1 2014-05-08 Richard Biener <rguenther@suse.de>
3 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
4 looking at TYPE_ARG_TYPES.
6 2014-05-08 Richard Biener <rguenther@suse.de>
8 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
9 pointer propagation special-case.
11 2014-05-08 Bin Cheng <bin.cheng@arm.com>
13 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
14 core part of address expressions.
16 2014-05-08 Alan Modra <amodra@gmail.com>
19 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
20 loads and stores when -mno-strict-align at any alignment.
21 (expand_block_clear): Similarly. Also correct calculation of
24 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
27 * tree-complex.c (expand_complex_move): Keep line info when expanding
29 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
30 of complex expression. Use new argument to display correct location
31 for values coming from phi statement.
32 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
33 (warn_uninitialized_phi): Pass location of phi argument to
35 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
36 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
38 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
40 * config/rs6000/predicates.md (indexed_address_mem): New.
41 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
42 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
43 fpstore_ux, fpstore_u.
44 (sign_extend, indexed, update): New.
46 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
47 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
48 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
49 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
50 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
51 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
52 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
53 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
54 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
55 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
56 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
57 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
58 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
59 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
60 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
62 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
63 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
64 *vsx_extract_<mode>_store): Adjust.
65 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
66 is_cracked_insn, insn_must_be_first_in_group,
67 insn_must_be_last_in_group): Adjust.
69 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
71 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
72 ppc440-fpstore): Adjust.
73 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
74 ppc476-fpstore): Adjust.
75 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
76 ppc601-fpstore): Adjust.
77 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
79 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
81 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
82 ppc7450-fpstore): Adjust.
83 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
84 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
85 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
87 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
88 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
89 cell-fpstore, cell-fpstore-update): Adjust.
90 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
91 ppce300c3_store, ppce300c3_fpstore): Adjust.
92 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
93 e500mc_fpstore): Adjust.
94 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
95 e500mc64_store, e500mc64_fpstore): Adjust.
96 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
97 e5500_fpstore): Adjust.
98 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
99 e6500_fpstore): Adjust.
100 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
102 * config/rs6000/power4.md (power4-load, power4-load-ext,
103 power4-load-ext-update, power4-load-ext-update-indexed,
104 power4-load-update-indexed, power4-load-update, power4-fpload,
105 power4-fpload-update, power4-store, power4-store-update,
106 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
108 * config/rs6000/power5.md (power5-load, power5-load-ext,
109 power5-load-ext-update, power5-load-ext-update-indexed,
110 power5-load-update-indexed, power5-load-update, power5-fpload,
111 power5-fpload-update, power5-store, power5-store-update,
112 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
114 * config/rs6000/power6.md (power6-load, power6-load-ext,
115 power6-load-update, power6-load-update-indexed,
116 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
117 power6-fpload-update, power6-store, power6-store-update,
118 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
120 * config/rs6000/power7.md (power7-load, power7-load-ext,
121 power7-load-update, power7-load-update-indexed,
122 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
123 power7-fpload-update, power7-store, power7-store-update,
124 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
126 * config/rs6000/power8.md (power8-load, power8-load-update,
127 power8-load-ext, power8-load-ext-update, power8-fpload,
128 power8-fpload-update, power8-store, power8-store-update-indexed,
129 power8-fpstore, power8-fpstore-update): Adjust.
130 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
132 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
133 titan_lsu_store, titan_lsu_fpstore): Adjust.
134 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
136 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
139 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
140 unrolled byte insns. Emit address increments after move insns.
142 2014-05-07 David Malcolm <dmalcolm@redhat.com>
144 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
145 const_gimple, rather than a gimple.
146 (gimple_call_builtin_p): Likewise, for the three variants.
148 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
149 (gimple_call_builtin_p): Likewise, for the three variants.
151 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
153 PR tree-optimization/61095
154 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
156 2014-05-07 Richard Biener <rguenther@suse.de>
158 PR tree-optimization/61034
159 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
160 (maybe_skip_until): Use translate to take into account
161 lattices when trying to do disambiguations.
162 (get_continuation_for_phi_1): Likewise.
163 (get_continuation_for_phi): Adjust for added translate
165 (walk_non_aliased_vuses): Likewise.
166 * tree-ssa-alias.h (get_continuation_for_phi): Adjust
168 (walk_non_aliased_vuses): Likewise.
169 (call_may_clobber_ref_p_1): Declare.
170 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also
171 disambiguate against calls. Stop early if we are
172 only supposed to disambiguate.
173 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
175 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
177 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
178 Emit an error when the function has arguments.
180 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
182 * cfgloop.h (unswitch_loops): Remove.
183 * doc/passes.texi: Remove references to loop-unswitch.c
184 * timevar.def (TV_LOOP_UNSWITCH): Remove.
186 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
188 * tree-vect-data-refs.c (vect_grouped_load_supported): New
189 check for loads group of length 3.
190 (vect_permute_load_chain): New permutations for loads group of
192 * tree-vect-stmts.c (vect_model_load_cost): Change cost
193 of vec_perm_shuffle for the new permutations.
195 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
197 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
198 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
199 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
200 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
201 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
202 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
203 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
204 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
206 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
208 * loop-unswitch.c: Delete.
210 2014-05-07 Richard Biener <rguenther@suse.de>
212 * config.gcc: Always set need_64bit_hwint to yes.
214 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
216 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
217 of using optimize_size.
219 2014-05-06 Mike Stump <mikestump@comcast.net>
221 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
223 2014-05-06 Joseph Myers <joseph@codesourcery.com>
225 * config/i386/sse.md (*mov<mode>_internal)
226 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
227 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
228 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
229 (*<code><mode>3, *andnot<mode>3<mask_name>)
230 (<mask_codefor><code><mode>3<mask_name>): Only consider
231 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
233 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
236 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
238 * lra-constraints.c (valid_address_p): Move earlier in file.
239 Add a constraint argument to the address_info version.
240 (satisfies_memory_constraint_p): New function.
241 (satisfies_address_constraint_p): Likewise.
242 (process_alt_operands, curr_insn_transform): Use them.
243 (process_address): Pass the constraint to valid_address_p when
244 checking address operands.
246 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
248 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
249 to their respective blocks. Fix inadvertent use of "node".
251 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
253 * emit-rtl.c (init_derived_machine_modes): New functionm, split
255 (init_emit_once): ...here.
256 * rtl.h (init_derived_machine_modes): Declare.
257 * toplev.c (do_compile): Call it even if no_backend.
259 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
260 Mike Stump <mikestump@comcast.net>
261 Richard Sandiford <rdsandiford@googlemail.com>
262 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
264 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
265 (rtx_equal_for_memref_p): Update comment.
266 (adjust_offset_for_component_ref): Use wide-int interfaces.
267 * builtins.c (get_object_alignment_2): Likewise.
268 (c_readstr): Likewise.
269 (target_char_cast): Add comment.
270 (determine_block_size): Use wide-int interfaces.
271 (expand_builtin_signbit): Likewise.
272 (fold_builtin_int_roundingfn): Likewise.
273 (fold_builtin_bitop): Likewise.
274 (fold_builtin_bswap): Likewise.
275 (fold_builtin_logarithm): Use signop.
276 (fold_builtin_pow): Likewise.
277 (fold_builtin_memory_op): Use wide-int interfaces.
278 (fold_builtin_object_size): Likewise.
279 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
280 nb_iterations_estimate.
281 (record_niter_bound): Use wide-int interfaces.
282 (get_estimated_loop_iterations_int): Likewise.
283 (get_estimated_loop_iterations): Likewise.
284 (get_max_loop_iterations): Likewise.
285 * cfgloop.h: Include wide-int.h.
286 (struct nb_iter_bound): Change bound to widest_int.
287 (struct loop): Change nb_iterations_upper_bound and
288 nb_iterations_estimate to widest_int.
289 (record_niter_bound): Switch to use widest_int.
290 (get_estimated_loop_iterations): Likewise.
291 (get_max_loop_iterations): Likewise.
292 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
294 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
295 * combine.c (try_combine): Likewise.
296 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
297 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
299 (aarch64_float_const_representable_p): Likewise.
300 * config/arc/arc.c: Include wide-int.h.
301 (arc_can_use_doloop_p): Use wide-int interfaces.
302 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
303 (vfp3_const_double_index): Likewise.
304 * config/avr/avr.c (avr_out_round): Likewise.
305 (avr_fold_builtin): Likewise.
306 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
307 (bfin_can_use_doloop_p): Likewise.
308 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
309 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
310 * config/i386/i386.c: Include wide-int.h.
311 (ix86_data_alignment): Use wide-int interfaces.
312 (ix86_local_alignment): Likewise.
313 (ix86_emit_swsqrtsf): Update real_from_integer.
314 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
315 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
316 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
317 (zero_constant): Likewise.
318 (input_operand): Likewise.
319 (splat_input_operand): Likewise.
320 (non_logical_cint_operand): Change const_double to const_wide_int.
321 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
322 (easy_altivec_constant): Remove comment.
323 (paired_expand_vector_init): Use CONSTANT_P.
324 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
325 (rs6000_emit_move): Update checks.
326 (rs6000_aggregate_candidate): Use wide-int interfaces.
327 (rs6000_expand_ternop_builtin): Likewise.
328 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
329 (rs6000_assemble_integer): Likewise.
330 (rs6000_hash_constant): Likewise.
331 (output_toc): Likewise.
332 (rs6000_rtx_costs): Likewise.
333 (rs6000_emit_swrsqrt); Update call to real_from_integer.
334 * config/rs6000/rs6000-c.c: Include wide-int.h.
335 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
336 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
337 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
338 Handle CONST_WIDE_INT.
339 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
340 Use tree_fits_uhwi_p.
341 * config/sparc/sparc.c: Include wide-int.h.
342 (sparc_fold_builtin): Use wide-int interfaces.
343 * config/vax/vax.c: Include wide-int.h.
344 (vax_float_literal): Use real_from_integer.
345 * coretypes.h (struct hwivec_def): New.
348 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
349 (equiv_constant): Handle CONST_WIDE_INT.
350 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
351 (cselib_hash_rtx): Handle CONST_WIDE_INT.
352 * dbxout.c (stabstr_U): Use wide-int interfaces.
353 (dbxout_type): Update to use cst_fits_shwi_p.
354 * defaults.h (LOG2_BITS_PER_UNIT): Define.
355 (TARGET_SUPPORTS_WIDE_INT): Add default.
356 * dfp.c: Include wide-int.h.
357 (decimal_real_to_integer2): Use wide-int interfaces and rename to
358 decimal_real_to_integer.
359 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
360 decimal_real_to_integer.
361 * doc/generic.texi (Constant expressions): Update for wide_int.
362 * doc/rtl.texi (const_double): Likewise.
363 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
364 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
365 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
366 (REAL_VALUE_FROM_INT): Remove.
367 (TARGET_SUPPORTS_WIDE_INT): New.
368 * doc/tm.texi: Regenerate.
369 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
370 * double-int.h: Include wide-int.h.
371 (struct wi::int_traits): New.
372 * dwarf2out.c (get_full_len): New.
373 (dw_val_equal_p): Add case dw_val_class_wide_int.
374 (size_of_loc_descr): Likewise.
375 (output_loc_operands): Likewise.
376 (insert_double): Remove.
377 (insert_wide_int): New.
379 (print_die): Add case dw_val_class_wide_int.
380 (attr_checksum): Likewise.
381 (attr_checksum_ordered): Likewise.
382 (same_dw_val_p): Likewise.
383 (size_of_die): Likewise.
384 (value_format): Likewise.
385 (output_die): Likewise.
386 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
388 (clz_loc_descriptor): Use wide-int interfaces.
389 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
390 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
391 (round_up_to_align): Use wide-int interfaces.
392 (field_byte_offset): Likewise.
393 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
394 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
395 CONST_DOUBLE handling. Use wide-int interfaces.
396 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
397 (gen_enumeration_type_die): Use add_AT_wide.
398 (hash_loc_operands): Add case dw_val_class_wide_int.
399 (compare_loc_operands): Likewise.
400 * dwarf2out.h: Include wide-int.h.
402 (enum dw_val_class): Add dw_val_class_wide_int.
403 (struct dw_val_struct): Add val_wide.
404 * emit-rtl.c (const_wide_int_htab): New.
405 (const_wide_int_htab_hash): New.
406 (const_wide_int_htab_eq): New.
407 (lookup_const_wide_int): New.
408 (const_double_htab_hash): Use wide-int interfaces.
409 (const_double_htab_eq): Likewise.
410 (rtx_to_double_int): Conditionally compile for wide-int.
411 (immed_double_int_const): Rename to immed_wide_int_const and
413 (immed_double_const): Conditionally compile for wide-int.
414 (init_emit_once): Use wide-int interfaces.
415 * explow.c (plus_constant): Likewise.
416 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
417 (lshift_value): Use wide-int interfaces.
418 (expand_mult): Likewise.
419 (choose_multiplier): Likewise.
420 (expand_smod_pow2): Likewise.
421 (make_tree): Likewise.
422 * expr.c (convert_modes): Consolidate handling of constants.
423 Use wide-int interfaces.
424 (emit_group_load_1): Add note.
425 (store_expr): Update comment.
426 (get_inner_reference): Use wide-int interfaces.
427 (expand_constructor): Update comment.
428 (expand_expr_real_2): Use wide-int interfaces.
429 (expand_expr_real_1): Likewise.
430 (reduce_to_bit_field_precision): Likewise.
431 (const_vector_from_tree): Likewise.
432 * final.c: Include wide-int-print.h.
433 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
434 * fixed-value.c: Include wide-int.h.
435 (fixed_from_string): Use wide-int interfaces.
436 (fixed_to_decimal): Likewise.
437 (fixed_convert_from_real): Likewise.
438 (real_convert_from_fixed): Likewise.
439 * fold-const.h (mem_ref_offset): Return an offset_int.
440 (div_if_zero_remainder): Remove code parameter.
441 * fold-const.c (div_if_zero_remainder): Remove code parameter.
442 Use wide-int interfaces.
443 (may_negate_without_overflow_p): Use wide-int interfaces.
444 (negate_expr_p): Likewise.
445 (fold_negate_expr): Likewise.
446 (int_const_binop_1): Likewise.
447 (const_binop): Likewise.
448 (fold_convert_const_int_from_int): Likewise.
449 (fold_convert_const_int_from_real): Likewise.
450 (fold_convert_const_int_from_fixed): Likewise.
451 (fold_convert_const_fixed_from_int): Likewise.
452 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
453 (sign_bit_p): Use wide-int interfaces.
454 (make_range_step): Likewise.
455 (build_range_check): Likewise. Pass an integer of the correct type
456 instead of using integer_one_node.
457 (range_predecessor): Pass an integer of the correct type instead
458 of using integer_one_node.
459 (range_successor): Likewise.
460 (merge_ranges): Likewise.
461 (unextend): Use wide-int interfaces.
462 (extract_muldiv_1): Likewise.
463 (fold_div_compare): Likewise.
464 (fold_single_bit_test): Likewise.
465 (fold_sign_changed_comparison): Likewise.
466 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
467 (fold_plusminus_mult_expr): Use wide-int interfaces.
468 (native_encode_int): Likewise.
469 (native_interpret_int): Likewise.
470 (fold_unary_loc): Likewise.
471 (pointer_may_wrap_p): Likewise.
472 (size_low_cst): Likewise.
473 (mask_with_tz): Likewise.
474 (fold_binary_loc): Likewise.
475 (fold_ternary_loc): Likewise.
476 (multiple_of_p): Likewise.
477 (tree_call_nonnegative_warnv_p): Update calls to
478 tree_int_cst_min_precision and real_from_integer.
479 (fold_negate_const): Use wide-int interfaces.
480 (fold_abs_const): Likewise.
481 (fold_relational_const): Use tree_int_cst_lt.
482 (round_up_loc): Use wide-int interfaces.
483 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
484 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
485 * gengtype.c: Remove include of double-int.h.
486 (do_typedef): Use wide-int interfaces.
487 (open_base_files): Add wide-int.h.
488 (main): Add offset_int and widest_int typedefs.
489 * gengtype-lex.l: Handle "^".
490 (CXX_KEYWORD): Add "static".
491 * gengtype-parse.c (require3): New.
492 (require_template_declaration): Handle constant template arguments
493 and nested templates.
494 * gengtype-state.c: Don't include "double-int.h".
495 * genpreds.c (write_one_predicate_function): Update comment.
496 (write_tm_constrs_h): Add check for hval and lval use in
498 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
499 (add_to_sequence): Likewise.
500 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
501 and const_double_operand.
502 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
504 * gimple-fold.c (get_base_constructor): Likewise.
505 (fold_array_ctor_reference): Likewise.
506 (fold_nonarray_ctor_reference): Likewise.
507 (fold_const_aggregate_ref_1): Likewise.
508 (gimple_val_nonnegative_real_p): Likewise.
509 (gimple_fold_indirect_ref): Likewise.
510 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
511 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
512 (struct slsr_cand_d): Change index to be widest_int.
513 (struct incr_info_d): Change incr to be widest_int.
514 (alloc_cand_and_find_basis): Use wide-int interfaces.
515 (slsr_process_phi): Likewise.
516 (backtrace_base_for_ref): Likewise. Return a widest_int.
517 (restructure_reference): Take a widest_int instead of a double_int.
518 (slsr_process_ref): Use wide-int interfaces.
519 (create_mul_ssa_cand): Likewise.
520 (create_mul_imm_cand): Likewise.
521 (create_add_ssa_cand): Likewise.
522 (create_add_imm_cand): Take a widest_int instead of a double_int.
523 (slsr_process_add): Use wide-int interfaces.
524 (slsr_process_cast): Likewise.
525 (slsr_process_copy): Likewise.
526 (dump_candidate): Likewise.
527 (dump_incr_vec): Likewise.
528 (replace_ref): Likewise.
529 (cand_increment): Likewise. Return a widest_int.
530 (cand_abs_increment): Likewise.
531 (replace_mult_candidate): Take a widest_int instead of a double_int.
532 (replace_unconditional_candidate): Use wide-int interfaces.
533 (incr_vec_index): Take a widest_int instead of a double_int.
534 (create_add_on_incoming_edge): Likewise.
535 (create_phi_basis): Use wide-int interfaces.
536 (replace_conditional_candidate): Likewise.
537 (record_increment): Take a widest_int instead of a double_int.
538 (record_phi_increments): Use wide-int interfaces.
539 (phi_incr_cost): Take a widest_int instead of a double_int.
540 (lowest_cost_path): Likewise.
541 (total_savings): Likewise.
542 (analyze_increments): Use wide-int interfaces.
543 (ncd_with_phi): Take a widest_int instead of a double_int.
544 (ncd_of_cand_and_phis): Likewise.
545 (nearest_common_dominator_for_cands): Likewise.
546 (insert_initializers): Use wide-int interfaces.
547 (all_phi_incrs_profitable): Likewise.
548 (replace_one_candidate): Likewise.
549 (replace_profitable_candidates): Likewise.
550 * godump.c: Include wide-int-print.h.
551 (go_output_typedef): Use wide-int interfaces.
552 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
553 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
554 (build_loop_iteration_domains): Likewise.
555 * hooks.h: Include wide-int.h rather than double-int.h.
556 (hook_bool_dint_dint_uint_bool_true): Delete.
557 (hook_bool_wint_wint_uint_bool_true): Declare.
558 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
559 (hook_bool_wint_wint_uint_bool_true): New.
560 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
562 (ubsan_expand_si_overflow_mul_check): Likewise.
563 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
564 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
565 (get_ancestor_addr_info): Likewise.
566 (ipa_modify_call_arguments): Likewise.
567 * loop-doloop.c (doloop_modify): Likewise.
568 (doloop_optimize): Likewise.
569 * loop-iv.c (iv_number_of_iterations): Likewise.
570 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
571 (unroll_loop_constant_iterations): Likewise.
572 (decide_unroll_runtime_iterations): Likewise.
573 (unroll_loop_runtime_iterations): Likewise.
574 (decide_peel_simple): Likewise.
575 (decide_unroll_stupid): Likewise.
576 * lto-streamer-in.c (streamer_read_wi): Add.
577 (input_cfg): Use wide-int interfaces.
578 (lto_input_tree_1): Likewise.
579 * lto-streamer-out.c (streamer_write_wi): Add.
580 (hash_tree): Use wide-int interfaces.
581 (output_cfg): Likewise.
582 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
583 (GTFILES): Add wide-int.h and signop.h.
584 (TAGS): Look for .cc files too.
585 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
586 * optabs.c (expand_subword_shift): Likewise.
587 (expand_doubleword_shift): Likewise.
588 (expand_absneg_bit): Likewise.
589 (expand_copysign_absneg): Likewise.
590 (expand_copysign_bit): Likewise.
591 * postreload.c (reload_cse_simplify_set): Likewise.
592 * predict.c (predict_iv_comparison): Likewise.
593 * pretty-print.h: Include wide-int-print.h.
595 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
596 * print-tree.c: Include wide-int-print.h.
597 (print_node_brief): Use wide-int interfaces.
598 (print_node): Likewise.
599 * read-rtl.c (validate_const_wide_int): New.
600 (read_rtx_code): Add CONST_WIDE_INT case.
601 * real.c: Include wide-int.h.
602 (real_to_integer2): Delete.
603 (real_to_integer): New function, returning a wide_int.
604 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
605 (ten_to_ptwo): Update call to real_from_integer.
606 (real_digit): Likewise.
607 * real.h: Include signop.h, wide-int.h and insn-modes.h.
608 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
609 (REAL_VALUE_TO_INT): Delete.
610 (real_to_integer): Declare a wide-int form.
611 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
612 * recog.c (const_int_operand): Improve comment.
613 (const_scalar_int_operand): New.
614 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
615 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
616 (split_double): Likewise.
617 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
618 (rtx_size): Likewise.
619 (rtx_alloc_stat_v): New.
620 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
621 (cwi_output_hex): New.
622 (iterative_hash_rtx): Handle CONST_WIDE_INT.
623 (cwi_check_failed_bounds): New.
624 * rtl.def (CONST_WIDE_INT): New.
625 * rtl.h: Include <utility> and wide-int.h.
626 (struct hwivec_def): New.
627 (CWI_GET_NUM_ELEM): New.
628 (CWI_PUT_NUM_ELEM): New.
629 (struct rtx_def): Add num_elem and hwiv.
630 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
631 (CASE_CONST_UNIQUE): Likewise.
632 (CASE_CONST_ANY): Likewise.
633 (CONST_SCALAR_INT_P): Likewise.
634 (CONST_WIDE_INT_P): New.
637 (cwi_check_failed_bounds): New.
640 (CONST_WIDE_INT_VEC) New.
641 (CONST_WIDE_INT_NUNITS) New.
642 (CONST_WIDE_INT_ELT) New.
643 (rtx_mode_t): New type.
644 (wi::int_traits <rtx_mode_t>): New.
646 (wi::min_value): New.
647 (wi::max_value): New.
649 (const_wide_int_alloc): New.
650 (immed_wide_int_const): New.
651 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
652 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
653 * signop.h: New file.
654 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
655 (simplify_const_unary_operation): Use wide-int interfaces.
656 (simplify_binary_operation_1): Likewise.
657 (simplify_const_binary_operation): Likewise.
658 (simplify_const_relational_operation): Likewise.
659 (simplify_immed_subreg): Likewise.
660 * stmt.c (expand_case): Likewise.
661 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
662 signop rather than a bool.
663 * stor-layout.c (layout_type): Use wide-int interfaces.
664 (initialize_sizetypes): Update calls to
665 set_min_and_max_values_for_integral_type.
666 (set_min_and_max_values_for_integral_type): Take a signop rather
667 than a bool. Use wide-int interfaces.
668 (fixup_signed_type): Update accordingly. Remove
669 HOST_BITS_PER_DOUBLE_INT limit.
670 (fixup_unsigned_type): Likewise.
671 * system.h (STATIC_CONSTANT_P): New.
672 (STATIC_ASSERT): New.
673 * target.def (can_use_doloop_p): Take widest_ints rather than
675 * target.h: Include wide-int.h rather than double-int.h.
676 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
678 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
679 rather than INT_CST_LT_UNSIGNED.
680 (can_use_doloop_if_innermost): Take widest_ints rather than
682 * tree-affine.c: Include wide-int-print.h.
683 (double_int_ext_for_comb): Delete.
684 (wide_int_ext_for_comb): New.
685 (aff_combination_zero): Use wide-int interfaces.
686 (aff_combination_const): Take a widest_int instead of a double_int.
687 (aff_combination_elt): Use wide-int interfaces.
688 (aff_combination_scale): Take a widest_int instead of a double_int.
689 (aff_combination_add_elt): Likewise.
690 (aff_combination_add_cst): Likewise.
691 (aff_combination_add): Use wide-int interfaces.
692 (aff_combination_convert): Likewise.
693 (tree_to_aff_combination): Likewise.
694 (add_elt_to_tree): Take a widest_int instead of a double_int.
695 (aff_combination_to_tree): Use wide-int interfaces.
696 (aff_combination_remove_elt): Likewise.
697 (aff_combination_add_product): Take a widest_int instead of
699 (aff_combination_mult): Use wide-int interfaces.
700 (aff_combination_expand): Likewise.
701 (double_int_constant_multiple_p): Delete.
702 (wide_int_constant_multiple_p): New.
703 (aff_combination_constant_multiple_p): Take a widest_int pointer
704 instead of a double_int pointer.
705 (print_aff): Use wide-int interfaces.
706 (get_inner_reference_aff): Take a widest_int pointer
707 instead of a double_int pointer.
708 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
709 * tree-affine.h: Include wide-int.h.
710 (struct aff_comb_elt): Change type of coef to widest_int.
711 (struct affine_tree_combination): Change type of offset to widest_int.
712 (double_int_ext_for_comb): Delete.
713 (wide_int_ext_for_comb): New.
714 (aff_combination_const): Use widest_int instead of double_int.
715 (aff_combination_scale): Likewise.
716 (aff_combination_add_elt): Likewise.
717 (aff_combination_constant_multiple_p): Likewise.
718 (get_inner_reference_aff): Likewise.
719 (aff_comb_cannot_overlap_p): Likewise.
720 (aff_combination_zero_p): Use wide-int interfaces.
721 * tree.c: Include tree.h.
722 (init_ttree): Use make_int_cst.
723 (tree_code_size): Removed code for INTEGER_CST case.
724 (tree_size): Add INTEGER_CST case.
725 (make_node_stat): Update comment.
726 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
727 (build_int_cst_type): Use wide-int interfaces.
728 (double_int_to_tree): Likewise.
729 (double_int_fits_to_tree_p): Delete.
730 (force_fit_type_double): Delete.
731 (force_fit_type): New.
732 (int_cst_hash_hash): Use wide-int interfaces.
733 (int_cst_hash_eq): Likewise.
734 (build_int_cst_wide): Delete.
735 (wide_int_to_tree): New.
736 (cache_integer_cst): Use wide-int interfaces.
737 (build_low_bits_mask): Likewise.
738 (cst_and_fits_in_hwi): Likewise.
739 (real_value_from_int_cst): Likewise.
740 (make_int_cst_stat): New.
741 (integer_zerop): Use wide_int interfaces.
742 (integer_onep): Likewise.
743 (integer_all_onesp): Likewise.
744 (integer_pow2p): Likewise.
745 (integer_nonzerop): Likewise.
746 (tree_log2): Likewise.
747 (tree_floor_log2): Likewise.
748 (tree_ctz): Likewise.
749 (int_size_in_bytes): Likewise.
750 (mem_ref_offset): Return an offset_int rather than a double_int.
751 (build_type_attribute_qual_variant): Use wide_int interfaces.
752 (type_hash_eq): Likewise
753 (tree_int_cst_equal): Likewise.
754 (tree_int_cst_lt): Delete.
755 (tree_int_cst_compare): Likewise.
756 (tree_fits_shwi_p): Use wide_int interfaces.
757 (tree_fits_uhwi_p): Likewise.
758 (tree_int_cst_sign_bit): Likewise.
759 (tree_int_cst_sgn): Likewise.
760 (tree_int_cst_min_precision): Take a signop rather than a bool.
761 (simple_cst_equal): Use wide_int interfaces.
762 (compare_tree_int): Likewise.
763 (iterative_hash_expr): Likewise.
764 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
766 (get_type_static_bounds): Use wide_int interfaces.
767 (tree_int_cst_elt_check_failed): New.
768 (build_common_tree_nodes): Reordered to set prec before filling in
770 (int_cst_value): Check cst_and_fits_in_hwi.
771 (widest_int_cst_value): Use wide_int interfaces.
772 (upper_bound_in_type): Likewise.
773 (lower_bound_in_type): Likewise.
774 (num_ending_zeros): Likewise.
775 (drop_tree_overflow): Likewise.
776 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
777 (gen_conditions_for_pow_cst_base): Likewise.
778 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
779 (group_case_labels_stmt): Use wide-int interfaces.
780 (verify_gimple_assign_binary): Likewise.
781 (print_loop): Likewise.
782 * tree-chrec.c (tree_fold_binomial): Likewise.
783 * tree-core.h (struct tree_base): Add int_length.
784 (struct tree_int_cst): Change rep of value.
785 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
786 (dr_may_alias_p): Likewise.
787 (max_stmt_executions_tree): Likewise.
788 * tree.def (INTEGER_CST): Update comment.
789 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
790 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
791 * tree-dump.c: Include wide-int.h and wide-int-print.h.
792 (dequeue_and_dump): Use wide-int interfaces.
793 * tree.h: Include wide-int.h.
794 (NULL_TREE): Moved to earlier loc in file.
795 (TREE_INT_CST_ELT_CHECK): New.
796 (tree_int_cst_elt_check_failed): New.
798 (TREE_INT_CST): Delete.
799 (TREE_INT_CST_LOW): Use wide-int interfaces.
800 (TREE_INT_CST_HIGH): Delete.
801 (TREE_INT_CST_NUNITS): New.
802 (TREE_INT_CST_EXT_NUNITS): Likewise.
803 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
804 (TREE_INT_CST_ELT): Likewise.
805 (INT_CST_LT): Delete.
806 (tree_int_cst_elt_check): New (two forms).
807 (type_code_size): Update comment.
808 (make_int_cst_stat, make_int_cst): New.
809 (tree_to_double_int): Delete.
810 (double_int_fits_to_tree_p): Delete.
811 (force_fit_type_double): Delete.
812 (build_int_cstu): Replace with out-of-line function.
813 (build_int_cst_wide): Delete.
814 (tree_int_cst_lt): Define inline.
815 (tree_int_cst_le): New.
816 (tree_int_cst_compare): Define inline.
817 (tree_int_cst_min_precision): Take a signop rather than a bool.
818 (wi::int_traits <const_tree>): New.
819 (wi::int_traits <tree>): New.
820 (wi::extended_tree): New.
821 (wi::int_traits <wi::extended_tree>): New.
822 (wi::to_widest): New.
823 (wi::to_offset): New.
824 (wi::fits_to_tree_p): New.
825 (wi::min_value): New.
826 (wi::max_value): New.
827 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
828 (copy_tree_body_r): Likewise.
829 * tree-object-size.c (compute_object_offset): Likewise.
830 (addr_object_size): Likewise.
831 * tree-predcom.c: Include wide-int-print.h.
832 (struct dref_d): Change type of offset to widest_int.
833 (dump_dref): Call wide-int printer.
834 (aff_combination_dr_offset): Use wide-int interfaces.
835 (determine_offset): Take a widest_int pointer rather than a
837 (split_data_refs_to_components): Use wide-int interfaces.
838 (suitable_component_p): Likewise.
839 (order_drefs): Likewise.
840 (add_ref_to_chain): Likewise.
841 (valid_initializer_p): Likewise.
842 (determine_roots_comp): Likewise.
843 * tree-pretty-print.c: Include wide-int-print.h.
844 (dump_generic_node): Use wide-int interfaces.
845 * tree-sra.c (sra_ipa_modify_expr): Likewise.
846 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
847 (move_fixed_address_to_symbol): Likewise.
848 (move_hint_to_base): Likewise.
849 (move_pointer_to_base): Likewise.
850 (move_variant_to_index): Likewise.
851 (most_expensive_mult_to_index): Likewise.
852 (addr_to_parts): Likewise.
853 (copy_ref_info): Likewise.
854 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
855 (indirect_refs_may_alias_p): Likewise.
856 (stmt_kills_ref_p_1): Likewise.
857 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
858 * tree-ssa-ccp.c: Update comment at top of file. Include
860 (struct prop_value_d): Change type of mask to widest_int.
861 (extend_mask): New function.
862 (dump_lattice_value): Use wide-int interfaces.
863 (get_default_value): Likewise.
864 (set_constant_value): Likewise.
865 (set_value_varying): Likewise.
866 (valid_lattice_transition): Likewise.
867 (set_lattice_value): Likewise.
868 (value_to_double_int): Delete.
869 (value_to_wide_int): New.
870 (get_value_from_alignment): Use wide-int interfaces.
871 (get_value_for_expr): Likewise.
872 (do_dbg_cnt): Likewise.
873 (ccp_finalize): Likewise.
874 (ccp_lattice_meet): Likewise.
875 (bit_value_unop_1): Use widest_ints rather than double_ints.
876 (bit_value_binop_1): Likewise.
877 (bit_value_unop): Use wide-int interfaces.
878 (bit_value_binop): Likewise.
879 (bit_value_assume_aligned): Likewise.
880 (evaluate_stmt): Likewise.
881 (ccp_fold_stmt): Likewise.
882 (visit_cond_stmt): Likewise.
883 (ccp_visit_stmt): Likewise.
884 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
885 (constant_pointer_difference): Likewise.
886 (associate_pointerplus): Likewise.
887 (combine_conversions): Likewise.
888 * tree-ssa-loop.h: Include wide-int.h.
889 (struct tree_niter_desc): Change type of max to widest_int.
890 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
891 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
892 (remove_redundant_iv_tests): Likewise.
893 (canonicalize_loop_induction_variables): Likewise.
894 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
895 (constant_multiple_of): Take a widest_int pointer instead of
896 a double_int pointer.
897 (get_computation_aff): Use wide-int interfaces.
898 (ptr_difference_cost): Likewise.
899 (difference_cost): Likewise.
900 (get_loop_invariant_expr_id): Likewise.
901 (get_computation_cost_at): Likewise.
902 (iv_elimination_compare_lt): Likewise.
903 (may_eliminate_iv): Likewise.
904 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
905 instead of double_int.
906 (max_loop_iterations): Likewise.
907 (max_stmt_executions): Likewise.
908 (estimated_stmt_executions): Likewise.
909 * tree-ssa-loop-niter.c: Include wide-int-print.h.
910 (split_to_var_and_offset): Use wide-int interfaces.
911 (determine_value_range): Likewise.
912 (bound_difference_of_offsetted_base): Likewise.
913 (bounds_add): Take a widest_int instead of a double_int.
914 (number_of_iterations_ne_max): Use wide-int interfaces.
915 (number_of_iterations_ne): Likewise.
916 (number_of_iterations_lt_to_ne): Likewise.
917 (assert_loop_rolls_lt): Likewise.
918 (number_of_iterations_lt): Likewise.
919 (number_of_iterations_le): Likewise.
920 (number_of_iterations_cond): Likewise.
921 (number_of_iterations_exit): Likewise.
922 (finite_loop_p): Likewise.
923 (derive_constant_upper_bound_assign): Likewise.
924 (derive_constant_upper_bound): Return a widest_int.
925 (derive_constant_upper_bound_ops): Likewise.
926 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
927 (record_estimate): Take a widest_int rather than a double_int.
928 (record_nonwrapping_iv): Use wide-int interfaces.
929 (double_int_cmp): Delete.
931 (bound_index): Take a widest_int rather than a double_int.
932 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
933 (maybe_lower_iteration_bound): Likewise.
934 (estimate_numbers_of_iterations_loop): Likewise.
935 (estimated_loop_iterations): Take a widest_int pointer than than
936 a double_int pointer.
937 (estimated_loop_iterations_int): Use wide-int interfaces.
938 (max_loop_iterations): Take a widest_int pointer than than
939 a double_int pointer.
940 (max_loop_iterations_int): Use wide-int interfaces.
941 (max_stmt_executions): Take a widest_int pointer than than
942 a double_int pointer.
943 (estimated_stmt_executions): Likewise.
944 (n_of_executions_at_most): Use wide-int interfaces.
945 (scev_probably_wraps_p): Likewise.
946 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
948 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
950 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
951 double_ints. Adjust for trailing_wide_ints <3> representation.
952 (set_nonzero_bits): Likewise.
953 (get_range_info): Return wide_ints rather than double_ints.
954 Adjust for trailing_wide_ints <3> representation.
955 (get_nonzero_bits): Likewise.
956 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
958 * tree-ssanames.h (struct range_info_def): Replace min, max and
959 nonzero_bits with a trailing_wide_ints <3>.
960 (set_range_info): Use wide_int_refs rather than double_ints.
961 (set_nonzero_bits): Likewise.
962 (get_range_info): Return wide_ints rather than double_ints.
963 (get_nonzero_bits): Likewise.
964 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
965 * tree-ssa-pre.c (phi_translate_1): Likewise.
966 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
967 (acceptable_pow_call): Likewise.
968 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
970 (vn_reference_fold_indirect): Likewise.
971 (vn_reference_maybe_forwprop_address): Likewise.
972 (valueize_refs_1): Likewise.
973 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
974 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
975 tree_int_cst_lt and tree_int_cst_le.
976 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
978 (streamer_alloc_tree): Likewise.
979 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
980 (streamer_write_tree_header): Likewise.
981 (streamer_write_integer_cst): Likewise.
982 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
983 (build_constructors): Likewise.
984 (array_value_type): Likewise.
985 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
986 (vect_check_gather): Likewise.
987 * tree-vect-generic.c (build_replicated_const): Likewise.
988 (expand_vector_divmod): Likewise.
989 * tree-vect-loop.c (vect_transform_loop): Likewise.
990 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
991 (vect_do_peeling_for_alignment): Likewise.
992 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
993 * tree-vrp.c: Include wide-int.h.
994 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
995 (extract_range_from_assert): Use wide-int interfaces.
996 (vrp_int_const_binop): Likewise.
997 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
999 (ranges_from_anti_range): Use wide-int interfaces.
1000 (quad_int_cmp): Delete.
1001 (quad_int_pair_sort): Likewise.
1002 (extract_range_from_binary_expr_1): Use wide-int interfaces.
1003 (extract_range_from_unary_expr_1): Likewise.
1004 (adjust_range_with_scev): Likewise.
1005 (masked_increment): Take and return wide_ints rather than double_ints.
1006 (register_edge_assert_for_2): Use wide-int interfaces.
1007 (check_array_ref): Likewise.
1008 (search_for_addr_array): Likewise.
1009 (maybe_set_nonzero_bits): Likewise.
1010 (union_ranges): Pass an integer of the correct type instead of
1011 using integer_one_node.
1012 (intersect_ranges): Likewise.
1013 (simplify_truth_ops_using_ranges): Likewise.
1014 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
1015 (range_fits_type_p): Likewise.
1016 (simplify_cond_using_ranges): Likewise. Take a signop rather than
1018 (simplify_conversion_using_ranges): Use wide-int interfaces.
1019 (simplify_float_conversion_using_ranges): Likewise.
1020 (vrp_finalize): Likewise.
1021 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
1022 (gimple_stringops_transform): Likewise.
1023 * varasm.c (decode_addr_const): Likewise.
1024 (const_hash_1): Likewise.
1025 (const_rtx_hash_1): Likewise
1026 (output_constant): Likewise.
1027 (array_size_for_constructor): Likewise.
1028 (output_constructor_regular_field): Likewise.
1029 (output_constructor_bitfield): Likewise.
1030 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
1031 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
1033 * gencheck.c: Define BITS_PER_UNIT.
1036 * wide-int-print.cc: New.
1037 * wide-int-print.h: New.
1039 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1041 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
1043 2014-05-06 Richard Biener <rguenther@suse.de>
1045 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
1046 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
1047 (TODO_verify_all): Adjust.
1048 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
1049 TODO_verify_stmts and TODO_verify_rtl_sharing.
1050 * bb-reorder.c: Likewise.
1051 * cfgexpand.c: Likewise.
1052 * cprop.c: Likewise.
1054 * function.c: Likewise.
1055 * fwprop.c: Likewise.
1057 * gimple-ssa-isolate-paths.c: Likewise.
1058 * gimple-ssa-strength-reduction.c: Likewise.
1059 * ipa-split.c: Likewise.
1060 * loop-init.c: Likewise.
1061 * loop-unroll.c: Likewise.
1062 * lower-subreg.c: Likewise.
1063 * modulo-sched.c: Likewise.
1064 * postreload-gcse.c: Likewise.
1065 * predict.c: Likewise.
1066 * recog.c: Likewise.
1067 * sched-rgn.c: Likewise.
1068 * store-motion.c: Likewise.
1069 * tracer.c: Likewise.
1070 * trans-mem.c: Likewise.
1071 * tree-call-cdce.c: Likewise.
1072 * tree-cfg.c: Likewise.
1073 * tree-cfgcleanup.c: Likewise.
1074 * tree-complex.c: Likewise.
1075 * tree-eh.c: Likewise.
1076 * tree-emutls.c: Likewise.
1077 * tree-if-conv.c: Likewise.
1078 * tree-into-ssa.c: Likewise.
1079 * tree-loop-distribution.c: Likewise.
1080 * tree-object-size.c: Likewise.
1081 * tree-parloops.c: Likewise.
1082 * tree-pass.h: Likewise.
1083 * tree-sra.c: Likewise.
1084 * tree-ssa-ccp.c: Likewise.
1085 * tree-ssa-copy.c: Likewise.
1086 * tree-ssa-copyrename.c: Likewise.
1087 * tree-ssa-dce.c: Likewise.
1088 * tree-ssa-dom.c: Likewise.
1089 * tree-ssa-dse.c: Likewise.
1090 * tree-ssa-forwprop.c: Likewise.
1091 * tree-ssa-ifcombine.c: Likewise.
1092 * tree-ssa-loop-ch.c: Likewise.
1093 * tree-ssa-loop-ivcanon.c: Likewise.
1094 * tree-ssa-loop.c: Likewise.
1095 * tree-ssa-math-opts.c: Likewise.
1096 * tree-ssa-phiopt.c: Likewise.
1097 * tree-ssa-phiprop.c: Likewise.
1098 * tree-ssa-pre.c: Likewise.
1099 * tree-ssa-reassoc.c: Likewise.
1100 * tree-ssa-sink.c: Likewise.
1101 * tree-ssa-strlen.c: Likewise.
1102 * tree-ssa-tail-merge.c: Likewise.
1103 * tree-ssa-uncprop.c: Likewise.
1104 * tree-switch-conversion.c: Likewise.
1105 * tree-tailcall.c: Likewise.
1106 * tree-vect-generic.c: Likewise.
1107 * tree-vectorizer.c: Likewise.
1108 * tree-vrp.c: Likewise.
1110 * var-tracking.c: Likewise.
1111 * bt-load.c: Likewise.
1112 * cfgcleanup.c: Likewise.
1113 * combine-stack-adj.c: Likewise.
1114 * combine.c: Likewise.
1115 * compare-elim.c: Likewise.
1116 * config/epiphany/resolve-sw-modes.c: Likewise.
1117 * config/i386/i386.c: Likewise.
1118 * config/mips/mips.c: Likewise.
1119 * config/s390/s390.c: Likewise.
1120 * config/sh/sh_treg_combine.cc: Likewise.
1121 * config/sparc/sparc.c: Likewise.
1124 * final.c: Likewise.
1125 * ifcvt.c: Likewise.
1126 * mode-switching.c: Likewise.
1127 * passes.c: Likewise.
1128 * postreload.c: Likewise.
1130 * reg-stack.c: Likewise.
1131 * regcprop.c: Likewise.
1132 * regrename.c: Likewise.
1135 2014-05-06 Richard Biener <rguenther@suse.de>
1138 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
1139 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
1141 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
1144 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
1146 2014-05-05 Radovan Obradovic <robradovic@mips.com>
1147 Tom de Vries <tom@codesourcery.com>
1149 * target.def (call_fusage_contains_non_callee_clobbers): New DEFHOOKPOD.
1150 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
1152 (@node Miscellaneous Register Hooks): New node.
1153 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
1154 * doc/tm.texi: Regenerate.
1156 2014-05-05 Marek Polacek <polacek@redhat.com>
1159 * opts.c (common_handle_option): Call error_at instead of warning_at.
1161 2014-05-05 Richard Biener <rguenther@suse.de>
1163 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
1164 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
1165 under the TODO_verify_il umbrella.
1167 2014-05-05 Richard Biener <rguenther@suse.de>
1169 * passes.c (execute_function_todo): Move TODO_verify_flow under
1170 the TODO_verify_ul umbrella.
1172 2014-05-05 Richard Biener <rguenther@suse.de>
1175 * fold-const.c (fold_binary_loc): Consistently avoid
1176 canonicalizing X & CST away from a CST that is the mask
1179 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1181 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
1182 int argument to enum machine_mode.
1183 (picochip_class_max_nregs): Ditto.
1184 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
1185 (picochip_class_max_nregs): Ditto.
1187 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1189 * target.def: Add new target hook.
1190 * doc/tm.texi: Regenerate.
1191 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
1192 * targhooks.c (default_keep_leaf_when_profiled): New function.
1194 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
1195 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
1197 2014-05-05 Bin Cheng <bin.cheng@arm.com>
1199 PR tree-optimization/60363
1200 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
1201 (copy_phi_args): New parameters. Call get_value_locus_in_path.
1202 (update_destination_phis): New parameter.
1203 (create_edge_and_update_destination_phis): Ditto.
1204 (ssa_fix_duplicate_block_edges): Pass new arguments.
1205 (thread_single_edge): Ditto.
1207 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
1209 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
1210 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
1211 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
1212 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
1213 Use RS6000_BTM_HARD_FLOAT.
1214 (BU_MISC_2): Likewise.
1215 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
1216 RS6000_BTM_HARD_FLOAT.
1217 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
1219 (rs6000_invalid_builtin): Add hard floating builtin support.
1220 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
1221 hard float builtins.
1222 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
1224 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1226 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
1227 Add missing function* argument.
1229 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1231 * lra-constraints.c (valid_address_p): Move earlier in file.
1232 Add a constraint argument to the address_info version.
1233 (satisfies_memory_constraint_p): New function.
1234 (satisfies_address_constraint_p): Likewise.
1235 (process_alt_operands, curr_insn_transform): Use them.
1236 (process_address): Pass the constraint to valid_address_p when
1237 checking address operands.
1239 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1241 * config/mips/mips.c (mips_isa_rev): New variable.
1242 (mips_set_architecture): Set it.
1243 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1245 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1246 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1247 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1248 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1249 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
1250 conditions in terms of mips_isa_rev.
1251 (mips_isa_rev): Declare.
1253 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1255 * config/sh/sh-mem.cc: Use tabs instead of spaces.
1256 (prob_unlikely, prob_likely): Make variables const.
1258 2014-05-03 Denis Chertykov <chertykov@gmail.com>
1260 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
1262 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1264 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
1266 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1268 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
1269 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
1270 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
1272 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
1273 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
1275 Replace usage of ROUND_REG with sh_round_reg.
1276 Use CEIL instead of ROUND_ADVANCE.
1278 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1281 * config/sh/sh.c: Include stdlib headers before everything else.
1283 2014-05-02 Jakub Jelinek <jakub@redhat.com>
1285 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
1286 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
1287 (gimplify_adjust_omp_clauses): Simd region is never
1288 directly nested in combined parallel. Instead, for linear
1289 with copyin/copyout, if in combined for simd loop, make decl
1290 firstprivate/lastprivate on OMP_FOR.
1291 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
1292 expand_omp_for_static_chunk): When setting endvar, also set
1293 fd->loop.v to the same value.
1295 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1297 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
1299 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
1301 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
1304 2014-05-02 Marek Polacek <polacek@redhat.com>
1306 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
1308 2014-05-02 Kito Cheng <kito@0xlab.org>
1310 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
1311 to a C expression marco.
1312 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
1313 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
1314 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
1315 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
1316 HONOR_REG_ALLOC_ORDER.
1317 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
1319 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1321 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
1323 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1325 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
1327 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
1329 * tree-if-conv.c (is_cond_scalar_reduction): New function.
1330 (convert_scalar_cond_reduction): Likewise.
1331 (predicate_scalar_phi): Add recognition and transformation
1332 of simple conditioanl reduction to be vectorizable.
1334 2014-05-01 Marek Polacek <polacek@redhat.com>
1337 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
1339 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
1341 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
1342 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
1343 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
1344 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
1345 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
1346 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
1347 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
1348 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
1350 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
1352 * config/arc/arc.opt (mlra): Move comment above option name
1353 to avoid mis-parsing as language options.
1355 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1357 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
1358 * config/sol2.h: ... here.
1359 * config/sol2-10.h: Remove.
1361 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
1362 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
1363 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
1364 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
1365 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
1366 * config/sol2.h: ... here.
1367 (SECTION_NAME_FORMAT): Don't redefine.
1368 (STARTFILE_ARCH32_SPEC): Rename to ...
1369 (STARTFILE_ARCH_SPEC): ... this.
1370 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
1371 * config/sparc/sol2.h: ... here.
1372 (SECTION_NAME_FORMAT): Don't undef.
1373 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
1374 (SUBTARGET_EXTRA_SPECS): Remove.
1375 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
1377 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
1378 (MD_STARTFILE_PREFIX): Remove.
1379 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
1380 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
1381 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
1382 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
1383 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
1384 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
1385 * config/i386/sol2.h: ... here.
1386 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
1387 * config/i386/sol2-bi.h: Remove.
1388 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
1389 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
1391 * config/i386/t-sol2-64: Rename to ...
1392 * config/i386/t-sol2: ... this.
1393 * config/sparc/t-sol2-64: Rename to ...
1394 * config/sparc/t-sol2: ... this.
1396 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
1397 sol2_tm_file_head, sol2_tm_file_tail.
1398 Include ${cpu_type}/sol2.h before sol2.h.
1400 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
1401 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
1402 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
1403 Reflect i386/t-sol2-64 renaming.
1404 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
1405 Reflect sparc/t-sol2-64 renaming.
1407 2014-04-30 Richard Biener <rguenther@suse.de>
1409 * passes.c (execute_function_todo): Move TODO_verify_stmts
1410 and TODO_verify_ssa under the TODO_verify_il umbrella.
1411 * tree-ssa.h (verify_ssa): Adjust prototype.
1412 * tree-ssa.c (verify_ssa): Add parameter to tell whether
1413 we should verify SSA operands.
1414 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
1415 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
1416 whether we should verify whether not throwing stmts have EH info.
1417 * graphite-scop-detection.c (create_sese_edges): Adjust.
1418 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
1419 * tree-eh.c (lower_try_finally_switch): Do not add the
1420 default case label twice.
1422 2014-04-30 Marek Polacek <polacek@redhat.com>
1424 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
1425 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
1426 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
1427 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
1429 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
1431 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
1432 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
1433 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
1434 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
1435 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
1436 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
1437 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
1438 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
1440 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1442 * tree-cfg.c (dump_function_to_file): Dump the return type of
1443 functions, in a line to itself before the function body, mimicking
1444 the layout of a C function.
1446 2014-04-29 Jakub Jelinek <jakub@redhat.com>
1448 PR tree-optimization/60971
1449 * tree-tailcall.c (process_assignment): Reject conversions which
1452 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
1454 * calls.c (initialize_argument_information): Always treat
1455 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
1456 (expand_call): Likewise.
1457 (emit_library_call_calue_1): Likewise.
1458 * expr.c (PUSH_ARGS_REVERSED): Do not define.
1459 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
1462 2014-04-29 Nick Clifton <nickc@redhat.com>
1464 * config/msp430/msp430.md (umulsidi): Fix typo.
1465 (mulhisi3): Enable even inside interrupt handlers.
1466 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
1467 bigger return address pushed in large mode.
1469 2014-04-29 Nick Clifton <nickc@redhat.com>
1471 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
1472 (arc_init_reg_tables): Use a machine_mode enum to iterate over
1474 * config/m32r/m32r.c (init_reg_tables): Likewise.
1475 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
1476 enum to hold the modes.
1478 2014-04-29 Richard Biener <rguenther@suse.de>
1480 * dominance.c (free_dominance_info): Add overload with
1482 (dom_info_state): Likewise.
1483 (dom_info_available_p): Likewise.
1484 * basic-block.h (free_dominance_info, dom_info_state,
1485 dom_info_available_p): Declare overloads.
1486 * passes.c (execute_function_todo): Verify that verifiers
1487 don't change dominator info state. Drop dominator info
1488 for IPA pass invocations.
1489 * cgraph.c (release_function_body): Restore asserts that
1490 dominator information is released.
1492 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
1494 * doc/invoke.texi: Fix typo.
1495 * tree-vrp.c: Fix typos.
1496 * gimple.c (infer_nonnull_range): Reorder operands of an &&
1499 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1501 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
1503 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1505 * config/aarch64/aarch64-builtins.c
1506 (aarch64_types_storestruct_lane_qualifiers): New.
1507 (TYPES_STORESTRUCT_LANE): Likewise.
1508 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
1509 (st3_lane): Likewise.
1510 (st4_lane): Likewise.
1511 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
1512 (vec_store_lanesci_lane<mode>): Likewise.
1513 (vec_store_lanesxi_lane<mode>): Likewise.
1514 (aarch64_st2_lane<VQ:mode>): Likewise.
1515 (aarch64_st3_lane<VQ:mode>): Likewise.
1516 (aarch64_st4_lane<VQ:mode>): Likewise.
1517 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
1518 * config/aarch64/arm_neon.h
1519 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
1520 use new macro arguments.
1521 (__ST3_LANE_FUNC): Likewise.
1522 (__ST4_LANE_FUNC): Likewise.
1523 * config/aarch64/iterators.md (V_TWO_ELEM): New.
1524 (V_THREE_ELEM): Likewise.
1525 (V_FOUR_ELEM): Likewise.
1527 2014-04-28 David Malcolm <dmalcolm@redhat.com>
1529 * doc/gimple.texi: Replace the description of the now-defunct
1530 union gimple_statement_d with a diagram showing the
1531 gimple_statement_base class hierarchy and its relationships to
1532 the GSS_ and GIMPLE_ enums.
1534 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1536 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
1537 * config/aarch64/aarch64.c
1538 (aarch64_cannot_change_mode_class): Weaken conditions.
1539 (aarch64_modes_tieable_p): New.
1540 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
1542 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
1544 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
1545 (loadsync_<mode>): Change mode.
1546 (load_quadpti, store_quadpti): New.
1547 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
1548 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
1550 2014-04-28 Martin Jambor <mjambor@suse.cz>
1552 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
1553 same alias type as the original statement.
1554 (subreplacement_assignment_data): New type.
1555 (handle_unscalarized_data_in_subtree): New type of parameter,
1556 generate new memory accesses with same alias type as the original
1558 (load_assign_lhs_subreplacements): Likewise.
1559 (sra_modify_constructor_assign): Generate new memory accesses with
1560 same alias type as the original statement.
1562 2014-04-28 Richard Biener <rguenther@suse.de>
1564 * tree-pass.h (TODO_verify_il): Define.
1565 (TODO_verify_all): Complete properly.
1566 * passes.c (execute_function_todo): Move existing loop-closed
1567 SSA verification under TODO_verify_il.
1568 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
1569 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
1570 Fix tree sharing issue.
1572 2014-04-28 Richard Biener <rguenther@suse.de>
1575 * builtins.def (DEF_C11_BUILTIN): Add.
1576 (BUILT_IN_ALIGNED_ALLOC): Likewise.
1577 * coretypes.h (enum function_class): Add function_c11_misc.
1578 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1579 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
1580 (call_may_clobber_ref_p_1): Likewise.
1581 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
1582 (mark_all_reaching_defs_necessary_1): Likewise.
1583 (propagate_necessity): Likewise.
1584 (eliminate_unnecessary_stmts): Likewise.
1585 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
1587 2014-04-28 Richard Biener <rguenther@suse.de>
1589 * tree-vrp.c (vrp_var_may_overflow): Remove.
1590 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
1591 with overflow immediately bump to one before that value and
1592 let iteration figure out overflow status.
1594 2014-04-28 Richard Biener <rguenther@suse.de>
1596 * configure.ac: Do valgrind header checks unconditionally.
1597 Add --enable-valgrind-annotations.
1598 * system.h: Guard valgrind header inclusion with
1599 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
1600 * alloc-pool.c (pool_alloc, pool_free): Use
1601 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
1602 to guard possibly dead code.
1603 * config.in: Regenerated.
1604 * configure: Likewise.
1606 2014-04-28 Jeff Law <law@redhat.com>
1608 PR tree-optimization/60902
1609 * tree-ssa-threadedge.c
1610 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
1611 over real defs when invalidating outputs from statements that do not
1612 produce useful outputs for threading.
1614 2014-04-28 Richard Biener <rguenther@suse.de>
1616 PR tree-optimization/60979
1617 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
1618 SCOPs that end in a block with a successor with abnormal
1621 2014-04-28 Richard Biener <rguenther@suse.de>
1623 * tree-pass.h (execute_pass_list): Adjust prototype.
1624 * passes.c (pass_manager::execute_early_local_passes):
1626 (do_per_function): Change callback signature, push all actual
1627 work to the callbals.
1628 (do_per_function_toporder): Likewise.
1629 (execute_function_dump): Adjust.
1630 (execute_function_todo): Likewise.
1631 (clear_last_verified): Likewise.
1632 (verify_curr_properties): Likewise.
1633 (update_properties_after_pass): Likewise.
1634 (execute_pass_list_1): Split out from ...
1635 (execute_pass_list): ... here. Adjust.
1636 (execute_ipa_pass_list): Likewise.
1637 * cgraphunit.c (cgraph_add_new_function): Adjust.
1638 (analyze_function): Likewise.
1639 (expand_function): Likewise.
1640 * cgraph.c (release_function_body): Free dominance info
1641 here instead of asserting it was magically freed elsewhere.
1643 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
1645 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
1646 * configure: Regenerate.
1647 * config/sparc/sparc.opt (muser-mode): New option.
1648 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
1650 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
1651 * doc/invoke.texi (SPARC options): Document -muser-mode.
1653 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
1655 * cselib.c (find_slot_memmode): Delete.
1656 (cselib_hasher): Change compare_type to a struct.
1657 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
1659 (preserve_constants_and_equivs): Adjust for new compare_type.
1660 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
1661 (wrap_constant): Delete.
1662 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
1664 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
1666 * doc/install.texi (Building with profile feedback): Remove
1669 2014-04-26 Tom de Vries <tom@codesourcery.com>
1671 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
1674 2014-04-25 Cary Coutant <ccoutant@google.com>
1677 * dwarf2out.c (should_move_die_to_comdat): A type definition
1678 can contain a subprogram definition, but don't move it to a
1680 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
1681 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
1683 (clone_tree_hash): Rename to...
1684 (clone_tree_partial): ...this; change callers. Copy
1685 DW_TAG_subprogram DIEs as declarations.
1686 (copy_decls_walk): Don't copy children of a declaration into a
1689 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
1692 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
1695 2014-04-25 Jiong Wang <jiong.wang@arm.com>
1697 * config/arm/predicates.md (call_insn_operand): Add long_call check.
1698 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
1700 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
1703 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1705 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
1708 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1710 PR tree-optimization/60930
1711 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
1712 creating a multiply candidate by folding two constant
1713 multiplicands when the result overflows.
1715 2014-04-25 Jakub Jelinek <jakub@redhat.com>
1717 PR tree-optimization/60960
1718 * tree-vect-generic.c (expand_vector_operation): Only call
1719 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
1721 2014-04-25 Tom de Vries <tom@codesourcery.com>
1723 * expr.c (clobber_reg_mode): New function.
1724 * expr.h (clobber_reg): New function.
1726 2014-04-25 Tom de Vries <tom@codesourcery.com>
1728 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
1731 2014-04-25 Radovan Obradovic <robradovic@mips.com>
1732 Tom de Vries <tom@codesourcery.com>
1734 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
1736 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
1737 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
1738 new argument to find_all_hard_reg_sets call.
1740 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1742 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
1743 Use HOST_WIDE_INT_C for mask literal.
1744 (aarch_rev16_shleft_mask_imm_p): Likewise.
1746 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
1749 * config/sparc/sparc.md (ashlsi3_extend): Delete.
1751 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
1753 PR preprocessor/56540
1754 * config/i386/i386-c.c (ix86_target_macros): Define
1755 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
1757 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1759 * configure.ac (tga_func): Remove.
1760 (LIB_TLS_SPEC): Remove.
1761 * configure: Regenerate.
1762 * config.in: Regenerate.
1763 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
1765 2014-04-25 Richard Biener <rguenther@suse.de>
1768 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
1769 call stmt use/clobber sets during stmt walk instead of
1770 walking the possibly incomplete set of caller edges.
1772 2014-04-25 Richard Biener <rguenther@suse.de>
1775 * passes.c (apply_ipa_transforms): Inline into only caller ...
1776 (execute_one_pass): ... here. Properly bring in function
1777 bodies for nodes we want to apply IPA transforms to.
1779 2014-04-24 Cong Hou <congh@google.com>
1781 PR tree-optimization/60896
1782 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
1783 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
1784 (vect_mark_pattern_stmts): Set the def type of all statements in
1785 PATTERN_DEF_SEQ as vect_internal_def.
1787 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
1789 * doc/extend.texi (PowerPC Built-in Functions): Document new
1790 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
1792 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
1794 * config/rs6000/predicates.md (const_0_to_3_operand): New
1795 predicate to match 0..3 integer constants.
1797 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
1798 to support adding miscellaneous builtin functions.
1799 (BU_DFP_MISC_2): Likewise.
1800 (BU_P7_MISC_1): Likewise.
1801 (BU_P7_MISC_2): Likewise.
1802 (BU_P8V_MISC_3): Likewise.
1803 (BU_MISC_1): Likewise.
1804 (BU_MISC_2): Likewise.
1805 (DIVWE): Add extended divide builtin functions.
1808 (DIVWEUO): Likewise.
1812 (DIVDEUO): Likewise.
1813 (DXEX): Add decimal floating-point builtin functions.
1816 (DDEDPDQ): Likewise.
1818 (DENBCDQ): Likewise.
1825 (CDTBCD): Add new BCD builtin functions.
1829 (BCDADD_LT): Likewise.
1830 (BCDADD_EQ): Likewise.
1831 (BCDADD_GT): Likewise.
1832 (BCDADD_OV): Likewise.
1834 (BCDSUB_LT): Likewise.
1835 (BCDSUB_EQ): Likewise.
1836 (BCDSUB_GT): Likewise.
1837 (BCDSUB_OV): Likewise.
1838 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
1839 (UNPACK_TD): Likewise.
1840 (PACK_TF): Likewise.
1841 (UNPACK_TF): Likewise.
1842 (UNPACK_TF_0): Likewise.
1843 (UNPACK_TF_1): Likewise.
1844 (PACK_V1TI): Likewise.
1845 (UNPACK_V1TI): Likewise.
1847 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
1848 support for decimal floating point builtin functions.
1849 (rs6000_expand_ternop_builtin): Add checks for the new builtin
1850 functions that take constant arguments.
1851 (rs6000_invalid_builtin): Add decimal floating point builtin
1853 (rs6000_init_builtins): Setup long double, _Decimal64, and
1854 _Decimal128 types for new builtin functions.
1855 (builtin_function_type): Set the unsigned flags appropriately for
1856 the new builtin functions.
1857 (rs6000_opt_masks): Add support for decimal floating point builtin
1860 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
1861 floating point builtin functions.
1862 (RS6000_BTM_COMMON): Likewise.
1863 (RS6000_BTI_long_double): Likewise.
1864 (RS6000_BTI_dfloat64): Likewise.
1865 (RS6000_BTI_dfloat128): Likewise.
1866 (long_double_type_internal_node): Likewise.
1867 (dfloat64_type_internal_node): Likewise.
1868 (dfloat128_type_internal_node): Likewise.
1870 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
1871 2.07 bcd arithmetic instructions.
1872 (UNSPEC_BCDSUB): Likewise.
1873 (UNSPEC_BCD_OVERFLOW): Likewise.
1874 (UNSPEC_BCD_ADD_SUB): Likewise.
1875 (bcd_add_sub): Likewise.
1876 (BCD_TEST): Likewise.
1877 (bcd<bcd_add_sub>): Likewise.
1878 (bcd<bcd_add_sub>_test): Likewise.
1879 (bcd<bcd_add_sub>_test2): Likewise.
1880 (bcd<bcd_add_sub>_<code>): Likewise.
1881 (peephole2 for combined bcd ops): Likewise.
1883 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
1884 decimal floating point builtin functions.
1885 (UNSPEC_DENBCD): Likewise.
1886 (UNSPEC_DXEX): Likewise.
1887 (UNSPEC_DIEX): Likewise.
1888 (UNSPEC_DSCLI): Likewise.
1889 (UNSPEC_DSCRI): Likewise.
1890 (D64_D128): Likewise.
1891 (dfp_suffix): Likewise.
1892 (dfp_ddedpd_<mode>): Likewise.
1893 (dfp_denbcd_<mode>): Likewise.
1894 (dfp_dxex_<mode>): Likewise.
1895 (dfp_diex_<mode>): Likewise.
1896 (dfp_dscli_<mode>): Likewise.
1897 (dfp_dscri_<mode>): Likewise.
1899 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
1901 (UNSPEC_CDTBCD): Likewise.
1902 (UNSPEC_CBCDTD): Likewise.
1903 (UNSPEC_DIVE): Add support for new extended divide builtin
1905 (UNSPEC_DIVEO): Likewise.
1906 (UNSPEC_DIVEU): Likewise.
1907 (UNSPEC_DIVEUO): Likewise.
1908 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
1909 pack/unpack 128-bit types.
1910 (UNSPEC_PACK_128BIT): Likewise.
1911 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
1912 (udiv<mode>3): Use idiv_ldiv mode attribute.
1913 (div<mode>3): Likewise.
1914 (addg6s): Add new BCD builtin functions.
1917 (UNSPEC_DIV_EXTEND): Add support for new extended divide
1919 (div_extend): Likewise.
1920 (div<div_extend>_<mode>"): Likewise.
1921 (FP128_64): Add support for new builtin functions to pack/unpack
1923 (unpack<mode>): Likewise.
1924 (unpacktf_0): Likewise.
1925 (unpacktf_1): Likewise.
1926 (unpack<mode>_dm): Likewise.
1927 (unpack<mode>_nodm): Likewise.
1928 (pack<mode>): Likewise.
1929 (unpackv1ti): Likewise.
1930 (packv1ti): Likewise.
1932 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
1934 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
1937 2014-04-24 Jakub Jelinek <jakub@redhat.com>
1939 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
1940 * gimplify.c (omp_is_private): Change last argument's type to int.
1941 Only diagnose lastprivate if the simd argument is 1, only diagnose
1942 linear if the simd argument is 2.
1943 (gimplify_omp_for): Adjust omp_is_private callers. When adding
1944 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
1945 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
1946 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
1947 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
1948 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
1949 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
1950 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
1951 * tree-nested.c (convert_nonlocal_omp_clauses,
1952 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
1954 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
1957 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
1960 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
1962 * flag-types.h (enum ivar_visibility): Add.
1964 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
1966 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
1967 function * argument.
1969 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
1971 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
1973 2014-04-24 Radovan Obradovic <robradovic@mips.com>
1974 Tom de Vries <tom@codesourcery.com>
1976 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
1977 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
1979 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
1980 * emit-rtl.c (try_split): Same.
1982 2014-04-24 Radovan Obradovic <robradovic@mips.com>
1983 Tom de Vries <tom@codesourcery.com>
1985 * common.opt (fuse-caller-save): New option.
1987 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
1989 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
1992 2014-04-24 Richard Biener <rguenther@suse.de>
1994 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
1995 during TER and instead use the sepops interface for expanding
1996 non-GIMPLE_SINGLE_RHS.
1998 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2000 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
2001 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
2003 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2005 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
2006 assembler 64-bit option.
2007 * configure: Regenerate.
2009 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2011 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
2012 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
2013 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
2014 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
2015 (TARGET_CRYPTO): Take TARGET_SIMD into account.
2017 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2019 * config/aarch64/aarch64-builtins.c
2020 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
2021 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
2022 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
2023 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
2025 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
2026 (Vrevsuff): New mode attribute.
2028 2014-04-24 Terry Guo <terry.guo@arm.com>
2030 * config/arm/arm.h (machine_function): Define variable
2031 after_arm_reorg here.
2032 * config/arm/arm.c (after_arm_reorg): Remove the definition.
2033 (arm_split_constant): Update the way to access variable
2036 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
2038 2014-04-23 Tom de Vries <tom@codesourcery.com>
2040 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
2042 2014-04-23 David Malcolm <dmalcolm@redhat.com>
2044 * is-a.h: Update comments to reflect the following changes to the
2045 "pointerness" of the API, making the template parameter match the
2046 return type, allowing use of is-a.h with typedefs of pointers.
2047 (is_a_helper::cast): Return a T rather then a pointer to a T, so
2048 that the return type matches the parameter to the is_a_helper.
2050 (dyn_cast): Likewise.
2052 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
2053 pointer from the is-a.h API.
2055 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
2056 (is_a_helper <cgraph_node *>::test): ...this, matching change to
2058 (is_a_helper <varpool_node>::test): Likewise, convert to...
2059 (is_a_helper <varpool_node *>::test): ...this.
2061 (varpool_first_variable): Update for removal of implicit pointer
2062 from the is-a.h API.
2063 (varpool_next_variable): Likewise.
2064 (varpool_first_static_initializer): Likewise.
2065 (varpool_next_static_initializer): Likewise.
2066 (varpool_first_defined_variable): Likewise.
2067 (varpool_next_defined_variable): Likewise.
2068 (cgraph_first_defined_function): Likewise.
2069 (cgraph_next_defined_function): Likewise.
2070 (cgraph_first_function): Likewise.
2071 (cgraph_next_function): Likewise.
2072 (cgraph_first_function_with_gimple_body): Likewise.
2073 (cgraph_next_function_with_gimple_body): Likewise.
2074 (cgraph_alias_target): Likewise.
2075 (varpool_alias_target): Likewise.
2076 (cgraph_function_or_thunk_node): Likewise.
2077 (varpool_variable_node): Likewise.
2078 (symtab_real_symbol_p): Likewise.
2079 * cgraphunit.c (referred_to_p): Likewise.
2080 (analyze_functions): Likewise.
2081 (handle_alias_pairs): Likewise.
2082 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2083 * gimple-ssa.h (gimple_vuse_op): Likewise.
2084 (gimple_vdef_op): Likewise.
2085 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
2086 * gimple.c (gimple_build_asm_1): Likewise.
2087 (gimple_build_try): Likewise.
2088 (gimple_build_resx): Likewise.
2089 (gimple_build_eh_dispatch): Likewise.
2090 (gimple_build_omp_for): Likewise.
2091 (gimple_omp_for_set_clauses): Likewise.
2093 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
2094 (is_a_helper <gimple_statement_asm *>::test): ...this.
2095 (is_a_helper <gimple_statement_bind>::test): Convert to...
2096 (is_a_helper <gimple_statement_bind *>::test): ...this.
2097 (is_a_helper <gimple_statement_call>::test): Convert to...
2098 (is_a_helper <gimple_statement_call *>::test): ...this.
2099 (is_a_helper <gimple_statement_catch>::test): Convert to...
2100 (is_a_helper <gimple_statement_catch *>::test): ...this.
2101 (is_a_helper <gimple_statement_resx>::test): Convert to...
2102 (is_a_helper <gimple_statement_resx *>::test): ...this.
2103 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
2104 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
2105 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
2106 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
2107 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
2108 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
2109 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
2110 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
2111 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
2112 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
2113 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
2114 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
2115 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
2116 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
2117 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
2118 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
2119 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
2120 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
2121 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
2122 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
2123 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
2124 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
2125 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
2126 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
2127 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
2128 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
2129 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
2130 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
2131 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
2132 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
2133 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
2134 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
2135 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
2136 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
2137 (is_a_helper <gimple_statement_phi>::test): Convert to...
2138 (is_a_helper <gimple_statement_phi *>::test): ...this.
2139 (is_a_helper <gimple_statement_transaction>::test): Convert to...
2140 (is_a_helper <gimple_statement_transaction *>::test): ...this.
2141 (is_a_helper <gimple_statement_try>::test): Convert to...
2142 (is_a_helper <gimple_statement_try *>::test): ...this.
2143 (is_a_helper <gimple_statement_wce>::test): Convert to...
2144 (is_a_helper <gimple_statement_wce *>::test): ...this.
2145 (is_a_helper <const gimple_statement_asm>::test): Convert to...
2146 (is_a_helper <const gimple_statement_asm *>::test): ...this.
2147 (is_a_helper <const gimple_statement_bind>::test): Convert to...
2148 (is_a_helper <const gimple_statement_bind *>::test): ...this.
2149 (is_a_helper <const gimple_statement_call>::test): Convert to...
2150 (is_a_helper <const gimple_statement_call *>::test): ...this.
2151 (is_a_helper <const gimple_statement_catch>::test): Convert to...
2152 (is_a_helper <const gimple_statement_catch *>::test): ...this.
2153 (is_a_helper <const gimple_statement_resx>::test): Convert to...
2154 (is_a_helper <const gimple_statement_resx *>::test): ...this.
2155 (is_a_helper <const gimple_statement_eh_dispatch>::test):
2157 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
2159 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
2161 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
2162 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
2164 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
2166 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
2168 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
2170 (is_a_helper <const gimple_statement_omp_return>::test): Convert
2172 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
2173 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
2175 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
2176 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
2178 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
2179 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
2180 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
2181 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
2183 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
2184 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
2186 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
2187 (is_a_helper <const gimple_statement_omp_target>::test): Convert
2189 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
2190 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
2192 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
2193 (is_a_helper <const gimple_statement_omp_single>::test): Convert
2195 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
2196 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
2198 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
2199 (is_a_helper <const gimple_statement_omp_task>::test): Convert
2201 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
2202 (is_a_helper <const gimple_statement_phi>::test): Convert to...
2203 (is_a_helper <const gimple_statement_phi *>::test): ...this.
2204 (is_a_helper <const gimple_statement_transaction>::test): Convert
2206 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
2207 (is_a_helper <const gimple_statement_with_ops>::test): Convert
2209 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
2210 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
2211 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
2212 (is_a_helper <const gimple_statement_with_memory_ops>::test):
2214 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
2216 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
2218 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
2220 (gimple_use_ops): Update for removal of implicit pointer from the
2222 (gimple_set_use_ops): Likewise.
2223 (gimple_vuse): Likewise.
2224 (gimple_vdef): Likewise.
2225 (gimple_vuse_ptr): Likewise.
2226 (gimple_vdef_ptr): Likewise.
2227 (gimple_set_vuse): Likewise.
2228 (gimple_set_vdef): Likewise.
2229 (gimple_omp_return_set_lhs): Likewise.
2230 (gimple_omp_return_lhs): Likewise.
2231 (gimple_omp_return_lhs_ptr): Likewise.
2232 (gimple_call_fntype): Likewise.
2233 (gimple_call_set_fntype): Likewise.
2234 (gimple_call_set_internal_fn): Likewise.
2235 (gimple_call_use_set): Likewise.
2236 (gimple_call_clobber_set): Likewise.
2237 (gimple_bind_vars): Likewise.
2238 (gimple_bind_set_vars): Likewise.
2239 (gimple_bind_body_ptr): Likewise.
2240 (gimple_bind_set_body): Likewise.
2241 (gimple_bind_add_stmt): Likewise.
2242 (gimple_bind_block): Likewise.
2243 (gimple_bind_set_block): Likewise.
2244 (gimple_asm_ninputs): Likewise.
2245 (gimple_asm_noutputs): Likewise.
2246 (gimple_asm_nclobbers): Likewise.
2247 (gimple_asm_nlabels): Likewise.
2248 (gimple_asm_input_op): Likewise.
2249 (gimple_asm_input_op_ptr): Likewise.
2250 (gimple_asm_output_op): Likewise.
2251 (gimple_asm_output_op_ptr): Likewise.
2252 (gimple_asm_set_output_op): Likewise.
2253 (gimple_asm_clobber_op): Likewise.
2254 (gimple_asm_set_clobber_op): Likewise.
2255 (gimple_asm_label_op): Likewise.
2256 (gimple_asm_set_label_op): Likewise.
2257 (gimple_asm_string): Likewise.
2258 (gimple_catch_types): Likewise.
2259 (gimple_catch_types_ptr): Likewise.
2260 (gimple_catch_handler_ptr): Likewise.
2261 (gimple_catch_set_types): Likewise.
2262 (gimple_catch_set_handler): Likewise.
2263 (gimple_eh_filter_types): Likewise.
2264 (gimple_eh_filter_types_ptr): Likewise.
2265 (gimple_eh_filter_failure_ptr): Likewise.
2266 (gimple_eh_filter_set_types): Likewise.
2267 (gimple_eh_filter_set_failure): Likewise.
2268 (gimple_eh_must_not_throw_fndecl): Likewise.
2269 (gimple_eh_must_not_throw_set_fndecl): Likewise.
2270 (gimple_eh_else_n_body_ptr): Likewise.
2271 (gimple_eh_else_e_body_ptr): Likewise.
2272 (gimple_eh_else_set_n_body): Likewise.
2273 (gimple_eh_else_set_e_body): Likewise.
2274 (gimple_try_eval_ptr): Likewise.
2275 (gimple_try_cleanup_ptr): Likewise.
2276 (gimple_try_set_eval): Likewise.
2277 (gimple_try_set_cleanup): Likewise.
2278 (gimple_wce_cleanup_ptr): Likewise.
2279 (gimple_wce_set_cleanup): Likewise.
2280 (gimple_phi_capacity): Likewise.
2281 (gimple_phi_num_args): Likewise.
2282 (gimple_phi_result): Likewise.
2283 (gimple_phi_result_ptr): Likewise.
2284 (gimple_phi_set_result): Likewise.
2285 (gimple_phi_arg): Likewise.
2286 (gimple_phi_set_arg): Likewise.
2287 (gimple_resx_region): Likewise.
2288 (gimple_resx_set_region): Likewise.
2289 (gimple_eh_dispatch_region): Likewise.
2290 (gimple_eh_dispatch_set_region): Likewise.
2291 (gimple_omp_critical_name): Likewise.
2292 (gimple_omp_critical_name_ptr): Likewise.
2293 (gimple_omp_critical_set_name): Likewise.
2294 (gimple_omp_for_clauses): Likewise.
2295 (gimple_omp_for_clauses_ptr): Likewise.
2296 (gimple_omp_for_set_clauses): Likewise.
2297 (gimple_omp_for_collapse): Likewise.
2298 (gimple_omp_for_index): Likewise.
2299 (gimple_omp_for_index_ptr): Likewise.
2300 (gimple_omp_for_set_index): Likewise.
2301 (gimple_omp_for_initial): Likewise.
2302 (gimple_omp_for_initial_ptr): Likewise.
2303 (gimple_omp_for_set_initial): Likewise.
2304 (gimple_omp_for_final): Likewise.
2305 (gimple_omp_for_final_ptr): Likewise.
2306 (gimple_omp_for_set_final): Likewise.
2307 (gimple_omp_for_incr): Likewise.
2308 (gimple_omp_for_incr_ptr): Likewise.
2309 (gimple_omp_for_set_incr): Likewise.
2310 (gimple_omp_for_pre_body_ptr): Likewise.
2311 (gimple_omp_for_set_pre_body): Likewise.
2312 (gimple_omp_parallel_clauses): Likewise.
2313 (gimple_omp_parallel_clauses_ptr): Likewise.
2314 (gimple_omp_parallel_set_clauses): Likewise.
2315 (gimple_omp_parallel_child_fn): Likewise.
2316 (gimple_omp_parallel_child_fn_ptr): Likewise.
2317 (gimple_omp_parallel_set_child_fn): Likewise.
2318 (gimple_omp_parallel_data_arg): Likewise.
2319 (gimple_omp_parallel_data_arg_ptr): Likewise.
2320 (gimple_omp_parallel_set_data_arg): Likewise.
2321 (gimple_omp_task_clauses): Likewise.
2322 (gimple_omp_task_clauses_ptr): Likewise.
2323 (gimple_omp_task_set_clauses): Likewise.
2324 (gimple_omp_task_child_fn): Likewise.
2325 (gimple_omp_task_child_fn_ptr): Likewise.
2326 (gimple_omp_task_set_child_fn): Likewise.
2327 (gimple_omp_task_data_arg): Likewise.
2328 (gimple_omp_task_data_arg_ptr): Likewise.
2329 (gimple_omp_task_set_data_arg): Likewise.
2330 (gimple_omp_taskreg_clauses): Likewise.
2331 (gimple_omp_taskreg_clauses_ptr): Likewise.
2332 (gimple_omp_taskreg_set_clauses): Likewise.
2333 (gimple_omp_taskreg_child_fn): Likewise.
2334 (gimple_omp_taskreg_child_fn_ptr): Likewise.
2335 (gimple_omp_taskreg_set_child_fn): Likewise.
2336 (gimple_omp_taskreg_data_arg): Likewise.
2337 (gimple_omp_taskreg_data_arg_ptr): Likewise.
2338 (gimple_omp_taskreg_set_data_arg): Likewise.
2339 (gimple_omp_task_copy_fn): Likewise.
2340 (gimple_omp_task_copy_fn_ptr): Likewise.
2341 (gimple_omp_task_set_copy_fn): Likewise.
2342 (gimple_omp_task_arg_size): Likewise.
2343 (gimple_omp_task_arg_size_ptr): Likewise.
2344 (gimple_omp_task_set_arg_size): Likewise.
2345 (gimple_omp_task_arg_align): Likewise.
2346 (gimple_omp_task_arg_align_ptr): Likewise.
2347 (gimple_omp_task_set_arg_align): Likewise.
2348 (gimple_omp_single_clauses): Likewise.
2349 (gimple_omp_single_clauses_ptr): Likewise.
2350 (gimple_omp_single_set_clauses): Likewise.
2351 (gimple_omp_target_clauses): Likewise.
2352 (gimple_omp_target_clauses_ptr): Likewise.
2353 (gimple_omp_target_set_clauses): Likewise.
2354 (gimple_omp_target_child_fn): Likewise.
2355 (gimple_omp_target_child_fn_ptr): Likewise.
2356 (gimple_omp_target_set_child_fn): Likewise.
2357 (gimple_omp_target_data_arg): Likewise.
2358 (gimple_omp_target_data_arg_ptr): Likewise.
2359 (gimple_omp_target_set_data_arg): Likewise.
2360 (gimple_omp_teams_clauses): Likewise.
2361 (gimple_omp_teams_clauses_ptr): Likewise.
2362 (gimple_omp_teams_set_clauses): Likewise.
2363 (gimple_omp_sections_clauses): Likewise.
2364 (gimple_omp_sections_clauses_ptr): Likewise.
2365 (gimple_omp_sections_set_clauses): Likewise.
2366 (gimple_omp_sections_control): Likewise.
2367 (gimple_omp_sections_control_ptr): Likewise.
2368 (gimple_omp_sections_set_control): Likewise.
2369 (gimple_omp_for_set_cond): Likewise.
2370 (gimple_omp_for_cond): Likewise.
2371 (gimple_omp_atomic_store_set_val): Likewise.
2372 (gimple_omp_atomic_store_val): Likewise.
2373 (gimple_omp_atomic_store_val_ptr): Likewise.
2374 (gimple_omp_atomic_load_set_lhs): Likewise.
2375 (gimple_omp_atomic_load_lhs): Likewise.
2376 (gimple_omp_atomic_load_lhs_ptr): Likewise.
2377 (gimple_omp_atomic_load_set_rhs): Likewise.
2378 (gimple_omp_atomic_load_rhs): Likewise.
2379 (gimple_omp_atomic_load_rhs_ptr): Likewise.
2380 (gimple_omp_continue_control_def): Likewise.
2381 (gimple_omp_continue_control_def_ptr): Likewise.
2382 (gimple_omp_continue_set_control_def): Likewise.
2383 (gimple_omp_continue_control_use): Likewise.
2384 (gimple_omp_continue_control_use_ptr): Likewise.
2385 (gimple_omp_continue_set_control_use): Likewise.
2386 (gimple_transaction_body_ptr): Likewise.
2387 (gimple_transaction_label): Likewise.
2388 (gimple_transaction_label_ptr): Likewise.
2389 (gimple_transaction_set_body): Likewise.
2390 (gimple_transaction_set_label): Likewise.
2392 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
2393 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2394 * ipa-ref.c (ipa_record_reference): Likewise.
2395 * ipa-reference.c (analyze_function): Likewise.
2396 (ipa_reference_write_optimization_summary): Likewise.
2397 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
2398 (address_taken_from_non_vtable_p): Likewise.
2399 (comdat_can_be_unshared_p_1): Likewise.
2400 * lto-cgraph.c (lto_output_ref): Likewise.
2401 (add_references): Likewise.
2402 (compute_ltrans_boundary): Likewise.
2403 (output_symtab): Likewise.
2404 (input_ref): Likewise.
2405 (input_cgraph_1): Likewise.
2406 (output_cgraph_opt_summary): Likewise.
2407 * lto-streamer-out.c (lto_output): Likewise.
2408 (output_symbol_p): Likewise.
2409 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
2410 (lsei_start_function_in_partition): Likewise.
2411 (lsei_next_variable_in_partition): Likewise.
2412 (lsei_start_variable_in_partition): Likewise.
2413 * symtab.c (insert_to_assembler_name_hash): Likewise.
2414 (unlink_from_assembler_name_hash): Likewise.
2415 (symtab_unregister_node): Likewise.
2416 (symtab_remove_node): Likewise.
2417 (dump_symtab_node): Likewise.
2418 (verify_symtab_base): Likewise.
2419 (verify_symtab_node): Likewise.
2420 (symtab_make_decl_local): Likewise.
2421 (symtab_alias_ultimate_target): Likewise.
2422 (symtab_resolve_alias): Likewise.
2423 (symtab_get_symbol_partitioning_class): Likewise.
2424 * tree-phinodes.c (allocate_phi_node): Likewise.
2425 (reserve_phi_args_for_new_edge): Likewise.
2426 (remove_phi_args): Likewise.
2427 * varpool.c (varpool_node_for_asm): Likewise.
2428 (varpool_remove_unreferenced_decls): Likewise.
2430 2014-04-23 Jeff Law <law@redhat.com>
2432 PR tree-optimization/60902
2433 * tree-ssa-threadedge.c
2434 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
2435 invalidate outputs from statements that do not produce useful
2436 outputs for threading.
2438 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2440 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
2441 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
2442 machine descriptions for Stack Smashing Protector.
2444 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
2446 * aarch64.md (<optab>_rol<mode>3): New pattern.
2447 (<optab>_rolsi3_uxtw): Likewise.
2448 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
2450 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
2452 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
2453 (arm_cortex_a12_tune): Likewise.
2455 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2457 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
2459 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2461 * config/arm/arm.md (arm_rev16si2): New pattern.
2462 (arm_rev16si2_alt): Likewise.
2463 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
2465 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2467 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
2468 (rev16<mode>2_alt): Likewise.
2469 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
2470 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
2471 (aarch_rev16_shleft_mask_imm_p): Likewise.
2472 (aarch_rev16_p_1): Likewise.
2473 (aarch_rev16_p): Likewise.
2474 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
2475 (aarch_rev16_shright_mask_imm_p): Likewise.
2476 (aarch_rev16_shleft_mask_imm_p): Likewise.
2478 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2480 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
2481 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
2483 (cortex_a53_extra_costs): Likewise.
2484 (cortex_a57_extra_costs): Likewise.
2485 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
2486 (cortexa7_extra_costs): Likewise.
2487 (cortexa8_extra_costs): Likewise.
2488 (cortexa12_extra_costs): Likewise.
2489 (cortexa15_extra_costs): Likewise.
2490 (v7m_extra_costs): Likewise.
2491 (arm_new_rtx_costs): Handle BSWAP.
2493 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2495 * config/arm/arm.c (cortexa8_extra_costs): New table.
2496 (arm_cortex_a8_tune): New tuning struct.
2497 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
2499 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2501 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
2503 2014-04-23 Richard Biener <rguenther@suse.de>
2505 * Makefile.in (OBJS): Remove loop-unswitch.o.
2506 * tree-pass.h (make_pass_rtl_unswitch): Remove.
2507 * passes.def (pass_rtl_unswitch): Likewise.
2508 * loop-init.c (gate_rtl_unswitch): Likewise.
2509 (rtl_unswitch): Likewise.
2510 (pass_data_rtl_unswitch): Likewise.
2511 (pass_rtl_unswitch): Likewise.
2512 (make_pass_rtl_unswitch): Likewise.
2513 * rtl.h (reversed_condition): Likewise.
2514 (compare_and_jump_seq): Likewise.
2515 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
2517 * loop-unroll.c (compare_and_jump_seq): Likewise.
2519 2014-04-23 Richard Biener <rguenther@suse.de>
2521 PR tree-optimization/60903
2522 * tree-ssa-loop-im.c (analyze_memory_references): Remove
2523 commented code block.
2524 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
2525 loop flags to newly created BBs and edges.
2527 2014-04-23 Nick Clifton <nickc@redhat.com>
2529 * config/msp430/msp430.c (msp430_handle_option): Move function
2531 (msp430_option_override): Simplify mcu and mcpu option handling.
2532 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
2533 support for -mhwmult command line option.
2534 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
2535 -mhwmult command line option.
2536 (msp430_hwmult_enabled): Delete.
2537 (msp43o_output_labelref): Add support for -mhwmult command line
2539 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
2540 (umulsidi3): Likewise.
2541 * config/msp430/msp430.opt (mmcu): Add Report attribute.
2542 (mcpu, mlarge, msmall): Likewise.
2543 (mhwmult): New option.
2544 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
2546 (msp430_is_f5_mcu): Remove prototype.
2547 (msp430_use_f5_series_hwmult): Add prototype.
2548 * config/msp430/msp430-opts.h: New file.
2549 * common/config/msp430: New directory.
2550 * common/config/msp430/msp430-common.c: New file.
2551 * config.gcc (msp430): Remove target_has_targetm_common.
2552 * doc/invoke.texi: Document -mhwmult command line option.
2554 2014-04-23 Nick Clifton <nickc@redhat.com>
2556 * config/i386/cygwin.h (ENDFILE_SPEC): Include
2557 default-manifest.o if it can be found in the search path.
2558 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
2560 2014-04-23 Terry Guo <terry.guo@arm.com>
2562 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
2564 2014-04-23 Richard Biener <rguenther@suse.de>
2567 * tree-inline.c (declare_return_variable): Use mark_addressable.
2569 2014-04-23 Richard Biener <rguenther@suse.de>
2572 * loop-init.c (loop_optimizer_init): Make sure to apply
2573 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
2575 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2578 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
2580 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
2581 if flag_sanitize_undefined_trap_on_error.
2582 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
2583 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
2584 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
2585 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
2586 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
2587 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
2588 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
2589 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
2590 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
2591 * ubsan.c (ubsan_instrument_unreachable): Return
2592 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
2593 (ubsan_expand_null_ifn): Emit __builtin_trap ()
2594 if flag_sanitize_undefined_trap_on_error and
2595 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
2596 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
2597 instrument_bool_enum_load): Emit __builtin_trap () if
2598 flag_sanitize_undefined_trap_on_error and
2599 __builtin_handle_*_abort () if !flag_sanitize_recover.
2600 * doc/invoke.texi (-fsanitize-recover,
2601 -fsanitize-undefined-trap-on-error): Document.
2603 2014-04-22 Christian Bruel <christian.bruel@st.com>
2605 * config/sh/sh.md (mov<mode>): Replace movQIHI.
2606 Force immediates to SImode.
2608 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
2610 * config/nios2/nios2.md (UNSPEC_ROUND): New.
2612 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
2613 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
2614 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
2615 (nios2_fpu_insn): Add entry for round.
2616 (N2FPU_NO_ERRNO_P): Define.
2617 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
2619 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
2621 2014-04-22 Richard Henderson <rth@redhat.com>
2623 * config/aarch64/aarch64 (addti3, subti3): New expanders.
2624 (add<GPI>3_compare0): Remove leading * from name.
2625 (add<GPI>3_carryin): Likewise.
2626 (sub<GPI>3_compare0): Likewise.
2627 (sub<GPI>3_carryin): Likewise.
2628 (<su_optab>mulditi3): New expander.
2629 (multi3): New expander.
2630 (madd<GPI>): Remove leading * from name.
2632 2014-04-22 Martin Jambor <mjambor@suse.cz>
2634 * cgraphclones.c (cgraph_function_versioning): Copy
2635 ipa_transforms_to_apply instead of asserting it is empty.
2637 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
2640 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
2641 on count_exp to get mode.
2643 2014-04-22 Andrew Pinski <apinski@cavium.com>
2645 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
2646 Handle TLS for ILP32.
2647 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
2648 (tlsie_small_<mode>): this and handle PTR.
2649 (tlsie_small_sidi): New pattern.
2650 (tlsle_small): Change to an expand to handle ILP32.
2651 (tlsle_small_<mode>): New pattern.
2652 (tlsdesc_small): Rename to ...
2653 (tlsdesc_small_<mode>): this and handle PTR.
2655 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2657 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
2659 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2661 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2662 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
2663 (aarch64_types_signed_poly_qualifiers): Likewise.
2664 (aarch64_types_unsigned_signed_qualifiers): Likewise.
2665 (aarch64_types_poly_signed_qualifiers): Likewise.
2666 (TYPES_REINTERP_SS): Type macro added.
2667 (TYPES_REINTERP_SU): Likewise.
2668 (TYPES_REINTERP_SP): Likewise.
2669 (TYPES_REINTERP_US): Likewise.
2670 (TYPES_REINTERP_PS): Likewise.
2671 (aarch64_fold_builtin): New expression folding added.
2672 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
2673 Declarations removed.
2674 (REINTERP_SS): Declarations added.
2675 (REINTERP_US): Likewise.
2676 (REINTERP_PS): Likewise.
2677 (REINTERP_SU): Likewise.
2678 (REINTERP_SP): Likewise.
2679 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
2680 (vreinterpretq_p8_f64): Likewise.
2681 (vreinterpret_p16_f64): Likewise.
2682 (vreinterpretq_p16_f64): Likewise.
2683 (vreinterpret_f32_f64): Likewise.
2684 (vreinterpretq_f32_f64): Likewise.
2685 (vreinterpret_f64_f32): Likewise.
2686 (vreinterpret_f64_p8): Likewise.
2687 (vreinterpret_f64_p16): Likewise.
2688 (vreinterpret_f64_s8): Likewise.
2689 (vreinterpret_f64_s16): Likewise.
2690 (vreinterpret_f64_s32): Likewise.
2691 (vreinterpret_f64_s64): Likewise.
2692 (vreinterpret_f64_u8): Likewise.
2693 (vreinterpret_f64_u16): Likewise.
2694 (vreinterpret_f64_u32): Likewise.
2695 (vreinterpret_f64_u64): Likewise.
2696 (vreinterpretq_f64_f32): Likewise.
2697 (vreinterpretq_f64_p8): Likewise.
2698 (vreinterpretq_f64_p16): Likewise.
2699 (vreinterpretq_f64_s8): Likewise.
2700 (vreinterpretq_f64_s16): Likewise.
2701 (vreinterpretq_f64_s32): Likewise.
2702 (vreinterpretq_f64_s64): Likewise.
2703 (vreinterpretq_f64_u8): Likewise.
2704 (vreinterpretq_f64_u16): Likewise.
2705 (vreinterpretq_f64_u32): Likewise.
2706 (vreinterpretq_f64_u64): Likewise.
2707 (vreinterpret_s64_f64): Likewise.
2708 (vreinterpretq_s64_f64): Likewise.
2709 (vreinterpret_u64_f64): Likewise.
2710 (vreinterpretq_u64_f64): Likewise.
2711 (vreinterpret_s8_f64): Likewise.
2712 (vreinterpretq_s8_f64): Likewise.
2713 (vreinterpret_s16_f64): Likewise.
2714 (vreinterpretq_s16_f64): Likewise.
2715 (vreinterpret_s32_f64): Likewise.
2716 (vreinterpretq_s32_f64): Likewise.
2717 (vreinterpret_u8_f64): Likewise.
2718 (vreinterpretq_u8_f64): Likewise.
2719 (vreinterpret_u16_f64): Likewise.
2720 (vreinterpretq_u16_f64): Likewise.
2721 (vreinterpret_u32_f64): Likewise.
2722 (vreinterpretq_u32_f64): Likewise.
2724 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2726 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2727 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
2728 (vreinterpret_p8_s8): Likewise.
2729 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
2730 (vreinterpret_p8_s16): Likewise.
2731 (vreinterpret_p8_s32): Likewise.
2732 (vreinterpret_p8_s64): Likewise.
2733 (vreinterpret_p8_f32): Likewise.
2734 (vreinterpret_p8_u8): Likewise.
2735 (vreinterpret_p8_u16): Likewise.
2736 (vreinterpret_p8_u32): Likewise.
2737 (vreinterpret_p8_u64): Likewise.
2738 (vreinterpret_p8_p16): Likewise.
2739 (vreinterpretq_p8_s8): Likewise.
2740 (vreinterpretq_p8_s16): Likewise.
2741 (vreinterpretq_p8_s32): Likewise.
2742 (vreinterpretq_p8_s64): Likewise.
2743 (vreinterpretq_p8_f32): Likewise.
2744 (vreinterpretq_p8_u8): Likewise.
2745 (vreinterpretq_p8_u16): Likewise.
2746 (vreinterpretq_p8_u32): Likewise.
2747 (vreinterpretq_p8_u64): Likewise.
2748 (vreinterpretq_p8_p16): Likewise.
2749 (vreinterpret_p16_s8): Likewise.
2750 (vreinterpret_p16_s16): Likewise.
2751 (vreinterpret_p16_s32): Likewise.
2752 (vreinterpret_p16_s64): Likewise.
2753 (vreinterpret_p16_f32): Likewise.
2754 (vreinterpret_p16_u8): Likewise.
2755 (vreinterpret_p16_u16): Likewise.
2756 (vreinterpret_p16_u32): Likewise.
2757 (vreinterpret_p16_u64): Likewise.
2758 (vreinterpret_p16_p8): Likewise.
2759 (vreinterpretq_p16_s8): Likewise.
2760 (vreinterpretq_p16_s16): Likewise.
2761 (vreinterpretq_p16_s32): Likewise.
2762 (vreinterpretq_p16_s64): Likewise.
2763 (vreinterpretq_p16_f32): Likewise.
2764 (vreinterpretq_p16_u8): Likewise.
2765 (vreinterpretq_p16_u16): Likewise.
2766 (vreinterpretq_p16_u32): Likewise.
2767 (vreinterpretq_p16_u64): Likewise.
2768 (vreinterpretq_p16_p8): Likewise.
2769 (vreinterpret_f32_s8): Likewise.
2770 (vreinterpret_f32_s16): Likewise.
2771 (vreinterpret_f32_s32): Likewise.
2772 (vreinterpret_f32_s64): Likewise.
2773 (vreinterpret_f32_u8): Likewise.
2774 (vreinterpret_f32_u16): Likewise.
2775 (vreinterpret_f32_u32): Likewise.
2776 (vreinterpret_f32_u64): Likewise.
2777 (vreinterpret_f32_p8): Likewise.
2778 (vreinterpret_f32_p16): Likewise.
2779 (vreinterpretq_f32_s8): Likewise.
2780 (vreinterpretq_f32_s16): Likewise.
2781 (vreinterpretq_f32_s32): Likewise.
2782 (vreinterpretq_f32_s64): Likewise.
2783 (vreinterpretq_f32_u8): Likewise.
2784 (vreinterpretq_f32_u16): Likewise.
2785 (vreinterpretq_f32_u32): Likewise.
2786 (vreinterpretq_f32_u64): Likewise.
2787 (vreinterpretq_f32_p8): Likewise.
2788 (vreinterpretq_f32_p16): Likewise.
2789 (vreinterpret_s64_s8): Likewise.
2790 (vreinterpret_s64_s16): Likewise.
2791 (vreinterpret_s64_s32): Likewise.
2792 (vreinterpret_s64_f32): Likewise.
2793 (vreinterpret_s64_u8): Likewise.
2794 (vreinterpret_s64_u16): Likewise.
2795 (vreinterpret_s64_u32): Likewise.
2796 (vreinterpret_s64_u64): Likewise.
2797 (vreinterpret_s64_p8): Likewise.
2798 (vreinterpret_s64_p16): Likewise.
2799 (vreinterpretq_s64_s8): Likewise.
2800 (vreinterpretq_s64_s16): Likewise.
2801 (vreinterpretq_s64_s32): Likewise.
2802 (vreinterpretq_s64_f32): Likewise.
2803 (vreinterpretq_s64_u8): Likewise.
2804 (vreinterpretq_s64_u16): Likewise.
2805 (vreinterpretq_s64_u32): Likewise.
2806 (vreinterpretq_s64_u64): Likewise.
2807 (vreinterpretq_s64_p8): Likewise.
2808 (vreinterpretq_s64_p16): Likewise.
2809 (vreinterpret_u64_s8): Likewise.
2810 (vreinterpret_u64_s16): Likewise.
2811 (vreinterpret_u64_s32): Likewise.
2812 (vreinterpret_u64_s64): Likewise.
2813 (vreinterpret_u64_f32): Likewise.
2814 (vreinterpret_u64_u8): Likewise.
2815 (vreinterpret_u64_u16): Likewise.
2816 (vreinterpret_u64_u32): Likewise.
2817 (vreinterpret_u64_p8): Likewise.
2818 (vreinterpret_u64_p16): Likewise.
2819 (vreinterpretq_u64_s8): Likewise.
2820 (vreinterpretq_u64_s16): Likewise.
2821 (vreinterpretq_u64_s32): Likewise.
2822 (vreinterpretq_u64_s64): Likewise.
2823 (vreinterpretq_u64_f32): Likewise.
2824 (vreinterpretq_u64_u8): Likewise.
2825 (vreinterpretq_u64_u16): Likewise.
2826 (vreinterpretq_u64_u32): Likewise.
2827 (vreinterpretq_u64_p8): Likewise.
2828 (vreinterpretq_u64_p16): Likewise.
2829 (vreinterpret_s8_s16): Likewise.
2830 (vreinterpret_s8_s32): Likewise.
2831 (vreinterpret_s8_s64): Likewise.
2832 (vreinterpret_s8_f32): Likewise.
2833 (vreinterpret_s8_u8): Likewise.
2834 (vreinterpret_s8_u16): Likewise.
2835 (vreinterpret_s8_u32): Likewise.
2836 (vreinterpret_s8_u64): Likewise.
2837 (vreinterpret_s8_p8): Likewise.
2838 (vreinterpret_s8_p16): Likewise.
2839 (vreinterpretq_s8_s16): Likewise.
2840 (vreinterpretq_s8_s32): Likewise.
2841 (vreinterpretq_s8_s64): Likewise.
2842 (vreinterpretq_s8_f32): Likewise.
2843 (vreinterpretq_s8_u8): Likewise.
2844 (vreinterpretq_s8_u16): Likewise.
2845 (vreinterpretq_s8_u32): Likewise.
2846 (vreinterpretq_s8_u64): Likewise.
2847 (vreinterpretq_s8_p8): Likewise.
2848 (vreinterpretq_s8_p16): Likewise.
2849 (vreinterpret_s16_s8): Likewise.
2850 (vreinterpret_s16_s32): Likewise.
2851 (vreinterpret_s16_s64): Likewise.
2852 (vreinterpret_s16_f32): Likewise.
2853 (vreinterpret_s16_u8): Likewise.
2854 (vreinterpret_s16_u16): Likewise.
2855 (vreinterpret_s16_u32): Likewise.
2856 (vreinterpret_s16_u64): Likewise.
2857 (vreinterpret_s16_p8): Likewise.
2858 (vreinterpret_s16_p16): Likewise.
2859 (vreinterpretq_s16_s8): Likewise.
2860 (vreinterpretq_s16_s32): Likewise.
2861 (vreinterpretq_s16_s64): Likewise.
2862 (vreinterpretq_s16_f32): Likewise.
2863 (vreinterpretq_s16_u8): Likewise.
2864 (vreinterpretq_s16_u16): Likewise.
2865 (vreinterpretq_s16_u32): Likewise.
2866 (vreinterpretq_s16_u64): Likewise.
2867 (vreinterpretq_s16_p8): Likewise.
2868 (vreinterpretq_s16_p16): Likewise.
2869 (vreinterpret_s32_s8): Likewise.
2870 (vreinterpret_s32_s16): Likewise.
2871 (vreinterpret_s32_s64): Likewise.
2872 (vreinterpret_s32_f32): Likewise.
2873 (vreinterpret_s32_u8): Likewise.
2874 (vreinterpret_s32_u16): Likewise.
2875 (vreinterpret_s32_u32): Likewise.
2876 (vreinterpret_s32_u64): Likewise.
2877 (vreinterpret_s32_p8): Likewise.
2878 (vreinterpret_s32_p16): Likewise.
2879 (vreinterpretq_s32_s8): Likewise.
2880 (vreinterpretq_s32_s16): Likewise.
2881 (vreinterpretq_s32_s64): Likewise.
2882 (vreinterpretq_s32_f32): Likewise.
2883 (vreinterpretq_s32_u8): Likewise.
2884 (vreinterpretq_s32_u16): Likewise.
2885 (vreinterpretq_s32_u32): Likewise.
2886 (vreinterpretq_s32_u64): Likewise.
2887 (vreinterpretq_s32_p8): Likewise.
2888 (vreinterpretq_s32_p16): Likewise.
2889 (vreinterpret_u8_s8): Likewise.
2890 (vreinterpret_u8_s16): Likewise.
2891 (vreinterpret_u8_s32): Likewise.
2892 (vreinterpret_u8_s64): Likewise.
2893 (vreinterpret_u8_f32): Likewise.
2894 (vreinterpret_u8_u16): Likewise.
2895 (vreinterpret_u8_u32): Likewise.
2896 (vreinterpret_u8_u64): Likewise.
2897 (vreinterpret_u8_p8): Likewise.
2898 (vreinterpret_u8_p16): Likewise.
2899 (vreinterpretq_u8_s8): Likewise.
2900 (vreinterpretq_u8_s16): Likewise.
2901 (vreinterpretq_u8_s32): Likewise.
2902 (vreinterpretq_u8_s64): Likewise.
2903 (vreinterpretq_u8_f32): Likewise.
2904 (vreinterpretq_u8_u16): Likewise.
2905 (vreinterpretq_u8_u32): Likewise.
2906 (vreinterpretq_u8_u64): Likewise.
2907 (vreinterpretq_u8_p8): Likewise.
2908 (vreinterpretq_u8_p16): Likewise.
2909 (vreinterpret_u16_s8): Likewise.
2910 (vreinterpret_u16_s16): Likewise.
2911 (vreinterpret_u16_s32): Likewise.
2912 (vreinterpret_u16_s64): Likewise.
2913 (vreinterpret_u16_f32): Likewise.
2914 (vreinterpret_u16_u8): Likewise.
2915 (vreinterpret_u16_u32): Likewise.
2916 (vreinterpret_u16_u64): Likewise.
2917 (vreinterpret_u16_p8): Likewise.
2918 (vreinterpret_u16_p16): Likewise.
2919 (vreinterpretq_u16_s8): Likewise.
2920 (vreinterpretq_u16_s16): Likewise.
2921 (vreinterpretq_u16_s32): Likewise.
2922 (vreinterpretq_u16_s64): Likewise.
2923 (vreinterpretq_u16_f32): Likewise.
2924 (vreinterpretq_u16_u8): Likewise.
2925 (vreinterpretq_u16_u32): Likewise.
2926 (vreinterpretq_u16_u64): Likewise.
2927 (vreinterpretq_u16_p8): Likewise.
2928 (vreinterpretq_u16_p16): Likewise.
2929 (vreinterpret_u32_s8): Likewise.
2930 (vreinterpret_u32_s16): Likewise.
2931 (vreinterpret_u32_s32): Likewise.
2932 (vreinterpret_u32_s64): Likewise.
2933 (vreinterpret_u32_f32): Likewise.
2934 (vreinterpret_u32_u8): Likewise.
2935 (vreinterpret_u32_u16): Likewise.
2936 (vreinterpret_u32_u64): Likewise.
2937 (vreinterpret_u32_p8): Likewise.
2938 (vreinterpret_u32_p16): Likewise.
2939 (vreinterpretq_u32_s8): Likewise.
2940 (vreinterpretq_u32_s16): Likewise.
2941 (vreinterpretq_u32_s32): Likewise.
2942 (vreinterpretq_u32_s64): Likewise.
2943 (vreinterpretq_u32_f32): Likewise.
2944 (vreinterpretq_u32_u8): Likewise.
2945 (vreinterpretq_u32_u16): Likewise.
2946 (vreinterpretq_u32_u64): Likewise.
2947 (vreinterpretq_u32_p8): Likewise.
2948 (vreinterpretq_u32_p16): Likewise.
2950 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2952 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
2954 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
2957 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
2958 (vqnegd_s64): Likewise.
2959 (vqabs_s64): Likewise.
2960 (vqabsd_s64): Likewise.
2962 2014-04-22 Richard Henderson <rth@redhat.com>
2964 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
2965 computation to the top of the loop.
2967 2014-04-22 Renlin <renlin.li@arm.com>
2968 Jiong Wang <jiong.wang@arm.com>
2970 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
2971 * config/aarch64/aarch64.c (aarch64_layout_frame)
2972 (aarch64_initial_elimination_offset): Likewise.
2974 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
2976 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
2979 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
2981 * machmode.h (bitwise_mode_for_mode): Declare.
2982 * stor-layout.h (bitwise_type_for_mode): Likewise.
2983 * stor-layout.c (bitwise_mode_for_mode): New function.
2984 (bitwise_type_for_mode): Likewise.
2985 * builtins.c (fold_builtin_memory_op): Use it instead of
2986 int_mode_for_mode and build_nonstandard_integer_type.
2988 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2990 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
2991 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
2992 (*-*-solaris2*): Simplify.
2993 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
2994 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
2995 *-*-solaris2.9* handling.
2997 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
2999 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
3000 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
3002 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
3003 * configure: Regenerate.
3005 * config/i386/sol2-9.h: Remove.
3007 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
3008 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
3009 Remove Solaris 9 references.
3011 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
3013 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
3014 (floatuns<GPI:mode><GPF:mode>2): Remove.
3015 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
3016 and floatuns conversions.
3017 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
3018 and floatuns conversions.
3019 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
3020 (w1,w2): New mode attributes for inequal width conversions.
3022 2014-04-22 Renlin Li <Renlin.Li@arm.com>
3024 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
3025 the output asm format.
3027 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
3029 * config/aarch64/aarch64-simd.md
3030 (aarch64_cm<optab>di): Always split.
3031 (*aarch64_cm<optab>di): New.
3032 (aarch64_cmtstdi): Always split.
3033 (*aarch64_cmtstdi): New.
3035 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3037 PR tree-optimization/60823
3038 * omp-low.c (ipa_simd_modify_function_body): Go through
3039 all SSA_NAMEs and for those refering to vector arguments
3040 which are going to be replaced adjust SSA_NAME_VAR and,
3041 if it is a default definition, change it into a non-default
3042 definition assigned at the beginning of function from new_decl.
3043 (ipa_simd_modify_stmt_ops): Rewritten.
3044 * tree-dfa.c (set_ssa_default_def): When removing default def,
3045 check for NULL loc instead of NULL *loc.
3047 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3049 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
3050 restrictions on core registers for DImode values in Thumb2.
3052 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3054 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
3055 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
3057 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3059 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
3060 (*iordi_notzesidi_di): Likewise.
3061 (*iordi_notsesidi_di): Likewise.
3063 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3065 * config/arm/arm-protos.h (tune_params): New struct members.
3066 * config/arm/arm.c: Initialise tune_params per processor.
3067 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
3068 for speed, based on new tune_params.
3070 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3072 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
3074 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
3076 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
3078 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
3079 * config/aarch64/arm_neon.h (vrnd_f64): Added.
3080 (vrnda_f64): Likewise.
3081 (vrndi_f64): Likewise.
3082 (vrndm_f64): Likewise.
3083 (vrndn_f64): Likewise.
3084 (vrndp_f64): Likewise.
3085 (vrndx_f64): Likewise.
3087 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3089 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
3090 GET_MODE_SIZE argument is enum machine_mode.
3092 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3095 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
3096 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
3098 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
3101 * asan.c (asan_emit_stack_protection): Force the base to align to
3102 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
3103 appropriate bits if STRICT_ALIGNMENT.
3104 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
3106 (expand_used_vars): Leave a space in the stack frame for alignment
3107 if STRICT_ALIGNMENT.
3109 2014-04-21 David Malcolm <dmalcolm@redhat.com>
3111 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
3113 (gimple_store_p): Likewise.
3114 (gimple_assign_load_p): Likewise.
3115 (gimple_assign_cast_p): Likewise.
3116 (gimple_clobber_p): Likewise.
3118 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
3119 rather than a gimple.
3120 (gimple_assign_cast_p): Likewise.
3122 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
3125 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
3126 If mode is DDmode and TARGET_E500_DOUBLE allow move.
3128 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
3129 more debug information for E500 if -mdebug=reg.
3131 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
3134 * config/i386/i386.c (ix86_expand_builtin)
3135 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
3136 register for target RTX.
3137 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
3139 2014-04-18 Cong Hou <congh@google.com>
3141 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
3142 the widen-mult pattern by handling two operands with different sizes,
3143 and operands whose size is smaller than half of the result type.
3145 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3147 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
3148 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
3149 (do_estimate_edge_time): Compute it.
3150 * ipa-inline.c (want_inline_small_function_p): Bypass
3151 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
3153 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3155 * ipa-inline.c (spec_rem): New static variable.
3156 (dump_overall_stats): New function.
3157 (dump_inline_stats): New function.
3159 2014-04-18 Richard Henderson <rth@redhat.com>
3161 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
3162 to GET_MODE_SIZE, not a reg_class_t.
3164 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3166 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
3167 (vsx_xxmrglw_<mode>): Likewise.
3169 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3172 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
3173 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
3174 (rs6000_init_hard_regno_mode_ok): Likewise.
3176 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
3178 * ipa-inline.c (inline_small_functions): Account only non-cold
3180 * doc/invoke.texi (inline-unit-growth): Update documentation.
3182 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
3184 * config/rs6000/rs6000.md (addti3, subti3): New.
3186 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
3189 * config/i386/i386.c (ix86_expand_clear): Remove outdated
3190 comment. Check optimize_insn_for_size_p instead of
3191 optimize_insn_for_speed_p.
3193 2014-04-17 Martin Jambor <mjambor@suse.cz>
3195 * gimple-iterator.c (gsi_start_edge): New function.
3196 * gimple-iterator.h (gsi_start_edge): Declare.
3197 * tree-sra.c (single_non_eh_succ): New function.
3198 (disqualify_ops_if_throwing_stmt): Renamed to
3199 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
3200 having one non-EH successor BB.
3201 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
3202 generate loads into replacements.
3203 (sra_modify_assign): Likewise and and also use the simple path for
3205 (sra_modify_function_body): Commit statements on edges.
3207 2014-04-17 Richard Biener <rguenther@suse.de>
3210 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
3211 comparison results and add clarifying comment.
3213 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3215 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
3216 (blank_mode): Initialize it.
3217 (emit_mode_size_inline, emit_mode_nunits_inline,
3218 emit_mode_inner_inline): New functions.
3219 (emit_insn_modes_h): Call them and surround their output with
3220 #if GCC_VERSION >= 4001 ... #endif.
3221 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
3222 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
3223 mode_* arrays if the argument is __builtin_constant_p.
3224 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
3225 is enum machine_mode.
3227 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3229 * passes.c (opt_pass::execute): Adjust.
3230 (pass_manager::execute_pass_mode_switching): Likewise.
3231 (early_local_passes::execute): Likewise.
3232 (execute_one_pass): Pass cfun to the pass's execute method.
3233 * tree-pass.h (opt_pass::execute): Add function * argument.
3234 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3235 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3236 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3237 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3238 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3239 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
3240 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
3241 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3242 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3243 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3244 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3245 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3246 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3247 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3248 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3249 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3250 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3251 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3252 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3253 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3254 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3255 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3256 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3257 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3258 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3259 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3260 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3261 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3262 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3265 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3267 * passes.c (opt_pass::gate): Take function * argument.
3268 (gate_all_early_local_passes): Merge into
3269 (early_local_passes::gate): this.
3270 (gate_all_early_optimizations): Merge into
3271 (all_early_optimizations::gate): this.
3272 (gate_all_optimizations): Mege into
3273 (all_optimizations::gate): this.
3274 (gate_all_optimizations_g): Merge into
3275 (all_optimizations_g::gate): this.
3276 (gate_rest_of_compilation): Mege into
3277 (rest_of_compilation::gate): this.
3278 (gate_postreload): Merge into
3279 (postreload::gate): this.
3280 (dump_one_pass): Pass cfun to the pass's gate method.
3281 (execute_ipa_summary_passes): Likewise.
3282 (execute_one_pass): Likewise.
3283 (ipa_write_summaries_2): Likewise.
3284 (ipa_write_optimization_summaries_1): Likewise.
3285 (ipa_read_summaries_1): Likewise.
3286 (ipa_read_optimization_summaries_1): Likewise.
3287 (execute_ipa_stmt_fixups): Likewise.
3288 * tree-pass.h (opt_pass::gate): Add function * argument.
3289 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
3290 combine-stack-adj.c, combine.c, compare-elim.c,
3291 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3292 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
3293 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
3294 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
3295 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3296 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3297 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3298 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
3299 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
3300 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
3301 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3302 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3303 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
3304 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3305 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3306 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3307 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3308 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3309 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3310 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3311 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3312 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3313 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3314 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
3315 var-tracking.c, vtable-verify.c, web.c: Adjust.
3317 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3319 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
3320 * configure: Regenerate.
3322 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3324 * passes.c (dump_one_pass): don't check pass->has_gate.
3325 (execute_ipa_summary_passes): Likewise.
3326 (execute_one_pass): Likewise.
3327 (ipa_write_summaries_2): Likewise.
3328 (ipa_write_optimization_summaries_1): Likewise.
3329 (ipa_read_optimization_summaries_1): Likewise.
3330 (execute_ipa_stmt_fixups): Likewise.
3331 * tree-pass.h (pass_data::has_gate): Remove.
3332 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3333 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3334 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3335 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3336 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3337 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
3338 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
3339 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
3340 gimple-low.c, gimple-ssa-isolate-paths.c,
3341 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
3342 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3343 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3344 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
3345 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
3346 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
3347 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
3348 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3349 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
3350 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3351 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3352 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3353 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3354 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3355 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3356 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3357 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3358 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3359 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3360 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3361 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3362 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3365 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3367 * pass_manager.h (pass_manager::register_dump_files_1): Remove
3369 * passes.c (pass_manager::register_dump_files_1): Merge into
3370 (pass_manager::register_dump_files): this, and remove its handling of
3371 properties since the pass always has the properties anyway.
3372 (pass_manager::pass_manager): Adjust.
3374 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3376 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
3377 * passes.c (pass_manager::register_dump_files_1): Remove dead code
3378 dealing with properties.
3379 (pass_manager::register_dump_files): Adjust.
3381 2014-03-20 Mark Wielaard <mjw@redhat.com>
3383 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
3384 then represent the bound as normal constant value.
3386 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3389 Forward port from 4.8 branch
3390 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
3392 * config/i386/bmiintrin.h (_blsi_u32): New.
3396 (_blsmsk_u32): Ditto.
3397 (_blsmsk_u64): Ditto.
3398 (_tzcnt_u32): Ditto.
3399 (_tzcnt_u64): Ditto.
3401 2014-04-17 Kito Cheng <kito@0xlab.org>
3403 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
3405 2014-04-17 Richard Biener <rguenther@suse.de>
3408 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
3409 boolean results for comparisons.
3411 2014-04-17 Richard Biener <rguenther@suse.de>
3413 PR tree-optimization/60836
3414 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
3415 initial PHI args to be gimple values.
3417 2014-04-17 Richard Biener <rguenther@suse.de>
3419 PR tree-optimization/60841
3420 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
3421 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
3422 of stmts to SLP build.
3423 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
3424 (vect_analyze_slp): Likewise.
3425 (vect_analyze_slp_instance): Likewise.
3426 (vect_build_slp_tree): Limit overall SLP tree growth.
3427 * tree-vectorizer.h (vect_analyze_data_refs,
3428 vect_analyze_slp): Adjust prototypes.
3430 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3432 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
3435 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3437 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
3438 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
3439 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
3440 for TARGET_SLOW_PSHUFB
3442 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3444 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
3445 * config/i386/i386.c (intel_cost): Ditto.
3447 2014-04-17 Joey Ye <joey.ye@arm.com>
3449 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
3451 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3453 * opts.c (common_handle_option): Disable -fipa-reference coorectly
3456 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3458 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
3459 (type_all_derivations_known_p): New predicate.
3460 (type_all_ctors_visible_p): New predicate.
3461 (type_possibly_instantiated_p): New predicate.
3462 (get_odr_type): Compute all_derivations_known.
3463 (dump_odr_type): Dump the flag.
3464 (maybe_record_type): Cleanup.
3465 (record_target_from_binfo): Add bases_to_consider array;
3466 record bases for types w/o instances and skip CXX destructor.
3467 (possible_polymorphic_call_targets_1): Add bases_to_consider
3468 and consider_construction parameters; check if type may have instance.
3469 (get_polymorphic_call_info): Set maybe_in_construction to true
3470 when we know nothing.
3471 (record_targets_from_bases): Skip CXX destructors; they are
3472 never called for types in construction.
3473 (possible_polymorphic_call_targets): Do not record target when
3474 type may not have instance.
3476 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3479 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
3480 external aliases alive, too.
3482 2014-04-16 Andrew Pinski <apinski@cavium.com>
3484 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
3487 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3489 * final.c (compute_alignments): Do not apply loop alignment to a block
3490 falling through to the exit.
3492 2014-04-16 Catherine Moore <clm@codesourcery.com>
3494 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
3495 Adjust constraints for microMIPS store patterns.
3497 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3499 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
3501 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3503 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
3504 (append_use): Run at -O0.
3505 (append_vdef): Likewise.
3506 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
3507 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
3509 2014-04-16 Jakub Jelinek <jakub@redhat.com>
3511 PR tree-optimization/60844
3512 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
3513 (propagate_op_to_single_use, remove_visited_stmt_chain,
3514 linearize_expr, repropagate_negates, reassociate_bb): Use it
3515 instead of gsi_remove.
3517 2014-04-16 Martin Jambor <mjambor@suse.cz>
3519 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
3520 ipa_transforms_to_apply.
3521 (cgraph_function_versioning): Assert that old_node has empty
3522 ipa_transforms_to_apply.
3523 * trans-mem.c (ipa_tm_create_version): Likewise.
3524 * tree-inline.c (tree_function_versioning): Do not duplicate
3525 ipa_transforms_to_apply.
3527 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3530 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
3532 Pass necessary as flags on 64-bit Solaris/x86.
3533 Use lowercase relocs for x86_64-*-*.
3534 * configure: Regenerate.
3536 2014-04-15 Jan Hubicka <jh@suse.cz>
3538 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
3539 (maybe_record_node, likely_target_p): Use it.
3541 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3544 Revert following patch
3546 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3549 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3550 software floating point or no floating point registers, do not
3551 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3552 in GPRs that occurs after we tested for GPRs that would never be
3555 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3556 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3557 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3558 specifically allow DDmode, since that does not use the SPE SIMD
3561 2014-03-21 Mark Wielaard <mjw@redhat.com>
3563 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
3564 as unsigned or int depending on type and value used.
3566 2014-04-15 Richard Biener <rguenther@suse.de>
3568 PR rtl-optimization/56965
3569 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
3570 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
3572 * alias.c (true_dependence_1): Do not call
3573 nonoverlapping_component_refs_p.
3574 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
3575 nonoverlapping_component_refs_p.
3576 (indirect_refs_may_alias_p): Likewise.
3578 2014-04-15 Teresa Johnson <tejohnson@google.com>
3580 * cfg.c (dump_bb_info): Fix flags check.
3581 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
3583 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3585 PR rtl-optimization/60663
3586 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
3589 2014-04-15 Richard Biener <rguenther@suse.de>
3591 * lto-streamer.h (LTO_major_version): Bump to 4.
3593 2014-04-15 Richard Biener <rguenther@suse.de>
3595 * common.opt (lto_partition_model): New enum.
3596 (flto-partition=): Merge separate options with a single with argument,
3597 add -flto-partition=one support.
3598 * flag-types.h (enum lto_partition_model): Declare.
3599 * opts.c (finish_options): Remove duplicate -flto-partition=
3601 * lto-wrapper.c (run_gcc): Adjust.
3603 2014-04-15 Richard Biener <rguenther@suse.de>
3605 * alias.c (ncr_compar): New function.
3606 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
3608 2014-04-15 Richard Biener <rguenther@suse.de>
3610 * alias.c (record_component_aliases): Do not walk BINFOs.
3612 2014-04-15 Richard Biener <rguenther@suse.de>
3614 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
3615 Add struct function argument and adjust.
3616 (find_func_aliases_for_call): Likewise.
3617 (find_func_aliases): Likewise.
3618 (find_func_clobbers): Likewise.
3619 (intra_create_variable_infos): Likewise.
3620 (compute_points_to_sets): Likewise.
3621 (ipa_pta_execute): Adjust. Do not push/pop cfun.
3623 2014-04-15 Richard Biener <rguenther@suse.de>
3625 * tree.c (iterative_hash_expr): Use enum tree_code_class
3626 to store TREE_CODE_CLASS.
3627 (tree_block): Likewise.
3628 (tree_set_block): Likewise.
3629 * tree.h (fold_build_pointer_plus_loc): Use
3630 convert_to_ptrofftype_loc.
3632 2014-04-15 Jakub Jelinek <jakub@redhat.com>
3635 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
3638 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
3640 * cfgloop.h (struct loop): Move force_vectorize down.
3641 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
3642 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
3643 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
3644 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
3645 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
3646 * tree-core.h (enum annot_expr_kind): Add new kind values.
3647 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
3648 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
3650 * tree.def (ANNOTATE_EXPR): Tweak comment.
3652 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3654 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
3657 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
3659 * tree.h (TYPE_IDENTIFIER): Declare.
3660 * tree.c (subrange_type_for_debug_p): Use it.
3661 * godump.c (go_format_type): Likewise.
3662 * dwarf2out.c (is_cxx_auto, modified_type_die,
3663 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
3664 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
3666 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3669 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
3671 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
3673 * config/i386/i386.c (examine_argument): Return bool. Return true if
3674 parameter should be passed in memory.
3675 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
3676 (construct_container): Update calls to examine_argument.
3677 (function_arg_advance_64): Ditto.
3678 (return_in_memory_32): Merge with ix86_return_in_memory.
3679 (return_in_memory_64): Ditto.
3680 (return_in_memory_ms_64): Ditto.
3682 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3684 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
3685 * coverage.c (coverage_compute_profile_id): Handle externally visible
3688 2014-04-14 Martin Jambor <mjambor@suse.cz>
3690 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
3691 DECL_DISREGARD_INLINE_LIMITS functions.
3693 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3696 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
3698 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3701 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
3702 optimize_insn_for_speed_p instead of
3703 optimize_function_for_speed_p.
3705 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
3707 * doc/invoke.texi (free): Document AArch64.
3709 2014-04-14 Richard Biener <rguenther@suse.de>
3711 PR tree-optimization/60042
3712 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
3713 (insert_into_preds_of_block): Do not prevent PHI insertion
3714 for REFERENCE exprs here ...
3715 (eliminate_dom_walker::before_dom_children): ... but prevent
3716 their use here under similar conditions when applied to the
3717 IL after PRE optimizations.
3719 2014-04-14 Richard Biener <rguenther@suse.de>
3721 * passes.def: Move early points-to after early SRA.
3723 2014-04-14 Richard Biener <rguenther@suse.de>
3725 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
3726 check for which sign-changes we allow when forwarding
3727 a converted value into a switch.
3729 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3731 * stor-layout.c (place_field): Finalize non-constant offset for the
3734 2014-04-14 Richard Biener <rguenther@suse.de>
3736 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
3738 (expand_switch_using_bit_tests_p): Likewise.
3739 (process_switch): Compute and pass on speed_p based on the
3741 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
3742 optimize_bb_for_speed_p.
3744 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3746 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
3747 * function.h (struct function): Rename has_force_vect_loops into
3748 has_force_vectorize_loops.
3749 * lto-streamer-in.c (input_cfg): Adjust for renaming.
3750 (input_struct_function_base): Likewise.
3751 * lto-streamer-out.c (output_cfg): Likewise.
3752 (output_struct_function_base): Likewise.
3753 * omp-low.c (expand_omp_simd): Likewise.
3754 * tree-cfg.c (move_sese_region_to_fn): Likewise.
3755 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
3756 (version_loop_for_if_conversion): Likewise.
3757 (tree_if_conversion): Likewise.
3758 (main_tree_if_conversion): Likewise.
3759 (gate_tree_if_conversion): Likewise.
3760 * tree-inline.c (copy_loops): Likewise.
3761 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
3762 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
3763 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
3764 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
3765 * tree-vectorizer.c (vectorize_loops): Likewise.
3766 * tree-vectorizer.h (unlimited_cost_model): Likewise.
3768 2014-04-14 Richard Biener <rguenther@suse.de>
3771 * lto-streamer-out.c (wrap_refs): New function.
3772 (lto_output): Wrap symbol references in global initializes in
3773 type-preserving MEM_REFs.
3775 2014-04-14 Christian Bruel <christian.bruel@st.com>
3777 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
3779 2014-04-14 Christian Bruel <christian.bruel@st.com>
3781 * config/sh/sh.md (setmemqi): New expand pattern.
3782 * config/sh/sh.h (CLEAR_RATIO): Define.
3783 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
3784 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
3786 2014-04-14 Richard Biener <rguenther@suse.de>
3789 * fold-const.c (negate_expr_p): Don't negate directional rounding
3791 (fold_negate_expr): Likewise.
3793 2014-04-14 Richard Biener <rguenther@suse.de>
3795 PR tree-optimization/59817
3796 PR tree-optimization/60453
3797 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
3798 recursion to catch all CHRECs in the scalar evolution and restrict
3799 the predicate for the remains appropriately.
3801 2014-04-12 Catherine Moore <clm@codesourcery.com>
3803 * config/mips/constraints.md: Add new register constraint "kb".
3804 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
3805 (*movhi_internal): Likewise.
3806 (*movqi_internal): Likewise.
3807 * config/mips/mips.h (M16_STORE_REGS): New register class.
3808 (REG_CLASS_NAMES): Add M16_STORE_REGS.
3809 (REG_CLASS_CONTENTS): Likewise.
3810 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
3812 2014-04-11 Tobias Burnus <burnus@net-b.de>
3815 * doc/invoke.texi (-Wformat-signedness): Document it.
3816 (Wformat=2): Mention that this enables -Wformat-signedness.
3818 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3820 * common/config/epiphany/epiphany-common.c
3821 (epiphany_option_optimization_table): Enable section anchors by
3822 default at -O1 or higher.
3823 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
3824 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
3825 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
3826 carries no extra cost.
3827 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
3828 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
3829 * config/epiphany/predicates.md (memclob_operand): New predicate.
3830 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
3831 Use memclob_operand predicate and X constraint for operand 3.
3833 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3835 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
3836 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
3839 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3841 PR rtl-optimization/60651
3842 * mode-switching.c (optimize_mode_switching): Make sure to emit
3843 sets of a lower numbered entity before sets of a higher numbered
3844 entity to a mode of the same or lower priority.
3845 When creating a seginfo for a basic block that starts with a code
3846 label, move the insertion point past the code label.
3847 (new_seginfo): Document and enforce requirement that
3848 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
3849 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
3850 * doc/tm.texi: Regenerate.
3852 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
3855 * config/arc/arc.c (arc_save_restore): Fix assert typo.
3857 2013-04-11 Jakub Jelinek <jakub@redhat.com>
3859 * BASE-VER: Set to 4.10.0.
3861 2014-04-11 Tobias Burnus <burnus@net-b.de>
3864 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
3865 * doc/gcc.texi (Service): Update description in the @menu
3866 * doc/invoke.texi (Option Summary): Remove misplaced and
3869 2014-04-11 Steve Ellcey <sellcey@mips.com>
3870 Jakub Jelinek <jakub@redhat.com>
3873 * expr.c (convert_move): Use emit_store_flag_force instead of
3874 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
3877 2014-04-11 Richard Biener <rguenther@suse.de>
3880 * varasm.c (assemble_alias): Avoid endless error reporting
3881 recursion by setting TREE_ASM_WRITTEN.
3883 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3885 * config/s390/s390.md: Add a splitter for NOT rtx.
3887 2014-04-11 Jakub Jelinek <jakub@redhat.com>
3889 PR rtl-optimization/60663
3890 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
3892 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
3893 Jakub Jelinek <jakub@redhat.com>
3896 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
3897 flag from decl_node to node.
3899 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3902 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
3903 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
3904 ameliorating the cases where it can be.
3906 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
3909 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
3911 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3912 (loadsync_<mode>): Change mode.
3913 (load_quadpti, store_quadpti): New.
3914 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3915 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3916 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
3918 2014-04-09 Cong Hou <congh@google.com>
3921 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
3924 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3926 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
3927 instead of vnor to exploit possible fusion opportunity in the
3929 (altivec_expand_vec_perm_const_le): Likewise.
3931 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
3933 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3934 (loadsync_<mode>): Change mode.
3935 (load_quadpti, store_quadpti): New.
3936 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3937 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3939 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
3942 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
3943 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
3944 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
3946 2014-04-08 Richard Biener <rguenther@suse.de>
3949 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
3950 a 64bit widest int print double-int similar to on HWI64 hosts.
3952 2014-04-08 Richard Biener <rguenther@suse.de>
3954 PR tree-optimization/60785
3955 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
3956 default defs properly.
3958 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
3960 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
3961 (Weffc++): Likewise.
3963 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
3965 * ipa-devirt.c (maybe_record_node): When node is not recorded,
3966 set completep to false rather than true.
3968 2014-04-07 Douglas B Rupp <rupp@adacore.com>
3971 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
3972 ARM_TARGET2_DWARF_FORMAT.
3974 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
3977 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
3978 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
3981 2014-04-07 Richard Biener <rguenther@suse.de>
3983 PR tree-optimization/60766
3984 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
3985 (may_eliminate_iv): Convert cand_value_at result to desired type.
3987 2014-04-07 Jason Merrill <jason@redhat.com>
3990 * common.opt (-fno-gnu-unique): Add.
3991 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
3993 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3995 * haifa-sched.c: Fix outdated function reference and minor
3996 grammar errors in introductory comment.
3998 2014-04-07 Richard Biener <rguenther@suse.de>
4001 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
4003 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
4005 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
4008 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
4009 size accounting for thunks.
4010 (pa_asm_output_mi_thunk): Use final_start_function() and
4011 final_end_function() to output function start and end directives.
4013 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4015 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
4016 device specific ISA/ feature information. Remove short_sp and
4017 errata_skip ds. Add avr_device_specific_features enum to have device
4019 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
4020 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
4021 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
4022 updated device specific info.
4023 * config/avr/avr-mcus.def: Merge device specific details to
4024 dev_attribute field.
4025 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
4027 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
4028 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
4029 assembler if RMW isa supported by current device.
4030 * config/avr/genmultilib.awk: Update as device info structure changed.
4031 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
4033 2014-04-04 Cong Hou <congh@google.com>
4035 PR tree-optimization/60656
4036 * tree-vect-stmts.c (supportable_widening_operation):
4037 Fix a bug that elements in a vector with vect_used_by_reduction
4038 property are incorrectly reordered when the operation on it is not
4039 consistant with the one in reduction operation.
4041 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
4043 PR rtl-optimization/60155
4044 * gcse.c (record_set_data): New function.
4045 (single_set_gcse): New function.
4046 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
4047 (hoist_code): Likewise.
4048 (get_pressure_class_and_nregs): Likewise.
4050 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
4052 * explow.c (probe_stack_range): Emit a final optimization blockage.
4054 2014-04-04 Anthony Green <green@moxielogic.com>
4056 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
4059 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
4062 * lto-cgraph.c (input_overwrite_node): Check that partitioning
4063 flags are set only during streaming.
4064 * ipa.c (process_references, walk_polymorphic_call_targets,
4065 symtab_remove_unreachable_nodes): Drop bodies of always inline
4066 after early inlining.
4067 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
4069 2014-04-04 Jakub Jelinek <jakub@redhat.com>
4070 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4073 * dwarf2out.c (const_ok_for_output_1): Reject expressions
4076 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4079 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
4081 (cortex_a53_fdivd): Likewise.
4083 2014-04-04 Martin Jambor <mjambor@suse.cz>
4086 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
4088 * cgraph.c (clone_of_p): Also return true if thunks match.
4089 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
4090 cgraph_function_or_thunk_node and an obsolete comment.
4091 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
4093 (build_function_decl_skip_args): Likewise.
4094 (set_new_clone_decl_and_node_flags): New function.
4095 (duplicate_thunk_for_node): Likewise.
4096 (redirect_edge_duplicating_thunks): Likewise.
4097 (cgraph_clone_node): New parameter args_to_skip, pass it to
4098 redirect_edge_duplicating_thunks which is called instead of
4099 cgraph_redirect_edge_callee.
4100 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
4101 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
4103 2014-04-04 Jeff Law <law@redhat.com>
4106 * config/arm/predicates.md (const_int_I_operand): New predicate.
4107 (const_int_M_operand): Similarly.
4108 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
4110 (insv_t2, extv_reg, extzv_t2): Likewise.
4111 (load_multiple_with_writeback): Similarly for const_int_I_operand.
4112 (pop_multiple_with_writeback_and_return): Likewise.
4113 (vfp_pop_multiple_with_writeback): Likewise
4115 2014-04-04 Richard Biener <rguenther@suse.de>
4118 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
4119 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
4121 * gimplify.h (gimple_add_tmp_var_fn): Declare.
4122 * gimplify.c (gimple_add_tmp_var_fn): New function.
4123 * gimple-expr.h (create_tmp_reg_fn): Declare.
4124 * gimple-expr.c (create_tmp_reg_fn): New function.
4125 * gimple-low.c (record_vars_into): Don't change cfun.
4126 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
4127 code generation without cfun.
4129 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
4132 * Makefile.in (install-driver): Fix shell scripting.
4134 2014-04-03 Cong Hou <congh@google.com>
4136 PR tree-optimization/60505
4137 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
4138 threshold of number of iterations below which no vectorization
4140 * tree-vect-loop.c (new_loop_vec_info):
4141 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
4142 * tree-vect-loop.c (vect_analyze_loop_operations):
4143 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
4144 * tree-vect-loop.c (vect_transform_loop):
4145 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
4146 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
4147 of iterations of the loop and see if we should build the epilogue.
4149 2014-04-03 Richard Biener <rguenther@suse.de>
4151 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
4152 (streamer_tree_cache_create): Adjust.
4153 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
4154 to allow optional nodes array.
4155 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
4156 (streamer_tree_cache_append): Likewise.
4157 (streamer_tree_cache_create): Create nodes array optionally
4158 as specified by parameter.
4159 * lto-streamer-out.c (create_output_block): Avoid maintaining
4160 the node array in the writer cache.
4161 (DFS_write_tree): Remove assertion.
4162 (produce_asm_for_decls): Free the out decl state hash table early.
4163 * lto-streamer-in.c (lto_data_in_create): Adjust for
4164 streamer_tree_cache_create prototype change.
4166 2014-04-03 Richard Biener <rguenther@suse.de>
4168 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
4169 set TREE_CHAIN to NULL_TREE.
4171 2014-04-03 Richard Biener <rguenther@suse.de>
4173 PR tree-optimization/60740
4174 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
4175 over all GIMPLE_COND operands.
4177 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
4179 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
4180 (Weffc++): Remove Scott's numbering, merge lists and reference
4183 2014-04-03 Nick Clifton <nickc@redhat.com>
4185 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
4188 2014-04-03 Martin Jambor <mjambor@suse.cz>
4190 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
4191 mention gcc_unreachable before failing.
4192 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
4195 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
4198 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
4199 inconsistent code and instead mark the context inconsistent.
4200 (possible_polymorphic_call_targets): For inconsistent contexts
4201 return empty complete list.
4203 2014-04-02 Anthony Green <green@moxielogic.com>
4205 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
4206 (extendqisi2, extendhisi2): Define.
4207 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
4208 (WCHAR_TYPE): Change to unsigned int.
4210 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4212 PR tree-optimization/60733
4213 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
4214 insertion point for PHI candidates to be the end of the feeding
4215 block for the PHI argument.
4217 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
4219 PR rtl-optimization/60650
4220 * lra-constraints.c (process_alt_operands): Decrease reject for
4221 earlyclobber matching.
4223 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4225 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
4227 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4229 * config/spu/spu.c (pad_bb): Do not crash when the last
4230 insn is CODE_FOR_blockage.
4232 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4234 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
4235 lies outside the target mode.
4237 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4240 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4241 software floating point or no floating point registers, do not
4242 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4243 in GPRs that occurs after we tested for GPRs that would never be
4246 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4247 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4248 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4249 specifically allow DDmode, since that does not use the SPE SIMD
4252 2014-04-02 Richard Biener <rguenther@suse.de>
4255 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4256 MODE_INTs. Properly use negv_optab.
4257 (expand_abs): Likewise.
4259 2014-04-02 Richard Biener <rguenther@suse.de>
4262 * Makefile.in (install-driver): Guard extra installs with special
4265 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4267 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4270 2014-04-01 Richard Henderson <rth@redhat.com>
4273 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
4274 alternative enabled before register allocation.
4276 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
4278 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
4279 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
4281 (nios2_large_got_address): Remove unneeded 'sym' parameter.
4282 (nios2_got_address): Update nios2_large_got_address call site.
4283 (nios2_delegitimize_address): New function.
4284 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
4285 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
4286 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
4288 2014-04-01 Martin Husemann <martin@duskware.de>
4290 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
4293 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
4295 PR rtl-optimization/60604
4296 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
4297 check from register_operand.
4298 (register_operand): Redefine in terms of general_operand.
4299 (nonmemory_operand): Use register_operand for the non-constant cases.
4301 2014-04-01 Richard Biener <rguenther@suse.de>
4303 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
4305 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
4307 * doc/invoke.texi (mapp-regs): Clarify.
4309 2014-03-31 Ulrich Drepper <drepper@gmail.com>
4311 * config/i386/avx512fintrin.h (__v32hi): Define type.
4312 (__v64qi): Likewise.
4313 (_mm512_set1_epi8): Define.
4314 (_mm512_set1_epi16): Define.
4315 (_mm512_set4_epi32): Define.
4316 (_mm512_set4_epi64): Define.
4317 (_mm512_set4_pd): Define.
4318 (_mm512_set4_ps): Define.
4319 (_mm512_setr4_epi64): Define.
4320 (_mm512_setr4_epi32): Define.
4321 (_mm512_setr4_pd): Define.
4322 (_mm512_setr4_ps): Define.
4323 (_mm512_setzero_epi32): Define.
4325 2014-03-31 Martin Jambor <mjambor@suse.cz>
4328 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
4329 callsite_arguments_match_p. Updated all callers. Also check types of
4330 corresponding formal parameters and actual arguments.
4331 (not_all_callers_have_enough_arguments_p) Renamed to
4332 some_callers_have_mismatched_arguments_p.
4334 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
4336 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
4338 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
4341 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
4344 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
4346 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
4347 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
4349 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
4350 Use FMAMODE_NOVF512 mode iterator.
4351 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
4352 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
4353 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
4354 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
4356 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
4357 Use VF_128_256 mode iterator.
4358 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
4361 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4363 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
4364 static chain if needed.
4366 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4369 * lra-constraints.c (index_part_to_reg): New.
4370 (process_address): Use it.
4372 2014-03-27 Jeff Law <law@redhat.com>
4373 Jakub Jelinek <jakub@redhat.com>
4376 * expr.c (do_tablejump): Use simplify_gen_binary rather than
4377 gen_rtx_{PLUS,MULT} to build up the address expression.
4379 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
4380 creating non-canonical RTL.
4382 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4385 * ipa-inline.c (want_inline_small_function_p): Short circuit large
4386 functions; reorganize to make cheap checks first.
4387 (inline_small_functions): Do not estimate growth when dumping;
4389 * ipa-inline.h (inline_summary): Add min_size.
4390 (growth_likely_positive): New function.
4391 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
4392 (set_cond_stmt_execution_predicate): Cleanup.
4393 (estimate_edge_size_and_time): Compute min_size.
4394 (estimate_calls_size_and_time): Likewise.
4395 (estimate_node_size_and_time): Likewise.
4396 (inline_update_overall_summary): Update min_size.
4397 (do_estimate_edge_time): Likewise.
4398 (do_estimate_edge_size): Update.
4399 (do_estimate_edge_hints): Update.
4400 (growth_likely_positive): New function.
4402 2014-03-28 Jakub Jelinek <jakub@redhat.com>
4405 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
4406 also if addr has VOIDmode.
4408 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4410 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
4411 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
4413 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
4414 instructions as well as AdvancedSIMD loads.
4416 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4418 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
4419 Use crypto_aese type.
4420 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
4421 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
4422 crypto_aese, crypto_aesmc. Move to types.md.
4423 * config/arm/types.md (crypto_aes): Split into crypto_aese,
4425 * config/arm/iterators.md (crypto_type): Likewise.
4427 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4429 * cgraph.c: Include expr.h and tree-dfa.h.
4430 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
4433 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4436 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
4437 regs from checking multi-reg pseudos.
4439 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4441 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
4443 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4445 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
4446 if it would clobber the stack pointer, even temporarily.
4448 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
4450 * mode-switching.c: Make small adjustments to the top comment.
4452 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4454 * config/rs6000/constraints.md (wD constraint): New constraint to
4455 match the constant integer to get the top DImode/DFmode out of a
4456 vector in a VSX register.
4458 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
4459 match the constant integer to get the top DImode/DFmode out of a
4460 vector in a VSX register.
4462 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
4465 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
4468 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
4469 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
4471 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
4472 Optimize vec_extract of 64-bit values, where the value being
4473 extracted is in the top word, where we can use scalar
4474 instructions. Add direct move and store support. Combine the big
4475 endian/little endian vector select load support into a single insn.
4476 (vsx_extract_<mode>_internal1): Likewise.
4477 (vsx_extract_<mode>_internal2): Likewise.
4478 (vsx_extract_<mode>_load): Likewise.
4479 (vsx_extract_<mode>_store): Likewise.
4480 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
4481 combined into vsx_extract_<mode>_load.
4482 (vsx_extract_<mode>_one_le): Likewise.
4484 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
4485 define the top 64-bit vector element.
4487 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
4490 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4491 Document vec_vbpermq builtin.
4494 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
4495 enable use of xxsldwi and xxpermdi builtin functions.
4496 (vec_xxpermdi): Likewise.
4498 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4499 Document use of vec_xxsldwi and vec_xxpermdi builtins.
4501 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
4503 PR rtl-optimization/60650
4504 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
4506 (find_spills_for): New.
4507 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
4508 Spill all pseudos on the second iteration.
4510 2014-03-27 Marek Polacek <polacek@redhat.com>
4513 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
4516 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4518 * config/s390/s390.c (s390_can_use_return_insn): Check for
4519 call-saved FPRs on 31 bit.
4521 2014-03-27 Jakub Jelinek <jakub@redhat.com>
4524 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
4525 if they need regimplification, just drop them instead of
4526 calling gimple_regimplify_operands on them.
4528 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
4531 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
4532 (aarch64_frame_pointer_required): Adjust logic.
4533 (aarch64_can_eliminate): Adjust logic.
4534 (aarch64_override_options_after_change): Adjust logic.
4536 2014-03-27 Dehao Chen <dehao@google.com>
4538 * ipa-inline.c (early_inliner): Update node's inline info.
4540 2014-03-26 Dehao Chen <dehao@google.com>
4542 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
4543 compiler inserted conditional jumps for NAN float check.
4545 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4547 * ubsan.h (ubsan_create_data): Change second argument's type
4548 to const location_t *.
4549 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
4551 (ubsan_create_data): Change second argument to const location_t *PLOC.
4552 Create Loc field whenever PLOC is non-NULL.
4553 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
4554 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
4558 * real.c (real_to_integer2): Change type of low to UHWI.
4560 2014-03-26 Tobias Burnus <burnus@net-b.de>
4562 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
4563 (CILK_SELF_SPECS): New define.
4564 (driver_self_specs): Use it.
4566 2014-03-26 Richard Biener <rguenther@suse.de>
4568 * tree-pretty-print.c (percent_K_format): Implement special
4569 case for LTO and its stripped down BLOCK tree.
4571 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4574 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
4576 * tree-vrp.c (simplify_internal_call_using_ranges): If only
4577 one range is range_int_cst_p, but not both, at least optimize
4578 addition/subtraction of 0 and multiplication by 0 or 1.
4579 * gimple-fold.c (gimple_fold_call): Fold
4580 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
4581 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
4582 INTEGER_CSTs, try to fold at least x * 0 and y - y.
4584 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
4586 PR rtl-optimization/60452
4587 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
4588 <case REG>: Return 1 for invalid offsets from the frame pointer.
4590 2014-03-26 Marek Polacek <polacek@redhat.com>
4593 * doc/extend.texi (C Extensions): Mention variable-length arrays in
4596 2014-03-26 Marek Polacek <polacek@redhat.com>
4599 * doc/extend.texi (Designated Inits): Describe what happens to omitted
4602 2014-03-26 Marek Polacek <polacek@redhat.com>
4605 * ira-color.c (update_conflict_hard_regno_costs): Perform the
4606 multiplication in unsigned type.
4608 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4610 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
4612 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4614 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
4616 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4619 * cif-code.def (UNREACHABLE) New code.
4620 * ipa-inline.c (inline_small_functions): Skip edges to
4621 __builtlin_unreachable.
4622 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
4623 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
4624 predicate to __bulitin_unreachable.
4625 (set_cond_stmt_execution_predicate): Fix issue when
4626 invert_tree_comparison returns ERROR_MARK.
4627 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
4628 propagate to inline clones.
4629 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
4631 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
4632 * cgraphclones.c (cgraph_clone_node): If call destination is already
4633 ureachable, do not redirect it back.
4634 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
4637 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4639 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
4640 Do not modify inline clones.
4642 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4644 * config/i386/i386.md (general_sext_operand): New mode attr.
4645 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
4646 don't generate (sign_extend (const_int)).
4647 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
4648 operands[2]. Use We constraint instead of <i> and
4649 <general_sext_operand> predicate instead of <general_operand>.
4650 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
4651 * config/i386/constraints.md (We): New constraint.
4652 * config/i386/predicates.md (x86_64_sext_operand,
4653 sext_operand): New predicates.
4655 2014-03-25 Martin Jambor <mjambor@suse.cz>
4658 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
4659 inconsistent devirtualizations to __builtin_unreachable.
4661 2014-03-25 Marek Polacek <polacek@redhat.com>
4664 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
4666 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
4668 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
4669 order of elements for big-endian.
4671 2014-03-25 Richard Biener <rguenther@suse.de>
4674 * gimplify-me.c (gimple_regimplify_operands): Update the
4677 2014-03-25 Martin Jambor <mjambor@suse.cz>
4680 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
4681 (lto_output_varpool_node): Likewise.
4682 (input_overwrite_node): Likewise.
4683 (input_varpool_node): Likewise.
4685 2014-03-25 Richard Biener <rguenther@suse.de>
4687 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
4688 (run_gcc): Likewise.
4690 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4692 * combine.c (simplify_compare_const): Add MODE argument.
4693 Handle mode_width 0 as very large mode_width.
4694 (try_combine, simplify_comparison): Adjust callers.
4696 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
4697 type to avoid signed integer overflow.
4698 * explow.c (plus_constant): Likewise.
4700 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
4702 * doc/generic.texi: Correct typos.
4704 2014-03-24 Tobias Burnus <burnus@net-b.de>
4706 * doc/invoke.texi (-flto): Expand section about
4707 using static libraries with LTO.
4709 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4711 PR rtl-optimization/60501
4712 * optabs.def (addptr3_optab): New optab.
4713 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
4714 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
4715 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
4717 * lra.c (emit_add3_insn): Use the addptr pattern if available.
4719 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
4721 2014-03-24 Ulrich Drepper <drepper@gmail.com>
4723 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
4726 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
4727 (_mm256_undefined_ps): Define.
4728 (_mm256_undefined_pd): Define.
4729 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
4730 (_mm_undefined_pd): Define.
4731 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
4732 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
4733 (_mm512_undefined_ps): Define.
4734 (_mm512_undefined_pd): Define.
4735 Use _mm*_undefined_*.
4736 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
4738 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
4740 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
4741 (lshr_simd): DI mode added.
4742 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
4743 (aarch64_ushr_simddi): Likewise.
4744 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
4745 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
4746 (vshrd_n_u64): Likewise.
4748 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4750 * Makefile.in (s-macro_list): Depend on cc1.
4752 2014-03-23 Teresa Johnson <tejohnson@google.com>
4754 * ipa-utils.c (ipa_print_order): Use specified dump file.
4756 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
4758 PR rtl-optimization/60601
4759 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
4761 * gcc.c (eval_spec_function): Initialize save_growing_value.
4763 2014-03-22 Jakub Jelinek <jakub@redhat.com>
4766 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
4767 code == MINUS_EXPR, never swap op0 with op1.
4769 * toplev.c (init_local_tick): Avoid signed integer multiplication
4771 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
4772 shift by first operand's bitsize.
4774 2014-03-21 Jakub Jelinek <jakub@redhat.com>
4777 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
4779 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
4780 TARGET_ISA_64BIT_P(x).
4782 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4784 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
4785 pattern for vector nor instead of subtract from splat(-1).
4786 (altivec_expand_vec_perm_const_le): Likewise.
4788 2014-03-21 Richard Henderson <rth@twiddle.net>
4791 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
4792 related insns after epilogue_completed.
4794 2014-03-21 Martin Jambor <mjambor@suse.cz>
4797 * cgraph.h (symtab_node): New flag body_removed.
4798 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
4799 when removing bodies.
4800 * symtab.c (dump_symtab_base): Dump body_removed flag.
4801 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
4802 had their bodies removed.
4804 2014-03-21 Martin Jambor <mjambor@suse.cz>
4807 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
4810 2014-03-21 Richard Biener <rguenther@suse.de>
4812 PR tree-optimization/60577
4813 * tree-core.h (struct tree_base): Document nothrow_flag use
4814 in VAR_DECL_NONALIASED.
4815 * tree.h (VAR_DECL_NONALIASED): New.
4816 (may_be_aliased): Adjust.
4817 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
4819 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4821 * expr.c (expand_expr_real_1): Remove outdated comment.
4823 2014-03-20 Jakub Jelinek <jakub@redhat.com>
4826 * ira.c (adjust_cleared_regs): Call copy_rtx on
4827 *reg_equiv[REGNO (loc)].src_p before passing it to
4828 simplify_replace_fn_rtx.
4831 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
4832 into CONST, put pic register as first operand of PLUS. Use
4833 gen_const_mem for both 32-bit and 64-bit PIC got loads.
4835 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4837 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
4839 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4841 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
4842 around for store forwarding issue in the FPU on the UT699.
4843 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
4844 loads and operations if -mfix-ut699 is specified.
4845 (divtf3_hq): Tweak attribute.
4846 (sqrttf2_hq): Likewise.
4848 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4850 * calls.c (store_one_arg): Remove incorrect const qualification on the
4851 type of the temporary.
4852 * cfgexpand.c (expand_return): Likewise.
4853 * expr.c (expand_constructor): Likewise.
4854 (expand_expr_real_1): Likewise.
4856 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4858 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
4861 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
4864 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
4866 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
4868 * config/arm/aarch-common-protos.h
4869 (alu_cost_table): Fix spelling of "extend".
4870 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
4872 2014-03-19 Richard Biener <rguenther@suse.de>
4875 * tree-core.h (tree_type_common): Re-order pointer members
4876 to reduce recursion depth during GC walks.
4878 2014-03-19 Marek Polacek <polacek@redhat.com>
4881 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
4882 before accessing it.
4884 2014-03-19 Richard Biener <rguenther@suse.de>
4887 * lto-streamer-in.c (input_function): In WPA stage do not drop
4890 2014-03-19 Jakub Jelinek <jakub@redhat.com>
4892 PR tree-optimization/60559
4893 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
4894 with build_zero_cst assignment.
4896 2014-03-18 Kai Tietz <ktietz@redhat.com>
4898 PR rtl-optimization/56356
4899 * sdbout.c (sdbout_parms): Verify that parms'
4900 incoming argument is valid.
4901 (sdbout_reg_parms): Likewise.
4903 2014-03-18 Richard Henderson <rth@redhat.com>
4906 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
4907 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
4908 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
4910 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
4912 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
4913 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
4914 Italicize plugin event names in description. Explain that
4915 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
4916 Remind that no GCC functions should be called after PLUGIN_FINISH.
4917 Explain what pragmas with expansion are.
4919 2014-03-18 Martin Liska <mliska@suse.cz>
4921 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
4922 gimple call statement is update.
4923 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
4924 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
4926 2014-03-18 Jakub Jelinek <jakub@redhat.com>
4929 * ubsan.c (ubsan_instrument_unreachable): Call
4930 initialize_sanitizer_builtins.
4931 (ubsan_pass): Likewise.
4934 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
4935 varpool_finalize_decl instead of rest_of_decl_compilation.
4937 2014-03-18 Richard Biener <rguenther@suse.de>
4939 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
4940 by using bitmap_and_compl instead of bitmap_and_compl_into.
4941 (df_rd_transfer_function): Likewise.
4943 2014-03-18 Richard Biener <rguenther@suse.de>
4945 * doc/lto.texi (fresolution): Fix typo.
4947 2014-03-18 Richard Biener <rguenther@suse.de>
4949 * doc/invoke.texi (flto): Update for changes in 4.9.
4951 2014-03-18 Richard Biener <rguenther@suse.de>
4953 * doc/loop.texi: Remove section on the removed lambda framework.
4954 Update loop docs with recent changes in preserving loop structure.
4956 2014-03-18 Richard Biener <rguenther@suse.de>
4958 * doc/lto.texi (-fresolution): Document.
4960 2014-03-18 Richard Biener <rguenther@suse.de>
4962 * doc/contrib.texi: Adjust my name.
4964 2014-03-18 Jakub Jelinek <jakub@redhat.com>
4967 * internal-fn.c: Include diagnostic-core.h.
4968 (expand_BUILTIN_EXPECT): New function.
4969 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
4970 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
4971 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
4972 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
4974 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
4975 Revert 3 argument __builtin_expect code.
4976 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
4977 * gimple-fold.c (gimple_fold_call): Likewise.
4978 * tree.h (fold_builtin_expect): New prototype.
4979 * builtins.c (build_builtin_expect_predicate): Add predictor
4980 argument, if non-NULL, create 3 argument __builtin_expect.
4981 (fold_builtin_expect): No longer static. Add ARG2 argument,
4982 pass it through to build_builtin_expect_predicate.
4983 (fold_builtin_2): Adjust caller.
4984 (fold_builtin_3): Handle BUILT_IN_EXPECT.
4985 * internal-fn.def (BUILTIN_EXPECT): New.
4987 2014-03-18 Tobias Burnus <burnus@net-b.de>
4990 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
4991 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
4992 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
4994 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
4997 * predict.c (combine_predictions_for_bb): Fix up formatting.
4998 (expr_expected_value_1, expr_expected_value): Add predictor argument,
4999 fill what it points to if non-NULL.
5000 (tree_predict_by_opcode): Adjust caller, use the predictor.
5001 * predict.def (PRED_COMPARE_AND_SWAP): Add.
5003 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
5005 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
5006 proper constant for the store mode.
5008 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
5010 * symtab.c (change_decl_assembler_name): Fix transparent alias
5013 2014-03-16 Renlin Li <Renlin.Li@arm.com>
5015 * config/aarch64/aarch64.c: Correct the comments about the
5016 aarch64 stack layout.
5018 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
5020 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
5021 check for GF_OMP_FOR_KIND_FOR.
5023 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
5025 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
5026 ymm and zmm register names.
5028 2014-03-17 Jakub Jelinek <jakub@redhat.com>
5031 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
5032 note creation for the 2010-08-31 changes.
5034 2014-03-17 Marek Polacek <polacek@redhat.com>
5037 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
5038 as -fno-tree-loop-vectorize.
5039 (expand_omp_simd): Likewise.
5041 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
5043 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
5044 (eligible_for_call_delay): New prototype.
5045 * config/sparc/sparc.c (tls_call_delay): Rename into...
5046 (eligible_for_call_delay): ...this. Return false if the instruction
5047 cannot be put in the delay slot of a branch.
5048 (eligible_for_restore_insn): Simplify.
5049 (eligible_for_return_delay): Return false if the instruction cannot be
5050 put in the delay slot of a branch and simplify.
5051 (eligible_for_sibcall_delay): Return false if the instruction cannot be
5052 put in the delay slot of a branch.
5053 * config/sparc/sparc.md (fix_ut699): New attribute.
5054 (tls_call_delay): Delete.
5055 (in_call_delay): Reimplement.
5056 (eligible_for_sibcall_delay): Rename into...
5057 (in_sibcall_delay): ...this.
5058 (eligible_for_return_delay): Rename into...
5059 (in_return_delay): ...this.
5060 (in_branch_delay): Reimplement.
5061 (in_uncond_branch_delay): Delete.
5062 (in_annul_branch_delay): Delete.
5064 2014-03-14 Richard Henderson <rth@redhat.com>
5067 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
5068 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
5069 (*floathi<X87MODEF>2_i387_with_temp): Remove.
5070 (floathi splitters): Remove.
5071 (float<SWI48x>xf2): New pattern.
5072 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
5073 code that tried to handle DImode for 32-bit, but which was excluded
5074 by the pattern's condition. Drop allocation of stack temporary.
5075 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
5076 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
5077 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
5078 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
5079 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
5080 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
5081 (*float<SWI48><MODEF>2_sse_interunit): Remove.
5082 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
5083 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
5084 (*float<SWI48x><X87MODEF>2_i387): Remove.
5085 (all float _with_temp splitters): Remove.
5086 (*float<SWI48x><MODEF>2_i387): New pattern.
5087 (*float<SWI48><MODEF>2_sse): New pattern.
5088 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
5089 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
5091 2014-03-14 Jakub Jelinek <jakub@redhat.com>
5092 Marek Polacek <polacek@redhat.com>
5095 * common.opt (dump_base_name_prefixed): New Variable.
5096 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
5097 if x_dump_base_name_prefixed is already set, set it at the end.
5099 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
5101 PR rtl-optimization/60508
5102 * lra-constraints.c (get_reload_reg): Add new parameter
5104 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
5105 Pass the new parameter values.
5107 2014-03-14 Richard Biener <rguenther@suse.de>
5109 * common.opt: Revert unintented changes from r205065.
5112 2014-03-14 Richard Biener <rguenther@suse.de>
5115 * cfghooks.c (split_block): Properly adjust all loops the
5116 block was a latch of.
5118 2014-03-14 Martin Jambor <mjambor@suse.cz>
5121 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
5124 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
5127 * config/avr/avr.c (avr_set_current_function): Pass function name
5128 through default_strip_name_encoding before sanity checking instead
5129 of skipping the first char of the assembler name.
5131 2014-03-13 Richard Henderson <rth@redhat.com>
5134 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
5135 (ix86_force_to_memory, ix86_free_from_memory): Remove.
5136 * config/i386/i386-protos.h: Likewise.
5137 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
5138 in the expander instead of a splitter.
5139 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
5140 any possibility of requiring a memory.
5141 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
5142 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
5143 (fp branch splitters): Update for ix86_split_fp_branch.
5144 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
5145 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
5146 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
5147 (*fop_<MODEF>_2_i387): Remove f/r alternative.
5148 (*fop_<MODEF>_3_i387): Likewise.
5149 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
5150 (splitters for the fop_* register patterns): Remove.
5151 (fscalexf4_i387): Rename from *fscalexf4_i387.
5152 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
5154 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5156 PR tree-optimization/59779
5157 * tree-dfa.c (get_ref_base_and_extent): Use double_int
5158 type for bitsize and maxsize instead of HOST_WIDE_INT.
5160 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
5162 PR rtl-optimization/57320
5163 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
5164 the CFG after thread_prologue_and_epilogue_insns.
5166 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
5168 PR rtl-optimization/57189
5169 * lra-constraints.c (process_alt_operands): Disfavor spilling
5172 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
5174 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
5176 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5178 PR tree-optimization/59025
5180 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
5181 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
5183 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
5186 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
5187 calls of avr_out_plus_1.
5189 2014-03-13 Bin Cheng <bin.cheng@arm.com>
5191 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
5192 BB's single pred and update the father loop's latch info later.
5194 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5196 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
5200 (VEC_base): Likewise.
5201 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
5202 registers, we need to swap double words in little endian mode.
5204 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
5205 to be a container mode for 128-bit integer operations added in ISA
5206 2.07. Unlike TImode and PTImode, the preferred register set is
5207 the Altivec/VMX registers for the 128-bit operations.
5209 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
5211 (rs6000_split_128bit_ok_p): Likewise.
5213 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
5214 macros for creating ISA 2.07 normal and overloaded builtin
5215 functions with 3 arguments.
5216 (BU_P8V_OVERLOAD_3): Likewise.
5217 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
5218 for use as overloaded functions.
5219 (VPERM_1TI_UNS): Likewise.
5220 (VSEL_1TI): Likewise.
5221 (VSEL_1TI_UNS): Likewise.
5222 (ST_INTERNAL_1ti): Likewise.
5223 (LD_INTERNAL_1ti): Likewise.
5224 (XXSEL_1TI): Likewise.
5225 (XXSEL_1TI_UNS): Likewise.
5226 (VPERM_1TI): Likewise.
5227 (VPERM_1TI_UNS): Likewise.
5228 (XXPERMDI_1TI): Likewise.
5229 (SET_1TI): Likewise.
5230 (LXVD2X_V1TI): Likewise.
5231 (STXVD2X_V1TI): Likewise.
5232 (VEC_INIT_V1TI): Likewise.
5233 (VEC_SET_V1TI): Likewise.
5234 (VEC_EXT_V1TI): Likewise.
5235 (EQV_V1TI): Likewise.
5236 (NAND_V1TI): Likewise.
5237 (ORC_V1TI): Likewise.
5238 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
5239 added in ISA 2.07. Add both normal 'altivec' builtins, and the
5241 (VADDUQM): Likewise.
5242 (VSUBCUQ): Likewise.
5243 (VADDEUQM): Likewise.
5244 (VADDECUQ): Likewise.
5245 (VSUBEUQM): Likewise.
5246 (VSUBECUQ): Likewise.
5248 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5249 __int128_t and __uint128_t types.
5250 (__uint128_type): Likewise.
5251 (altivec_categorize_keyword): Add support for vector __int128_t,
5252 vector __uint128_t, vector __int128, and vector unsigned __int128
5253 as a container type for TImode operations that need to be done in
5254 VSX/Altivec registers.
5255 (rs6000_macro_to_expand): Likewise.
5256 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5257 to support 128-bit integer instructions vaddcuq, vadduqm,
5258 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5259 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5261 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5262 for V1TImode, and set up preferences to use VSX/Altivec registers.
5263 Setup VSX reload handlers.
5264 (rs6000_debug_reg_global): Likewise.
5265 (rs6000_init_hard_regno_mode_ok): Likewise.
5266 (rs6000_preferred_simd_mode): Likewise.
5267 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5268 (easy_altivec_constant): Likewise.
5269 (output_vec_const_move): Likewise.
5270 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5272 (rs6000_expand_vector_extract): Likewise.
5273 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
5275 (rs6000_const_vec): Add support for V1TImode.
5276 (rs6000_emit_le_vsx_load): Swap double words when loading or
5277 storing TImode/V1TImode.
5278 (rs6000_emit_le_vsx_store): Likewise.
5279 (rs6000_emit_le_vsx_move): Likewise.
5280 (rs6000_emit_move): Add support for V1TImode.
5281 (altivec_expand_ld_builtin): Likewise.
5282 (altivec_expand_st_builtin): Likewise.
5283 (altivec_expand_vec_init_builtin): Likewise.
5284 (altivec_expand_builtin): Likewise.
5285 (rs6000_init_builtins): Add support for V1TImode type. Add
5286 support for ISA 2.07 128-bit integer builtins. Define type names
5287 for the VSX/Altivec vector types.
5288 (altivec_init_builtins): Add support for overloaded vector
5289 functions with V1TImode type.
5290 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
5291 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
5293 (rs6000_split_128bit_ok_p): Likewise.
5294 (rs6000_handle_altivec_attribute): Create V1TImode from vector
5295 __int128_t and vector __uint128_t.
5297 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
5298 and mode attributes.
5305 (VS_scalar): Likewise.
5306 (VS_double): Likewise.
5307 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
5309 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
5310 we support the ISA 2.07 128-bit integer arithmetic instructions.
5311 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
5312 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
5313 and TImode types for use with the builtin functions.
5314 (V1TI_type_node): Likewise.
5315 (unsigned_V1TI_type_node): Likewise.
5316 (intTI_type_internal_node): Likewise.
5317 (uintTI_type_internal_node): Likewise.
5319 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
5320 128-bit builtin functions.
5321 (UNSPEC_VADDEUQM): Likewise.
5322 (UNSPEC_VADDECUQ): Likewise.
5323 (UNSPEC_VSUBCUQ): Likewise.
5324 (UNSPEC_VSUBEUQM): Likewise.
5325 (UNSPEC_VSUBECUQ): Likewise.
5326 (VM): Add V1TImode to vector mode iterators.
5328 (VI_unit): Likewise.
5329 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
5330 (altivec_vaddcuq): Likewise.
5331 (altivec_vsubuqm): Likewise.
5332 (altivec_vsubcuq): Likewise.
5333 (altivec_vaddeuqm): Likewise.
5334 (altivec_vaddecuq): Likewise.
5335 (altivec_vsubeuqm): Likewise.
5336 (altivec_vsubecuq): Likewise.
5338 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
5340 (BOOL_128): Likewise.
5341 (BOOL_REGS_OUTPUT): Likewise.
5342 (BOOL_REGS_OP1): Likewise.
5343 (BOOL_REGS_OP2): Likewise.
5344 (BOOL_REGS_UNARY): Likewise.
5345 (BOOL_REGS_AND_CR0): Likewise.
5347 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
5348 128-bit integer builtin support.
5349 (vec_vadduqm): Likewise.
5350 (vec_vaddecuq): Likewise.
5351 (vec_vaddeuqm): Likewise.
5352 (vec_vsubecuq): Likewise.
5353 (vec_vsubeuqm): Likewise.
5354 (vec_vsubcuq): Likewise.
5355 (vec_vsubuqm): Likewise.
5357 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5358 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
5359 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
5360 128-bit integer add/subtract to ISA 2.07.
5362 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
5364 * config/arc/arc.c (arc_predicate_delay_insns):
5365 Fix third argument passed to conditionalize_nonjump.
5367 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
5369 * config/aarch64/aarch64-builtins.c
5370 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
5371 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
5372 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
5373 instead of __builtin_lfloor.
5374 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
5376 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5378 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
5379 (tree_ssa_ifcombine_bb_1): New function.
5380 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
5381 is an empty forwarder block to then_bb or vice versa and then_bb
5382 and else_bb are effectively swapped.
5384 2014-03-12 Christian Bruel <christian.bruel@st.com>
5387 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
5388 REG_CFA_DEF_CFA note.
5389 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
5390 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
5392 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
5394 PR tree-optimization/60454
5395 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
5397 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5399 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
5400 Do not define target_cpu_default2 to generic.
5401 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
5402 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
5403 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
5405 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5406 Marc Glisse <marc.glisse@inria.fr>
5408 PR tree-optimization/60502
5409 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
5410 instead of build_low_bits_mask.
5412 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5415 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
5416 if there are multiple uses, but op doesn't live on E edge.
5417 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
5418 clobber stmts before __builtin_unreachable.
5420 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
5422 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
5423 hard_frame_pointer_rtx.
5424 * cse.c (cse_insn): Remove volatile check.
5425 * cselib.c (cselib_process_insn): Likewise.
5426 * dse.c (scan_insn): Likewise.
5428 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5430 * config/arc/arc.c (conditionalize_nonjump): New function,
5432 (arc_ifcvt): ... this.
5433 (arc_predicate_delay_insns): Use it.
5435 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5437 * config/arc/predicates.md (extend_operand): During/after reload,
5438 allow const_int_operand.
5439 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
5440 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
5441 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
5443 (umulsi3_highpart_i): Likewise.
5445 2014-03-11 Richard Biener <rguenther@suse.de>
5447 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
5448 Add asserts to guard possible wrong-code bugs.
5450 2014-03-11 Richard Biener <rguenther@suse.de>
5452 PR tree-optimization/60429
5453 PR tree-optimization/60485
5454 * tree-ssa-structalias.c (set_union_with_increment): Properly
5455 take into account all fields that overlap the shifted vars.
5456 (do_sd_constraint): Likewise.
5457 (do_ds_constraint): Likewise.
5458 (get_constraint_for_ptr_offset): Likewise.
5460 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
5462 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
5463 (nios2_compute_frame_layout):
5464 Add calculation of cfun->machine->fp_save_offset.
5465 (nios2_expand_prologue): Correct setting of frame pointer register
5467 (nios2_expand_epilogue): Update recovery of stack pointer from
5468 frame pointer accordingly.
5469 (nios2_initial_elimination_offset): Update calculation of offset
5470 for eliminating to HARD_FRAME_POINTER_REGNUM.
5472 2014-03-10 Jakub Jelinek <jakub@redhat.com>
5475 * ipa.c (symtab_remove_unreachable_nodes): Don't call
5476 cgraph_get_create_node on VAR_DECLs.
5478 2014-03-10 Richard Biener <rguenther@suse.de>
5481 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
5483 2014-03-08 Douglas B Rupp <rupp@gnat.com>
5485 * config/vms/vms.opt (vms_float_format): New variable.
5487 2014-03-08 Tobias Burnus <burnus@net-b.de>
5489 * doc/invoke.texi (-fcilkplus): Update implementation status.
5491 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
5492 Richard Biener <rguenther@suse.de>
5494 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
5495 consistently accross all TUs.
5496 (run_gcc): Enable -fshort-double automatically at link at link-time
5497 and disallow override.
5499 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
5502 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
5503 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
5504 if they can't be used.
5506 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5508 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
5509 for Solaris 11/x86 ld.
5510 * configure: Regenerate.
5512 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5514 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
5515 (LIB_TLS_SPEC): Save as ld_tls_libs.
5516 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
5517 (HAVE_AS_IX86_TLSLDM): New test.
5518 * configure, config.in: Regenerate.
5519 * config/i386/i386.c (legitimize_tls_address): Fall back to
5520 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
5521 cannot support TLS_MODEL_LOCAL_DYNAMIC.
5522 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
5523 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
5525 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
5527 * common.opt (fira-loop-pressure): Mark as optimization.
5529 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
5531 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
5532 an OpenMP mappable type.
5534 2014-03-06 Matthias Klose <doko@ubuntu.com>
5536 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
5537 MULTILIB_OSDIRNAMES is not defined.
5539 2014-03-06 Jakub Jelinek <jakub@redhat.com>
5540 Meador Inge <meadori@codesourcery.com>
5543 * config/arm/arm.c (arm_tls_symbol_p): Remove.
5544 (arm_legitimize_address): Call legitimize_tls_address for any
5545 arm_tls_referenced_p expression, handle constant addend. Call it
5546 before testing for !TARGET_ARM.
5547 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
5549 2014-03-06 Richard Biener <rguenther@suse.de>
5555 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5557 * tree-streamer.c (record_common_node): Assert we don't record
5558 nodes with type double.
5559 (preload_common_node): Skip type double, complex double and double
5560 pointer since it is now frontend dependent due to fshort-double option.
5562 2014-03-06 Richard Biener <rguenther@suse.de>
5564 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
5565 or -fno-lto is specified and the linker has full plugin support.
5566 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
5567 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
5568 * lto-wrapper.c (merge_and_complain): Merge compile-time
5569 optimization levels.
5570 (run_gcc): And pass it through to the link options.
5572 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
5576 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5578 * cselib.c (remove_useless_values): Skip to avoid quadratic
5579 behavior if the condition moved from...
5580 (cselib_process_insn): ... here holds.
5582 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5585 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
5586 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
5589 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
5590 (TM_H): Add x86-tune.def.
5592 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5594 * config/aarch64/aarch64.c (generic_tunings):
5595 Use cortexa57_extra_costs.
5597 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5600 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
5601 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
5602 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
5605 2014-03-04 Heiher <r@hev.cc>
5607 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
5608 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
5610 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
5612 * config/i386/predicates.md (const2356_operand): Change to ...
5613 (const2367_operand): ... this.
5614 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
5616 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5617 (*avx512pf_scatterpf<mode>sf): Ditto.
5618 (avx512pf_scatterpf<mode>df): Ditto.
5619 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5620 (*avx512pf_scatterpf<mode>df): Ditto.
5621 * config/i386/i386.c (ix86_expand_builtin): Update
5622 incorrect hint operand error message.
5624 2014-03-04 Richard Biener <rguenther@suse.de>
5626 * lto-section-in.c (lto_get_section_data): Fix const cast.
5628 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5630 * tree-streamer.c (record_common_node): Assert we don't record
5631 nodes with type double.
5632 (preload_common_node): Skip type double, complex double and double
5633 pointer since it is now frontend dependent due to fshort-double option.
5635 2014-03-04 Richard Biener <rguenther@suse.de>
5638 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
5639 (lto_input_toplevel_asms): Likewise.
5640 * lto-section-in.c (lto_get_section_data): Instead do it here
5643 2014-03-04 Richard Biener <rguenther@suse.de>
5645 PR tree-optimization/60382
5646 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
5647 dead PHIs a reduction.
5649 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
5651 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
5653 (_mm_prefetch): Move out of GCC target("sse") pragma.
5654 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
5655 GCC target("prfchw") pragma.
5656 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
5658 * config/i386/i386.c (ix86_option_override_internal): Enable
5659 -mprfchw with -mprefetchwt1.
5661 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5663 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
5666 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5668 * opts.h (CL_PCH_IGNORE): Define.
5669 * targhooks.c (option_affects_pch_p):
5670 Return false for options that have CL_PCH_IGNORE set.
5671 * opt-functions.awk: Process PchIgnore.
5672 * doc/options.texi: Document PchIgnore.
5674 * config/arc/arc.opt (misize): Add PchIgnore property.
5676 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5678 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
5679 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
5680 constraint on constants to permit them being loaded into
5681 GENERAL_REGS or BASE_REGS.
5683 2014-03-03 Nick Clifton <nickc@redhat.com>
5685 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
5686 anti-cacnonical alternatives.
5687 (negandhi3_real): New pattern.
5688 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
5690 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5692 * config/avr/avr-mcus.def: Remove atxmega16x1.
5693 * config/avr/avr-tables.opt: Regenerate.
5694 * config/avr/t-multilib: Regenerate.
5695 * doc/avr-mmcu.texi: Regenerate.
5697 2014-03-03 Tobias Grosser <tobias@grosser.es>
5698 Mircea Namolaru <mircea.namolaru@inria.fr>
5700 PR tree-optimization/58028
5701 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
5704 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5706 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
5707 not handled by recognizers.
5709 2014-03-03 Jakub Jelinek <jakub@redhat.com>
5712 * function.c (expand_function_end): Don't emit
5713 clobber_return_register sequence if clobber_after is a BARRIER.
5714 * cfgexpand.c (construct_exit_block): Append instructions before
5715 return_label to prev_bb.
5717 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5719 * config/rs6000/constraints.md: Document reserved use of "wc".
5721 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5724 * ipa.c (function_and_variable_visibility): When dissolving comdat
5725 group, also set all symbols to local.
5727 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5732 2013-12-14 Jan Hubicka <jh@suse.cz>
5734 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
5736 2014-03-02 Jon Beniston <jon@beniston.com>
5742 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
5743 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
5744 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
5745 (simple_return, *simple_return): New patterns
5746 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
5747 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
5749 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
5751 * dwarf2out.c (gen_subprogram_die): Tidy.
5753 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
5756 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
5757 (*mov_t_msb_neg_negc): ... this new insn.
5759 2014-02-28 Jason Merrill <jason@redhat.com>
5762 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
5765 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
5768 * dwarf2out.c (decltype_auto_die): New static.
5769 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
5770 (gen_type_die_with_usage): Handle 'decltype(auto)'.
5771 (is_cxx_auto): Likewise.
5773 2014-02-28 Ian Bolton <ian.bolton@arm.com>
5775 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
5776 we are not using general regs only.
5778 2014-02-28 Richard Biener <rguenther@suse.de>
5781 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
5782 previous fix and only allow to remove trivial pre-headers
5783 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
5784 (remove_forwarder_block): Properly update the latch of a loop.
5786 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5789 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
5790 (cselib_preserved_hash_table): New.
5791 (preserve_constants_and_equivs): Move preserved vals to it.
5792 (cselib_find_slot): Look it up first.
5793 (cselib_init): Initialize it.
5794 (cselib_finish): Release it.
5795 (dump_cselib_table): Dump it.
5797 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5800 * cselib.c (remove_useless_values): Skip to avoid quadratic
5801 behavior if the condition moved from...
5802 (cselib_process_insn): ... here holds.
5804 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5807 * var-tracking.c (vt_initialize): Apply the same condition to
5808 preserve the CFA base value.
5810 2014-02-28 Joey Ye <joey.ye@arm.com>
5813 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
5814 if reload in progress or completed.
5816 2014-02-28 Tobias Burnus <burnus@net-b.de>
5819 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
5822 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
5824 * doc/tm.texi.in (Condition Code Status): Update documention for
5825 relative locations of cc0-setter and cc0-user.
5827 2014-02-27 Jeff Law <law@redhat.com>
5829 PR rtl-optimization/52714
5830 * combine.c (try_combine): When splitting an unrecognized PARALLEL
5831 into two independent simple sets, if I3 is a jump, ensure the
5832 pattern we place into I3 is a (set (pc) ...).
5834 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
5835 Jeff Law <law@redhat.com>
5837 PR rtl-optimization/49847
5838 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
5839 are in different blocks.
5840 * doc/tm.texi (Condition Code Status): Update documention for
5841 relative locations of cc0-setter and cc0-user.
5843 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
5846 * lra.c (lra_emit_add): Check SUBREG too.
5848 2014-02-27 Andreas Schwab <schwab@suse.de>
5850 * config/m68k/m68k.c (m68k_option_override): Disable
5851 -flive-range-shrinkage for classic m68k.
5852 (m68k_override_options_after_change): Likewise.
5854 2014-02-27 Marek Polacek <polacek@redhat.com>
5857 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
5858 -Wmaybe-uninitialized.
5860 2014-02-27 Alan Modra <amodra@gmail.com>
5863 * reload1.c (emit_input_reload_insns): When reload_override_in,
5864 set old to rl->in_reg when rl->in_reg is a subreg.
5866 2014-02-26 Richard Biener <rguenther@suse.de>
5869 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
5871 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
5873 * common/config/i386/predicates.md (const1256_operand): Remove.
5874 (const2356_operand): New.
5875 (const_1_to_2_operand): Remove.
5876 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5877 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5878 (*avx512pf_gatherpf<mode>sf): Ditto.
5879 (avx512pf_gatherpf<mode>df): Ditto.
5880 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5881 (*avx512pf_gatherpf<mode>df): Ditto.
5882 (avx512pf_scatterpf<mode>sf): Ditto.
5883 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5884 (*avx512pf_scatterpf<mode>sf): Ditto.
5885 (avx512pf_scatterpf<mode>df): Ditto.
5886 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5887 (*avx512pf_scatterpf<mode>df): Ditto.
5888 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
5890 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
5892 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
5893 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
5894 (_mm512_mask_testn_epi64_mask): Move to ...
5895 * config/i386/avx512cdintrin.h: Here.
5896 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
5897 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
5898 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
5899 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
5900 TARGET_AVX512F from TARGET_AVX512CD.
5902 2014-02-26 Richard Biener <rguenther@suse.de>
5905 * ipa.c (walk_polymorphic_call_targets): Properly guard
5906 call to inline_update_overall_summary.
5908 2014-02-26 Bin Cheng <bin.cheng@arm.com>
5911 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
5912 and latches only if requested. Fix latch if it is removed.
5913 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
5914 LOOPS_HAVE_PREHEADERS.
5916 2014-02-25 Andrew Pinski <apinski@cavium.com>
5918 * builtins.c (expand_builtin_thread_pointer): Create a new target
5919 when the target is NULL.
5921 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
5923 PR rtl-optimization/60317
5924 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
5925 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
5926 * lra-assigns.c: Include params.h.
5927 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
5928 other reload pseudos considerations.
5930 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5932 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
5933 to use canonical form for nor<mode>3.
5935 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5938 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
5941 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
5943 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
5944 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
5945 (ix86_handle_option): Handle OPT_mprefetchwt1.
5946 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
5947 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
5949 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
5950 OPTION_MASK_ISA_PREFETCHWT1.
5951 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
5952 (PTA_PREFETCHWT1): New.
5953 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
5954 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
5955 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
5956 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
5957 (*prefetch_avx512pf_<mode>_: Change into ...
5958 (*prefetch_prefetchwt1_<mode>: This.
5959 * config/i386/i386.opt (mprefetchwt1): New.
5960 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
5961 (_mm_prefetch): Handle intent to write.
5962 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
5964 2014-02-25 Richard Biener <rguenther@suse.de>
5967 * emit-rtl.c (mem_attrs_htab): Remove.
5968 (mem_attrs_htab_hash): Likewise.
5969 (mem_attrs_htab_eq): Likewise.
5970 (set_mem_attrs): Always allocate new mem-attrs when something changed.
5971 (init_emit_once): Do not allocate mem_attrs_htab.
5973 2014-02-25 Richard Biener <rguenther@suse.de>
5976 * lto-opts.c (lto_write_options): Output non-explicit conservative
5977 -fwrapv, -fno-trapv and -fno-strict-overflow.
5978 * lto-wrapper.c (merge_and_complain): Handle merging those options.
5979 (run_gcc): And pass them through.
5981 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
5983 * sel-sched.c (calculate_new_fences): New parameter ptime.
5984 Calculate it as a maximum over all fence cycles.
5985 (sel_sched_region_2): Adjust the call to calculate_new_fences.
5986 Print the final schedule timing when sched_verbose.
5988 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
5990 PR rtl-optimization/60292
5991 * sel-sched.c (fill_vec_av_set): Do not reset target availability
5992 bit fot the fence instruction.
5994 2014-02-24 Alangi Derick <alangiderick@gmail.com>
5996 * calls.h: Fix typo in comment.
5998 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
6000 * config/pa/pa.c (pa_output_move_double): Don't valididate when
6001 adjusting offsetable addresses.
6003 2014-02-24 Guozhi Wei <carrot@google.com>
6005 * sparseset.h (sparseset_pop): Fix the wrong index.
6007 2014-02-24 Walter Lee <walt@tilera.com>
6009 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
6010 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
6012 * common/config/tilegx/tilegx-common.c
6013 (TARGET_DEFAULT_TARGET_FLAGS): Define.
6014 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
6016 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
6017 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
6018 (tilegx_gimplify_va_arg_expr): Handle big endian.
6019 (tilegx_expand_unaligned_load): Ditto.
6020 (tilegx_expand_unaligned_store): Ditto.
6021 (TARGET_RETURN_IN_MSB): New.
6022 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
6023 (TARGET_ENDIAN_DEFAULT): New.
6024 (TARGET_BIG_ENDIAN): Handle big endian.
6025 (BYTES_BIG_ENDIAN): Ditto.
6026 (WORDS_BIG_ENDIAN): Ditto.
6027 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
6030 * config/tilegx/tilegx.md (extv): Handle big endian.
6032 (insn_st<n>): Ditto.
6033 (insn_st<n>_add<bitsuffix>): Ditto.
6034 (insn_stnt<n>): Ditto.
6035 (insn_stnt<n>_add<bitsuffix>):Ditto.
6036 (vec_interleave_highv8qi): Handle big endian.
6037 (vec_interleave_highv8qi_be): New.
6038 (vec_interleave_highv8qi_le): New.
6039 (insn_v1int_h): Handle big endian.
6040 (vec_interleave_lowv8qi): Handle big endian.
6041 (vec_interleave_lowv8qi_be): New.
6042 (vec_interleave_lowv8qi_le): New.
6043 (insn_v1int_l): Handle big endian.
6044 (vec_interleave_highv4hi): Handle big endian.
6045 (vec_interleave_highv4hi_be): New.
6046 (vec_interleave_highv4hi_le): New.
6047 (insn_v2int_h): Handle big endian.
6048 (vec_interleave_lowv4hi): Handle big endian.
6049 (vec_interleave_lowv4hi_be): New.
6050 (vec_interleave_lowv4hi_le): New.
6051 (insn_v2int_l): Handle big endian.
6052 (vec_interleave_highv2si): Handle big endian.
6053 (vec_interleave_highv2si_be): New.
6054 (vec_interleave_highv2si_le): New.
6055 (insn_v4int_h): Handle big endian.
6056 (vec_interleave_lowv2si): Handle big endian.
6057 (vec_interleave_lowv2si_be): New.
6058 (vec_interleave_lowv2si_le): New.
6059 (insn_v4int_l): Handle big endian.
6060 * config/tilegx/tilegx.opt (mbig-endian): New option.
6061 (mlittle-endian): New option.
6062 * doc/install.texi: Document tilegxbe-linux.
6063 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
6065 2014-02-24 Martin Jambor <mjambor@suse.cz>
6068 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
6069 there are no parameter descriptors.
6071 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
6073 PR rtl-optimization/60268
6074 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
6075 initialization to ...
6076 (sched_rgn_init): ... here.
6077 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
6079 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6081 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
6084 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
6086 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
6089 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6091 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
6092 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
6094 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6096 * config/microblaze/predicates.md: Add cmp_op predicate.
6097 * config/microblaze/microblaze.md: Add branch_compare instruction
6098 which uses cmp_op predicate and emits cmp insn before branch.
6099 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
6100 to microblaze_expand_conditional_branch and consolidate logic.
6101 (microblaze_expand_conditional_branch): emit branch_compare
6102 insn instead of handling cmp op separate from branch insn.
6104 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6106 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
6109 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6111 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
6112 define_insn with define_expand and new define_insn
6113 *altivec_lve<VI_char>x_internal.
6114 (altivec_stve<VI_char>x): Replace define_insn with define_expand
6115 and new define_insn *altivec_stve<VI_char>x_internal.
6116 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
6118 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
6120 (altivec_expand_stvex_be): New function.
6122 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
6124 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
6125 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
6126 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
6127 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
6129 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
6132 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
6133 instead of emit_move_insn.
6135 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6137 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
6139 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
6140 gen_altivec_vsumsws.
6142 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6144 * config/rs6000/altivec.md (altivec_lvxl): Rename as
6145 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
6146 (altivec_lvxl_<mode>): New define_expand incorporating
6147 -maltivec=be semantics where needed.
6148 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
6149 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
6150 semantics where needed.
6151 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
6152 (altivec_stvx_<mode>): New define_expand incorporating
6153 -maltivec=be semantics where needed.
6154 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
6155 VM2 iterator instead of V4SI.
6156 (altivec_stvxl_<mode>): New define_expand incorporating
6157 -maltivec=be semantics where needed.
6158 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
6159 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
6160 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
6161 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
6162 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
6163 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
6164 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
6165 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
6166 ALTIVEC_BUILTIN_STVXL.
6167 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
6168 (altivec_expand_stvx_be): Likewise.
6169 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
6170 (altivec_expand_lvx_be): Likewise.
6171 (altivec_expand_stvx_be): Likewise.
6172 (altivec_expand_builtin): Add cases for
6173 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
6174 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
6175 (altivec_init_builtins): Add definitions for
6176 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
6177 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
6179 2014-02-21 Catherine Moore <clm@codesourcery.com>
6181 * doc/invoke.texi (mvirt, mno-virt): Document.
6182 * config/mips/mips.opt (mvirt): New option.
6183 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
6185 2014-02-21 Richard Biener <rguenther@suse.de>
6187 PR tree-optimization/60276
6188 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
6189 (STMT_VINFO_MIN_NEG_DIST): New macro.
6190 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
6191 STMT_VINFO_MIN_NEG_DIST.
6192 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
6193 made for negative dependence distances still hold.
6195 2014-02-21 Richard Biener <rguenther@suse.de>
6198 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
6199 DECL_INITIAL for globals not in the current function context.
6201 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6203 PR tree-optimization/56490
6204 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
6205 * tree-ssa-uninit.c: Include params.h.
6206 (compute_control_dep_chain): Add num_calls argument, return false
6207 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
6208 num_calls to recursive call.
6209 (find_predicates): Change dep_chain into normal array,
6210 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
6211 variable and adjust compute_control_dep_chain caller.
6212 (find_def_preds): Likewise.
6214 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
6216 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
6217 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
6219 2014-02-21 Nick Clifton <nickc@redhat.com>
6221 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
6222 (pushhi1): Likewise.
6223 (popqi1): Add mode to pre_dec.
6226 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6228 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
6229 mode for mask of V8SFmode permutation.
6231 2014-02-20 Richard Henderson <rth@redhat.com>
6234 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
6235 a new pseudo for OLDVAL.
6237 2014-02-20 Jakub Jelinek <jakub@redhat.com>
6240 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
6241 gen_reg_rtx if d->testing_p.
6242 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
6243 if d->testing_p and we will certainly return true.
6244 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6247 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6249 * emit-rtl.c (gen_reg_rtx): Assert that
6250 crtl->emit.regno_pointer_align_length is non-zero.
6252 2014-02-20 Richard Henderson <rth@redhat.com>
6255 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6256 on failure the store back into EXPECT.
6258 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6259 Sandra Loosemore <sandra@codesourcery.com>
6261 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6262 * config/nios2/nios2.c (nios2_function_profiler): Add
6263 -fPIC (flag_pic == 2) support.
6264 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6265 (nios2_large_offset_p): New function.
6266 (nios2_unspec_reloc_p): Move up position, update to use
6267 nios2_large_offset_p.
6268 (nios2_unspec_address): Remove function.
6269 (nios2_unspec_offset): New function.
6270 (nios2_large_got_address): New function.
6271 (nios2_got_address): Add large offset support.
6272 (nios2_legitimize_tls_address): Update usage of removed and new
6274 (nios2_symbol_binds_local_p): New function.
6275 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
6276 (nios2_legitimize_address): Update to use nios2_large_offset_p.
6277 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
6278 (nios2_print_operand): Merge H/L processing, add hiadj/lo
6279 processing for (const (unspec ...)).
6280 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
6282 2014-02-20 Richard Biener <rguenther@suse.de>
6284 * tree-cfg.c (replace_uses_by): Mark altered BBs before
6285 doing the substitution.
6286 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
6288 2014-02-20 Martin Jambor <mjambor@suse.cz>
6291 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
6292 info when checking whether lattices are bottom.
6294 2014-02-20 Richard Biener <rguenther@suse.de>
6297 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
6300 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
6303 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
6304 parameter specifying the scaling.
6305 (inline_call): Update.
6306 (want_inline_recursively): Guard division by zero.
6307 (recursive_inlining): Update.
6308 * ipa-inline.h (clone_inlined_nodes): Update.
6310 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6313 * config/i386/i386.c (classify_argument): Pass structures of size
6314 64 bytes or less in register.
6316 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6317 Kirill Yukhin <kirill.yukhin@intel.com>
6319 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
6320 (_mm_rcp28_round_ss): Ditto.
6321 (_mm_rsqrt28_round_sd): Ditto.
6322 (_mm_rsqrt28_round_ss): Ditto.
6323 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
6324 (_mm_rcp14_round_ss): Ditto.
6325 (_mm_rsqrt14_round_sd): Ditto.
6326 (_mm_rsqrt14_round_ss): Ditto.
6327 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
6328 the first input operand, get rid of match_dup.
6329 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
6331 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
6333 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
6334 operand as the first input operand, set type attribute.
6335 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
6337 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
6338 operand as the first input operand, set type attribute.
6340 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6342 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
6345 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
6348 * config/i386/i386.c (construct_container): Remove TFmode check
6349 for X86_64_INTEGER_CLASS.
6351 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
6354 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
6355 only when -Wpsabi is enabled.
6357 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
6360 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
6361 passing arrays in registers are the same as for structs, so remove the
6362 special case for them.
6364 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
6366 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
6367 destination type, extract only the valid bits if the source type is not
6368 integral and has a different mode.
6370 2014-02-19 Richard Biener <rguenther@suse.de>
6373 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
6376 2014-02-19 Richard Biener <rguenther@suse.de>
6379 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
6380 (ipa_modify_call_arguments): Emit an argument load explicitely and
6381 preserve virtual SSA form there and for the replacement call.
6382 Do not update SSA form nor free dominance info.
6384 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6386 * ipa.c (function_and_variable_visibility): Also clear WEAK
6387 flag when disolving COMDAT_GROUP.
6389 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6391 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
6392 * ipa-prop.c (ipa_set_jf_known_type): Return early when
6394 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
6395 do more sanity checks.
6396 (detect_type_change): Return true when giving up early.
6397 (compute_complex_assign_jump_func): Fix type parameter of
6398 ipa_set_ancestor_jf.
6399 (compute_complex_ancestor_jump_func): Likewise.
6400 (update_jump_functions_after_inlining): Fix updating of
6402 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
6404 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6406 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
6407 inline clones when edge disappears.
6409 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6412 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
6413 Split 64-bit moves into 2 patterns. Do not allow the use of
6414 direct move for TDmode in little endian, since the decimal value
6415 has little endian bytes within a word, but the 64-bit pieces are
6416 ordered in a big endian fashion, and normal subreg's of TDmode are
6418 (mov<mode>_64bit_dm): Likewise.
6419 (movtd_64bit_nodm): Likewise.
6421 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6423 PR tree-optimization/60174
6424 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
6425 statement of an SSA_NAME that occurs in an abnormal PHI node.
6427 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6430 * final.c (SEEN_BB): Remove.
6431 (SEEN_NOTE, SEEN_EMITTED): Renumber.
6432 (final_scan_insn): Don't force_source_line on second
6433 NOTE_INSN_BASIC_BLOCK.
6435 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
6438 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
6439 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
6440 (type_natural_mode): Warn ABI change when %zmm register is not
6441 available for AVX512F vector value passing.
6443 2014-02-18 Kai Tietz <ktietz@redhat.com>
6446 * config/i386/i386.c (ix86_expand_prologue): Use value in
6447 rax register as displacement when restoring %r10 or %rax.
6448 Fix wrong offset when restoring both registers.
6450 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6452 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
6453 assertion with conditional return.
6455 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6456 Uros Bizjak <ubizjak@gmail.com>
6459 * config/i386/driver-i386.c (host_detect_local_cpu): If
6460 YMM state is not saved by the OS, also clear has_f16c. Move
6461 CPUID 0x80000001 handling before YMM state saving checking.
6463 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
6465 PR rtl-optimization/58960
6466 * haifa-sched.c (alloc_global_sched_pressure_data): New,
6467 factored out from ...
6468 (sched_init): ... here.
6469 (free_global_sched_pressure_data): New, factored out from ...
6470 (sched_finish): ... here.
6471 * sched-int.h (free_global_sched_pressure_data): Declare.
6472 * sched-rgn.c (nr_regions_initial): New static global.
6473 (haifa_find_rgns): Initialize it.
6474 (schedule_region): Disable sched-pressure for the newly
6477 2014-02-17 Richard Biener <rguenther@suse.de>
6479 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
6480 release SSA defs of pattern stmts.
6482 2014-02-17 Richard Biener <rguenther@suse.de>
6484 * tree-inline.c (expand_call_inline): Release the virtual
6485 operand defined by the call we are about to inline.
6487 2014-02-17 Richard Biener <rguenther@suse.de>
6489 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
6491 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
6492 Ilya Tocar <ilya.tocar@intel.com>
6494 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
6495 arguments order in builtin.
6496 (_mm512_permutexvar_epi64): Ditto.
6497 (_mm512_mask_permutexvar_epi64): Ditto
6498 (_mm512_maskz_permutexvar_epi32): Ditto
6499 (_mm512_permutexvar_epi32): Ditto
6500 (_mm512_mask_permutexvar_epi32): Ditto
6502 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6504 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
6505 (p8_vmrgow): Likewise.
6507 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6509 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
6512 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
6515 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
6516 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
6517 into 64-bit and 32-bit moves. On 64-bit moves, add support for
6518 using direct move instructions on ISA 2.07. Also adjust
6519 instruction length for 64-bit.
6520 (mov<mode>_64bit, TFmode/TDmode): Likewise.
6521 (mov<mode>_32bit, TFmode/TDmode): Likewise.
6523 2014-02-15 Alan Modra <amodra@gmail.com>
6527 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
6528 find_replacement on parts of insn rtl that might be reloaded.
6530 2014-02-15 Richard Biener <rguenther@suse.de>
6532 PR tree-optimization/60183
6533 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
6534 (tree_ssa_phiprop): Calculate and free post-dominators.
6536 2014-02-14 Jeff Law <law@redhat.com>
6538 PR rtl-optimization/60131
6539 * ree.c (get_extended_src_reg): New function.
6540 (combine_reaching_defs): Use it rather than assuming location of REG.
6541 (find_and_remove_re): Verify first operand of extension is
6542 a REG before adding the insns to the copy list.
6544 2014-02-14 Roland McGrath <mcgrathr@google.com>
6546 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
6547 * configure: Regenerated.
6548 * config.in: Regenerated.
6549 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
6550 instead of ASM_SHORT.
6552 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
6553 Richard Earnshaw <rearnsha@arm.com>
6555 PR rtl-optimization/59535
6556 * lra-constraints.c (process_alt_operands): Encourage alternative
6557 when unassigned pseudo class is superset of the alternative class.
6558 (inherit_reload_reg): Don't inherit when optimizing for code size.
6559 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
6560 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
6561 modes not less than 4 for Thumb1.
6563 2014-02-14 Kyle McMartin <kyle@redhat.com>
6566 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
6568 2014-02-14 Richard Biener <rguenther@suse.de>
6570 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
6571 (get_frame_arg): Drop the assert with langhook types_compatible_p.
6572 Do not strip INDIRECT_REFs.
6574 2014-02-14 Richard Biener <rguenther@suse.de>
6577 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
6578 DECL_FUNCTION_SPECIFIC_TARGET.
6579 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
6580 * tree-streamer-out.c (pack_ts_target_option): Remove.
6581 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
6582 (write_ts_function_decl_tree_pointers): Do not stream
6583 DECL_FUNCTION_SPECIFIC_TARGET.
6584 * tree-streamer-in.c (unpack_ts_target_option): Remove.
6585 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
6586 (lto_input_ts_function_decl_tree_pointers): Do not stream
6587 DECL_FUNCTION_SPECIFIC_TARGET.
6589 2014-02-14 Jakub Jelinek <jakub@redhat.com>
6591 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
6592 (get_initial_def_for_induction, vectorizable_induction): Ignore
6593 debug stmts when looking for exit_phi.
6594 (vectorizable_live_operation): Fix up condition.
6596 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6598 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
6599 nreverse() because it changes the content of original tree list.
6601 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6603 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
6604 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
6606 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6608 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
6609 GNU coding standards.
6611 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6614 * dwarf2out.c (gen_subprogram_die): Don't call
6615 add_calling_convention_attribute if subr_die is old_die.
6617 2014-02-13 Sharad Singhai <singhai@google.com>
6619 * doc/optinfo.texi: Fix order of nodes.
6621 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
6623 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
6624 operands[2], not operands[3].
6626 2014-02-13 Richard Biener <rguenther@suse.de>
6629 * doc/install.texi (ISL): Update recommended version to 0.12.2,
6630 mention the possibility of an in-tree build.
6631 (CLooG): Update recommended version to 0.18.1, mention the
6632 possibility of an in-tree build and clarify that the ISL
6633 bundled with CLooG does not work.
6635 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6638 * expr.c (compress_float_constant): If x is a hard register,
6639 extend into a pseudo and then move to x.
6641 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
6643 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
6644 caused by bad second argument to warning_at() with -mhotpatch and
6645 nested functions (e.g. with gfortran).
6647 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
6649 * opts.c (option_name): Remove "enabled by default" rider.
6651 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
6653 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
6655 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
6656 Uros Bizjak <ubizjak@gmail.com>
6659 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
6660 * configure: Regenerated.
6662 2014-02-12 Richard Biener <rguenther@suse.de>
6664 * vec.c (vec_prefix::calculate_allocation): Move as
6665 inline variant to vec.h.
6666 (vec_prefix::calculate_allocation_1): New out-of-line version.
6667 * vec.h (vec_prefix::calculate_allocation_1): Declare.
6668 (vec_prefix::m_has_auto_buf): Rename to ...
6669 (vec_prefix::m_using_auto_storage): ... this.
6670 (vec_prefix::calculate_allocation): Inline the easy cases
6671 and dispatch to calculate_allocation_1 which doesn't need the
6673 (va_heap::reserve): Use gcc_checking_assert.
6674 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
6675 m_using_auto_storage.
6676 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
6678 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
6679 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
6680 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
6682 2014-02-12 Richard Biener <rguenther@suse.de>
6684 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
6685 when we found a dependence.
6687 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
6689 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
6691 (maybe_fold_stmt): ... into this new function.
6692 * omp-low.c (lower_omp): Update comment.
6694 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
6697 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
6700 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
6702 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
6703 identifiers in comments.
6704 (cortexa53_extra_costs): Likewise.
6705 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
6706 (cortexa7_extra_costs): Likewise.
6707 (cortexa12_extra_costs): Likewise.
6708 (cortexa15_extra_costs): Likewise.
6709 (v7m_extra_costs): Likewise.
6711 2014-02-12 Richard Biener <rguenther@suse.de>
6714 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
6715 of posix_memalign being successful.
6716 (lower_stmt): Restrict lowering of posix_memalign to when
6717 -ftree-bit-ccp is enabled.
6719 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6721 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
6723 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
6725 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
6727 PR rtl-optimization/60116
6728 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
6729 other_insn once the combination has been validated.
6731 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
6734 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
6736 * ipa-devirt.c: Include demangle.h
6737 (odr_violation_reported): New static variable.
6738 (add_type_duplicate): Update odr_violations.
6739 (maybe_record_node): Add completep parameter; update it.
6740 (record_target_from_binfo): Add COMPLETEP parameter;
6741 update it as needed.
6742 (possible_polymorphic_call_targets_1): Likewise.
6743 (struct polymorphic_call_target_d): Add nonconstruction_targets;
6744 rename FINAL to COMPLETE.
6745 (record_targets_from_bases): Sanity check we found the binfo;
6746 fix COMPLETEP updating.
6747 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
6748 parameter, fix computing of COMPLETEP.
6749 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
6750 at LTO time do demangling.
6751 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
6752 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
6754 (gimple_get_virt_method_for_binfo): Likewise.
6755 * gimple-fold.h (gimple_get_virt_method_for_binfo,
6756 gimple_get_virt_method_for_vtable): Update prototypes.
6758 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
6761 * genautomata.c (add_presence_absence): Fix typo with
6762 {final_}presence_list.
6764 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
6767 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
6768 for VSX/Altivec vectors that land in GPR registers.
6770 2014-02-11 Richard Henderson <rth@redhat.com>
6771 Jakub Jelinek <jakub@redhat.com>
6774 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
6775 around drhs if type conversion to lacc->type is not useless.
6777 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6779 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
6781 (cortex-a57.cortex-a53): Likewise.
6782 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
6784 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6786 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
6789 2014-02-11 Renlin Li <Renlin.Li@arm.com>
6791 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
6792 add_options_for_arm_vfp3.
6794 2014-02-11 Jeff Law <law@redhat.com>
6797 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
6798 object with an undesirable mode.
6800 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6803 * config/i386/sol2-9.h: New file.
6804 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
6805 *-*-solaris2.9*): Use it.
6807 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6809 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
6810 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
6812 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6814 * config/microblaze/microblaze.c: Extend mcpu version format
6816 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
6818 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
6820 2014-02-10 Richard Henderson <rth@redhat.com>
6823 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
6824 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
6825 ms-abi vs -mno-accumulate-outgoing-args.
6826 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
6827 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
6830 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
6833 * cfgexpand.c (expand_asm_operands): Attach source location to
6834 ASM_INPUT rtx objects.
6835 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
6837 2014-02-10 Nick Clifton <nickc@redhat.com>
6839 * config/mn10300/mn10300.c (popcount): New function.
6840 (mn10300_expand_prologue): Include saved registers in stack usage
6843 2014-02-10 Jeff Law <law@redhat.com>
6846 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
6847 when changing the SET_DEST of a prior insn to avoid an input reload.
6849 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6851 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
6852 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
6853 -mcall-openbsd, or -mcall-linux.
6854 (CC1_ENDIAN_BIG_SPEC): Remove.
6855 (CC1_ENDIAN_LITTLE_SPEC): Remove.
6856 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
6857 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
6858 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
6859 and %cc1_endian_default.
6860 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
6862 2014-02-10 Richard Biener <rguenther@suse.de>
6864 PR tree-optimization/60115
6865 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
6866 MEM_REF handling. Properly verify that the accesses are not
6867 out of the objects bound.
6869 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6871 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
6874 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
6876 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
6877 proper constants and fix formatting.
6878 (possible_polymorphic_call_targets): Fix formatting.
6880 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
6881 Ilya Tocar <ilya.tocar@intel.com>
6883 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
6884 (_mm512_loadu_epi32): Renamed into...
6885 (_mm512_loadu_si512): This.
6886 (_mm512_storeu_epi32): Renamed into...
6887 (_mm512_storeu_si512): This.
6888 (_mm512_maskz_ceil_ps): Removed.
6889 (_mm512_maskz_ceil_pd): Ditto.
6890 (_mm512_maskz_floor_ps): Ditto.
6891 (_mm512_maskz_floor_pd): Ditto.
6892 (_mm512_floor_round_ps): Ditto.
6893 (_mm512_floor_round_pd): Ditto.
6894 (_mm512_ceil_round_ps): Ditto.
6895 (_mm512_ceil_round_pd): Ditto.
6896 (_mm512_mask_floor_round_ps): Ditto.
6897 (_mm512_mask_floor_round_pd): Ditto.
6898 (_mm512_mask_ceil_round_ps): Ditto.
6899 (_mm512_mask_ceil_round_pd): Ditto.
6900 (_mm512_maskz_floor_round_ps): Ditto.
6901 (_mm512_maskz_floor_round_pd): Ditto.
6902 (_mm512_maskz_ceil_round_ps): Ditto.
6903 (_mm512_maskz_ceil_round_pd): Ditto.
6904 (_mm512_expand_pd): Ditto.
6905 (_mm512_expand_ps): Ditto.
6906 * config/i386/i386.c (ix86_builtins): Remove
6907 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
6908 (bdesc_args): Ditto.
6909 * config/i386/predicates.md (const1256_operand): New.
6910 (const_1_to_2_operand): Ditto.
6911 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6912 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6913 (*avx512pf_gatherpf<mode>sf): Ditto.
6914 (avx512pf_gatherpf<mode>df): Ditto.
6915 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6916 (*avx512pf_gatherpf<mode>df): Ditto.
6917 (avx512pf_scatterpf<mode>sf): Ditto.
6918 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6919 (*avx512pf_scatterpf<mode>sf): Ditto.
6920 (avx512pf_scatterpf<mode>df): Ditto.
6921 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6922 (*avx512pf_scatterpf<mode>df): Ditto.
6923 (avx512f_expand<mode>): Removed.
6924 (<shift_insn><mode>3<mask_name>): Change predicate type.
6926 2014-02-08 Jakub Jelinek <jakub@redhat.com>
6928 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
6929 not at the end of datarefs vector use ordered_remove to avoid
6930 reordering datarefs vector.
6933 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
6934 mark local addressable non-static vars as GOVD_PRIVATE
6935 instead of GOVD_LOCAL.
6936 * omp-low.c (lower_omp_for): Move gimple_bind_vars
6937 and BLOCK_VARS of gimple_bind_block to new_stmt rather
6941 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
6942 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
6943 assume_aligned or alloc_align attributes.
6944 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
6945 arguments. Handle also assume_aligned and alloc_align attributes.
6946 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
6947 calls to functions with assume_aligned or alloc_align attributes.
6948 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
6950 2014-02-08 Terry Guo <terry.guo@arm.com>
6952 * doc/invoke.texi: Document ARM -march=armv7e-m.
6954 2014-02-08 Jakub Jelinek <jakub@redhat.com>
6956 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
6957 flag on __cilkrts_rethrow builtin.
6960 * ipa-cp.c (determine_versionability): Fail at -O0
6961 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
6962 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
6965 2014-02-04 Jakub Jelinek <jakub@redhat.com>
6968 * tree-inline.c (copy_forbidden): Fail for
6969 __attribute__((optimize (0))) functions.
6971 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
6973 * varpool.c: Include pointer-set.h.
6974 (varpool_remove_unreferenced_decls): Variables in other partitions
6975 will not be output; be however careful to not lose information
6978 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
6980 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
6981 lookup in the vtable constructor.
6983 2014-02-07 Jeff Law <law@redhat.com>
6986 * config/m68k/m68k.md (ashldi_extsi): Turn into a
6987 define_insn_and_split.
6989 * ipa-inline.c (inline_small_functions): Fix typos.
6991 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
6993 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
6994 (s390_can_use_return_insn): Declare.
6995 * config/s390/s390.h (EPILOGUE_USES): Define.
6996 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
6998 (s390_chunkify_start): Handle return JUMP_LABELs.
6999 (s390_early_mach): Emit a main_pool instruction on the entry edge.
7000 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
7001 (s390_can_use_return_insn): New functions.
7002 (s390_fix_long_loop_prediction): Handle conditional returns.
7003 (TARGET_SET_UP_BY_PROLOGUE): Define.
7004 * config/s390/s390.md (ANY_RETURN): New code iterator.
7005 (*creturn, *csimple_return, return, simple_return): New patterns.
7007 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7009 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
7010 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
7011 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
7012 REG_CFA_RESTORE list when deciding not to restore a register.
7014 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7016 * config/s390/s390.c: Include tree-pass.h and context.h.
7017 (s390_early_mach): New function, split out from...
7018 (s390_emit_prologue): ...here.
7019 (pass_data_s390_early_mach): New pass structure.
7020 (pass_s390_early_mach): New class.
7021 (s390_option_override): Create and register early_mach pass.
7022 Move to end of file.
7024 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7026 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
7027 to match for the exit block.
7029 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7031 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
7032 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
7033 Reject misaligned operands.
7035 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7037 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
7039 2014-02-07 Richard Biener <rguenther@suse.de>
7042 * gimple-low.c (lower_builtin_posix_memalign): New function.
7043 (lower_stmt): Call it to lower posix_memalign in a way
7044 to make alignment info accessible.
7046 2014-02-07 Jakub Jelinek <jakub@redhat.com>
7049 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
7050 __builtin_setjmp_receiver.
7052 2014-02-07 Richard Biener <rguenther@suse.de>
7055 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
7056 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
7057 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7058 Handle BUILT_IN_POSIX_MEMALIGN.
7059 (find_func_clobbers): Likewise.
7060 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
7061 (call_may_clobber_ref_p_1): Likewise.
7063 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7066 * ipa-devirt.c (record_target_from_binfo): Remove overactive
7069 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7072 * lto-cgraph.c (lto_output_node): Use
7073 symtab_get_symbol_partitioning_class.
7074 (lto_output_varpool_node): likewise.
7075 (symtab_get_symbol_partitioning_class): Move here from
7077 * cgraph.h (symbol_partitioning_class): Likewise.
7078 (symtab_get_symbol_partitioning_class): Declare.
7080 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7082 * ggc.h (ggc_internal_cleared_alloc): New macro.
7083 * vec.h (vec_safe_copy): Handle memory stats.
7084 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
7085 * target-globals.c (save_target_globals): Likewise.
7087 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7090 * expr.c (emit_move_resolve_push): Export; be bit more selective
7091 on when to clear alias set.
7092 * expr.h (emit_move_resolve_push): Declare.
7093 * function.h (struct function): Add tail_call_marked.
7094 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
7095 * config/i386/i386-protos.h (ix86_expand_push): Remove.
7096 * config/i386/i386.md (TImode move expander): De not call
7098 (FP push expanders): Preserve memory attributes.
7099 * config/i386/sse.md (push<mode>1): Remove.
7100 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
7101 (ix86_expand_push): Remove.
7102 * config/i386/mmx.md (push<mode>1): Remove.
7104 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7106 PR rtl-optimization/60030
7107 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
7108 lopart with paradoxical subreg before shifting it up by hprec.
7110 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7112 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
7113 Remove extra newline at end of file.
7114 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
7115 (arm_issue_rate): Handle cortexa57.
7116 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
7117 (cortex-a57.cortex-a53): Likewise.
7119 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7122 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
7123 don't record in REG_FRAME_RELATED_EXPR registers not set in that
7125 (arm_expand_prologue): Adjust all callers.
7126 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
7127 info, registers also at the lowest numbered registers side. Use
7128 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
7132 * var-tracking.c (adjust_mems): Before adding a SET to
7133 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
7135 2014-02-06 Alan Modra <amodra@gmail.com>
7138 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
7139 change SDmode to DDmode when lra_in_progress.
7141 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7144 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
7145 free_data_ref on the dr first, and before goto again also set dr
7146 to the next dr. For simd_lane_access, free old datarefs[i] before
7147 overwriting it. For get_vectype_for_scalar_type failure, don't
7148 free_data_ref if simd_lane_access.
7150 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
7153 * tree.h (opts_for_fn): New inline function.
7154 (opt_for_fn): Define.
7155 * config/i386/i386.c (ix86_function_regparm): Use
7156 opt_for_fn (decl, optimize) instead of optimize.
7158 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
7160 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
7161 for SYMBOL_REF in large memory model.
7163 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7165 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
7167 (cortex-a57): Likewise.
7168 (cortex-a57.cortex-a53): Likewise.
7170 2014-02-06 Yury Gribov <y.gribov@samsung.com>
7171 Kugan Vivekanandarajah <kuganv@linaro.org>
7173 * config/arm/arm.c (arm_vector_alignment_reachable): Check
7175 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
7177 2014-02-06 Richard Biener <rguenther@suse.de>
7179 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
7180 set_loop_copy and initialize_original_copy_tables.
7182 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
7184 * config/aarch64/aarch64-simd.md
7185 (aarch64_ashr_simddi): Change QI to SI.
7187 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7188 Jakub Jelinek <jakub@redhat.com>
7191 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
7194 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7196 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
7197 CODE_FOR_altivec_vpku[hw]um to
7198 CODE_FOR_altivec_vpku[hw]um_direct.
7199 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
7200 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
7201 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
7202 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
7204 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7206 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
7207 generation for -maltivec=be.
7208 (altivec_vsumsws): Simplify redundant test.
7210 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7212 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
7213 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
7214 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
7215 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
7216 gen_altivec_vpkuwum.
7217 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
7219 (altivec_vpks<VI_char>ss): Likewise.
7220 (altivec_vpks<VI_char>us): Likewise.
7221 (altivec_vpku<VI_char>us): Likewise.
7222 (altivec_vpku<VI_char>um): Likewise.
7223 (altivec_vpku<VI_char>um_direct): New (copy of
7224 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
7226 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
7227 target is little endian and -maltivec=be is not specified.
7228 (*altivec_vupkhs<VU_char>_direct): New (copy of
7229 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
7230 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
7231 target is little endian and -maltivec=be is not specified.
7232 (*altivec_vupkls<VU_char>_direct): New (copy of
7233 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
7234 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
7235 little endian and -maltivec=be is not specified.
7236 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
7237 little endian and -maltivec=be is not specified.
7239 2014-02-05 Richard Henderson <rth@redhat.com>
7242 * combine-stack-adj.c: Revert r206943.
7243 * sched-int.h (struct deps_desc): Add last_args_size.
7244 * sched-deps.c (init_deps): Initialize it.
7245 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7246 contain REG_ARGS_SIZE notes.
7248 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7250 * lto-cgraph.c (asm_nodes_output): Make global.
7251 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7252 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7253 (driver_handle_option): Handle OPT_fwpa.
7255 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7258 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7259 a comment typo and formatting issue. If odr_hash hasn't been
7260 created, return vNULL and set *completep to false.
7263 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7264 bb with no successors.
7266 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7269 * doc/invoke.texi (-march): Clarify documentation for ARM.
7273 2014-02-05 Richard Biener <rguenther@suse.de>
7275 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
7276 when not vectorizing because of too many alias checks.
7277 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
7278 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
7280 2014-02-05 Nick Clifton <nickc@redhat.com>
7282 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
7283 accept extended registers in any mode when compiling for the MN10300.
7285 2014-02-05 Yury Gribov <y.gribov@samsung.com>
7287 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
7288 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
7289 sanitization attributes.
7290 (can_inline_edge_p): Likewise.
7291 (sanitize_attrs_match_for_inline_p): New function.
7293 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7295 * ipa-prop.c (detect_type_change): Shor circuit testing of
7296 type changes on THIS pointer.
7298 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
7301 * config/pa/pa.c (legitimize_tls_address): Return original address
7302 if not passed a SYMBOL_REF rtx.
7303 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
7305 (pa_emit_move_sequence): Simplify TLS source operands.
7306 (pa_legitimate_constant_p): Reject all TLS constants.
7307 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
7308 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
7310 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7312 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
7313 groups when we know they are controlled by LTO.
7314 * varasm.c (default_binds_local_p_1): If object is in other partition,
7315 it will be resolved locally.
7317 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7319 * config/host-linux.c (linux_gt_pch_use_address): Don't
7320 use SSIZE_MAX because it is not always defined.
7322 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
7325 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
7326 threshold for pseudo splitting.
7327 (update_ebb_live_info): Process call argument hard registers and
7328 hard registers from insn definition too.
7329 (max_small_class_regs_num): New constant.
7330 (inherit_in_ebb): Update live hard regs through EBBs. Update
7331 reloads_num only for small register classes. Don't split for
7334 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
7337 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
7340 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7342 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
7345 2014-02-04 Marek Polacek <polacek@redhat.com>
7347 * gdbinit.in (pel): Define.
7349 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7351 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
7354 2014-02-04 Richard Biener <rguenther@suse.de>
7357 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
7358 in function context local.
7359 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
7360 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
7361 similar to LTO_imported_decl_ref.
7363 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7365 PR tree-optimization/60002
7366 * cgraphclones.c (build_function_decl_skip_args): Clear
7369 PR tree-optimization/60023
7370 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
7371 false to gsi_replace.
7372 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
7373 has been in some EH region and vec_stmt could throw, add
7374 vec_stmt into the same EH region.
7375 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
7376 has no lhs, ignore it.
7377 * internal-fn.c (expand_MASK_LOAD): Likewise.
7380 * tree-inline.c (copy_forbidden): Fail for
7381 __attribute__((optimize (0))) functions.
7384 * omp-low.c (simd_clone_struct_copy): If from->inbranch
7385 is set, copy one less argument.
7386 (expand_simd_clones): Don't subtract clone_info->inbranch
7387 from simd_clone_struct_alloc argument.
7389 PR rtl-optimization/57915
7390 * recog.c (simplify_while_replacing): If all unary/binary/relational
7391 operation arguments are constant, attempt to simplify those.
7394 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
7395 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
7397 2014-02-04 Richard Biener <rguenther@suse.de>
7399 PR tree-optimization/60012
7400 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
7401 TBAA disambiguation to all DDRs.
7403 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7406 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
7407 (LINK_SPEC): Use it for -shared, -shared-libgcc.
7409 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7412 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
7414 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7416 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
7417 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
7419 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7422 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
7423 to figure out targets of polymorphic calls with known decl.
7424 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7425 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
7426 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
7427 (get_polymorphic_call_info): ... here.
7428 (get_polymorphic_call_info_from_invariant): New function.
7430 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7432 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
7433 lookup via vtable pointer; check for type consistency
7434 and turn inconsitent facts into UNREACHABLE.
7435 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7436 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
7437 type inconsistent querries; return UNREACHABLE instead.
7439 2014-02-03 Richard Henderson <rth@twiddle.net>
7442 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
7443 already processed this node.
7444 (normalize_one_pred_1): Pass along mark_set.
7445 (normalize_one_pred): Create and destroy a pointer_set_t.
7446 (normalize_one_pred_chain): Likewise.
7448 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
7450 PR gcov-profile/58602
7451 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
7453 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7456 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
7457 -fno-devirtualize; try to devirtualize by the knowledge of
7458 virtual table pointer given by aggregate propagation.
7459 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7460 (ipa_print_node_jump_functions): Dump also offset that
7461 is relevant for polymorphic calls.
7462 (determine_known_aggregate_parts): Add arg_type parameter; use it
7463 instead of determining the type from pointer type.
7464 (ipa_compute_jump_functions_for_edge): Update call of
7465 determine_known_aggregate_parts.
7466 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
7467 (gimple_get_virt_method_for_binfo): ... here; simplify using
7468 vtable_pointer_value_to_vtable.
7469 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
7470 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
7471 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
7472 (vtable_pointer_value_to_vtable): Break out from ...; handle also
7474 (vtable_pointer_value_to_binfo): ... here.
7475 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
7477 2014-02-03 Teresa Johnson <tejohnson@google.com>
7479 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
7480 redef of outer loop index variable.
7482 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
7486 * doc/extend.texi (Function Attributes): Typo.
7488 2014-02-03 Cong Hou <congh@google.com>
7490 PR tree-optimization/60000
7491 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
7492 if the vectorized statement is a store. A store statement can only
7493 appear at the end of pattern statements.
7495 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7497 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
7498 (ix86_option_override_internal): Default long double to 64-bit for
7499 32-bit Bionic and to 128-bit for 64-bit Bionic.
7501 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
7502 TARGET_LONG_DOUBLE_128 is true.
7503 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
7505 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
7506 (mlong-double-64): Negate -mlong-double-128.
7507 (mlong-double-128): New option.
7509 * config/i386/i386-c.c (ix86_target_macros): Define
7510 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
7512 * doc/invoke.texi: Document -mlong-double-128.
7514 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7516 PR rtl-optimization/60024
7517 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
7519 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
7521 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
7523 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7525 PR rtl-optimization/57662
7526 * sel-sched.c (code_motion_path_driver): Do not mark already not
7527 existing blocks in the visiting bitmap.
7529 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7531 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
7532 on the insn being emitted.
7534 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
7535 Will Deacon <will.deacon@arm.com>
7537 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
7539 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7541 * config/arm/arm-tables.opt: Regenerate.
7543 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7545 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
7546 for vector types other than V16QImode.
7547 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
7548 define_expand, and call altivec_expand_vec_perm_le when producing
7549 code with little endian element order.
7550 (*altivec_vperm_<mode>_internal): New insn having previous
7551 behavior of altivec_vperm_<mode>.
7552 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
7553 altivec_expand_vec_perm_le when producing code with little endian
7555 (*altivec_vperm_<mode>_uns_internal): New insn having previous
7556 behavior of altivec_vperm_<mode>_uns.
7558 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7560 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
7561 (altivec_vsumsws): Add handling for -maltivec=be with a little
7563 (altivec_vsumsws_direct): New.
7564 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
7565 gen_altivec_vsumsws.
7567 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7569 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
7570 vtable_pointer_value_to_binfo): New functions.
7571 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
7572 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
7574 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
7576 * config/nios2/nios2.md (load_got_register): Initialize GOT
7577 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
7578 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
7580 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7582 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
7583 preserverd by passthrough, do not propagate the type.
7585 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7587 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
7588 (mips_atomic_assign_expand_fenv): New function.
7589 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
7591 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7593 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
7594 (__builtin_mips_set_fcsr): Likewise.
7595 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
7596 MIPS_USI_FTYPE_VOID.
7597 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
7598 (mips16_expand_set_fcsr): Likewise.
7599 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
7600 (mips16_set_fcsr_stub): Likewise.
7601 (mips16_get_fcsr_one_only_stub): New class.
7602 (mips16_set_fcsr_one_only_stub): Likewise.
7603 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
7604 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
7605 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
7606 (hard_float): New availability predicate.
7607 (mips_builtins): Add get_fcsr and set_fcsr.
7608 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
7609 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
7610 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
7611 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
7612 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
7615 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7617 * config/mips/mips.c (mips_one_only_stub): New class.
7618 (mips_need_mips16_rdhwr_p): Replace with...
7619 (mips16_rdhwr_stub): ...this new variable.
7620 (mips16_stub_call_address): New function.
7621 (mips16_rdhwr_one_only_stub): New class.
7622 (mips_expand_thread_pointer): Use mips16_stub_call_address.
7623 (mips_output_mips16_rdhwr): Delete.
7624 (mips_finish_stub): New function.
7625 (mips_code_end): Use it to handle rdhwr stubs.
7627 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
7630 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
7631 when calculating size of integer atomic types.
7633 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
7635 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
7637 2014-02-01 Jakub Jelinek <jakub@redhat.com>
7639 PR tree-optimization/60003
7640 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
7641 * profile.c (branch_prob): Use gimple_call_builtin_p
7642 to check for BUILT_IN_SETJMP_RECEIVER.
7643 * tree-inline.c (copy_bb): Call notice_special_calls.
7645 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
7648 * lra-constraints.c (process_alt_operands): Update reload_sum only
7651 2014-01-31 Richard Henderson <rth@redhat.com>
7654 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
7655 until after else_eh is processed.
7657 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7659 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
7660 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
7661 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
7662 in smmintrin.h, remove them.
7663 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
7664 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
7665 * config/i386/i386.md (ROUND_SAE): Fix value.
7666 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
7667 (const48_operand): New.
7668 * config/i386/subst.md (round), (round_expand): Use
7669 const_4_or_8_to_11_operand.
7670 (round_saeonly), (round_saeonly_expand): Use const48_operand.
7672 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7674 * config/i386/constraints.md (Yk): Swap meaning with k.
7675 * config/i386/i386.md (movhi_internal): Change Yk to k.
7676 (movqi_internal): Ditto.
7677 (*k<logic><mode>): Ditto.
7680 (kandn<mode>): Ditto.
7681 (*<code>hi_1): Ditto.
7682 (*<code>qi_1): Ditto.
7683 (kxnor<mode>): Ditto.
7684 (kortestzhi): Ditto.
7685 (kortestchi): Ditto.
7687 (*one_cmplhi2_1): Ditto.
7688 (*one_cmplqi2_1): Ditto.
7689 * config/i386/sse.md (): Change k to Yk.
7690 (avx512f_load<mode>_mask): Ditto.
7691 (avx512f_blendm<mode>): Ditto.
7692 (avx512f_store<mode>_mask): Ditto.
7693 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
7694 (avx512f_storedqu<mode>_mask): Ditto.
7695 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
7697 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
7698 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
7699 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
7700 (avx512f_maskcmp<mode>3): Ditto.
7701 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
7702 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
7703 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
7704 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
7705 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
7706 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
7707 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
7708 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
7709 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
7710 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
7711 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
7712 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
7713 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
7714 (vec_extract_lo_<mode>_maskm): Ditto.
7715 (vec_extract_hi_<mode>_maskm): Ditto.
7716 (avx512f_vternlog<mode>_mask): Ditto.
7717 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
7718 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
7719 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
7720 (avx512f_<code>v8div16qi2_mask): Ditto.
7721 (avx512f_<code>v8div16qi2_mask_store): Ditto.
7722 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
7723 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
7724 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
7725 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
7726 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7727 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7728 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7729 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7730 (avx512cd_maskb_vec_dupv8di): Ditto.
7731 (avx512cd_maskw_vec_dupv16si): Ditto.
7732 (avx512f_vpermi2var<mode>3_maskz): Ditto.
7733 (avx512f_vpermi2var<mode>3_mask): Ditto.
7734 (avx512f_vpermi2var<mode>3_mask): Ditto.
7735 (avx512f_vpermt2var<mode>3_maskz): Ditto.
7736 (*avx512f_gathersi<mode>): Ditto.
7737 (*avx512f_gathersi<mode>_2): Ditto.
7738 (*avx512f_gatherdi<mode>): Ditto.
7739 (*avx512f_gatherdi<mode>_2): Ditto.
7740 (*avx512f_scattersi<mode>): Ditto.
7741 (*avx512f_scatterdi<mode>): Ditto.
7742 (avx512f_compress<mode>_mask): Ditto.
7743 (avx512f_compressstore<mode>_mask): Ditto.
7744 (avx512f_expand<mode>_mask): Ditto.
7745 * config/i386/subst.md (mask): Change k to Yk.
7746 (mask_scalar_merge): Ditto.
7749 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
7751 * doc/extend.texi (Vector Extensions): Document ?: in C++.
7753 2014-01-31 Richard Biener <rguenther@suse.de>
7756 * builtins.c (fold_builtin_memory_op): Make sure to not
7757 use a floating-point mode or a boolean or enumeral type for
7760 2014-01-30 DJ Delorie <dj@redhat.com>
7762 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
7763 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
7764 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
7765 whenever main() has an epilogue.
7767 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7769 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
7770 unused variable "field".
7771 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
7772 (vsx_mergeh_<mode>): Likewise.
7773 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
7774 (altivec_vmrghh): Likewise.
7775 (altivec_vmrghw): Likewise.
7776 (altivec_vmrglb): Likewise.
7777 (altivec_vmrglh): Likewise.
7778 (altivec_vmrglw): Likewise.
7779 (altivec_vspltb): Add missing uses.
7780 (altivec_vsplth): Likewise.
7781 (altivec_vspltw): Likewise.
7782 (altivec_vspltsf): Likewise.
7784 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7787 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
7788 frame related instructions.
7790 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
7792 PR rtl-optimization/59959
7793 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
7794 any reload of register whose subreg is invalid.
7796 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7798 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
7799 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
7800 Add missing return type - void.
7802 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7804 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
7805 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
7806 remove element index adjustment for endian (now handled in vsx.md
7808 (altivec_expand_vec_perm_const): Use
7809 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
7810 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
7811 (vsx_xxspltw_<mode>): Adjust element index for little endian.
7812 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
7813 define_expand and a new define_insn *altivec_vspltb_internal;
7814 adjust for -maltivec=be on a little endian target.
7815 (altivec_vspltb_direct): New.
7816 (altivec_vsplth): Divide into a define_expand and a new
7817 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
7818 little endian target.
7819 (altivec_vsplth_direct): New.
7820 (altivec_vspltw): Divide into a define_expand and a new
7821 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
7822 little endian target.
7823 (altivec_vspltw_direct): New.
7824 (altivec_vspltsf): Divide into a define_expand and a new
7825 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
7826 a little endian target.
7828 2014-01-30 Richard Biener <rguenther@suse.de>
7830 PR tree-optimization/59993
7831 * tree-ssa-forwprop.c (associate_pointerplus): Check we
7832 can propagate form the earlier stmt and avoid the transform
7833 when the intermediate result is needed.
7835 2014-01-30 Alangi Derick <alangiderick@gmail.com>
7837 * README.Portability: Fix typo.
7839 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
7841 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
7842 comparison_operator with ordered_comparison_operator.
7844 2014-01-30 Nick Clifton <nickc@redhat.com>
7846 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
7847 Rename to mn10300_store_multiple_regs.
7848 * config/mn10300/mn10300.c: Likewise.
7849 * config/mn10300/mn10300.md (store_movm): Fix typo: call
7850 store_multiple_regs.
7851 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
7852 Call mn10300_store_multiple_regs.
7854 2014-01-30 Nick Clifton <nickc@redhat.com>
7855 DJ Delorie <dj@redhat.com>
7857 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
7858 %fp 2 to keep registers after it properly word-aligned.
7859 (rl78_alloc_physical_registers_umul): Handle the case where both
7860 input operands are the same.
7862 2014-01-30 Richard Biener <rguenther@suse.de>
7864 PR tree-optimization/59903
7865 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
7868 2014-01-30 Jason Merrill <jason@redhat.com>
7871 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
7874 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
7876 2014-01-30 Richard Biener <rguenther@suse.de>
7878 PR tree-optimization/59951
7879 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
7881 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
7884 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
7885 SFmode to DFmode case.
7887 2014-01-29 DJ Delorie <dj@redhat.com>
7889 * config/msp430/msp430.opt (-minrt): New.
7890 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
7892 (ENDFILE_SPEC): Likewise.
7894 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
7896 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
7897 (estimate_function_body_sizes): Use it.
7899 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
7902 * dwarf2out.c (is_cxx_auto): New.
7903 (is_base_type): Use it.
7904 (gen_type_die_with_usage): Likewise.
7906 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7908 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
7909 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
7910 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
7911 -maltivec=be with LE targets.
7912 (vsx_mergeh_<mode>): Likewise.
7913 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
7914 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
7915 (altivec_vmrghb): Replace with define_expand and new
7916 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
7917 (altivec_vmrghb_direct): New define_insn.
7918 (altivec_vmrghh): Replace with define_expand and new
7919 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
7920 (altivec_vmrghh_direct): New define_insn.
7921 (altivec_vmrghw): Replace with define_expand and new
7922 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
7923 (altivec_vmrghw_direct): New define_insn.
7924 (*altivec_vmrghsf): Adjust for endianness.
7925 (altivec_vmrglb): Replace with define_expand and new
7926 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
7927 (altivec_vmrglb_direct): New define_insn.
7928 (altivec_vmrglh): Replace with define_expand and new
7929 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
7930 (altivec_vmrglh_direct): New define_insn.
7931 (altivec_vmrglw): Replace with define_expand and new
7932 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
7933 (altivec_vmrglw_direct): New define_insn.
7934 (*altivec_vmrglsf): Adjust for endianness.
7935 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
7936 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
7937 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
7938 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
7939 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
7940 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
7941 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
7942 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
7944 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
7946 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
7947 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
7950 2014-01-29 Richard Biener <rguenther@suse.de>
7952 PR tree-optimization/58742
7953 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
7954 associate_pointerplus_align.
7955 (associate_pointerplus_diff): New function.
7956 (associate_pointerplus): Likewise. Call associate_pointerplus_align
7957 and associate_pointerplus_diff.
7959 2014-01-29 Richard Biener <rguenther@suse.de>
7961 * lto-streamer.h (LTO_major_version): Bump to 3.
7962 (LTO_minor_version): Reset to 0.
7964 2014-01-29 Renlin Li <Renlin.Li@arm.com>
7966 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
7967 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
7968 (arm_file_start): Generate correct asm header for armv7ve.
7969 * config/arm/bpabi.h: Add multilib support for armv7ve.
7970 * config/arm/driver-arm.c: Change the architectures of cortex-a7
7971 and cortex-a15 to armv7ve.
7972 * config/arm/t-aprofile: Add multilib support for armv7ve.
7973 * doc/invoke.texi: Document -march=armv7ve.
7975 2014-01-29 Richard Biener <rguenther@suse.de>
7977 PR tree-optimization/58742
7978 * tree-ssa-forwprop.c (associate_plusminus): Return true
7979 if we changed sth, defer EH cleanup to ...
7980 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
7981 (simplify_mult): New function.
7983 2014-01-29 Jakub Jelinek <jakub@redhat.com>
7986 PR tree-optimization/59920
7987 * tree.c (build_common_builtin_nodes): Remove
7988 __builtin_setjmp_dispatcher initialization.
7989 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
7990 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
7991 instead of gsi_after_labels + manually skipping debug stmts.
7992 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
7993 ignore bbs with IFN_ABNORMAL_DISPATCHER.
7994 * tree-inline.c (copy_edges_for_bb): Remove
7995 can_make_abnormal_goto argument, instead add abnormal_goto_dest
7996 argument. Ignore computed_goto_p stmts. Don't call
7997 make_abnormal_goto_edges. If a call might need abnormal edges
7998 for non-local gotos, see if it already has an edge to
7999 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
8000 with true argument, don't do anything then, otherwise add
8001 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
8002 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
8004 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
8005 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
8006 (lower_stmt): Don't set data->calls_builtin_setjmp.
8007 (lower_builtin_setjmp): Adjust comment.
8008 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
8009 * tree-cfg.c (found_computed_goto): Remove.
8010 (factor_computed_gotos): Remove.
8011 (make_goto_expr_edges): Return bool, true for computed gotos.
8012 Don't call make_abnormal_goto_edges.
8013 (build_gimple_cfg): Don't set found_computed_goto, don't call
8014 factor_computed_gotos.
8015 (computed_goto_p): No longer static.
8016 (make_blocks): Don't set found_computed_goto.
8017 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
8018 (make_edges): If make_goto_expr_edges returns true, push bb
8019 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
8020 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
8021 vector. Record mapping between bbs and OpenMP regions if there
8022 are any, adjust make_gimple_omp_edges caller. Call
8023 handle_abnormal_edges.
8024 (make_abnormal_goto_edges): Remove.
8025 * tree-cfg.h (make_abnormal_goto_edges): Remove.
8026 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
8027 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
8028 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
8029 * internal-fn.def (ABNORMAL_DISPATCHER): New.
8030 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
8031 filling *region also set *region_idx to (*region)->entry->index.
8034 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
8035 For REGs set ORIGINAL_REGNO.
8037 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
8039 * doc/md.texi: Mention that a target shouldn't implement
8040 vec_widen_(s|u)mul_even/odd pair if it is less efficient
8043 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8045 PR tree-optimization/59594
8046 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
8047 a copy of the datarefs vector rather than the vector itself.
8049 2014-01-28 Jason Merrill <jason@redhat.com>
8052 * dwarf2out.c (auto_die): New static.
8053 (gen_type_die_with_usage): Handle C++1y 'auto'.
8054 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
8057 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
8060 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
8061 (SPEC_X32): Likewise.
8062 (SPEC_64): Likewise.
8063 * config/i386/i386.c (ix86_option_override_internal): Turn off
8064 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
8066 (x86_file_start): Output .code16gcc for TARGET_16BIT.
8067 * config/i386/i386.h (TARGET_16BIT): New macro.
8068 (TARGET_16BIT_P): Likewise.
8069 * config/i386/i386.opt: Add m16.
8070 * doc/invoke.texi: Document -m16.
8072 2014-01-28 Jakub Jelinek <jakub@redhat.com>
8074 PR preprocessor/59935
8075 * input.c (location_get_source_line): Bail out on when line number
8076 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
8078 2014-01-28 Richard Biener <rguenther@suse.de>
8080 PR tree-optimization/58742
8081 * tree-ssa-forwprop.c (associate_plusminus): Handle
8082 pointer subtraction of the form (T)(P + A) - (T)P.
8084 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8086 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
8089 2014-01-28 Richard Biener <rguenther@suse.de>
8092 2014-01-28 Richard Biener <rguenther@suse.de>
8094 PR rtl-optimization/45364
8095 PR rtl-optimization/59890
8096 * var-tracking.c (local_get_addr_clear_given_value): Handle
8097 already cleared slot.
8098 (val_reset): Handle not allocated local_get_addr_cache.
8099 (vt_find_locations): Use post-order on the inverted CFG.
8101 2014-01-28 Richard Biener <rguenther@suse.de>
8103 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
8105 2014-01-28 Richard Biener <rguenther@suse.de>
8107 PR rtl-optimization/45364
8108 PR rtl-optimization/59890
8109 * var-tracking.c (local_get_addr_clear_given_value): Handle
8110 already cleared slot.
8111 (val_reset): Handle not allocated local_get_addr_cache.
8112 (vt_find_locations): Use post-order on the inverted CFG.
8114 2014-01-28 Alan Modra <amodra@gmail.com>
8116 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
8117 * configure.ac <recursive call for build != host>: Define
8118 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
8119 and LD_FOR_BUILD too.
8120 * configure: Regenerate.
8122 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
8124 * config/i386/i386.c (get_builtin_code_for_version): Separate
8125 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
8126 Broadwell from Haswell.
8128 2014-01-27 Steve Ellcey <sellcey@mips.com>
8130 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
8131 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
8132 * config/mips/mips.c (mips_option_override): Change setting
8134 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
8135 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
8136 Change from Mask to Var.
8138 2014-01-27 Jeff Law <law@redhat.com>
8140 * ipa-inline.c (inline_small_functions): Fix typo.
8142 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
8144 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
8145 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
8146 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
8147 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
8148 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
8149 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
8150 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
8151 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
8152 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
8153 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
8154 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
8155 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
8156 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
8157 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
8158 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
8159 (_mm512_storeu_epi64): Ditto.
8160 (_mm512_cmpge_epi32_mask): Ditto.
8161 (_mm512_cmpge_epu32_mask): Ditto.
8162 (_mm512_cmpge_epi64_mask): Ditto.
8163 (_mm512_cmpge_epu64_mask): Ditto.
8164 (_mm512_cmple_epi32_mask): Ditto.
8165 (_mm512_cmple_epu32_mask): Ditto.
8166 (_mm512_cmple_epi64_mask): Ditto.
8167 (_mm512_cmple_epu64_mask): Ditto.
8168 (_mm512_cmplt_epi32_mask): Ditto.
8169 (_mm512_cmplt_epu32_mask): Ditto.
8170 (_mm512_cmplt_epi64_mask): Ditto.
8171 (_mm512_cmplt_epu64_mask): Ditto.
8172 (_mm512_cmpneq_epi32_mask): Ditto.
8173 (_mm512_cmpneq_epu32_mask): Ditto.
8174 (_mm512_cmpneq_epi64_mask): Ditto.
8175 (_mm512_cmpneq_epu64_mask): Ditto.
8176 (_mm512_expand_pd): Ditto.
8177 (_mm512_expand_ps): Ditto.
8178 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
8179 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
8180 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
8181 * config/i386/i386.c (ix86_builtins): Add
8182 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
8183 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
8184 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
8185 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
8186 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
8187 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
8188 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
8189 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
8190 IX86_BUILTIN_PMOVUSQW512_MEM.
8191 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
8192 __builtin_ia32_pmovsqd512mem_mask,
8193 __builtin_ia32_pmovqd512mem_mask,
8194 __builtin_ia32_pmovusqw512mem_mask,
8195 __builtin_ia32_pmovsqw512mem_mask,
8196 __builtin_ia32_pmovqw512mem_mask,
8197 __builtin_ia32_pmovusdw512mem_mask,
8198 __builtin_ia32_pmovsdw512mem_mask,
8199 __builtin_ia32_pmovdw512mem_mask,
8200 __builtin_ia32_pmovqb512mem_mask,
8201 __builtin_ia32_pmovusqb512mem_mask,
8202 __builtin_ia32_pmovsqb512mem_mask,
8203 __builtin_ia32_pmovusdb512mem_mask,
8204 __builtin_ia32_pmovsdb512mem_mask,
8205 __builtin_ia32_pmovdb512mem_mask.
8206 (bdesc_args): Add __builtin_ia32_expanddf512,
8207 __builtin_ia32_expandsf512.
8208 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
8209 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
8210 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
8211 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
8212 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
8213 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
8214 (avx512f_<code>v8div16qi2_mask_store): This.
8215 (avx512f_expand<mode>): New.
8217 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
8219 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
8221 (_mm512_mask_prefetch_i64gather_pd): Ditto.
8222 (_mm512_prefetch_i32scatter_pd): Ditto.
8223 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
8224 (_mm512_prefetch_i64scatter_pd): Ditto.
8225 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
8226 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
8227 (_mm512_mask_prefetch_i64gather_ps): Ditto.
8228 (_mm512_prefetch_i32scatter_ps): Ditto.
8229 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
8230 (_mm512_prefetch_i64scatter_ps): Ditto.
8231 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
8232 * config/i386/i386-builtin-types.def: Define
8233 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
8234 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
8235 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
8236 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
8237 IX86_BUILTIN_SCATTERPFQPD.
8238 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
8239 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
8240 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
8241 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
8242 __builtin_ia32_scatterpfqps.
8243 (ix86_expand_builtin): Expand new built-ins.
8244 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8245 fix memory access data type.
8246 (*avx512pf_gatherpf<mode>_mask): Ditto.
8247 (*avx512pf_gatherpf<mode>): Ditto.
8248 (avx512pf_scatterpf<mode>): Ditto.
8249 (*avx512pf_scatterpf<mode>_mask): Ditto.
8250 (*avx512pf_scatterpf<mode>): Ditto.
8251 (GATHER_SCATTER_SF_MEM_MODE): New.
8252 (avx512pf_gatherpf<mode>df): Ditto.
8253 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8254 (*avx512pf_scatterpf<mode>df): Ditto.
8256 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8259 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8260 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8263 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8265 * common/config/arm/arm-common.c
8266 (arm_rewrite_mcpu): Handle multiple names.
8268 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8270 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8272 * gimple-builder.h (create_gimple_tmp): Delete.
8274 2014-01-27 Christian Bruel <christian.bruel@st.com>
8276 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
8279 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
8281 * config/pa/pa.md (call): Generate indirect long calls to non-local
8282 functions when outputing 32-bit code.
8283 (call_value): Likewise except for special call to buggy powf function.
8285 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
8286 portable runtime and PIC indirect calls.
8287 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
8288 and PIC call sequences. Use ldo instead of blr to set return register
8289 in PIC call sequence.
8291 2014-01-25 Walter Lee <walt@tilera.com>
8293 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
8294 avoid clobbering a live register.
8296 2014-01-25 Walter Lee <walt@tilera.com>
8298 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
8299 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
8300 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
8301 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
8303 2014-01-25 Walter Lee <walt@tilera.com>
8305 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
8306 arguments on even registers.
8307 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
8309 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
8310 (BIGGEST_ALIGNMENT): Ditto.
8311 (BIGGEST_FIELD_ALIGNMENT): Ditto.
8313 2014-01-25 Walter Lee <walt@tilera.com>
8315 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
8316 insns before bundling.
8317 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8319 2014-01-25 Walter Lee <walt@tilera.com>
8321 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
8322 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
8323 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
8325 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8327 * config/mips/constraints.md (kl): Delete.
8328 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
8329 define expands, using...
8330 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
8331 instructions for MIPS16.
8332 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
8333 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
8335 2014-01-25 Walter Lee <walt@tilera.com>
8337 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
8341 2014-01-25 Walter Lee <walt@tilera.com>
8343 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
8344 (TARGET_EXPAND_TO_RTL_HOOK): Define.
8346 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8348 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
8351 2014-01-25 Jakub Jelinek <jakub@redhat.com>
8353 * print-rtl.c (in_call_function_usage): New var.
8354 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
8355 EXPR_LIST mode as mode and not as reg note name.
8358 * cfgloopmanip.c (copy_loop_info): If
8359 loop->warned_aggressive_loop_optimizations, make sure
8360 the flag is set in target loop too.
8362 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
8364 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
8366 * builtins.def: Likewise.
8367 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
8368 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
8369 * ira.c (ira_setup_eliminable_regset): Likewise.
8370 * omp-low.c (gate_expand_omp): Likewise.
8371 (execute_lower_omp): Likewise.
8372 (diagnose_sb_0): Likewise.
8373 (gate_diagnose_omp_blocks): Likewise.
8374 (simd_clone_clauses_extract): Likewise.
8377 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8379 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
8380 correction for little endian...
8381 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
8384 2014-01-24 Jeff Law <law@redhat.com>
8386 PR tree-optimization/59919
8387 * tree-vrp.c (find_assert_locations_1): Do not register asserts
8388 for non-returning calls.
8390 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
8392 * common/config/aarch64/aarch64-common.c
8393 (aarch64_rewrite_mcpu): Handle multiple names.
8394 * config/aarch64/aarch64.h
8395 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8397 2014-01-24 Dodji Seketeli <dodji@redhat.com>
8399 * input.c (add_file_to_cache_tab): Handle the case where fopen
8402 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
8405 * config/i386/i386.md (pushsf splitter): Get stack adjustment
8406 from push operand if code of push isn't PRE_DEC.
8408 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8411 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
8412 -mquad-memory-atomic. Update -mquad-memory documentation to say
8413 it is only used for non-atomic loads/stores.
8415 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
8416 -mquad-memory or -mquad-memory-atomic switches.
8418 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
8419 -mquad-memory-atomic to ISA 2.07 support.
8421 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
8422 to separate support of normal quad word memory operations (ldq, stq)
8423 from the atomic quad word memory operations.
8425 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
8426 support to separate non-atomic quad word operations from atomic
8427 quad word operations. Disable non-atomic quad word operations in
8428 little endian mode so that we don't have to swap words after the
8429 load and before the store.
8430 (quad_load_store_p): Add comment about atomic quad word support.
8431 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
8432 options printed with -mdebug=reg.
8434 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
8435 -mquad-memory-atomic as the test for whether we have quad word
8436 atomic instructions.
8437 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
8438 or -mp8-vector are used, allow byte/half-word atomic operations.
8440 * config/rs6000/sync.md (load_lockedti): Insure that the address
8441 is a proper indexed or indirect address for the lqarx instruction.
8442 On little endian systems, swap the hi/lo registers after the lqarx
8444 (load_lockedpti): Use indexed_or_indirect_operand predicate to
8445 insure the address is valid for the lqarx instruction.
8446 (store_conditionalti): Insure that the address is a proper indexed
8447 or indirect address for the stqcrx. instruction. On little endian
8448 systems, swap the hi/lo registers before doing the stqcrx.
8450 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
8451 insure the address is valid for the stqcrx. instruction.
8453 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8454 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
8455 type of quad memory support is available.
8457 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
8460 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
8461 there is a danger of looping.
8463 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
8465 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8466 force flag_ira_loop_pressure if set via command line.
8468 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8470 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
8471 (ashr_simd): New builtin handling DI mode.
8472 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
8473 (aarch64_sshr_simddi): New match pattern.
8474 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
8475 (vshrd_n_s64): Likewise.
8476 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
8478 2014-01-23 Nick Clifton <nickc@redhat.com>
8480 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
8481 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
8482 favour of mcu specific scripts.
8483 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
8486 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8487 Alex Velenko <Alex.Velenko@arm.com>
8489 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
8490 (vaddv_s16): Likewise.
8491 (vaddv_s32): Likewise.
8492 (vaddv_u8): Likewise.
8493 (vaddv_u16): Likewise.
8494 (vaddv_u32): Likewise.
8495 (vaddvq_s8): Likewise.
8496 (vaddvq_s16): Likewise.
8497 (vaddvq_s32): Likewise.
8498 (vaddvq_s64): Likewise.
8499 (vaddvq_u8): Likewise.
8500 (vaddvq_u16): Likewise.
8501 (vaddvq_u32): Likewise.
8502 (vaddvq_u64): Likewise.
8503 (vaddv_f32): Likewise.
8504 (vaddvq_f32): Likewise.
8505 (vaddvq_f64): Likewise.
8506 (vmaxv_f32): Likewise.
8507 (vmaxv_s8): Likewise.
8508 (vmaxv_s16): Likewise.
8509 (vmaxv_s32): Likewise.
8510 (vmaxv_u8): Likewise.
8511 (vmaxv_u16): Likewise.
8512 (vmaxv_u32): Likewise.
8513 (vmaxvq_f32): Likewise.
8514 (vmaxvq_f64): Likewise.
8515 (vmaxvq_s8): Likewise.
8516 (vmaxvq_s16): Likewise.
8517 (vmaxvq_s32): Likewise.
8518 (vmaxvq_u8): Likewise.
8519 (vmaxvq_u16): Likewise.
8520 (vmaxvq_u32): Likewise.
8521 (vmaxnmv_f32): Likewise.
8522 (vmaxnmvq_f32): Likewise.
8523 (vmaxnmvq_f64): Likewise.
8524 (vminv_f32): Likewise.
8525 (vminv_s8): Likewise.
8526 (vminv_s16): Likewise.
8527 (vminv_s32): Likewise.
8528 (vminv_u8): Likewise.
8529 (vminv_u16): Likewise.
8530 (vminv_u32): Likewise.
8531 (vminvq_f32): Likewise.
8532 (vminvq_f64): Likewise.
8533 (vminvq_s8): Likewise.
8534 (vminvq_s16): Likewise.
8535 (vminvq_s32): Likewise.
8536 (vminvq_u8): Likewise.
8537 (vminvq_u16): Likewise.
8538 (vminvq_u32): Likewise.
8539 (vminnmv_f32): Likewise.
8540 (vminnmvq_f32): Likewise.
8541 (vminnmvq_f64): Likewise.
8543 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8545 * config/aarch64/aarch64-simd.md
8546 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
8547 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
8548 (*aarch64_mul3_elt<mode>): Likewise.
8549 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
8550 (*aarch64_mul3_elt_to_64v2df): Likewise.
8551 (*aarch64_mla_elt<mode>): Likewise.
8552 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8553 (*aarch64_mls_elt<mode>): Likewise.
8554 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8555 (*aarch64_fma4_elt<mode>): Likewise.
8556 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8557 (*aarch64_fma4_elt_to_64v2df): Likewise.
8558 (*aarch64_fnma4_elt<mode>): Likewise.
8559 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8560 (*aarch64_fnma4_elt_to_64v2df): Likewise.
8561 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
8562 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
8563 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8564 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8565 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
8566 (aarch64_sqdmull_lane<mode>_internal): Likewise.
8567 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
8569 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
8571 * config/aarch64/aarch64-simd.md
8572 (aarch64_be_checked_get_lane<mode>): New define_expand.
8573 * config/aarch64/aarch64-simd-builtins.def
8574 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
8575 New builtin definition.
8576 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
8577 Use new safe be builtin.
8579 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8581 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
8583 (aarch64_be_st1<mode>): Likewise.
8584 (aarch_ld1<VALL:mode>): Define_expand modified.
8585 (aarch_st1<VALL:mode>): Likewise.
8586 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
8587 (UNSPEC_ST1): Likewise.
8589 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
8591 * config/microblaze/microblaze.md: Add trap insn and attribute
8593 2014-01-23 Dodji Seketeli <dodji@redhat.com>
8595 PR preprocessor/58580
8596 * input.h (location_get_source_line): Take an additional line_size
8598 (void diagnostics_file_cache_fini): Declare new function.
8599 * input.c (struct fcache): New type.
8600 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
8601 New static constants.
8602 (diagnostic_file_cache_init, total_lines_num)
8603 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
8604 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
8605 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
8606 (get_next_line, read_next_line, goto_next_line, read_line_num):
8607 New static function definitions.
8608 (diagnostic_file_cache_fini): New function.
8609 (location_get_source_line): Take an additional output line_len
8610 parameter. Re-write using lookup_or_add_file_to_cache_tab and
8612 * diagnostic.c (diagnostic_finish): Call
8613 diagnostic_file_cache_fini.
8614 (adjust_line): Take an additional input parameter for the length
8615 of the line, rather than calculating it with strlen.
8616 (diagnostic_show_locus): Adjust the use of
8617 location_get_source_line and adjust_line with respect to their new
8618 signature. While displaying a line now, do not stop at the first
8619 null byte. Rather, display the zero byte as a space and keep
8620 going until we reach the size of the line.
8621 * Makefile.in: Add vec.o to OBJS-libcommon
8623 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8624 Ilya Tocar <ilya.tocar@intel.com>
8626 * config/i386/avx512fintrin.h (_mm512_kmov): New.
8627 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
8628 (__builtin_ia32_kmov16): Ditto.
8629 * config/i386/i386.md (UNSPEC_KMOV): New.
8632 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8634 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
8635 (_mm512_storeu_si512): Ditto.
8637 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
8640 * rtl.h (get_referenced_operands): Declare.
8641 * recog.c (get_referenced_operands): New function.
8642 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
8643 operands have been referenced when recording LO_SUM references.
8645 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
8647 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
8649 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8651 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
8652 Enable for generic and recent AMD targets.
8654 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8656 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
8657 ARG_SIZE note when adjustment was eliminated.
8659 2014-01-22 Jeff Law <law@redhat.com>
8661 PR tree-optimization/59597
8662 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
8663 in file. Accept new argument REGISTERING and use it to modify
8664 dump output appropriately.
8665 (register_jump_thread): Corresponding changes.
8666 (mark_threaded_blocks): Reinstate code to cancel unprofitable
8667 thread paths involving joiner blocks. Add code to dump cancelled
8668 jump threading paths.
8670 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
8672 PR rtl-optimization/59477
8673 * lra-constraints.c (inherit_in_ebb): Process call for living hard
8674 regs. Update reloads_num and potential_reload_hard_regs for all insns.
8676 2014-01-22 Tom Tromey <tromey@redhat.com>
8678 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
8680 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
8682 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8684 PR rtl-optimization/59896
8685 * lra-constraints.c (process_alt_operands): Check unused note for
8686 matched operands of insn with no output reloads.
8688 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
8690 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
8691 (mips_move_from_gpr_cost): Likewise.
8693 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8695 PR rtl-optimization/59858
8696 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
8697 ira_class_hard_regs_num.
8698 (process_alt_operands): Increase reject for dying matched operand.
8700 2014-01-21 Jakub Jelinek <jakub@redhat.com>
8703 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
8704 smaller than size, perform several stores or loads and stores
8705 at dst + count - size to store or copy all of size bytes, rather
8706 than just last modesize bytes.
8708 2014-01-20 DJ Delorie <dj@redhat.com>
8710 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
8711 that CLOBBERs are REGs before propogating their values.
8713 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
8716 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
8717 (cgraph_inline_failed_type): New function.
8718 * cgraph.h (DEFCIFCODE): Add type.
8719 (cgraph_inline_failed_type_t): New enum.
8720 (cgraph_inline_failed_type): New prototype.
8721 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
8722 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
8723 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
8724 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
8725 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
8726 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
8727 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
8728 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
8729 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
8730 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
8731 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
8732 OPTIMIZATION_MISMATCH.
8733 * tree-inline.c (expand_call_inline): Emit errors during
8734 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
8736 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8739 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
8740 mode attribute in insn output.
8742 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
8744 * output.h (output_constant): Delete.
8745 * varasm.c (output_constant): Make private.
8747 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
8749 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
8751 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8754 * tree.h (fold_builtin_strcat): New prototype.
8755 * builtins.c (fold_builtin_strcat): No longer static. Add len
8756 argument, if non-NULL, don't call c_strlen. Optimize
8757 directly into __builtin_memcpy instead of __builtin_strcpy.
8758 (fold_builtin_2): Adjust fold_builtin_strcat caller.
8759 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
8761 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8763 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8764 for SImode_address_operand operands, having only a REG argument.
8766 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
8768 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
8769 loader name using mbig-endian.
8770 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
8772 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8774 * doc/invoke.texi (-march): Clarify documentation for AArch64.
8778 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
8780 * config/aarch64/aarch64-protos.h
8781 (aarch64_cannot_change_mode_class_ptr): Declare.
8782 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
8783 aarch64_cannot_change_mode_class_ptr): New.
8784 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
8785 backend hook aarch64_cannot_change_mode_class.
8787 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8789 * common/config/aarch64/aarch64-common.c
8790 (aarch64_handle_option): Don't handle any option order logic here.
8791 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
8792 selected_cpu, warn on architecture version mismatch.
8793 (aarch64_override_options): Fix parsing order for option strings.
8795 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8796 Iain Sandoe <iain@codesourcery.com>
8799 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
8800 warning. Amend comment to reflect current functionality.
8802 2014-01-20 Richard Biener <rguenther@suse.de>
8805 * builtins.c (fold_builtin_strcat): Remove case better handled
8806 by tree-ssa-strlen.c.
8808 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
8810 * config/aarch64/aarch64.opt
8811 (mcpu, march, mtune): Make case-insensitive.
8813 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8816 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8817 if operands[1] is a REG or ZERO_EXTEND of a REG.
8819 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
8821 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
8823 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
8825 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
8826 long non-pic millicode calls.
8828 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8830 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
8832 2014-01-19 Kito Cheng <kito@0xlab.org>
8834 * builtins.c (expand_movstr): Check movstr expand done or fail.
8836 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8837 H.J. Lu <hongjiu.lu@intel.com>
8840 * config/i386/i386.md (*lea<mode>): Zero-extend return register
8841 to DImode for zero-extended addresses.
8843 2014-01-19 Jakub Jelinek <jakub@redhat.com>
8845 PR rtl-optimization/57763
8846 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
8847 on the new indirect jump_insn and increment LABEL_NUSES (label).
8849 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
8853 * config.gcc (x86_archs): New variable.
8854 (x86_64_archs): Likewise.
8855 (x86_cpus): Likewise.
8856 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
8857 --with-arch/--with-cpu= options.
8858 Support --with-arch=/--with-cpu={nehalem,westmere,
8859 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
8861 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8863 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
8864 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
8866 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
8868 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
8870 2014-01-18 Jakub Jelinek <jakub@redhat.com>
8873 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
8874 clear cpp_get_options (parse_in)->warn_unused_macros for
8875 ix86_target_macros_internal with cpp_define.
8877 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
8879 * jump.c (delete_related_insns): Keep (use (insn))s.
8880 * reorg.c (redundant_insn): Check for barriers too.
8882 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8884 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
8886 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
8888 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
8889 call to $$dyncall when TARGET_LONG_CALLS is true.
8891 2014-01-17 Jeff Law <law@redhat.com>
8893 * ree.c (combine_set_extension): Temporarily disable test for
8894 changing number of hard registers.
8896 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
8899 * ipa-inline-analysis.c (inline_free_summary):
8900 Do not free summary of aliases.
8902 2014-01-17 Jakub Jelinek <jakub@redhat.com>
8905 * gimplify.c (gimplify_expr): Use create_tmp_var
8906 instead of create_tmp_var_raw. If cond doesn't have
8907 integral type, don't add the IFN_ANNOTATE builtin at all.
8909 2014-01-17 Martin Jambor <mjambor@suse.cz>
8912 * ipa-cp.c (prev_edge_clone): New variable.
8913 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
8914 Also resize prev_edge_clone vector.
8915 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
8916 (ipcp_edge_removal_hook): New function.
8917 (ipcp_driver): Register ipcp_edge_removal_hook.
8919 2014-01-17 Andrew Pinski <apinski@cavium.com>
8920 Steve Ellcey <sellcey@mips.com>
8923 * config/mips/mips.c (mips_print_operand): Check operand mode instead
8926 2014-01-17 Jeff Law <law@redhat.com>
8929 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
8930 so that pass_ccp runs first.
8932 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8934 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
8935 (ix86_adjust_cost): Use !TARGET_XXX.
8936 (do_reorder_for_imul): Likewise.
8937 (swap_top_of_ready_list): Likewise.
8938 (ix86_sched_reorder): Likewise.
8940 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
8942 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
8943 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
8944 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
8945 (intel_memset): New. Duplicate slm_memset.
8946 (intel_cost): New. Duplicate slm_cost.
8947 (m_INTEL): New macro.
8948 (processor_target_table): Add "intel".
8949 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
8950 with PROCESSOR_INTEL for "intel".
8951 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
8952 PROCESSOR_SILVERMONT.
8953 (ix86_issue_rate): Likewise.
8954 (ix86_adjust_cost): Likewise.
8955 (ia32_multipass_dfa_lookahead): Likewise.
8956 (swap_top_of_ready_list): Likewise.
8957 (ix86_sched_reorder): Likewise.
8958 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
8959 instead of TARGET_OPT_AGU.
8960 * config/i386/i386.h (TARGET_INTEL): New.
8961 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
8962 (processor_type): Add PROCESSOR_INTEL.
8963 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
8964 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
8966 2014-01-17 Marek Polacek <polacek@redhat.com>
8969 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
8972 2014-01-17 Richard Biener <rguenther@suse.de>
8974 PR tree-optimization/46590
8975 * opts.c (default_options_table): Add entries for
8976 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
8977 all enabled at -O1 but not for -Og.
8978 * common.opt (fbranch-count-reg): Remove Init(1).
8979 (fmove-loop-invariants): Likewise.
8980 (ftree-pta): Likewise.
8982 2014-01-17 Jakub Jelinek <jakub@redhat.com>
8984 * config/i386/i386.c (ix86_data_alignment): For compatibility with
8985 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
8986 decls to at least the GCC 4.8 used alignments.
8989 * tree-nested.c (convert_nonlocal_reference_stmt,
8990 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
8991 of GIMPLE_BIND stmts, adjust associated decls.
8993 2014-01-17 Richard Biener <rguenther@suse.de>
8995 PR tree-optimization/46590
8996 * vec.h (vec<>::bseach): New member function implementing
8997 binary search according to C89 bsearch.
8998 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
8999 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
9000 bitmap pointer again. Make accesses_in_loop a flat array.
9001 (mem_ref_obstack): New global.
9002 (outermost_indep_loop): Adjust for mem_ref->stored changes.
9003 (mark_ref_stored): Likewise.
9004 (ref_indep_loop_p_2): Likewise.
9005 (set_ref_stored_in_loop): New helper function.
9006 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
9007 (memref_free): Adjust.
9008 (record_mem_ref_loc): Simplify.
9009 (gather_mem_refs_stmt): Adjust.
9010 (sort_locs_in_loop_postorder_cmp): New function.
9011 (analyze_memory_references): Sort accesses_in_loop after
9012 loop postorder number.
9013 (find_ref_loc_in_loop_cmp): New function.
9014 (for_all_locs_in_loop): Find relevant cluster of locs in
9015 accesses_in_loop and iterate without recursion.
9016 (execute_sm): Avoid uninit warning.
9017 (struct ref_always_accessed): Simplify.
9018 (ref_always_accessed::operator ()): Likewise.
9019 (ref_always_accessed_p): Likewise.
9020 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
9021 loop postorder numbers here.
9022 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
9025 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9028 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
9029 on decls for which assemble_alias has been called.
9031 2014-01-17 Nick Clifton <nickc@redhat.com>
9033 * config/msp430/msp430.opt: (mcpu): New option.
9034 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
9035 (msp430_option_override): Parse target_cpu. If the MCU name
9036 matches a generic string, clear target_mcu.
9037 (msp430_attr): Allow numeric interrupt values up to 63.
9038 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
9039 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
9041 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
9043 * config/msp430/msp430.md (popm): Use %J rather than %I.
9044 (addsi3): Use msp430_nonimmediate_operand for operand 2.
9045 (addhi_cy_i): Use immediate_operand for operand 2.
9046 * doc/invoke.texi: Document -mcpu option.
9048 2014-01-17 Richard Biener <rguenther@suse.de>
9050 PR rtl-optimization/38518
9051 * df.h (df_analyze_loop): Declare.
9052 * df-core.c: Include cfgloop.h.
9053 (df_analyze_1): Split out main part of df_analyze.
9054 (df_analyze): Adjust.
9055 (loop_inverted_post_order_compute): New function.
9056 (loop_post_order_compute): Likewise.
9057 (df_analyze_loop): New function avoiding whole-function
9059 * loop-invariant.c (find_defs): Use df_analyze_loop.
9060 (find_invariants): Adjust.
9061 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
9063 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
9065 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
9066 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
9068 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
9070 * ipa-ref.c (ipa_remove_stmt_references): Fix references
9071 traversal when removing references.
9073 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
9076 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
9078 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
9081 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
9082 pushing a reload for an autoinc when we had previously reloaded an
9083 inner part of the address.
9085 2014-01-16 Jakub Jelinek <jakub@redhat.com>
9087 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
9089 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
9090 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
9091 when not giving up or versioning for alias only because of
9093 (vect_analyze_data_ref_dependences): Set to true.
9094 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
9096 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
9097 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
9101 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
9104 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
9105 operand 0 predicate for gathers, use a new pseudo as subtarget.
9107 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9110 * lra-constraints.c (process_alt_operands): Add printing debug info.
9111 Check absence of input/output reloads for matched operands too.
9113 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9115 PR rtl-optimization/59835
9116 * ira.c (ira_init_register_move_cost): Increase cost for
9119 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
9121 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
9123 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
9126 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
9127 non-register objects. Use gen_(high/low)part more consistently.
9130 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9133 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
9134 endian support, remove tests for WORDS_BIG_ENDIAN.
9135 (p8_mfvsrd_3_<mode>): Likewise.
9136 (reload_gpr_from_vsx<mode>): Likewise.
9137 (reload_gpr_from_vsxsf): Likewise.
9138 (p8_mfvsrd_4_disf): Likewise.
9140 2014-01-16 Richard Biener <rguenther@suse.de>
9142 PR rtl-optimization/46590
9143 * lcm.c (compute_antinout_edge): Use postorder iteration.
9144 (compute_laterin): Use inverted postorder iteration.
9146 2014-01-16 Nick Clifton <nickc@redhat.com>
9149 * varasm.c (output_constant): Return the number of bytes actually
9151 (output_constructor_array_range): Update the field size with the
9152 number of bytes emitted by output_constant.
9153 (output_constructor_regular_field): Likewise. Also do not
9154 complain if the total number of bytes emitted is now greater
9155 than the expected fieldpos.
9156 * output.h (output_constant): Update prototype and descriptive comment.
9158 2014-01-16 Marek Polacek <polacek@redhat.com>
9161 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
9162 it is error_mark_node.
9164 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
9166 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
9167 VALID_AVX256_REG_OR_OI_MODE.
9169 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
9171 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
9172 current procedure should be profiled.
9174 2014-01-15 Andrew Pinski <apinski@cavium.com>
9176 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
9177 of moving from/to the STACK_REG register class.
9179 2014-01-15 Richard Henderson <rth@redhat.com>
9182 * reginfo.c (global_regs_decl): Globalize.
9183 * rtl.h (global_regs_decl): Declare.
9184 * ira.c (do_reload): Diagnose frame_pointer_needed and it
9185 reserved via global_regs.
9187 2014-01-15 Teresa Johnson <tejohnson@google.com>
9189 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
9191 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9193 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
9194 and vmulosh rather than call gen_vec_widen_smult_*.
9195 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
9196 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
9197 (vec_widen_smult_even_v16qi): Likewise.
9198 (vec_widen_umult_even_v8hi): Likewise.
9199 (vec_widen_smult_even_v8hi): Likewise.
9200 (vec_widen_umult_odd_v16qi): Likewise.
9201 (vec_widen_smult_odd_v16qi): Likewise.
9202 (vec_widen_umult_odd_v8hi): Likewise.
9203 (vec_widen_smult_odd_v8hi): Likewise.
9204 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
9205 vmuloub rather than call gen_vec_widen_umult_*.
9206 (vec_widen_umult_lo_v16qi): Likewise.
9207 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
9208 vmulosb rather than call gen_vec_widen_smult_*.
9209 (vec_widen_smult_lo_v16qi): Likewise.
9210 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
9211 rather than call gen_vec_widen_umult_*.
9212 (vec_widen_umult_lo_v8hi): Likewise.
9213 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
9214 rather than call gen_vec_widen_smult_*.
9215 (vec_widen_smult_lo_v8hi): Likewise.
9217 2014-01-15 Jeff Law <law@redhat.com>
9219 PR tree-optimization/59747
9220 * ree.c (find_and_remove_re): Properly handle case where a second
9221 eliminated extension requires widening a copy created for elimination
9222 of a prior extension.
9223 (combine_set_extension): Ensure that the number of hard regs needed
9224 for a destination register does not change when we widen it.
9226 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
9228 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
9229 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
9230 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
9231 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
9232 (avr-*-rtems*): Likewise.
9233 (bfin*-rtems*): Likewise.
9234 (moxie-*-rtems*): Likewise.
9235 (h8300-*-rtems*): Likewise.
9236 (i[34567]86-*-rtems*): Likewise.
9237 (lm32-*-rtems*): Likewise.
9238 (m32r-*-rtems*): Likewise.
9239 (m68k-*-rtems*): Likewise.
9240 (microblaze*-*-rtems*): Likewise.
9241 (mips*-*-rtems*): Likewise.
9242 (powerpc-*-rtems*): Likewise.
9243 (sh-*-rtems*): Likewise.
9244 (sparc-*-rtems*): Likewise.
9245 (sparc64-*-rtems*): Likewise.
9246 (v850-*-rtems*): Likewise.
9247 (m32c-*-rtems*): Likewise.
9249 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9251 PR rtl-optimization/59511
9252 * ira.c (ira_init_register_move_cost): Use memory costs for some
9253 cases of register move cost calculations.
9254 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9255 instead of BB frequency.
9256 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9257 * lra-assigns.c (find_hard_regno_for): Ditto.
9259 2014-01-15 Richard Biener <rguenther@suse.de>
9261 PR tree-optimization/59822
9262 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9263 (vectorizable_load): Use it to hoist defs of uses of invariant
9264 loads out of the loop.
9266 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9267 Kugan Vivekanandarajah <kuganv@linaro.org>
9270 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9273 2014-01-15 Richard Biener <rguenther@suse.de>
9275 PR rtl-optimization/59802
9276 * lcm.c (compute_available): Use inverted postorder to seed
9277 the initial worklist.
9279 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9282 * config/s390/s390.c (s390_preferred_reload_class): Don't return
9283 ADDR_REGS for invalid symrefs in non-PIC code.
9285 2014-01-15 Jakub Jelinek <jakub@redhat.com>
9288 * builtins.c (determine_block_size): Initialize *probable_max_size
9289 even if len_rtx is CONST_INT.
9291 2014-01-14 Andrew Pinski <apinski@cavium.com>
9293 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
9294 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
9295 (cortexa53_tunings): Likewise.
9296 (aarch64_sched_issue_rate): New function.
9297 (TARGET_SCHED_ISSUE_RATE): Define.
9299 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9301 * ira-costs.c (find_costs_and_classes): Add missed
9302 ira_init_register_move_cost_if_necessary.
9304 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9307 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
9309 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
9312 * config/i386/i386.c (type_natural_mode): Add a bool parameter
9313 to indicate if type is used for function return value. Warn ABI
9314 change if the vector mode isn't available for function return value.
9315 (ix86_function_arg_advance): Pass false to type_natural_mode.
9316 (ix86_function_arg): Likewise.
9317 (ix86_gimplify_va_arg): Likewise.
9318 (function_arg_32): Don't warn ABI change.
9319 (ix86_function_value): Pass true to type_natural_mode.
9320 (ix86_return_in_memory): Likewise.
9321 (ix86_struct_value_rtx): Removed.
9322 (TARGET_STRUCT_VALUE_RTX): Likewise.
9324 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9326 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
9327 converting a conditional jump into a conditional return.
9329 2014-01-14 Richard Biener <rguenther@suse.de>
9331 PR tree-optimization/58921
9332 PR tree-optimization/59006
9333 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
9334 hoisting invariant stmts.
9335 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
9336 invariant loads on the preheader edge if possible.
9338 2014-01-14 Joey Ye <joey.ye@arm.com>
9340 * doc/plugin.texi (Building GCC plugins): Update to C++.
9342 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
9344 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
9345 (_mm_rcp28_round_ss): Ditto.
9346 (_mm_rsqrt28_round_sd): Ditto.
9347 (_mm_rsqrt28_round_ss): Ditto.
9348 (_mm_rcp28_sd): Ditto.
9349 (_mm_rcp28_ss): Ditto.
9350 (_mm_rsqrt28_sd): Ditto.
9351 (_mm_rsqrt28_ss): Ditto.
9352 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
9353 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
9354 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
9355 (IX86_BUILTIN_RCP28SD): Ditto.
9356 (IX86_BUILTIN_RCP28SS): Ditto.
9357 (IX86_BUILTIN_RSQRT28SD): Ditto.
9358 (IX86_BUILTIN_RSQRT28SS): Ditto.
9359 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
9360 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
9361 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
9362 (ix86_expand_special_args_builtin): Expand new FTYPE.
9363 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
9364 (srcp14<mode>): Make insn unary.
9365 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
9366 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
9367 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
9368 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
9369 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
9370 Fix rounding: make it SAE only.
9371 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9373 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9375 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
9376 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
9377 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
9378 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
9379 (round_saeonly_mask_scalar_operand4): Ditto.
9380 (round_saeonly_mask_scalar_op3): Ditto.
9381 (round_saeonly_mask_scalar_op4): Ditto.
9383 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9385 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9386 Implement -maltivec=be for vec_insert and vec_extract.
9388 2014-01-10 DJ Delorie <dj@redhat.com>
9390 * config/msp430/msp430.md (call_internal): Don't allow memory
9391 references with SP as the base register.
9392 (call_value_internal): Likewise.
9393 * config/msp430/constraints.md (Yc): New. For memory references
9394 that don't use SP as a base register.
9396 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
9397 "an integer without a # prefix"
9398 * config/msp430/msp430.md (epilogue_helper): Use it.
9400 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9403 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
9404 AVX512F gather builtins.
9405 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
9406 on gather decls with INTEGER_TYPE masktype.
9407 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
9408 directly into the builtin rather than hoisting it before loop.
9410 PR tree-optimization/59387
9411 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
9412 (scev_const_prop): If folded_casts and type has undefined overflow,
9413 use force_gimple_operand instead of force_gimple_operand_gsi and
9414 for each added stmt if it is assign with
9415 arith_code_with_undefined_signed_overflow, call
9416 rewrite_to_defined_overflow.
9417 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
9418 gimple-fold.h instead.
9419 (arith_code_with_undefined_signed_overflow,
9420 rewrite_to_defined_overflow): Moved to ...
9421 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
9422 rewrite_to_defined_overflow): ... here. No longer static.
9423 Include gimplify-me.h.
9424 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
9425 rewrite_to_defined_overflow): New prototypes.
9427 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9429 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
9431 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
9433 * builtins.c (get_object_alignment_2): Minor tweak.
9434 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
9436 2014-01-13 Christian Bruel <christian.bruel@st.com>
9438 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
9439 optimized non constant lengths.
9441 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9444 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
9445 load as __atomic_load_N if possible.
9447 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9449 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
9451 (rs6000_expand_builtin): Adjust call.
9453 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9456 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
9457 * config/rs6000/rs6000.c: Include target-globals.h.
9458 (rs6000_set_current_function): Instead of doing target_reinit
9459 unconditionally, use save_target_globals_default_opts and
9460 restore_target_globals.
9462 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
9464 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
9465 (rs6000_expand_builtin): Handle mffs and mtfsf.
9466 (rs6000_init_builtins): Define mffs and mtfsf.
9467 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
9468 (rs6000_mffs): New pattern.
9469 (rs6000_mtfsf): New pattern.
9471 2014-01-11 Bin Cheng <bin.cheng@arm.com>
9473 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
9474 Start narrowing with START. Apply candidate-use pair
9475 and check overall cost in narrowing.
9476 (iv_ca_prune): Pass new argument.
9478 2014-01-10 Jeff Law <law@redhat.com>
9481 * ree.c (combine_reaching_defs): Ensure the defining statement
9482 occurs before the extension when optimizing extensions with
9483 different source and destination hard registers.
9485 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9488 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
9489 vtables into the type inheritance graph.
9491 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9493 PR rtl-optimization/59754
9494 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
9495 modes in the REGNO != REGNO case.
9497 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9499 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
9501 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9503 PR tree-optimization/59745
9504 * tree-predcom.c (tree_predictive_commoning_loop): Call
9505 free_affine_expand_cache if giving up because components is NULL.
9507 * target-globals.c (save_target_globals): Allocate < 4KB structs using
9508 GC in payload of target_globals struct instead of allocating them on
9509 the heap and the larger structs separately using GC.
9510 * target-globals.h (struct target_globals): Make regs, hard_regs,
9511 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
9512 of GTY((skip)) and change type to void *.
9513 (reset_target_globals): Cast loads from those fields to corresponding
9516 2014-01-10 Steve Ellcey <sellcey@mips.com>
9519 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
9520 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
9521 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
9523 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
9526 * aarch64-modes.def (CC_Zmode): New flags mode.
9527 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
9528 represents an equality.
9529 (aarch64_get_condition_code): Handle CC_Zmode.
9530 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
9532 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9534 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
9535 extraction in good case.
9537 2014-01-10 Richard Biener <rguenther@suse.de>
9539 PR tree-optimization/59374
9540 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
9541 checking after SLP discovery. Mark stmts not participating
9542 in any SLP instance properly.
9544 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9546 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
9547 when handling a SET rtx.
9549 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9551 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
9552 (cortex-a57): Likewise.
9553 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
9555 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9557 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
9558 non-iwmmxt builtins.
9560 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9564 * ipa-devirt.c record_target_from_binfo): Take as argument
9565 stack of binfos and lookup matching one for virtual inheritance.
9566 (possible_polymorphic_call_targets_1): Update.
9568 2014-01-10 Huacai Chen <chenhc@lemote.com>
9570 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
9571 kernel strings for Loongson-2E/2F/3A.
9573 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9576 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
9577 is_gimple_call before calling gimple_call_internal_p.
9579 2014-01-09 Steve Ellcey <sellcey@mips.com>
9581 * Makefile.in (TREE_FLOW_H): Remove.
9582 (TREE_SSA_H): Add file names from tree-flow.h.
9583 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
9584 * tree.h: Remove tree-flow.h reference.
9585 * hash-table.h: Remove tree-flow.h reference.
9586 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
9587 reference with tree-ssa-loop.h.
9589 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9591 * doc/invoke.texi: Add -maltivec={be,le} options, and document
9592 default element-order behavior for -maltivec.
9593 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
9594 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
9595 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
9596 when targeting big endian, at least for now.
9597 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
9599 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9602 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
9603 var satisfies use_register_for_decl, just take into account type
9604 alignment, rather than decl alignment.
9606 PR tree-optimization/59622
9607 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
9608 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
9609 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
9610 Don't devirtualize for inplace at all. For targets.length () == 1,
9611 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
9613 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9615 * config/i386/i386.md (cpu): Remove the unused btver1.
9617 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9619 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
9621 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9624 * tree-core.h (struct target_globals): New forward declaration.
9625 (struct tree_target_option): Add globals field.
9626 * tree.h (TREE_TARGET_GLOBALS): Define.
9627 (prepare_target_option_nodes_for_pch): New prototype.
9628 * target-globals.h (struct target_globals): Define even if
9630 * tree.c (prepare_target_option_node_for_pch,
9631 prepare_target_option_nodes_for_pch): New functions.
9632 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
9633 * config/i386/i386.c: Include target-globals.h.
9634 (ix86_set_current_function): Instead of doing target_reinit
9635 unconditionally, use save_target_globals_default_opts and
9636 restore_target_globals.
9638 2014-01-09 Richard Biener <rguenther@suse.de>
9640 PR tree-optimization/59715
9641 * tree-cfg.h (split_critical_edges): Declare.
9642 * tree-cfg.c (split_critical_edges): Export.
9643 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
9645 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
9647 * cfgexpand.c (expand_stack_vars): Optionally disable
9648 asan stack protection.
9649 (expand_used_vars): Likewise.
9650 (partition_stack_vars): Likewise.
9651 * asan.c (asan_emit_stack_protection): Optionally disable
9652 after return stack usage.
9653 (instrument_derefs): Optionally disable memory access instrumentation.
9654 (instrument_builtin_call): Likewise.
9655 (instrument_strlen_call): Likewise.
9656 (asan_protect_global): Optionally disable global variables protection.
9657 * doc/invoke.texi: Added doc for new options.
9658 * params.def: Added new options.
9659 * params.h: Likewise.
9661 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9663 PR rtl-optimization/59724
9664 * ifcvt.c (cond_exec_process_if_block): Don't call
9665 flow_find_head_matching_sequence with 0 longest_match.
9666 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
9667 non-active insns if !stop_after.
9668 (try_head_merge_bb): Revert 2014-01-07 changes.
9670 2014-01-08 Jeff Law <law@redhat.com>
9672 * ree.c (get_sub_rtx): New function, extracted from...
9673 (merge_def_and_ext): Here.
9674 (combine_reaching_defs): Use get_sub_rtx.
9676 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
9678 * cgraph.h (varpool_variable_node): Do not choke on null node.
9680 2014-01-08 Catherine Moore <clm@codesourcery.com>
9682 * config/mips/mips.md (simple_return): Attempt to use JRC
9684 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
9686 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9688 PR rtl-optimization/59137
9689 * reorg.c (steal_delay_list_from_target): Call update_block for
9691 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
9693 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9695 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
9696 two duplicate entries.
9698 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9701 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
9703 * config/mips/mips.c (mips_truncated_op_cost): New function.
9704 (mips_rtx_costs): Adjust test for BADDU.
9705 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
9707 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
9709 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
9710 (*baddu_si): ...this new pattern.
9712 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9715 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
9717 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9720 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
9722 (expand_expr, expand_normal): Adjust.
9723 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
9724 inner_reference_p. Use inner_reference_p to expand inner references.
9725 (store_expr): Adjust.
9726 * cfgexpand.c (expand_call_stmt): Adjust.
9728 2014-01-08 Rong Xu <xur@google.com>
9730 * gcov-io.c (gcov_var): Move from gcov-io.h.
9731 (gcov_position): Ditto.
9732 (gcov_is_error): Ditto.
9733 (gcov_rewrite): Ditto.
9734 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
9735 only part to libgcc/libgcov.h.
9737 2014-01-08 Marek Polacek <polacek@redhat.com>
9740 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
9742 2014-01-08 Marek Polacek <polacek@redhat.com>
9745 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
9747 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9749 PR rtl-optimization/59649
9750 * stor-layout.c (get_mode_bounds): For BImode return
9751 0 and STORE_FLAG_VALUE.
9753 2014-01-08 Richard Biener <rguenther@suse.de>
9756 * gimple.h (is_gimple_builtin_call): Remove.
9757 (gimple_builtin_call_types_compatible_p): New.
9758 (gimple_call_builtin_p): New overload.
9759 * gimple.c (is_gimple_builtin_call): Remove.
9760 (validate_call): Rename to ...
9761 (gimple_builtin_call_types_compatible_p): ... this and export. Also
9763 (validate_type): New static function.
9764 (gimple_call_builtin_p): New overload and adjust.
9765 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
9766 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
9767 (gimple_fold_stmt_to_constant_1): Likewise.
9768 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
9770 2014-01-08 Richard Biener <rguenther@suse.de>
9773 * gimplify.c (gimplify_expr): Gimplify register-register type
9774 VIEW_CONVERT_EXPRs to separate stmts.
9776 2014-01-07 Jeff Law <law@redhat.com>
9779 * ree.c (combine_set_extension): Handle case where source
9780 and destination registers in an extension insn are different.
9781 (combine_reaching_defs): Allow source and destination registers
9782 in extension to be different under limited circumstances.
9783 (add_removable_extension): Remove restriction that the
9784 source and destination registers in the extension are the same.
9785 (find_and_remove_re): Emit a copy from the extension's
9786 destination to its source after the defining insn if
9787 the source and destination registers are different.
9790 * ifcvt.c (merge_if_block): If we are merging a block with more than
9791 one successor with a block with no successors, remove any BARRIER
9792 after the second block.
9794 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
9796 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
9798 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
9801 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
9802 for 14-bit register offsets when INT14_OK_STRICT is false.
9804 2014-01-07 Roland Stigge <stigge@antcom.de>
9805 Michael Meissner <meissner@linux.vnet.ibm.com>
9808 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
9809 Only check TFmode for SPE constants. Don't check TImode or TDmode.
9811 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
9813 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
9816 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
9818 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
9819 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
9820 rtx is const0_rtx or not.
9822 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
9825 * target-globals.c (save_target_globals): Remove this_fn_optab
9827 * toplev.c: Include optabs.h.
9828 (target_reinit): Temporarily restore the global options if another
9829 set of options are in force.
9831 2014-01-07 Jakub Jelinek <jakub@redhat.com>
9833 PR rtl-optimization/58668
9834 * cfgcleanup.c (flow_find_cross_jump): Don't count
9835 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
9836 to determine what is counted.
9837 (flow_find_head_matching_sequence): Use active_insn_p to determine
9839 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
9841 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
9842 determine what is counted.
9844 PR tree-optimization/59643
9845 * tree-predcom.c (split_data_refs_to_components): If one dr is
9846 read and one write, determine_offset fails and the write isn't
9847 in the bad component, just put the read into the bad component.
9849 2014-01-07 Mike Stump <mikestump@comcast.net>
9850 Jakub Jelinek <jakub@redhat.com>
9853 * tree-core.h (struct tree_optimization_option): Change optabs
9854 type from unsigned char * to void *.
9855 * optabs.c (init_tree_optimization_optabs): Adjust
9856 TREE_OPTIMIZATION_OPTABS initialization.
9858 2014-01-06 Jakub Jelinek <jakub@redhat.com>
9861 * config/i386/i386.h (struct machine_function): Add
9862 no_drap_save_restore field.
9863 * config/i386/i386.c (ix86_save_reg): Use
9864 !cfun->machine->no_drap_save_restore instead of
9865 crtl->stack_realign_needed.
9866 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
9867 this function clears frame_pointer_needed. Set
9868 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
9869 and DRAP reg is needed.
9871 2014-01-06 Marek Polacek <polacek@redhat.com>
9874 * doc/implement-c.texi: Mention that other integer types are
9875 permitted as bit-field types in strictly conforming mode.
9877 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
9879 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
9882 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
9884 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
9886 2014-01-06 Martin Jambor <mjambor@suse.cz>
9889 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
9891 * ipa-prop.c (ipa_print_node_params): Fix indentation.
9893 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
9897 * var-tracking.c (add_stores): Preserve the value of the source even if
9898 we don't record the store.
9900 2014-01-06 Terry Guo <terry.guo@arm.com>
9902 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
9904 2014-01-05 Iain Sandoe <iain@codesourcery.com>
9907 * config/darwin.c (darwin_function_section): Adjust return values to
9908 correspond to optimisation changes made in r206070.
9910 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
9912 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
9913 from prefetch_block tune setting.
9914 (nocona_cost): Correct size of prefetch block to 64.
9916 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
9918 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
9919 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
9920 used to save the static chain register in the computation of the offset
9921 from which the FP registers need to be restored.
9923 2014-01-04 Jakub Jelinek <jakub@redhat.com>
9925 PR tree-optimization/59519
9926 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
9927 ICE if get_current_def (current_new_name) is already non-NULL, as long
9928 as it is a phi result of some other phi in *new_exit_bb that has
9931 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
9932 or vmovdqu* for misaligned_operand.
9933 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
9934 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
9935 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
9936 aligned_mem for AVX512F masked aligned load and store builtins and for
9939 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
9941 PR tree-optimization/59651
9942 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
9943 Address range for negative step should be added by TYPE_SIZE_UNIT.
9945 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
9947 * config/m68k/m68k.c (handle_move_double): Handle pushes with
9948 overlapping registers also for registers other than the stack pointer.
9950 2014-01-03 Marek Polacek <polacek@redhat.com>
9953 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
9956 2014-01-03 Jakub Jelinek <jakub@redhat.com>
9959 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
9962 * config/i386/i386.md (MODE_SIZE): New mode attribute.
9963 (push splitter): Use <P:MODE_SIZE> instead of
9964 GET_MODE_SIZE (<P:MODE>mode).
9965 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
9966 (mov -1, reg peephole2): Likewise.
9967 * config/i386/sse.md (*mov<mode>_internal,
9968 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
9969 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
9970 *<code><mode>3, *andnot<mode>3<mask_name>,
9971 <mask_codefor><code><mode>3<mask_name>): Likewise.
9972 * config/i386/subst.md (mask_mode512bit_condition,
9973 sd_mask_mode512bit_condition): Likewise.
9975 2014-01-02 Xinliang David Li <davidxl@google.com>
9977 PR tree-optimization/59303
9978 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
9979 (dump_predicates): Better output format.
9980 (pred_equal_p): New function.
9981 (is_neq_relop_p): Ditto.
9982 (is_neq_zero_form_p): Ditto.
9983 (pred_expr_equal_p): Ditto.
9984 (pred_neg_p): Ditto.
9985 (simplify_pred): Ditto.
9986 (simplify_preds_2): Ditto.
9987 (simplify_preds_3): Ditto.
9988 (simplify_preds_4): Ditto.
9989 (simplify_preds): Ditto.
9991 (push_to_worklist): Ditto.
9992 (get_pred_info_from_cmp): Ditto.
9993 (is_degenerated_phi): Ditto.
9994 (normalize_one_pred_1): Ditto.
9995 (normalize_one_pred): Ditto.
9996 (normalize_one_pred_chain): Ditto.
9997 (normalize_preds): Ditto.
9998 (normalize_cond_1): Remove function.
9999 (normalize_cond): Ditto.
10000 (is_gcond_subset_of): Ditto.
10001 (is_subset_of_any): Ditto.
10002 (is_or_set_subset_of): Ditto.
10003 (is_and_set_subset_of): Ditto.
10004 (is_norm_cond_subset_of): Ditto.
10005 (pred_chain_length_cmp): Ditto.
10006 (convert_control_dep_chain_into_preds): Type change.
10007 (find_predicates): Ditto.
10008 (find_def_preds): Ditto.
10009 (destroy_predicates_vecs): Ditto.
10010 (find_matching_predicates_in_rest_chains): Ditto.
10011 (use_pred_not_overlap_with_undef_path_pred): Ditto.
10012 (is_pred_expr_subset): Ditto.
10013 (is_pred_chain_subset_of): Ditto.
10014 (is_included_in): Ditto.
10015 (is_superset_of): Ditto.
10017 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10019 Update copyright years.
10021 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10023 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
10024 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
10025 config/arc/arc.md, config/arc/arc.opt,
10026 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
10027 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
10028 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
10029 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
10030 config/linux-protos.h, config/linux.c, config/winnt-c.c,
10031 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
10032 vtable-verify.c, vtable-verify.h: Use the standard form for the
10035 2014-01-02 Tobias Burnus <burnus@net-b.de>
10037 * gcc.c (process_command): Update copyright notice dates.
10038 * gcov-dump.c: Ditto.
10040 * doc/cpp.texi: Bump @copying's copyright year.
10041 * doc/cppinternals.texi: Ditto.
10042 * doc/gcc.texi: Ditto.
10043 * doc/gccint.texi: Ditto.
10044 * doc/gcov.texi: Ditto.
10045 * doc/install.texi: Ditto.
10046 * doc/invoke.texi: Ditto.
10048 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10050 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
10052 2014-01-01 Jakub Jelinek <jakub@redhat.com>
10054 * config/i386/sse.md (*mov<mode>_internal): Guard
10055 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
10057 PR rtl-optimization/59647
10058 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
10059 new_rtx into UNSIGNED_FLOAT rtxes.
10061 Copyright (C) 2014 Free Software Foundation, Inc.
10063 Copying and distribution of this file, with or without modification,
10064 are permitted in any medium without royalty provided the copyright
10065 notice and this notice are preserved.