
8357 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
8358 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
8359 * ira.c (ira_setup_eliminable_regset): Likewise.
8360 * omp-low.c (gate_expand_omp): Likewise.
8361 (execute_lower_omp): Likewise.
8362 (diagnose_sb_0): Likewise.
8363 (gate_diagnose_omp_blocks): Likewise.
8364 (simd_clone_clauses_extract): Likewise.
8367 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8369 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
8370 correction for little endian...
8371 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
8374 2014-01-24 Jeff Law <law@redhat.com>
8376 PR tree-optimization/59919
8377 * tree-vrp.c (find_assert_locations_1): Do not register asserts
8378 for non-returning calls.
8380 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
8382 * common/config/aarch64/aarch64-common.c
8383 (aarch64_rewrite_mcpu): Handle multiple names.
8384 * config/aarch64/aarch64.h
8385 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8387 2014-01-24 Dodji Seketeli <dodji@redhat.com>
8389 * input.c (add_file_to_cache_tab): Handle the case where fopen
8392 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
8395 * config/i386/i386.md (pushsf splitter): Get stack adjustment
8396 from push operand if code of push isn't PRE_DEC.
8398 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8401 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
8402 -mquad-memory-atomic. Update -mquad-memory documentation to say
8403 it is only used for non-atomic loads/stores.
8405 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
8406 -mquad-memory or -mquad-memory-atomic switches.
8408 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
8409 -mquad-memory-atomic to ISA 2.07 support.
8411 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
8412 to separate support of normal quad word memory operations (ldq, stq)
8413 from the atomic quad word memory operations.
8415 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
8416 support to separate non-atomic quad word operations from atomic
8417 quad word operations. Disable non-atomic quad word operations in
8418 little endian mode so that we don't have to swap words after the
8419 load and before the store.
8420 (quad_load_store_p): Add comment about atomic quad word support.
8421 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
8422 options printed with -mdebug=reg.
8424 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
8425 -mquad-memory-atomic as the test for whether we have quad word
8426 atomic instructions.
8427 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
8428 or -mp8-vector are used, allow byte/half-word atomic operations.
8430 * config/rs6000/sync.md (load_lockedti): Insure that the address
8431 is a proper indexed or indirect address for the lqarx instruction.
8432 On little endian systems, swap the hi/lo registers after the lqarx
8434 (load_lockedpti): Use indexed_or_indirect_operand predicate to
8435 insure the address is valid for the lqarx instruction.
8436 (store_conditionalti): Insure that the address is a proper indexed
8437 or indirect address for the stqcrx. instruction. On little endian
8438 systems, swap the hi/lo registers before doing the stqcrx.
8440 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
8441 insure the address is valid for the stqcrx. instruction.
8443 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8444 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
8445 type of quad memory support is available.
8447 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
8450 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
8451 there is a danger of looping.
8453 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
8455 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8456 force flag_ira_loop_pressure if set via command line.
8458 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8460 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
8461 (ashr_simd): New builtin handling DI mode.
8462 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
8463 (aarch64_sshr_simddi): New match pattern.
8464 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
8465 (vshrd_n_s64): Likewise.
8466 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
8468 2014-01-23 Nick Clifton <nickc@redhat.com>
8470 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
8471 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
8472 favour of mcu specific scripts.
8473 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
8476 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8477 Alex Velenko <Alex.Velenko@arm.com>
8479 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
8480 (vaddv_s16): Likewise.
8481 (vaddv_s32): Likewise.
8482 (vaddv_u8): Likewise.
8483 (vaddv_u16): Likewise.
8484 (vaddv_u32): Likewise.
8485 (vaddvq_s8): Likewise.
8486 (vaddvq_s16): Likewise.
8487 (vaddvq_s32): Likewise.
8488 (vaddvq_s64): Likewise.
8489 (vaddvq_u8): Likewise.
8490 (vaddvq_u16): Likewise.
8491 (vaddvq_u32): Likewise.
8492 (vaddvq_u64): Likewise.
8493 (vaddv_f32): Likewise.
8494 (vaddvq_f32): Likewise.
8495 (vaddvq_f64): Likewise.
8496 (vmaxv_f32): Likewise.
8497 (vmaxv_s8): Likewise.
8498 (vmaxv_s16): Likewise.
8499 (vmaxv_s32): Likewise.
8500 (vmaxv_u8): Likewise.
8501 (vmaxv_u16): Likewise.
8502 (vmaxv_u32): Likewise.
8503 (vmaxvq_f32): Likewise.
8504 (vmaxvq_f64): Likewise.
8505 (vmaxvq_s8): Likewise.
8506 (vmaxvq_s16): Likewise.
8507 (vmaxvq_s32): Likewise.
8508 (vmaxvq_u8): Likewise.
8509 (vmaxvq_u16): Likewise.
8510 (vmaxvq_u32): Likewise.
8511 (vmaxnmv_f32): Likewise.
8512 (vmaxnmvq_f32): Likewise.
8513 (vmaxnmvq_f64): Likewise.
8514 (vminv_f32): Likewise.
8515 (vminv_s8): Likewise.
8516 (vminv_s16): Likewise.
8517 (vminv_s32): Likewise.
8518 (vminv_u8): Likewise.
8519 (vminv_u16): Likewise.
8520 (vminv_u32): Likewise.
8521 (vminvq_f32): Likewise.
8522 (vminvq_f64): Likewise.
8523 (vminvq_s8): Likewise.
8524 (vminvq_s16): Likewise.
8525 (vminvq_s32): Likewise.
8526 (vminvq_u8): Likewise.
8527 (vminvq_u16): Likewise.
8528 (vminvq_u32): Likewise.
8529 (vminnmv_f32): Likewise.
8530 (vminnmvq_f32): Likewise.
8531 (vminnmvq_f64): Likewise.
8533 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8535 * config/aarch64/aarch64-simd.md
8536 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
8537 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
8538 (*aarch64_mul3_elt<mode>): Likewise.
8539 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
8540 (*aarch64_mul3_elt_to_64v2df): Likewise.
8541 (*aarch64_mla_elt<mode>): Likewise.
8542 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8543 (*aarch64_mls_elt<mode>): Likewise.
8544 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8545 (*aarch64_fma4_elt<mode>): Likewise.
8546 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8547 (*aarch64_fma4_elt_to_64v2df): Likewise.
8548 (*aarch64_fnma4_elt<mode>): Likewise.
8549 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8550 (*aarch64_fnma4_elt_to_64v2df): Likewise.
8551 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
8552 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
8553 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8554 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8555 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
8556 (aarch64_sqdmull_lane<mode>_internal): Likewise.
8557 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
8559 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
8561 * config/aarch64/aarch64-simd.md
8562 (aarch64_be_checked_get_lane<mode>): New define_expand.
8563 * config/aarch64/aarch64-simd-builtins.def
8564 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
8565 New builtin definition.
8566 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
8567 Use new safe be builtin.
8569 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8571 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
8573 (aarch64_be_st1<mode>): Likewise.
8574 (aarch_ld1<VALL:mode>): Define_expand modified.
8575 (aarch_st1<VALL:mode>): Likewise.
8576 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
8577 (UNSPEC_ST1): Likewise.
8579 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
8581 * config/microblaze/microblaze.md: Add trap insn and attribute
8583 2014-01-23 Dodji Seketeli <dodji@redhat.com>
8585 PR preprocessor/58580
8586 * input.h (location_get_source_line): Take an additional line_size
8588 (void diagnostics_file_cache_fini): Declare new function.
8589 * input.c (struct fcache): New type.
8590 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
8591 New static constants.
8592 (diagnostic_file_cache_init, total_lines_num)
8593 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
8594 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
8595 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
8596 (get_next_line, read_next_line, goto_next_line, read_line_num):
8597 New static function definitions.
8598 (diagnostic_file_cache_fini): New function.
8599 (location_get_source_line): Take an additional output line_len
8600 parameter. Re-write using lookup_or_add_file_to_cache_tab and
8602 * diagnostic.c (diagnostic_finish): Call
8603 diagnostic_file_cache_fini.
8604 (adjust_line): Take an additional input parameter for the length
8605 of the line, rather than calculating it with strlen.
8606 (diagnostic_show_locus): Adjust the use of
8607 location_get_source_line and adjust_line with respect to their new
8608 signature. While displaying a line now, do not stop at the first
8609 null byte. Rather, display the zero byte as a space and keep
8610 going until we reach the size of the line.
8611 * Makefile.in: Add vec.o to OBJS-libcommon
8613 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8614 Ilya Tocar <ilya.tocar@intel.com>
8616 * config/i386/avx512fintrin.h (_mm512_kmov): New.
8617 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
8618 (__builtin_ia32_kmov16): Ditto.
8619 * config/i386/i386.md (UNSPEC_KMOV): New.
8622 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8624 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
8625 (_mm512_storeu_si512): Ditto.
8627 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
8630 * rtl.h (get_referenced_operands): Declare.
8631 * recog.c (get_referenced_operands): New function.
8632 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
8633 operands have been referenced when recording LO_SUM references.
8635 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
8637 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
8639 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8641 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
8642 Enable for generic and recent AMD targets.
8644 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8646 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
8647 ARG_SIZE note when adjustment was eliminated.
8649 2014-01-22 Jeff Law <law@redhat.com>
8651 PR tree-optimization/59597
8652 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
8653 in file. Accept new argument REGISTERING and use it to modify
8654 dump output appropriately.
8655 (register_jump_thread): Corresponding changes.
8656 (mark_threaded_blocks): Reinstate code to cancel unprofitable
8657 thread paths involving joiner blocks. Add code to dump cancelled
8658 jump threading paths.
8660 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
8662 PR rtl-optimization/59477
8663 * lra-constraints.c (inherit_in_ebb): Process call for living hard
8664 regs. Update reloads_num and potential_reload_hard_regs for all insns.
8666 2014-01-22 Tom Tromey <tromey@redhat.com>
8668 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
8670 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
8672 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8674 PR rtl-optimization/59896
8675 * lra-constraints.c (process_alt_operands): Check unused note for
8676 matched operands of insn with no output reloads.
8678 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
8680 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
8681 (mips_move_from_gpr_cost): Likewise.
8683 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8685 PR rtl-optimization/59858
8686 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
8687 ira_class_hard_regs_num.
8688 (process_alt_operands): Increase reject for dying matched operand.
8690 2014-01-21 Jakub Jelinek <jakub@redhat.com>
8693 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
8694 smaller than size, perform several stores or loads and stores
8695 at dst + count - size to store or copy all of size bytes, rather
8696 than just last modesize bytes.
8698 2014-01-20 DJ Delorie <dj@redhat.com>
8700 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
8701 that CLOBBERs are REGs before propogating their values.
8703 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
8706 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
8707 (cgraph_inline_failed_type): New function.
8708 * cgraph.h (DEFCIFCODE): Add type.
8709 (cgraph_inline_failed_type_t): New enum.
8710 (cgraph_inline_failed_type): New prototype.
8711 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
8712 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
8713 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
8714 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
8715 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
8716 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
8717 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
8718 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
8719 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
8720 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
8721 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
8722 OPTIMIZATION_MISMATCH.
8723 * tree-inline.c (expand_call_inline): Emit errors during
8724 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
8726 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8729 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
8730 mode attribute in insn output.
8732 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
8734 * output.h (output_constant): Delete.
8735 * varasm.c (output_constant): Make private.
8737 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
8739 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
8741 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8744 * tree.h (fold_builtin_strcat): New prototype.
8745 * builtins.c (fold_builtin_strcat): No longer static. Add len
8746 argument, if non-NULL, don't call c_strlen. Optimize
8747 directly into __builtin_memcpy instead of __builtin_strcpy.
8748 (fold_builtin_2): Adjust fold_builtin_strcat caller.
8749 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
8751 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8753 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8754 for SImode_address_operand operands, having only a REG argument.
8756 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
8758 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
8759 loader name using mbig-endian.
8760 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
8762 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8764 * doc/invoke.texi (-march): Clarify documentation for AArch64.
8768 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
8770 * config/aarch64/aarch64-protos.h
8771 (aarch64_cannot_change_mode_class_ptr): Declare.
8772 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
8773 aarch64_cannot_change_mode_class_ptr): New.
8774 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
8775 backend hook aarch64_cannot_change_mode_class.
8777 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8779 * common/config/aarch64/aarch64-common.c
8780 (aarch64_handle_option): Don't handle any option order logic here.
8781 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
8782 selected_cpu, warn on architecture version mismatch.
8783 (aarch64_override_options): Fix parsing order for option strings.
8785 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8786 Iain Sandoe <iain@codesourcery.com>
8789 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
8790 warning. Amend comment to reflect current functionality.
8792 2014-01-20 Richard Biener <rguenther@suse.de>
8795 * builtins.c (fold_builtin_strcat): Remove case better handled
8796 by tree-ssa-strlen.c.
8798 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
8800 * config/aarch64/aarch64.opt
8801 (mcpu, march, mtune): Make case-insensitive.
8803 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8806 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8807 if operands[1] is a REG or ZERO_EXTEND of a REG.
8809 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
8811 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
8813 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
8815 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
8816 long non-pic millicode calls.
8818 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8820 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
8822 2014-01-19 Kito Cheng <kito@0xlab.org>
8824 * builtins.c (expand_movstr): Check movstr expand done or fail.
8826 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8827 H.J. Lu <hongjiu.lu@intel.com>
8830 * config/i386/i386.md (*lea<mode>): Zero-extend return register
8831 to DImode for zero-extended addresses.
8833 2014-01-19 Jakub Jelinek <jakub@redhat.com>
8835 PR rtl-optimization/57763
8836 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
8837 on the new indirect jump_insn and increment LABEL_NUSES (label).
8839 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
8843 * config.gcc (x86_archs): New variable.
8844 (x86_64_archs): Likewise.
8845 (x86_cpus): Likewise.
8846 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
8847 --with-arch/--with-cpu= options.
8848 Support --with-arch=/--with-cpu={nehalem,westmere,
8849 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
8851 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8853 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
8854 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
8856 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8858 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
8860 2014-01-18 Jakub Jelinek <jakub@redhat.com>
8863 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
8864 clear cpp_get_options (parse_in)->warn_unused_macros for
8865 ix86_target_macros_internal with cpp_define.
8867 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
8869 * jump.c (delete_related_insns): Keep (use (insn))s.
8870 * reorg.c (redundant_insn): Check for barriers too.
8872 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8874 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
8876 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
8878 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
8879 call to $$dyncall when TARGET_LONG_CALLS is true.
8881 2014-01-17 Jeff Law <law@redhat.com>
8883 * ree.c (combine_set_extension): Temporarily disable test for
8884 changing number of hard registers.
8886 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
8889 * ipa-inline-analysis.c (inline_free_summary):
8890 Do not free summary of aliases.
8892 2014-01-17 Jakub Jelinek <jakub@redhat.com>
8895 * gimplify.c (gimplify_expr): Use create_tmp_var
8896 instead of create_tmp_var_raw. If cond doesn't have
8897 integral type, don't add the IFN_ANNOTATE builtin at all.
8899 2014-01-17 Martin Jambor <mjambor@suse.cz>
8902 * ipa-cp.c (prev_edge_clone): New variable.
8903 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
8904 Also resize prev_edge_clone vector.
8905 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
8906 (ipcp_edge_removal_hook): New function.
8907 (ipcp_driver): Register ipcp_edge_removal_hook.
8909 2014-01-17 Andrew Pinski <apinski@cavium.com>
8910 Steve Ellcey <sellcey@mips.com>
8913 * config/mips/mips.c (mips_print_operand): Check operand mode instead
8916 2014-01-17 Jeff Law <law@redhat.com>
8919 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
8920 so that pass_ccp runs first.
8922 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8924 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
8925 (ix86_adjust_cost): Use !TARGET_XXX.
8926 (do_reorder_for_imul): Likewise.
8927 (swap_top_of_ready_list): Likewise.
8928 (ix86_sched_reorder): Likewise.
8930 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8932 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
8933 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
8934 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
8935 (intel_memset): New. Duplicate slm_memset.
8936 (intel_cost): New. Duplicate slm_cost.
8937 (m_INTEL): New macro.
8938 (processor_target_table): Add "intel".
8939 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
8940 with PROCESSOR_INTEL for "intel".
8941 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
8942 PROCESSOR_SILVERMONT.
8943 (ix86_issue_rate): Likewise.
8944 (ix86_adjust_cost): Likewise.
8945 (ia32_multipass_dfa_lookahead): Likewise.
8946 (swap_top_of_ready_list): Likewise.
8947 (ix86_sched_reorder): Likewise.
8948 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
8949 instead of TARGET_OPT_AGU.
8950 * config/i386/i386.h (TARGET_INTEL): New.
8951 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
8952 (processor_type): Add PROCESSOR_INTEL.
8953 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
8954 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
8956 2014-01-17 Marek Polacek <polacek@redhat.com>
8959 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
8962 2014-01-17 Richard Biener <rguenther@suse.de>
8964 PR tree-optimization/46590
8965 * opts.c (default_options_table): Add entries for
8966 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
8967 all enabled at -O1 but not for -Og.
8968 * common.opt (fbranch-count-reg): Remove Init(1).
8969 (fmove-loop-invariants): Likewise.
8970 (ftree-pta): Likewise.
8972 2014-01-17 Jakub Jelinek <jakub@redhat.com>
8974 * config/i386/i386.c (ix86_data_alignment): For compatibility with
8975 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
8976 decls to at least the GCC 4.8 used alignments.
8979 * tree-nested.c (convert_nonlocal_reference_stmt,
8980 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
8981 of GIMPLE_BIND stmts, adjust associated decls.
8983 2014-01-17 Richard Biener <rguenther@suse.de>
8985 PR tree-optimization/46590
8986 * vec.h (vec<>::bseach): New member function implementing
8987 binary search according to C89 bsearch.
8988 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
8989 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
8990 bitmap pointer again. Make accesses_in_loop a flat array.
8991 (mem_ref_obstack): New global.
8992 (outermost_indep_loop): Adjust for mem_ref->stored changes.
8993 (mark_ref_stored): Likewise.
8994 (ref_indep_loop_p_2): Likewise.
8995 (set_ref_stored_in_loop): New helper function.
8996 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
8997 (memref_free): Adjust.
8998 (record_mem_ref_loc): Simplify.
8999 (gather_mem_refs_stmt): Adjust.
9000 (sort_locs_in_loop_postorder_cmp): New function.
9001 (analyze_memory_references): Sort accesses_in_loop after
9002 loop postorder number.
9003 (find_ref_loc_in_loop_cmp): New function.
9004 (for_all_locs_in_loop): Find relevant cluster of locs in
9005 accesses_in_loop and iterate without recursion.
9006 (execute_sm): Avoid uninit warning.
9007 (struct ref_always_accessed): Simplify.
9008 (ref_always_accessed::operator ()): Likewise.
9009 (ref_always_accessed_p): Likewise.
9010 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
9011 loop postorder numbers here.
9012 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
9015 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9018 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
9019 on decls for which assemble_alias has been called.
9021 2014-01-17 Nick Clifton <nickc@redhat.com>
9023 * config/msp430/msp430.opt: (mcpu): New option.
9024 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
9025 (msp430_option_override): Parse target_cpu. If the MCU name
9026 matches a generic string, clear target_mcu.
9027 (msp430_attr): Allow numeric interrupt values up to 63.
9028 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
9029 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
9031 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
9033 * config/msp430/msp430.md (popm): Use %J rather than %I.
9034 (addsi3): Use msp430_nonimmediate_operand for operand 2.
9035 (addhi_cy_i): Use immediate_operand for operand 2.
9036 * doc/invoke.texi: Document -mcpu option.
9038 2014-01-17 Richard Biener <rguenther@suse.de>
9040 PR rtl-optimization/38518
9041 * df.h (df_analyze_loop): Declare.
9042 * df-core.c: Include cfgloop.h.
9043 (df_analyze_1): Split out main part of df_analyze.
9044 (df_analyze): Adjust.
9045 (loop_inverted_post_order_compute): New function.
9046 (loop_post_order_compute): Likewise.
9047 (df_analyze_loop): New function avoiding whole-function
9049 * loop-invariant.c (find_defs): Use df_analyze_loop.
9050 (find_invariants): Adjust.
9051 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
9053 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
9055 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
9056 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
9058 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
9060 * ipa-ref.c (ipa_remove_stmt_references): Fix references
9061 traversal when removing references.
9063 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
9066 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
9068 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
9071 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
9072 pushing a reload for an autoinc when we had previously reloaded an
9073 inner part of the address.
9075 2014-01-16 Jakub Jelinek <jakub@redhat.com>
9077 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
9079 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
9080 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
9081 when not giving up or versioning for alias only because of
9083 (vect_analyze_data_ref_dependences): Set to true.
9084 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
9086 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
9087 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
9091 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
9094 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
9095 operand 0 predicate for gathers, use a new pseudo as subtarget.
9097 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9100 * lra-constraints.c (process_alt_operands): Add printing debug info.
9101 Check absence of input/output reloads for matched operands too.
9103 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9105 PR rtl-optimization/59835
9106 * ira.c (ira_init_register_move_cost): Increase cost for
9109 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
9111 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
9113 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
9116 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
9117 non-register objects. Use gen_(high/low)part more consistently.
9120 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9123 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
9124 endian support, remove tests for WORDS_BIG_ENDIAN.
9125 (p8_mfvsrd_3_<mode>): Likewise.
9126 (reload_gpr_from_vsx<mode>): Likewise.
9127 (reload_gpr_from_vsxsf): Likewise.
9128 (p8_mfvsrd_4_disf): Likewise.
9130 2014-01-16 Richard Biener <rguenther@suse.de>
9132 PR rtl-optimization/46590
9133 * lcm.c (compute_antinout_edge): Use postorder iteration.
9134 (compute_laterin): Use inverted postorder iteration.
9136 2014-01-16 Nick Clifton <nickc@redhat.com>
9139 * varasm.c (output_constant): Return the number of bytes actually
9141 (output_constructor_array_range): Update the field size with the
9142 number of bytes emitted by output_constant.
9143 (output_constructor_regular_field): Likewise. Also do not
9144 complain if the total number of bytes emitted is now greater
9145 than the expected fieldpos.
9146 * output.h (output_constant): Update prototype and descriptive comment.
9148 2014-01-16 Marek Polacek <polacek@redhat.com>
9151 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
9152 it is error_mark_node.
9154 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
9156 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
9157 VALID_AVX256_REG_OR_OI_MODE.
9159 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
9161 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
9162 current procedure should be profiled.
9164 2014-01-15 Andrew Pinski <apinski@cavium.com>
9166 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
9167 of moving from/to the STACK_REG register class.
9169 2014-01-15 Richard Henderson <rth@redhat.com>
9172 * reginfo.c (global_regs_decl): Globalize.
9173 * rtl.h (global_regs_decl): Declare.
9174 * ira.c (do_reload): Diagnose frame_pointer_needed and it
9175 reserved via global_regs.
9177 2014-01-15 Teresa Johnson <tejohnson@google.com>
9179 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
9181 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9183 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
9184 and vmulosh rather than call gen_vec_widen_smult_*.
9185 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
9186 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
9187 (vec_widen_smult_even_v16qi): Likewise.
9188 (vec_widen_umult_even_v8hi): Likewise.
9189 (vec_widen_smult_even_v8hi): Likewise.
9190 (vec_widen_umult_odd_v16qi): Likewise.
9191 (vec_widen_smult_odd_v16qi): Likewise.
9192 (vec_widen_umult_odd_v8hi): Likewise.
9193 (vec_widen_smult_odd_v8hi): Likewise.
9194 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
9195 vmuloub rather than call gen_vec_widen_umult_*.
9196 (vec_widen_umult_lo_v16qi): Likewise.
9197 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
9198 vmulosb rather than call gen_vec_widen_smult_*.
9199 (vec_widen_smult_lo_v16qi): Likewise.
9200 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
9201 rather than call gen_vec_widen_umult_*.
9202 (vec_widen_umult_lo_v8hi): Likewise.
9203 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
9204 rather than call gen_vec_widen_smult_*.
9205 (vec_widen_smult_lo_v8hi): Likewise.
9207 2014-01-15 Jeff Law <law@redhat.com>
9209 PR tree-optimization/59747
9210 * ree.c (find_and_remove_re): Properly handle case where a second
9211 eliminated extension requires widening a copy created for elimination
9212 of a prior extension.
9213 (combine_set_extension): Ensure that the number of hard regs needed
9214 for a destination register does not change when we widen it.
9216 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
9218 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
9219 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
9220 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
9221 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
9222 (avr-*-rtems*): Likewise.
9223 (bfin*-rtems*): Likewise.
9224 (moxie-*-rtems*): Likewise.
9225 (h8300-*-rtems*): Likewise.
9226 (i[34567]86-*-rtems*): Likewise.
9227 (lm32-*-rtems*): Likewise.
9228 (m32r-*-rtems*): Likewise.
9229 (m68k-*-rtems*): Likewise.
9230 (microblaze*-*-rtems*): Likewise.
9231 (mips*-*-rtems*): Likewise.
9232 (powerpc-*-rtems*): Likewise.
9233 (sh-*-rtems*): Likewise.
9234 (sparc-*-rtems*): Likewise.
9235 (sparc64-*-rtems*): Likewise.
9236 (v850-*-rtems*): Likewise.
9237 (m32c-*-rtems*): Likewise.
9239 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9241 PR rtl-optimization/59511
9242 * ira.c (ira_init_register_move_cost): Use memory costs for some
9243 cases of register move cost calculations.
9244 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9245 instead of BB frequency.
9246 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9247 * lra-assigns.c (find_hard_regno_for): Ditto.
9249 2014-01-15 Richard Biener <rguenther@suse.de>
9251 PR tree-optimization/59822
9252 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9253 (vectorizable_load): Use it to hoist defs of uses of invariant
9254 loads out of the loop.
9256 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9257 Kugan Vivekanandarajah <kuganv@linaro.org>
9260 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9263 2014-01-15 Richard Biener <rguenther@suse.de>
9265 PR rtl-optimization/59802
9266 * lcm.c (compute_available): Use inverted postorder to seed
9267 the initial worklist.
9269 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9272 * config/s390/s390.c (s390_preferred_reload_class): Don't return
9273 ADDR_REGS for invalid symrefs in non-PIC code.
9275 2014-01-15 Jakub Jelinek <jakub@redhat.com>
9278 * builtins.c (determine_block_size): Initialize *probable_max_size
9279 even if len_rtx is CONST_INT.
9281 2014-01-14 Andrew Pinski <apinski@cavium.com>
9283 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
9284 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
9285 (cortexa53_tunings): Likewise.
9286 (aarch64_sched_issue_rate): New function.
9287 (TARGET_SCHED_ISSUE_RATE): Define.
9289 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9291 * ira-costs.c (find_costs_and_classes): Add missed
9292 ira_init_register_move_cost_if_necessary.
9294 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9297 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
9299 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
9302 * config/i386/i386.c (type_natural_mode): Add a bool parameter
9303 to indicate if type is used for function return value. Warn ABI
9304 change if the vector mode isn't available for function return value.
9305 (ix86_function_arg_advance): Pass false to type_natural_mode.
9306 (ix86_function_arg): Likewise.
9307 (ix86_gimplify_va_arg): Likewise.
9308 (function_arg_32): Don't warn ABI change.
9309 (ix86_function_value): Pass true to type_natural_mode.
9310 (ix86_return_in_memory): Likewise.
9311 (ix86_struct_value_rtx): Removed.
9312 (TARGET_STRUCT_VALUE_RTX): Likewise.
9314 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9316 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
9317 converting a conditional jump into a conditional return.
9319 2014-01-14 Richard Biener <rguenther@suse.de>
9321 PR tree-optimization/58921
9322 PR tree-optimization/59006
9323 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
9324 hoisting invariant stmts.
9325 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
9326 invariant loads on the preheader edge if possible.
9328 2014-01-14 Joey Ye <joey.ye@arm.com>
9330 * doc/plugin.texi (Building GCC plugins): Update to C++.
9332 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
9334 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
9335 (_mm_rcp28_round_ss): Ditto.
9336 (_mm_rsqrt28_round_sd): Ditto.
9337 (_mm_rsqrt28_round_ss): Ditto.
9338 (_mm_rcp28_sd): Ditto.
9339 (_mm_rcp28_ss): Ditto.
9340 (_mm_rsqrt28_sd): Ditto.
9341 (_mm_rsqrt28_ss): Ditto.
9342 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
9343 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
9344 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
9345 (IX86_BUILTIN_RCP28SD): Ditto.
9346 (IX86_BUILTIN_RCP28SS): Ditto.
9347 (IX86_BUILTIN_RSQRT28SD): Ditto.
9348 (IX86_BUILTIN_RSQRT28SS): Ditto.
9349 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
9350 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
9351 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
9352 (ix86_expand_special_args_builtin): Expand new FTYPE.
9353 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
9354 (srcp14<mode>): Make insn unary.
9355 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
9356 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
9357 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
9358 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
9359 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
9360 Fix rounding: make it SAE only.
9361 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9363 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9365 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
9366 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
9367 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
9368 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
9369 (round_saeonly_mask_scalar_operand4): Ditto.
9370 (round_saeonly_mask_scalar_op3): Ditto.
9371 (round_saeonly_mask_scalar_op4): Ditto.
9373 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9375 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9376 Implement -maltivec=be for vec_insert and vec_extract.
9378 2014-01-10 DJ Delorie <dj@redhat.com>
9380 * config/msp430/msp430.md (call_internal): Don't allow memory
9381 references with SP as the base register.
9382 (call_value_internal): Likewise.
9383 * config/msp430/constraints.md (Yc): New. For memory references
9384 that don't use SP as a base register.
9386 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
9387 "an integer without a # prefix"
9388 * config/msp430/msp430.md (epilogue_helper): Use it.
9390 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9393 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
9394 AVX512F gather builtins.
9395 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
9396 on gather decls with INTEGER_TYPE masktype.
9397 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
9398 directly into the builtin rather than hoisting it before loop.
9400 PR tree-optimization/59387
9401 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
9402 (scev_const_prop): If folded_casts and type has undefined overflow,
9403 use force_gimple_operand instead of force_gimple_operand_gsi and
9404 for each added stmt if it is assign with
9405 arith_code_with_undefined_signed_overflow, call
9406 rewrite_to_defined_overflow.
9407 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
9408 gimple-fold.h instead.
9409 (arith_code_with_undefined_signed_overflow,
9410 rewrite_to_defined_overflow): Moved to ...
9411 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
9412 rewrite_to_defined_overflow): ... here. No longer static.
9413 Include gimplify-me.h.
9414 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
9415 rewrite_to_defined_overflow): New prototypes.
9417 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9419 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
9421 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
9423 * builtins.c (get_object_alignment_2): Minor tweak.
9424 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
9426 2014-01-13 Christian Bruel <christian.bruel@st.com>
9428 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
9429 optimized non constant lengths.
9431 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9434 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
9435 load as __atomic_load_N if possible.
9437 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9439 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
9441 (rs6000_expand_builtin): Adjust call.
9443 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9446 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
9447 * config/rs6000/rs6000.c: Include target-globals.h.
9448 (rs6000_set_current_function): Instead of doing target_reinit
9449 unconditionally, use save_target_globals_default_opts and
9450 restore_target_globals.
9452 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
9454 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
9455 (rs6000_expand_builtin): Handle mffs and mtfsf.
9456 (rs6000_init_builtins): Define mffs and mtfsf.
9457 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
9458 (rs6000_mffs): New pattern.
9459 (rs6000_mtfsf): New pattern.
9461 2014-01-11 Bin Cheng <bin.cheng@arm.com>
9463 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
9464 Start narrowing with START. Apply candidate-use pair
9465 and check overall cost in narrowing.
9466 (iv_ca_prune): Pass new argument.
9468 2014-01-10 Jeff Law <law@redhat.com>
9471 * ree.c (combine_reaching_defs): Ensure the defining statement
9472 occurs before the extension when optimizing extensions with
9473 different source and destination hard registers.
9475 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9478 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
9479 vtables into the type inheritance graph.
9481 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9483 PR rtl-optimization/59754
9484 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
9485 modes in the REGNO != REGNO case.
9487 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9489 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
9491 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9493 PR tree-optimization/59745
9494 * tree-predcom.c (tree_predictive_commoning_loop): Call
9495 free_affine_expand_cache if giving up because components is NULL.
9497 * target-globals.c (save_target_globals): Allocate < 4KB structs using
9498 GC in payload of target_globals struct instead of allocating them on
9499 the heap and the larger structs separately using GC.
9500 * target-globals.h (struct target_globals): Make regs, hard_regs,
9501 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
9502 of GTY((skip)) and change type to void *.
9503 (reset_target_globals): Cast loads from those fields to corresponding
9506 2014-01-10 Steve Ellcey <sellcey@mips.com>
9509 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
9510 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
9511 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
9513 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
9516 * aarch64-modes.def (CC_Zmode): New flags mode.
9517 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
9518 represents an equality.
9519 (aarch64_get_condition_code): Handle CC_Zmode.
9520 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
9522 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9524 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
9525 extraction in good case.
9527 2014-01-10 Richard Biener <rguenther@suse.de>
9529 PR tree-optimization/59374
9530 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
9531 checking after SLP discovery. Mark stmts not participating
9532 in any SLP instance properly.
9534 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9536 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
9537 when handling a SET rtx.
9539 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9541 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
9542 (cortex-a57): Likewise.
9543 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
9545 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9547 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
9548 non-iwmmxt builtins.
9550 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9554 * ipa-devirt.c record_target_from_binfo): Take as argument
9555 stack of binfos and lookup matching one for virtual inheritance.
9556 (possible_polymorphic_call_targets_1): Update.
9558 2014-01-10 Huacai Chen <chenhc@lemote.com>
9560 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
9561 kernel strings for Loongson-2E/2F/3A.
9563 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9566 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
9567 is_gimple_call before calling gimple_call_internal_p.
9569 2014-01-09 Steve Ellcey <sellcey@mips.com>
9571 * Makefile.in (TREE_FLOW_H): Remove.
9572 (TREE_SSA_H): Add file names from tree-flow.h.
9573 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
9574 * tree.h: Remove tree-flow.h reference.
9575 * hash-table.h: Remove tree-flow.h reference.
9576 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
9577 reference with tree-ssa-loop.h.
9579 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9581 * doc/invoke.texi: Add -maltivec={be,le} options, and document
9582 default element-order behavior for -maltivec.
9583 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
9584 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
9585 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
9586 when targeting big endian, at least for now.
9587 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
9589 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9592 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
9593 var satisfies use_register_for_decl, just take into account type
9594 alignment, rather than decl alignment.
9596 PR tree-optimization/59622
9597 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
9598 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
9599 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
9600 Don't devirtualize for inplace at all. For targets.length () == 1,
9601 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
9603 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9605 * config/i386/i386.md (cpu): Remove the unused btver1.
9607 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9609 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
9611 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9614 * tree-core.h (struct target_globals): New forward declaration.
9615 (struct tree_target_option): Add globals field.
9616 * tree.h (TREE_TARGET_GLOBALS): Define.
9617 (prepare_target_option_nodes_for_pch): New prototype.
9618 * target-globals.h (struct target_globals): Define even if
9620 * tree.c (prepare_target_option_node_for_pch,
9621 prepare_target_option_nodes_for_pch): New functions.
9622 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
9623 * config/i386/i386.c: Include target-globals.h.
9624 (ix86_set_current_function): Instead of doing target_reinit
9625 unconditionally, use save_target_globals_default_opts and
9626 restore_target_globals.
9628 2014-01-09 Richard Biener <rguenther@suse.de>
9630 PR tree-optimization/59715
9631 * tree-cfg.h (split_critical_edges): Declare.
9632 * tree-cfg.c (split_critical_edges): Export.
9633 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
9635 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
9637 * cfgexpand.c (expand_stack_vars): Optionally disable
9638 asan stack protection.
9639 (expand_used_vars): Likewise.
9640 (partition_stack_vars): Likewise.
9641 * asan.c (asan_emit_stack_protection): Optionally disable
9642 after return stack usage.
9643 (instrument_derefs): Optionally disable memory access instrumentation.
9644 (instrument_builtin_call): Likewise.
9645 (instrument_strlen_call): Likewise.
9646 (asan_protect_global): Optionally disable global variables protection.
9647 * doc/invoke.texi: Added doc for new options.
9648 * params.def: Added new options.
9649 * params.h: Likewise.
9651 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9653 PR rtl-optimization/59724
9654 * ifcvt.c (cond_exec_process_if_block): Don't call
9655 flow_find_head_matching_sequence with 0 longest_match.
9656 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
9657 non-active insns if !stop_after.
9658 (try_head_merge_bb): Revert 2014-01-07 changes.
9660 2014-01-08 Jeff Law <law@redhat.com>
9662 * ree.c (get_sub_rtx): New function, extracted from...
9663 (merge_def_and_ext): Here.
9664 (combine_reaching_defs): Use get_sub_rtx.
9666 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
9668 * cgraph.h (varpool_variable_node): Do not choke on null node.
9670 2014-01-08 Catherine Moore <clm@codesourcery.com>
9672 * config/mips/mips.md (simple_return): Attempt to use JRC
9674 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
9676 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9678 PR rtl-optimization/59137
9679 * reorg.c (steal_delay_list_from_target): Call update_block for
9681 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
9683 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9685 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
9686 two duplicate entries.
9688 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9691 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
9693 * config/mips/mips.c (mips_truncated_op_cost): New function.
9694 (mips_rtx_costs): Adjust test for BADDU.
9695 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
9697 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
9699 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
9700 (*baddu_si): ...this new pattern.
9702 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9705 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
9707 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9710 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
9712 (expand_expr, expand_normal): Adjust.
9713 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
9714 inner_reference_p. Use inner_reference_p to expand inner references.
9715 (store_expr): Adjust.
9716 * cfgexpand.c (expand_call_stmt): Adjust.
9718 2014-01-08 Rong Xu <xur@google.com>
9720 * gcov-io.c (gcov_var): Move from gcov-io.h.
9721 (gcov_position): Ditto.
9722 (gcov_is_error): Ditto.
9723 (gcov_rewrite): Ditto.
9724 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
9725 only part to libgcc/libgcov.h.
9727 2014-01-08 Marek Polacek <polacek@redhat.com>
9730 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
9732 2014-01-08 Marek Polacek <polacek@redhat.com>
9735 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
9737 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9739 PR rtl-optimization/59649
9740 * stor-layout.c (get_mode_bounds): For BImode return
9741 0 and STORE_FLAG_VALUE.
9743 2014-01-08 Richard Biener <rguenther@suse.de>
9746 * gimple.h (is_gimple_builtin_call): Remove.
9747 (gimple_builtin_call_types_compatible_p): New.
9748 (gimple_call_builtin_p): New overload.
9749 * gimple.c (is_gimple_builtin_call): Remove.
9750 (validate_call): Rename to ...
9751 (gimple_builtin_call_types_compatible_p): ... this and export. Also
9753 (validate_type): New static function.
9754 (gimple_call_builtin_p): New overload and adjust.
9755 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
9756 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
9757 (gimple_fold_stmt_to_constant_1): Likewise.
9758 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
9760 2014-01-08 Richard Biener <rguenther@suse.de>
9763 * gimplify.c (gimplify_expr): Gimplify register-register type
9764 VIEW_CONVERT_EXPRs to separate stmts.
9766 2014-01-07 Jeff Law <law@redhat.com>
9769 * ree.c (combine_set_extension): Handle case where source
9770 and destination registers in an extension insn are different.
9771 (combine_reaching_defs): Allow source and destination registers
9772 in extension to be different under limited circumstances.
9773 (add_removable_extension): Remove restriction that the
9774 source and destination registers in the extension are the same.
9775 (find_and_remove_re): Emit a copy from the extension's
9776 destination to its source after the defining insn if
9777 the source and destination registers are different.
9780 * ifcvt.c (merge_if_block): If we are merging a block with more than
9781 one successor with a block with no successors, remove any BARRIER
9782 after the second block.
9784 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
9786 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
9788 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
9791 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
9792 for 14-bit register offsets when INT14_OK_STRICT is false.
9794 2014-01-07 Roland Stigge <stigge@antcom.de>
9795 Michael Meissner <meissner@linux.vnet.ibm.com>
9798 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
9799 Only check TFmode for SPE constants. Don't check TImode or TDmode.
9801 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
9803 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
9806 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
9808 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
9809 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
9810 rtx is const0_rtx or not.
9812 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
9815 * target-globals.c (save_target_globals): Remove this_fn_optab
9817 * toplev.c: Include optabs.h.
9818 (target_reinit): Temporarily restore the global options if another
9819 set of options are in force.
9821 2014-01-07 Jakub Jelinek <jakub@redhat.com>
9823 PR rtl-optimization/58668
9824 * cfgcleanup.c (flow_find_cross_jump): Don't count
9825 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
9826 to determine what is counted.
9827 (flow_find_head_matching_sequence): Use active_insn_p to determine
9829 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
9831 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
9832 determine what is counted.
9834 PR tree-optimization/59643
9835 * tree-predcom.c (split_data_refs_to_components): If one dr is
9836 read and one write, determine_offset fails and the write isn't
9837 in the bad component, just put the read into the bad component.
9839 2014-01-07 Mike Stump <mikestump@comcast.net>
9840 Jakub Jelinek <jakub@redhat.com>
9843 * tree-core.h (struct tree_optimization_option): Change optabs
9844 type from unsigned char * to void *.
9845 * optabs.c (init_tree_optimization_optabs): Adjust
9846 TREE_OPTIMIZATION_OPTABS initialization.
9848 2014-01-06 Jakub Jelinek <jakub@redhat.com>
9851 * config/i386/i386.h (struct machine_function): Add
9852 no_drap_save_restore field.
9853 * config/i386/i386.c (ix86_save_reg): Use
9854 !cfun->machine->no_drap_save_restore instead of
9855 crtl->stack_realign_needed.
9856 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
9857 this function clears frame_pointer_needed. Set
9858 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
9859 and DRAP reg is needed.
9861 2014-01-06 Marek Polacek <polacek@redhat.com>
9864 * doc/implement-c.texi: Mention that other integer types are
9865 permitted as bit-field types in strictly conforming mode.
9867 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
9869 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
9872 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
9874 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
9876 2014-01-06 Martin Jambor <mjambor@suse.cz>
9879 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
9881 * ipa-prop.c (ipa_print_node_params): Fix indentation.
9883 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
9887 * var-tracking.c (add_stores): Preserve the value of the source even if
9888 we don't record the store.
9890 2014-01-06 Terry Guo <terry.guo@arm.com>
9892 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
9894 2014-01-05 Iain Sandoe <iain@codesourcery.com>
9897 * config/darwin.c (darwin_function_section): Adjust return values to
9898 correspond to optimisation changes made in r206070.
9900 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
9902 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
9903 from prefetch_block tune setting.
9904 (nocona_cost): Correct size of prefetch block to 64.
9906 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
9908 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
9909 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
9910 used to save the static chain register in the computation of the offset
9911 from which the FP registers need to be restored.
9913 2014-01-04 Jakub Jelinek <jakub@redhat.com>
9915 PR tree-optimization/59519
9916 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
9917 ICE if get_current_def (current_new_name) is already non-NULL, as long
9918 as it is a phi result of some other phi in *new_exit_bb that has
9921 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
9922 or vmovdqu* for misaligned_operand.
9923 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
9924 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
9925 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
9926 aligned_mem for AVX512F masked aligned load and store builtins and for
9929 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
9931 PR tree-optimization/59651
9932 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
9933 Address range for negative step should be added by TYPE_SIZE_UNIT.
9935 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
9937 * config/m68k/m68k.c (handle_move_double): Handle pushes with
9938 overlapping registers also for registers other than the stack pointer.
9940 2014-01-03 Marek Polacek <polacek@redhat.com>
9943 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
9946 2014-01-03 Jakub Jelinek <jakub@redhat.com>
9949 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
9952 * config/i386/i386.md (MODE_SIZE): New mode attribute.
9953 (push splitter): Use <P:MODE_SIZE> instead of
9954 GET_MODE_SIZE (<P:MODE>mode).
9955 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
9956 (mov -1, reg peephole2): Likewise.
9957 * config/i386/sse.md (*mov<mode>_internal,
9958 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
9959 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
9960 *<code><mode>3, *andnot<mode>3<mask_name>,
9961 <mask_codefor><code><mode>3<mask_name>): Likewise.
9962 * config/i386/subst.md (mask_mode512bit_condition,
9963 sd_mask_mode512bit_condition): Likewise.
9965 2014-01-02 Xinliang David Li <davidxl@google.com>
9967 PR tree-optimization/59303
9968 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
9969 (dump_predicates): Better output format.
9970 (pred_equal_p): New function.
9971 (is_neq_relop_p): Ditto.
9972 (is_neq_zero_form_p): Ditto.
9973 (pred_expr_equal_p): Ditto.
9974 (pred_neg_p): Ditto.
9975 (simplify_pred): Ditto.
9976 (simplify_preds_2): Ditto.
9977 (simplify_preds_3): Ditto.
9978 (simplify_preds_4): Ditto.
9979 (simplify_preds): Ditto.
9981 (push_to_worklist): Ditto.
9982 (get_pred_info_from_cmp): Ditto.
9983 (is_degenerated_phi): Ditto.
9984 (normalize_one_pred_1): Ditto.
9985 (normalize_one_pred): Ditto.
9986 (normalize_one_pred_chain): Ditto.
9987 (normalize_preds): Ditto.
9988 (normalize_cond_1): Remove function.
9989 (normalize_cond): Ditto.
9990 (is_gcond_subset_of): Ditto.
9991 (is_subset_of_any): Ditto.
9992 (is_or_set_subset_of): Ditto.
9993 (is_and_set_subset_of): Ditto.
9994 (is_norm_cond_subset_of): Ditto.
9995 (pred_chain_length_cmp): Ditto.
9996 (convert_control_dep_chain_into_preds): Type change.
9997 (find_predicates): Ditto.
9998 (find_def_preds): Ditto.
9999 (destroy_predicates_vecs): Ditto.
10000 (find_matching_predicates_in_rest_chains): Ditto.
10001 (use_pred_not_overlap_with_undef_path_pred): Ditto.
10002 (is_pred_expr_subset): Ditto.
10003 (is_pred_chain_subset_of): Ditto.
10004 (is_included_in): Ditto.
10005 (is_superset_of): Ditto.
10007 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10009 Update copyright years.
10011 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10013 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
10014 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
10015 config/arc/arc.md, config/arc/arc.opt,
10016 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
10017 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
10018 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
10019 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
10020 config/linux-protos.h, config/linux.c, config/winnt-c.c,
10021 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
10022 vtable-verify.c, vtable-verify.h: Use the standard form for the
10025 2014-01-02 Tobias Burnus <burnus@net-b.de>
10027 * gcc.c (process_command): Update copyright notice dates.
10028 * gcov-dump.c: Ditto.
10030 * doc/cpp.texi: Bump @copying's copyright year.
10031 * doc/cppinternals.texi: Ditto.
10032 * doc/gcc.texi: Ditto.
10033 * doc/gccint.texi: Ditto.
10034 * doc/gcov.texi: Ditto.
10035 * doc/install.texi: Ditto.
10036 * doc/invoke.texi: Ditto.
10038 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10040 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
10042 2014-01-01 Jakub Jelinek <jakub@redhat.com>
10044 * config/i386/sse.md (*mov<mode>_internal): Guard
10045 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
10047 PR rtl-optimization/59647
10048 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
10049 new_rtx into UNSIGNED_FLOAT rtxes.
10051 Copyright (C) 2014 Free Software Foundation, Inc.
10053 Copying and distribution of this file, with or without modification,
10054 are permitted in any medium without royalty provided the copyright
10055 notice and this notice are preserved.