1 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3 * wide-int.cc (UTItype): Define.
4 (UDWtype): Define for appropriate W_TYPE_SIZE.
6 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
8 PR tree-optimization/59100
9 * tree-ssa-phiopt.c: Include tree-inline.h.
10 (neutral_element_p, absorbing_element_p): New functions.
11 (value_replacement): Handle conditional binary operations with a
12 neutral or absorbing element.
14 2014-05-08 Richard Biener <rguenther@suse.de>
16 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
17 folding the expression.
18 (valueize_expr): Remove.
19 (visit_reference_op_load): Do not valueize the result of
21 (simplify_binary_expression): Likewise.
22 (simplify_unary_expression): Likewise.
24 2014-05-08 Richard Biener <rguenther@suse.de>
26 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
27 looking at TYPE_ARG_TYPES.
29 2014-05-08 Richard Biener <rguenther@suse.de>
31 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
32 pointer propagation special-case.
34 2014-05-08 Bin Cheng <bin.cheng@arm.com>
36 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
37 core part of address expressions.
39 2014-05-08 Alan Modra <amodra@gmail.com>
42 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
43 loads and stores when -mno-strict-align at any alignment.
44 (expand_block_clear): Similarly. Also correct calculation of
47 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
50 * tree-complex.c (expand_complex_move): Keep line info when expanding
52 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
53 of complex expression. Use new argument to display correct location
54 for values coming from phi statement.
55 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
56 (warn_uninitialized_phi): Pass location of phi argument to
58 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
59 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
61 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
63 * config/rs6000/predicates.md (indexed_address_mem): New.
64 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
65 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
66 fpstore_ux, fpstore_u.
67 (sign_extend, indexed, update): New.
69 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
70 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
71 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
72 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
73 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
74 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
75 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
76 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
77 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
78 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
79 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
80 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
81 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
82 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
83 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
85 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
86 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
87 *vsx_extract_<mode>_store): Adjust.
88 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
89 is_cracked_insn, insn_must_be_first_in_group,
90 insn_must_be_last_in_group): Adjust.
92 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
94 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
95 ppc440-fpstore): Adjust.
96 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
97 ppc476-fpstore): Adjust.
98 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
99 ppc601-fpstore): Adjust.
100 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
102 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
104 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
105 ppc7450-fpstore): Adjust.
106 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
107 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
108 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
110 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
111 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
112 cell-fpstore, cell-fpstore-update): Adjust.
113 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
114 ppce300c3_store, ppce300c3_fpstore): Adjust.
115 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
116 e500mc_fpstore): Adjust.
117 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
118 e500mc64_store, e500mc64_fpstore): Adjust.
119 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
120 e5500_fpstore): Adjust.
121 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
122 e6500_fpstore): Adjust.
123 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
125 * config/rs6000/power4.md (power4-load, power4-load-ext,
126 power4-load-ext-update, power4-load-ext-update-indexed,
127 power4-load-update-indexed, power4-load-update, power4-fpload,
128 power4-fpload-update, power4-store, power4-store-update,
129 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
131 * config/rs6000/power5.md (power5-load, power5-load-ext,
132 power5-load-ext-update, power5-load-ext-update-indexed,
133 power5-load-update-indexed, power5-load-update, power5-fpload,
134 power5-fpload-update, power5-store, power5-store-update,
135 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
137 * config/rs6000/power6.md (power6-load, power6-load-ext,
138 power6-load-update, power6-load-update-indexed,
139 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
140 power6-fpload-update, power6-store, power6-store-update,
141 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
143 * config/rs6000/power7.md (power7-load, power7-load-ext,
144 power7-load-update, power7-load-update-indexed,
145 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
146 power7-fpload-update, power7-store, power7-store-update,
147 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
149 * config/rs6000/power8.md (power8-load, power8-load-update,
150 power8-load-ext, power8-load-ext-update, power8-fpload,
151 power8-fpload-update, power8-store, power8-store-update-indexed,
152 power8-fpstore, power8-fpstore-update): Adjust.
153 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
155 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
156 titan_lsu_store, titan_lsu_fpstore): Adjust.
157 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
159 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
162 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
163 unrolled byte insns. Emit address increments after move insns.
165 2014-05-07 David Malcolm <dmalcolm@redhat.com>
167 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
168 const_gimple, rather than a gimple.
169 (gimple_call_builtin_p): Likewise, for the three variants.
171 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
172 (gimple_call_builtin_p): Likewise, for the three variants.
174 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
176 PR tree-optimization/61095
177 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
179 2014-05-07 Richard Biener <rguenther@suse.de>
181 PR tree-optimization/61034
182 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
183 (maybe_skip_until): Use translate to take into account
184 lattices when trying to do disambiguations.
185 (get_continuation_for_phi_1): Likewise.
186 (get_continuation_for_phi): Adjust for added translate
188 (walk_non_aliased_vuses): Likewise.
189 * tree-ssa-alias.h (get_continuation_for_phi): Adjust
191 (walk_non_aliased_vuses): Likewise.
192 (call_may_clobber_ref_p_1): Declare.
193 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also
194 disambiguate against calls. Stop early if we are
195 only supposed to disambiguate.
196 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
198 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
200 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
201 Emit an error when the function has arguments.
203 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
205 * cfgloop.h (unswitch_loops): Remove.
206 * doc/passes.texi: Remove references to loop-unswitch.c
207 * timevar.def (TV_LOOP_UNSWITCH): Remove.
209 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
211 * tree-vect-data-refs.c (vect_grouped_load_supported): New
212 check for loads group of length 3.
213 (vect_permute_load_chain): New permutations for loads group of
215 * tree-vect-stmts.c (vect_model_load_cost): Change cost
216 of vec_perm_shuffle for the new permutations.
218 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
220 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
221 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
222 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
223 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
224 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
225 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
226 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
227 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
229 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
231 * loop-unswitch.c: Delete.
233 2014-05-07 Richard Biener <rguenther@suse.de>
235 * config.gcc: Always set need_64bit_hwint to yes.
237 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
239 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
240 of using optimize_size.
242 2014-05-06 Mike Stump <mikestump@comcast.net>
244 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
246 2014-05-06 Joseph Myers <joseph@codesourcery.com>
248 * config/i386/sse.md (*mov<mode>_internal)
249 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
250 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
251 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
252 (*<code><mode>3, *andnot<mode>3<mask_name>)
253 (<mask_codefor><code><mode>3<mask_name>): Only consider
254 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
256 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
259 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
261 * lra-constraints.c (valid_address_p): Move earlier in file.
262 Add a constraint argument to the address_info version.
263 (satisfies_memory_constraint_p): New function.
264 (satisfies_address_constraint_p): Likewise.
265 (process_alt_operands, curr_insn_transform): Use them.
266 (process_address): Pass the constraint to valid_address_p when
267 checking address operands.
269 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
271 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
272 to their respective blocks. Fix inadvertent use of "node".
274 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
276 * emit-rtl.c (init_derived_machine_modes): New functionm, split
278 (init_emit_once): ...here.
279 * rtl.h (init_derived_machine_modes): Declare.
280 * toplev.c (do_compile): Call it even if no_backend.
282 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
283 Mike Stump <mikestump@comcast.net>
284 Richard Sandiford <rdsandiford@googlemail.com>
285 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
287 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
288 (rtx_equal_for_memref_p): Update comment.
289 (adjust_offset_for_component_ref): Use wide-int interfaces.
290 * builtins.c (get_object_alignment_2): Likewise.
291 (c_readstr): Likewise.
292 (target_char_cast): Add comment.
293 (determine_block_size): Use wide-int interfaces.
294 (expand_builtin_signbit): Likewise.
295 (fold_builtin_int_roundingfn): Likewise.
296 (fold_builtin_bitop): Likewise.
297 (fold_builtin_bswap): Likewise.
298 (fold_builtin_logarithm): Use signop.
299 (fold_builtin_pow): Likewise.
300 (fold_builtin_memory_op): Use wide-int interfaces.
301 (fold_builtin_object_size): Likewise.
302 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
303 nb_iterations_estimate.
304 (record_niter_bound): Use wide-int interfaces.
305 (get_estimated_loop_iterations_int): Likewise.
306 (get_estimated_loop_iterations): Likewise.
307 (get_max_loop_iterations): Likewise.
308 * cfgloop.h: Include wide-int.h.
309 (struct nb_iter_bound): Change bound to widest_int.
310 (struct loop): Change nb_iterations_upper_bound and
311 nb_iterations_estimate to widest_int.
312 (record_niter_bound): Switch to use widest_int.
313 (get_estimated_loop_iterations): Likewise.
314 (get_max_loop_iterations): Likewise.
315 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
317 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
318 * combine.c (try_combine): Likewise.
319 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
320 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
322 (aarch64_float_const_representable_p): Likewise.
323 * config/arc/arc.c: Include wide-int.h.
324 (arc_can_use_doloop_p): Use wide-int interfaces.
325 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
326 (vfp3_const_double_index): Likewise.
327 * config/avr/avr.c (avr_out_round): Likewise.
328 (avr_fold_builtin): Likewise.
329 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
330 (bfin_can_use_doloop_p): Likewise.
331 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
332 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
333 * config/i386/i386.c: Include wide-int.h.
334 (ix86_data_alignment): Use wide-int interfaces.
335 (ix86_local_alignment): Likewise.
336 (ix86_emit_swsqrtsf): Update real_from_integer.
337 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
338 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
339 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
340 (zero_constant): Likewise.
341 (input_operand): Likewise.
342 (splat_input_operand): Likewise.
343 (non_logical_cint_operand): Change const_double to const_wide_int.
344 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
345 (easy_altivec_constant): Remove comment.
346 (paired_expand_vector_init): Use CONSTANT_P.
347 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
348 (rs6000_emit_move): Update checks.
349 (rs6000_aggregate_candidate): Use wide-int interfaces.
350 (rs6000_expand_ternop_builtin): Likewise.
351 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
352 (rs6000_assemble_integer): Likewise.
353 (rs6000_hash_constant): Likewise.
354 (output_toc): Likewise.
355 (rs6000_rtx_costs): Likewise.
356 (rs6000_emit_swrsqrt); Update call to real_from_integer.
357 * config/rs6000/rs6000-c.c: Include wide-int.h.
358 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
359 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
360 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
361 Handle CONST_WIDE_INT.
362 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
363 Use tree_fits_uhwi_p.
364 * config/sparc/sparc.c: Include wide-int.h.
365 (sparc_fold_builtin): Use wide-int interfaces.
366 * config/vax/vax.c: Include wide-int.h.
367 (vax_float_literal): Use real_from_integer.
368 * coretypes.h (struct hwivec_def): New.
371 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
372 (equiv_constant): Handle CONST_WIDE_INT.
373 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
374 (cselib_hash_rtx): Handle CONST_WIDE_INT.
375 * dbxout.c (stabstr_U): Use wide-int interfaces.
376 (dbxout_type): Update to use cst_fits_shwi_p.
377 * defaults.h (LOG2_BITS_PER_UNIT): Define.
378 (TARGET_SUPPORTS_WIDE_INT): Add default.
379 * dfp.c: Include wide-int.h.
380 (decimal_real_to_integer2): Use wide-int interfaces and rename to
381 decimal_real_to_integer.
382 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
383 decimal_real_to_integer.
384 * doc/generic.texi (Constant expressions): Update for wide_int.
385 * doc/rtl.texi (const_double): Likewise.
386 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
387 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
388 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
389 (REAL_VALUE_FROM_INT): Remove.
390 (TARGET_SUPPORTS_WIDE_INT): New.
391 * doc/tm.texi: Regenerate.
392 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
393 * double-int.h: Include wide-int.h.
394 (struct wi::int_traits): New.
395 * dwarf2out.c (get_full_len): New.
396 (dw_val_equal_p): Add case dw_val_class_wide_int.
397 (size_of_loc_descr): Likewise.
398 (output_loc_operands): Likewise.
399 (insert_double): Remove.
400 (insert_wide_int): New.
402 (print_die): Add case dw_val_class_wide_int.
403 (attr_checksum): Likewise.
404 (attr_checksum_ordered): Likewise.
405 (same_dw_val_p): Likewise.
406 (size_of_die): Likewise.
407 (value_format): Likewise.
408 (output_die): Likewise.
409 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
411 (clz_loc_descriptor): Use wide-int interfaces.
412 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
413 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
414 (round_up_to_align): Use wide-int interfaces.
415 (field_byte_offset): Likewise.
416 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
417 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
418 CONST_DOUBLE handling. Use wide-int interfaces.
419 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
420 (gen_enumeration_type_die): Use add_AT_wide.
421 (hash_loc_operands): Add case dw_val_class_wide_int.
422 (compare_loc_operands): Likewise.
423 * dwarf2out.h: Include wide-int.h.
425 (enum dw_val_class): Add dw_val_class_wide_int.
426 (struct dw_val_struct): Add val_wide.
427 * emit-rtl.c (const_wide_int_htab): New.
428 (const_wide_int_htab_hash): New.
429 (const_wide_int_htab_eq): New.
430 (lookup_const_wide_int): New.
431 (const_double_htab_hash): Use wide-int interfaces.
432 (const_double_htab_eq): Likewise.
433 (rtx_to_double_int): Conditionally compile for wide-int.
434 (immed_double_int_const): Rename to immed_wide_int_const and
436 (immed_double_const): Conditionally compile for wide-int.
437 (init_emit_once): Use wide-int interfaces.
438 * explow.c (plus_constant): Likewise.
439 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
440 (lshift_value): Use wide-int interfaces.
441 (expand_mult): Likewise.
442 (choose_multiplier): Likewise.
443 (expand_smod_pow2): Likewise.
444 (make_tree): Likewise.
445 * expr.c (convert_modes): Consolidate handling of constants.
446 Use wide-int interfaces.
447 (emit_group_load_1): Add note.
448 (store_expr): Update comment.
449 (get_inner_reference): Use wide-int interfaces.
450 (expand_constructor): Update comment.
451 (expand_expr_real_2): Use wide-int interfaces.
452 (expand_expr_real_1): Likewise.
453 (reduce_to_bit_field_precision): Likewise.
454 (const_vector_from_tree): Likewise.
455 * final.c: Include wide-int-print.h.
456 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
457 * fixed-value.c: Include wide-int.h.
458 (fixed_from_string): Use wide-int interfaces.
459 (fixed_to_decimal): Likewise.
460 (fixed_convert_from_real): Likewise.
461 (real_convert_from_fixed): Likewise.
462 * fold-const.h (mem_ref_offset): Return an offset_int.
463 (div_if_zero_remainder): Remove code parameter.
464 * fold-const.c (div_if_zero_remainder): Remove code parameter.
465 Use wide-int interfaces.
466 (may_negate_without_overflow_p): Use wide-int interfaces.
467 (negate_expr_p): Likewise.
468 (fold_negate_expr): Likewise.
469 (int_const_binop_1): Likewise.
470 (const_binop): Likewise.
471 (fold_convert_const_int_from_int): Likewise.
472 (fold_convert_const_int_from_real): Likewise.
473 (fold_convert_const_int_from_fixed): Likewise.
474 (fold_convert_const_fixed_from_int): Likewise.
475 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
476 (sign_bit_p): Use wide-int interfaces.
477 (make_range_step): Likewise.
478 (build_range_check): Likewise. Pass an integer of the correct type
479 instead of using integer_one_node.
480 (range_predecessor): Pass an integer of the correct type instead
481 of using integer_one_node.
482 (range_successor): Likewise.
483 (merge_ranges): Likewise.
484 (unextend): Use wide-int interfaces.
485 (extract_muldiv_1): Likewise.
486 (fold_div_compare): Likewise.
487 (fold_single_bit_test): Likewise.
488 (fold_sign_changed_comparison): Likewise.
489 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
490 (fold_plusminus_mult_expr): Use wide-int interfaces.
491 (native_encode_int): Likewise.
492 (native_interpret_int): Likewise.
493 (fold_unary_loc): Likewise.
494 (pointer_may_wrap_p): Likewise.
495 (size_low_cst): Likewise.
496 (mask_with_tz): Likewise.
497 (fold_binary_loc): Likewise.
498 (fold_ternary_loc): Likewise.
499 (multiple_of_p): Likewise.
500 (tree_call_nonnegative_warnv_p): Update calls to
501 tree_int_cst_min_precision and real_from_integer.
502 (fold_negate_const): Use wide-int interfaces.
503 (fold_abs_const): Likewise.
504 (fold_relational_const): Use tree_int_cst_lt.
505 (round_up_loc): Use wide-int interfaces.
506 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
507 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
508 * gengtype.c: Remove include of double-int.h.
509 (do_typedef): Use wide-int interfaces.
510 (open_base_files): Add wide-int.h.
511 (main): Add offset_int and widest_int typedefs.
512 * gengtype-lex.l: Handle "^".
513 (CXX_KEYWORD): Add "static".
514 * gengtype-parse.c (require3): New.
515 (require_template_declaration): Handle constant template arguments
516 and nested templates.
517 * gengtype-state.c: Don't include "double-int.h".
518 * genpreds.c (write_one_predicate_function): Update comment.
519 (write_tm_constrs_h): Add check for hval and lval use in
521 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
522 (add_to_sequence): Likewise.
523 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
524 and const_double_operand.
525 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
527 * gimple-fold.c (get_base_constructor): Likewise.
528 (fold_array_ctor_reference): Likewise.
529 (fold_nonarray_ctor_reference): Likewise.
530 (fold_const_aggregate_ref_1): Likewise.
531 (gimple_val_nonnegative_real_p): Likewise.
532 (gimple_fold_indirect_ref): Likewise.
533 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
534 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
535 (struct slsr_cand_d): Change index to be widest_int.
536 (struct incr_info_d): Change incr to be widest_int.
537 (alloc_cand_and_find_basis): Use wide-int interfaces.
538 (slsr_process_phi): Likewise.
539 (backtrace_base_for_ref): Likewise. Return a widest_int.
540 (restructure_reference): Take a widest_int instead of a double_int.
541 (slsr_process_ref): Use wide-int interfaces.
542 (create_mul_ssa_cand): Likewise.
543 (create_mul_imm_cand): Likewise.
544 (create_add_ssa_cand): Likewise.
545 (create_add_imm_cand): Take a widest_int instead of a double_int.
546 (slsr_process_add): Use wide-int interfaces.
547 (slsr_process_cast): Likewise.
548 (slsr_process_copy): Likewise.
549 (dump_candidate): Likewise.
550 (dump_incr_vec): Likewise.
551 (replace_ref): Likewise.
552 (cand_increment): Likewise. Return a widest_int.
553 (cand_abs_increment): Likewise.
554 (replace_mult_candidate): Take a widest_int instead of a double_int.
555 (replace_unconditional_candidate): Use wide-int interfaces.
556 (incr_vec_index): Take a widest_int instead of a double_int.
557 (create_add_on_incoming_edge): Likewise.
558 (create_phi_basis): Use wide-int interfaces.
559 (replace_conditional_candidate): Likewise.
560 (record_increment): Take a widest_int instead of a double_int.
561 (record_phi_increments): Use wide-int interfaces.
562 (phi_incr_cost): Take a widest_int instead of a double_int.
563 (lowest_cost_path): Likewise.
564 (total_savings): Likewise.
565 (analyze_increments): Use wide-int interfaces.
566 (ncd_with_phi): Take a widest_int instead of a double_int.
567 (ncd_of_cand_and_phis): Likewise.
568 (nearest_common_dominator_for_cands): Likewise.
569 (insert_initializers): Use wide-int interfaces.
570 (all_phi_incrs_profitable): Likewise.
571 (replace_one_candidate): Likewise.
572 (replace_profitable_candidates): Likewise.
573 * godump.c: Include wide-int-print.h.
574 (go_output_typedef): Use wide-int interfaces.
575 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
576 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
577 (build_loop_iteration_domains): Likewise.
578 * hooks.h: Include wide-int.h rather than double-int.h.
579 (hook_bool_dint_dint_uint_bool_true): Delete.
580 (hook_bool_wint_wint_uint_bool_true): Declare.
581 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
582 (hook_bool_wint_wint_uint_bool_true): New.
583 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
585 (ubsan_expand_si_overflow_mul_check): Likewise.
586 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
587 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
588 (get_ancestor_addr_info): Likewise.
589 (ipa_modify_call_arguments): Likewise.
590 * loop-doloop.c (doloop_modify): Likewise.
591 (doloop_optimize): Likewise.
592 * loop-iv.c (iv_number_of_iterations): Likewise.
593 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
594 (unroll_loop_constant_iterations): Likewise.
595 (decide_unroll_runtime_iterations): Likewise.
596 (unroll_loop_runtime_iterations): Likewise.
597 (decide_peel_simple): Likewise.
598 (decide_unroll_stupid): Likewise.
599 * lto-streamer-in.c (streamer_read_wi): Add.
600 (input_cfg): Use wide-int interfaces.
601 (lto_input_tree_1): Likewise.
602 * lto-streamer-out.c (streamer_write_wi): Add.
603 (hash_tree): Use wide-int interfaces.
604 (output_cfg): Likewise.
605 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
606 (GTFILES): Add wide-int.h and signop.h.
607 (TAGS): Look for .cc files too.
608 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
609 * optabs.c (expand_subword_shift): Likewise.
610 (expand_doubleword_shift): Likewise.
611 (expand_absneg_bit): Likewise.
612 (expand_copysign_absneg): Likewise.
613 (expand_copysign_bit): Likewise.
614 * postreload.c (reload_cse_simplify_set): Likewise.
615 * predict.c (predict_iv_comparison): Likewise.
616 * pretty-print.h: Include wide-int-print.h.
618 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
619 * print-tree.c: Include wide-int-print.h.
620 (print_node_brief): Use wide-int interfaces.
621 (print_node): Likewise.
622 * read-rtl.c (validate_const_wide_int): New.
623 (read_rtx_code): Add CONST_WIDE_INT case.
624 * real.c: Include wide-int.h.
625 (real_to_integer2): Delete.
626 (real_to_integer): New function, returning a wide_int.
627 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
628 (ten_to_ptwo): Update call to real_from_integer.
629 (real_digit): Likewise.
630 * real.h: Include signop.h, wide-int.h and insn-modes.h.
631 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
632 (REAL_VALUE_TO_INT): Delete.
633 (real_to_integer): Declare a wide-int form.
634 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
635 * recog.c (const_int_operand): Improve comment.
636 (const_scalar_int_operand): New.
637 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
638 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
639 (split_double): Likewise.
640 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
641 (rtx_size): Likewise.
642 (rtx_alloc_stat_v): New.
643 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
644 (cwi_output_hex): New.
645 (iterative_hash_rtx): Handle CONST_WIDE_INT.
646 (cwi_check_failed_bounds): New.
647 * rtl.def (CONST_WIDE_INT): New.
648 * rtl.h: Include <utility> and wide-int.h.
649 (struct hwivec_def): New.
650 (CWI_GET_NUM_ELEM): New.
651 (CWI_PUT_NUM_ELEM): New.
652 (struct rtx_def): Add num_elem and hwiv.
653 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
654 (CASE_CONST_UNIQUE): Likewise.
655 (CASE_CONST_ANY): Likewise.
656 (CONST_SCALAR_INT_P): Likewise.
657 (CONST_WIDE_INT_P): New.
660 (cwi_check_failed_bounds): New.
663 (CONST_WIDE_INT_VEC) New.
664 (CONST_WIDE_INT_NUNITS) New.
665 (CONST_WIDE_INT_ELT) New.
666 (rtx_mode_t): New type.
667 (wi::int_traits <rtx_mode_t>): New.
669 (wi::min_value): New.
670 (wi::max_value): New.
672 (const_wide_int_alloc): New.
673 (immed_wide_int_const): New.
674 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
675 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
676 * signop.h: New file.
677 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
678 (simplify_const_unary_operation): Use wide-int interfaces.
679 (simplify_binary_operation_1): Likewise.
680 (simplify_const_binary_operation): Likewise.
681 (simplify_const_relational_operation): Likewise.
682 (simplify_immed_subreg): Likewise.
683 * stmt.c (expand_case): Likewise.
684 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
685 signop rather than a bool.
686 * stor-layout.c (layout_type): Use wide-int interfaces.
687 (initialize_sizetypes): Update calls to
688 set_min_and_max_values_for_integral_type.
689 (set_min_and_max_values_for_integral_type): Take a signop rather
690 than a bool. Use wide-int interfaces.
691 (fixup_signed_type): Update accordingly. Remove
692 HOST_BITS_PER_DOUBLE_INT limit.
693 (fixup_unsigned_type): Likewise.
694 * system.h (STATIC_CONSTANT_P): New.
695 (STATIC_ASSERT): New.
696 * target.def (can_use_doloop_p): Take widest_ints rather than
698 * target.h: Include wide-int.h rather than double-int.h.
699 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
701 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
702 rather than INT_CST_LT_UNSIGNED.
703 (can_use_doloop_if_innermost): Take widest_ints rather than
705 * tree-affine.c: Include wide-int-print.h.
706 (double_int_ext_for_comb): Delete.
707 (wide_int_ext_for_comb): New.
708 (aff_combination_zero): Use wide-int interfaces.
709 (aff_combination_const): Take a widest_int instead of a double_int.
710 (aff_combination_elt): Use wide-int interfaces.
711 (aff_combination_scale): Take a widest_int instead of a double_int.
712 (aff_combination_add_elt): Likewise.
713 (aff_combination_add_cst): Likewise.
714 (aff_combination_add): Use wide-int interfaces.
715 (aff_combination_convert): Likewise.
716 (tree_to_aff_combination): Likewise.
717 (add_elt_to_tree): Take a widest_int instead of a double_int.
718 (aff_combination_to_tree): Use wide-int interfaces.
719 (aff_combination_remove_elt): Likewise.
720 (aff_combination_add_product): Take a widest_int instead of
722 (aff_combination_mult): Use wide-int interfaces.
723 (aff_combination_expand): Likewise.
724 (double_int_constant_multiple_p): Delete.
725 (wide_int_constant_multiple_p): New.
726 (aff_combination_constant_multiple_p): Take a widest_int pointer
727 instead of a double_int pointer.
728 (print_aff): Use wide-int interfaces.
729 (get_inner_reference_aff): Take a widest_int pointer
730 instead of a double_int pointer.
731 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
732 * tree-affine.h: Include wide-int.h.
733 (struct aff_comb_elt): Change type of coef to widest_int.
734 (struct affine_tree_combination): Change type of offset to widest_int.
735 (double_int_ext_for_comb): Delete.
736 (wide_int_ext_for_comb): New.
737 (aff_combination_const): Use widest_int instead of double_int.
738 (aff_combination_scale): Likewise.
739 (aff_combination_add_elt): Likewise.
740 (aff_combination_constant_multiple_p): Likewise.
741 (get_inner_reference_aff): Likewise.
742 (aff_comb_cannot_overlap_p): Likewise.
743 (aff_combination_zero_p): Use wide-int interfaces.
744 * tree.c: Include tree.h.
745 (init_ttree): Use make_int_cst.
746 (tree_code_size): Removed code for INTEGER_CST case.
747 (tree_size): Add INTEGER_CST case.
748 (make_node_stat): Update comment.
749 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
750 (build_int_cst_type): Use wide-int interfaces.
751 (double_int_to_tree): Likewise.
752 (double_int_fits_to_tree_p): Delete.
753 (force_fit_type_double): Delete.
754 (force_fit_type): New.
755 (int_cst_hash_hash): Use wide-int interfaces.
756 (int_cst_hash_eq): Likewise.
757 (build_int_cst_wide): Delete.
758 (wide_int_to_tree): New.
759 (cache_integer_cst): Use wide-int interfaces.
760 (build_low_bits_mask): Likewise.
761 (cst_and_fits_in_hwi): Likewise.
762 (real_value_from_int_cst): Likewise.
763 (make_int_cst_stat): New.
764 (integer_zerop): Use wide_int interfaces.
765 (integer_onep): Likewise.
766 (integer_all_onesp): Likewise.
767 (integer_pow2p): Likewise.
768 (integer_nonzerop): Likewise.
769 (tree_log2): Likewise.
770 (tree_floor_log2): Likewise.
771 (tree_ctz): Likewise.
772 (int_size_in_bytes): Likewise.
773 (mem_ref_offset): Return an offset_int rather than a double_int.
774 (build_type_attribute_qual_variant): Use wide_int interfaces.
775 (type_hash_eq): Likewise
776 (tree_int_cst_equal): Likewise.
777 (tree_int_cst_lt): Delete.
778 (tree_int_cst_compare): Likewise.
779 (tree_fits_shwi_p): Use wide_int interfaces.
780 (tree_fits_uhwi_p): Likewise.
781 (tree_int_cst_sign_bit): Likewise.
782 (tree_int_cst_sgn): Likewise.
783 (tree_int_cst_min_precision): Take a signop rather than a bool.
784 (simple_cst_equal): Use wide_int interfaces.
785 (compare_tree_int): Likewise.
786 (iterative_hash_expr): Likewise.
787 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
789 (get_type_static_bounds): Use wide_int interfaces.
790 (tree_int_cst_elt_check_failed): New.
791 (build_common_tree_nodes): Reordered to set prec before filling in
793 (int_cst_value): Check cst_and_fits_in_hwi.
794 (widest_int_cst_value): Use wide_int interfaces.
795 (upper_bound_in_type): Likewise.
796 (lower_bound_in_type): Likewise.
797 (num_ending_zeros): Likewise.
798 (drop_tree_overflow): Likewise.
799 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
800 (gen_conditions_for_pow_cst_base): Likewise.
801 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
802 (group_case_labels_stmt): Use wide-int interfaces.
803 (verify_gimple_assign_binary): Likewise.
804 (print_loop): Likewise.
805 * tree-chrec.c (tree_fold_binomial): Likewise.
806 * tree-core.h (struct tree_base): Add int_length.
807 (struct tree_int_cst): Change rep of value.
808 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
809 (dr_may_alias_p): Likewise.
810 (max_stmt_executions_tree): Likewise.
811 * tree.def (INTEGER_CST): Update comment.
812 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
813 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
814 * tree-dump.c: Include wide-int.h and wide-int-print.h.
815 (dequeue_and_dump): Use wide-int interfaces.
816 * tree.h: Include wide-int.h.
817 (NULL_TREE): Moved to earlier loc in file.
818 (TREE_INT_CST_ELT_CHECK): New.
819 (tree_int_cst_elt_check_failed): New.
821 (TREE_INT_CST): Delete.
822 (TREE_INT_CST_LOW): Use wide-int interfaces.
823 (TREE_INT_CST_HIGH): Delete.
824 (TREE_INT_CST_NUNITS): New.
825 (TREE_INT_CST_EXT_NUNITS): Likewise.
826 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
827 (TREE_INT_CST_ELT): Likewise.
828 (INT_CST_LT): Delete.
829 (tree_int_cst_elt_check): New (two forms).
830 (type_code_size): Update comment.
831 (make_int_cst_stat, make_int_cst): New.
832 (tree_to_double_int): Delete.
833 (double_int_fits_to_tree_p): Delete.
834 (force_fit_type_double): Delete.
835 (build_int_cstu): Replace with out-of-line function.
836 (build_int_cst_wide): Delete.
837 (tree_int_cst_lt): Define inline.
838 (tree_int_cst_le): New.
839 (tree_int_cst_compare): Define inline.
840 (tree_int_cst_min_precision): Take a signop rather than a bool.
841 (wi::int_traits <const_tree>): New.
842 (wi::int_traits <tree>): New.
843 (wi::extended_tree): New.
844 (wi::int_traits <wi::extended_tree>): New.
845 (wi::to_widest): New.
846 (wi::to_offset): New.
847 (wi::fits_to_tree_p): New.
848 (wi::min_value): New.
849 (wi::max_value): New.
850 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
851 (copy_tree_body_r): Likewise.
852 * tree-object-size.c (compute_object_offset): Likewise.
853 (addr_object_size): Likewise.
854 * tree-predcom.c: Include wide-int-print.h.
855 (struct dref_d): Change type of offset to widest_int.
856 (dump_dref): Call wide-int printer.
857 (aff_combination_dr_offset): Use wide-int interfaces.
858 (determine_offset): Take a widest_int pointer rather than a
860 (split_data_refs_to_components): Use wide-int interfaces.
861 (suitable_component_p): Likewise.
862 (order_drefs): Likewise.
863 (add_ref_to_chain): Likewise.
864 (valid_initializer_p): Likewise.
865 (determine_roots_comp): Likewise.
866 * tree-pretty-print.c: Include wide-int-print.h.
867 (dump_generic_node): Use wide-int interfaces.
868 * tree-sra.c (sra_ipa_modify_expr): Likewise.
869 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
870 (move_fixed_address_to_symbol): Likewise.
871 (move_hint_to_base): Likewise.
872 (move_pointer_to_base): Likewise.
873 (move_variant_to_index): Likewise.
874 (most_expensive_mult_to_index): Likewise.
875 (addr_to_parts): Likewise.
876 (copy_ref_info): Likewise.
877 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
878 (indirect_refs_may_alias_p): Likewise.
879 (stmt_kills_ref_p_1): Likewise.
880 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
881 * tree-ssa-ccp.c: Update comment at top of file. Include
883 (struct prop_value_d): Change type of mask to widest_int.
884 (extend_mask): New function.
885 (dump_lattice_value): Use wide-int interfaces.
886 (get_default_value): Likewise.
887 (set_constant_value): Likewise.
888 (set_value_varying): Likewise.
889 (valid_lattice_transition): Likewise.
890 (set_lattice_value): Likewise.
891 (value_to_double_int): Delete.
892 (value_to_wide_int): New.
893 (get_value_from_alignment): Use wide-int interfaces.
894 (get_value_for_expr): Likewise.
895 (do_dbg_cnt): Likewise.
896 (ccp_finalize): Likewise.
897 (ccp_lattice_meet): Likewise.
898 (bit_value_unop_1): Use widest_ints rather than double_ints.
899 (bit_value_binop_1): Likewise.
900 (bit_value_unop): Use wide-int interfaces.
901 (bit_value_binop): Likewise.
902 (bit_value_assume_aligned): Likewise.
903 (evaluate_stmt): Likewise.
904 (ccp_fold_stmt): Likewise.
905 (visit_cond_stmt): Likewise.
906 (ccp_visit_stmt): Likewise.
907 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
908 (constant_pointer_difference): Likewise.
909 (associate_pointerplus): Likewise.
910 (combine_conversions): Likewise.
911 * tree-ssa-loop.h: Include wide-int.h.
912 (struct tree_niter_desc): Change type of max to widest_int.
913 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
914 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
915 (remove_redundant_iv_tests): Likewise.
916 (canonicalize_loop_induction_variables): Likewise.
917 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
918 (constant_multiple_of): Take a widest_int pointer instead of
919 a double_int pointer.
920 (get_computation_aff): Use wide-int interfaces.
921 (ptr_difference_cost): Likewise.
922 (difference_cost): Likewise.
923 (get_loop_invariant_expr_id): Likewise.
924 (get_computation_cost_at): Likewise.
925 (iv_elimination_compare_lt): Likewise.
926 (may_eliminate_iv): Likewise.
927 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
928 instead of double_int.
929 (max_loop_iterations): Likewise.
930 (max_stmt_executions): Likewise.
931 (estimated_stmt_executions): Likewise.
932 * tree-ssa-loop-niter.c: Include wide-int-print.h.
933 (split_to_var_and_offset): Use wide-int interfaces.
934 (determine_value_range): Likewise.
935 (bound_difference_of_offsetted_base): Likewise.
936 (bounds_add): Take a widest_int instead of a double_int.
937 (number_of_iterations_ne_max): Use wide-int interfaces.
938 (number_of_iterations_ne): Likewise.
939 (number_of_iterations_lt_to_ne): Likewise.
940 (assert_loop_rolls_lt): Likewise.
941 (number_of_iterations_lt): Likewise.
942 (number_of_iterations_le): Likewise.
943 (number_of_iterations_cond): Likewise.
944 (number_of_iterations_exit): Likewise.
945 (finite_loop_p): Likewise.
946 (derive_constant_upper_bound_assign): Likewise.
947 (derive_constant_upper_bound): Return a widest_int.
948 (derive_constant_upper_bound_ops): Likewise.
949 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
950 (record_estimate): Take a widest_int rather than a double_int.
951 (record_nonwrapping_iv): Use wide-int interfaces.
952 (double_int_cmp): Delete.
954 (bound_index): Take a widest_int rather than a double_int.
955 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
956 (maybe_lower_iteration_bound): Likewise.
957 (estimate_numbers_of_iterations_loop): Likewise.
958 (estimated_loop_iterations): Take a widest_int pointer than than
959 a double_int pointer.
960 (estimated_loop_iterations_int): Use wide-int interfaces.
961 (max_loop_iterations): Take a widest_int pointer than than
962 a double_int pointer.
963 (max_loop_iterations_int): Use wide-int interfaces.
964 (max_stmt_executions): Take a widest_int pointer than than
965 a double_int pointer.
966 (estimated_stmt_executions): Likewise.
967 (n_of_executions_at_most): Use wide-int interfaces.
968 (scev_probably_wraps_p): Likewise.
969 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
971 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
973 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
974 double_ints. Adjust for trailing_wide_ints <3> representation.
975 (set_nonzero_bits): Likewise.
976 (get_range_info): Return wide_ints rather than double_ints.
977 Adjust for trailing_wide_ints <3> representation.
978 (get_nonzero_bits): Likewise.
979 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
981 * tree-ssanames.h (struct range_info_def): Replace min, max and
982 nonzero_bits with a trailing_wide_ints <3>.
983 (set_range_info): Use wide_int_refs rather than double_ints.
984 (set_nonzero_bits): Likewise.
985 (get_range_info): Return wide_ints rather than double_ints.
986 (get_nonzero_bits): Likewise.
987 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
988 * tree-ssa-pre.c (phi_translate_1): Likewise.
989 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
990 (acceptable_pow_call): Likewise.
991 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
993 (vn_reference_fold_indirect): Likewise.
994 (vn_reference_maybe_forwprop_address): Likewise.
995 (valueize_refs_1): Likewise.
996 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
997 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
998 tree_int_cst_lt and tree_int_cst_le.
999 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
1001 (streamer_alloc_tree): Likewise.
1002 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
1003 (streamer_write_tree_header): Likewise.
1004 (streamer_write_integer_cst): Likewise.
1005 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
1006 (build_constructors): Likewise.
1007 (array_value_type): Likewise.
1008 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
1009 (vect_check_gather): Likewise.
1010 * tree-vect-generic.c (build_replicated_const): Likewise.
1011 (expand_vector_divmod): Likewise.
1012 * tree-vect-loop.c (vect_transform_loop): Likewise.
1013 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
1014 (vect_do_peeling_for_alignment): Likewise.
1015 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
1016 * tree-vrp.c: Include wide-int.h.
1017 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
1018 (extract_range_from_assert): Use wide-int interfaces.
1019 (vrp_int_const_binop): Likewise.
1020 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
1021 double_int pointers.
1022 (ranges_from_anti_range): Use wide-int interfaces.
1023 (quad_int_cmp): Delete.
1024 (quad_int_pair_sort): Likewise.
1025 (extract_range_from_binary_expr_1): Use wide-int interfaces.
1026 (extract_range_from_unary_expr_1): Likewise.
1027 (adjust_range_with_scev): Likewise.
1028 (masked_increment): Take and return wide_ints rather than double_ints.
1029 (register_edge_assert_for_2): Use wide-int interfaces.
1030 (check_array_ref): Likewise.
1031 (search_for_addr_array): Likewise.
1032 (maybe_set_nonzero_bits): Likewise.
1033 (union_ranges): Pass an integer of the correct type instead of
1034 using integer_one_node.
1035 (intersect_ranges): Likewise.
1036 (simplify_truth_ops_using_ranges): Likewise.
1037 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
1038 (range_fits_type_p): Likewise.
1039 (simplify_cond_using_ranges): Likewise. Take a signop rather than
1041 (simplify_conversion_using_ranges): Use wide-int interfaces.
1042 (simplify_float_conversion_using_ranges): Likewise.
1043 (vrp_finalize): Likewise.
1044 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
1045 (gimple_stringops_transform): Likewise.
1046 * varasm.c (decode_addr_const): Likewise.
1047 (const_hash_1): Likewise.
1048 (const_rtx_hash_1): Likewise
1049 (output_constant): Likewise.
1050 (array_size_for_constructor): Likewise.
1051 (output_constructor_regular_field): Likewise.
1052 (output_constructor_bitfield): Likewise.
1053 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
1054 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
1056 * gencheck.c: Define BITS_PER_UNIT.
1059 * wide-int-print.cc: New.
1060 * wide-int-print.h: New.
1062 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1064 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
1066 2014-05-06 Richard Biener <rguenther@suse.de>
1068 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
1069 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
1070 (TODO_verify_all): Adjust.
1071 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
1072 TODO_verify_stmts and TODO_verify_rtl_sharing.
1073 * bb-reorder.c: Likewise.
1074 * cfgexpand.c: Likewise.
1075 * cprop.c: Likewise.
1077 * function.c: Likewise.
1078 * fwprop.c: Likewise.
1080 * gimple-ssa-isolate-paths.c: Likewise.
1081 * gimple-ssa-strength-reduction.c: Likewise.
1082 * ipa-split.c: Likewise.
1083 * loop-init.c: Likewise.
1084 * loop-unroll.c: Likewise.
1085 * lower-subreg.c: Likewise.
1086 * modulo-sched.c: Likewise.
1087 * postreload-gcse.c: Likewise.
1088 * predict.c: Likewise.
1089 * recog.c: Likewise.
1090 * sched-rgn.c: Likewise.
1091 * store-motion.c: Likewise.
1092 * tracer.c: Likewise.
1093 * trans-mem.c: Likewise.
1094 * tree-call-cdce.c: Likewise.
1095 * tree-cfg.c: Likewise.
1096 * tree-cfgcleanup.c: Likewise.
1097 * tree-complex.c: Likewise.
1098 * tree-eh.c: Likewise.
1099 * tree-emutls.c: Likewise.
1100 * tree-if-conv.c: Likewise.
1101 * tree-into-ssa.c: Likewise.
1102 * tree-loop-distribution.c: Likewise.
1103 * tree-object-size.c: Likewise.
1104 * tree-parloops.c: Likewise.
1105 * tree-pass.h: Likewise.
1106 * tree-sra.c: Likewise.
1107 * tree-ssa-ccp.c: Likewise.
1108 * tree-ssa-copy.c: Likewise.
1109 * tree-ssa-copyrename.c: Likewise.
1110 * tree-ssa-dce.c: Likewise.
1111 * tree-ssa-dom.c: Likewise.
1112 * tree-ssa-dse.c: Likewise.
1113 * tree-ssa-forwprop.c: Likewise.
1114 * tree-ssa-ifcombine.c: Likewise.
1115 * tree-ssa-loop-ch.c: Likewise.
1116 * tree-ssa-loop-ivcanon.c: Likewise.
1117 * tree-ssa-loop.c: Likewise.
1118 * tree-ssa-math-opts.c: Likewise.
1119 * tree-ssa-phiopt.c: Likewise.
1120 * tree-ssa-phiprop.c: Likewise.
1121 * tree-ssa-pre.c: Likewise.
1122 * tree-ssa-reassoc.c: Likewise.
1123 * tree-ssa-sink.c: Likewise.
1124 * tree-ssa-strlen.c: Likewise.
1125 * tree-ssa-tail-merge.c: Likewise.
1126 * tree-ssa-uncprop.c: Likewise.
1127 * tree-switch-conversion.c: Likewise.
1128 * tree-tailcall.c: Likewise.
1129 * tree-vect-generic.c: Likewise.
1130 * tree-vectorizer.c: Likewise.
1131 * tree-vrp.c: Likewise.
1133 * var-tracking.c: Likewise.
1134 * bt-load.c: Likewise.
1135 * cfgcleanup.c: Likewise.
1136 * combine-stack-adj.c: Likewise.
1137 * combine.c: Likewise.
1138 * compare-elim.c: Likewise.
1139 * config/epiphany/resolve-sw-modes.c: Likewise.
1140 * config/i386/i386.c: Likewise.
1141 * config/mips/mips.c: Likewise.
1142 * config/s390/s390.c: Likewise.
1143 * config/sh/sh_treg_combine.cc: Likewise.
1144 * config/sparc/sparc.c: Likewise.
1147 * final.c: Likewise.
1148 * ifcvt.c: Likewise.
1149 * mode-switching.c: Likewise.
1150 * passes.c: Likewise.
1151 * postreload.c: Likewise.
1153 * reg-stack.c: Likewise.
1154 * regcprop.c: Likewise.
1155 * regrename.c: Likewise.
1158 2014-05-06 Richard Biener <rguenther@suse.de>
1161 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
1162 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
1164 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
1167 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
1169 2014-05-05 Radovan Obradovic <robradovic@mips.com>
1170 Tom de Vries <tom@codesourcery.com>
1172 * target.def (call_fusage_contains_non_callee_clobbers): New DEFHOOKPOD.
1173 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
1175 (@node Miscellaneous Register Hooks): New node.
1176 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
1177 * doc/tm.texi: Regenerate.
1179 2014-05-05 Marek Polacek <polacek@redhat.com>
1182 * opts.c (common_handle_option): Call error_at instead of warning_at.
1184 2014-05-05 Richard Biener <rguenther@suse.de>
1186 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
1187 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
1188 under the TODO_verify_il umbrella.
1190 2014-05-05 Richard Biener <rguenther@suse.de>
1192 * passes.c (execute_function_todo): Move TODO_verify_flow under
1193 the TODO_verify_ul umbrella.
1195 2014-05-05 Richard Biener <rguenther@suse.de>
1198 * fold-const.c (fold_binary_loc): Consistently avoid
1199 canonicalizing X & CST away from a CST that is the mask
1202 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1204 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
1205 int argument to enum machine_mode.
1206 (picochip_class_max_nregs): Ditto.
1207 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
1208 (picochip_class_max_nregs): Ditto.
1210 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1212 * target.def: Add new target hook.
1213 * doc/tm.texi: Regenerate.
1214 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
1215 * targhooks.c (default_keep_leaf_when_profiled): New function.
1217 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
1218 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
1220 2014-05-05 Bin Cheng <bin.cheng@arm.com>
1222 PR tree-optimization/60363
1223 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
1224 (copy_phi_args): New parameters. Call get_value_locus_in_path.
1225 (update_destination_phis): New parameter.
1226 (create_edge_and_update_destination_phis): Ditto.
1227 (ssa_fix_duplicate_block_edges): Pass new arguments.
1228 (thread_single_edge): Ditto.
1230 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
1232 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
1233 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
1234 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
1235 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
1236 Use RS6000_BTM_HARD_FLOAT.
1237 (BU_MISC_2): Likewise.
1238 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
1239 RS6000_BTM_HARD_FLOAT.
1240 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
1242 (rs6000_invalid_builtin): Add hard floating builtin support.
1243 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
1244 hard float builtins.
1245 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
1247 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1249 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
1250 Add missing function* argument.
1252 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1254 * lra-constraints.c (valid_address_p): Move earlier in file.
1255 Add a constraint argument to the address_info version.
1256 (satisfies_memory_constraint_p): New function.
1257 (satisfies_address_constraint_p): Likewise.
1258 (process_alt_operands, curr_insn_transform): Use them.
1259 (process_address): Pass the constraint to valid_address_p when
1260 checking address operands.
1262 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1264 * config/mips/mips.c (mips_isa_rev): New variable.
1265 (mips_set_architecture): Set it.
1266 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1268 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1269 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1270 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1271 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1272 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
1273 conditions in terms of mips_isa_rev.
1274 (mips_isa_rev): Declare.
1276 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1278 * config/sh/sh-mem.cc: Use tabs instead of spaces.
1279 (prob_unlikely, prob_likely): Make variables const.
1281 2014-05-03 Denis Chertykov <chertykov@gmail.com>
1283 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
1285 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1287 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
1289 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1291 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
1292 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
1293 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
1295 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
1296 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
1298 Replace usage of ROUND_REG with sh_round_reg.
1299 Use CEIL instead of ROUND_ADVANCE.
1301 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1304 * config/sh/sh.c: Include stdlib headers before everything else.
1306 2014-05-02 Jakub Jelinek <jakub@redhat.com>
1308 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
1309 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
1310 (gimplify_adjust_omp_clauses): Simd region is never
1311 directly nested in combined parallel. Instead, for linear
1312 with copyin/copyout, if in combined for simd loop, make decl
1313 firstprivate/lastprivate on OMP_FOR.
1314 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
1315 expand_omp_for_static_chunk): When setting endvar, also set
1316 fd->loop.v to the same value.
1318 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1320 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
1322 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
1324 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
1327 2014-05-02 Marek Polacek <polacek@redhat.com>
1329 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
1331 2014-05-02 Kito Cheng <kito@0xlab.org>
1333 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
1334 to a C expression marco.
1335 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
1336 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
1337 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
1338 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
1339 HONOR_REG_ALLOC_ORDER.
1340 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
1342 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1344 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
1346 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1348 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
1350 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
1352 * tree-if-conv.c (is_cond_scalar_reduction): New function.
1353 (convert_scalar_cond_reduction): Likewise.
1354 (predicate_scalar_phi): Add recognition and transformation
1355 of simple conditioanl reduction to be vectorizable.
1357 2014-05-01 Marek Polacek <polacek@redhat.com>
1360 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
1362 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
1364 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
1365 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
1366 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
1367 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
1368 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
1369 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
1370 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
1371 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
1373 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
1375 * config/arc/arc.opt (mlra): Move comment above option name
1376 to avoid mis-parsing as language options.
1378 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1380 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
1381 * config/sol2.h: ... here.
1382 * config/sol2-10.h: Remove.
1384 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
1385 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
1386 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
1387 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
1388 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
1389 * config/sol2.h: ... here.
1390 (SECTION_NAME_FORMAT): Don't redefine.
1391 (STARTFILE_ARCH32_SPEC): Rename to ...
1392 (STARTFILE_ARCH_SPEC): ... this.
1393 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
1394 * config/sparc/sol2.h: ... here.
1395 (SECTION_NAME_FORMAT): Don't undef.
1396 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
1397 (SUBTARGET_EXTRA_SPECS): Remove.
1398 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
1400 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
1401 (MD_STARTFILE_PREFIX): Remove.
1402 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
1403 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
1404 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
1405 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
1406 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
1407 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
1408 * config/i386/sol2.h: ... here.
1409 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
1410 * config/i386/sol2-bi.h: Remove.
1411 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
1412 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
1414 * config/i386/t-sol2-64: Rename to ...
1415 * config/i386/t-sol2: ... this.
1416 * config/sparc/t-sol2-64: Rename to ...
1417 * config/sparc/t-sol2: ... this.
1419 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
1420 sol2_tm_file_head, sol2_tm_file_tail.
1421 Include ${cpu_type}/sol2.h before sol2.h.
1423 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
1424 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
1425 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
1426 Reflect i386/t-sol2-64 renaming.
1427 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
1428 Reflect sparc/t-sol2-64 renaming.
1430 2014-04-30 Richard Biener <rguenther@suse.de>
1432 * passes.c (execute_function_todo): Move TODO_verify_stmts
1433 and TODO_verify_ssa under the TODO_verify_il umbrella.
1434 * tree-ssa.h (verify_ssa): Adjust prototype.
1435 * tree-ssa.c (verify_ssa): Add parameter to tell whether
1436 we should verify SSA operands.
1437 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
1438 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
1439 whether we should verify whether not throwing stmts have EH info.
1440 * graphite-scop-detection.c (create_sese_edges): Adjust.
1441 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
1442 * tree-eh.c (lower_try_finally_switch): Do not add the
1443 default case label twice.
1445 2014-04-30 Marek Polacek <polacek@redhat.com>
1447 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
1448 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
1449 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
1450 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
1452 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
1454 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
1455 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
1456 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
1457 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
1458 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
1459 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
1460 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
1461 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
1463 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1465 * tree-cfg.c (dump_function_to_file): Dump the return type of
1466 functions, in a line to itself before the function body, mimicking
1467 the layout of a C function.
1469 2014-04-29 Jakub Jelinek <jakub@redhat.com>
1471 PR tree-optimization/60971
1472 * tree-tailcall.c (process_assignment): Reject conversions which
1475 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
1477 * calls.c (initialize_argument_information): Always treat
1478 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
1479 (expand_call): Likewise.
1480 (emit_library_call_calue_1): Likewise.
1481 * expr.c (PUSH_ARGS_REVERSED): Do not define.
1482 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
1485 2014-04-29 Nick Clifton <nickc@redhat.com>
1487 * config/msp430/msp430.md (umulsidi): Fix typo.
1488 (mulhisi3): Enable even inside interrupt handlers.
1489 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
1490 bigger return address pushed in large mode.
1492 2014-04-29 Nick Clifton <nickc@redhat.com>
1494 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
1495 (arc_init_reg_tables): Use a machine_mode enum to iterate over
1497 * config/m32r/m32r.c (init_reg_tables): Likewise.
1498 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
1499 enum to hold the modes.
1501 2014-04-29 Richard Biener <rguenther@suse.de>
1503 * dominance.c (free_dominance_info): Add overload with
1505 (dom_info_state): Likewise.
1506 (dom_info_available_p): Likewise.
1507 * basic-block.h (free_dominance_info, dom_info_state,
1508 dom_info_available_p): Declare overloads.
1509 * passes.c (execute_function_todo): Verify that verifiers
1510 don't change dominator info state. Drop dominator info
1511 for IPA pass invocations.
1512 * cgraph.c (release_function_body): Restore asserts that
1513 dominator information is released.
1515 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
1517 * doc/invoke.texi: Fix typo.
1518 * tree-vrp.c: Fix typos.
1519 * gimple.c (infer_nonnull_range): Reorder operands of an &&
1522 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1524 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
1526 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1528 * config/aarch64/aarch64-builtins.c
1529 (aarch64_types_storestruct_lane_qualifiers): New.
1530 (TYPES_STORESTRUCT_LANE): Likewise.
1531 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
1532 (st3_lane): Likewise.
1533 (st4_lane): Likewise.
1534 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
1535 (vec_store_lanesci_lane<mode>): Likewise.
1536 (vec_store_lanesxi_lane<mode>): Likewise.
1537 (aarch64_st2_lane<VQ:mode>): Likewise.
1538 (aarch64_st3_lane<VQ:mode>): Likewise.
1539 (aarch64_st4_lane<VQ:mode>): Likewise.
1540 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
1541 * config/aarch64/arm_neon.h
1542 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
1543 use new macro arguments.
1544 (__ST3_LANE_FUNC): Likewise.
1545 (__ST4_LANE_FUNC): Likewise.
1546 * config/aarch64/iterators.md (V_TWO_ELEM): New.
1547 (V_THREE_ELEM): Likewise.
1548 (V_FOUR_ELEM): Likewise.
1550 2014-04-28 David Malcolm <dmalcolm@redhat.com>
1552 * doc/gimple.texi: Replace the description of the now-defunct
1553 union gimple_statement_d with a diagram showing the
1554 gimple_statement_base class hierarchy and its relationships to
1555 the GSS_ and GIMPLE_ enums.
1557 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1559 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
1560 * config/aarch64/aarch64.c
1561 (aarch64_cannot_change_mode_class): Weaken conditions.
1562 (aarch64_modes_tieable_p): New.
1563 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
1565 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
1567 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
1568 (loadsync_<mode>): Change mode.
1569 (load_quadpti, store_quadpti): New.
1570 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
1571 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
1573 2014-04-28 Martin Jambor <mjambor@suse.cz>
1575 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
1576 same alias type as the original statement.
1577 (subreplacement_assignment_data): New type.
1578 (handle_unscalarized_data_in_subtree): New type of parameter,
1579 generate new memory accesses with same alias type as the original
1581 (load_assign_lhs_subreplacements): Likewise.
1582 (sra_modify_constructor_assign): Generate new memory accesses with
1583 same alias type as the original statement.
1585 2014-04-28 Richard Biener <rguenther@suse.de>
1587 * tree-pass.h (TODO_verify_il): Define.
1588 (TODO_verify_all): Complete properly.
1589 * passes.c (execute_function_todo): Move existing loop-closed
1590 SSA verification under TODO_verify_il.
1591 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
1592 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
1593 Fix tree sharing issue.
1595 2014-04-28 Richard Biener <rguenther@suse.de>
1598 * builtins.def (DEF_C11_BUILTIN): Add.
1599 (BUILT_IN_ALIGNED_ALLOC): Likewise.
1600 * coretypes.h (enum function_class): Add function_c11_misc.
1601 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1602 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
1603 (call_may_clobber_ref_p_1): Likewise.
1604 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
1605 (mark_all_reaching_defs_necessary_1): Likewise.
1606 (propagate_necessity): Likewise.
1607 (eliminate_unnecessary_stmts): Likewise.
1608 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
1610 2014-04-28 Richard Biener <rguenther@suse.de>
1612 * tree-vrp.c (vrp_var_may_overflow): Remove.
1613 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
1614 with overflow immediately bump to one before that value and
1615 let iteration figure out overflow status.
1617 2014-04-28 Richard Biener <rguenther@suse.de>
1619 * configure.ac: Do valgrind header checks unconditionally.
1620 Add --enable-valgrind-annotations.
1621 * system.h: Guard valgrind header inclusion with
1622 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
1623 * alloc-pool.c (pool_alloc, pool_free): Use
1624 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
1625 to guard possibly dead code.
1626 * config.in: Regenerated.
1627 * configure: Likewise.
1629 2014-04-28 Jeff Law <law@redhat.com>
1631 PR tree-optimization/60902
1632 * tree-ssa-threadedge.c
1633 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
1634 over real defs when invalidating outputs from statements that do not
1635 produce useful outputs for threading.
1637 2014-04-28 Richard Biener <rguenther@suse.de>
1639 PR tree-optimization/60979
1640 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
1641 SCOPs that end in a block with a successor with abnormal
1644 2014-04-28 Richard Biener <rguenther@suse.de>
1646 * tree-pass.h (execute_pass_list): Adjust prototype.
1647 * passes.c (pass_manager::execute_early_local_passes):
1649 (do_per_function): Change callback signature, push all actual
1650 work to the callbals.
1651 (do_per_function_toporder): Likewise.
1652 (execute_function_dump): Adjust.
1653 (execute_function_todo): Likewise.
1654 (clear_last_verified): Likewise.
1655 (verify_curr_properties): Likewise.
1656 (update_properties_after_pass): Likewise.
1657 (execute_pass_list_1): Split out from ...
1658 (execute_pass_list): ... here. Adjust.
1659 (execute_ipa_pass_list): Likewise.
1660 * cgraphunit.c (cgraph_add_new_function): Adjust.
1661 (analyze_function): Likewise.
1662 (expand_function): Likewise.
1663 * cgraph.c (release_function_body): Free dominance info
1664 here instead of asserting it was magically freed elsewhere.
1666 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
1668 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
1669 * configure: Regenerate.
1670 * config/sparc/sparc.opt (muser-mode): New option.
1671 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
1673 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
1674 * doc/invoke.texi (SPARC options): Document -muser-mode.
1676 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
1678 * cselib.c (find_slot_memmode): Delete.
1679 (cselib_hasher): Change compare_type to a struct.
1680 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
1682 (preserve_constants_and_equivs): Adjust for new compare_type.
1683 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
1684 (wrap_constant): Delete.
1685 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
1687 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
1689 * doc/install.texi (Building with profile feedback): Remove
1692 2014-04-26 Tom de Vries <tom@codesourcery.com>
1694 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
1697 2014-04-25 Cary Coutant <ccoutant@google.com>
1700 * dwarf2out.c (should_move_die_to_comdat): A type definition
1701 can contain a subprogram definition, but don't move it to a
1703 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
1704 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
1706 (clone_tree_hash): Rename to...
1707 (clone_tree_partial): ...this; change callers. Copy
1708 DW_TAG_subprogram DIEs as declarations.
1709 (copy_decls_walk): Don't copy children of a declaration into a
1712 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
1715 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
1718 2014-04-25 Jiong Wang <jiong.wang@arm.com>
1720 * config/arm/predicates.md (call_insn_operand): Add long_call check.
1721 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
1723 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
1726 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1728 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
1731 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1733 PR tree-optimization/60930
1734 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
1735 creating a multiply candidate by folding two constant
1736 multiplicands when the result overflows.
1738 2014-04-25 Jakub Jelinek <jakub@redhat.com>
1740 PR tree-optimization/60960
1741 * tree-vect-generic.c (expand_vector_operation): Only call
1742 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
1744 2014-04-25 Tom de Vries <tom@codesourcery.com>
1746 * expr.c (clobber_reg_mode): New function.
1747 * expr.h (clobber_reg): New function.
1749 2014-04-25 Tom de Vries <tom@codesourcery.com>
1751 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
1754 2014-04-25 Radovan Obradovic <robradovic@mips.com>
1755 Tom de Vries <tom@codesourcery.com>
1757 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
1759 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
1760 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
1761 new argument to find_all_hard_reg_sets call.
1763 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1765 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
1766 Use HOST_WIDE_INT_C for mask literal.
1767 (aarch_rev16_shleft_mask_imm_p): Likewise.
1769 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
1772 * config/sparc/sparc.md (ashlsi3_extend): Delete.
1774 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
1776 PR preprocessor/56540
1777 * config/i386/i386-c.c (ix86_target_macros): Define
1778 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
1780 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1782 * configure.ac (tga_func): Remove.
1783 (LIB_TLS_SPEC): Remove.
1784 * configure: Regenerate.
1785 * config.in: Regenerate.
1786 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
1788 2014-04-25 Richard Biener <rguenther@suse.de>
1791 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
1792 call stmt use/clobber sets during stmt walk instead of
1793 walking the possibly incomplete set of caller edges.
1795 2014-04-25 Richard Biener <rguenther@suse.de>
1798 * passes.c (apply_ipa_transforms): Inline into only caller ...
1799 (execute_one_pass): ... here. Properly bring in function
1800 bodies for nodes we want to apply IPA transforms to.
1802 2014-04-24 Cong Hou <congh@google.com>
1804 PR tree-optimization/60896
1805 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
1806 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
1807 (vect_mark_pattern_stmts): Set the def type of all statements in
1808 PATTERN_DEF_SEQ as vect_internal_def.
1810 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
1812 * doc/extend.texi (PowerPC Built-in Functions): Document new
1813 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
1815 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
1817 * config/rs6000/predicates.md (const_0_to_3_operand): New
1818 predicate to match 0..3 integer constants.
1820 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
1821 to support adding miscellaneous builtin functions.
1822 (BU_DFP_MISC_2): Likewise.
1823 (BU_P7_MISC_1): Likewise.
1824 (BU_P7_MISC_2): Likewise.
1825 (BU_P8V_MISC_3): Likewise.
1826 (BU_MISC_1): Likewise.
1827 (BU_MISC_2): Likewise.
1828 (DIVWE): Add extended divide builtin functions.
1831 (DIVWEUO): Likewise.
1835 (DIVDEUO): Likewise.
1836 (DXEX): Add decimal floating-point builtin functions.
1839 (DDEDPDQ): Likewise.
1841 (DENBCDQ): Likewise.
1848 (CDTBCD): Add new BCD builtin functions.
1852 (BCDADD_LT): Likewise.
1853 (BCDADD_EQ): Likewise.
1854 (BCDADD_GT): Likewise.
1855 (BCDADD_OV): Likewise.
1857 (BCDSUB_LT): Likewise.
1858 (BCDSUB_EQ): Likewise.
1859 (BCDSUB_GT): Likewise.
1860 (BCDSUB_OV): Likewise.
1861 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
1862 (UNPACK_TD): Likewise.
1863 (PACK_TF): Likewise.
1864 (UNPACK_TF): Likewise.
1865 (UNPACK_TF_0): Likewise.
1866 (UNPACK_TF_1): Likewise.
1867 (PACK_V1TI): Likewise.
1868 (UNPACK_V1TI): Likewise.
1870 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
1871 support for decimal floating point builtin functions.
1872 (rs6000_expand_ternop_builtin): Add checks for the new builtin
1873 functions that take constant arguments.
1874 (rs6000_invalid_builtin): Add decimal floating point builtin
1876 (rs6000_init_builtins): Setup long double, _Decimal64, and
1877 _Decimal128 types for new builtin functions.
1878 (builtin_function_type): Set the unsigned flags appropriately for
1879 the new builtin functions.
1880 (rs6000_opt_masks): Add support for decimal floating point builtin
1883 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
1884 floating point builtin functions.
1885 (RS6000_BTM_COMMON): Likewise.
1886 (RS6000_BTI_long_double): Likewise.
1887 (RS6000_BTI_dfloat64): Likewise.
1888 (RS6000_BTI_dfloat128): Likewise.
1889 (long_double_type_internal_node): Likewise.
1890 (dfloat64_type_internal_node): Likewise.
1891 (dfloat128_type_internal_node): Likewise.
1893 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
1894 2.07 bcd arithmetic instructions.
1895 (UNSPEC_BCDSUB): Likewise.
1896 (UNSPEC_BCD_OVERFLOW): Likewise.
1897 (UNSPEC_BCD_ADD_SUB): Likewise.
1898 (bcd_add_sub): Likewise.
1899 (BCD_TEST): Likewise.
1900 (bcd<bcd_add_sub>): Likewise.
1901 (bcd<bcd_add_sub>_test): Likewise.
1902 (bcd<bcd_add_sub>_test2): Likewise.
1903 (bcd<bcd_add_sub>_<code>): Likewise.
1904 (peephole2 for combined bcd ops): Likewise.
1906 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
1907 decimal floating point builtin functions.
1908 (UNSPEC_DENBCD): Likewise.
1909 (UNSPEC_DXEX): Likewise.
1910 (UNSPEC_DIEX): Likewise.
1911 (UNSPEC_DSCLI): Likewise.
1912 (UNSPEC_DSCRI): Likewise.
1913 (D64_D128): Likewise.
1914 (dfp_suffix): Likewise.
1915 (dfp_ddedpd_<mode>): Likewise.
1916 (dfp_denbcd_<mode>): Likewise.
1917 (dfp_dxex_<mode>): Likewise.
1918 (dfp_diex_<mode>): Likewise.
1919 (dfp_dscli_<mode>): Likewise.
1920 (dfp_dscri_<mode>): Likewise.
1922 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
1924 (UNSPEC_CDTBCD): Likewise.
1925 (UNSPEC_CBCDTD): Likewise.
1926 (UNSPEC_DIVE): Add support for new extended divide builtin
1928 (UNSPEC_DIVEO): Likewise.
1929 (UNSPEC_DIVEU): Likewise.
1930 (UNSPEC_DIVEUO): Likewise.
1931 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
1932 pack/unpack 128-bit types.
1933 (UNSPEC_PACK_128BIT): Likewise.
1934 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
1935 (udiv<mode>3): Use idiv_ldiv mode attribute.
1936 (div<mode>3): Likewise.
1937 (addg6s): Add new BCD builtin functions.
1940 (UNSPEC_DIV_EXTEND): Add support for new extended divide
1942 (div_extend): Likewise.
1943 (div<div_extend>_<mode>"): Likewise.
1944 (FP128_64): Add support for new builtin functions to pack/unpack
1946 (unpack<mode>): Likewise.
1947 (unpacktf_0): Likewise.
1948 (unpacktf_1): Likewise.
1949 (unpack<mode>_dm): Likewise.
1950 (unpack<mode>_nodm): Likewise.
1951 (pack<mode>): Likewise.
1952 (unpackv1ti): Likewise.
1953 (packv1ti): Likewise.
1955 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
1957 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
1960 2014-04-24 Jakub Jelinek <jakub@redhat.com>
1962 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
1963 * gimplify.c (omp_is_private): Change last argument's type to int.
1964 Only diagnose lastprivate if the simd argument is 1, only diagnose
1965 linear if the simd argument is 2.
1966 (gimplify_omp_for): Adjust omp_is_private callers. When adding
1967 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
1968 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
1969 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
1970 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
1971 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
1972 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
1973 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
1974 * tree-nested.c (convert_nonlocal_omp_clauses,
1975 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
1977 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
1980 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
1983 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
1985 * flag-types.h (enum ivar_visibility): Add.
1987 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
1989 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
1990 function * argument.
1992 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
1994 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
1996 2014-04-24 Radovan Obradovic <robradovic@mips.com>
1997 Tom de Vries <tom@codesourcery.com>
1999 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
2000 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
2002 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
2003 * emit-rtl.c (try_split): Same.
2005 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2006 Tom de Vries <tom@codesourcery.com>
2008 * common.opt (fuse-caller-save): New option.
2010 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
2012 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
2015 2014-04-24 Richard Biener <rguenther@suse.de>
2017 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
2018 during TER and instead use the sepops interface for expanding
2019 non-GIMPLE_SINGLE_RHS.
2021 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2023 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
2024 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
2026 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2028 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
2029 assembler 64-bit option.
2030 * configure: Regenerate.
2032 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2034 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
2035 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
2036 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
2037 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
2038 (TARGET_CRYPTO): Take TARGET_SIMD into account.
2040 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2042 * config/aarch64/aarch64-builtins.c
2043 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
2044 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
2045 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
2046 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
2048 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
2049 (Vrevsuff): New mode attribute.
2051 2014-04-24 Terry Guo <terry.guo@arm.com>
2053 * config/arm/arm.h (machine_function): Define variable
2054 after_arm_reorg here.
2055 * config/arm/arm.c (after_arm_reorg): Remove the definition.
2056 (arm_split_constant): Update the way to access variable
2059 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
2061 2014-04-23 Tom de Vries <tom@codesourcery.com>
2063 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
2065 2014-04-23 David Malcolm <dmalcolm@redhat.com>
2067 * is-a.h: Update comments to reflect the following changes to the
2068 "pointerness" of the API, making the template parameter match the
2069 return type, allowing use of is-a.h with typedefs of pointers.
2070 (is_a_helper::cast): Return a T rather then a pointer to a T, so
2071 that the return type matches the parameter to the is_a_helper.
2073 (dyn_cast): Likewise.
2075 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
2076 pointer from the is-a.h API.
2078 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
2079 (is_a_helper <cgraph_node *>::test): ...this, matching change to
2081 (is_a_helper <varpool_node>::test): Likewise, convert to...
2082 (is_a_helper <varpool_node *>::test): ...this.
2084 (varpool_first_variable): Update for removal of implicit pointer
2085 from the is-a.h API.
2086 (varpool_next_variable): Likewise.
2087 (varpool_first_static_initializer): Likewise.
2088 (varpool_next_static_initializer): Likewise.
2089 (varpool_first_defined_variable): Likewise.
2090 (varpool_next_defined_variable): Likewise.
2091 (cgraph_first_defined_function): Likewise.
2092 (cgraph_next_defined_function): Likewise.
2093 (cgraph_first_function): Likewise.
2094 (cgraph_next_function): Likewise.
2095 (cgraph_first_function_with_gimple_body): Likewise.
2096 (cgraph_next_function_with_gimple_body): Likewise.
2097 (cgraph_alias_target): Likewise.
2098 (varpool_alias_target): Likewise.
2099 (cgraph_function_or_thunk_node): Likewise.
2100 (varpool_variable_node): Likewise.
2101 (symtab_real_symbol_p): Likewise.
2102 * cgraphunit.c (referred_to_p): Likewise.
2103 (analyze_functions): Likewise.
2104 (handle_alias_pairs): Likewise.
2105 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2106 * gimple-ssa.h (gimple_vuse_op): Likewise.
2107 (gimple_vdef_op): Likewise.
2108 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
2109 * gimple.c (gimple_build_asm_1): Likewise.
2110 (gimple_build_try): Likewise.
2111 (gimple_build_resx): Likewise.
2112 (gimple_build_eh_dispatch): Likewise.
2113 (gimple_build_omp_for): Likewise.
2114 (gimple_omp_for_set_clauses): Likewise.
2116 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
2117 (is_a_helper <gimple_statement_asm *>::test): ...this.
2118 (is_a_helper <gimple_statement_bind>::test): Convert to...
2119 (is_a_helper <gimple_statement_bind *>::test): ...this.
2120 (is_a_helper <gimple_statement_call>::test): Convert to...
2121 (is_a_helper <gimple_statement_call *>::test): ...this.
2122 (is_a_helper <gimple_statement_catch>::test): Convert to...
2123 (is_a_helper <gimple_statement_catch *>::test): ...this.
2124 (is_a_helper <gimple_statement_resx>::test): Convert to...
2125 (is_a_helper <gimple_statement_resx *>::test): ...this.
2126 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
2127 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
2128 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
2129 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
2130 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
2131 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
2132 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
2133 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
2134 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
2135 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
2136 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
2137 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
2138 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
2139 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
2140 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
2141 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
2142 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
2143 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
2144 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
2145 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
2146 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
2147 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
2148 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
2149 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
2150 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
2151 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
2152 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
2153 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
2154 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
2155 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
2156 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
2157 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
2158 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
2159 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
2160 (is_a_helper <gimple_statement_phi>::test): Convert to...
2161 (is_a_helper <gimple_statement_phi *>::test): ...this.
2162 (is_a_helper <gimple_statement_transaction>::test): Convert to...
2163 (is_a_helper <gimple_statement_transaction *>::test): ...this.
2164 (is_a_helper <gimple_statement_try>::test): Convert to...
2165 (is_a_helper <gimple_statement_try *>::test): ...this.
2166 (is_a_helper <gimple_statement_wce>::test): Convert to...
2167 (is_a_helper <gimple_statement_wce *>::test): ...this.
2168 (is_a_helper <const gimple_statement_asm>::test): Convert to...
2169 (is_a_helper <const gimple_statement_asm *>::test): ...this.
2170 (is_a_helper <const gimple_statement_bind>::test): Convert to...
2171 (is_a_helper <const gimple_statement_bind *>::test): ...this.
2172 (is_a_helper <const gimple_statement_call>::test): Convert to...
2173 (is_a_helper <const gimple_statement_call *>::test): ...this.
2174 (is_a_helper <const gimple_statement_catch>::test): Convert to...
2175 (is_a_helper <const gimple_statement_catch *>::test): ...this.
2176 (is_a_helper <const gimple_statement_resx>::test): Convert to...
2177 (is_a_helper <const gimple_statement_resx *>::test): ...this.
2178 (is_a_helper <const gimple_statement_eh_dispatch>::test):
2180 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
2182 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
2184 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
2185 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
2187 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
2189 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
2191 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
2193 (is_a_helper <const gimple_statement_omp_return>::test): Convert
2195 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
2196 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
2198 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
2199 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
2201 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
2202 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
2203 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
2204 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
2206 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
2207 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
2209 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
2210 (is_a_helper <const gimple_statement_omp_target>::test): Convert
2212 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
2213 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
2215 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
2216 (is_a_helper <const gimple_statement_omp_single>::test): Convert
2218 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
2219 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
2221 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
2222 (is_a_helper <const gimple_statement_omp_task>::test): Convert
2224 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
2225 (is_a_helper <const gimple_statement_phi>::test): Convert to...
2226 (is_a_helper <const gimple_statement_phi *>::test): ...this.
2227 (is_a_helper <const gimple_statement_transaction>::test): Convert
2229 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
2230 (is_a_helper <const gimple_statement_with_ops>::test): Convert
2232 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
2233 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
2234 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
2235 (is_a_helper <const gimple_statement_with_memory_ops>::test):
2237 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
2239 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
2241 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
2243 (gimple_use_ops): Update for removal of implicit pointer from the
2245 (gimple_set_use_ops): Likewise.
2246 (gimple_vuse): Likewise.
2247 (gimple_vdef): Likewise.
2248 (gimple_vuse_ptr): Likewise.
2249 (gimple_vdef_ptr): Likewise.
2250 (gimple_set_vuse): Likewise.
2251 (gimple_set_vdef): Likewise.
2252 (gimple_omp_return_set_lhs): Likewise.
2253 (gimple_omp_return_lhs): Likewise.
2254 (gimple_omp_return_lhs_ptr): Likewise.
2255 (gimple_call_fntype): Likewise.
2256 (gimple_call_set_fntype): Likewise.
2257 (gimple_call_set_internal_fn): Likewise.
2258 (gimple_call_use_set): Likewise.
2259 (gimple_call_clobber_set): Likewise.
2260 (gimple_bind_vars): Likewise.
2261 (gimple_bind_set_vars): Likewise.
2262 (gimple_bind_body_ptr): Likewise.
2263 (gimple_bind_set_body): Likewise.
2264 (gimple_bind_add_stmt): Likewise.
2265 (gimple_bind_block): Likewise.
2266 (gimple_bind_set_block): Likewise.
2267 (gimple_asm_ninputs): Likewise.
2268 (gimple_asm_noutputs): Likewise.
2269 (gimple_asm_nclobbers): Likewise.
2270 (gimple_asm_nlabels): Likewise.
2271 (gimple_asm_input_op): Likewise.
2272 (gimple_asm_input_op_ptr): Likewise.
2273 (gimple_asm_output_op): Likewise.
2274 (gimple_asm_output_op_ptr): Likewise.
2275 (gimple_asm_set_output_op): Likewise.
2276 (gimple_asm_clobber_op): Likewise.
2277 (gimple_asm_set_clobber_op): Likewise.
2278 (gimple_asm_label_op): Likewise.
2279 (gimple_asm_set_label_op): Likewise.
2280 (gimple_asm_string): Likewise.
2281 (gimple_catch_types): Likewise.
2282 (gimple_catch_types_ptr): Likewise.
2283 (gimple_catch_handler_ptr): Likewise.
2284 (gimple_catch_set_types): Likewise.
2285 (gimple_catch_set_handler): Likewise.
2286 (gimple_eh_filter_types): Likewise.
2287 (gimple_eh_filter_types_ptr): Likewise.
2288 (gimple_eh_filter_failure_ptr): Likewise.
2289 (gimple_eh_filter_set_types): Likewise.
2290 (gimple_eh_filter_set_failure): Likewise.
2291 (gimple_eh_must_not_throw_fndecl): Likewise.
2292 (gimple_eh_must_not_throw_set_fndecl): Likewise.
2293 (gimple_eh_else_n_body_ptr): Likewise.
2294 (gimple_eh_else_e_body_ptr): Likewise.
2295 (gimple_eh_else_set_n_body): Likewise.
2296 (gimple_eh_else_set_e_body): Likewise.
2297 (gimple_try_eval_ptr): Likewise.
2298 (gimple_try_cleanup_ptr): Likewise.
2299 (gimple_try_set_eval): Likewise.
2300 (gimple_try_set_cleanup): Likewise.
2301 (gimple_wce_cleanup_ptr): Likewise.
2302 (gimple_wce_set_cleanup): Likewise.
2303 (gimple_phi_capacity): Likewise.
2304 (gimple_phi_num_args): Likewise.
2305 (gimple_phi_result): Likewise.
2306 (gimple_phi_result_ptr): Likewise.
2307 (gimple_phi_set_result): Likewise.
2308 (gimple_phi_arg): Likewise.
2309 (gimple_phi_set_arg): Likewise.
2310 (gimple_resx_region): Likewise.
2311 (gimple_resx_set_region): Likewise.
2312 (gimple_eh_dispatch_region): Likewise.
2313 (gimple_eh_dispatch_set_region): Likewise.
2314 (gimple_omp_critical_name): Likewise.
2315 (gimple_omp_critical_name_ptr): Likewise.
2316 (gimple_omp_critical_set_name): Likewise.
2317 (gimple_omp_for_clauses): Likewise.
2318 (gimple_omp_for_clauses_ptr): Likewise.
2319 (gimple_omp_for_set_clauses): Likewise.
2320 (gimple_omp_for_collapse): Likewise.
2321 (gimple_omp_for_index): Likewise.
2322 (gimple_omp_for_index_ptr): Likewise.
2323 (gimple_omp_for_set_index): Likewise.
2324 (gimple_omp_for_initial): Likewise.
2325 (gimple_omp_for_initial_ptr): Likewise.
2326 (gimple_omp_for_set_initial): Likewise.
2327 (gimple_omp_for_final): Likewise.
2328 (gimple_omp_for_final_ptr): Likewise.
2329 (gimple_omp_for_set_final): Likewise.
2330 (gimple_omp_for_incr): Likewise.
2331 (gimple_omp_for_incr_ptr): Likewise.
2332 (gimple_omp_for_set_incr): Likewise.
2333 (gimple_omp_for_pre_body_ptr): Likewise.
2334 (gimple_omp_for_set_pre_body): Likewise.
2335 (gimple_omp_parallel_clauses): Likewise.
2336 (gimple_omp_parallel_clauses_ptr): Likewise.
2337 (gimple_omp_parallel_set_clauses): Likewise.
2338 (gimple_omp_parallel_child_fn): Likewise.
2339 (gimple_omp_parallel_child_fn_ptr): Likewise.
2340 (gimple_omp_parallel_set_child_fn): Likewise.
2341 (gimple_omp_parallel_data_arg): Likewise.
2342 (gimple_omp_parallel_data_arg_ptr): Likewise.
2343 (gimple_omp_parallel_set_data_arg): Likewise.
2344 (gimple_omp_task_clauses): Likewise.
2345 (gimple_omp_task_clauses_ptr): Likewise.
2346 (gimple_omp_task_set_clauses): Likewise.
2347 (gimple_omp_task_child_fn): Likewise.
2348 (gimple_omp_task_child_fn_ptr): Likewise.
2349 (gimple_omp_task_set_child_fn): Likewise.
2350 (gimple_omp_task_data_arg): Likewise.
2351 (gimple_omp_task_data_arg_ptr): Likewise.
2352 (gimple_omp_task_set_data_arg): Likewise.
2353 (gimple_omp_taskreg_clauses): Likewise.
2354 (gimple_omp_taskreg_clauses_ptr): Likewise.
2355 (gimple_omp_taskreg_set_clauses): Likewise.
2356 (gimple_omp_taskreg_child_fn): Likewise.
2357 (gimple_omp_taskreg_child_fn_ptr): Likewise.
2358 (gimple_omp_taskreg_set_child_fn): Likewise.
2359 (gimple_omp_taskreg_data_arg): Likewise.
2360 (gimple_omp_taskreg_data_arg_ptr): Likewise.
2361 (gimple_omp_taskreg_set_data_arg): Likewise.
2362 (gimple_omp_task_copy_fn): Likewise.
2363 (gimple_omp_task_copy_fn_ptr): Likewise.
2364 (gimple_omp_task_set_copy_fn): Likewise.
2365 (gimple_omp_task_arg_size): Likewise.
2366 (gimple_omp_task_arg_size_ptr): Likewise.
2367 (gimple_omp_task_set_arg_size): Likewise.
2368 (gimple_omp_task_arg_align): Likewise.
2369 (gimple_omp_task_arg_align_ptr): Likewise.
2370 (gimple_omp_task_set_arg_align): Likewise.
2371 (gimple_omp_single_clauses): Likewise.
2372 (gimple_omp_single_clauses_ptr): Likewise.
2373 (gimple_omp_single_set_clauses): Likewise.
2374 (gimple_omp_target_clauses): Likewise.
2375 (gimple_omp_target_clauses_ptr): Likewise.
2376 (gimple_omp_target_set_clauses): Likewise.
2377 (gimple_omp_target_child_fn): Likewise.
2378 (gimple_omp_target_child_fn_ptr): Likewise.
2379 (gimple_omp_target_set_child_fn): Likewise.
2380 (gimple_omp_target_data_arg): Likewise.
2381 (gimple_omp_target_data_arg_ptr): Likewise.
2382 (gimple_omp_target_set_data_arg): Likewise.
2383 (gimple_omp_teams_clauses): Likewise.
2384 (gimple_omp_teams_clauses_ptr): Likewise.
2385 (gimple_omp_teams_set_clauses): Likewise.
2386 (gimple_omp_sections_clauses): Likewise.
2387 (gimple_omp_sections_clauses_ptr): Likewise.
2388 (gimple_omp_sections_set_clauses): Likewise.
2389 (gimple_omp_sections_control): Likewise.
2390 (gimple_omp_sections_control_ptr): Likewise.
2391 (gimple_omp_sections_set_control): Likewise.
2392 (gimple_omp_for_set_cond): Likewise.
2393 (gimple_omp_for_cond): Likewise.
2394 (gimple_omp_atomic_store_set_val): Likewise.
2395 (gimple_omp_atomic_store_val): Likewise.
2396 (gimple_omp_atomic_store_val_ptr): Likewise.
2397 (gimple_omp_atomic_load_set_lhs): Likewise.
2398 (gimple_omp_atomic_load_lhs): Likewise.
2399 (gimple_omp_atomic_load_lhs_ptr): Likewise.
2400 (gimple_omp_atomic_load_set_rhs): Likewise.
2401 (gimple_omp_atomic_load_rhs): Likewise.
2402 (gimple_omp_atomic_load_rhs_ptr): Likewise.
2403 (gimple_omp_continue_control_def): Likewise.
2404 (gimple_omp_continue_control_def_ptr): Likewise.
2405 (gimple_omp_continue_set_control_def): Likewise.
2406 (gimple_omp_continue_control_use): Likewise.
2407 (gimple_omp_continue_control_use_ptr): Likewise.
2408 (gimple_omp_continue_set_control_use): Likewise.
2409 (gimple_transaction_body_ptr): Likewise.
2410 (gimple_transaction_label): Likewise.
2411 (gimple_transaction_label_ptr): Likewise.
2412 (gimple_transaction_set_body): Likewise.
2413 (gimple_transaction_set_label): Likewise.
2415 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
2416 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2417 * ipa-ref.c (ipa_record_reference): Likewise.
2418 * ipa-reference.c (analyze_function): Likewise.
2419 (ipa_reference_write_optimization_summary): Likewise.
2420 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
2421 (address_taken_from_non_vtable_p): Likewise.
2422 (comdat_can_be_unshared_p_1): Likewise.
2423 * lto-cgraph.c (lto_output_ref): Likewise.
2424 (add_references): Likewise.
2425 (compute_ltrans_boundary): Likewise.
2426 (output_symtab): Likewise.
2427 (input_ref): Likewise.
2428 (input_cgraph_1): Likewise.
2429 (output_cgraph_opt_summary): Likewise.
2430 * lto-streamer-out.c (lto_output): Likewise.
2431 (output_symbol_p): Likewise.
2432 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
2433 (lsei_start_function_in_partition): Likewise.
2434 (lsei_next_variable_in_partition): Likewise.
2435 (lsei_start_variable_in_partition): Likewise.
2436 * symtab.c (insert_to_assembler_name_hash): Likewise.
2437 (unlink_from_assembler_name_hash): Likewise.
2438 (symtab_unregister_node): Likewise.
2439 (symtab_remove_node): Likewise.
2440 (dump_symtab_node): Likewise.
2441 (verify_symtab_base): Likewise.
2442 (verify_symtab_node): Likewise.
2443 (symtab_make_decl_local): Likewise.
2444 (symtab_alias_ultimate_target): Likewise.
2445 (symtab_resolve_alias): Likewise.
2446 (symtab_get_symbol_partitioning_class): Likewise.
2447 * tree-phinodes.c (allocate_phi_node): Likewise.
2448 (reserve_phi_args_for_new_edge): Likewise.
2449 (remove_phi_args): Likewise.
2450 * varpool.c (varpool_node_for_asm): Likewise.
2451 (varpool_remove_unreferenced_decls): Likewise.
2453 2014-04-23 Jeff Law <law@redhat.com>
2455 PR tree-optimization/60902
2456 * tree-ssa-threadedge.c
2457 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
2458 invalidate outputs from statements that do not produce useful
2459 outputs for threading.
2461 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2463 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
2464 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
2465 machine descriptions for Stack Smashing Protector.
2467 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
2469 * aarch64.md (<optab>_rol<mode>3): New pattern.
2470 (<optab>_rolsi3_uxtw): Likewise.
2471 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
2473 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
2475 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
2476 (arm_cortex_a12_tune): Likewise.
2478 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2480 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
2482 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2484 * config/arm/arm.md (arm_rev16si2): New pattern.
2485 (arm_rev16si2_alt): Likewise.
2486 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
2488 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2490 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
2491 (rev16<mode>2_alt): Likewise.
2492 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
2493 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
2494 (aarch_rev16_shleft_mask_imm_p): Likewise.
2495 (aarch_rev16_p_1): Likewise.
2496 (aarch_rev16_p): Likewise.
2497 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
2498 (aarch_rev16_shright_mask_imm_p): Likewise.
2499 (aarch_rev16_shleft_mask_imm_p): Likewise.
2501 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2503 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
2504 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
2506 (cortex_a53_extra_costs): Likewise.
2507 (cortex_a57_extra_costs): Likewise.
2508 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
2509 (cortexa7_extra_costs): Likewise.
2510 (cortexa8_extra_costs): Likewise.
2511 (cortexa12_extra_costs): Likewise.
2512 (cortexa15_extra_costs): Likewise.
2513 (v7m_extra_costs): Likewise.
2514 (arm_new_rtx_costs): Handle BSWAP.
2516 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2518 * config/arm/arm.c (cortexa8_extra_costs): New table.
2519 (arm_cortex_a8_tune): New tuning struct.
2520 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
2522 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2524 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
2526 2014-04-23 Richard Biener <rguenther@suse.de>
2528 * Makefile.in (OBJS): Remove loop-unswitch.o.
2529 * tree-pass.h (make_pass_rtl_unswitch): Remove.
2530 * passes.def (pass_rtl_unswitch): Likewise.
2531 * loop-init.c (gate_rtl_unswitch): Likewise.
2532 (rtl_unswitch): Likewise.
2533 (pass_data_rtl_unswitch): Likewise.
2534 (pass_rtl_unswitch): Likewise.
2535 (make_pass_rtl_unswitch): Likewise.
2536 * rtl.h (reversed_condition): Likewise.
2537 (compare_and_jump_seq): Likewise.
2538 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
2540 * loop-unroll.c (compare_and_jump_seq): Likewise.
2542 2014-04-23 Richard Biener <rguenther@suse.de>
2544 PR tree-optimization/60903
2545 * tree-ssa-loop-im.c (analyze_memory_references): Remove
2546 commented code block.
2547 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
2548 loop flags to newly created BBs and edges.
2550 2014-04-23 Nick Clifton <nickc@redhat.com>
2552 * config/msp430/msp430.c (msp430_handle_option): Move function
2554 (msp430_option_override): Simplify mcu and mcpu option handling.
2555 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
2556 support for -mhwmult command line option.
2557 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
2558 -mhwmult command line option.
2559 (msp430_hwmult_enabled): Delete.
2560 (msp43o_output_labelref): Add support for -mhwmult command line
2562 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
2563 (umulsidi3): Likewise.
2564 * config/msp430/msp430.opt (mmcu): Add Report attribute.
2565 (mcpu, mlarge, msmall): Likewise.
2566 (mhwmult): New option.
2567 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
2569 (msp430_is_f5_mcu): Remove prototype.
2570 (msp430_use_f5_series_hwmult): Add prototype.
2571 * config/msp430/msp430-opts.h: New file.
2572 * common/config/msp430: New directory.
2573 * common/config/msp430/msp430-common.c: New file.
2574 * config.gcc (msp430): Remove target_has_targetm_common.
2575 * doc/invoke.texi: Document -mhwmult command line option.
2577 2014-04-23 Nick Clifton <nickc@redhat.com>
2579 * config/i386/cygwin.h (ENDFILE_SPEC): Include
2580 default-manifest.o if it can be found in the search path.
2581 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
2583 2014-04-23 Terry Guo <terry.guo@arm.com>
2585 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
2587 2014-04-23 Richard Biener <rguenther@suse.de>
2590 * tree-inline.c (declare_return_variable): Use mark_addressable.
2592 2014-04-23 Richard Biener <rguenther@suse.de>
2595 * loop-init.c (loop_optimizer_init): Make sure to apply
2596 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
2598 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2601 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
2603 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
2604 if flag_sanitize_undefined_trap_on_error.
2605 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
2606 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
2607 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
2608 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
2609 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
2610 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
2611 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
2612 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
2613 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
2614 * ubsan.c (ubsan_instrument_unreachable): Return
2615 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
2616 (ubsan_expand_null_ifn): Emit __builtin_trap ()
2617 if flag_sanitize_undefined_trap_on_error and
2618 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
2619 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
2620 instrument_bool_enum_load): Emit __builtin_trap () if
2621 flag_sanitize_undefined_trap_on_error and
2622 __builtin_handle_*_abort () if !flag_sanitize_recover.
2623 * doc/invoke.texi (-fsanitize-recover,
2624 -fsanitize-undefined-trap-on-error): Document.
2626 2014-04-22 Christian Bruel <christian.bruel@st.com>
2628 * config/sh/sh.md (mov<mode>): Replace movQIHI.
2629 Force immediates to SImode.
2631 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
2633 * config/nios2/nios2.md (UNSPEC_ROUND): New.
2635 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
2636 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
2637 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
2638 (nios2_fpu_insn): Add entry for round.
2639 (N2FPU_NO_ERRNO_P): Define.
2640 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
2642 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
2644 2014-04-22 Richard Henderson <rth@redhat.com>
2646 * config/aarch64/aarch64 (addti3, subti3): New expanders.
2647 (add<GPI>3_compare0): Remove leading * from name.
2648 (add<GPI>3_carryin): Likewise.
2649 (sub<GPI>3_compare0): Likewise.
2650 (sub<GPI>3_carryin): Likewise.
2651 (<su_optab>mulditi3): New expander.
2652 (multi3): New expander.
2653 (madd<GPI>): Remove leading * from name.
2655 2014-04-22 Martin Jambor <mjambor@suse.cz>
2657 * cgraphclones.c (cgraph_function_versioning): Copy
2658 ipa_transforms_to_apply instead of asserting it is empty.
2660 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
2663 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
2664 on count_exp to get mode.
2666 2014-04-22 Andrew Pinski <apinski@cavium.com>
2668 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
2669 Handle TLS for ILP32.
2670 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
2671 (tlsie_small_<mode>): this and handle PTR.
2672 (tlsie_small_sidi): New pattern.
2673 (tlsle_small): Change to an expand to handle ILP32.
2674 (tlsle_small_<mode>): New pattern.
2675 (tlsdesc_small): Rename to ...
2676 (tlsdesc_small_<mode>): this and handle PTR.
2678 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2680 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
2682 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2684 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2685 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
2686 (aarch64_types_signed_poly_qualifiers): Likewise.
2687 (aarch64_types_unsigned_signed_qualifiers): Likewise.
2688 (aarch64_types_poly_signed_qualifiers): Likewise.
2689 (TYPES_REINTERP_SS): Type macro added.
2690 (TYPES_REINTERP_SU): Likewise.
2691 (TYPES_REINTERP_SP): Likewise.
2692 (TYPES_REINTERP_US): Likewise.
2693 (TYPES_REINTERP_PS): Likewise.
2694 (aarch64_fold_builtin): New expression folding added.
2695 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
2696 Declarations removed.
2697 (REINTERP_SS): Declarations added.
2698 (REINTERP_US): Likewise.
2699 (REINTERP_PS): Likewise.
2700 (REINTERP_SU): Likewise.
2701 (REINTERP_SP): Likewise.
2702 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
2703 (vreinterpretq_p8_f64): Likewise.
2704 (vreinterpret_p16_f64): Likewise.
2705 (vreinterpretq_p16_f64): Likewise.
2706 (vreinterpret_f32_f64): Likewise.
2707 (vreinterpretq_f32_f64): Likewise.
2708 (vreinterpret_f64_f32): Likewise.
2709 (vreinterpret_f64_p8): Likewise.
2710 (vreinterpret_f64_p16): Likewise.
2711 (vreinterpret_f64_s8): Likewise.
2712 (vreinterpret_f64_s16): Likewise.
2713 (vreinterpret_f64_s32): Likewise.
2714 (vreinterpret_f64_s64): Likewise.
2715 (vreinterpret_f64_u8): Likewise.
2716 (vreinterpret_f64_u16): Likewise.
2717 (vreinterpret_f64_u32): Likewise.
2718 (vreinterpret_f64_u64): Likewise.
2719 (vreinterpretq_f64_f32): Likewise.
2720 (vreinterpretq_f64_p8): Likewise.
2721 (vreinterpretq_f64_p16): Likewise.
2722 (vreinterpretq_f64_s8): Likewise.
2723 (vreinterpretq_f64_s16): Likewise.
2724 (vreinterpretq_f64_s32): Likewise.
2725 (vreinterpretq_f64_s64): Likewise.
2726 (vreinterpretq_f64_u8): Likewise.
2727 (vreinterpretq_f64_u16): Likewise.
2728 (vreinterpretq_f64_u32): Likewise.
2729 (vreinterpretq_f64_u64): Likewise.
2730 (vreinterpret_s64_f64): Likewise.
2731 (vreinterpretq_s64_f64): Likewise.
2732 (vreinterpret_u64_f64): Likewise.
2733 (vreinterpretq_u64_f64): Likewise.
2734 (vreinterpret_s8_f64): Likewise.
2735 (vreinterpretq_s8_f64): Likewise.
2736 (vreinterpret_s16_f64): Likewise.
2737 (vreinterpretq_s16_f64): Likewise.
2738 (vreinterpret_s32_f64): Likewise.
2739 (vreinterpretq_s32_f64): Likewise.
2740 (vreinterpret_u8_f64): Likewise.
2741 (vreinterpretq_u8_f64): Likewise.
2742 (vreinterpret_u16_f64): Likewise.
2743 (vreinterpretq_u16_f64): Likewise.
2744 (vreinterpret_u32_f64): Likewise.
2745 (vreinterpretq_u32_f64): Likewise.
2747 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2749 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2750 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
2751 (vreinterpret_p8_s8): Likewise.
2752 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
2753 (vreinterpret_p8_s16): Likewise.
2754 (vreinterpret_p8_s32): Likewise.
2755 (vreinterpret_p8_s64): Likewise.
2756 (vreinterpret_p8_f32): Likewise.
2757 (vreinterpret_p8_u8): Likewise.
2758 (vreinterpret_p8_u16): Likewise.
2759 (vreinterpret_p8_u32): Likewise.
2760 (vreinterpret_p8_u64): Likewise.
2761 (vreinterpret_p8_p16): Likewise.
2762 (vreinterpretq_p8_s8): Likewise.
2763 (vreinterpretq_p8_s16): Likewise.
2764 (vreinterpretq_p8_s32): Likewise.
2765 (vreinterpretq_p8_s64): Likewise.
2766 (vreinterpretq_p8_f32): Likewise.
2767 (vreinterpretq_p8_u8): Likewise.
2768 (vreinterpretq_p8_u16): Likewise.
2769 (vreinterpretq_p8_u32): Likewise.
2770 (vreinterpretq_p8_u64): Likewise.
2771 (vreinterpretq_p8_p16): Likewise.
2772 (vreinterpret_p16_s8): Likewise.
2773 (vreinterpret_p16_s16): Likewise.
2774 (vreinterpret_p16_s32): Likewise.
2775 (vreinterpret_p16_s64): Likewise.
2776 (vreinterpret_p16_f32): Likewise.
2777 (vreinterpret_p16_u8): Likewise.
2778 (vreinterpret_p16_u16): Likewise.
2779 (vreinterpret_p16_u32): Likewise.
2780 (vreinterpret_p16_u64): Likewise.
2781 (vreinterpret_p16_p8): Likewise.
2782 (vreinterpretq_p16_s8): Likewise.
2783 (vreinterpretq_p16_s16): Likewise.
2784 (vreinterpretq_p16_s32): Likewise.
2785 (vreinterpretq_p16_s64): Likewise.
2786 (vreinterpretq_p16_f32): Likewise.
2787 (vreinterpretq_p16_u8): Likewise.
2788 (vreinterpretq_p16_u16): Likewise.
2789 (vreinterpretq_p16_u32): Likewise.
2790 (vreinterpretq_p16_u64): Likewise.
2791 (vreinterpretq_p16_p8): Likewise.
2792 (vreinterpret_f32_s8): Likewise.
2793 (vreinterpret_f32_s16): Likewise.
2794 (vreinterpret_f32_s32): Likewise.
2795 (vreinterpret_f32_s64): Likewise.
2796 (vreinterpret_f32_u8): Likewise.
2797 (vreinterpret_f32_u16): Likewise.
2798 (vreinterpret_f32_u32): Likewise.
2799 (vreinterpret_f32_u64): Likewise.
2800 (vreinterpret_f32_p8): Likewise.
2801 (vreinterpret_f32_p16): Likewise.
2802 (vreinterpretq_f32_s8): Likewise.
2803 (vreinterpretq_f32_s16): Likewise.
2804 (vreinterpretq_f32_s32): Likewise.
2805 (vreinterpretq_f32_s64): Likewise.
2806 (vreinterpretq_f32_u8): Likewise.
2807 (vreinterpretq_f32_u16): Likewise.
2808 (vreinterpretq_f32_u32): Likewise.
2809 (vreinterpretq_f32_u64): Likewise.
2810 (vreinterpretq_f32_p8): Likewise.
2811 (vreinterpretq_f32_p16): Likewise.
2812 (vreinterpret_s64_s8): Likewise.
2813 (vreinterpret_s64_s16): Likewise.
2814 (vreinterpret_s64_s32): Likewise.
2815 (vreinterpret_s64_f32): Likewise.
2816 (vreinterpret_s64_u8): Likewise.
2817 (vreinterpret_s64_u16): Likewise.
2818 (vreinterpret_s64_u32): Likewise.
2819 (vreinterpret_s64_u64): Likewise.
2820 (vreinterpret_s64_p8): Likewise.
2821 (vreinterpret_s64_p16): Likewise.
2822 (vreinterpretq_s64_s8): Likewise.
2823 (vreinterpretq_s64_s16): Likewise.
2824 (vreinterpretq_s64_s32): Likewise.
2825 (vreinterpretq_s64_f32): Likewise.
2826 (vreinterpretq_s64_u8): Likewise.
2827 (vreinterpretq_s64_u16): Likewise.
2828 (vreinterpretq_s64_u32): Likewise.
2829 (vreinterpretq_s64_u64): Likewise.
2830 (vreinterpretq_s64_p8): Likewise.
2831 (vreinterpretq_s64_p16): Likewise.
2832 (vreinterpret_u64_s8): Likewise.
2833 (vreinterpret_u64_s16): Likewise.
2834 (vreinterpret_u64_s32): Likewise.
2835 (vreinterpret_u64_s64): Likewise.
2836 (vreinterpret_u64_f32): Likewise.
2837 (vreinterpret_u64_u8): Likewise.
2838 (vreinterpret_u64_u16): Likewise.
2839 (vreinterpret_u64_u32): Likewise.
2840 (vreinterpret_u64_p8): Likewise.
2841 (vreinterpret_u64_p16): Likewise.
2842 (vreinterpretq_u64_s8): Likewise.
2843 (vreinterpretq_u64_s16): Likewise.
2844 (vreinterpretq_u64_s32): Likewise.
2845 (vreinterpretq_u64_s64): Likewise.
2846 (vreinterpretq_u64_f32): Likewise.
2847 (vreinterpretq_u64_u8): Likewise.
2848 (vreinterpretq_u64_u16): Likewise.
2849 (vreinterpretq_u64_u32): Likewise.
2850 (vreinterpretq_u64_p8): Likewise.
2851 (vreinterpretq_u64_p16): Likewise.
2852 (vreinterpret_s8_s16): Likewise.
2853 (vreinterpret_s8_s32): Likewise.
2854 (vreinterpret_s8_s64): Likewise.
2855 (vreinterpret_s8_f32): Likewise.
2856 (vreinterpret_s8_u8): Likewise.
2857 (vreinterpret_s8_u16): Likewise.
2858 (vreinterpret_s8_u32): Likewise.
2859 (vreinterpret_s8_u64): Likewise.
2860 (vreinterpret_s8_p8): Likewise.
2861 (vreinterpret_s8_p16): Likewise.
2862 (vreinterpretq_s8_s16): Likewise.
2863 (vreinterpretq_s8_s32): Likewise.
2864 (vreinterpretq_s8_s64): Likewise.
2865 (vreinterpretq_s8_f32): Likewise.
2866 (vreinterpretq_s8_u8): Likewise.
2867 (vreinterpretq_s8_u16): Likewise.
2868 (vreinterpretq_s8_u32): Likewise.
2869 (vreinterpretq_s8_u64): Likewise.
2870 (vreinterpretq_s8_p8): Likewise.
2871 (vreinterpretq_s8_p16): Likewise.
2872 (vreinterpret_s16_s8): Likewise.
2873 (vreinterpret_s16_s32): Likewise.
2874 (vreinterpret_s16_s64): Likewise.
2875 (vreinterpret_s16_f32): Likewise.
2876 (vreinterpret_s16_u8): Likewise.
2877 (vreinterpret_s16_u16): Likewise.
2878 (vreinterpret_s16_u32): Likewise.
2879 (vreinterpret_s16_u64): Likewise.
2880 (vreinterpret_s16_p8): Likewise.
2881 (vreinterpret_s16_p16): Likewise.
2882 (vreinterpretq_s16_s8): Likewise.
2883 (vreinterpretq_s16_s32): Likewise.
2884 (vreinterpretq_s16_s64): Likewise.
2885 (vreinterpretq_s16_f32): Likewise.
2886 (vreinterpretq_s16_u8): Likewise.
2887 (vreinterpretq_s16_u16): Likewise.
2888 (vreinterpretq_s16_u32): Likewise.
2889 (vreinterpretq_s16_u64): Likewise.
2890 (vreinterpretq_s16_p8): Likewise.
2891 (vreinterpretq_s16_p16): Likewise.
2892 (vreinterpret_s32_s8): Likewise.
2893 (vreinterpret_s32_s16): Likewise.
2894 (vreinterpret_s32_s64): Likewise.
2895 (vreinterpret_s32_f32): Likewise.
2896 (vreinterpret_s32_u8): Likewise.
2897 (vreinterpret_s32_u16): Likewise.
2898 (vreinterpret_s32_u32): Likewise.
2899 (vreinterpret_s32_u64): Likewise.
2900 (vreinterpret_s32_p8): Likewise.
2901 (vreinterpret_s32_p16): Likewise.
2902 (vreinterpretq_s32_s8): Likewise.
2903 (vreinterpretq_s32_s16): Likewise.
2904 (vreinterpretq_s32_s64): Likewise.
2905 (vreinterpretq_s32_f32): Likewise.
2906 (vreinterpretq_s32_u8): Likewise.
2907 (vreinterpretq_s32_u16): Likewise.
2908 (vreinterpretq_s32_u32): Likewise.
2909 (vreinterpretq_s32_u64): Likewise.
2910 (vreinterpretq_s32_p8): Likewise.
2911 (vreinterpretq_s32_p16): Likewise.
2912 (vreinterpret_u8_s8): Likewise.
2913 (vreinterpret_u8_s16): Likewise.
2914 (vreinterpret_u8_s32): Likewise.
2915 (vreinterpret_u8_s64): Likewise.
2916 (vreinterpret_u8_f32): Likewise.
2917 (vreinterpret_u8_u16): Likewise.
2918 (vreinterpret_u8_u32): Likewise.
2919 (vreinterpret_u8_u64): Likewise.
2920 (vreinterpret_u8_p8): Likewise.
2921 (vreinterpret_u8_p16): Likewise.
2922 (vreinterpretq_u8_s8): Likewise.
2923 (vreinterpretq_u8_s16): Likewise.
2924 (vreinterpretq_u8_s32): Likewise.
2925 (vreinterpretq_u8_s64): Likewise.
2926 (vreinterpretq_u8_f32): Likewise.
2927 (vreinterpretq_u8_u16): Likewise.
2928 (vreinterpretq_u8_u32): Likewise.
2929 (vreinterpretq_u8_u64): Likewise.
2930 (vreinterpretq_u8_p8): Likewise.
2931 (vreinterpretq_u8_p16): Likewise.
2932 (vreinterpret_u16_s8): Likewise.
2933 (vreinterpret_u16_s16): Likewise.
2934 (vreinterpret_u16_s32): Likewise.
2935 (vreinterpret_u16_s64): Likewise.
2936 (vreinterpret_u16_f32): Likewise.
2937 (vreinterpret_u16_u8): Likewise.
2938 (vreinterpret_u16_u32): Likewise.
2939 (vreinterpret_u16_u64): Likewise.
2940 (vreinterpret_u16_p8): Likewise.
2941 (vreinterpret_u16_p16): Likewise.
2942 (vreinterpretq_u16_s8): Likewise.
2943 (vreinterpretq_u16_s16): Likewise.
2944 (vreinterpretq_u16_s32): Likewise.
2945 (vreinterpretq_u16_s64): Likewise.
2946 (vreinterpretq_u16_f32): Likewise.
2947 (vreinterpretq_u16_u8): Likewise.
2948 (vreinterpretq_u16_u32): Likewise.
2949 (vreinterpretq_u16_u64): Likewise.
2950 (vreinterpretq_u16_p8): Likewise.
2951 (vreinterpretq_u16_p16): Likewise.
2952 (vreinterpret_u32_s8): Likewise.
2953 (vreinterpret_u32_s16): Likewise.
2954 (vreinterpret_u32_s32): Likewise.
2955 (vreinterpret_u32_s64): Likewise.
2956 (vreinterpret_u32_f32): Likewise.
2957 (vreinterpret_u32_u8): Likewise.
2958 (vreinterpret_u32_u16): Likewise.
2959 (vreinterpret_u32_u64): Likewise.
2960 (vreinterpret_u32_p8): Likewise.
2961 (vreinterpret_u32_p16): Likewise.
2962 (vreinterpretq_u32_s8): Likewise.
2963 (vreinterpretq_u32_s16): Likewise.
2964 (vreinterpretq_u32_s32): Likewise.
2965 (vreinterpretq_u32_s64): Likewise.
2966 (vreinterpretq_u32_f32): Likewise.
2967 (vreinterpretq_u32_u8): Likewise.
2968 (vreinterpretq_u32_u16): Likewise.
2969 (vreinterpretq_u32_u64): Likewise.
2970 (vreinterpretq_u32_p8): Likewise.
2971 (vreinterpretq_u32_p16): Likewise.
2973 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2975 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
2977 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
2980 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
2981 (vqnegd_s64): Likewise.
2982 (vqabs_s64): Likewise.
2983 (vqabsd_s64): Likewise.
2985 2014-04-22 Richard Henderson <rth@redhat.com>
2987 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
2988 computation to the top of the loop.
2990 2014-04-22 Renlin <renlin.li@arm.com>
2991 Jiong Wang <jiong.wang@arm.com>
2993 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
2994 * config/aarch64/aarch64.c (aarch64_layout_frame)
2995 (aarch64_initial_elimination_offset): Likewise.
2997 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
2999 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
3002 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
3004 * machmode.h (bitwise_mode_for_mode): Declare.
3005 * stor-layout.h (bitwise_type_for_mode): Likewise.
3006 * stor-layout.c (bitwise_mode_for_mode): New function.
3007 (bitwise_type_for_mode): Likewise.
3008 * builtins.c (fold_builtin_memory_op): Use it instead of
3009 int_mode_for_mode and build_nonstandard_integer_type.
3011 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3013 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
3014 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
3015 (*-*-solaris2*): Simplify.
3016 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
3017 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
3018 *-*-solaris2.9* handling.
3020 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
3022 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
3023 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
3025 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
3026 * configure: Regenerate.
3028 * config/i386/sol2-9.h: Remove.
3030 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
3031 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
3032 Remove Solaris 9 references.
3034 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
3036 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
3037 (floatuns<GPI:mode><GPF:mode>2): Remove.
3038 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
3039 and floatuns conversions.
3040 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
3041 and floatuns conversions.
3042 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
3043 (w1,w2): New mode attributes for inequal width conversions.
3045 2014-04-22 Renlin Li <Renlin.Li@arm.com>
3047 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
3048 the output asm format.
3050 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
3052 * config/aarch64/aarch64-simd.md
3053 (aarch64_cm<optab>di): Always split.
3054 (*aarch64_cm<optab>di): New.
3055 (aarch64_cmtstdi): Always split.
3056 (*aarch64_cmtstdi): New.
3058 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3060 PR tree-optimization/60823
3061 * omp-low.c (ipa_simd_modify_function_body): Go through
3062 all SSA_NAMEs and for those refering to vector arguments
3063 which are going to be replaced adjust SSA_NAME_VAR and,
3064 if it is a default definition, change it into a non-default
3065 definition assigned at the beginning of function from new_decl.
3066 (ipa_simd_modify_stmt_ops): Rewritten.
3067 * tree-dfa.c (set_ssa_default_def): When removing default def,
3068 check for NULL loc instead of NULL *loc.
3070 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3072 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
3073 restrictions on core registers for DImode values in Thumb2.
3075 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3077 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
3078 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
3080 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3082 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
3083 (*iordi_notzesidi_di): Likewise.
3084 (*iordi_notsesidi_di): Likewise.
3086 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3088 * config/arm/arm-protos.h (tune_params): New struct members.
3089 * config/arm/arm.c: Initialise tune_params per processor.
3090 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
3091 for speed, based on new tune_params.
3093 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3095 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
3097 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
3099 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
3101 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
3102 * config/aarch64/arm_neon.h (vrnd_f64): Added.
3103 (vrnda_f64): Likewise.
3104 (vrndi_f64): Likewise.
3105 (vrndm_f64): Likewise.
3106 (vrndn_f64): Likewise.
3107 (vrndp_f64): Likewise.
3108 (vrndx_f64): Likewise.
3110 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3112 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
3113 GET_MODE_SIZE argument is enum machine_mode.
3115 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3118 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
3119 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
3121 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
3124 * asan.c (asan_emit_stack_protection): Force the base to align to
3125 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
3126 appropriate bits if STRICT_ALIGNMENT.
3127 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
3129 (expand_used_vars): Leave a space in the stack frame for alignment
3130 if STRICT_ALIGNMENT.
3132 2014-04-21 David Malcolm <dmalcolm@redhat.com>
3134 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
3136 (gimple_store_p): Likewise.
3137 (gimple_assign_load_p): Likewise.
3138 (gimple_assign_cast_p): Likewise.
3139 (gimple_clobber_p): Likewise.
3141 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
3142 rather than a gimple.
3143 (gimple_assign_cast_p): Likewise.
3145 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
3148 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
3149 If mode is DDmode and TARGET_E500_DOUBLE allow move.
3151 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
3152 more debug information for E500 if -mdebug=reg.
3154 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
3157 * config/i386/i386.c (ix86_expand_builtin)
3158 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
3159 register for target RTX.
3160 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
3162 2014-04-18 Cong Hou <congh@google.com>
3164 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
3165 the widen-mult pattern by handling two operands with different sizes,
3166 and operands whose size is smaller than half of the result type.
3168 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3170 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
3171 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
3172 (do_estimate_edge_time): Compute it.
3173 * ipa-inline.c (want_inline_small_function_p): Bypass
3174 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
3176 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3178 * ipa-inline.c (spec_rem): New static variable.
3179 (dump_overall_stats): New function.
3180 (dump_inline_stats): New function.
3182 2014-04-18 Richard Henderson <rth@redhat.com>
3184 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
3185 to GET_MODE_SIZE, not a reg_class_t.
3187 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3189 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
3190 (vsx_xxmrglw_<mode>): Likewise.
3192 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3195 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
3196 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
3197 (rs6000_init_hard_regno_mode_ok): Likewise.
3199 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
3201 * ipa-inline.c (inline_small_functions): Account only non-cold
3203 * doc/invoke.texi (inline-unit-growth): Update documentation.
3205 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
3207 * config/rs6000/rs6000.md (addti3, subti3): New.
3209 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
3212 * config/i386/i386.c (ix86_expand_clear): Remove outdated
3213 comment. Check optimize_insn_for_size_p instead of
3214 optimize_insn_for_speed_p.
3216 2014-04-17 Martin Jambor <mjambor@suse.cz>
3218 * gimple-iterator.c (gsi_start_edge): New function.
3219 * gimple-iterator.h (gsi_start_edge): Declare.
3220 * tree-sra.c (single_non_eh_succ): New function.
3221 (disqualify_ops_if_throwing_stmt): Renamed to
3222 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
3223 having one non-EH successor BB.
3224 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
3225 generate loads into replacements.
3226 (sra_modify_assign): Likewise and and also use the simple path for
3228 (sra_modify_function_body): Commit statements on edges.
3230 2014-04-17 Richard Biener <rguenther@suse.de>
3233 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
3234 comparison results and add clarifying comment.
3236 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3238 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
3239 (blank_mode): Initialize it.
3240 (emit_mode_size_inline, emit_mode_nunits_inline,
3241 emit_mode_inner_inline): New functions.
3242 (emit_insn_modes_h): Call them and surround their output with
3243 #if GCC_VERSION >= 4001 ... #endif.
3244 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
3245 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
3246 mode_* arrays if the argument is __builtin_constant_p.
3247 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
3248 is enum machine_mode.
3250 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3252 * passes.c (opt_pass::execute): Adjust.
3253 (pass_manager::execute_pass_mode_switching): Likewise.
3254 (early_local_passes::execute): Likewise.
3255 (execute_one_pass): Pass cfun to the pass's execute method.
3256 * tree-pass.h (opt_pass::execute): Add function * argument.
3257 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3258 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3259 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3260 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3261 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3262 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
3263 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
3264 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3265 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3266 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3267 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3268 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3269 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3270 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3271 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3272 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3273 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3274 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3275 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3276 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3277 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3278 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3279 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3280 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3281 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3282 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3283 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3284 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3285 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3288 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3290 * passes.c (opt_pass::gate): Take function * argument.
3291 (gate_all_early_local_passes): Merge into
3292 (early_local_passes::gate): this.
3293 (gate_all_early_optimizations): Merge into
3294 (all_early_optimizations::gate): this.
3295 (gate_all_optimizations): Mege into
3296 (all_optimizations::gate): this.
3297 (gate_all_optimizations_g): Merge into
3298 (all_optimizations_g::gate): this.
3299 (gate_rest_of_compilation): Mege into
3300 (rest_of_compilation::gate): this.
3301 (gate_postreload): Merge into
3302 (postreload::gate): this.
3303 (dump_one_pass): Pass cfun to the pass's gate method.
3304 (execute_ipa_summary_passes): Likewise.
3305 (execute_one_pass): Likewise.
3306 (ipa_write_summaries_2): Likewise.
3307 (ipa_write_optimization_summaries_1): Likewise.
3308 (ipa_read_summaries_1): Likewise.
3309 (ipa_read_optimization_summaries_1): Likewise.
3310 (execute_ipa_stmt_fixups): Likewise.
3311 * tree-pass.h (opt_pass::gate): Add function * argument.
3312 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
3313 combine-stack-adj.c, combine.c, compare-elim.c,
3314 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3315 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
3316 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
3317 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
3318 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3319 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3320 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3321 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
3322 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
3323 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
3324 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3325 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3326 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
3327 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3328 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3329 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3330 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3331 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3332 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3333 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3334 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3335 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3336 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3337 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
3338 var-tracking.c, vtable-verify.c, web.c: Adjust.
3340 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3342 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
3343 * configure: Regenerate.
3345 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3347 * passes.c (dump_one_pass): don't check pass->has_gate.
3348 (execute_ipa_summary_passes): Likewise.
3349 (execute_one_pass): Likewise.
3350 (ipa_write_summaries_2): Likewise.
3351 (ipa_write_optimization_summaries_1): Likewise.
3352 (ipa_read_optimization_summaries_1): Likewise.
3353 (execute_ipa_stmt_fixups): Likewise.
3354 * tree-pass.h (pass_data::has_gate): Remove.
3355 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3356 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3357 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3358 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3359 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3360 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
3361 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
3362 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
3363 gimple-low.c, gimple-ssa-isolate-paths.c,
3364 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
3365 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3366 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3367 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
3368 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
3369 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
3370 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
3371 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3372 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
3373 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3374 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3375 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3376 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3377 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3378 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3379 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3380 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3381 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3382 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3383 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3384 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3385 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3388 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3390 * pass_manager.h (pass_manager::register_dump_files_1): Remove
3392 * passes.c (pass_manager::register_dump_files_1): Merge into
3393 (pass_manager::register_dump_files): this, and remove its handling of
3394 properties since the pass always has the properties anyway.
3395 (pass_manager::pass_manager): Adjust.
3397 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3399 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
3400 * passes.c (pass_manager::register_dump_files_1): Remove dead code
3401 dealing with properties.
3402 (pass_manager::register_dump_files): Adjust.
3404 2014-03-20 Mark Wielaard <mjw@redhat.com>
3406 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
3407 then represent the bound as normal constant value.
3409 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3412 Forward port from 4.8 branch
3413 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
3415 * config/i386/bmiintrin.h (_blsi_u32): New.
3419 (_blsmsk_u32): Ditto.
3420 (_blsmsk_u64): Ditto.
3421 (_tzcnt_u32): Ditto.
3422 (_tzcnt_u64): Ditto.
3424 2014-04-17 Kito Cheng <kito@0xlab.org>
3426 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
3428 2014-04-17 Richard Biener <rguenther@suse.de>
3431 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
3432 boolean results for comparisons.
3434 2014-04-17 Richard Biener <rguenther@suse.de>
3436 PR tree-optimization/60836
3437 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
3438 initial PHI args to be gimple values.
3440 2014-04-17 Richard Biener <rguenther@suse.de>
3442 PR tree-optimization/60841
3443 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
3444 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
3445 of stmts to SLP build.
3446 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
3447 (vect_analyze_slp): Likewise.
3448 (vect_analyze_slp_instance): Likewise.
3449 (vect_build_slp_tree): Limit overall SLP tree growth.
3450 * tree-vectorizer.h (vect_analyze_data_refs,
3451 vect_analyze_slp): Adjust prototypes.
3453 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3455 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
3458 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3460 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
3461 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
3462 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
3463 for TARGET_SLOW_PSHUFB
3465 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3467 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
3468 * config/i386/i386.c (intel_cost): Ditto.
3470 2014-04-17 Joey Ye <joey.ye@arm.com>
3472 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
3474 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3476 * opts.c (common_handle_option): Disable -fipa-reference coorectly
3479 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3481 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
3482 (type_all_derivations_known_p): New predicate.
3483 (type_all_ctors_visible_p): New predicate.
3484 (type_possibly_instantiated_p): New predicate.
3485 (get_odr_type): Compute all_derivations_known.
3486 (dump_odr_type): Dump the flag.
3487 (maybe_record_type): Cleanup.
3488 (record_target_from_binfo): Add bases_to_consider array;
3489 record bases for types w/o instances and skip CXX destructor.
3490 (possible_polymorphic_call_targets_1): Add bases_to_consider
3491 and consider_construction parameters; check if type may have instance.
3492 (get_polymorphic_call_info): Set maybe_in_construction to true
3493 when we know nothing.
3494 (record_targets_from_bases): Skip CXX destructors; they are
3495 never called for types in construction.
3496 (possible_polymorphic_call_targets): Do not record target when
3497 type may not have instance.
3499 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3502 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
3503 external aliases alive, too.
3505 2014-04-16 Andrew Pinski <apinski@cavium.com>
3507 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
3510 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3512 * final.c (compute_alignments): Do not apply loop alignment to a block
3513 falling through to the exit.
3515 2014-04-16 Catherine Moore <clm@codesourcery.com>
3517 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
3518 Adjust constraints for microMIPS store patterns.
3520 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3522 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
3524 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3526 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
3527 (append_use): Run at -O0.
3528 (append_vdef): Likewise.
3529 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
3530 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
3532 2014-04-16 Jakub Jelinek <jakub@redhat.com>
3534 PR tree-optimization/60844
3535 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
3536 (propagate_op_to_single_use, remove_visited_stmt_chain,
3537 linearize_expr, repropagate_negates, reassociate_bb): Use it
3538 instead of gsi_remove.
3540 2014-04-16 Martin Jambor <mjambor@suse.cz>
3542 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
3543 ipa_transforms_to_apply.
3544 (cgraph_function_versioning): Assert that old_node has empty
3545 ipa_transforms_to_apply.
3546 * trans-mem.c (ipa_tm_create_version): Likewise.
3547 * tree-inline.c (tree_function_versioning): Do not duplicate
3548 ipa_transforms_to_apply.
3550 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3553 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
3555 Pass necessary as flags on 64-bit Solaris/x86.
3556 Use lowercase relocs for x86_64-*-*.
3557 * configure: Regenerate.
3559 2014-04-15 Jan Hubicka <jh@suse.cz>
3561 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
3562 (maybe_record_node, likely_target_p): Use it.
3564 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3567 Revert following patch
3569 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3572 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3573 software floating point or no floating point registers, do not
3574 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3575 in GPRs that occurs after we tested for GPRs that would never be
3578 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3579 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3580 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3581 specifically allow DDmode, since that does not use the SPE SIMD
3584 2014-03-21 Mark Wielaard <mjw@redhat.com>
3586 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
3587 as unsigned or int depending on type and value used.
3589 2014-04-15 Richard Biener <rguenther@suse.de>
3591 PR rtl-optimization/56965
3592 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
3593 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
3595 * alias.c (true_dependence_1): Do not call
3596 nonoverlapping_component_refs_p.
3597 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
3598 nonoverlapping_component_refs_p.
3599 (indirect_refs_may_alias_p): Likewise.
3601 2014-04-15 Teresa Johnson <tejohnson@google.com>
3603 * cfg.c (dump_bb_info): Fix flags check.
3604 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
3606 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3608 PR rtl-optimization/60663
3609 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
3612 2014-04-15 Richard Biener <rguenther@suse.de>
3614 * lto-streamer.h (LTO_major_version): Bump to 4.
3616 2014-04-15 Richard Biener <rguenther@suse.de>
3618 * common.opt (lto_partition_model): New enum.
3619 (flto-partition=): Merge separate options with a single with argument,
3620 add -flto-partition=one support.
3621 * flag-types.h (enum lto_partition_model): Declare.
3622 * opts.c (finish_options): Remove duplicate -flto-partition=
3624 * lto-wrapper.c (run_gcc): Adjust.
3626 2014-04-15 Richard Biener <rguenther@suse.de>
3628 * alias.c (ncr_compar): New function.
3629 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
3631 2014-04-15 Richard Biener <rguenther@suse.de>
3633 * alias.c (record_component_aliases): Do not walk BINFOs.
3635 2014-04-15 Richard Biener <rguenther@suse.de>
3637 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
3638 Add struct function argument and adjust.
3639 (find_func_aliases_for_call): Likewise.
3640 (find_func_aliases): Likewise.
3641 (find_func_clobbers): Likewise.
3642 (intra_create_variable_infos): Likewise.
3643 (compute_points_to_sets): Likewise.
3644 (ipa_pta_execute): Adjust. Do not push/pop cfun.
3646 2014-04-15 Richard Biener <rguenther@suse.de>
3648 * tree.c (iterative_hash_expr): Use enum tree_code_class
3649 to store TREE_CODE_CLASS.
3650 (tree_block): Likewise.
3651 (tree_set_block): Likewise.
3652 * tree.h (fold_build_pointer_plus_loc): Use
3653 convert_to_ptrofftype_loc.
3655 2014-04-15 Jakub Jelinek <jakub@redhat.com>
3658 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
3661 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
3663 * cfgloop.h (struct loop): Move force_vectorize down.
3664 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
3665 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
3666 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
3667 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
3668 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
3669 * tree-core.h (enum annot_expr_kind): Add new kind values.
3670 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
3671 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
3673 * tree.def (ANNOTATE_EXPR): Tweak comment.
3675 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3677 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
3680 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
3682 * tree.h (TYPE_IDENTIFIER): Declare.
3683 * tree.c (subrange_type_for_debug_p): Use it.
3684 * godump.c (go_format_type): Likewise.
3685 * dwarf2out.c (is_cxx_auto, modified_type_die,
3686 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
3687 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
3689 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3692 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
3694 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
3696 * config/i386/i386.c (examine_argument): Return bool. Return true if
3697 parameter should be passed in memory.
3698 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
3699 (construct_container): Update calls to examine_argument.
3700 (function_arg_advance_64): Ditto.
3701 (return_in_memory_32): Merge with ix86_return_in_memory.
3702 (return_in_memory_64): Ditto.
3703 (return_in_memory_ms_64): Ditto.
3705 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3707 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
3708 * coverage.c (coverage_compute_profile_id): Handle externally visible
3711 2014-04-14 Martin Jambor <mjambor@suse.cz>
3713 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
3714 DECL_DISREGARD_INLINE_LIMITS functions.
3716 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3719 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
3721 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3724 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
3725 optimize_insn_for_speed_p instead of
3726 optimize_function_for_speed_p.
3728 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
3730 * doc/invoke.texi (free): Document AArch64.
3732 2014-04-14 Richard Biener <rguenther@suse.de>
3734 PR tree-optimization/60042
3735 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
3736 (insert_into_preds_of_block): Do not prevent PHI insertion
3737 for REFERENCE exprs here ...
3738 (eliminate_dom_walker::before_dom_children): ... but prevent
3739 their use here under similar conditions when applied to the
3740 IL after PRE optimizations.
3742 2014-04-14 Richard Biener <rguenther@suse.de>
3744 * passes.def: Move early points-to after early SRA.
3746 2014-04-14 Richard Biener <rguenther@suse.de>
3748 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
3749 check for which sign-changes we allow when forwarding
3750 a converted value into a switch.
3752 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3754 * stor-layout.c (place_field): Finalize non-constant offset for the
3757 2014-04-14 Richard Biener <rguenther@suse.de>
3759 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
3761 (expand_switch_using_bit_tests_p): Likewise.
3762 (process_switch): Compute and pass on speed_p based on the
3764 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
3765 optimize_bb_for_speed_p.
3767 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3769 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
3770 * function.h (struct function): Rename has_force_vect_loops into
3771 has_force_vectorize_loops.
3772 * lto-streamer-in.c (input_cfg): Adjust for renaming.
3773 (input_struct_function_base): Likewise.
3774 * lto-streamer-out.c (output_cfg): Likewise.
3775 (output_struct_function_base): Likewise.
3776 * omp-low.c (expand_omp_simd): Likewise.
3777 * tree-cfg.c (move_sese_region_to_fn): Likewise.
3778 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
3779 (version_loop_for_if_conversion): Likewise.
3780 (tree_if_conversion): Likewise.
3781 (main_tree_if_conversion): Likewise.
3782 (gate_tree_if_conversion): Likewise.
3783 * tree-inline.c (copy_loops): Likewise.
3784 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
3785 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
3786 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
3787 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
3788 * tree-vectorizer.c (vectorize_loops): Likewise.
3789 * tree-vectorizer.h (unlimited_cost_model): Likewise.
3791 2014-04-14 Richard Biener <rguenther@suse.de>
3794 * lto-streamer-out.c (wrap_refs): New function.
3795 (lto_output): Wrap symbol references in global initializes in
3796 type-preserving MEM_REFs.
3798 2014-04-14 Christian Bruel <christian.bruel@st.com>
3800 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
3802 2014-04-14 Christian Bruel <christian.bruel@st.com>
3804 * config/sh/sh.md (setmemqi): New expand pattern.
3805 * config/sh/sh.h (CLEAR_RATIO): Define.
3806 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
3807 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
3809 2014-04-14 Richard Biener <rguenther@suse.de>
3812 * fold-const.c (negate_expr_p): Don't negate directional rounding
3814 (fold_negate_expr): Likewise.
3816 2014-04-14 Richard Biener <rguenther@suse.de>
3818 PR tree-optimization/59817
3819 PR tree-optimization/60453
3820 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
3821 recursion to catch all CHRECs in the scalar evolution and restrict
3822 the predicate for the remains appropriately.
3824 2014-04-12 Catherine Moore <clm@codesourcery.com>
3826 * config/mips/constraints.md: Add new register constraint "kb".
3827 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
3828 (*movhi_internal): Likewise.
3829 (*movqi_internal): Likewise.
3830 * config/mips/mips.h (M16_STORE_REGS): New register class.
3831 (REG_CLASS_NAMES): Add M16_STORE_REGS.
3832 (REG_CLASS_CONTENTS): Likewise.
3833 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
3835 2014-04-11 Tobias Burnus <burnus@net-b.de>
3838 * doc/invoke.texi (-Wformat-signedness): Document it.
3839 (Wformat=2): Mention that this enables -Wformat-signedness.
3841 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3843 * common/config/epiphany/epiphany-common.c
3844 (epiphany_option_optimization_table): Enable section anchors by
3845 default at -O1 or higher.
3846 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
3847 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
3848 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
3849 carries no extra cost.
3850 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
3851 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
3852 * config/epiphany/predicates.md (memclob_operand): New predicate.
3853 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
3854 Use memclob_operand predicate and X constraint for operand 3.
3856 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3858 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
3859 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
3862 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3864 PR rtl-optimization/60651
3865 * mode-switching.c (optimize_mode_switching): Make sure to emit
3866 sets of a lower numbered entity before sets of a higher numbered
3867 entity to a mode of the same or lower priority.
3868 When creating a seginfo for a basic block that starts with a code
3869 label, move the insertion point past the code label.
3870 (new_seginfo): Document and enforce requirement that
3871 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
3872 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
3873 * doc/tm.texi: Regenerate.
3875 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
3878 * config/arc/arc.c (arc_save_restore): Fix assert typo.
3880 2013-04-11 Jakub Jelinek <jakub@redhat.com>
3882 * BASE-VER: Set to 4.10.0.
3884 2014-04-11 Tobias Burnus <burnus@net-b.de>
3887 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
3888 * doc/gcc.texi (Service): Update description in the @menu
3889 * doc/invoke.texi (Option Summary): Remove misplaced and
3892 2014-04-11 Steve Ellcey <sellcey@mips.com>
3893 Jakub Jelinek <jakub@redhat.com>
3896 * expr.c (convert_move): Use emit_store_flag_force instead of
3897 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
3900 2014-04-11 Richard Biener <rguenther@suse.de>
3903 * varasm.c (assemble_alias): Avoid endless error reporting
3904 recursion by setting TREE_ASM_WRITTEN.
3906 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3908 * config/s390/s390.md: Add a splitter for NOT rtx.
3910 2014-04-11 Jakub Jelinek <jakub@redhat.com>
3912 PR rtl-optimization/60663
3913 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
3915 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
3916 Jakub Jelinek <jakub@redhat.com>
3919 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
3920 flag from decl_node to node.
3922 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3925 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
3926 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
3927 ameliorating the cases where it can be.
3929 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
3932 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
3934 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3935 (loadsync_<mode>): Change mode.
3936 (load_quadpti, store_quadpti): New.
3937 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3938 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3939 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
3941 2014-04-09 Cong Hou <congh@google.com>
3944 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
3947 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3949 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
3950 instead of vnor to exploit possible fusion opportunity in the
3952 (altivec_expand_vec_perm_const_le): Likewise.
3954 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
3956 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3957 (loadsync_<mode>): Change mode.
3958 (load_quadpti, store_quadpti): New.
3959 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3960 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3962 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
3965 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
3966 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
3967 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
3969 2014-04-08 Richard Biener <rguenther@suse.de>
3972 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
3973 a 64bit widest int print double-int similar to on HWI64 hosts.
3975 2014-04-08 Richard Biener <rguenther@suse.de>
3977 PR tree-optimization/60785
3978 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
3979 default defs properly.
3981 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
3983 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
3984 (Weffc++): Likewise.
3986 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
3988 * ipa-devirt.c (maybe_record_node): When node is not recorded,
3989 set completep to false rather than true.
3991 2014-04-07 Douglas B Rupp <rupp@adacore.com>
3994 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
3995 ARM_TARGET2_DWARF_FORMAT.
3997 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
4000 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
4001 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
4004 2014-04-07 Richard Biener <rguenther@suse.de>
4006 PR tree-optimization/60766
4007 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
4008 (may_eliminate_iv): Convert cand_value_at result to desired type.
4010 2014-04-07 Jason Merrill <jason@redhat.com>
4013 * common.opt (-fno-gnu-unique): Add.
4014 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
4016 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4018 * haifa-sched.c: Fix outdated function reference and minor
4019 grammar errors in introductory comment.
4021 2014-04-07 Richard Biener <rguenther@suse.de>
4024 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
4026 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
4028 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
4031 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
4032 size accounting for thunks.
4033 (pa_asm_output_mi_thunk): Use final_start_function() and
4034 final_end_function() to output function start and end directives.
4036 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4038 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
4039 device specific ISA/ feature information. Remove short_sp and
4040 errata_skip ds. Add avr_device_specific_features enum to have device
4042 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
4043 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
4044 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
4045 updated device specific info.
4046 * config/avr/avr-mcus.def: Merge device specific details to
4047 dev_attribute field.
4048 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
4050 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
4051 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
4052 assembler if RMW isa supported by current device.
4053 * config/avr/genmultilib.awk: Update as device info structure changed.
4054 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
4056 2014-04-04 Cong Hou <congh@google.com>
4058 PR tree-optimization/60656
4059 * tree-vect-stmts.c (supportable_widening_operation):
4060 Fix a bug that elements in a vector with vect_used_by_reduction
4061 property are incorrectly reordered when the operation on it is not
4062 consistant with the one in reduction operation.
4064 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
4066 PR rtl-optimization/60155
4067 * gcse.c (record_set_data): New function.
4068 (single_set_gcse): New function.
4069 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
4070 (hoist_code): Likewise.
4071 (get_pressure_class_and_nregs): Likewise.
4073 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
4075 * explow.c (probe_stack_range): Emit a final optimization blockage.
4077 2014-04-04 Anthony Green <green@moxielogic.com>
4079 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
4082 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
4085 * lto-cgraph.c (input_overwrite_node): Check that partitioning
4086 flags are set only during streaming.
4087 * ipa.c (process_references, walk_polymorphic_call_targets,
4088 symtab_remove_unreachable_nodes): Drop bodies of always inline
4089 after early inlining.
4090 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
4092 2014-04-04 Jakub Jelinek <jakub@redhat.com>
4093 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4096 * dwarf2out.c (const_ok_for_output_1): Reject expressions
4099 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4102 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
4104 (cortex_a53_fdivd): Likewise.
4106 2014-04-04 Martin Jambor <mjambor@suse.cz>
4109 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
4111 * cgraph.c (clone_of_p): Also return true if thunks match.
4112 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
4113 cgraph_function_or_thunk_node and an obsolete comment.
4114 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
4116 (build_function_decl_skip_args): Likewise.
4117 (set_new_clone_decl_and_node_flags): New function.
4118 (duplicate_thunk_for_node): Likewise.
4119 (redirect_edge_duplicating_thunks): Likewise.
4120 (cgraph_clone_node): New parameter args_to_skip, pass it to
4121 redirect_edge_duplicating_thunks which is called instead of
4122 cgraph_redirect_edge_callee.
4123 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
4124 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
4126 2014-04-04 Jeff Law <law@redhat.com>
4129 * config/arm/predicates.md (const_int_I_operand): New predicate.
4130 (const_int_M_operand): Similarly.
4131 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
4133 (insv_t2, extv_reg, extzv_t2): Likewise.
4134 (load_multiple_with_writeback): Similarly for const_int_I_operand.
4135 (pop_multiple_with_writeback_and_return): Likewise.
4136 (vfp_pop_multiple_with_writeback): Likewise
4138 2014-04-04 Richard Biener <rguenther@suse.de>
4141 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
4142 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
4144 * gimplify.h (gimple_add_tmp_var_fn): Declare.
4145 * gimplify.c (gimple_add_tmp_var_fn): New function.
4146 * gimple-expr.h (create_tmp_reg_fn): Declare.
4147 * gimple-expr.c (create_tmp_reg_fn): New function.
4148 * gimple-low.c (record_vars_into): Don't change cfun.
4149 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
4150 code generation without cfun.
4152 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
4155 * Makefile.in (install-driver): Fix shell scripting.
4157 2014-04-03 Cong Hou <congh@google.com>
4159 PR tree-optimization/60505
4160 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
4161 threshold of number of iterations below which no vectorization
4163 * tree-vect-loop.c (new_loop_vec_info):
4164 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
4165 * tree-vect-loop.c (vect_analyze_loop_operations):
4166 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
4167 * tree-vect-loop.c (vect_transform_loop):
4168 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
4169 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
4170 of iterations of the loop and see if we should build the epilogue.
4172 2014-04-03 Richard Biener <rguenther@suse.de>
4174 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
4175 (streamer_tree_cache_create): Adjust.
4176 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
4177 to allow optional nodes array.
4178 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
4179 (streamer_tree_cache_append): Likewise.
4180 (streamer_tree_cache_create): Create nodes array optionally
4181 as specified by parameter.
4182 * lto-streamer-out.c (create_output_block): Avoid maintaining
4183 the node array in the writer cache.
4184 (DFS_write_tree): Remove assertion.
4185 (produce_asm_for_decls): Free the out decl state hash table early.
4186 * lto-streamer-in.c (lto_data_in_create): Adjust for
4187 streamer_tree_cache_create prototype change.
4189 2014-04-03 Richard Biener <rguenther@suse.de>
4191 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
4192 set TREE_CHAIN to NULL_TREE.
4194 2014-04-03 Richard Biener <rguenther@suse.de>
4196 PR tree-optimization/60740
4197 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
4198 over all GIMPLE_COND operands.
4200 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
4202 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
4203 (Weffc++): Remove Scott's numbering, merge lists and reference
4206 2014-04-03 Nick Clifton <nickc@redhat.com>
4208 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
4211 2014-04-03 Martin Jambor <mjambor@suse.cz>
4213 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
4214 mention gcc_unreachable before failing.
4215 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
4218 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
4221 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
4222 inconsistent code and instead mark the context inconsistent.
4223 (possible_polymorphic_call_targets): For inconsistent contexts
4224 return empty complete list.
4226 2014-04-02 Anthony Green <green@moxielogic.com>
4228 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
4229 (extendqisi2, extendhisi2): Define.
4230 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
4231 (WCHAR_TYPE): Change to unsigned int.
4233 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4235 PR tree-optimization/60733
4236 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
4237 insertion point for PHI candidates to be the end of the feeding
4238 block for the PHI argument.
4240 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
4242 PR rtl-optimization/60650
4243 * lra-constraints.c (process_alt_operands): Decrease reject for
4244 earlyclobber matching.
4246 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4248 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
4250 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4252 * config/spu/spu.c (pad_bb): Do not crash when the last
4253 insn is CODE_FOR_blockage.
4255 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4257 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
4258 lies outside the target mode.
4260 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4263 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4264 software floating point or no floating point registers, do not
4265 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4266 in GPRs that occurs after we tested for GPRs that would never be
4269 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4270 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4271 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4272 specifically allow DDmode, since that does not use the SPE SIMD
4275 2014-04-02 Richard Biener <rguenther@suse.de>
4278 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4279 MODE_INTs. Properly use negv_optab.
4280 (expand_abs): Likewise.
4282 2014-04-02 Richard Biener <rguenther@suse.de>
4285 * Makefile.in (install-driver): Guard extra installs with special
4288 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4290 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4293 2014-04-01 Richard Henderson <rth@redhat.com>
4296 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
4297 alternative enabled before register allocation.
4299 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
4301 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
4302 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
4304 (nios2_large_got_address): Remove unneeded 'sym' parameter.
4305 (nios2_got_address): Update nios2_large_got_address call site.
4306 (nios2_delegitimize_address): New function.
4307 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
4308 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
4309 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
4311 2014-04-01 Martin Husemann <martin@duskware.de>
4313 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
4316 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
4318 PR rtl-optimization/60604
4319 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
4320 check from register_operand.
4321 (register_operand): Redefine in terms of general_operand.
4322 (nonmemory_operand): Use register_operand for the non-constant cases.
4324 2014-04-01 Richard Biener <rguenther@suse.de>
4326 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
4328 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
4330 * doc/invoke.texi (mapp-regs): Clarify.
4332 2014-03-31 Ulrich Drepper <drepper@gmail.com>
4334 * config/i386/avx512fintrin.h (__v32hi): Define type.
4335 (__v64qi): Likewise.
4336 (_mm512_set1_epi8): Define.
4337 (_mm512_set1_epi16): Define.
4338 (_mm512_set4_epi32): Define.
4339 (_mm512_set4_epi64): Define.
4340 (_mm512_set4_pd): Define.
4341 (_mm512_set4_ps): Define.
4342 (_mm512_setr4_epi64): Define.
4343 (_mm512_setr4_epi32): Define.
4344 (_mm512_setr4_pd): Define.
4345 (_mm512_setr4_ps): Define.
4346 (_mm512_setzero_epi32): Define.
4348 2014-03-31 Martin Jambor <mjambor@suse.cz>
4351 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
4352 callsite_arguments_match_p. Updated all callers. Also check types of
4353 corresponding formal parameters and actual arguments.
4354 (not_all_callers_have_enough_arguments_p) Renamed to
4355 some_callers_have_mismatched_arguments_p.
4357 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
4359 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
4361 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
4364 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
4367 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
4369 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
4370 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
4372 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
4373 Use FMAMODE_NOVF512 mode iterator.
4374 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
4375 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
4376 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
4377 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
4379 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
4380 Use VF_128_256 mode iterator.
4381 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
4384 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4386 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
4387 static chain if needed.
4389 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4392 * lra-constraints.c (index_part_to_reg): New.
4393 (process_address): Use it.
4395 2014-03-27 Jeff Law <law@redhat.com>
4396 Jakub Jelinek <jakub@redhat.com>
4399 * expr.c (do_tablejump): Use simplify_gen_binary rather than
4400 gen_rtx_{PLUS,MULT} to build up the address expression.
4402 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
4403 creating non-canonical RTL.
4405 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4408 * ipa-inline.c (want_inline_small_function_p): Short circuit large
4409 functions; reorganize to make cheap checks first.
4410 (inline_small_functions): Do not estimate growth when dumping;
4412 * ipa-inline.h (inline_summary): Add min_size.
4413 (growth_likely_positive): New function.
4414 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
4415 (set_cond_stmt_execution_predicate): Cleanup.
4416 (estimate_edge_size_and_time): Compute min_size.
4417 (estimate_calls_size_and_time): Likewise.
4418 (estimate_node_size_and_time): Likewise.
4419 (inline_update_overall_summary): Update min_size.
4420 (do_estimate_edge_time): Likewise.
4421 (do_estimate_edge_size): Update.
4422 (do_estimate_edge_hints): Update.
4423 (growth_likely_positive): New function.
4425 2014-03-28 Jakub Jelinek <jakub@redhat.com>
4428 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
4429 also if addr has VOIDmode.
4431 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4433 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
4434 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
4436 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
4437 instructions as well as AdvancedSIMD loads.
4439 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4441 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
4442 Use crypto_aese type.
4443 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
4444 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
4445 crypto_aese, crypto_aesmc. Move to types.md.
4446 * config/arm/types.md (crypto_aes): Split into crypto_aese,
4448 * config/arm/iterators.md (crypto_type): Likewise.
4450 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4452 * cgraph.c: Include expr.h and tree-dfa.h.
4453 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
4456 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4459 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
4460 regs from checking multi-reg pseudos.
4462 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4464 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
4466 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4468 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
4469 if it would clobber the stack pointer, even temporarily.
4471 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
4473 * mode-switching.c: Make small adjustments to the top comment.
4475 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4477 * config/rs6000/constraints.md (wD constraint): New constraint to
4478 match the constant integer to get the top DImode/DFmode out of a
4479 vector in a VSX register.
4481 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
4482 match the constant integer to get the top DImode/DFmode out of a
4483 vector in a VSX register.
4485 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
4488 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
4491 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
4492 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
4494 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
4495 Optimize vec_extract of 64-bit values, where the value being
4496 extracted is in the top word, where we can use scalar
4497 instructions. Add direct move and store support. Combine the big
4498 endian/little endian vector select load support into a single insn.
4499 (vsx_extract_<mode>_internal1): Likewise.
4500 (vsx_extract_<mode>_internal2): Likewise.
4501 (vsx_extract_<mode>_load): Likewise.
4502 (vsx_extract_<mode>_store): Likewise.
4503 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
4504 combined into vsx_extract_<mode>_load.
4505 (vsx_extract_<mode>_one_le): Likewise.
4507 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
4508 define the top 64-bit vector element.
4510 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
4513 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4514 Document vec_vbpermq builtin.
4517 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
4518 enable use of xxsldwi and xxpermdi builtin functions.
4519 (vec_xxpermdi): Likewise.
4521 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4522 Document use of vec_xxsldwi and vec_xxpermdi builtins.
4524 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
4526 PR rtl-optimization/60650
4527 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
4529 (find_spills_for): New.
4530 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
4531 Spill all pseudos on the second iteration.
4533 2014-03-27 Marek Polacek <polacek@redhat.com>
4536 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
4539 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4541 * config/s390/s390.c (s390_can_use_return_insn): Check for
4542 call-saved FPRs on 31 bit.
4544 2014-03-27 Jakub Jelinek <jakub@redhat.com>
4547 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
4548 if they need regimplification, just drop them instead of
4549 calling gimple_regimplify_operands on them.
4551 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
4554 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
4555 (aarch64_frame_pointer_required): Adjust logic.
4556 (aarch64_can_eliminate): Adjust logic.
4557 (aarch64_override_options_after_change): Adjust logic.
4559 2014-03-27 Dehao Chen <dehao@google.com>
4561 * ipa-inline.c (early_inliner): Update node's inline info.
4563 2014-03-26 Dehao Chen <dehao@google.com>
4565 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
4566 compiler inserted conditional jumps for NAN float check.
4568 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4570 * ubsan.h (ubsan_create_data): Change second argument's type
4571 to const location_t *.
4572 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
4574 (ubsan_create_data): Change second argument to const location_t *PLOC.
4575 Create Loc field whenever PLOC is non-NULL.
4576 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
4577 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
4581 * real.c (real_to_integer2): Change type of low to UHWI.
4583 2014-03-26 Tobias Burnus <burnus@net-b.de>
4585 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
4586 (CILK_SELF_SPECS): New define.
4587 (driver_self_specs): Use it.
4589 2014-03-26 Richard Biener <rguenther@suse.de>
4591 * tree-pretty-print.c (percent_K_format): Implement special
4592 case for LTO and its stripped down BLOCK tree.
4594 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4597 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
4599 * tree-vrp.c (simplify_internal_call_using_ranges): If only
4600 one range is range_int_cst_p, but not both, at least optimize
4601 addition/subtraction of 0 and multiplication by 0 or 1.
4602 * gimple-fold.c (gimple_fold_call): Fold
4603 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
4604 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
4605 INTEGER_CSTs, try to fold at least x * 0 and y - y.
4607 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
4609 PR rtl-optimization/60452
4610 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
4611 <case REG>: Return 1 for invalid offsets from the frame pointer.
4613 2014-03-26 Marek Polacek <polacek@redhat.com>
4616 * doc/extend.texi (C Extensions): Mention variable-length arrays in
4619 2014-03-26 Marek Polacek <polacek@redhat.com>
4622 * doc/extend.texi (Designated Inits): Describe what happens to omitted
4625 2014-03-26 Marek Polacek <polacek@redhat.com>
4628 * ira-color.c (update_conflict_hard_regno_costs): Perform the
4629 multiplication in unsigned type.
4631 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4633 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
4635 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4637 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
4639 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4642 * cif-code.def (UNREACHABLE) New code.
4643 * ipa-inline.c (inline_small_functions): Skip edges to
4644 __builtlin_unreachable.
4645 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
4646 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
4647 predicate to __bulitin_unreachable.
4648 (set_cond_stmt_execution_predicate): Fix issue when
4649 invert_tree_comparison returns ERROR_MARK.
4650 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
4651 propagate to inline clones.
4652 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
4654 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
4655 * cgraphclones.c (cgraph_clone_node): If call destination is already
4656 ureachable, do not redirect it back.
4657 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
4660 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4662 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
4663 Do not modify inline clones.
4665 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4667 * config/i386/i386.md (general_sext_operand): New mode attr.
4668 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
4669 don't generate (sign_extend (const_int)).
4670 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
4671 operands[2]. Use We constraint instead of <i> and
4672 <general_sext_operand> predicate instead of <general_operand>.
4673 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
4674 * config/i386/constraints.md (We): New constraint.
4675 * config/i386/predicates.md (x86_64_sext_operand,
4676 sext_operand): New predicates.
4678 2014-03-25 Martin Jambor <mjambor@suse.cz>
4681 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
4682 inconsistent devirtualizations to __builtin_unreachable.
4684 2014-03-25 Marek Polacek <polacek@redhat.com>
4687 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
4689 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
4691 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
4692 order of elements for big-endian.
4694 2014-03-25 Richard Biener <rguenther@suse.de>
4697 * gimplify-me.c (gimple_regimplify_operands): Update the
4700 2014-03-25 Martin Jambor <mjambor@suse.cz>
4703 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
4704 (lto_output_varpool_node): Likewise.
4705 (input_overwrite_node): Likewise.
4706 (input_varpool_node): Likewise.
4708 2014-03-25 Richard Biener <rguenther@suse.de>
4710 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
4711 (run_gcc): Likewise.
4713 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4715 * combine.c (simplify_compare_const): Add MODE argument.
4716 Handle mode_width 0 as very large mode_width.
4717 (try_combine, simplify_comparison): Adjust callers.
4719 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
4720 type to avoid signed integer overflow.
4721 * explow.c (plus_constant): Likewise.
4723 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
4725 * doc/generic.texi: Correct typos.
4727 2014-03-24 Tobias Burnus <burnus@net-b.de>
4729 * doc/invoke.texi (-flto): Expand section about
4730 using static libraries with LTO.
4732 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4734 PR rtl-optimization/60501
4735 * optabs.def (addptr3_optab): New optab.
4736 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
4737 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
4738 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
4740 * lra.c (emit_add3_insn): Use the addptr pattern if available.
4742 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
4744 2014-03-24 Ulrich Drepper <drepper@gmail.com>
4746 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
4749 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
4750 (_mm256_undefined_ps): Define.
4751 (_mm256_undefined_pd): Define.
4752 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
4753 (_mm_undefined_pd): Define.
4754 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
4755 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
4756 (_mm512_undefined_ps): Define.
4757 (_mm512_undefined_pd): Define.
4758 Use _mm*_undefined_*.
4759 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
4761 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
4763 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
4764 (lshr_simd): DI mode added.
4765 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
4766 (aarch64_ushr_simddi): Likewise.
4767 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
4768 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
4769 (vshrd_n_u64): Likewise.
4771 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4773 * Makefile.in (s-macro_list): Depend on cc1.
4775 2014-03-23 Teresa Johnson <tejohnson@google.com>
4777 * ipa-utils.c (ipa_print_order): Use specified dump file.
4779 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
4781 PR rtl-optimization/60601
4782 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
4784 * gcc.c (eval_spec_function): Initialize save_growing_value.
4786 2014-03-22 Jakub Jelinek <jakub@redhat.com>
4789 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
4790 code == MINUS_EXPR, never swap op0 with op1.
4792 * toplev.c (init_local_tick): Avoid signed integer multiplication
4794 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
4795 shift by first operand's bitsize.
4797 2014-03-21 Jakub Jelinek <jakub@redhat.com>
4800 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
4802 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
4803 TARGET_ISA_64BIT_P(x).
4805 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4807 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
4808 pattern for vector nor instead of subtract from splat(-1).
4809 (altivec_expand_vec_perm_const_le): Likewise.
4811 2014-03-21 Richard Henderson <rth@twiddle.net>
4814 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
4815 related insns after epilogue_completed.
4817 2014-03-21 Martin Jambor <mjambor@suse.cz>
4820 * cgraph.h (symtab_node): New flag body_removed.
4821 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
4822 when removing bodies.
4823 * symtab.c (dump_symtab_base): Dump body_removed flag.
4824 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
4825 had their bodies removed.
4827 2014-03-21 Martin Jambor <mjambor@suse.cz>
4830 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
4833 2014-03-21 Richard Biener <rguenther@suse.de>
4835 PR tree-optimization/60577
4836 * tree-core.h (struct tree_base): Document nothrow_flag use
4837 in VAR_DECL_NONALIASED.
4838 * tree.h (VAR_DECL_NONALIASED): New.
4839 (may_be_aliased): Adjust.
4840 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
4842 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4844 * expr.c (expand_expr_real_1): Remove outdated comment.
4846 2014-03-20 Jakub Jelinek <jakub@redhat.com>
4849 * ira.c (adjust_cleared_regs): Call copy_rtx on
4850 *reg_equiv[REGNO (loc)].src_p before passing it to
4851 simplify_replace_fn_rtx.
4854 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
4855 into CONST, put pic register as first operand of PLUS. Use
4856 gen_const_mem for both 32-bit and 64-bit PIC got loads.
4858 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4860 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
4862 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4864 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
4865 around for store forwarding issue in the FPU on the UT699.
4866 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
4867 loads and operations if -mfix-ut699 is specified.
4868 (divtf3_hq): Tweak attribute.
4869 (sqrttf2_hq): Likewise.
4871 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4873 * calls.c (store_one_arg): Remove incorrect const qualification on the
4874 type of the temporary.
4875 * cfgexpand.c (expand_return): Likewise.
4876 * expr.c (expand_constructor): Likewise.
4877 (expand_expr_real_1): Likewise.
4879 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4881 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
4884 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
4887 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
4889 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
4891 * config/arm/aarch-common-protos.h
4892 (alu_cost_table): Fix spelling of "extend".
4893 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
4895 2014-03-19 Richard Biener <rguenther@suse.de>
4898 * tree-core.h (tree_type_common): Re-order pointer members
4899 to reduce recursion depth during GC walks.
4901 2014-03-19 Marek Polacek <polacek@redhat.com>
4904 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
4905 before accessing it.
4907 2014-03-19 Richard Biener <rguenther@suse.de>
4910 * lto-streamer-in.c (input_function): In WPA stage do not drop
4913 2014-03-19 Jakub Jelinek <jakub@redhat.com>
4915 PR tree-optimization/60559
4916 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
4917 with build_zero_cst assignment.
4919 2014-03-18 Kai Tietz <ktietz@redhat.com>
4921 PR rtl-optimization/56356
4922 * sdbout.c (sdbout_parms): Verify that parms'
4923 incoming argument is valid.
4924 (sdbout_reg_parms): Likewise.
4926 2014-03-18 Richard Henderson <rth@redhat.com>
4929 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
4930 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
4931 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
4933 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
4935 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
4936 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
4937 Italicize plugin event names in description. Explain that
4938 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
4939 Remind that no GCC functions should be called after PLUGIN_FINISH.
4940 Explain what pragmas with expansion are.
4942 2014-03-18 Martin Liska <mliska@suse.cz>
4944 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
4945 gimple call statement is update.
4946 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
4947 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
4949 2014-03-18 Jakub Jelinek <jakub@redhat.com>
4952 * ubsan.c (ubsan_instrument_unreachable): Call
4953 initialize_sanitizer_builtins.
4954 (ubsan_pass): Likewise.
4957 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
4958 varpool_finalize_decl instead of rest_of_decl_compilation.
4960 2014-03-18 Richard Biener <rguenther@suse.de>
4962 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
4963 by using bitmap_and_compl instead of bitmap_and_compl_into.
4964 (df_rd_transfer_function): Likewise.
4966 2014-03-18 Richard Biener <rguenther@suse.de>
4968 * doc/lto.texi (fresolution): Fix typo.
4970 2014-03-18 Richard Biener <rguenther@suse.de>
4972 * doc/invoke.texi (flto): Update for changes in 4.9.
4974 2014-03-18 Richard Biener <rguenther@suse.de>
4976 * doc/loop.texi: Remove section on the removed lambda framework.
4977 Update loop docs with recent changes in preserving loop structure.
4979 2014-03-18 Richard Biener <rguenther@suse.de>
4981 * doc/lto.texi (-fresolution): Document.
4983 2014-03-18 Richard Biener <rguenther@suse.de>
4985 * doc/contrib.texi: Adjust my name.
4987 2014-03-18 Jakub Jelinek <jakub@redhat.com>
4990 * internal-fn.c: Include diagnostic-core.h.
4991 (expand_BUILTIN_EXPECT): New function.
4992 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
4993 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
4994 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
4995 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
4997 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
4998 Revert 3 argument __builtin_expect code.
4999 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
5000 * gimple-fold.c (gimple_fold_call): Likewise.
5001 * tree.h (fold_builtin_expect): New prototype.
5002 * builtins.c (build_builtin_expect_predicate): Add predictor
5003 argument, if non-NULL, create 3 argument __builtin_expect.
5004 (fold_builtin_expect): No longer static. Add ARG2 argument,
5005 pass it through to build_builtin_expect_predicate.
5006 (fold_builtin_2): Adjust caller.
5007 (fold_builtin_3): Handle BUILT_IN_EXPECT.
5008 * internal-fn.def (BUILTIN_EXPECT): New.
5010 2014-03-18 Tobias Burnus <burnus@net-b.de>
5013 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
5014 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
5015 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
5017 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
5020 * predict.c (combine_predictions_for_bb): Fix up formatting.
5021 (expr_expected_value_1, expr_expected_value): Add predictor argument,
5022 fill what it points to if non-NULL.
5023 (tree_predict_by_opcode): Adjust caller, use the predictor.
5024 * predict.def (PRED_COMPARE_AND_SWAP): Add.
5026 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
5028 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
5029 proper constant for the store mode.
5031 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
5033 * symtab.c (change_decl_assembler_name): Fix transparent alias
5036 2014-03-16 Renlin Li <Renlin.Li@arm.com>
5038 * config/aarch64/aarch64.c: Correct the comments about the
5039 aarch64 stack layout.
5041 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
5043 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
5044 check for GF_OMP_FOR_KIND_FOR.
5046 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
5048 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
5049 ymm and zmm register names.
5051 2014-03-17 Jakub Jelinek <jakub@redhat.com>
5054 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
5055 note creation for the 2010-08-31 changes.
5057 2014-03-17 Marek Polacek <polacek@redhat.com>
5060 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
5061 as -fno-tree-loop-vectorize.
5062 (expand_omp_simd): Likewise.
5064 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
5066 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
5067 (eligible_for_call_delay): New prototype.
5068 * config/sparc/sparc.c (tls_call_delay): Rename into...
5069 (eligible_for_call_delay): ...this. Return false if the instruction
5070 cannot be put in the delay slot of a branch.
5071 (eligible_for_restore_insn): Simplify.
5072 (eligible_for_return_delay): Return false if the instruction cannot be
5073 put in the delay slot of a branch and simplify.
5074 (eligible_for_sibcall_delay): Return false if the instruction cannot be
5075 put in the delay slot of a branch.
5076 * config/sparc/sparc.md (fix_ut699): New attribute.
5077 (tls_call_delay): Delete.
5078 (in_call_delay): Reimplement.
5079 (eligible_for_sibcall_delay): Rename into...
5080 (in_sibcall_delay): ...this.
5081 (eligible_for_return_delay): Rename into...
5082 (in_return_delay): ...this.
5083 (in_branch_delay): Reimplement.
5084 (in_uncond_branch_delay): Delete.
5085 (in_annul_branch_delay): Delete.
5087 2014-03-14 Richard Henderson <rth@redhat.com>
5090 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
5091 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
5092 (*floathi<X87MODEF>2_i387_with_temp): Remove.
5093 (floathi splitters): Remove.
5094 (float<SWI48x>xf2): New pattern.
5095 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
5096 code that tried to handle DImode for 32-bit, but which was excluded
5097 by the pattern's condition. Drop allocation of stack temporary.
5098 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
5099 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
5100 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
5101 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
5102 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
5103 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
5104 (*float<SWI48><MODEF>2_sse_interunit): Remove.
5105 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
5106 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
5107 (*float<SWI48x><X87MODEF>2_i387): Remove.
5108 (all float _with_temp splitters): Remove.
5109 (*float<SWI48x><MODEF>2_i387): New pattern.
5110 (*float<SWI48><MODEF>2_sse): New pattern.
5111 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
5112 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
5114 2014-03-14 Jakub Jelinek <jakub@redhat.com>
5115 Marek Polacek <polacek@redhat.com>
5118 * common.opt (dump_base_name_prefixed): New Variable.
5119 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
5120 if x_dump_base_name_prefixed is already set, set it at the end.
5122 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
5124 PR rtl-optimization/60508
5125 * lra-constraints.c (get_reload_reg): Add new parameter
5127 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
5128 Pass the new parameter values.
5130 2014-03-14 Richard Biener <rguenther@suse.de>
5132 * common.opt: Revert unintented changes from r205065.
5135 2014-03-14 Richard Biener <rguenther@suse.de>
5138 * cfghooks.c (split_block): Properly adjust all loops the
5139 block was a latch of.
5141 2014-03-14 Martin Jambor <mjambor@suse.cz>
5144 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
5147 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
5150 * config/avr/avr.c (avr_set_current_function): Pass function name
5151 through default_strip_name_encoding before sanity checking instead
5152 of skipping the first char of the assembler name.
5154 2014-03-13 Richard Henderson <rth@redhat.com>
5157 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
5158 (ix86_force_to_memory, ix86_free_from_memory): Remove.
5159 * config/i386/i386-protos.h: Likewise.
5160 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
5161 in the expander instead of a splitter.
5162 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
5163 any possibility of requiring a memory.
5164 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
5165 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
5166 (fp branch splitters): Update for ix86_split_fp_branch.
5167 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
5168 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
5169 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
5170 (*fop_<MODEF>_2_i387): Remove f/r alternative.
5171 (*fop_<MODEF>_3_i387): Likewise.
5172 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
5173 (splitters for the fop_* register patterns): Remove.
5174 (fscalexf4_i387): Rename from *fscalexf4_i387.
5175 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
5177 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5179 PR tree-optimization/59779
5180 * tree-dfa.c (get_ref_base_and_extent): Use double_int
5181 type for bitsize and maxsize instead of HOST_WIDE_INT.
5183 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
5185 PR rtl-optimization/57320
5186 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
5187 the CFG after thread_prologue_and_epilogue_insns.
5189 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
5191 PR rtl-optimization/57189
5192 * lra-constraints.c (process_alt_operands): Disfavor spilling
5195 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
5197 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
5199 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5201 PR tree-optimization/59025
5203 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
5204 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
5206 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
5209 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
5210 calls of avr_out_plus_1.
5212 2014-03-13 Bin Cheng <bin.cheng@arm.com>
5214 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
5215 BB's single pred and update the father loop's latch info later.
5217 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5219 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
5223 (VEC_base): Likewise.
5224 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
5225 registers, we need to swap double words in little endian mode.
5227 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
5228 to be a container mode for 128-bit integer operations added in ISA
5229 2.07. Unlike TImode and PTImode, the preferred register set is
5230 the Altivec/VMX registers for the 128-bit operations.
5232 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
5234 (rs6000_split_128bit_ok_p): Likewise.
5236 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
5237 macros for creating ISA 2.07 normal and overloaded builtin
5238 functions with 3 arguments.
5239 (BU_P8V_OVERLOAD_3): Likewise.
5240 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
5241 for use as overloaded functions.
5242 (VPERM_1TI_UNS): Likewise.
5243 (VSEL_1TI): Likewise.
5244 (VSEL_1TI_UNS): Likewise.
5245 (ST_INTERNAL_1ti): Likewise.
5246 (LD_INTERNAL_1ti): Likewise.
5247 (XXSEL_1TI): Likewise.
5248 (XXSEL_1TI_UNS): Likewise.
5249 (VPERM_1TI): Likewise.
5250 (VPERM_1TI_UNS): Likewise.
5251 (XXPERMDI_1TI): Likewise.
5252 (SET_1TI): Likewise.
5253 (LXVD2X_V1TI): Likewise.
5254 (STXVD2X_V1TI): Likewise.
5255 (VEC_INIT_V1TI): Likewise.
5256 (VEC_SET_V1TI): Likewise.
5257 (VEC_EXT_V1TI): Likewise.
5258 (EQV_V1TI): Likewise.
5259 (NAND_V1TI): Likewise.
5260 (ORC_V1TI): Likewise.
5261 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
5262 added in ISA 2.07. Add both normal 'altivec' builtins, and the
5264 (VADDUQM): Likewise.
5265 (VSUBCUQ): Likewise.
5266 (VADDEUQM): Likewise.
5267 (VADDECUQ): Likewise.
5268 (VSUBEUQM): Likewise.
5269 (VSUBECUQ): Likewise.
5271 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5272 __int128_t and __uint128_t types.
5273 (__uint128_type): Likewise.
5274 (altivec_categorize_keyword): Add support for vector __int128_t,
5275 vector __uint128_t, vector __int128, and vector unsigned __int128
5276 as a container type for TImode operations that need to be done in
5277 VSX/Altivec registers.
5278 (rs6000_macro_to_expand): Likewise.
5279 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5280 to support 128-bit integer instructions vaddcuq, vadduqm,
5281 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5282 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5284 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5285 for V1TImode, and set up preferences to use VSX/Altivec registers.
5286 Setup VSX reload handlers.
5287 (rs6000_debug_reg_global): Likewise.
5288 (rs6000_init_hard_regno_mode_ok): Likewise.
5289 (rs6000_preferred_simd_mode): Likewise.
5290 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5291 (easy_altivec_constant): Likewise.
5292 (output_vec_const_move): Likewise.
5293 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5295 (rs6000_expand_vector_extract): Likewise.
5296 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
5298 (rs6000_const_vec): Add support for V1TImode.
5299 (rs6000_emit_le_vsx_load): Swap double words when loading or
5300 storing TImode/V1TImode.
5301 (rs6000_emit_le_vsx_store): Likewise.
5302 (rs6000_emit_le_vsx_move): Likewise.
5303 (rs6000_emit_move): Add support for V1TImode.
5304 (altivec_expand_ld_builtin): Likewise.
5305 (altivec_expand_st_builtin): Likewise.
5306 (altivec_expand_vec_init_builtin): Likewise.
5307 (altivec_expand_builtin): Likewise.
5308 (rs6000_init_builtins): Add support for V1TImode type. Add
5309 support for ISA 2.07 128-bit integer builtins. Define type names
5310 for the VSX/Altivec vector types.
5311 (altivec_init_builtins): Add support for overloaded vector
5312 functions with V1TImode type.
5313 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
5314 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
5316 (rs6000_split_128bit_ok_p): Likewise.
5317 (rs6000_handle_altivec_attribute): Create V1TImode from vector
5318 __int128_t and vector __uint128_t.
5320 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
5321 and mode attributes.
5328 (VS_scalar): Likewise.
5329 (VS_double): Likewise.
5330 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
5332 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
5333 we support the ISA 2.07 128-bit integer arithmetic instructions.
5334 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
5335 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
5336 and TImode types for use with the builtin functions.
5337 (V1TI_type_node): Likewise.
5338 (unsigned_V1TI_type_node): Likewise.
5339 (intTI_type_internal_node): Likewise.
5340 (uintTI_type_internal_node): Likewise.
5342 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
5343 128-bit builtin functions.
5344 (UNSPEC_VADDEUQM): Likewise.
5345 (UNSPEC_VADDECUQ): Likewise.
5346 (UNSPEC_VSUBCUQ): Likewise.
5347 (UNSPEC_VSUBEUQM): Likewise.
5348 (UNSPEC_VSUBECUQ): Likewise.
5349 (VM): Add V1TImode to vector mode iterators.
5351 (VI_unit): Likewise.
5352 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
5353 (altivec_vaddcuq): Likewise.
5354 (altivec_vsubuqm): Likewise.
5355 (altivec_vsubcuq): Likewise.
5356 (altivec_vaddeuqm): Likewise.
5357 (altivec_vaddecuq): Likewise.
5358 (altivec_vsubeuqm): Likewise.
5359 (altivec_vsubecuq): Likewise.
5361 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
5363 (BOOL_128): Likewise.
5364 (BOOL_REGS_OUTPUT): Likewise.
5365 (BOOL_REGS_OP1): Likewise.
5366 (BOOL_REGS_OP2): Likewise.
5367 (BOOL_REGS_UNARY): Likewise.
5368 (BOOL_REGS_AND_CR0): Likewise.
5370 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
5371 128-bit integer builtin support.
5372 (vec_vadduqm): Likewise.
5373 (vec_vaddecuq): Likewise.
5374 (vec_vaddeuqm): Likewise.
5375 (vec_vsubecuq): Likewise.
5376 (vec_vsubeuqm): Likewise.
5377 (vec_vsubcuq): Likewise.
5378 (vec_vsubuqm): Likewise.
5380 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5381 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
5382 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
5383 128-bit integer add/subtract to ISA 2.07.
5385 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
5387 * config/arc/arc.c (arc_predicate_delay_insns):
5388 Fix third argument passed to conditionalize_nonjump.
5390 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
5392 * config/aarch64/aarch64-builtins.c
5393 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
5394 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
5395 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
5396 instead of __builtin_lfloor.
5397 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
5399 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5401 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
5402 (tree_ssa_ifcombine_bb_1): New function.
5403 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
5404 is an empty forwarder block to then_bb or vice versa and then_bb
5405 and else_bb are effectively swapped.
5407 2014-03-12 Christian Bruel <christian.bruel@st.com>
5410 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
5411 REG_CFA_DEF_CFA note.
5412 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
5413 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
5415 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
5417 PR tree-optimization/60454
5418 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
5420 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5422 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
5423 Do not define target_cpu_default2 to generic.
5424 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
5425 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
5426 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
5428 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5429 Marc Glisse <marc.glisse@inria.fr>
5431 PR tree-optimization/60502
5432 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
5433 instead of build_low_bits_mask.
5435 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5438 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
5439 if there are multiple uses, but op doesn't live on E edge.
5440 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
5441 clobber stmts before __builtin_unreachable.
5443 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
5445 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
5446 hard_frame_pointer_rtx.
5447 * cse.c (cse_insn): Remove volatile check.
5448 * cselib.c (cselib_process_insn): Likewise.
5449 * dse.c (scan_insn): Likewise.
5451 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5453 * config/arc/arc.c (conditionalize_nonjump): New function,
5455 (arc_ifcvt): ... this.
5456 (arc_predicate_delay_insns): Use it.
5458 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5460 * config/arc/predicates.md (extend_operand): During/after reload,
5461 allow const_int_operand.
5462 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
5463 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
5464 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
5466 (umulsi3_highpart_i): Likewise.
5468 2014-03-11 Richard Biener <rguenther@suse.de>
5470 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
5471 Add asserts to guard possible wrong-code bugs.
5473 2014-03-11 Richard Biener <rguenther@suse.de>
5475 PR tree-optimization/60429
5476 PR tree-optimization/60485
5477 * tree-ssa-structalias.c (set_union_with_increment): Properly
5478 take into account all fields that overlap the shifted vars.
5479 (do_sd_constraint): Likewise.
5480 (do_ds_constraint): Likewise.
5481 (get_constraint_for_ptr_offset): Likewise.
5483 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
5485 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
5486 (nios2_compute_frame_layout):
5487 Add calculation of cfun->machine->fp_save_offset.
5488 (nios2_expand_prologue): Correct setting of frame pointer register
5490 (nios2_expand_epilogue): Update recovery of stack pointer from
5491 frame pointer accordingly.
5492 (nios2_initial_elimination_offset): Update calculation of offset
5493 for eliminating to HARD_FRAME_POINTER_REGNUM.
5495 2014-03-10 Jakub Jelinek <jakub@redhat.com>
5498 * ipa.c (symtab_remove_unreachable_nodes): Don't call
5499 cgraph_get_create_node on VAR_DECLs.
5501 2014-03-10 Richard Biener <rguenther@suse.de>
5504 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
5506 2014-03-08 Douglas B Rupp <rupp@gnat.com>
5508 * config/vms/vms.opt (vms_float_format): New variable.
5510 2014-03-08 Tobias Burnus <burnus@net-b.de>
5512 * doc/invoke.texi (-fcilkplus): Update implementation status.
5514 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
5515 Richard Biener <rguenther@suse.de>
5517 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
5518 consistently accross all TUs.
5519 (run_gcc): Enable -fshort-double automatically at link at link-time
5520 and disallow override.
5522 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
5525 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
5526 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
5527 if they can't be used.
5529 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5531 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
5532 for Solaris 11/x86 ld.
5533 * configure: Regenerate.
5535 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5537 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
5538 (LIB_TLS_SPEC): Save as ld_tls_libs.
5539 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
5540 (HAVE_AS_IX86_TLSLDM): New test.
5541 * configure, config.in: Regenerate.
5542 * config/i386/i386.c (legitimize_tls_address): Fall back to
5543 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
5544 cannot support TLS_MODEL_LOCAL_DYNAMIC.
5545 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
5546 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
5548 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
5550 * common.opt (fira-loop-pressure): Mark as optimization.
5552 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
5554 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
5555 an OpenMP mappable type.
5557 2014-03-06 Matthias Klose <doko@ubuntu.com>
5559 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
5560 MULTILIB_OSDIRNAMES is not defined.
5562 2014-03-06 Jakub Jelinek <jakub@redhat.com>
5563 Meador Inge <meadori@codesourcery.com>
5566 * config/arm/arm.c (arm_tls_symbol_p): Remove.
5567 (arm_legitimize_address): Call legitimize_tls_address for any
5568 arm_tls_referenced_p expression, handle constant addend. Call it
5569 before testing for !TARGET_ARM.
5570 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
5572 2014-03-06 Richard Biener <rguenther@suse.de>
5578 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5580 * tree-streamer.c (record_common_node): Assert we don't record
5581 nodes with type double.
5582 (preload_common_node): Skip type double, complex double and double
5583 pointer since it is now frontend dependent due to fshort-double option.
5585 2014-03-06 Richard Biener <rguenther@suse.de>
5587 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
5588 or -fno-lto is specified and the linker has full plugin support.
5589 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
5590 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
5591 * lto-wrapper.c (merge_and_complain): Merge compile-time
5592 optimization levels.
5593 (run_gcc): And pass it through to the link options.
5595 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
5599 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5601 * cselib.c (remove_useless_values): Skip to avoid quadratic
5602 behavior if the condition moved from...
5603 (cselib_process_insn): ... here holds.
5605 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5608 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
5609 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
5612 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
5613 (TM_H): Add x86-tune.def.
5615 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5617 * config/aarch64/aarch64.c (generic_tunings):
5618 Use cortexa57_extra_costs.
5620 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5623 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
5624 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
5625 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
5628 2014-03-04 Heiher <r@hev.cc>
5630 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
5631 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
5633 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
5635 * config/i386/predicates.md (const2356_operand): Change to ...
5636 (const2367_operand): ... this.
5637 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
5639 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5640 (*avx512pf_scatterpf<mode>sf): Ditto.
5641 (avx512pf_scatterpf<mode>df): Ditto.
5642 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5643 (*avx512pf_scatterpf<mode>df): Ditto.
5644 * config/i386/i386.c (ix86_expand_builtin): Update
5645 incorrect hint operand error message.
5647 2014-03-04 Richard Biener <rguenther@suse.de>
5649 * lto-section-in.c (lto_get_section_data): Fix const cast.
5651 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5653 * tree-streamer.c (record_common_node): Assert we don't record
5654 nodes with type double.
5655 (preload_common_node): Skip type double, complex double and double
5656 pointer since it is now frontend dependent due to fshort-double option.
5658 2014-03-04 Richard Biener <rguenther@suse.de>
5661 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
5662 (lto_input_toplevel_asms): Likewise.
5663 * lto-section-in.c (lto_get_section_data): Instead do it here
5666 2014-03-04 Richard Biener <rguenther@suse.de>
5668 PR tree-optimization/60382
5669 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
5670 dead PHIs a reduction.
5672 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
5674 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
5676 (_mm_prefetch): Move out of GCC target("sse") pragma.
5677 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
5678 GCC target("prfchw") pragma.
5679 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
5681 * config/i386/i386.c (ix86_option_override_internal): Enable
5682 -mprfchw with -mprefetchwt1.
5684 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5686 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
5689 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5691 * opts.h (CL_PCH_IGNORE): Define.
5692 * targhooks.c (option_affects_pch_p):
5693 Return false for options that have CL_PCH_IGNORE set.
5694 * opt-functions.awk: Process PchIgnore.
5695 * doc/options.texi: Document PchIgnore.
5697 * config/arc/arc.opt (misize): Add PchIgnore property.
5699 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5701 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
5702 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
5703 constraint on constants to permit them being loaded into
5704 GENERAL_REGS or BASE_REGS.
5706 2014-03-03 Nick Clifton <nickc@redhat.com>
5708 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
5709 anti-cacnonical alternatives.
5710 (negandhi3_real): New pattern.
5711 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
5713 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5715 * config/avr/avr-mcus.def: Remove atxmega16x1.
5716 * config/avr/avr-tables.opt: Regenerate.
5717 * config/avr/t-multilib: Regenerate.
5718 * doc/avr-mmcu.texi: Regenerate.
5720 2014-03-03 Tobias Grosser <tobias@grosser.es>
5721 Mircea Namolaru <mircea.namolaru@inria.fr>
5723 PR tree-optimization/58028
5724 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
5727 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5729 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
5730 not handled by recognizers.
5732 2014-03-03 Jakub Jelinek <jakub@redhat.com>
5735 * function.c (expand_function_end): Don't emit
5736 clobber_return_register sequence if clobber_after is a BARRIER.
5737 * cfgexpand.c (construct_exit_block): Append instructions before
5738 return_label to prev_bb.
5740 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5742 * config/rs6000/constraints.md: Document reserved use of "wc".
5744 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5747 * ipa.c (function_and_variable_visibility): When dissolving comdat
5748 group, also set all symbols to local.
5750 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5755 2013-12-14 Jan Hubicka <jh@suse.cz>
5757 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
5759 2014-03-02 Jon Beniston <jon@beniston.com>
5765 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
5766 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
5767 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
5768 (simple_return, *simple_return): New patterns
5769 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
5770 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
5772 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
5774 * dwarf2out.c (gen_subprogram_die): Tidy.
5776 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
5779 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
5780 (*mov_t_msb_neg_negc): ... this new insn.
5782 2014-02-28 Jason Merrill <jason@redhat.com>
5785 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
5788 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
5791 * dwarf2out.c (decltype_auto_die): New static.
5792 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
5793 (gen_type_die_with_usage): Handle 'decltype(auto)'.
5794 (is_cxx_auto): Likewise.
5796 2014-02-28 Ian Bolton <ian.bolton@arm.com>
5798 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
5799 we are not using general regs only.
5801 2014-02-28 Richard Biener <rguenther@suse.de>
5804 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
5805 previous fix and only allow to remove trivial pre-headers
5806 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
5807 (remove_forwarder_block): Properly update the latch of a loop.
5809 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5812 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
5813 (cselib_preserved_hash_table): New.
5814 (preserve_constants_and_equivs): Move preserved vals to it.
5815 (cselib_find_slot): Look it up first.
5816 (cselib_init): Initialize it.
5817 (cselib_finish): Release it.
5818 (dump_cselib_table): Dump it.
5820 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5823 * cselib.c (remove_useless_values): Skip to avoid quadratic
5824 behavior if the condition moved from...
5825 (cselib_process_insn): ... here holds.
5827 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5830 * var-tracking.c (vt_initialize): Apply the same condition to
5831 preserve the CFA base value.
5833 2014-02-28 Joey Ye <joey.ye@arm.com>
5836 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
5837 if reload in progress or completed.
5839 2014-02-28 Tobias Burnus <burnus@net-b.de>
5842 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
5845 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
5847 * doc/tm.texi.in (Condition Code Status): Update documention for
5848 relative locations of cc0-setter and cc0-user.
5850 2014-02-27 Jeff Law <law@redhat.com>
5852 PR rtl-optimization/52714
5853 * combine.c (try_combine): When splitting an unrecognized PARALLEL
5854 into two independent simple sets, if I3 is a jump, ensure the
5855 pattern we place into I3 is a (set (pc) ...).
5857 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
5858 Jeff Law <law@redhat.com>
5860 PR rtl-optimization/49847
5861 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
5862 are in different blocks.
5863 * doc/tm.texi (Condition Code Status): Update documention for
5864 relative locations of cc0-setter and cc0-user.
5866 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
5869 * lra.c (lra_emit_add): Check SUBREG too.
5871 2014-02-27 Andreas Schwab <schwab@suse.de>
5873 * config/m68k/m68k.c (m68k_option_override): Disable
5874 -flive-range-shrinkage for classic m68k.
5875 (m68k_override_options_after_change): Likewise.
5877 2014-02-27 Marek Polacek <polacek@redhat.com>
5880 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
5881 -Wmaybe-uninitialized.
5883 2014-02-27 Alan Modra <amodra@gmail.com>
5886 * reload1.c (emit_input_reload_insns): When reload_override_in,
5887 set old to rl->in_reg when rl->in_reg is a subreg.
5889 2014-02-26 Richard Biener <rguenther@suse.de>
5892 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
5894 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
5896 * common/config/i386/predicates.md (const1256_operand): Remove.
5897 (const2356_operand): New.
5898 (const_1_to_2_operand): Remove.
5899 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5900 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5901 (*avx512pf_gatherpf<mode>sf): Ditto.
5902 (avx512pf_gatherpf<mode>df): Ditto.
5903 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5904 (*avx512pf_gatherpf<mode>df): Ditto.
5905 (avx512pf_scatterpf<mode>sf): Ditto.
5906 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5907 (*avx512pf_scatterpf<mode>sf): Ditto.
5908 (avx512pf_scatterpf<mode>df): Ditto.
5909 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5910 (*avx512pf_scatterpf<mode>df): Ditto.
5911 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
5913 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
5915 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
5916 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
5917 (_mm512_mask_testn_epi64_mask): Move to ...
5918 * config/i386/avx512cdintrin.h: Here.
5919 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
5920 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
5921 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
5922 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
5923 TARGET_AVX512F from TARGET_AVX512CD.
5925 2014-02-26 Richard Biener <rguenther@suse.de>
5928 * ipa.c (walk_polymorphic_call_targets): Properly guard
5929 call to inline_update_overall_summary.
5931 2014-02-26 Bin Cheng <bin.cheng@arm.com>
5934 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
5935 and latches only if requested. Fix latch if it is removed.
5936 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
5937 LOOPS_HAVE_PREHEADERS.
5939 2014-02-25 Andrew Pinski <apinski@cavium.com>
5941 * builtins.c (expand_builtin_thread_pointer): Create a new target
5942 when the target is NULL.
5944 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
5946 PR rtl-optimization/60317
5947 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
5948 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
5949 * lra-assigns.c: Include params.h.
5950 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
5951 other reload pseudos considerations.
5953 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5955 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
5956 to use canonical form for nor<mode>3.
5958 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5961 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
5964 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
5966 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
5967 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
5968 (ix86_handle_option): Handle OPT_mprefetchwt1.
5969 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
5970 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
5972 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
5973 OPTION_MASK_ISA_PREFETCHWT1.
5974 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
5975 (PTA_PREFETCHWT1): New.
5976 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
5977 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
5978 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
5979 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
5980 (*prefetch_avx512pf_<mode>_: Change into ...
5981 (*prefetch_prefetchwt1_<mode>: This.
5982 * config/i386/i386.opt (mprefetchwt1): New.
5983 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
5984 (_mm_prefetch): Handle intent to write.
5985 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
5987 2014-02-25 Richard Biener <rguenther@suse.de>
5990 * emit-rtl.c (mem_attrs_htab): Remove.
5991 (mem_attrs_htab_hash): Likewise.
5992 (mem_attrs_htab_eq): Likewise.
5993 (set_mem_attrs): Always allocate new mem-attrs when something changed.
5994 (init_emit_once): Do not allocate mem_attrs_htab.
5996 2014-02-25 Richard Biener <rguenther@suse.de>
5999 * lto-opts.c (lto_write_options): Output non-explicit conservative
6000 -fwrapv, -fno-trapv and -fno-strict-overflow.
6001 * lto-wrapper.c (merge_and_complain): Handle merging those options.
6002 (run_gcc): And pass them through.
6004 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6006 * sel-sched.c (calculate_new_fences): New parameter ptime.
6007 Calculate it as a maximum over all fence cycles.
6008 (sel_sched_region_2): Adjust the call to calculate_new_fences.
6009 Print the final schedule timing when sched_verbose.
6011 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6013 PR rtl-optimization/60292
6014 * sel-sched.c (fill_vec_av_set): Do not reset target availability
6015 bit fot the fence instruction.
6017 2014-02-24 Alangi Derick <alangiderick@gmail.com>
6019 * calls.h: Fix typo in comment.
6021 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
6023 * config/pa/pa.c (pa_output_move_double): Don't valididate when
6024 adjusting offsetable addresses.
6026 2014-02-24 Guozhi Wei <carrot@google.com>
6028 * sparseset.h (sparseset_pop): Fix the wrong index.
6030 2014-02-24 Walter Lee <walt@tilera.com>
6032 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
6033 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
6035 * common/config/tilegx/tilegx-common.c
6036 (TARGET_DEFAULT_TARGET_FLAGS): Define.
6037 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
6039 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
6040 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
6041 (tilegx_gimplify_va_arg_expr): Handle big endian.
6042 (tilegx_expand_unaligned_load): Ditto.
6043 (tilegx_expand_unaligned_store): Ditto.
6044 (TARGET_RETURN_IN_MSB): New.
6045 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
6046 (TARGET_ENDIAN_DEFAULT): New.
6047 (TARGET_BIG_ENDIAN): Handle big endian.
6048 (BYTES_BIG_ENDIAN): Ditto.
6049 (WORDS_BIG_ENDIAN): Ditto.
6050 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
6053 * config/tilegx/tilegx.md (extv): Handle big endian.
6055 (insn_st<n>): Ditto.
6056 (insn_st<n>_add<bitsuffix>): Ditto.
6057 (insn_stnt<n>): Ditto.
6058 (insn_stnt<n>_add<bitsuffix>):Ditto.
6059 (vec_interleave_highv8qi): Handle big endian.
6060 (vec_interleave_highv8qi_be): New.
6061 (vec_interleave_highv8qi_le): New.
6062 (insn_v1int_h): Handle big endian.
6063 (vec_interleave_lowv8qi): Handle big endian.
6064 (vec_interleave_lowv8qi_be): New.
6065 (vec_interleave_lowv8qi_le): New.
6066 (insn_v1int_l): Handle big endian.
6067 (vec_interleave_highv4hi): Handle big endian.
6068 (vec_interleave_highv4hi_be): New.
6069 (vec_interleave_highv4hi_le): New.
6070 (insn_v2int_h): Handle big endian.
6071 (vec_interleave_lowv4hi): Handle big endian.
6072 (vec_interleave_lowv4hi_be): New.
6073 (vec_interleave_lowv4hi_le): New.
6074 (insn_v2int_l): Handle big endian.
6075 (vec_interleave_highv2si): Handle big endian.
6076 (vec_interleave_highv2si_be): New.
6077 (vec_interleave_highv2si_le): New.
6078 (insn_v4int_h): Handle big endian.
6079 (vec_interleave_lowv2si): Handle big endian.
6080 (vec_interleave_lowv2si_be): New.
6081 (vec_interleave_lowv2si_le): New.
6082 (insn_v4int_l): Handle big endian.
6083 * config/tilegx/tilegx.opt (mbig-endian): New option.
6084 (mlittle-endian): New option.
6085 * doc/install.texi: Document tilegxbe-linux.
6086 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
6088 2014-02-24 Martin Jambor <mjambor@suse.cz>
6091 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
6092 there are no parameter descriptors.
6094 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
6096 PR rtl-optimization/60268
6097 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
6098 initialization to ...
6099 (sched_rgn_init): ... here.
6100 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
6102 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6104 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
6107 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
6109 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
6112 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6114 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
6115 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
6117 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6119 * config/microblaze/predicates.md: Add cmp_op predicate.
6120 * config/microblaze/microblaze.md: Add branch_compare instruction
6121 which uses cmp_op predicate and emits cmp insn before branch.
6122 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
6123 to microblaze_expand_conditional_branch and consolidate logic.
6124 (microblaze_expand_conditional_branch): emit branch_compare
6125 insn instead of handling cmp op separate from branch insn.
6127 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6129 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
6132 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6134 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
6135 define_insn with define_expand and new define_insn
6136 *altivec_lve<VI_char>x_internal.
6137 (altivec_stve<VI_char>x): Replace define_insn with define_expand
6138 and new define_insn *altivec_stve<VI_char>x_internal.
6139 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
6141 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
6143 (altivec_expand_stvex_be): New function.
6145 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
6147 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
6148 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
6149 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
6150 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
6152 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
6155 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
6156 instead of emit_move_insn.
6158 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6160 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
6162 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
6163 gen_altivec_vsumsws.
6165 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6167 * config/rs6000/altivec.md (altivec_lvxl): Rename as
6168 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
6169 (altivec_lvxl_<mode>): New define_expand incorporating
6170 -maltivec=be semantics where needed.
6171 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
6172 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
6173 semantics where needed.
6174 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
6175 (altivec_stvx_<mode>): New define_expand incorporating
6176 -maltivec=be semantics where needed.
6177 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
6178 VM2 iterator instead of V4SI.
6179 (altivec_stvxl_<mode>): New define_expand incorporating
6180 -maltivec=be semantics where needed.
6181 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
6182 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
6183 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
6184 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
6185 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
6186 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
6187 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
6188 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
6189 ALTIVEC_BUILTIN_STVXL.
6190 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
6191 (altivec_expand_stvx_be): Likewise.
6192 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
6193 (altivec_expand_lvx_be): Likewise.
6194 (altivec_expand_stvx_be): Likewise.
6195 (altivec_expand_builtin): Add cases for
6196 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
6197 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
6198 (altivec_init_builtins): Add definitions for
6199 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
6200 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
6202 2014-02-21 Catherine Moore <clm@codesourcery.com>
6204 * doc/invoke.texi (mvirt, mno-virt): Document.
6205 * config/mips/mips.opt (mvirt): New option.
6206 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
6208 2014-02-21 Richard Biener <rguenther@suse.de>
6210 PR tree-optimization/60276
6211 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
6212 (STMT_VINFO_MIN_NEG_DIST): New macro.
6213 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
6214 STMT_VINFO_MIN_NEG_DIST.
6215 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
6216 made for negative dependence distances still hold.
6218 2014-02-21 Richard Biener <rguenther@suse.de>
6221 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
6222 DECL_INITIAL for globals not in the current function context.
6224 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6226 PR tree-optimization/56490
6227 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
6228 * tree-ssa-uninit.c: Include params.h.
6229 (compute_control_dep_chain): Add num_calls argument, return false
6230 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
6231 num_calls to recursive call.
6232 (find_predicates): Change dep_chain into normal array,
6233 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
6234 variable and adjust compute_control_dep_chain caller.
6235 (find_def_preds): Likewise.
6237 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
6239 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
6240 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
6242 2014-02-21 Nick Clifton <nickc@redhat.com>
6244 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
6245 (pushhi1): Likewise.
6246 (popqi1): Add mode to pre_dec.
6249 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6251 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
6252 mode for mask of V8SFmode permutation.
6254 2014-02-20 Richard Henderson <rth@redhat.com>
6257 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
6258 a new pseudo for OLDVAL.
6260 2014-02-20 Jakub Jelinek <jakub@redhat.com>
6263 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
6264 gen_reg_rtx if d->testing_p.
6265 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
6266 if d->testing_p and we will certainly return true.
6267 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6270 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6272 * emit-rtl.c (gen_reg_rtx): Assert that
6273 crtl->emit.regno_pointer_align_length is non-zero.
6275 2014-02-20 Richard Henderson <rth@redhat.com>
6278 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6279 on failure the store back into EXPECT.
6281 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6282 Sandra Loosemore <sandra@codesourcery.com>
6284 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6285 * config/nios2/nios2.c (nios2_function_profiler): Add
6286 -fPIC (flag_pic == 2) support.
6287 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6288 (nios2_large_offset_p): New function.
6289 (nios2_unspec_reloc_p): Move up position, update to use
6290 nios2_large_offset_p.
6291 (nios2_unspec_address): Remove function.
6292 (nios2_unspec_offset): New function.
6293 (nios2_large_got_address): New function.
6294 (nios2_got_address): Add large offset support.
6295 (nios2_legitimize_tls_address): Update usage of removed and new
6297 (nios2_symbol_binds_local_p): New function.
6298 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
6299 (nios2_legitimize_address): Update to use nios2_large_offset_p.
6300 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
6301 (nios2_print_operand): Merge H/L processing, add hiadj/lo
6302 processing for (const (unspec ...)).
6303 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
6305 2014-02-20 Richard Biener <rguenther@suse.de>
6307 * tree-cfg.c (replace_uses_by): Mark altered BBs before
6308 doing the substitution.
6309 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
6311 2014-02-20 Martin Jambor <mjambor@suse.cz>
6314 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
6315 info when checking whether lattices are bottom.
6317 2014-02-20 Richard Biener <rguenther@suse.de>
6320 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
6323 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
6326 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
6327 parameter specifying the scaling.
6328 (inline_call): Update.
6329 (want_inline_recursively): Guard division by zero.
6330 (recursive_inlining): Update.
6331 * ipa-inline.h (clone_inlined_nodes): Update.
6333 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6336 * config/i386/i386.c (classify_argument): Pass structures of size
6337 64 bytes or less in register.
6339 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6340 Kirill Yukhin <kirill.yukhin@intel.com>
6342 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
6343 (_mm_rcp28_round_ss): Ditto.
6344 (_mm_rsqrt28_round_sd): Ditto.
6345 (_mm_rsqrt28_round_ss): Ditto.
6346 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
6347 (_mm_rcp14_round_ss): Ditto.
6348 (_mm_rsqrt14_round_sd): Ditto.
6349 (_mm_rsqrt14_round_ss): Ditto.
6350 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
6351 the first input operand, get rid of match_dup.
6352 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
6354 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
6356 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
6357 operand as the first input operand, set type attribute.
6358 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
6360 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
6361 operand as the first input operand, set type attribute.
6363 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6365 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
6368 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
6371 * config/i386/i386.c (construct_container): Remove TFmode check
6372 for X86_64_INTEGER_CLASS.
6374 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
6377 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
6378 only when -Wpsabi is enabled.
6380 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
6383 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
6384 passing arrays in registers are the same as for structs, so remove the
6385 special case for them.
6387 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
6389 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
6390 destination type, extract only the valid bits if the source type is not
6391 integral and has a different mode.
6393 2014-02-19 Richard Biener <rguenther@suse.de>
6396 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
6399 2014-02-19 Richard Biener <rguenther@suse.de>
6402 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
6403 (ipa_modify_call_arguments): Emit an argument load explicitely and
6404 preserve virtual SSA form there and for the replacement call.
6405 Do not update SSA form nor free dominance info.
6407 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6409 * ipa.c (function_and_variable_visibility): Also clear WEAK
6410 flag when disolving COMDAT_GROUP.
6412 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6414 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
6415 * ipa-prop.c (ipa_set_jf_known_type): Return early when
6417 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
6418 do more sanity checks.
6419 (detect_type_change): Return true when giving up early.
6420 (compute_complex_assign_jump_func): Fix type parameter of
6421 ipa_set_ancestor_jf.
6422 (compute_complex_ancestor_jump_func): Likewise.
6423 (update_jump_functions_after_inlining): Fix updating of
6425 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
6427 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6429 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
6430 inline clones when edge disappears.
6432 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6435 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
6436 Split 64-bit moves into 2 patterns. Do not allow the use of
6437 direct move for TDmode in little endian, since the decimal value
6438 has little endian bytes within a word, but the 64-bit pieces are
6439 ordered in a big endian fashion, and normal subreg's of TDmode are
6441 (mov<mode>_64bit_dm): Likewise.
6442 (movtd_64bit_nodm): Likewise.
6444 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6446 PR tree-optimization/60174
6447 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
6448 statement of an SSA_NAME that occurs in an abnormal PHI node.
6450 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6453 * final.c (SEEN_BB): Remove.
6454 (SEEN_NOTE, SEEN_EMITTED): Renumber.
6455 (final_scan_insn): Don't force_source_line on second
6456 NOTE_INSN_BASIC_BLOCK.
6458 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
6461 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
6462 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
6463 (type_natural_mode): Warn ABI change when %zmm register is not
6464 available for AVX512F vector value passing.
6466 2014-02-18 Kai Tietz <ktietz@redhat.com>
6469 * config/i386/i386.c (ix86_expand_prologue): Use value in
6470 rax register as displacement when restoring %r10 or %rax.
6471 Fix wrong offset when restoring both registers.
6473 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6475 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
6476 assertion with conditional return.
6478 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6479 Uros Bizjak <ubizjak@gmail.com>
6482 * config/i386/driver-i386.c (host_detect_local_cpu): If
6483 YMM state is not saved by the OS, also clear has_f16c. Move
6484 CPUID 0x80000001 handling before YMM state saving checking.
6486 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
6488 PR rtl-optimization/58960
6489 * haifa-sched.c (alloc_global_sched_pressure_data): New,
6490 factored out from ...
6491 (sched_init): ... here.
6492 (free_global_sched_pressure_data): New, factored out from ...
6493 (sched_finish): ... here.
6494 * sched-int.h (free_global_sched_pressure_data): Declare.
6495 * sched-rgn.c (nr_regions_initial): New static global.
6496 (haifa_find_rgns): Initialize it.
6497 (schedule_region): Disable sched-pressure for the newly
6500 2014-02-17 Richard Biener <rguenther@suse.de>
6502 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
6503 release SSA defs of pattern stmts.
6505 2014-02-17 Richard Biener <rguenther@suse.de>
6507 * tree-inline.c (expand_call_inline): Release the virtual
6508 operand defined by the call we are about to inline.
6510 2014-02-17 Richard Biener <rguenther@suse.de>
6512 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
6514 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
6515 Ilya Tocar <ilya.tocar@intel.com>
6517 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
6518 arguments order in builtin.
6519 (_mm512_permutexvar_epi64): Ditto.
6520 (_mm512_mask_permutexvar_epi64): Ditto
6521 (_mm512_maskz_permutexvar_epi32): Ditto
6522 (_mm512_permutexvar_epi32): Ditto
6523 (_mm512_mask_permutexvar_epi32): Ditto
6525 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6527 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
6528 (p8_vmrgow): Likewise.
6530 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6532 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
6535 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
6538 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
6539 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
6540 into 64-bit and 32-bit moves. On 64-bit moves, add support for
6541 using direct move instructions on ISA 2.07. Also adjust
6542 instruction length for 64-bit.
6543 (mov<mode>_64bit, TFmode/TDmode): Likewise.
6544 (mov<mode>_32bit, TFmode/TDmode): Likewise.
6546 2014-02-15 Alan Modra <amodra@gmail.com>
6550 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
6551 find_replacement on parts of insn rtl that might be reloaded.
6553 2014-02-15 Richard Biener <rguenther@suse.de>
6555 PR tree-optimization/60183
6556 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
6557 (tree_ssa_phiprop): Calculate and free post-dominators.
6559 2014-02-14 Jeff Law <law@redhat.com>
6561 PR rtl-optimization/60131
6562 * ree.c (get_extended_src_reg): New function.
6563 (combine_reaching_defs): Use it rather than assuming location of REG.
6564 (find_and_remove_re): Verify first operand of extension is
6565 a REG before adding the insns to the copy list.
6567 2014-02-14 Roland McGrath <mcgrathr@google.com>
6569 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
6570 * configure: Regenerated.
6571 * config.in: Regenerated.
6572 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
6573 instead of ASM_SHORT.
6575 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
6576 Richard Earnshaw <rearnsha@arm.com>
6578 PR rtl-optimization/59535
6579 * lra-constraints.c (process_alt_operands): Encourage alternative
6580 when unassigned pseudo class is superset of the alternative class.
6581 (inherit_reload_reg): Don't inherit when optimizing for code size.
6582 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
6583 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
6584 modes not less than 4 for Thumb1.
6586 2014-02-14 Kyle McMartin <kyle@redhat.com>
6589 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
6591 2014-02-14 Richard Biener <rguenther@suse.de>
6593 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
6594 (get_frame_arg): Drop the assert with langhook types_compatible_p.
6595 Do not strip INDIRECT_REFs.
6597 2014-02-14 Richard Biener <rguenther@suse.de>
6600 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
6601 DECL_FUNCTION_SPECIFIC_TARGET.
6602 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
6603 * tree-streamer-out.c (pack_ts_target_option): Remove.
6604 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
6605 (write_ts_function_decl_tree_pointers): Do not stream
6606 DECL_FUNCTION_SPECIFIC_TARGET.
6607 * tree-streamer-in.c (unpack_ts_target_option): Remove.
6608 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
6609 (lto_input_ts_function_decl_tree_pointers): Do not stream
6610 DECL_FUNCTION_SPECIFIC_TARGET.
6612 2014-02-14 Jakub Jelinek <jakub@redhat.com>
6614 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
6615 (get_initial_def_for_induction, vectorizable_induction): Ignore
6616 debug stmts when looking for exit_phi.
6617 (vectorizable_live_operation): Fix up condition.
6619 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6621 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
6622 nreverse() because it changes the content of original tree list.
6624 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6626 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
6627 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
6629 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6631 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
6632 GNU coding standards.
6634 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6637 * dwarf2out.c (gen_subprogram_die): Don't call
6638 add_calling_convention_attribute if subr_die is old_die.
6640 2014-02-13 Sharad Singhai <singhai@google.com>
6642 * doc/optinfo.texi: Fix order of nodes.
6644 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
6646 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
6647 operands[2], not operands[3].
6649 2014-02-13 Richard Biener <rguenther@suse.de>
6652 * doc/install.texi (ISL): Update recommended version to 0.12.2,
6653 mention the possibility of an in-tree build.
6654 (CLooG): Update recommended version to 0.18.1, mention the
6655 possibility of an in-tree build and clarify that the ISL
6656 bundled with CLooG does not work.
6658 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6661 * expr.c (compress_float_constant): If x is a hard register,
6662 extend into a pseudo and then move to x.
6664 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
6666 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
6667 caused by bad second argument to warning_at() with -mhotpatch and
6668 nested functions (e.g. with gfortran).
6670 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
6672 * opts.c (option_name): Remove "enabled by default" rider.
6674 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
6676 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
6678 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
6679 Uros Bizjak <ubizjak@gmail.com>
6682 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
6683 * configure: Regenerated.
6685 2014-02-12 Richard Biener <rguenther@suse.de>
6687 * vec.c (vec_prefix::calculate_allocation): Move as
6688 inline variant to vec.h.
6689 (vec_prefix::calculate_allocation_1): New out-of-line version.
6690 * vec.h (vec_prefix::calculate_allocation_1): Declare.
6691 (vec_prefix::m_has_auto_buf): Rename to ...
6692 (vec_prefix::m_using_auto_storage): ... this.
6693 (vec_prefix::calculate_allocation): Inline the easy cases
6694 and dispatch to calculate_allocation_1 which doesn't need the
6696 (va_heap::reserve): Use gcc_checking_assert.
6697 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
6698 m_using_auto_storage.
6699 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
6701 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
6702 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
6703 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
6705 2014-02-12 Richard Biener <rguenther@suse.de>
6707 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
6708 when we found a dependence.
6710 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
6712 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
6714 (maybe_fold_stmt): ... into this new function.
6715 * omp-low.c (lower_omp): Update comment.
6717 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
6720 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
6723 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
6725 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
6726 identifiers in comments.
6727 (cortexa53_extra_costs): Likewise.
6728 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
6729 (cortexa7_extra_costs): Likewise.
6730 (cortexa12_extra_costs): Likewise.
6731 (cortexa15_extra_costs): Likewise.
6732 (v7m_extra_costs): Likewise.
6734 2014-02-12 Richard Biener <rguenther@suse.de>
6737 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
6738 of posix_memalign being successful.
6739 (lower_stmt): Restrict lowering of posix_memalign to when
6740 -ftree-bit-ccp is enabled.
6742 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6744 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
6746 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
6748 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
6750 PR rtl-optimization/60116
6751 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
6752 other_insn once the combination has been validated.
6754 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
6757 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
6759 * ipa-devirt.c: Include demangle.h
6760 (odr_violation_reported): New static variable.
6761 (add_type_duplicate): Update odr_violations.
6762 (maybe_record_node): Add completep parameter; update it.
6763 (record_target_from_binfo): Add COMPLETEP parameter;
6764 update it as needed.
6765 (possible_polymorphic_call_targets_1): Likewise.
6766 (struct polymorphic_call_target_d): Add nonconstruction_targets;
6767 rename FINAL to COMPLETE.
6768 (record_targets_from_bases): Sanity check we found the binfo;
6769 fix COMPLETEP updating.
6770 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
6771 parameter, fix computing of COMPLETEP.
6772 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
6773 at LTO time do demangling.
6774 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
6775 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
6777 (gimple_get_virt_method_for_binfo): Likewise.
6778 * gimple-fold.h (gimple_get_virt_method_for_binfo,
6779 gimple_get_virt_method_for_vtable): Update prototypes.
6781 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
6784 * genautomata.c (add_presence_absence): Fix typo with
6785 {final_}presence_list.
6787 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
6790 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
6791 for VSX/Altivec vectors that land in GPR registers.
6793 2014-02-11 Richard Henderson <rth@redhat.com>
6794 Jakub Jelinek <jakub@redhat.com>
6797 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
6798 around drhs if type conversion to lacc->type is not useless.
6800 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6802 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
6804 (cortex-a57.cortex-a53): Likewise.
6805 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
6807 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6809 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
6812 2014-02-11 Renlin Li <Renlin.Li@arm.com>
6814 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
6815 add_options_for_arm_vfp3.
6817 2014-02-11 Jeff Law <law@redhat.com>
6820 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
6821 object with an undesirable mode.
6823 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6826 * config/i386/sol2-9.h: New file.
6827 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
6828 *-*-solaris2.9*): Use it.
6830 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6832 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
6833 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
6835 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6837 * config/microblaze/microblaze.c: Extend mcpu version format
6839 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
6841 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
6843 2014-02-10 Richard Henderson <rth@redhat.com>
6846 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
6847 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
6848 ms-abi vs -mno-accumulate-outgoing-args.
6849 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
6850 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
6853 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
6856 * cfgexpand.c (expand_asm_operands): Attach source location to
6857 ASM_INPUT rtx objects.
6858 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
6860 2014-02-10 Nick Clifton <nickc@redhat.com>
6862 * config/mn10300/mn10300.c (popcount): New function.
6863 (mn10300_expand_prologue): Include saved registers in stack usage
6866 2014-02-10 Jeff Law <law@redhat.com>
6869 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
6870 when changing the SET_DEST of a prior insn to avoid an input reload.
6872 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6874 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
6875 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
6876 -mcall-openbsd, or -mcall-linux.
6877 (CC1_ENDIAN_BIG_SPEC): Remove.
6878 (CC1_ENDIAN_LITTLE_SPEC): Remove.
6879 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
6880 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
6881 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
6882 and %cc1_endian_default.
6883 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
6885 2014-02-10 Richard Biener <rguenther@suse.de>
6887 PR tree-optimization/60115
6888 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
6889 MEM_REF handling. Properly verify that the accesses are not
6890 out of the objects bound.
6892 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6894 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
6897 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
6899 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
6900 proper constants and fix formatting.
6901 (possible_polymorphic_call_targets): Fix formatting.
6903 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
6904 Ilya Tocar <ilya.tocar@intel.com>
6906 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
6907 (_mm512_loadu_epi32): Renamed into...
6908 (_mm512_loadu_si512): This.
6909 (_mm512_storeu_epi32): Renamed into...
6910 (_mm512_storeu_si512): This.
6911 (_mm512_maskz_ceil_ps): Removed.
6912 (_mm512_maskz_ceil_pd): Ditto.
6913 (_mm512_maskz_floor_ps): Ditto.
6914 (_mm512_maskz_floor_pd): Ditto.
6915 (_mm512_floor_round_ps): Ditto.
6916 (_mm512_floor_round_pd): Ditto.
6917 (_mm512_ceil_round_ps): Ditto.
6918 (_mm512_ceil_round_pd): Ditto.
6919 (_mm512_mask_floor_round_ps): Ditto.
6920 (_mm512_mask_floor_round_pd): Ditto.
6921 (_mm512_mask_ceil_round_ps): Ditto.
6922 (_mm512_mask_ceil_round_pd): Ditto.
6923 (_mm512_maskz_floor_round_ps): Ditto.
6924 (_mm512_maskz_floor_round_pd): Ditto.
6925 (_mm512_maskz_ceil_round_ps): Ditto.
6926 (_mm512_maskz_ceil_round_pd): Ditto.
6927 (_mm512_expand_pd): Ditto.
6928 (_mm512_expand_ps): Ditto.
6929 * config/i386/i386.c (ix86_builtins): Remove
6930 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
6931 (bdesc_args): Ditto.
6932 * config/i386/predicates.md (const1256_operand): New.
6933 (const_1_to_2_operand): Ditto.
6934 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6935 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6936 (*avx512pf_gatherpf<mode>sf): Ditto.
6937 (avx512pf_gatherpf<mode>df): Ditto.
6938 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6939 (*avx512pf_gatherpf<mode>df): Ditto.
6940 (avx512pf_scatterpf<mode>sf): Ditto.
6941 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6942 (*avx512pf_scatterpf<mode>sf): Ditto.
6943 (avx512pf_scatterpf<mode>df): Ditto.
6944 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6945 (*avx512pf_scatterpf<mode>df): Ditto.
6946 (avx512f_expand<mode>): Removed.
6947 (<shift_insn><mode>3<mask_name>): Change predicate type.
6949 2014-02-08 Jakub Jelinek <jakub@redhat.com>
6951 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
6952 not at the end of datarefs vector use ordered_remove to avoid
6953 reordering datarefs vector.
6956 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
6957 mark local addressable non-static vars as GOVD_PRIVATE
6958 instead of GOVD_LOCAL.
6959 * omp-low.c (lower_omp_for): Move gimple_bind_vars
6960 and BLOCK_VARS of gimple_bind_block to new_stmt rather
6964 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
6965 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
6966 assume_aligned or alloc_align attributes.
6967 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
6968 arguments. Handle also assume_aligned and alloc_align attributes.
6969 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
6970 calls to functions with assume_aligned or alloc_align attributes.
6971 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
6973 2014-02-08 Terry Guo <terry.guo@arm.com>
6975 * doc/invoke.texi: Document ARM -march=armv7e-m.
6977 2014-02-08 Jakub Jelinek <jakub@redhat.com>
6979 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
6980 flag on __cilkrts_rethrow builtin.
6983 * ipa-cp.c (determine_versionability): Fail at -O0
6984 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
6985 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
6988 2014-02-04 Jakub Jelinek <jakub@redhat.com>
6991 * tree-inline.c (copy_forbidden): Fail for
6992 __attribute__((optimize (0))) functions.
6994 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
6996 * varpool.c: Include pointer-set.h.
6997 (varpool_remove_unreferenced_decls): Variables in other partitions
6998 will not be output; be however careful to not lose information
7001 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7003 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
7004 lookup in the vtable constructor.
7006 2014-02-07 Jeff Law <law@redhat.com>
7009 * config/m68k/m68k.md (ashldi_extsi): Turn into a
7010 define_insn_and_split.
7012 * ipa-inline.c (inline_small_functions): Fix typos.
7014 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7016 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
7017 (s390_can_use_return_insn): Declare.
7018 * config/s390/s390.h (EPILOGUE_USES): Define.
7019 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
7021 (s390_chunkify_start): Handle return JUMP_LABELs.
7022 (s390_early_mach): Emit a main_pool instruction on the entry edge.
7023 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
7024 (s390_can_use_return_insn): New functions.
7025 (s390_fix_long_loop_prediction): Handle conditional returns.
7026 (TARGET_SET_UP_BY_PROLOGUE): Define.
7027 * config/s390/s390.md (ANY_RETURN): New code iterator.
7028 (*creturn, *csimple_return, return, simple_return): New patterns.
7030 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7032 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
7033 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
7034 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
7035 REG_CFA_RESTORE list when deciding not to restore a register.
7037 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7039 * config/s390/s390.c: Include tree-pass.h and context.h.
7040 (s390_early_mach): New function, split out from...
7041 (s390_emit_prologue): ...here.
7042 (pass_data_s390_early_mach): New pass structure.
7043 (pass_s390_early_mach): New class.
7044 (s390_option_override): Create and register early_mach pass.
7045 Move to end of file.
7047 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7049 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
7050 to match for the exit block.
7052 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7054 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
7055 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
7056 Reject misaligned operands.
7058 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7060 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
7062 2014-02-07 Richard Biener <rguenther@suse.de>
7065 * gimple-low.c (lower_builtin_posix_memalign): New function.
7066 (lower_stmt): Call it to lower posix_memalign in a way
7067 to make alignment info accessible.
7069 2014-02-07 Jakub Jelinek <jakub@redhat.com>
7072 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
7073 __builtin_setjmp_receiver.
7075 2014-02-07 Richard Biener <rguenther@suse.de>
7078 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
7079 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
7080 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7081 Handle BUILT_IN_POSIX_MEMALIGN.
7082 (find_func_clobbers): Likewise.
7083 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
7084 (call_may_clobber_ref_p_1): Likewise.
7086 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7089 * ipa-devirt.c (record_target_from_binfo): Remove overactive
7092 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7095 * lto-cgraph.c (lto_output_node): Use
7096 symtab_get_symbol_partitioning_class.
7097 (lto_output_varpool_node): likewise.
7098 (symtab_get_symbol_partitioning_class): Move here from
7100 * cgraph.h (symbol_partitioning_class): Likewise.
7101 (symtab_get_symbol_partitioning_class): Declare.
7103 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7105 * ggc.h (ggc_internal_cleared_alloc): New macro.
7106 * vec.h (vec_safe_copy): Handle memory stats.
7107 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
7108 * target-globals.c (save_target_globals): Likewise.
7110 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7113 * expr.c (emit_move_resolve_push): Export; be bit more selective
7114 on when to clear alias set.
7115 * expr.h (emit_move_resolve_push): Declare.
7116 * function.h (struct function): Add tail_call_marked.
7117 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
7118 * config/i386/i386-protos.h (ix86_expand_push): Remove.
7119 * config/i386/i386.md (TImode move expander): De not call
7121 (FP push expanders): Preserve memory attributes.
7122 * config/i386/sse.md (push<mode>1): Remove.
7123 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
7124 (ix86_expand_push): Remove.
7125 * config/i386/mmx.md (push<mode>1): Remove.
7127 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7129 PR rtl-optimization/60030
7130 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
7131 lopart with paradoxical subreg before shifting it up by hprec.
7133 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7135 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
7136 Remove extra newline at end of file.
7137 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
7138 (arm_issue_rate): Handle cortexa57.
7139 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
7140 (cortex-a57.cortex-a53): Likewise.
7142 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7145 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
7146 don't record in REG_FRAME_RELATED_EXPR registers not set in that
7148 (arm_expand_prologue): Adjust all callers.
7149 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
7150 info, registers also at the lowest numbered registers side. Use
7151 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
7155 * var-tracking.c (adjust_mems): Before adding a SET to
7156 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
7158 2014-02-06 Alan Modra <amodra@gmail.com>
7161 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
7162 change SDmode to DDmode when lra_in_progress.
7164 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7167 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
7168 free_data_ref on the dr first, and before goto again also set dr
7169 to the next dr. For simd_lane_access, free old datarefs[i] before
7170 overwriting it. For get_vectype_for_scalar_type failure, don't
7171 free_data_ref if simd_lane_access.
7173 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
7176 * tree.h (opts_for_fn): New inline function.
7177 (opt_for_fn): Define.
7178 * config/i386/i386.c (ix86_function_regparm): Use
7179 opt_for_fn (decl, optimize) instead of optimize.
7181 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
7183 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
7184 for SYMBOL_REF in large memory model.
7186 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7188 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
7190 (cortex-a57): Likewise.
7191 (cortex-a57.cortex-a53): Likewise.
7193 2014-02-06 Yury Gribov <y.gribov@samsung.com>
7194 Kugan Vivekanandarajah <kuganv@linaro.org>
7196 * config/arm/arm.c (arm_vector_alignment_reachable): Check
7198 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
7200 2014-02-06 Richard Biener <rguenther@suse.de>
7202 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
7203 set_loop_copy and initialize_original_copy_tables.
7205 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
7207 * config/aarch64/aarch64-simd.md
7208 (aarch64_ashr_simddi): Change QI to SI.
7210 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7211 Jakub Jelinek <jakub@redhat.com>
7214 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
7217 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7219 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
7220 CODE_FOR_altivec_vpku[hw]um to
7221 CODE_FOR_altivec_vpku[hw]um_direct.
7222 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
7223 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
7224 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
7225 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
7227 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7229 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
7230 generation for -maltivec=be.
7231 (altivec_vsumsws): Simplify redundant test.
7233 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7235 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
7236 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
7237 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
7238 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
7239 gen_altivec_vpkuwum.
7240 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
7242 (altivec_vpks<VI_char>ss): Likewise.
7243 (altivec_vpks<VI_char>us): Likewise.
7244 (altivec_vpku<VI_char>us): Likewise.
7245 (altivec_vpku<VI_char>um): Likewise.
7246 (altivec_vpku<VI_char>um_direct): New (copy of
7247 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
7249 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
7250 target is little endian and -maltivec=be is not specified.
7251 (*altivec_vupkhs<VU_char>_direct): New (copy of
7252 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
7253 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
7254 target is little endian and -maltivec=be is not specified.
7255 (*altivec_vupkls<VU_char>_direct): New (copy of
7256 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
7257 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
7258 little endian and -maltivec=be is not specified.
7259 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
7260 little endian and -maltivec=be is not specified.
7262 2014-02-05 Richard Henderson <rth@redhat.com>
7265 * combine-stack-adj.c: Revert r206943.
7266 * sched-int.h (struct deps_desc): Add last_args_size.
7267 * sched-deps.c (init_deps): Initialize it.
7268 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7269 contain REG_ARGS_SIZE notes.
7271 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7273 * lto-cgraph.c (asm_nodes_output): Make global.
7274 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7275 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7276 (driver_handle_option): Handle OPT_fwpa.
7278 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7281 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7282 a comment typo and formatting issue. If odr_hash hasn't been
7283 created, return vNULL and set *completep to false.
7286 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7287 bb with no successors.
7289 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7292 * doc/invoke.texi (-march): Clarify documentation for ARM.
7296 2014-02-05 Richard Biener <rguenther@suse.de>
7298 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
7299 when not vectorizing because of too many alias checks.
7300 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
7301 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
7303 2014-02-05 Nick Clifton <nickc@redhat.com>
7305 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
7306 accept extended registers in any mode when compiling for the MN10300.
7308 2014-02-05 Yury Gribov <y.gribov@samsung.com>
7310 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
7311 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
7312 sanitization attributes.
7313 (can_inline_edge_p): Likewise.
7314 (sanitize_attrs_match_for_inline_p): New function.
7316 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7318 * ipa-prop.c (detect_type_change): Shor circuit testing of
7319 type changes on THIS pointer.
7321 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
7324 * config/pa/pa.c (legitimize_tls_address): Return original address
7325 if not passed a SYMBOL_REF rtx.
7326 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
7328 (pa_emit_move_sequence): Simplify TLS source operands.
7329 (pa_legitimate_constant_p): Reject all TLS constants.
7330 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
7331 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
7333 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7335 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
7336 groups when we know they are controlled by LTO.
7337 * varasm.c (default_binds_local_p_1): If object is in other partition,
7338 it will be resolved locally.
7340 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7342 * config/host-linux.c (linux_gt_pch_use_address): Don't
7343 use SSIZE_MAX because it is not always defined.
7345 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
7348 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
7349 threshold for pseudo splitting.
7350 (update_ebb_live_info): Process call argument hard registers and
7351 hard registers from insn definition too.
7352 (max_small_class_regs_num): New constant.
7353 (inherit_in_ebb): Update live hard regs through EBBs. Update
7354 reloads_num only for small register classes. Don't split for
7357 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
7360 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
7363 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7365 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
7368 2014-02-04 Marek Polacek <polacek@redhat.com>
7370 * gdbinit.in (pel): Define.
7372 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7374 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
7377 2014-02-04 Richard Biener <rguenther@suse.de>
7380 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
7381 in function context local.
7382 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
7383 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
7384 similar to LTO_imported_decl_ref.
7386 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7388 PR tree-optimization/60002
7389 * cgraphclones.c (build_function_decl_skip_args): Clear
7392 PR tree-optimization/60023
7393 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
7394 false to gsi_replace.
7395 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
7396 has been in some EH region and vec_stmt could throw, add
7397 vec_stmt into the same EH region.
7398 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
7399 has no lhs, ignore it.
7400 * internal-fn.c (expand_MASK_LOAD): Likewise.
7403 * tree-inline.c (copy_forbidden): Fail for
7404 __attribute__((optimize (0))) functions.
7407 * omp-low.c (simd_clone_struct_copy): If from->inbranch
7408 is set, copy one less argument.
7409 (expand_simd_clones): Don't subtract clone_info->inbranch
7410 from simd_clone_struct_alloc argument.
7412 PR rtl-optimization/57915
7413 * recog.c (simplify_while_replacing): If all unary/binary/relational
7414 operation arguments are constant, attempt to simplify those.
7417 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
7418 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
7420 2014-02-04 Richard Biener <rguenther@suse.de>
7422 PR tree-optimization/60012
7423 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
7424 TBAA disambiguation to all DDRs.
7426 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7429 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
7430 (LINK_SPEC): Use it for -shared, -shared-libgcc.
7432 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7435 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
7437 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7439 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
7440 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
7442 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7445 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
7446 to figure out targets of polymorphic calls with known decl.
7447 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7448 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
7449 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
7450 (get_polymorphic_call_info): ... here.
7451 (get_polymorphic_call_info_from_invariant): New function.
7453 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7455 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
7456 lookup via vtable pointer; check for type consistency
7457 and turn inconsitent facts into UNREACHABLE.
7458 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7459 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
7460 type inconsistent querries; return UNREACHABLE instead.
7462 2014-02-03 Richard Henderson <rth@twiddle.net>
7465 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
7466 already processed this node.
7467 (normalize_one_pred_1): Pass along mark_set.
7468 (normalize_one_pred): Create and destroy a pointer_set_t.
7469 (normalize_one_pred_chain): Likewise.
7471 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
7473 PR gcov-profile/58602
7474 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
7476 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7479 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
7480 -fno-devirtualize; try to devirtualize by the knowledge of
7481 virtual table pointer given by aggregate propagation.
7482 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7483 (ipa_print_node_jump_functions): Dump also offset that
7484 is relevant for polymorphic calls.
7485 (determine_known_aggregate_parts): Add arg_type parameter; use it
7486 instead of determining the type from pointer type.
7487 (ipa_compute_jump_functions_for_edge): Update call of
7488 determine_known_aggregate_parts.
7489 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
7490 (gimple_get_virt_method_for_binfo): ... here; simplify using
7491 vtable_pointer_value_to_vtable.
7492 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
7493 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
7494 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
7495 (vtable_pointer_value_to_vtable): Break out from ...; handle also
7497 (vtable_pointer_value_to_binfo): ... here.
7498 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
7500 2014-02-03 Teresa Johnson <tejohnson@google.com>
7502 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
7503 redef of outer loop index variable.
7505 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
7509 * doc/extend.texi (Function Attributes): Typo.
7511 2014-02-03 Cong Hou <congh@google.com>
7513 PR tree-optimization/60000
7514 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
7515 if the vectorized statement is a store. A store statement can only
7516 appear at the end of pattern statements.
7518 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7520 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
7521 (ix86_option_override_internal): Default long double to 64-bit for
7522 32-bit Bionic and to 128-bit for 64-bit Bionic.
7524 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
7525 TARGET_LONG_DOUBLE_128 is true.
7526 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
7528 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
7529 (mlong-double-64): Negate -mlong-double-128.
7530 (mlong-double-128): New option.
7532 * config/i386/i386-c.c (ix86_target_macros): Define
7533 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
7535 * doc/invoke.texi: Document -mlong-double-128.
7537 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7539 PR rtl-optimization/60024
7540 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
7542 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
7544 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
7546 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7548 PR rtl-optimization/57662
7549 * sel-sched.c (code_motion_path_driver): Do not mark already not
7550 existing blocks in the visiting bitmap.
7552 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7554 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
7555 on the insn being emitted.
7557 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
7558 Will Deacon <will.deacon@arm.com>
7560 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
7562 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7564 * config/arm/arm-tables.opt: Regenerate.
7566 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7568 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
7569 for vector types other than V16QImode.
7570 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
7571 define_expand, and call altivec_expand_vec_perm_le when producing
7572 code with little endian element order.
7573 (*altivec_vperm_<mode>_internal): New insn having previous
7574 behavior of altivec_vperm_<mode>.
7575 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
7576 altivec_expand_vec_perm_le when producing code with little endian
7578 (*altivec_vperm_<mode>_uns_internal): New insn having previous
7579 behavior of altivec_vperm_<mode>_uns.
7581 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7583 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
7584 (altivec_vsumsws): Add handling for -maltivec=be with a little
7586 (altivec_vsumsws_direct): New.
7587 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
7588 gen_altivec_vsumsws.
7590 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7592 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
7593 vtable_pointer_value_to_binfo): New functions.
7594 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
7595 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
7597 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
7599 * config/nios2/nios2.md (load_got_register): Initialize GOT
7600 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
7601 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
7603 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7605 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
7606 preserverd by passthrough, do not propagate the type.
7608 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7610 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
7611 (mips_atomic_assign_expand_fenv): New function.
7612 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
7614 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7616 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
7617 (__builtin_mips_set_fcsr): Likewise.
7618 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
7619 MIPS_USI_FTYPE_VOID.
7620 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
7621 (mips16_expand_set_fcsr): Likewise.
7622 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
7623 (mips16_set_fcsr_stub): Likewise.
7624 (mips16_get_fcsr_one_only_stub): New class.
7625 (mips16_set_fcsr_one_only_stub): Likewise.
7626 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
7627 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
7628 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
7629 (hard_float): New availability predicate.
7630 (mips_builtins): Add get_fcsr and set_fcsr.
7631 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
7632 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
7633 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
7634 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
7635 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
7638 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7640 * config/mips/mips.c (mips_one_only_stub): New class.
7641 (mips_need_mips16_rdhwr_p): Replace with...
7642 (mips16_rdhwr_stub): ...this new variable.
7643 (mips16_stub_call_address): New function.
7644 (mips16_rdhwr_one_only_stub): New class.
7645 (mips_expand_thread_pointer): Use mips16_stub_call_address.
7646 (mips_output_mips16_rdhwr): Delete.
7647 (mips_finish_stub): New function.
7648 (mips_code_end): Use it to handle rdhwr stubs.
7650 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
7653 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
7654 when calculating size of integer atomic types.
7656 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
7658 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
7660 2014-02-01 Jakub Jelinek <jakub@redhat.com>
7662 PR tree-optimization/60003
7663 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
7664 * profile.c (branch_prob): Use gimple_call_builtin_p
7665 to check for BUILT_IN_SETJMP_RECEIVER.
7666 * tree-inline.c (copy_bb): Call notice_special_calls.
7668 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
7671 * lra-constraints.c (process_alt_operands): Update reload_sum only
7674 2014-01-31 Richard Henderson <rth@redhat.com>
7677 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
7678 until after else_eh is processed.
7680 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7682 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
7683 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
7684 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
7685 in smmintrin.h, remove them.
7686 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
7687 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
7688 * config/i386/i386.md (ROUND_SAE): Fix value.
7689 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
7690 (const48_operand): New.
7691 * config/i386/subst.md (round), (round_expand): Use
7692 const_4_or_8_to_11_operand.
7693 (round_saeonly), (round_saeonly_expand): Use const48_operand.
7695 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7697 * config/i386/constraints.md (Yk): Swap meaning with k.
7698 * config/i386/i386.md (movhi_internal): Change Yk to k.
7699 (movqi_internal): Ditto.
7700 (*k<logic><mode>): Ditto.
7703 (kandn<mode>): Ditto.
7704 (*<code>hi_1): Ditto.
7705 (*<code>qi_1): Ditto.
7706 (kxnor<mode>): Ditto.
7707 (kortestzhi): Ditto.
7708 (kortestchi): Ditto.
7710 (*one_cmplhi2_1): Ditto.
7711 (*one_cmplqi2_1): Ditto.
7712 * config/i386/sse.md (): Change k to Yk.
7713 (avx512f_load<mode>_mask): Ditto.
7714 (avx512f_blendm<mode>): Ditto.
7715 (avx512f_store<mode>_mask): Ditto.
7716 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
7717 (avx512f_storedqu<mode>_mask): Ditto.
7718 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
7720 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
7721 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
7722 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
7723 (avx512f_maskcmp<mode>3): Ditto.
7724 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
7725 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
7726 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
7727 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
7728 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
7729 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
7730 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
7731 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
7732 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
7733 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
7734 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
7735 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
7736 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
7737 (vec_extract_lo_<mode>_maskm): Ditto.
7738 (vec_extract_hi_<mode>_maskm): Ditto.
7739 (avx512f_vternlog<mode>_mask): Ditto.
7740 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
7741 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
7742 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
7743 (avx512f_<code>v8div16qi2_mask): Ditto.
7744 (avx512f_<code>v8div16qi2_mask_store): Ditto.
7745 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
7746 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
7747 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
7748 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
7749 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7750 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7751 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7752 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7753 (avx512cd_maskb_vec_dupv8di): Ditto.
7754 (avx512cd_maskw_vec_dupv16si): Ditto.
7755 (avx512f_vpermi2var<mode>3_maskz): Ditto.
7756 (avx512f_vpermi2var<mode>3_mask): Ditto.
7757 (avx512f_vpermi2var<mode>3_mask): Ditto.
7758 (avx512f_vpermt2var<mode>3_maskz): Ditto.
7759 (*avx512f_gathersi<mode>): Ditto.
7760 (*avx512f_gathersi<mode>_2): Ditto.
7761 (*avx512f_gatherdi<mode>): Ditto.
7762 (*avx512f_gatherdi<mode>_2): Ditto.
7763 (*avx512f_scattersi<mode>): Ditto.
7764 (*avx512f_scatterdi<mode>): Ditto.
7765 (avx512f_compress<mode>_mask): Ditto.
7766 (avx512f_compressstore<mode>_mask): Ditto.
7767 (avx512f_expand<mode>_mask): Ditto.
7768 * config/i386/subst.md (mask): Change k to Yk.
7769 (mask_scalar_merge): Ditto.
7772 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
7774 * doc/extend.texi (Vector Extensions): Document ?: in C++.
7776 2014-01-31 Richard Biener <rguenther@suse.de>
7779 * builtins.c (fold_builtin_memory_op): Make sure to not
7780 use a floating-point mode or a boolean or enumeral type for
7783 2014-01-30 DJ Delorie <dj@redhat.com>
7785 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
7786 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
7787 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
7788 whenever main() has an epilogue.
7790 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7792 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
7793 unused variable "field".
7794 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
7795 (vsx_mergeh_<mode>): Likewise.
7796 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
7797 (altivec_vmrghh): Likewise.
7798 (altivec_vmrghw): Likewise.
7799 (altivec_vmrglb): Likewise.
7800 (altivec_vmrglh): Likewise.
7801 (altivec_vmrglw): Likewise.
7802 (altivec_vspltb): Add missing uses.
7803 (altivec_vsplth): Likewise.
7804 (altivec_vspltw): Likewise.
7805 (altivec_vspltsf): Likewise.
7807 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7810 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
7811 frame related instructions.
7813 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
7815 PR rtl-optimization/59959
7816 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
7817 any reload of register whose subreg is invalid.
7819 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7821 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
7822 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
7823 Add missing return type - void.
7825 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7827 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
7828 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
7829 remove element index adjustment for endian (now handled in vsx.md
7831 (altivec_expand_vec_perm_const): Use
7832 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
7833 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
7834 (vsx_xxspltw_<mode>): Adjust element index for little endian.
7835 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
7836 define_expand and a new define_insn *altivec_vspltb_internal;
7837 adjust for -maltivec=be on a little endian target.
7838 (altivec_vspltb_direct): New.
7839 (altivec_vsplth): Divide into a define_expand and a new
7840 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
7841 little endian target.
7842 (altivec_vsplth_direct): New.
7843 (altivec_vspltw): Divide into a define_expand and a new
7844 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
7845 little endian target.
7846 (altivec_vspltw_direct): New.
7847 (altivec_vspltsf): Divide into a define_expand and a new
7848 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
7849 a little endian target.
7851 2014-01-30 Richard Biener <rguenther@suse.de>
7853 PR tree-optimization/59993
7854 * tree-ssa-forwprop.c (associate_pointerplus): Check we
7855 can propagate form the earlier stmt and avoid the transform
7856 when the intermediate result is needed.
7858 2014-01-30 Alangi Derick <alangiderick@gmail.com>
7860 * README.Portability: Fix typo.
7862 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
7864 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
7865 comparison_operator with ordered_comparison_operator.
7867 2014-01-30 Nick Clifton <nickc@redhat.com>
7869 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
7870 Rename to mn10300_store_multiple_regs.
7871 * config/mn10300/mn10300.c: Likewise.
7872 * config/mn10300/mn10300.md (store_movm): Fix typo: call
7873 store_multiple_regs.
7874 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
7875 Call mn10300_store_multiple_regs.
7877 2014-01-30 Nick Clifton <nickc@redhat.com>
7878 DJ Delorie <dj@redhat.com>
7880 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
7881 %fp 2 to keep registers after it properly word-aligned.
7882 (rl78_alloc_physical_registers_umul): Handle the case where both
7883 input operands are the same.
7885 2014-01-30 Richard Biener <rguenther@suse.de>
7887 PR tree-optimization/59903
7888 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
7891 2014-01-30 Jason Merrill <jason@redhat.com>
7894 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
7897 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
7899 2014-01-30 Richard Biener <rguenther@suse.de>
7901 PR tree-optimization/59951
7902 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
7904 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
7907 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
7908 SFmode to DFmode case.
7910 2014-01-29 DJ Delorie <dj@redhat.com>
7912 * config/msp430/msp430.opt (-minrt): New.
7913 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
7915 (ENDFILE_SPEC): Likewise.
7917 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
7919 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
7920 (estimate_function_body_sizes): Use it.
7922 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
7925 * dwarf2out.c (is_cxx_auto): New.
7926 (is_base_type): Use it.
7927 (gen_type_die_with_usage): Likewise.
7929 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7931 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
7932 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
7933 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
7934 -maltivec=be with LE targets.
7935 (vsx_mergeh_<mode>): Likewise.
7936 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
7937 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
7938 (altivec_vmrghb): Replace with define_expand and new
7939 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
7940 (altivec_vmrghb_direct): New define_insn.
7941 (altivec_vmrghh): Replace with define_expand and new
7942 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
7943 (altivec_vmrghh_direct): New define_insn.
7944 (altivec_vmrghw): Replace with define_expand and new
7945 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
7946 (altivec_vmrghw_direct): New define_insn.
7947 (*altivec_vmrghsf): Adjust for endianness.
7948 (altivec_vmrglb): Replace with define_expand and new
7949 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
7950 (altivec_vmrglb_direct): New define_insn.
7951 (altivec_vmrglh): Replace with define_expand and new
7952 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
7953 (altivec_vmrglh_direct): New define_insn.
7954 (altivec_vmrglw): Replace with define_expand and new
7955 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
7956 (altivec_vmrglw_direct): New define_insn.
7957 (*altivec_vmrglsf): Adjust for endianness.
7958 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
7959 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
7960 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
7961 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
7962 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
7963 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
7964 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
7965 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
7967 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
7969 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
7970 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
7973 2014-01-29 Richard Biener <rguenther@suse.de>
7975 PR tree-optimization/58742
7976 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
7977 associate_pointerplus_align.
7978 (associate_pointerplus_diff): New function.
7979 (associate_pointerplus): Likewise. Call associate_pointerplus_align
7980 and associate_pointerplus_diff.
7982 2014-01-29 Richard Biener <rguenther@suse.de>
7984 * lto-streamer.h (LTO_major_version): Bump to 3.
7985 (LTO_minor_version): Reset to 0.
7987 2014-01-29 Renlin Li <Renlin.Li@arm.com>
7989 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
7990 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
7991 (arm_file_start): Generate correct asm header for armv7ve.
7992 * config/arm/bpabi.h: Add multilib support for armv7ve.
7993 * config/arm/driver-arm.c: Change the architectures of cortex-a7
7994 and cortex-a15 to armv7ve.
7995 * config/arm/t-aprofile: Add multilib support for armv7ve.
7996 * doc/invoke.texi: Document -march=armv7ve.
7998 2014-01-29 Richard Biener <rguenther@suse.de>
8000 PR tree-optimization/58742
8001 * tree-ssa-forwprop.c (associate_plusminus): Return true
8002 if we changed sth, defer EH cleanup to ...
8003 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
8004 (simplify_mult): New function.
8006 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8009 PR tree-optimization/59920
8010 * tree.c (build_common_builtin_nodes): Remove
8011 __builtin_setjmp_dispatcher initialization.
8012 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
8013 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
8014 instead of gsi_after_labels + manually skipping debug stmts.
8015 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
8016 ignore bbs with IFN_ABNORMAL_DISPATCHER.
8017 * tree-inline.c (copy_edges_for_bb): Remove
8018 can_make_abnormal_goto argument, instead add abnormal_goto_dest
8019 argument. Ignore computed_goto_p stmts. Don't call
8020 make_abnormal_goto_edges. If a call might need abnormal edges
8021 for non-local gotos, see if it already has an edge to
8022 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
8023 with true argument, don't do anything then, otherwise add
8024 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
8025 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
8027 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
8028 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
8029 (lower_stmt): Don't set data->calls_builtin_setjmp.
8030 (lower_builtin_setjmp): Adjust comment.
8031 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
8032 * tree-cfg.c (found_computed_goto): Remove.
8033 (factor_computed_gotos): Remove.
8034 (make_goto_expr_edges): Return bool, true for computed gotos.
8035 Don't call make_abnormal_goto_edges.
8036 (build_gimple_cfg): Don't set found_computed_goto, don't call
8037 factor_computed_gotos.
8038 (computed_goto_p): No longer static.
8039 (make_blocks): Don't set found_computed_goto.
8040 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
8041 (make_edges): If make_goto_expr_edges returns true, push bb
8042 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
8043 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
8044 vector. Record mapping between bbs and OpenMP regions if there
8045 are any, adjust make_gimple_omp_edges caller. Call
8046 handle_abnormal_edges.
8047 (make_abnormal_goto_edges): Remove.
8048 * tree-cfg.h (make_abnormal_goto_edges): Remove.
8049 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
8050 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
8051 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
8052 * internal-fn.def (ABNORMAL_DISPATCHER): New.
8053 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
8054 filling *region also set *region_idx to (*region)->entry->index.
8057 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
8058 For REGs set ORIGINAL_REGNO.
8060 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
8062 * doc/md.texi: Mention that a target shouldn't implement
8063 vec_widen_(s|u)mul_even/odd pair if it is less efficient
8066 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8068 PR tree-optimization/59594
8069 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
8070 a copy of the datarefs vector rather than the vector itself.
8072 2014-01-28 Jason Merrill <jason@redhat.com>
8075 * dwarf2out.c (auto_die): New static.
8076 (gen_type_die_with_usage): Handle C++1y 'auto'.
8077 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
8080 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
8083 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
8084 (SPEC_X32): Likewise.
8085 (SPEC_64): Likewise.
8086 * config/i386/i386.c (ix86_option_override_internal): Turn off
8087 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
8089 (x86_file_start): Output .code16gcc for TARGET_16BIT.
8090 * config/i386/i386.h (TARGET_16BIT): New macro.
8091 (TARGET_16BIT_P): Likewise.
8092 * config/i386/i386.opt: Add m16.
8093 * doc/invoke.texi: Document -m16.
8095 2014-01-28 Jakub Jelinek <jakub@redhat.com>
8097 PR preprocessor/59935
8098 * input.c (location_get_source_line): Bail out on when line number
8099 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
8101 2014-01-28 Richard Biener <rguenther@suse.de>
8103 PR tree-optimization/58742
8104 * tree-ssa-forwprop.c (associate_plusminus): Handle
8105 pointer subtraction of the form (T)(P + A) - (T)P.
8107 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8109 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
8112 2014-01-28 Richard Biener <rguenther@suse.de>
8115 2014-01-28 Richard Biener <rguenther@suse.de>
8117 PR rtl-optimization/45364
8118 PR rtl-optimization/59890
8119 * var-tracking.c (local_get_addr_clear_given_value): Handle
8120 already cleared slot.
8121 (val_reset): Handle not allocated local_get_addr_cache.
8122 (vt_find_locations): Use post-order on the inverted CFG.
8124 2014-01-28 Richard Biener <rguenther@suse.de>
8126 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
8128 2014-01-28 Richard Biener <rguenther@suse.de>
8130 PR rtl-optimization/45364
8131 PR rtl-optimization/59890
8132 * var-tracking.c (local_get_addr_clear_given_value): Handle
8133 already cleared slot.
8134 (val_reset): Handle not allocated local_get_addr_cache.
8135 (vt_find_locations): Use post-order on the inverted CFG.
8137 2014-01-28 Alan Modra <amodra@gmail.com>
8139 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
8140 * configure.ac <recursive call for build != host>: Define
8141 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
8142 and LD_FOR_BUILD too.
8143 * configure: Regenerate.
8145 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
8147 * config/i386/i386.c (get_builtin_code_for_version): Separate
8148 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
8149 Broadwell from Haswell.
8151 2014-01-27 Steve Ellcey <sellcey@mips.com>
8153 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
8154 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
8155 * config/mips/mips.c (mips_option_override): Change setting
8157 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
8158 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
8159 Change from Mask to Var.
8161 2014-01-27 Jeff Law <law@redhat.com>
8163 * ipa-inline.c (inline_small_functions): Fix typo.
8165 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
8167 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
8168 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
8169 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
8170 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
8171 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
8172 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
8173 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
8174 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
8175 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
8176 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
8177 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
8178 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
8179 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
8180 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
8181 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
8182 (_mm512_storeu_epi64): Ditto.
8183 (_mm512_cmpge_epi32_mask): Ditto.
8184 (_mm512_cmpge_epu32_mask): Ditto.
8185 (_mm512_cmpge_epi64_mask): Ditto.
8186 (_mm512_cmpge_epu64_mask): Ditto.
8187 (_mm512_cmple_epi32_mask): Ditto.
8188 (_mm512_cmple_epu32_mask): Ditto.
8189 (_mm512_cmple_epi64_mask): Ditto.
8190 (_mm512_cmple_epu64_mask): Ditto.
8191 (_mm512_cmplt_epi32_mask): Ditto.
8192 (_mm512_cmplt_epu32_mask): Ditto.
8193 (_mm512_cmplt_epi64_mask): Ditto.
8194 (_mm512_cmplt_epu64_mask): Ditto.
8195 (_mm512_cmpneq_epi32_mask): Ditto.
8196 (_mm512_cmpneq_epu32_mask): Ditto.
8197 (_mm512_cmpneq_epi64_mask): Ditto.
8198 (_mm512_cmpneq_epu64_mask): Ditto.
8199 (_mm512_expand_pd): Ditto.
8200 (_mm512_expand_ps): Ditto.
8201 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
8202 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
8203 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
8204 * config/i386/i386.c (ix86_builtins): Add
8205 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
8206 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
8207 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
8208 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
8209 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
8210 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
8211 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
8212 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
8213 IX86_BUILTIN_PMOVUSQW512_MEM.
8214 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
8215 __builtin_ia32_pmovsqd512mem_mask,
8216 __builtin_ia32_pmovqd512mem_mask,
8217 __builtin_ia32_pmovusqw512mem_mask,
8218 __builtin_ia32_pmovsqw512mem_mask,
8219 __builtin_ia32_pmovqw512mem_mask,
8220 __builtin_ia32_pmovusdw512mem_mask,
8221 __builtin_ia32_pmovsdw512mem_mask,
8222 __builtin_ia32_pmovdw512mem_mask,
8223 __builtin_ia32_pmovqb512mem_mask,
8224 __builtin_ia32_pmovusqb512mem_mask,
8225 __builtin_ia32_pmovsqb512mem_mask,
8226 __builtin_ia32_pmovusdb512mem_mask,
8227 __builtin_ia32_pmovsdb512mem_mask,
8228 __builtin_ia32_pmovdb512mem_mask.
8229 (bdesc_args): Add __builtin_ia32_expanddf512,
8230 __builtin_ia32_expandsf512.
8231 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
8232 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
8233 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
8234 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
8235 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
8236 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
8237 (avx512f_<code>v8div16qi2_mask_store): This.
8238 (avx512f_expand<mode>): New.
8240 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
8242 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
8244 (_mm512_mask_prefetch_i64gather_pd): Ditto.
8245 (_mm512_prefetch_i32scatter_pd): Ditto.
8246 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
8247 (_mm512_prefetch_i64scatter_pd): Ditto.
8248 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
8249 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
8250 (_mm512_mask_prefetch_i64gather_ps): Ditto.
8251 (_mm512_prefetch_i32scatter_ps): Ditto.
8252 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
8253 (_mm512_prefetch_i64scatter_ps): Ditto.
8254 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
8255 * config/i386/i386-builtin-types.def: Define
8256 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
8257 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
8258 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
8259 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
8260 IX86_BUILTIN_SCATTERPFQPD.
8261 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
8262 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
8263 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
8264 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
8265 __builtin_ia32_scatterpfqps.
8266 (ix86_expand_builtin): Expand new built-ins.
8267 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8268 fix memory access data type.
8269 (*avx512pf_gatherpf<mode>_mask): Ditto.
8270 (*avx512pf_gatherpf<mode>): Ditto.
8271 (avx512pf_scatterpf<mode>): Ditto.
8272 (*avx512pf_scatterpf<mode>_mask): Ditto.
8273 (*avx512pf_scatterpf<mode>): Ditto.
8274 (GATHER_SCATTER_SF_MEM_MODE): New.
8275 (avx512pf_gatherpf<mode>df): Ditto.
8276 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8277 (*avx512pf_scatterpf<mode>df): Ditto.
8279 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8282 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8283 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8286 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8288 * common/config/arm/arm-common.c
8289 (arm_rewrite_mcpu): Handle multiple names.
8291 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8293 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8295 * gimple-builder.h (create_gimple_tmp): Delete.
8297 2014-01-27 Christian Bruel <christian.bruel@st.com>
8299 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
8302 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
8304 * config/pa/pa.md (call): Generate indirect long calls to non-local
8305 functions when outputing 32-bit code.
8306 (call_value): Likewise except for special call to buggy powf function.
8308 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
8309 portable runtime and PIC indirect calls.
8310 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
8311 and PIC call sequences. Use ldo instead of blr to set return register
8312 in PIC call sequence.
8314 2014-01-25 Walter Lee <walt@tilera.com>
8316 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
8317 avoid clobbering a live register.
8319 2014-01-25 Walter Lee <walt@tilera.com>
8321 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
8322 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
8323 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
8324 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
8326 2014-01-25 Walter Lee <walt@tilera.com>
8328 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
8329 arguments on even registers.
8330 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
8332 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
8333 (BIGGEST_ALIGNMENT): Ditto.
8334 (BIGGEST_FIELD_ALIGNMENT): Ditto.
8336 2014-01-25 Walter Lee <walt@tilera.com>
8338 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
8339 insns before bundling.
8340 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8342 2014-01-25 Walter Lee <walt@tilera.com>
8344 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
8345 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
8346 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
8348 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8350 * config/mips/constraints.md (kl): Delete.
8351 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
8352 define expands, using...
8353 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
8354 instructions for MIPS16.
8355 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
8356 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
8358 2014-01-25 Walter Lee <walt@tilera.com>
8360 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
8364 2014-01-25 Walter Lee <walt@tilera.com>
8366 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
8367 (TARGET_EXPAND_TO_RTL_HOOK): Define.
8369 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8371 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
8374 2014-01-25 Jakub Jelinek <jakub@redhat.com>
8376 * print-rtl.c (in_call_function_usage): New var.
8377 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
8378 EXPR_LIST mode as mode and not as reg note name.
8381 * cfgloopmanip.c (copy_loop_info): If
8382 loop->warned_aggressive_loop_optimizations, make sure
8383 the flag is set in target loop too.
8385 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
8387 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
8389 * builtins.def: Likewise.
8390 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
8391 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
8392 * ira.c (ira_setup_eliminable_regset): Likewise.
8393 * omp-low.c (gate_expand_omp): Likewise.
8394 (execute_lower_omp): Likewise.
8395 (diagnose_sb_0): Likewise.
8396 (gate_diagnose_omp_blocks): Likewise.
8397 (simd_clone_clauses_extract): Likewise.
8400 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8402 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
8403 correction for little endian...
8404 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
8407 2014-01-24 Jeff Law <law@redhat.com>
8409 PR tree-optimization/59919
8410 * tree-vrp.c (find_assert_locations_1): Do not register asserts
8411 for non-returning calls.
8413 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
8415 * common/config/aarch64/aarch64-common.c
8416 (aarch64_rewrite_mcpu): Handle multiple names.
8417 * config/aarch64/aarch64.h
8418 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8420 2014-01-24 Dodji Seketeli <dodji@redhat.com>
8422 * input.c (add_file_to_cache_tab): Handle the case where fopen
8425 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
8428 * config/i386/i386.md (pushsf splitter): Get stack adjustment
8429 from push operand if code of push isn't PRE_DEC.
8431 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8434 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
8435 -mquad-memory-atomic. Update -mquad-memory documentation to say
8436 it is only used for non-atomic loads/stores.
8438 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
8439 -mquad-memory or -mquad-memory-atomic switches.
8441 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
8442 -mquad-memory-atomic to ISA 2.07 support.
8444 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
8445 to separate support of normal quad word memory operations (ldq, stq)
8446 from the atomic quad word memory operations.
8448 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
8449 support to separate non-atomic quad word operations from atomic
8450 quad word operations. Disable non-atomic quad word operations in
8451 little endian mode so that we don't have to swap words after the
8452 load and before the store.
8453 (quad_load_store_p): Add comment about atomic quad word support.
8454 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
8455 options printed with -mdebug=reg.
8457 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
8458 -mquad-memory-atomic as the test for whether we have quad word
8459 atomic instructions.
8460 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
8461 or -mp8-vector are used, allow byte/half-word atomic operations.
8463 * config/rs6000/sync.md (load_lockedti): Insure that the address
8464 is a proper indexed or indirect address for the lqarx instruction.
8465 On little endian systems, swap the hi/lo registers after the lqarx
8467 (load_lockedpti): Use indexed_or_indirect_operand predicate to
8468 insure the address is valid for the lqarx instruction.
8469 (store_conditionalti): Insure that the address is a proper indexed
8470 or indirect address for the stqcrx. instruction. On little endian
8471 systems, swap the hi/lo registers before doing the stqcrx.
8473 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
8474 insure the address is valid for the stqcrx. instruction.
8476 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8477 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
8478 type of quad memory support is available.
8480 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
8483 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
8484 there is a danger of looping.
8486 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
8488 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8489 force flag_ira_loop_pressure if set via command line.
8491 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8493 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
8494 (ashr_simd): New builtin handling DI mode.
8495 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
8496 (aarch64_sshr_simddi): New match pattern.
8497 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
8498 (vshrd_n_s64): Likewise.
8499 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
8501 2014-01-23 Nick Clifton <nickc@redhat.com>
8503 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
8504 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
8505 favour of mcu specific scripts.
8506 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
8509 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8510 Alex Velenko <Alex.Velenko@arm.com>
8512 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
8513 (vaddv_s16): Likewise.
8514 (vaddv_s32): Likewise.
8515 (vaddv_u8): Likewise.
8516 (vaddv_u16): Likewise.
8517 (vaddv_u32): Likewise.
8518 (vaddvq_s8): Likewise.
8519 (vaddvq_s16): Likewise.
8520 (vaddvq_s32): Likewise.
8521 (vaddvq_s64): Likewise.
8522 (vaddvq_u8): Likewise.
8523 (vaddvq_u16): Likewise.
8524 (vaddvq_u32): Likewise.
8525 (vaddvq_u64): Likewise.
8526 (vaddv_f32): Likewise.
8527 (vaddvq_f32): Likewise.
8528 (vaddvq_f64): Likewise.
8529 (vmaxv_f32): Likewise.
8530 (vmaxv_s8): Likewise.
8531 (vmaxv_s16): Likewise.
8532 (vmaxv_s32): Likewise.
8533 (vmaxv_u8): Likewise.
8534 (vmaxv_u16): Likewise.
8535 (vmaxv_u32): Likewise.
8536 (vmaxvq_f32): Likewise.
8537 (vmaxvq_f64): Likewise.
8538 (vmaxvq_s8): Likewise.
8539 (vmaxvq_s16): Likewise.
8540 (vmaxvq_s32): Likewise.
8541 (vmaxvq_u8): Likewise.
8542 (vmaxvq_u16): Likewise.
8543 (vmaxvq_u32): Likewise.
8544 (vmaxnmv_f32): Likewise.
8545 (vmaxnmvq_f32): Likewise.
8546 (vmaxnmvq_f64): Likewise.
8547 (vminv_f32): Likewise.
8548 (vminv_s8): Likewise.
8549 (vminv_s16): Likewise.
8550 (vminv_s32): Likewise.
8551 (vminv_u8): Likewise.
8552 (vminv_u16): Likewise.
8553 (vminv_u32): Likewise.
8554 (vminvq_f32): Likewise.
8555 (vminvq_f64): Likewise.
8556 (vminvq_s8): Likewise.
8557 (vminvq_s16): Likewise.
8558 (vminvq_s32): Likewise.
8559 (vminvq_u8): Likewise.
8560 (vminvq_u16): Likewise.
8561 (vminvq_u32): Likewise.
8562 (vminnmv_f32): Likewise.
8563 (vminnmvq_f32): Likewise.
8564 (vminnmvq_f64): Likewise.
8566 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8568 * config/aarch64/aarch64-simd.md
8569 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
8570 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
8571 (*aarch64_mul3_elt<mode>): Likewise.
8572 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
8573 (*aarch64_mul3_elt_to_64v2df): Likewise.
8574 (*aarch64_mla_elt<mode>): Likewise.
8575 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8576 (*aarch64_mls_elt<mode>): Likewise.
8577 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8578 (*aarch64_fma4_elt<mode>): Likewise.
8579 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8580 (*aarch64_fma4_elt_to_64v2df): Likewise.
8581 (*aarch64_fnma4_elt<mode>): Likewise.
8582 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8583 (*aarch64_fnma4_elt_to_64v2df): Likewise.
8584 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
8585 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
8586 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8587 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8588 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
8589 (aarch64_sqdmull_lane<mode>_internal): Likewise.
8590 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
8592 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
8594 * config/aarch64/aarch64-simd.md
8595 (aarch64_be_checked_get_lane<mode>): New define_expand.
8596 * config/aarch64/aarch64-simd-builtins.def
8597 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
8598 New builtin definition.
8599 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
8600 Use new safe be builtin.
8602 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8604 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
8606 (aarch64_be_st1<mode>): Likewise.
8607 (aarch_ld1<VALL:mode>): Define_expand modified.
8608 (aarch_st1<VALL:mode>): Likewise.
8609 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
8610 (UNSPEC_ST1): Likewise.
8612 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
8614 * config/microblaze/microblaze.md: Add trap insn and attribute
8616 2014-01-23 Dodji Seketeli <dodji@redhat.com>
8618 PR preprocessor/58580
8619 * input.h (location_get_source_line): Take an additional line_size
8621 (void diagnostics_file_cache_fini): Declare new function.
8622 * input.c (struct fcache): New type.
8623 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
8624 New static constants.
8625 (diagnostic_file_cache_init, total_lines_num)
8626 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
8627 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
8628 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
8629 (get_next_line, read_next_line, goto_next_line, read_line_num):
8630 New static function definitions.
8631 (diagnostic_file_cache_fini): New function.
8632 (location_get_source_line): Take an additional output line_len
8633 parameter. Re-write using lookup_or_add_file_to_cache_tab and
8635 * diagnostic.c (diagnostic_finish): Call
8636 diagnostic_file_cache_fini.
8637 (adjust_line): Take an additional input parameter for the length
8638 of the line, rather than calculating it with strlen.
8639 (diagnostic_show_locus): Adjust the use of
8640 location_get_source_line and adjust_line with respect to their new
8641 signature. While displaying a line now, do not stop at the first
8642 null byte. Rather, display the zero byte as a space and keep
8643 going until we reach the size of the line.
8644 * Makefile.in: Add vec.o to OBJS-libcommon
8646 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8647 Ilya Tocar <ilya.tocar@intel.com>
8649 * config/i386/avx512fintrin.h (_mm512_kmov): New.
8650 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
8651 (__builtin_ia32_kmov16): Ditto.
8652 * config/i386/i386.md (UNSPEC_KMOV): New.
8655 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8657 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
8658 (_mm512_storeu_si512): Ditto.
8660 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
8663 * rtl.h (get_referenced_operands): Declare.
8664 * recog.c (get_referenced_operands): New function.
8665 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
8666 operands have been referenced when recording LO_SUM references.
8668 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
8670 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
8672 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8674 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
8675 Enable for generic and recent AMD targets.
8677 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8679 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
8680 ARG_SIZE note when adjustment was eliminated.
8682 2014-01-22 Jeff Law <law@redhat.com>
8684 PR tree-optimization/59597
8685 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
8686 in file. Accept new argument REGISTERING and use it to modify
8687 dump output appropriately.
8688 (register_jump_thread): Corresponding changes.
8689 (mark_threaded_blocks): Reinstate code to cancel unprofitable
8690 thread paths involving joiner blocks. Add code to dump cancelled
8691 jump threading paths.
8693 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
8695 PR rtl-optimization/59477
8696 * lra-constraints.c (inherit_in_ebb): Process call for living hard
8697 regs. Update reloads_num and potential_reload_hard_regs for all insns.
8699 2014-01-22 Tom Tromey <tromey@redhat.com>
8701 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
8703 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
8705 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8707 PR rtl-optimization/59896
8708 * lra-constraints.c (process_alt_operands): Check unused note for
8709 matched operands of insn with no output reloads.
8711 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
8713 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
8714 (mips_move_from_gpr_cost): Likewise.
8716 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8718 PR rtl-optimization/59858
8719 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
8720 ira_class_hard_regs_num.
8721 (process_alt_operands): Increase reject for dying matched operand.
8723 2014-01-21 Jakub Jelinek <jakub@redhat.com>
8726 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
8727 smaller than size, perform several stores or loads and stores
8728 at dst + count - size to store or copy all of size bytes, rather
8729 than just last modesize bytes.
8731 2014-01-20 DJ Delorie <dj@redhat.com>
8733 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
8734 that CLOBBERs are REGs before propogating their values.
8736 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
8739 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
8740 (cgraph_inline_failed_type): New function.
8741 * cgraph.h (DEFCIFCODE): Add type.
8742 (cgraph_inline_failed_type_t): New enum.
8743 (cgraph_inline_failed_type): New prototype.
8744 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
8745 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
8746 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
8747 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
8748 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
8749 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
8750 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
8751 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
8752 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
8753 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
8754 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
8755 OPTIMIZATION_MISMATCH.
8756 * tree-inline.c (expand_call_inline): Emit errors during
8757 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
8759 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8762 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
8763 mode attribute in insn output.
8765 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
8767 * output.h (output_constant): Delete.
8768 * varasm.c (output_constant): Make private.
8770 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
8772 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
8774 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8777 * tree.h (fold_builtin_strcat): New prototype.
8778 * builtins.c (fold_builtin_strcat): No longer static. Add len
8779 argument, if non-NULL, don't call c_strlen. Optimize
8780 directly into __builtin_memcpy instead of __builtin_strcpy.
8781 (fold_builtin_2): Adjust fold_builtin_strcat caller.
8782 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
8784 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8786 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8787 for SImode_address_operand operands, having only a REG argument.
8789 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
8791 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
8792 loader name using mbig-endian.
8793 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
8795 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8797 * doc/invoke.texi (-march): Clarify documentation for AArch64.
8801 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
8803 * config/aarch64/aarch64-protos.h
8804 (aarch64_cannot_change_mode_class_ptr): Declare.
8805 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
8806 aarch64_cannot_change_mode_class_ptr): New.
8807 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
8808 backend hook aarch64_cannot_change_mode_class.
8810 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8812 * common/config/aarch64/aarch64-common.c
8813 (aarch64_handle_option): Don't handle any option order logic here.
8814 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
8815 selected_cpu, warn on architecture version mismatch.
8816 (aarch64_override_options): Fix parsing order for option strings.
8818 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8819 Iain Sandoe <iain@codesourcery.com>
8822 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
8823 warning. Amend comment to reflect current functionality.
8825 2014-01-20 Richard Biener <rguenther@suse.de>
8828 * builtins.c (fold_builtin_strcat): Remove case better handled
8829 by tree-ssa-strlen.c.
8831 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
8833 * config/aarch64/aarch64.opt
8834 (mcpu, march, mtune): Make case-insensitive.
8836 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8839 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8840 if operands[1] is a REG or ZERO_EXTEND of a REG.
8842 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
8844 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
8846 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
8848 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
8849 long non-pic millicode calls.
8851 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8853 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
8855 2014-01-19 Kito Cheng <kito@0xlab.org>
8857 * builtins.c (expand_movstr): Check movstr expand done or fail.
8859 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8860 H.J. Lu <hongjiu.lu@intel.com>
8863 * config/i386/i386.md (*lea<mode>): Zero-extend return register
8864 to DImode for zero-extended addresses.
8866 2014-01-19 Jakub Jelinek <jakub@redhat.com>
8868 PR rtl-optimization/57763
8869 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
8870 on the new indirect jump_insn and increment LABEL_NUSES (label).
8872 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
8876 * config.gcc (x86_archs): New variable.
8877 (x86_64_archs): Likewise.
8878 (x86_cpus): Likewise.
8879 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
8880 --with-arch/--with-cpu= options.
8881 Support --with-arch=/--with-cpu={nehalem,westmere,
8882 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
8884 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8886 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
8887 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
8889 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8891 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
8893 2014-01-18 Jakub Jelinek <jakub@redhat.com>
8896 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
8897 clear cpp_get_options (parse_in)->warn_unused_macros for
8898 ix86_target_macros_internal with cpp_define.
8900 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
8902 * jump.c (delete_related_insns): Keep (use (insn))s.
8903 * reorg.c (redundant_insn): Check for barriers too.
8905 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8907 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
8909 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
8911 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
8912 call to $$dyncall when TARGET_LONG_CALLS is true.
8914 2014-01-17 Jeff Law <law@redhat.com>
8916 * ree.c (combine_set_extension): Temporarily disable test for
8917 changing number of hard registers.
8919 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
8922 * ipa-inline-analysis.c (inline_free_summary):
8923 Do not free summary of aliases.
8925 2014-01-17 Jakub Jelinek <jakub@redhat.com>
8928 * gimplify.c (gimplify_expr): Use create_tmp_var
8929 instead of create_tmp_var_raw. If cond doesn't have
8930 integral type, don't add the IFN_ANNOTATE builtin at all.
8932 2014-01-17 Martin Jambor <mjambor@suse.cz>
8935 * ipa-cp.c (prev_edge_clone): New variable.
8936 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
8937 Also resize prev_edge_clone vector.
8938 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
8939 (ipcp_edge_removal_hook): New function.
8940 (ipcp_driver): Register ipcp_edge_removal_hook.
8942 2014-01-17 Andrew Pinski <apinski@cavium.com>
8943 Steve Ellcey <sellcey@mips.com>
8946 * config/mips/mips.c (mips_print_operand): Check operand mode instead
8949 2014-01-17 Jeff Law <law@redhat.com>
8952 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
8953 so that pass_ccp runs first.
8955 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8957 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
8958 (ix86_adjust_cost): Use !TARGET_XXX.
8959 (do_reorder_for_imul): Likewise.
8960 (swap_top_of_ready_list): Likewise.
8961 (ix86_sched_reorder): Likewise.
8963 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8965 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
8966 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
8967 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
8968 (intel_memset): New. Duplicate slm_memset.
8969 (intel_cost): New. Duplicate slm_cost.
8970 (m_INTEL): New macro.
8971 (processor_target_table): Add "intel".
8972 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
8973 with PROCESSOR_INTEL for "intel".
8974 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
8975 PROCESSOR_SILVERMONT.
8976 (ix86_issue_rate): Likewise.
8977 (ix86_adjust_cost): Likewise.
8978 (ia32_multipass_dfa_lookahead): Likewise.
8979 (swap_top_of_ready_list): Likewise.
8980 (ix86_sched_reorder): Likewise.
8981 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
8982 instead of TARGET_OPT_AGU.
8983 * config/i386/i386.h (TARGET_INTEL): New.
8984 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
8985 (processor_type): Add PROCESSOR_INTEL.
8986 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
8987 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
8989 2014-01-17 Marek Polacek <polacek@redhat.com>
8992 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
8995 2014-01-17 Richard Biener <rguenther@suse.de>
8997 PR tree-optimization/46590
8998 * opts.c (default_options_table): Add entries for
8999 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
9000 all enabled at -O1 but not for -Og.
9001 * common.opt (fbranch-count-reg): Remove Init(1).
9002 (fmove-loop-invariants): Likewise.
9003 (ftree-pta): Likewise.
9005 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9007 * config/i386/i386.c (ix86_data_alignment): For compatibility with
9008 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
9009 decls to at least the GCC 4.8 used alignments.
9012 * tree-nested.c (convert_nonlocal_reference_stmt,
9013 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
9014 of GIMPLE_BIND stmts, adjust associated decls.
9016 2014-01-17 Richard Biener <rguenther@suse.de>
9018 PR tree-optimization/46590
9019 * vec.h (vec<>::bseach): New member function implementing
9020 binary search according to C89 bsearch.
9021 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
9022 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
9023 bitmap pointer again. Make accesses_in_loop a flat array.
9024 (mem_ref_obstack): New global.
9025 (outermost_indep_loop): Adjust for mem_ref->stored changes.
9026 (mark_ref_stored): Likewise.
9027 (ref_indep_loop_p_2): Likewise.
9028 (set_ref_stored_in_loop): New helper function.
9029 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
9030 (memref_free): Adjust.
9031 (record_mem_ref_loc): Simplify.
9032 (gather_mem_refs_stmt): Adjust.
9033 (sort_locs_in_loop_postorder_cmp): New function.
9034 (analyze_memory_references): Sort accesses_in_loop after
9035 loop postorder number.
9036 (find_ref_loc_in_loop_cmp): New function.
9037 (for_all_locs_in_loop): Find relevant cluster of locs in
9038 accesses_in_loop and iterate without recursion.
9039 (execute_sm): Avoid uninit warning.
9040 (struct ref_always_accessed): Simplify.
9041 (ref_always_accessed::operator ()): Likewise.
9042 (ref_always_accessed_p): Likewise.
9043 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
9044 loop postorder numbers here.
9045 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
9048 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9051 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
9052 on decls for which assemble_alias has been called.
9054 2014-01-17 Nick Clifton <nickc@redhat.com>
9056 * config/msp430/msp430.opt: (mcpu): New option.
9057 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
9058 (msp430_option_override): Parse target_cpu. If the MCU name
9059 matches a generic string, clear target_mcu.
9060 (msp430_attr): Allow numeric interrupt values up to 63.
9061 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
9062 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
9064 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
9066 * config/msp430/msp430.md (popm): Use %J rather than %I.
9067 (addsi3): Use msp430_nonimmediate_operand for operand 2.
9068 (addhi_cy_i): Use immediate_operand for operand 2.
9069 * doc/invoke.texi: Document -mcpu option.
9071 2014-01-17 Richard Biener <rguenther@suse.de>
9073 PR rtl-optimization/38518
9074 * df.h (df_analyze_loop): Declare.
9075 * df-core.c: Include cfgloop.h.
9076 (df_analyze_1): Split out main part of df_analyze.
9077 (df_analyze): Adjust.
9078 (loop_inverted_post_order_compute): New function.
9079 (loop_post_order_compute): Likewise.
9080 (df_analyze_loop): New function avoiding whole-function
9082 * loop-invariant.c (find_defs): Use df_analyze_loop.
9083 (find_invariants): Adjust.
9084 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
9086 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
9088 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
9089 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
9091 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
9093 * ipa-ref.c (ipa_remove_stmt_references): Fix references
9094 traversal when removing references.
9096 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
9099 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
9101 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
9104 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
9105 pushing a reload for an autoinc when we had previously reloaded an
9106 inner part of the address.
9108 2014-01-16 Jakub Jelinek <jakub@redhat.com>
9110 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
9112 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
9113 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
9114 when not giving up or versioning for alias only because of
9116 (vect_analyze_data_ref_dependences): Set to true.
9117 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
9119 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
9120 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
9124 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
9127 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
9128 operand 0 predicate for gathers, use a new pseudo as subtarget.
9130 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9133 * lra-constraints.c (process_alt_operands): Add printing debug info.
9134 Check absence of input/output reloads for matched operands too.
9136 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9138 PR rtl-optimization/59835
9139 * ira.c (ira_init_register_move_cost): Increase cost for
9142 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
9144 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
9146 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
9149 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
9150 non-register objects. Use gen_(high/low)part more consistently.
9153 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9156 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
9157 endian support, remove tests for WORDS_BIG_ENDIAN.
9158 (p8_mfvsrd_3_<mode>): Likewise.
9159 (reload_gpr_from_vsx<mode>): Likewise.
9160 (reload_gpr_from_vsxsf): Likewise.
9161 (p8_mfvsrd_4_disf): Likewise.
9163 2014-01-16 Richard Biener <rguenther@suse.de>
9165 PR rtl-optimization/46590
9166 * lcm.c (compute_antinout_edge): Use postorder iteration.
9167 (compute_laterin): Use inverted postorder iteration.
9169 2014-01-16 Nick Clifton <nickc@redhat.com>
9172 * varasm.c (output_constant): Return the number of bytes actually
9174 (output_constructor_array_range): Update the field size with the
9175 number of bytes emitted by output_constant.
9176 (output_constructor_regular_field): Likewise. Also do not
9177 complain if the total number of bytes emitted is now greater
9178 than the expected fieldpos.
9179 * output.h (output_constant): Update prototype and descriptive comment.
9181 2014-01-16 Marek Polacek <polacek@redhat.com>
9184 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
9185 it is error_mark_node.
9187 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
9189 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
9190 VALID_AVX256_REG_OR_OI_MODE.
9192 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
9194 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
9195 current procedure should be profiled.
9197 2014-01-15 Andrew Pinski <apinski@cavium.com>
9199 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
9200 of moving from/to the STACK_REG register class.
9202 2014-01-15 Richard Henderson <rth@redhat.com>
9205 * reginfo.c (global_regs_decl): Globalize.
9206 * rtl.h (global_regs_decl): Declare.
9207 * ira.c (do_reload): Diagnose frame_pointer_needed and it
9208 reserved via global_regs.
9210 2014-01-15 Teresa Johnson <tejohnson@google.com>
9212 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
9214 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9216 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
9217 and vmulosh rather than call gen_vec_widen_smult_*.
9218 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
9219 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
9220 (vec_widen_smult_even_v16qi): Likewise.
9221 (vec_widen_umult_even_v8hi): Likewise.
9222 (vec_widen_smult_even_v8hi): Likewise.
9223 (vec_widen_umult_odd_v16qi): Likewise.
9224 (vec_widen_smult_odd_v16qi): Likewise.
9225 (vec_widen_umult_odd_v8hi): Likewise.
9226 (vec_widen_smult_odd_v8hi): Likewise.
9227 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
9228 vmuloub rather than call gen_vec_widen_umult_*.
9229 (vec_widen_umult_lo_v16qi): Likewise.
9230 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
9231 vmulosb rather than call gen_vec_widen_smult_*.
9232 (vec_widen_smult_lo_v16qi): Likewise.
9233 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
9234 rather than call gen_vec_widen_umult_*.
9235 (vec_widen_umult_lo_v8hi): Likewise.
9236 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
9237 rather than call gen_vec_widen_smult_*.
9238 (vec_widen_smult_lo_v8hi): Likewise.
9240 2014-01-15 Jeff Law <law@redhat.com>
9242 PR tree-optimization/59747
9243 * ree.c (find_and_remove_re): Properly handle case where a second
9244 eliminated extension requires widening a copy created for elimination
9245 of a prior extension.
9246 (combine_set_extension): Ensure that the number of hard regs needed
9247 for a destination register does not change when we widen it.
9249 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
9251 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
9252 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
9253 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
9254 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
9255 (avr-*-rtems*): Likewise.
9256 (bfin*-rtems*): Likewise.
9257 (moxie-*-rtems*): Likewise.
9258 (h8300-*-rtems*): Likewise.
9259 (i[34567]86-*-rtems*): Likewise.
9260 (lm32-*-rtems*): Likewise.
9261 (m32r-*-rtems*): Likewise.
9262 (m68k-*-rtems*): Likewise.
9263 (microblaze*-*-rtems*): Likewise.
9264 (mips*-*-rtems*): Likewise.
9265 (powerpc-*-rtems*): Likewise.
9266 (sh-*-rtems*): Likewise.
9267 (sparc-*-rtems*): Likewise.
9268 (sparc64-*-rtems*): Likewise.
9269 (v850-*-rtems*): Likewise.
9270 (m32c-*-rtems*): Likewise.
9272 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9274 PR rtl-optimization/59511
9275 * ira.c (ira_init_register_move_cost): Use memory costs for some
9276 cases of register move cost calculations.
9277 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9278 instead of BB frequency.
9279 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9280 * lra-assigns.c (find_hard_regno_for): Ditto.
9282 2014-01-15 Richard Biener <rguenther@suse.de>
9284 PR tree-optimization/59822
9285 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9286 (vectorizable_load): Use it to hoist defs of uses of invariant
9287 loads out of the loop.
9289 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9290 Kugan Vivekanandarajah <kuganv@linaro.org>
9293 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9296 2014-01-15 Richard Biener <rguenther@suse.de>
9298 PR rtl-optimization/59802
9299 * lcm.c (compute_available): Use inverted postorder to seed
9300 the initial worklist.
9302 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9305 * config/s390/s390.c (s390_preferred_reload_class): Don't return
9306 ADDR_REGS for invalid symrefs in non-PIC code.
9308 2014-01-15 Jakub Jelinek <jakub@redhat.com>
9311 * builtins.c (determine_block_size): Initialize *probable_max_size
9312 even if len_rtx is CONST_INT.
9314 2014-01-14 Andrew Pinski <apinski@cavium.com>
9316 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
9317 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
9318 (cortexa53_tunings): Likewise.
9319 (aarch64_sched_issue_rate): New function.
9320 (TARGET_SCHED_ISSUE_RATE): Define.
9322 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9324 * ira-costs.c (find_costs_and_classes): Add missed
9325 ira_init_register_move_cost_if_necessary.
9327 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9330 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
9332 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
9335 * config/i386/i386.c (type_natural_mode): Add a bool parameter
9336 to indicate if type is used for function return value. Warn ABI
9337 change if the vector mode isn't available for function return value.
9338 (ix86_function_arg_advance): Pass false to type_natural_mode.
9339 (ix86_function_arg): Likewise.
9340 (ix86_gimplify_va_arg): Likewise.
9341 (function_arg_32): Don't warn ABI change.
9342 (ix86_function_value): Pass true to type_natural_mode.
9343 (ix86_return_in_memory): Likewise.
9344 (ix86_struct_value_rtx): Removed.
9345 (TARGET_STRUCT_VALUE_RTX): Likewise.
9347 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9349 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
9350 converting a conditional jump into a conditional return.
9352 2014-01-14 Richard Biener <rguenther@suse.de>
9354 PR tree-optimization/58921
9355 PR tree-optimization/59006
9356 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
9357 hoisting invariant stmts.
9358 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
9359 invariant loads on the preheader edge if possible.
9361 2014-01-14 Joey Ye <joey.ye@arm.com>
9363 * doc/plugin.texi (Building GCC plugins): Update to C++.
9365 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
9367 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
9368 (_mm_rcp28_round_ss): Ditto.
9369 (_mm_rsqrt28_round_sd): Ditto.
9370 (_mm_rsqrt28_round_ss): Ditto.
9371 (_mm_rcp28_sd): Ditto.
9372 (_mm_rcp28_ss): Ditto.
9373 (_mm_rsqrt28_sd): Ditto.
9374 (_mm_rsqrt28_ss): Ditto.
9375 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
9376 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
9377 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
9378 (IX86_BUILTIN_RCP28SD): Ditto.
9379 (IX86_BUILTIN_RCP28SS): Ditto.
9380 (IX86_BUILTIN_RSQRT28SD): Ditto.
9381 (IX86_BUILTIN_RSQRT28SS): Ditto.
9382 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
9383 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
9384 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
9385 (ix86_expand_special_args_builtin): Expand new FTYPE.
9386 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
9387 (srcp14<mode>): Make insn unary.
9388 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
9389 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
9390 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
9391 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
9392 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
9393 Fix rounding: make it SAE only.
9394 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9396 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9398 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
9399 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
9400 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
9401 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
9402 (round_saeonly_mask_scalar_operand4): Ditto.
9403 (round_saeonly_mask_scalar_op3): Ditto.
9404 (round_saeonly_mask_scalar_op4): Ditto.
9406 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9408 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9409 Implement -maltivec=be for vec_insert and vec_extract.
9411 2014-01-10 DJ Delorie <dj@redhat.com>
9413 * config/msp430/msp430.md (call_internal): Don't allow memory
9414 references with SP as the base register.
9415 (call_value_internal): Likewise.
9416 * config/msp430/constraints.md (Yc): New. For memory references
9417 that don't use SP as a base register.
9419 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
9420 "an integer without a # prefix"
9421 * config/msp430/msp430.md (epilogue_helper): Use it.
9423 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9426 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
9427 AVX512F gather builtins.
9428 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
9429 on gather decls with INTEGER_TYPE masktype.
9430 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
9431 directly into the builtin rather than hoisting it before loop.
9433 PR tree-optimization/59387
9434 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
9435 (scev_const_prop): If folded_casts and type has undefined overflow,
9436 use force_gimple_operand instead of force_gimple_operand_gsi and
9437 for each added stmt if it is assign with
9438 arith_code_with_undefined_signed_overflow, call
9439 rewrite_to_defined_overflow.
9440 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
9441 gimple-fold.h instead.
9442 (arith_code_with_undefined_signed_overflow,
9443 rewrite_to_defined_overflow): Moved to ...
9444 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
9445 rewrite_to_defined_overflow): ... here. No longer static.
9446 Include gimplify-me.h.
9447 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
9448 rewrite_to_defined_overflow): New prototypes.
9450 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9452 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
9454 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
9456 * builtins.c (get_object_alignment_2): Minor tweak.
9457 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
9459 2014-01-13 Christian Bruel <christian.bruel@st.com>
9461 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
9462 optimized non constant lengths.
9464 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9467 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
9468 load as __atomic_load_N if possible.
9470 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9472 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
9474 (rs6000_expand_builtin): Adjust call.
9476 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9479 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
9480 * config/rs6000/rs6000.c: Include target-globals.h.
9481 (rs6000_set_current_function): Instead of doing target_reinit
9482 unconditionally, use save_target_globals_default_opts and
9483 restore_target_globals.
9485 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
9487 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
9488 (rs6000_expand_builtin): Handle mffs and mtfsf.
9489 (rs6000_init_builtins): Define mffs and mtfsf.
9490 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
9491 (rs6000_mffs): New pattern.
9492 (rs6000_mtfsf): New pattern.
9494 2014-01-11 Bin Cheng <bin.cheng@arm.com>
9496 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
9497 Start narrowing with START. Apply candidate-use pair
9498 and check overall cost in narrowing.
9499 (iv_ca_prune): Pass new argument.
9501 2014-01-10 Jeff Law <law@redhat.com>
9504 * ree.c (combine_reaching_defs): Ensure the defining statement
9505 occurs before the extension when optimizing extensions with
9506 different source and destination hard registers.
9508 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9511 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
9512 vtables into the type inheritance graph.
9514 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9516 PR rtl-optimization/59754
9517 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
9518 modes in the REGNO != REGNO case.
9520 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9522 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
9524 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9526 PR tree-optimization/59745
9527 * tree-predcom.c (tree_predictive_commoning_loop): Call
9528 free_affine_expand_cache if giving up because components is NULL.
9530 * target-globals.c (save_target_globals): Allocate < 4KB structs using
9531 GC in payload of target_globals struct instead of allocating them on
9532 the heap and the larger structs separately using GC.
9533 * target-globals.h (struct target_globals): Make regs, hard_regs,
9534 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
9535 of GTY((skip)) and change type to void *.
9536 (reset_target_globals): Cast loads from those fields to corresponding
9539 2014-01-10 Steve Ellcey <sellcey@mips.com>
9542 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
9543 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
9544 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
9546 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
9549 * aarch64-modes.def (CC_Zmode): New flags mode.
9550 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
9551 represents an equality.
9552 (aarch64_get_condition_code): Handle CC_Zmode.
9553 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
9555 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9557 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
9558 extraction in good case.
9560 2014-01-10 Richard Biener <rguenther@suse.de>
9562 PR tree-optimization/59374
9563 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
9564 checking after SLP discovery. Mark stmts not participating
9565 in any SLP instance properly.
9567 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9569 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
9570 when handling a SET rtx.
9572 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9574 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
9575 (cortex-a57): Likewise.
9576 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
9578 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9580 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
9581 non-iwmmxt builtins.
9583 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9587 * ipa-devirt.c record_target_from_binfo): Take as argument
9588 stack of binfos and lookup matching one for virtual inheritance.
9589 (possible_polymorphic_call_targets_1): Update.
9591 2014-01-10 Huacai Chen <chenhc@lemote.com>
9593 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
9594 kernel strings for Loongson-2E/2F/3A.
9596 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9599 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
9600 is_gimple_call before calling gimple_call_internal_p.
9602 2014-01-09 Steve Ellcey <sellcey@mips.com>
9604 * Makefile.in (TREE_FLOW_H): Remove.
9605 (TREE_SSA_H): Add file names from tree-flow.h.
9606 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
9607 * tree.h: Remove tree-flow.h reference.
9608 * hash-table.h: Remove tree-flow.h reference.
9609 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
9610 reference with tree-ssa-loop.h.
9612 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9614 * doc/invoke.texi: Add -maltivec={be,le} options, and document
9615 default element-order behavior for -maltivec.
9616 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
9617 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
9618 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
9619 when targeting big endian, at least for now.
9620 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
9622 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9625 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
9626 var satisfies use_register_for_decl, just take into account type
9627 alignment, rather than decl alignment.
9629 PR tree-optimization/59622
9630 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
9631 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
9632 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
9633 Don't devirtualize for inplace at all. For targets.length () == 1,
9634 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
9636 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9638 * config/i386/i386.md (cpu): Remove the unused btver1.
9640 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9642 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
9644 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9647 * tree-core.h (struct target_globals): New forward declaration.
9648 (struct tree_target_option): Add globals field.
9649 * tree.h (TREE_TARGET_GLOBALS): Define.
9650 (prepare_target_option_nodes_for_pch): New prototype.
9651 * target-globals.h (struct target_globals): Define even if
9653 * tree.c (prepare_target_option_node_for_pch,
9654 prepare_target_option_nodes_for_pch): New functions.
9655 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
9656 * config/i386/i386.c: Include target-globals.h.
9657 (ix86_set_current_function): Instead of doing target_reinit
9658 unconditionally, use save_target_globals_default_opts and
9659 restore_target_globals.
9661 2014-01-09 Richard Biener <rguenther@suse.de>
9663 PR tree-optimization/59715
9664 * tree-cfg.h (split_critical_edges): Declare.
9665 * tree-cfg.c (split_critical_edges): Export.
9666 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
9668 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
9670 * cfgexpand.c (expand_stack_vars): Optionally disable
9671 asan stack protection.
9672 (expand_used_vars): Likewise.
9673 (partition_stack_vars): Likewise.
9674 * asan.c (asan_emit_stack_protection): Optionally disable
9675 after return stack usage.
9676 (instrument_derefs): Optionally disable memory access instrumentation.
9677 (instrument_builtin_call): Likewise.
9678 (instrument_strlen_call): Likewise.
9679 (asan_protect_global): Optionally disable global variables protection.
9680 * doc/invoke.texi: Added doc for new options.
9681 * params.def: Added new options.
9682 * params.h: Likewise.
9684 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9686 PR rtl-optimization/59724
9687 * ifcvt.c (cond_exec_process_if_block): Don't call
9688 flow_find_head_matching_sequence with 0 longest_match.
9689 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
9690 non-active insns if !stop_after.
9691 (try_head_merge_bb): Revert 2014-01-07 changes.
9693 2014-01-08 Jeff Law <law@redhat.com>
9695 * ree.c (get_sub_rtx): New function, extracted from...
9696 (merge_def_and_ext): Here.
9697 (combine_reaching_defs): Use get_sub_rtx.
9699 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
9701 * cgraph.h (varpool_variable_node): Do not choke on null node.
9703 2014-01-08 Catherine Moore <clm@codesourcery.com>
9705 * config/mips/mips.md (simple_return): Attempt to use JRC
9707 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
9709 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9711 PR rtl-optimization/59137
9712 * reorg.c (steal_delay_list_from_target): Call update_block for
9714 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
9716 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9718 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
9719 two duplicate entries.
9721 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9724 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
9726 * config/mips/mips.c (mips_truncated_op_cost): New function.
9727 (mips_rtx_costs): Adjust test for BADDU.
9728 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
9730 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
9732 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
9733 (*baddu_si): ...this new pattern.
9735 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9738 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
9740 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9743 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
9745 (expand_expr, expand_normal): Adjust.
9746 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
9747 inner_reference_p. Use inner_reference_p to expand inner references.
9748 (store_expr): Adjust.
9749 * cfgexpand.c (expand_call_stmt): Adjust.
9751 2014-01-08 Rong Xu <xur@google.com>
9753 * gcov-io.c (gcov_var): Move from gcov-io.h.
9754 (gcov_position): Ditto.
9755 (gcov_is_error): Ditto.
9756 (gcov_rewrite): Ditto.
9757 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
9758 only part to libgcc/libgcov.h.
9760 2014-01-08 Marek Polacek <polacek@redhat.com>
9763 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
9765 2014-01-08 Marek Polacek <polacek@redhat.com>
9768 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
9770 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9772 PR rtl-optimization/59649
9773 * stor-layout.c (get_mode_bounds): For BImode return
9774 0 and STORE_FLAG_VALUE.
9776 2014-01-08 Richard Biener <rguenther@suse.de>
9779 * gimple.h (is_gimple_builtin_call): Remove.
9780 (gimple_builtin_call_types_compatible_p): New.
9781 (gimple_call_builtin_p): New overload.
9782 * gimple.c (is_gimple_builtin_call): Remove.
9783 (validate_call): Rename to ...
9784 (gimple_builtin_call_types_compatible_p): ... this and export. Also
9786 (validate_type): New static function.
9787 (gimple_call_builtin_p): New overload and adjust.
9788 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
9789 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
9790 (gimple_fold_stmt_to_constant_1): Likewise.
9791 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
9793 2014-01-08 Richard Biener <rguenther@suse.de>
9796 * gimplify.c (gimplify_expr): Gimplify register-register type
9797 VIEW_CONVERT_EXPRs to separate stmts.
9799 2014-01-07 Jeff Law <law@redhat.com>
9802 * ree.c (combine_set_extension): Handle case where source
9803 and destination registers in an extension insn are different.
9804 (combine_reaching_defs): Allow source and destination registers
9805 in extension to be different under limited circumstances.
9806 (add_removable_extension): Remove restriction that the
9807 source and destination registers in the extension are the same.
9808 (find_and_remove_re): Emit a copy from the extension's
9809 destination to its source after the defining insn if
9810 the source and destination registers are different.
9813 * ifcvt.c (merge_if_block): If we are merging a block with more than
9814 one successor with a block with no successors, remove any BARRIER
9815 after the second block.
9817 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
9819 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
9821 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
9824 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
9825 for 14-bit register offsets when INT14_OK_STRICT is false.
9827 2014-01-07 Roland Stigge <stigge@antcom.de>
9828 Michael Meissner <meissner@linux.vnet.ibm.com>
9831 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
9832 Only check TFmode for SPE constants. Don't check TImode or TDmode.
9834 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
9836 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
9839 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
9841 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
9842 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
9843 rtx is const0_rtx or not.
9845 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
9848 * target-globals.c (save_target_globals): Remove this_fn_optab
9850 * toplev.c: Include optabs.h.
9851 (target_reinit): Temporarily restore the global options if another
9852 set of options are in force.
9854 2014-01-07 Jakub Jelinek <jakub@redhat.com>
9856 PR rtl-optimization/58668
9857 * cfgcleanup.c (flow_find_cross_jump): Don't count
9858 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
9859 to determine what is counted.
9860 (flow_find_head_matching_sequence): Use active_insn_p to determine
9862 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
9864 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
9865 determine what is counted.
9867 PR tree-optimization/59643
9868 * tree-predcom.c (split_data_refs_to_components): If one dr is
9869 read and one write, determine_offset fails and the write isn't
9870 in the bad component, just put the read into the bad component.
9872 2014-01-07 Mike Stump <mikestump@comcast.net>
9873 Jakub Jelinek <jakub@redhat.com>
9876 * tree-core.h (struct tree_optimization_option): Change optabs
9877 type from unsigned char * to void *.
9878 * optabs.c (init_tree_optimization_optabs): Adjust
9879 TREE_OPTIMIZATION_OPTABS initialization.
9881 2014-01-06 Jakub Jelinek <jakub@redhat.com>
9884 * config/i386/i386.h (struct machine_function): Add
9885 no_drap_save_restore field.
9886 * config/i386/i386.c (ix86_save_reg): Use
9887 !cfun->machine->no_drap_save_restore instead of
9888 crtl->stack_realign_needed.
9889 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
9890 this function clears frame_pointer_needed. Set
9891 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
9892 and DRAP reg is needed.
9894 2014-01-06 Marek Polacek <polacek@redhat.com>
9897 * doc/implement-c.texi: Mention that other integer types are
9898 permitted as bit-field types in strictly conforming mode.
9900 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
9902 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
9905 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
9907 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
9909 2014-01-06 Martin Jambor <mjambor@suse.cz>
9912 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
9914 * ipa-prop.c (ipa_print_node_params): Fix indentation.
9916 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
9920 * var-tracking.c (add_stores): Preserve the value of the source even if
9921 we don't record the store.
9923 2014-01-06 Terry Guo <terry.guo@arm.com>
9925 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
9927 2014-01-05 Iain Sandoe <iain@codesourcery.com>
9930 * config/darwin.c (darwin_function_section): Adjust return values to
9931 correspond to optimisation changes made in r206070.
9933 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
9935 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
9936 from prefetch_block tune setting.
9937 (nocona_cost): Correct size of prefetch block to 64.
9939 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
9941 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
9942 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
9943 used to save the static chain register in the computation of the offset
9944 from which the FP registers need to be restored.
9946 2014-01-04 Jakub Jelinek <jakub@redhat.com>
9948 PR tree-optimization/59519
9949 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
9950 ICE if get_current_def (current_new_name) is already non-NULL, as long
9951 as it is a phi result of some other phi in *new_exit_bb that has
9954 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
9955 or vmovdqu* for misaligned_operand.
9956 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
9957 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
9958 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
9959 aligned_mem for AVX512F masked aligned load and store builtins and for
9962 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
9964 PR tree-optimization/59651
9965 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
9966 Address range for negative step should be added by TYPE_SIZE_UNIT.
9968 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
9970 * config/m68k/m68k.c (handle_move_double): Handle pushes with
9971 overlapping registers also for registers other than the stack pointer.
9973 2014-01-03 Marek Polacek <polacek@redhat.com>
9976 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
9979 2014-01-03 Jakub Jelinek <jakub@redhat.com>
9982 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
9985 * config/i386/i386.md (MODE_SIZE): New mode attribute.
9986 (push splitter): Use <P:MODE_SIZE> instead of
9987 GET_MODE_SIZE (<P:MODE>mode).
9988 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
9989 (mov -1, reg peephole2): Likewise.
9990 * config/i386/sse.md (*mov<mode>_internal,
9991 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
9992 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
9993 *<code><mode>3, *andnot<mode>3<mask_name>,
9994 <mask_codefor><code><mode>3<mask_name>): Likewise.
9995 * config/i386/subst.md (mask_mode512bit_condition,
9996 sd_mask_mode512bit_condition): Likewise.
9998 2014-01-02 Xinliang David Li <davidxl@google.com>
10000 PR tree-optimization/59303
10001 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
10002 (dump_predicates): Better output format.
10003 (pred_equal_p): New function.
10004 (is_neq_relop_p): Ditto.
10005 (is_neq_zero_form_p): Ditto.
10006 (pred_expr_equal_p): Ditto.
10007 (pred_neg_p): Ditto.
10008 (simplify_pred): Ditto.
10009 (simplify_preds_2): Ditto.
10010 (simplify_preds_3): Ditto.
10011 (simplify_preds_4): Ditto.
10012 (simplify_preds): Ditto.
10013 (push_pred): Ditto.
10014 (push_to_worklist): Ditto.
10015 (get_pred_info_from_cmp): Ditto.
10016 (is_degenerated_phi): Ditto.
10017 (normalize_one_pred_1): Ditto.
10018 (normalize_one_pred): Ditto.
10019 (normalize_one_pred_chain): Ditto.
10020 (normalize_preds): Ditto.
10021 (normalize_cond_1): Remove function.
10022 (normalize_cond): Ditto.
10023 (is_gcond_subset_of): Ditto.
10024 (is_subset_of_any): Ditto.
10025 (is_or_set_subset_of): Ditto.
10026 (is_and_set_subset_of): Ditto.
10027 (is_norm_cond_subset_of): Ditto.
10028 (pred_chain_length_cmp): Ditto.
10029 (convert_control_dep_chain_into_preds): Type change.
10030 (find_predicates): Ditto.
10031 (find_def_preds): Ditto.
10032 (destroy_predicates_vecs): Ditto.
10033 (find_matching_predicates_in_rest_chains): Ditto.
10034 (use_pred_not_overlap_with_undef_path_pred): Ditto.
10035 (is_pred_expr_subset): Ditto.
10036 (is_pred_chain_subset_of): Ditto.
10037 (is_included_in): Ditto.
10038 (is_superset_of): Ditto.
10040 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10042 Update copyright years.
10044 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10046 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
10047 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
10048 config/arc/arc.md, config/arc/arc.opt,
10049 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
10050 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
10051 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
10052 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
10053 config/linux-protos.h, config/linux.c, config/winnt-c.c,
10054 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
10055 vtable-verify.c, vtable-verify.h: Use the standard form for the
10058 2014-01-02 Tobias Burnus <burnus@net-b.de>
10060 * gcc.c (process_command): Update copyright notice dates.
10061 * gcov-dump.c: Ditto.
10063 * doc/cpp.texi: Bump @copying's copyright year.
10064 * doc/cppinternals.texi: Ditto.
10065 * doc/gcc.texi: Ditto.
10066 * doc/gccint.texi: Ditto.
10067 * doc/gcov.texi: Ditto.
10068 * doc/install.texi: Ditto.
10069 * doc/invoke.texi: Ditto.
10071 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10073 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
10075 2014-01-01 Jakub Jelinek <jakub@redhat.com>
10077 * config/i386/sse.md (*mov<mode>_internal): Guard
10078 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
10080 PR rtl-optimization/59647
10081 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
10082 new_rtx into UNSIGNED_FLOAT rtxes.
10084 Copyright (C) 2014 Free Software Foundation, Inc.
10086 Copying and distribution of this file, with or without modification,
10087 are permitted in any medium without royalty provided the copyright
10088 notice and this notice are preserved.