1 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
4 * config/alpha/alpha.c: Include gimple-iterator.h.
5 (alpha_gimple_fold_builtin): New function. Move
6 ALPHA_BUILTIN_UMULH folding from ...
7 (alpha_fold_builtin): ... here.
8 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
10 2014-05-08 Wei Mi <wmi@google.com>
13 * config/i386/i386.c (ix86_compute_frame_layout): Update
14 preferred_stack_boundary for call, expanded from tls descriptor.
15 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
16 to depend on SP register.
17 (*tls_local_dynamic_base_32_gnu): Ditto.
18 (*tls_local_dynamic_32_once): Ditto.
19 (tls_global_dynamic_64_<mode>): Set
20 ix86_tls_descriptor_calls_expanded_in_cfun.
21 (tls_local_dynamic_base_64_<mode>): Ditto.
22 (tls_global_dynamic_32): Set
23 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
24 to depend on SP register.
25 (tls_local_dynamic_base_32): Ditto.
27 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
29 * config/arm/arm_neon.h: Update comment.
30 * config/arm/neon-docgen.ml: Delete.
31 * config/arm/neon-gen.ml: Delete.
32 * doc/arm-neon-intrinsics.texi: Update comment.
34 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
36 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
38 (vand, vorr, veor, vorn, vbic): Remove.
39 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
41 (neon_vsub_unspec): Likewise.
42 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
44 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
46 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
56 (vaddq_s16): Likewise.
57 (vaddq_s32): Likewise.
58 (vaddq_s64): Likewise.
59 (vaddq_f32): Likewise.
61 (vaddq_u16): Likewise.
62 (vaddq_u32): Likewise.
63 (vaddq_u64): Likewise.
73 (vmulq_s16): Likewise.
74 (vmulq_s32): Likewise.
75 (vmulq_f32): Likewise.
77 (vmulq_u16): Likewise.
78 (vmulq_u32): Likewise.
89 (vsubq_s16): Likewise.
90 (vsubq_s32): Likewise.
91 (vsubq_s64): Likewise.
92 (vsubq_f32): Likewise.
94 (vsubq_u16): Likewise.
95 (vsubq_u32): Likewise.
96 (vsubq_u64): Likewise.
101 (vand_u16): Likewise.
102 (vand_u32): Likewise.
103 (vand_s64): Likewise.
104 (vand_u64): Likewise.
105 (vandq_s8): Likewise.
106 (vandq_s16): Likewise.
107 (vandq_s32): Likewise.
108 (vandq_s64): Likewise.
109 (vandq_u8): Likewise.
110 (vandq_u16): Likewise.
111 (vandq_u32): Likewise.
112 (vandq_u64): Likewise.
114 (vorr_s16): Likewise.
115 (vorr_s32): Likewise.
117 (vorr_u16): Likewise.
118 (vorr_u32): Likewise.
119 (vorr_s64): Likewise.
120 (vorr_u64): Likewise.
121 (vorrq_s8): Likewise.
122 (vorrq_s16): Likewise.
123 (vorrq_s32): Likewise.
124 (vorrq_s64): Likewise.
125 (vorrq_u8): Likewise.
126 (vorrq_u16): Likewise.
127 (vorrq_u32): Likewise.
128 (vorrq_u64): Likewise.
130 (veor_s16): Likewise.
131 (veor_s32): Likewise.
133 (veor_u16): Likewise.
134 (veor_u32): Likewise.
135 (veor_s64): Likewise.
136 (veor_u64): Likewise.
137 (veorq_s8): Likewise.
138 (veorq_s16): Likewise.
139 (veorq_s32): Likewise.
140 (veorq_s64): Likewise.
141 (veorq_u8): Likewise.
142 (veorq_u16): Likewise.
143 (veorq_u32): Likewise.
144 (veorq_u64): Likewise.
146 (vbic_s16): Likewise.
147 (vbic_s32): Likewise.
149 (vbic_u16): Likewise.
150 (vbic_u32): Likewise.
151 (vbic_s64): Likewise.
152 (vbic_u64): Likewise.
153 (vbicq_s8): Likewise.
154 (vbicq_s16): Likewise.
155 (vbicq_s32): Likewise.
156 (vbicq_s64): Likewise.
157 (vbicq_u8): Likewise.
158 (vbicq_u16): Likewise.
159 (vbicq_u32): Likewise.
160 (vbicq_u64): Likewise.
162 (vorn_s16): Likewise.
163 (vorn_s32): Likewise.
165 (vorn_u16): Likewise.
166 (vorn_u32): Likewise.
167 (vorn_s64): Likewise.
168 (vorn_u64): Likewise.
169 (vornq_s8): Likewise.
170 (vornq_s16): Likewise.
171 (vornq_s32): Likewise.
172 (vornq_s64): Likewise.
173 (vornq_u8): Likewise.
174 (vornq_u16): Likewise.
175 (vornq_u32): Likewise.
176 (vornq_u64): Likewise.
178 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
180 * wide-int.cc (UTItype): Define.
181 (UDWtype): Define for appropriate W_TYPE_SIZE.
183 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
185 PR tree-optimization/59100
186 * tree-ssa-phiopt.c: Include tree-inline.h.
187 (neutral_element_p, absorbing_element_p): New functions.
188 (value_replacement): Handle conditional binary operations with a
189 neutral or absorbing element.
191 2014-05-08 Richard Biener <rguenther@suse.de>
193 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
194 folding the expression.
195 (valueize_expr): Remove.
196 (visit_reference_op_load): Do not valueize the result of
198 (simplify_binary_expression): Likewise.
199 (simplify_unary_expression): Likewise.
201 2014-05-08 Richard Biener <rguenther@suse.de>
203 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
204 looking at TYPE_ARG_TYPES.
206 2014-05-08 Richard Biener <rguenther@suse.de>
208 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
209 pointer propagation special-case.
211 2014-05-08 Bin Cheng <bin.cheng@arm.com>
213 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
214 core part of address expressions.
216 2014-05-08 Alan Modra <amodra@gmail.com>
219 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
220 loads and stores when -mno-strict-align at any alignment.
221 (expand_block_clear): Similarly. Also correct calculation of
224 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
227 * tree-complex.c (expand_complex_move): Keep line info when expanding
229 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
230 of complex expression. Use new argument to display correct location
231 for values coming from phi statement.
232 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
233 (warn_uninitialized_phi): Pass location of phi argument to
235 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
236 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
238 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
240 * config/rs6000/predicates.md (indexed_address_mem): New.
241 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
242 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
243 fpstore_ux, fpstore_u.
244 (sign_extend, indexed, update): New.
245 (cell_micro): Adjust.
246 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
247 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
248 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
249 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
250 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
251 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
252 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
253 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
254 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
255 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
256 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
257 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
258 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
259 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
260 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
262 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
263 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
264 *vsx_extract_<mode>_store): Adjust.
265 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
266 is_cracked_insn, insn_must_be_first_in_group,
267 insn_must_be_last_in_group): Adjust.
269 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
271 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
272 ppc440-fpstore): Adjust.
273 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
274 ppc476-fpstore): Adjust.
275 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
276 ppc601-fpstore): Adjust.
277 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
279 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
281 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
282 ppc7450-fpstore): Adjust.
283 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
284 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
285 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
287 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
288 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
289 cell-fpstore, cell-fpstore-update): Adjust.
290 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
291 ppce300c3_store, ppce300c3_fpstore): Adjust.
292 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
293 e500mc_fpstore): Adjust.
294 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
295 e500mc64_store, e500mc64_fpstore): Adjust.
296 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
297 e5500_fpstore): Adjust.
298 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
299 e6500_fpstore): Adjust.
300 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
302 * config/rs6000/power4.md (power4-load, power4-load-ext,
303 power4-load-ext-update, power4-load-ext-update-indexed,
304 power4-load-update-indexed, power4-load-update, power4-fpload,
305 power4-fpload-update, power4-store, power4-store-update,
306 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
308 * config/rs6000/power5.md (power5-load, power5-load-ext,
309 power5-load-ext-update, power5-load-ext-update-indexed,
310 power5-load-update-indexed, power5-load-update, power5-fpload,
311 power5-fpload-update, power5-store, power5-store-update,
312 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
314 * config/rs6000/power6.md (power6-load, power6-load-ext,
315 power6-load-update, power6-load-update-indexed,
316 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
317 power6-fpload-update, power6-store, power6-store-update,
318 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
320 * config/rs6000/power7.md (power7-load, power7-load-ext,
321 power7-load-update, power7-load-update-indexed,
322 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
323 power7-fpload-update, power7-store, power7-store-update,
324 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
326 * config/rs6000/power8.md (power8-load, power8-load-update,
327 power8-load-ext, power8-load-ext-update, power8-fpload,
328 power8-fpload-update, power8-store, power8-store-update-indexed,
329 power8-fpstore, power8-fpstore-update): Adjust.
330 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
332 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
333 titan_lsu_store, titan_lsu_fpstore): Adjust.
334 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
336 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
339 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
340 unrolled byte insns. Emit address increments after move insns.
342 2014-05-07 David Malcolm <dmalcolm@redhat.com>
344 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
345 const_gimple, rather than a gimple.
346 (gimple_call_builtin_p): Likewise, for the three variants.
348 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
349 (gimple_call_builtin_p): Likewise, for the three variants.
351 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
353 PR tree-optimization/61095
354 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
356 2014-05-07 Richard Biener <rguenther@suse.de>
358 PR tree-optimization/61034
359 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
360 (maybe_skip_until): Use translate to take into account
361 lattices when trying to do disambiguations.
362 (get_continuation_for_phi_1): Likewise.
363 (get_continuation_for_phi): Adjust for added translate arguments.
364 (walk_non_aliased_vuses): Likewise.
365 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
366 (walk_non_aliased_vuses): Likewise.
367 (call_may_clobber_ref_p_1): Declare.
368 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
369 calls. Stop early if we are only supposed to disambiguate.
370 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
372 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
374 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
375 Emit an error when the function has arguments.
377 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
379 * cfgloop.h (unswitch_loops): Remove.
380 * doc/passes.texi: Remove references to loop-unswitch.c
381 * timevar.def (TV_LOOP_UNSWITCH): Remove.
383 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
385 * tree-vect-data-refs.c (vect_grouped_load_supported): New
386 check for loads group of length 3.
387 (vect_permute_load_chain): New permutations for loads group of
389 * tree-vect-stmts.c (vect_model_load_cost): Change cost
390 of vec_perm_shuffle for the new permutations.
392 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
394 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
395 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
396 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
397 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
398 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
399 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
400 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
401 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
403 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
405 * loop-unswitch.c: Delete.
407 2014-05-07 Richard Biener <rguenther@suse.de>
409 * config.gcc: Always set need_64bit_hwint to yes.
411 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
413 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
414 of using optimize_size.
416 2014-05-06 Mike Stump <mikestump@comcast.net>
418 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
420 2014-05-06 Joseph Myers <joseph@codesourcery.com>
422 * config/i386/sse.md (*mov<mode>_internal)
423 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
424 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
425 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
426 (*<code><mode>3, *andnot<mode>3<mask_name>)
427 (<mask_codefor><code><mode>3<mask_name>): Only consider
428 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
430 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
433 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
435 * lra-constraints.c (valid_address_p): Move earlier in file.
436 Add a constraint argument to the address_info version.
437 (satisfies_memory_constraint_p): New function.
438 (satisfies_address_constraint_p): Likewise.
439 (process_alt_operands, curr_insn_transform): Use them.
440 (process_address): Pass the constraint to valid_address_p when
441 checking address operands.
443 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
445 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
446 to their respective blocks. Fix inadvertent use of "node".
448 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
450 * emit-rtl.c (init_derived_machine_modes): New functionm, split
452 (init_emit_once): ...here.
453 * rtl.h (init_derived_machine_modes): Declare.
454 * toplev.c (do_compile): Call it even if no_backend.
456 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
457 Mike Stump <mikestump@comcast.net>
458 Richard Sandiford <rdsandiford@googlemail.com>
459 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
461 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
462 (rtx_equal_for_memref_p): Update comment.
463 (adjust_offset_for_component_ref): Use wide-int interfaces.
464 * builtins.c (get_object_alignment_2): Likewise.
465 (c_readstr): Likewise.
466 (target_char_cast): Add comment.
467 (determine_block_size): Use wide-int interfaces.
468 (expand_builtin_signbit): Likewise.
469 (fold_builtin_int_roundingfn): Likewise.
470 (fold_builtin_bitop): Likewise.
471 (fold_builtin_bswap): Likewise.
472 (fold_builtin_logarithm): Use signop.
473 (fold_builtin_pow): Likewise.
474 (fold_builtin_memory_op): Use wide-int interfaces.
475 (fold_builtin_object_size): Likewise.
476 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
477 nb_iterations_estimate.
478 (record_niter_bound): Use wide-int interfaces.
479 (get_estimated_loop_iterations_int): Likewise.
480 (get_estimated_loop_iterations): Likewise.
481 (get_max_loop_iterations): Likewise.
482 * cfgloop.h: Include wide-int.h.
483 (struct nb_iter_bound): Change bound to widest_int.
484 (struct loop): Change nb_iterations_upper_bound and
485 nb_iterations_estimate to widest_int.
486 (record_niter_bound): Switch to use widest_int.
487 (get_estimated_loop_iterations): Likewise.
488 (get_max_loop_iterations): Likewise.
489 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
491 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
492 * combine.c (try_combine): Likewise.
493 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
494 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
496 (aarch64_float_const_representable_p): Likewise.
497 * config/arc/arc.c: Include wide-int.h.
498 (arc_can_use_doloop_p): Use wide-int interfaces.
499 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
500 (vfp3_const_double_index): Likewise.
501 * config/avr/avr.c (avr_out_round): Likewise.
502 (avr_fold_builtin): Likewise.
503 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
504 (bfin_can_use_doloop_p): Likewise.
505 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
506 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
507 * config/i386/i386.c: Include wide-int.h.
508 (ix86_data_alignment): Use wide-int interfaces.
509 (ix86_local_alignment): Likewise.
510 (ix86_emit_swsqrtsf): Update real_from_integer.
511 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
512 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
513 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
514 (zero_constant): Likewise.
515 (input_operand): Likewise.
516 (splat_input_operand): Likewise.
517 (non_logical_cint_operand): Change const_double to const_wide_int.
518 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
519 (easy_altivec_constant): Remove comment.
520 (paired_expand_vector_init): Use CONSTANT_P.
521 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
522 (rs6000_emit_move): Update checks.
523 (rs6000_aggregate_candidate): Use wide-int interfaces.
524 (rs6000_expand_ternop_builtin): Likewise.
525 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
526 (rs6000_assemble_integer): Likewise.
527 (rs6000_hash_constant): Likewise.
528 (output_toc): Likewise.
529 (rs6000_rtx_costs): Likewise.
530 (rs6000_emit_swrsqrt); Update call to real_from_integer.
531 * config/rs6000/rs6000-c.c: Include wide-int.h.
532 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
533 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
534 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
535 Handle CONST_WIDE_INT.
536 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
537 Use tree_fits_uhwi_p.
538 * config/sparc/sparc.c: Include wide-int.h.
539 (sparc_fold_builtin): Use wide-int interfaces.
540 * config/vax/vax.c: Include wide-int.h.
541 (vax_float_literal): Use real_from_integer.
542 * coretypes.h (struct hwivec_def): New.
545 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
546 (equiv_constant): Handle CONST_WIDE_INT.
547 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
548 (cselib_hash_rtx): Handle CONST_WIDE_INT.
549 * dbxout.c (stabstr_U): Use wide-int interfaces.
550 (dbxout_type): Update to use cst_fits_shwi_p.
551 * defaults.h (LOG2_BITS_PER_UNIT): Define.
552 (TARGET_SUPPORTS_WIDE_INT): Add default.
553 * dfp.c: Include wide-int.h.
554 (decimal_real_to_integer2): Use wide-int interfaces and rename to
555 decimal_real_to_integer.
556 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
557 decimal_real_to_integer.
558 * doc/generic.texi (Constant expressions): Update for wide_int.
559 * doc/rtl.texi (const_double): Likewise.
560 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
561 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
562 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
563 (REAL_VALUE_FROM_INT): Remove.
564 (TARGET_SUPPORTS_WIDE_INT): New.
565 * doc/tm.texi: Regenerate.
566 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
567 * double-int.h: Include wide-int.h.
568 (struct wi::int_traits): New.
569 * dwarf2out.c (get_full_len): New.
570 (dw_val_equal_p): Add case dw_val_class_wide_int.
571 (size_of_loc_descr): Likewise.
572 (output_loc_operands): Likewise.
573 (insert_double): Remove.
574 (insert_wide_int): New.
576 (print_die): Add case dw_val_class_wide_int.
577 (attr_checksum): Likewise.
578 (attr_checksum_ordered): Likewise.
579 (same_dw_val_p): Likewise.
580 (size_of_die): Likewise.
581 (value_format): Likewise.
582 (output_die): Likewise.
583 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
585 (clz_loc_descriptor): Use wide-int interfaces.
586 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
587 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
588 (round_up_to_align): Use wide-int interfaces.
589 (field_byte_offset): Likewise.
590 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
591 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
592 CONST_DOUBLE handling. Use wide-int interfaces.
593 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
594 (gen_enumeration_type_die): Use add_AT_wide.
595 (hash_loc_operands): Add case dw_val_class_wide_int.
596 (compare_loc_operands): Likewise.
597 * dwarf2out.h: Include wide-int.h.
599 (enum dw_val_class): Add dw_val_class_wide_int.
600 (struct dw_val_struct): Add val_wide.
601 * emit-rtl.c (const_wide_int_htab): New.
602 (const_wide_int_htab_hash): New.
603 (const_wide_int_htab_eq): New.
604 (lookup_const_wide_int): New.
605 (const_double_htab_hash): Use wide-int interfaces.
606 (const_double_htab_eq): Likewise.
607 (rtx_to_double_int): Conditionally compile for wide-int.
608 (immed_double_int_const): Rename to immed_wide_int_const and
610 (immed_double_const): Conditionally compile for wide-int.
611 (init_emit_once): Use wide-int interfaces.
612 * explow.c (plus_constant): Likewise.
613 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
614 (lshift_value): Use wide-int interfaces.
615 (expand_mult): Likewise.
616 (choose_multiplier): Likewise.
617 (expand_smod_pow2): Likewise.
618 (make_tree): Likewise.
619 * expr.c (convert_modes): Consolidate handling of constants.
620 Use wide-int interfaces.
621 (emit_group_load_1): Add note.
622 (store_expr): Update comment.
623 (get_inner_reference): Use wide-int interfaces.
624 (expand_constructor): Update comment.
625 (expand_expr_real_2): Use wide-int interfaces.
626 (expand_expr_real_1): Likewise.
627 (reduce_to_bit_field_precision): Likewise.
628 (const_vector_from_tree): Likewise.
629 * final.c: Include wide-int-print.h.
630 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
631 * fixed-value.c: Include wide-int.h.
632 (fixed_from_string): Use wide-int interfaces.
633 (fixed_to_decimal): Likewise.
634 (fixed_convert_from_real): Likewise.
635 (real_convert_from_fixed): Likewise.
636 * fold-const.h (mem_ref_offset): Return an offset_int.
637 (div_if_zero_remainder): Remove code parameter.
638 * fold-const.c (div_if_zero_remainder): Remove code parameter.
639 Use wide-int interfaces.
640 (may_negate_without_overflow_p): Use wide-int interfaces.
641 (negate_expr_p): Likewise.
642 (fold_negate_expr): Likewise.
643 (int_const_binop_1): Likewise.
644 (const_binop): Likewise.
645 (fold_convert_const_int_from_int): Likewise.
646 (fold_convert_const_int_from_real): Likewise.
647 (fold_convert_const_int_from_fixed): Likewise.
648 (fold_convert_const_fixed_from_int): Likewise.
649 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
650 (sign_bit_p): Use wide-int interfaces.
651 (make_range_step): Likewise.
652 (build_range_check): Likewise. Pass an integer of the correct type
653 instead of using integer_one_node.
654 (range_predecessor): Pass an integer of the correct type instead
655 of using integer_one_node.
656 (range_successor): Likewise.
657 (merge_ranges): Likewise.
658 (unextend): Use wide-int interfaces.
659 (extract_muldiv_1): Likewise.
660 (fold_div_compare): Likewise.
661 (fold_single_bit_test): Likewise.
662 (fold_sign_changed_comparison): Likewise.
663 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
664 (fold_plusminus_mult_expr): Use wide-int interfaces.
665 (native_encode_int): Likewise.
666 (native_interpret_int): Likewise.
667 (fold_unary_loc): Likewise.
668 (pointer_may_wrap_p): Likewise.
669 (size_low_cst): Likewise.
670 (mask_with_tz): Likewise.
671 (fold_binary_loc): Likewise.
672 (fold_ternary_loc): Likewise.
673 (multiple_of_p): Likewise.
674 (tree_call_nonnegative_warnv_p): Update calls to
675 tree_int_cst_min_precision and real_from_integer.
676 (fold_negate_const): Use wide-int interfaces.
677 (fold_abs_const): Likewise.
678 (fold_relational_const): Use tree_int_cst_lt.
679 (round_up_loc): Use wide-int interfaces.
680 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
681 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
682 * gengtype.c: Remove include of double-int.h.
683 (do_typedef): Use wide-int interfaces.
684 (open_base_files): Add wide-int.h.
685 (main): Add offset_int and widest_int typedefs.
686 * gengtype-lex.l: Handle "^".
687 (CXX_KEYWORD): Add "static".
688 * gengtype-parse.c (require3): New.
689 (require_template_declaration): Handle constant template arguments
690 and nested templates.
691 * gengtype-state.c: Don't include "double-int.h".
692 * genpreds.c (write_one_predicate_function): Update comment.
693 (write_tm_constrs_h): Add check for hval and lval use in
695 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
696 (add_to_sequence): Likewise.
697 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
698 and const_double_operand.
699 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
701 * gimple-fold.c (get_base_constructor): Likewise.
702 (fold_array_ctor_reference): Likewise.
703 (fold_nonarray_ctor_reference): Likewise.
704 (fold_const_aggregate_ref_1): Likewise.
705 (gimple_val_nonnegative_real_p): Likewise.
706 (gimple_fold_indirect_ref): Likewise.
707 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
708 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
709 (struct slsr_cand_d): Change index to be widest_int.
710 (struct incr_info_d): Change incr to be widest_int.
711 (alloc_cand_and_find_basis): Use wide-int interfaces.
712 (slsr_process_phi): Likewise.
713 (backtrace_base_for_ref): Likewise. Return a widest_int.
714 (restructure_reference): Take a widest_int instead of a double_int.
715 (slsr_process_ref): Use wide-int interfaces.
716 (create_mul_ssa_cand): Likewise.
717 (create_mul_imm_cand): Likewise.
718 (create_add_ssa_cand): Likewise.
719 (create_add_imm_cand): Take a widest_int instead of a double_int.
720 (slsr_process_add): Use wide-int interfaces.
721 (slsr_process_cast): Likewise.
722 (slsr_process_copy): Likewise.
723 (dump_candidate): Likewise.
724 (dump_incr_vec): Likewise.
725 (replace_ref): Likewise.
726 (cand_increment): Likewise. Return a widest_int.
727 (cand_abs_increment): Likewise.
728 (replace_mult_candidate): Take a widest_int instead of a double_int.
729 (replace_unconditional_candidate): Use wide-int interfaces.
730 (incr_vec_index): Take a widest_int instead of a double_int.
731 (create_add_on_incoming_edge): Likewise.
732 (create_phi_basis): Use wide-int interfaces.
733 (replace_conditional_candidate): Likewise.
734 (record_increment): Take a widest_int instead of a double_int.
735 (record_phi_increments): Use wide-int interfaces.
736 (phi_incr_cost): Take a widest_int instead of a double_int.
737 (lowest_cost_path): Likewise.
738 (total_savings): Likewise.
739 (analyze_increments): Use wide-int interfaces.
740 (ncd_with_phi): Take a widest_int instead of a double_int.
741 (ncd_of_cand_and_phis): Likewise.
742 (nearest_common_dominator_for_cands): Likewise.
743 (insert_initializers): Use wide-int interfaces.
744 (all_phi_incrs_profitable): Likewise.
745 (replace_one_candidate): Likewise.
746 (replace_profitable_candidates): Likewise.
747 * godump.c: Include wide-int-print.h.
748 (go_output_typedef): Use wide-int interfaces.
749 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
750 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
751 (build_loop_iteration_domains): Likewise.
752 * hooks.h: Include wide-int.h rather than double-int.h.
753 (hook_bool_dint_dint_uint_bool_true): Delete.
754 (hook_bool_wint_wint_uint_bool_true): Declare.
755 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
756 (hook_bool_wint_wint_uint_bool_true): New.
757 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
759 (ubsan_expand_si_overflow_mul_check): Likewise.
760 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
761 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
762 (get_ancestor_addr_info): Likewise.
763 (ipa_modify_call_arguments): Likewise.
764 * loop-doloop.c (doloop_modify): Likewise.
765 (doloop_optimize): Likewise.
766 * loop-iv.c (iv_number_of_iterations): Likewise.
767 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
768 (unroll_loop_constant_iterations): Likewise.
769 (decide_unroll_runtime_iterations): Likewise.
770 (unroll_loop_runtime_iterations): Likewise.
771 (decide_peel_simple): Likewise.
772 (decide_unroll_stupid): Likewise.
773 * lto-streamer-in.c (streamer_read_wi): Add.
774 (input_cfg): Use wide-int interfaces.
775 (lto_input_tree_1): Likewise.
776 * lto-streamer-out.c (streamer_write_wi): Add.
777 (hash_tree): Use wide-int interfaces.
778 (output_cfg): Likewise.
779 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
780 (GTFILES): Add wide-int.h and signop.h.
781 (TAGS): Look for .cc files too.
782 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
783 * optabs.c (expand_subword_shift): Likewise.
784 (expand_doubleword_shift): Likewise.
785 (expand_absneg_bit): Likewise.
786 (expand_copysign_absneg): Likewise.
787 (expand_copysign_bit): Likewise.
788 * postreload.c (reload_cse_simplify_set): Likewise.
789 * predict.c (predict_iv_comparison): Likewise.
790 * pretty-print.h: Include wide-int-print.h.
792 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
793 * print-tree.c: Include wide-int-print.h.
794 (print_node_brief): Use wide-int interfaces.
795 (print_node): Likewise.
796 * read-rtl.c (validate_const_wide_int): New.
797 (read_rtx_code): Add CONST_WIDE_INT case.
798 * real.c: Include wide-int.h.
799 (real_to_integer2): Delete.
800 (real_to_integer): New function, returning a wide_int.
801 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
802 (ten_to_ptwo): Update call to real_from_integer.
803 (real_digit): Likewise.
804 * real.h: Include signop.h, wide-int.h and insn-modes.h.
805 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
806 (REAL_VALUE_TO_INT): Delete.
807 (real_to_integer): Declare a wide-int form.
808 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
809 * recog.c (const_int_operand): Improve comment.
810 (const_scalar_int_operand): New.
811 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
812 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
813 (split_double): Likewise.
814 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
815 (rtx_size): Likewise.
816 (rtx_alloc_stat_v): New.
817 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
818 (cwi_output_hex): New.
819 (iterative_hash_rtx): Handle CONST_WIDE_INT.
820 (cwi_check_failed_bounds): New.
821 * rtl.def (CONST_WIDE_INT): New.
822 * rtl.h: Include <utility> and wide-int.h.
823 (struct hwivec_def): New.
824 (CWI_GET_NUM_ELEM): New.
825 (CWI_PUT_NUM_ELEM): New.
826 (struct rtx_def): Add num_elem and hwiv.
827 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
828 (CASE_CONST_UNIQUE): Likewise.
829 (CASE_CONST_ANY): Likewise.
830 (CONST_SCALAR_INT_P): Likewise.
831 (CONST_WIDE_INT_P): New.
834 (cwi_check_failed_bounds): New.
837 (CONST_WIDE_INT_VEC) New.
838 (CONST_WIDE_INT_NUNITS) New.
839 (CONST_WIDE_INT_ELT) New.
840 (rtx_mode_t): New type.
841 (wi::int_traits <rtx_mode_t>): New.
843 (wi::min_value): New.
844 (wi::max_value): New.
846 (const_wide_int_alloc): New.
847 (immed_wide_int_const): New.
848 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
849 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
850 * signop.h: New file.
851 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
852 (simplify_const_unary_operation): Use wide-int interfaces.
853 (simplify_binary_operation_1): Likewise.
854 (simplify_const_binary_operation): Likewise.
855 (simplify_const_relational_operation): Likewise.
856 (simplify_immed_subreg): Likewise.
857 * stmt.c (expand_case): Likewise.
858 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
859 signop rather than a bool.
860 * stor-layout.c (layout_type): Use wide-int interfaces.
861 (initialize_sizetypes): Update calls to
862 set_min_and_max_values_for_integral_type.
863 (set_min_and_max_values_for_integral_type): Take a signop rather
864 than a bool. Use wide-int interfaces.
865 (fixup_signed_type): Update accordingly. Remove
866 HOST_BITS_PER_DOUBLE_INT limit.
867 (fixup_unsigned_type): Likewise.
868 * system.h (STATIC_CONSTANT_P): New.
869 (STATIC_ASSERT): New.
870 * target.def (can_use_doloop_p): Take widest_ints rather than
872 * target.h: Include wide-int.h rather than double-int.h.
873 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
875 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
876 rather than INT_CST_LT_UNSIGNED.
877 (can_use_doloop_if_innermost): Take widest_ints rather than
879 * tree-affine.c: Include wide-int-print.h.
880 (double_int_ext_for_comb): Delete.
881 (wide_int_ext_for_comb): New.
882 (aff_combination_zero): Use wide-int interfaces.
883 (aff_combination_const): Take a widest_int instead of a double_int.
884 (aff_combination_elt): Use wide-int interfaces.
885 (aff_combination_scale): Take a widest_int instead of a double_int.
886 (aff_combination_add_elt): Likewise.
887 (aff_combination_add_cst): Likewise.
888 (aff_combination_add): Use wide-int interfaces.
889 (aff_combination_convert): Likewise.
890 (tree_to_aff_combination): Likewise.
891 (add_elt_to_tree): Take a widest_int instead of a double_int.
892 (aff_combination_to_tree): Use wide-int interfaces.
893 (aff_combination_remove_elt): Likewise.
894 (aff_combination_add_product): Take a widest_int instead of
896 (aff_combination_mult): Use wide-int interfaces.
897 (aff_combination_expand): Likewise.
898 (double_int_constant_multiple_p): Delete.
899 (wide_int_constant_multiple_p): New.
900 (aff_combination_constant_multiple_p): Take a widest_int pointer
901 instead of a double_int pointer.
902 (print_aff): Use wide-int interfaces.
903 (get_inner_reference_aff): Take a widest_int pointer
904 instead of a double_int pointer.
905 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
906 * tree-affine.h: Include wide-int.h.
907 (struct aff_comb_elt): Change type of coef to widest_int.
908 (struct affine_tree_combination): Change type of offset to widest_int.
909 (double_int_ext_for_comb): Delete.
910 (wide_int_ext_for_comb): New.
911 (aff_combination_const): Use widest_int instead of double_int.
912 (aff_combination_scale): Likewise.
913 (aff_combination_add_elt): Likewise.
914 (aff_combination_constant_multiple_p): Likewise.
915 (get_inner_reference_aff): Likewise.
916 (aff_comb_cannot_overlap_p): Likewise.
917 (aff_combination_zero_p): Use wide-int interfaces.
918 * tree.c: Include tree.h.
919 (init_ttree): Use make_int_cst.
920 (tree_code_size): Removed code for INTEGER_CST case.
921 (tree_size): Add INTEGER_CST case.
922 (make_node_stat): Update comment.
923 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
924 (build_int_cst_type): Use wide-int interfaces.
925 (double_int_to_tree): Likewise.
926 (double_int_fits_to_tree_p): Delete.
927 (force_fit_type_double): Delete.
928 (force_fit_type): New.
929 (int_cst_hash_hash): Use wide-int interfaces.
930 (int_cst_hash_eq): Likewise.
931 (build_int_cst_wide): Delete.
932 (wide_int_to_tree): New.
933 (cache_integer_cst): Use wide-int interfaces.
934 (build_low_bits_mask): Likewise.
935 (cst_and_fits_in_hwi): Likewise.
936 (real_value_from_int_cst): Likewise.
937 (make_int_cst_stat): New.
938 (integer_zerop): Use wide_int interfaces.
939 (integer_onep): Likewise.
940 (integer_all_onesp): Likewise.
941 (integer_pow2p): Likewise.
942 (integer_nonzerop): Likewise.
943 (tree_log2): Likewise.
944 (tree_floor_log2): Likewise.
945 (tree_ctz): Likewise.
946 (int_size_in_bytes): Likewise.
947 (mem_ref_offset): Return an offset_int rather than a double_int.
948 (build_type_attribute_qual_variant): Use wide_int interfaces.
949 (type_hash_eq): Likewise
950 (tree_int_cst_equal): Likewise.
951 (tree_int_cst_lt): Delete.
952 (tree_int_cst_compare): Likewise.
953 (tree_fits_shwi_p): Use wide_int interfaces.
954 (tree_fits_uhwi_p): Likewise.
955 (tree_int_cst_sign_bit): Likewise.
956 (tree_int_cst_sgn): Likewise.
957 (tree_int_cst_min_precision): Take a signop rather than a bool.
958 (simple_cst_equal): Use wide_int interfaces.
959 (compare_tree_int): Likewise.
960 (iterative_hash_expr): Likewise.
961 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
963 (get_type_static_bounds): Use wide_int interfaces.
964 (tree_int_cst_elt_check_failed): New.
965 (build_common_tree_nodes): Reordered to set prec before filling in
967 (int_cst_value): Check cst_and_fits_in_hwi.
968 (widest_int_cst_value): Use wide_int interfaces.
969 (upper_bound_in_type): Likewise.
970 (lower_bound_in_type): Likewise.
971 (num_ending_zeros): Likewise.
972 (drop_tree_overflow): Likewise.
973 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
974 (gen_conditions_for_pow_cst_base): Likewise.
975 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
976 (group_case_labels_stmt): Use wide-int interfaces.
977 (verify_gimple_assign_binary): Likewise.
978 (print_loop): Likewise.
979 * tree-chrec.c (tree_fold_binomial): Likewise.
980 * tree-core.h (struct tree_base): Add int_length.
981 (struct tree_int_cst): Change rep of value.
982 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
983 (dr_may_alias_p): Likewise.
984 (max_stmt_executions_tree): Likewise.
985 * tree.def (INTEGER_CST): Update comment.
986 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
987 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
988 * tree-dump.c: Include wide-int.h and wide-int-print.h.
989 (dequeue_and_dump): Use wide-int interfaces.
990 * tree.h: Include wide-int.h.
991 (NULL_TREE): Moved to earlier loc in file.
992 (TREE_INT_CST_ELT_CHECK): New.
993 (tree_int_cst_elt_check_failed): New.
995 (TREE_INT_CST): Delete.
996 (TREE_INT_CST_LOW): Use wide-int interfaces.
997 (TREE_INT_CST_HIGH): Delete.
998 (TREE_INT_CST_NUNITS): New.
999 (TREE_INT_CST_EXT_NUNITS): Likewise.
1000 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
1001 (TREE_INT_CST_ELT): Likewise.
1002 (INT_CST_LT): Delete.
1003 (tree_int_cst_elt_check): New (two forms).
1004 (type_code_size): Update comment.
1005 (make_int_cst_stat, make_int_cst): New.
1006 (tree_to_double_int): Delete.
1007 (double_int_fits_to_tree_p): Delete.
1008 (force_fit_type_double): Delete.
1009 (build_int_cstu): Replace with out-of-line function.
1010 (build_int_cst_wide): Delete.
1011 (tree_int_cst_lt): Define inline.
1012 (tree_int_cst_le): New.
1013 (tree_int_cst_compare): Define inline.
1014 (tree_int_cst_min_precision): Take a signop rather than a bool.
1015 (wi::int_traits <const_tree>): New.
1016 (wi::int_traits <tree>): New.
1017 (wi::extended_tree): New.
1018 (wi::int_traits <wi::extended_tree>): New.
1019 (wi::to_widest): New.
1020 (wi::to_offset): New.
1021 (wi::fits_to_tree_p): New.
1022 (wi::min_value): New.
1023 (wi::max_value): New.
1024 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
1025 (copy_tree_body_r): Likewise.
1026 * tree-object-size.c (compute_object_offset): Likewise.
1027 (addr_object_size): Likewise.
1028 * tree-predcom.c: Include wide-int-print.h.
1029 (struct dref_d): Change type of offset to widest_int.
1030 (dump_dref): Call wide-int printer.
1031 (aff_combination_dr_offset): Use wide-int interfaces.
1032 (determine_offset): Take a widest_int pointer rather than a
1034 (split_data_refs_to_components): Use wide-int interfaces.
1035 (suitable_component_p): Likewise.
1036 (order_drefs): Likewise.
1037 (add_ref_to_chain): Likewise.
1038 (valid_initializer_p): Likewise.
1039 (determine_roots_comp): Likewise.
1040 * tree-pretty-print.c: Include wide-int-print.h.
1041 (dump_generic_node): Use wide-int interfaces.
1042 * tree-sra.c (sra_ipa_modify_expr): Likewise.
1043 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
1044 (move_fixed_address_to_symbol): Likewise.
1045 (move_hint_to_base): Likewise.
1046 (move_pointer_to_base): Likewise.
1047 (move_variant_to_index): Likewise.
1048 (most_expensive_mult_to_index): Likewise.
1049 (addr_to_parts): Likewise.
1050 (copy_ref_info): Likewise.
1051 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
1052 (indirect_refs_may_alias_p): Likewise.
1053 (stmt_kills_ref_p_1): Likewise.
1054 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
1055 * tree-ssa-ccp.c: Update comment at top of file. Include
1057 (struct prop_value_d): Change type of mask to widest_int.
1058 (extend_mask): New function.
1059 (dump_lattice_value): Use wide-int interfaces.
1060 (get_default_value): Likewise.
1061 (set_constant_value): Likewise.
1062 (set_value_varying): Likewise.
1063 (valid_lattice_transition): Likewise.
1064 (set_lattice_value): Likewise.
1065 (value_to_double_int): Delete.
1066 (value_to_wide_int): New.
1067 (get_value_from_alignment): Use wide-int interfaces.
1068 (get_value_for_expr): Likewise.
1069 (do_dbg_cnt): Likewise.
1070 (ccp_finalize): Likewise.
1071 (ccp_lattice_meet): Likewise.
1072 (bit_value_unop_1): Use widest_ints rather than double_ints.
1073 (bit_value_binop_1): Likewise.
1074 (bit_value_unop): Use wide-int interfaces.
1075 (bit_value_binop): Likewise.
1076 (bit_value_assume_aligned): Likewise.
1077 (evaluate_stmt): Likewise.
1078 (ccp_fold_stmt): Likewise.
1079 (visit_cond_stmt): Likewise.
1080 (ccp_visit_stmt): Likewise.
1081 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
1082 (constant_pointer_difference): Likewise.
1083 (associate_pointerplus): Likewise.
1084 (combine_conversions): Likewise.
1085 * tree-ssa-loop.h: Include wide-int.h.
1086 (struct tree_niter_desc): Change type of max to widest_int.
1087 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
1088 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
1089 (remove_redundant_iv_tests): Likewise.
1090 (canonicalize_loop_induction_variables): Likewise.
1091 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
1092 (constant_multiple_of): Take a widest_int pointer instead of
1093 a double_int pointer.
1094 (get_computation_aff): Use wide-int interfaces.
1095 (ptr_difference_cost): Likewise.
1096 (difference_cost): Likewise.
1097 (get_loop_invariant_expr_id): Likewise.
1098 (get_computation_cost_at): Likewise.
1099 (iv_elimination_compare_lt): Likewise.
1100 (may_eliminate_iv): Likewise.
1101 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
1102 instead of double_int.
1103 (max_loop_iterations): Likewise.
1104 (max_stmt_executions): Likewise.
1105 (estimated_stmt_executions): Likewise.
1106 * tree-ssa-loop-niter.c: Include wide-int-print.h.
1107 (split_to_var_and_offset): Use wide-int interfaces.
1108 (determine_value_range): Likewise.
1109 (bound_difference_of_offsetted_base): Likewise.
1110 (bounds_add): Take a widest_int instead of a double_int.
1111 (number_of_iterations_ne_max): Use wide-int interfaces.
1112 (number_of_iterations_ne): Likewise.
1113 (number_of_iterations_lt_to_ne): Likewise.
1114 (assert_loop_rolls_lt): Likewise.
1115 (number_of_iterations_lt): Likewise.
1116 (number_of_iterations_le): Likewise.
1117 (number_of_iterations_cond): Likewise.
1118 (number_of_iterations_exit): Likewise.
1119 (finite_loop_p): Likewise.
1120 (derive_constant_upper_bound_assign): Likewise.
1121 (derive_constant_upper_bound): Return a widest_int.
1122 (derive_constant_upper_bound_ops): Likewise.
1123 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
1124 (record_estimate): Take a widest_int rather than a double_int.
1125 (record_nonwrapping_iv): Use wide-int interfaces.
1126 (double_int_cmp): Delete.
1127 (wide_int_cmp): New.
1128 (bound_index): Take a widest_int rather than a double_int.
1129 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
1130 (maybe_lower_iteration_bound): Likewise.
1131 (estimate_numbers_of_iterations_loop): Likewise.
1132 (estimated_loop_iterations): Take a widest_int pointer than than
1133 a double_int pointer.
1134 (estimated_loop_iterations_int): Use wide-int interfaces.
1135 (max_loop_iterations): Take a widest_int pointer than than
1136 a double_int pointer.
1137 (max_loop_iterations_int): Use wide-int interfaces.
1138 (max_stmt_executions): Take a widest_int pointer than than
1139 a double_int pointer.
1140 (estimated_stmt_executions): Likewise.
1141 (n_of_executions_at_most): Use wide-int interfaces.
1142 (scev_probably_wraps_p): Likewise.
1143 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
1145 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
1147 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
1148 double_ints. Adjust for trailing_wide_ints <3> representation.
1149 (set_nonzero_bits): Likewise.
1150 (get_range_info): Return wide_ints rather than double_ints.
1151 Adjust for trailing_wide_ints <3> representation.
1152 (get_nonzero_bits): Likewise.
1153 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
1155 * tree-ssanames.h (struct range_info_def): Replace min, max and
1156 nonzero_bits with a trailing_wide_ints <3>.
1157 (set_range_info): Use wide_int_refs rather than double_ints.
1158 (set_nonzero_bits): Likewise.
1159 (get_range_info): Return wide_ints rather than double_ints.
1160 (get_nonzero_bits): Likewise.
1161 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
1162 * tree-ssa-pre.c (phi_translate_1): Likewise.
1163 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
1164 (acceptable_pow_call): Likewise.
1165 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
1167 (vn_reference_fold_indirect): Likewise.
1168 (vn_reference_maybe_forwprop_address): Likewise.
1169 (valueize_refs_1): Likewise.
1170 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
1171 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
1172 tree_int_cst_lt and tree_int_cst_le.
1173 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
1175 (streamer_alloc_tree): Likewise.
1176 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
1177 (streamer_write_tree_header): Likewise.
1178 (streamer_write_integer_cst): Likewise.
1179 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
1180 (build_constructors): Likewise.
1181 (array_value_type): Likewise.
1182 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
1183 (vect_check_gather): Likewise.
1184 * tree-vect-generic.c (build_replicated_const): Likewise.
1185 (expand_vector_divmod): Likewise.
1186 * tree-vect-loop.c (vect_transform_loop): Likewise.
1187 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
1188 (vect_do_peeling_for_alignment): Likewise.
1189 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
1190 * tree-vrp.c: Include wide-int.h.
1191 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
1192 (extract_range_from_assert): Use wide-int interfaces.
1193 (vrp_int_const_binop): Likewise.
1194 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
1195 double_int pointers.
1196 (ranges_from_anti_range): Use wide-int interfaces.
1197 (quad_int_cmp): Delete.
1198 (quad_int_pair_sort): Likewise.
1199 (extract_range_from_binary_expr_1): Use wide-int interfaces.
1200 (extract_range_from_unary_expr_1): Likewise.
1201 (adjust_range_with_scev): Likewise.
1202 (masked_increment): Take and return wide_ints rather than double_ints.
1203 (register_edge_assert_for_2): Use wide-int interfaces.
1204 (check_array_ref): Likewise.
1205 (search_for_addr_array): Likewise.
1206 (maybe_set_nonzero_bits): Likewise.
1207 (union_ranges): Pass an integer of the correct type instead of
1208 using integer_one_node.
1209 (intersect_ranges): Likewise.
1210 (simplify_truth_ops_using_ranges): Likewise.
1211 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
1212 (range_fits_type_p): Likewise.
1213 (simplify_cond_using_ranges): Likewise. Take a signop rather than
1215 (simplify_conversion_using_ranges): Use wide-int interfaces.
1216 (simplify_float_conversion_using_ranges): Likewise.
1217 (vrp_finalize): Likewise.
1218 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
1219 (gimple_stringops_transform): Likewise.
1220 * varasm.c (decode_addr_const): Likewise.
1221 (const_hash_1): Likewise.
1222 (const_rtx_hash_1): Likewise
1223 (output_constant): Likewise.
1224 (array_size_for_constructor): Likewise.
1225 (output_constructor_regular_field): Likewise.
1226 (output_constructor_bitfield): Likewise.
1227 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
1228 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
1230 * gencheck.c: Define BITS_PER_UNIT.
1233 * wide-int-print.cc: New.
1234 * wide-int-print.h: New.
1236 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1238 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
1240 2014-05-06 Richard Biener <rguenther@suse.de>
1242 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
1243 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
1244 (TODO_verify_all): Adjust.
1245 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
1246 TODO_verify_stmts and TODO_verify_rtl_sharing.
1247 * bb-reorder.c: Likewise.
1248 * cfgexpand.c: Likewise.
1249 * cprop.c: Likewise.
1251 * function.c: Likewise.
1252 * fwprop.c: Likewise.
1254 * gimple-ssa-isolate-paths.c: Likewise.
1255 * gimple-ssa-strength-reduction.c: Likewise.
1256 * ipa-split.c: Likewise.
1257 * loop-init.c: Likewise.
1258 * loop-unroll.c: Likewise.
1259 * lower-subreg.c: Likewise.
1260 * modulo-sched.c: Likewise.
1261 * postreload-gcse.c: Likewise.
1262 * predict.c: Likewise.
1263 * recog.c: Likewise.
1264 * sched-rgn.c: Likewise.
1265 * store-motion.c: Likewise.
1266 * tracer.c: Likewise.
1267 * trans-mem.c: Likewise.
1268 * tree-call-cdce.c: Likewise.
1269 * tree-cfg.c: Likewise.
1270 * tree-cfgcleanup.c: Likewise.
1271 * tree-complex.c: Likewise.
1272 * tree-eh.c: Likewise.
1273 * tree-emutls.c: Likewise.
1274 * tree-if-conv.c: Likewise.
1275 * tree-into-ssa.c: Likewise.
1276 * tree-loop-distribution.c: Likewise.
1277 * tree-object-size.c: Likewise.
1278 * tree-parloops.c: Likewise.
1279 * tree-pass.h: Likewise.
1280 * tree-sra.c: Likewise.
1281 * tree-ssa-ccp.c: Likewise.
1282 * tree-ssa-copy.c: Likewise.
1283 * tree-ssa-copyrename.c: Likewise.
1284 * tree-ssa-dce.c: Likewise.
1285 * tree-ssa-dom.c: Likewise.
1286 * tree-ssa-dse.c: Likewise.
1287 * tree-ssa-forwprop.c: Likewise.
1288 * tree-ssa-ifcombine.c: Likewise.
1289 * tree-ssa-loop-ch.c: Likewise.
1290 * tree-ssa-loop-ivcanon.c: Likewise.
1291 * tree-ssa-loop.c: Likewise.
1292 * tree-ssa-math-opts.c: Likewise.
1293 * tree-ssa-phiopt.c: Likewise.
1294 * tree-ssa-phiprop.c: Likewise.
1295 * tree-ssa-pre.c: Likewise.
1296 * tree-ssa-reassoc.c: Likewise.
1297 * tree-ssa-sink.c: Likewise.
1298 * tree-ssa-strlen.c: Likewise.
1299 * tree-ssa-tail-merge.c: Likewise.
1300 * tree-ssa-uncprop.c: Likewise.
1301 * tree-switch-conversion.c: Likewise.
1302 * tree-tailcall.c: Likewise.
1303 * tree-vect-generic.c: Likewise.
1304 * tree-vectorizer.c: Likewise.
1305 * tree-vrp.c: Likewise.
1307 * var-tracking.c: Likewise.
1308 * bt-load.c: Likewise.
1309 * cfgcleanup.c: Likewise.
1310 * combine-stack-adj.c: Likewise.
1311 * combine.c: Likewise.
1312 * compare-elim.c: Likewise.
1313 * config/epiphany/resolve-sw-modes.c: Likewise.
1314 * config/i386/i386.c: Likewise.
1315 * config/mips/mips.c: Likewise.
1316 * config/s390/s390.c: Likewise.
1317 * config/sh/sh_treg_combine.cc: Likewise.
1318 * config/sparc/sparc.c: Likewise.
1321 * final.c: Likewise.
1322 * ifcvt.c: Likewise.
1323 * mode-switching.c: Likewise.
1324 * passes.c: Likewise.
1325 * postreload.c: Likewise.
1327 * reg-stack.c: Likewise.
1328 * regcprop.c: Likewise.
1329 * regrename.c: Likewise.
1332 2014-05-06 Richard Biener <rguenther@suse.de>
1335 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
1336 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
1338 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
1341 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
1343 2014-05-05 Radovan Obradovic <robradovic@mips.com>
1344 Tom de Vries <tom@codesourcery.com>
1346 * target.def (call_fusage_contains_non_callee_clobbers): New
1348 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
1350 (@node Miscellaneous Register Hooks): New node.
1351 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
1352 * doc/tm.texi: Regenerate.
1354 2014-05-05 Marek Polacek <polacek@redhat.com>
1357 * opts.c (common_handle_option): Call error_at instead of warning_at.
1359 2014-05-05 Richard Biener <rguenther@suse.de>
1361 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
1362 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
1363 under the TODO_verify_il umbrella.
1365 2014-05-05 Richard Biener <rguenther@suse.de>
1367 * passes.c (execute_function_todo): Move TODO_verify_flow under
1368 the TODO_verify_ul umbrella.
1370 2014-05-05 Richard Biener <rguenther@suse.de>
1373 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
1374 X & CST away from a CST that is the mask of a mode.
1376 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1378 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
1379 int argument to enum machine_mode.
1380 (picochip_class_max_nregs): Ditto.
1381 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
1382 (picochip_class_max_nregs): Ditto.
1384 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1386 * target.def: Add new target hook.
1387 * doc/tm.texi: Regenerate.
1388 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
1389 * targhooks.c (default_keep_leaf_when_profiled): New function.
1391 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
1392 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
1394 2014-05-05 Bin Cheng <bin.cheng@arm.com>
1396 PR tree-optimization/60363
1397 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
1398 (copy_phi_args): New parameters. Call get_value_locus_in_path.
1399 (update_destination_phis): New parameter.
1400 (create_edge_and_update_destination_phis): Ditto.
1401 (ssa_fix_duplicate_block_edges): Pass new arguments.
1402 (thread_single_edge): Ditto.
1404 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
1406 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
1407 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
1408 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
1409 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
1410 Use RS6000_BTM_HARD_FLOAT.
1411 (BU_MISC_2): Likewise.
1412 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
1413 RS6000_BTM_HARD_FLOAT.
1414 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
1416 (rs6000_invalid_builtin): Add hard floating builtin support.
1417 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
1418 hard float builtins.
1419 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
1421 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1423 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
1424 Add missing function* argument.
1426 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1428 * lra-constraints.c (valid_address_p): Move earlier in file.
1429 Add a constraint argument to the address_info version.
1430 (satisfies_memory_constraint_p): New function.
1431 (satisfies_address_constraint_p): Likewise.
1432 (process_alt_operands, curr_insn_transform): Use them.
1433 (process_address): Pass the constraint to valid_address_p when
1434 checking address operands.
1436 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1438 * config/mips/mips.c (mips_isa_rev): New variable.
1439 (mips_set_architecture): Set it.
1440 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1442 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1443 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1444 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1445 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1446 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
1447 conditions in terms of mips_isa_rev.
1448 (mips_isa_rev): Declare.
1450 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1452 * config/sh/sh-mem.cc: Use tabs instead of spaces.
1453 (prob_unlikely, prob_likely): Make variables const.
1455 2014-05-03 Denis Chertykov <chertykov@gmail.com>
1457 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
1459 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1461 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
1463 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1465 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
1466 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
1467 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
1469 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
1470 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
1472 Replace usage of ROUND_REG with sh_round_reg.
1473 Use CEIL instead of ROUND_ADVANCE.
1475 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1478 * config/sh/sh.c: Include stdlib headers before everything else.
1480 2014-05-02 Jakub Jelinek <jakub@redhat.com>
1482 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
1483 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
1484 (gimplify_adjust_omp_clauses): Simd region is never
1485 directly nested in combined parallel. Instead, for linear
1486 with copyin/copyout, if in combined for simd loop, make decl
1487 firstprivate/lastprivate on OMP_FOR.
1488 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
1489 expand_omp_for_static_chunk): When setting endvar, also set
1490 fd->loop.v to the same value.
1492 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1494 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
1496 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
1498 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
1501 2014-05-02 Marek Polacek <polacek@redhat.com>
1503 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
1505 2014-05-02 Kito Cheng <kito@0xlab.org>
1507 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
1508 to a C expression marco.
1509 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
1510 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
1511 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
1512 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
1513 HONOR_REG_ALLOC_ORDER.
1514 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
1516 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1518 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
1520 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1522 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
1524 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
1526 * tree-if-conv.c (is_cond_scalar_reduction): New function.
1527 (convert_scalar_cond_reduction): Likewise.
1528 (predicate_scalar_phi): Add recognition and transformation
1529 of simple conditioanl reduction to be vectorizable.
1531 2014-05-01 Marek Polacek <polacek@redhat.com>
1534 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
1536 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
1538 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
1539 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
1540 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
1541 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
1542 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
1543 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
1544 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
1545 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
1547 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
1549 * config/arc/arc.opt (mlra): Move comment above option name
1550 to avoid mis-parsing as language options.
1552 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1554 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
1555 * config/sol2.h: ... here.
1556 * config/sol2-10.h: Remove.
1558 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
1559 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
1560 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
1561 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
1562 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
1563 * config/sol2.h: ... here.
1564 (SECTION_NAME_FORMAT): Don't redefine.
1565 (STARTFILE_ARCH32_SPEC): Rename to ...
1566 (STARTFILE_ARCH_SPEC): ... this.
1567 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
1568 * config/sparc/sol2.h: ... here.
1569 (SECTION_NAME_FORMAT): Don't undef.
1570 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
1571 (SUBTARGET_EXTRA_SPECS): Remove.
1572 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
1574 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
1575 (MD_STARTFILE_PREFIX): Remove.
1576 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
1577 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
1578 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
1579 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
1580 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
1581 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
1582 * config/i386/sol2.h: ... here.
1583 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
1584 * config/i386/sol2-bi.h: Remove.
1585 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
1586 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
1588 * config/i386/t-sol2-64: Rename to ...
1589 * config/i386/t-sol2: ... this.
1590 * config/sparc/t-sol2-64: Rename to ...
1591 * config/sparc/t-sol2: ... this.
1593 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
1594 sol2_tm_file_head, sol2_tm_file_tail.
1595 Include ${cpu_type}/sol2.h before sol2.h.
1597 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
1598 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
1599 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
1600 Reflect i386/t-sol2-64 renaming.
1601 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
1602 Reflect sparc/t-sol2-64 renaming.
1604 2014-04-30 Richard Biener <rguenther@suse.de>
1606 * passes.c (execute_function_todo): Move TODO_verify_stmts
1607 and TODO_verify_ssa under the TODO_verify_il umbrella.
1608 * tree-ssa.h (verify_ssa): Adjust prototype.
1609 * tree-ssa.c (verify_ssa): Add parameter to tell whether
1610 we should verify SSA operands.
1611 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
1612 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
1613 whether we should verify whether not throwing stmts have EH info.
1614 * graphite-scop-detection.c (create_sese_edges): Adjust.
1615 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
1616 * tree-eh.c (lower_try_finally_switch): Do not add the
1617 default case label twice.
1619 2014-04-30 Marek Polacek <polacek@redhat.com>
1621 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
1622 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
1623 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
1624 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
1626 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
1628 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
1629 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
1630 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
1631 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
1632 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
1633 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
1634 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
1635 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
1637 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1639 * tree-cfg.c (dump_function_to_file): Dump the return type of
1640 functions, in a line to itself before the function body, mimicking
1641 the layout of a C function.
1643 2014-04-29 Jakub Jelinek <jakub@redhat.com>
1645 PR tree-optimization/60971
1646 * tree-tailcall.c (process_assignment): Reject conversions which
1649 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
1651 * calls.c (initialize_argument_information): Always treat
1652 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
1653 (expand_call): Likewise.
1654 (emit_library_call_calue_1): Likewise.
1655 * expr.c (PUSH_ARGS_REVERSED): Do not define.
1656 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
1659 2014-04-29 Nick Clifton <nickc@redhat.com>
1661 * config/msp430/msp430.md (umulsidi): Fix typo.
1662 (mulhisi3): Enable even inside interrupt handlers.
1663 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
1664 bigger return address pushed in large mode.
1666 2014-04-29 Nick Clifton <nickc@redhat.com>
1668 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
1669 (arc_init_reg_tables): Use a machine_mode enum to iterate over
1671 * config/m32r/m32r.c (init_reg_tables): Likewise.
1672 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
1673 enum to hold the modes.
1675 2014-04-29 Richard Biener <rguenther@suse.de>
1677 * dominance.c (free_dominance_info): Add overload with
1679 (dom_info_state): Likewise.
1680 (dom_info_available_p): Likewise.
1681 * basic-block.h (free_dominance_info, dom_info_state,
1682 dom_info_available_p): Declare overloads.
1683 * passes.c (execute_function_todo): Verify that verifiers
1684 don't change dominator info state. Drop dominator info
1685 for IPA pass invocations.
1686 * cgraph.c (release_function_body): Restore asserts that
1687 dominator information is released.
1689 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
1691 * doc/invoke.texi: Fix typo.
1692 * tree-vrp.c: Fix typos.
1693 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
1695 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1697 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
1699 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1701 * config/aarch64/aarch64-builtins.c
1702 (aarch64_types_storestruct_lane_qualifiers): New.
1703 (TYPES_STORESTRUCT_LANE): Likewise.
1704 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
1705 (st3_lane): Likewise.
1706 (st4_lane): Likewise.
1707 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
1708 (vec_store_lanesci_lane<mode>): Likewise.
1709 (vec_store_lanesxi_lane<mode>): Likewise.
1710 (aarch64_st2_lane<VQ:mode>): Likewise.
1711 (aarch64_st3_lane<VQ:mode>): Likewise.
1712 (aarch64_st4_lane<VQ:mode>): Likewise.
1713 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
1714 * config/aarch64/arm_neon.h
1715 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
1716 use new macro arguments.
1717 (__ST3_LANE_FUNC): Likewise.
1718 (__ST4_LANE_FUNC): Likewise.
1719 * config/aarch64/iterators.md (V_TWO_ELEM): New.
1720 (V_THREE_ELEM): Likewise.
1721 (V_FOUR_ELEM): Likewise.
1723 2014-04-28 David Malcolm <dmalcolm@redhat.com>
1725 * doc/gimple.texi: Replace the description of the now-defunct
1726 union gimple_statement_d with a diagram showing the
1727 gimple_statement_base class hierarchy and its relationships to
1728 the GSS_ and GIMPLE_ enums.
1730 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1732 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
1733 * config/aarch64/aarch64.c
1734 (aarch64_cannot_change_mode_class): Weaken conditions.
1735 (aarch64_modes_tieable_p): New.
1736 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
1738 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
1740 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
1741 (loadsync_<mode>): Change mode.
1742 (load_quadpti, store_quadpti): New.
1743 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
1744 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
1746 2014-04-28 Martin Jambor <mjambor@suse.cz>
1748 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
1749 same alias type as the original statement.
1750 (subreplacement_assignment_data): New type.
1751 (handle_unscalarized_data_in_subtree): New type of parameter,
1752 generate new memory accesses with same alias type as the original
1754 (load_assign_lhs_subreplacements): Likewise.
1755 (sra_modify_constructor_assign): Generate new memory accesses with
1756 same alias type as the original statement.
1758 2014-04-28 Richard Biener <rguenther@suse.de>
1760 * tree-pass.h (TODO_verify_il): Define.
1761 (TODO_verify_all): Complete properly.
1762 * passes.c (execute_function_todo): Move existing loop-closed
1763 SSA verification under TODO_verify_il.
1764 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
1765 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
1766 Fix tree sharing issue.
1768 2014-04-28 Richard Biener <rguenther@suse.de>
1771 * builtins.def (DEF_C11_BUILTIN): Add.
1772 (BUILT_IN_ALIGNED_ALLOC): Likewise.
1773 * coretypes.h (enum function_class): Add function_c11_misc.
1774 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1775 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
1776 (call_may_clobber_ref_p_1): Likewise.
1777 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
1778 (mark_all_reaching_defs_necessary_1): Likewise.
1779 (propagate_necessity): Likewise.
1780 (eliminate_unnecessary_stmts): Likewise.
1781 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
1783 2014-04-28 Richard Biener <rguenther@suse.de>
1785 * tree-vrp.c (vrp_var_may_overflow): Remove.
1786 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
1787 with overflow immediately bump to one before that value and
1788 let iteration figure out overflow status.
1790 2014-04-28 Richard Biener <rguenther@suse.de>
1792 * configure.ac: Do valgrind header checks unconditionally.
1793 Add --enable-valgrind-annotations.
1794 * system.h: Guard valgrind header inclusion with
1795 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
1796 * alloc-pool.c (pool_alloc, pool_free): Use
1797 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
1798 to guard possibly dead code.
1799 * config.in: Regenerated.
1800 * configure: Likewise.
1802 2014-04-28 Jeff Law <law@redhat.com>
1804 PR tree-optimization/60902
1805 * tree-ssa-threadedge.c
1806 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
1807 over real defs when invalidating outputs from statements that do not
1808 produce useful outputs for threading.
1810 2014-04-28 Richard Biener <rguenther@suse.de>
1812 PR tree-optimization/60979
1813 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
1814 SCOPs that end in a block with a successor with abnormal
1817 2014-04-28 Richard Biener <rguenther@suse.de>
1819 * tree-pass.h (execute_pass_list): Adjust prototype.
1820 * passes.c (pass_manager::execute_early_local_passes): Adjust.
1821 (do_per_function): Change callback signature, push all actual
1822 work to the callbals.
1823 (do_per_function_toporder): Likewise.
1824 (execute_function_dump): Adjust.
1825 (execute_function_todo): Likewise.
1826 (clear_last_verified): Likewise.
1827 (verify_curr_properties): Likewise.
1828 (update_properties_after_pass): Likewise.
1829 (execute_pass_list_1): Split out from ...
1830 (execute_pass_list): ... here. Adjust.
1831 (execute_ipa_pass_list): Likewise.
1832 * cgraphunit.c (cgraph_add_new_function): Adjust.
1833 (analyze_function): Likewise.
1834 (expand_function): Likewise.
1835 * cgraph.c (release_function_body): Free dominance info
1836 here instead of asserting it was magically freed elsewhere.
1838 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
1840 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
1841 * configure: Regenerate.
1842 * config/sparc/sparc.opt (muser-mode): New option.
1843 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
1845 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
1846 * doc/invoke.texi (SPARC options): Document -muser-mode.
1848 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
1850 * cselib.c (find_slot_memmode): Delete.
1851 (cselib_hasher): Change compare_type to a struct.
1852 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
1854 (preserve_constants_and_equivs): Adjust for new compare_type.
1855 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
1856 (wrap_constant): Delete.
1857 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
1859 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
1861 * doc/install.texi (Building with profile feedback): Remove
1864 2014-04-26 Tom de Vries <tom@codesourcery.com>
1866 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
1869 2014-04-25 Cary Coutant <ccoutant@google.com>
1872 * dwarf2out.c (should_move_die_to_comdat): A type definition
1873 can contain a subprogram definition, but don't move it to a
1875 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
1876 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
1878 (clone_tree_hash): Rename to...
1879 (clone_tree_partial): ...this; change callers. Copy
1880 DW_TAG_subprogram DIEs as declarations.
1881 (copy_decls_walk): Don't copy children of a declaration into a
1884 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
1887 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
1890 2014-04-25 Jiong Wang <jiong.wang@arm.com>
1892 * config/arm/predicates.md (call_insn_operand): Add long_call check.
1893 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
1895 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
1898 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1900 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
1902 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1904 PR tree-optimization/60930
1905 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
1906 creating a multiply candidate by folding two constant
1907 multiplicands when the result overflows.
1909 2014-04-25 Jakub Jelinek <jakub@redhat.com>
1911 PR tree-optimization/60960
1912 * tree-vect-generic.c (expand_vector_operation): Only call
1913 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
1915 2014-04-25 Tom de Vries <tom@codesourcery.com>
1917 * expr.c (clobber_reg_mode): New function.
1918 * expr.h (clobber_reg): New function.
1920 2014-04-25 Tom de Vries <tom@codesourcery.com>
1922 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
1925 2014-04-25 Radovan Obradovic <robradovic@mips.com>
1926 Tom de Vries <tom@codesourcery.com>
1928 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
1930 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
1931 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
1932 new argument to find_all_hard_reg_sets call.
1934 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1936 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
1937 Use HOST_WIDE_INT_C for mask literal.
1938 (aarch_rev16_shleft_mask_imm_p): Likewise.
1940 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
1943 * config/sparc/sparc.md (ashlsi3_extend): Delete.
1945 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
1947 PR preprocessor/56540
1948 * config/i386/i386-c.c (ix86_target_macros): Define
1949 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
1951 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1953 * configure.ac (tga_func): Remove.
1954 (LIB_TLS_SPEC): Remove.
1955 * configure: Regenerate.
1956 * config.in: Regenerate.
1957 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
1959 2014-04-25 Richard Biener <rguenther@suse.de>
1962 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
1963 call stmt use/clobber sets during stmt walk instead of
1964 walking the possibly incomplete set of caller edges.
1966 2014-04-25 Richard Biener <rguenther@suse.de>
1969 * passes.c (apply_ipa_transforms): Inline into only caller ...
1970 (execute_one_pass): ... here. Properly bring in function
1971 bodies for nodes we want to apply IPA transforms to.
1973 2014-04-24 Cong Hou <congh@google.com>
1975 PR tree-optimization/60896
1976 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
1977 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
1978 (vect_mark_pattern_stmts): Set the def type of all statements in
1979 PATTERN_DEF_SEQ as vect_internal_def.
1981 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
1983 * doc/extend.texi (PowerPC Built-in Functions): Document new
1984 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
1985 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
1987 * config/rs6000/predicates.md (const_0_to_3_operand): New
1988 predicate to match 0..3 integer constants.
1990 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
1991 to support adding miscellaneous builtin functions.
1992 (BU_DFP_MISC_2): Likewise.
1993 (BU_P7_MISC_1): Likewise.
1994 (BU_P7_MISC_2): Likewise.
1995 (BU_P8V_MISC_3): Likewise.
1996 (BU_MISC_1): Likewise.
1997 (BU_MISC_2): Likewise.
1998 (DIVWE): Add extended divide builtin functions.
2001 (DIVWEUO): Likewise.
2005 (DIVDEUO): Likewise.
2006 (DXEX): Add decimal floating-point builtin functions.
2009 (DDEDPDQ): Likewise.
2011 (DENBCDQ): Likewise.
2018 (CDTBCD): Add new BCD builtin functions.
2022 (BCDADD_LT): Likewise.
2023 (BCDADD_EQ): Likewise.
2024 (BCDADD_GT): Likewise.
2025 (BCDADD_OV): Likewise.
2027 (BCDSUB_LT): Likewise.
2028 (BCDSUB_EQ): Likewise.
2029 (BCDSUB_GT): Likewise.
2030 (BCDSUB_OV): Likewise.
2031 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
2032 (UNPACK_TD): Likewise.
2033 (PACK_TF): Likewise.
2034 (UNPACK_TF): Likewise.
2035 (UNPACK_TF_0): Likewise.
2036 (UNPACK_TF_1): Likewise.
2037 (PACK_V1TI): Likewise.
2038 (UNPACK_V1TI): Likewise.
2040 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
2041 support for decimal floating point builtin functions.
2042 (rs6000_expand_ternop_builtin): Add checks for the new builtin
2043 functions that take constant arguments.
2044 (rs6000_invalid_builtin): Add decimal floating point builtin support.
2045 (rs6000_init_builtins): Setup long double, _Decimal64, and
2046 _Decimal128 types for new builtin functions.
2047 (builtin_function_type): Set the unsigned flags appropriately for
2048 the new builtin functions.
2049 (rs6000_opt_masks): Add support for decimal floating point builtin
2052 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
2053 floating point builtin functions.
2054 (RS6000_BTM_COMMON): Likewise.
2055 (RS6000_BTI_long_double): Likewise.
2056 (RS6000_BTI_dfloat64): Likewise.
2057 (RS6000_BTI_dfloat128): Likewise.
2058 (long_double_type_internal_node): Likewise.
2059 (dfloat64_type_internal_node): Likewise.
2060 (dfloat128_type_internal_node): Likewise.
2062 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
2063 2.07 bcd arithmetic instructions.
2064 (UNSPEC_BCDSUB): Likewise.
2065 (UNSPEC_BCD_OVERFLOW): Likewise.
2066 (UNSPEC_BCD_ADD_SUB): Likewise.
2067 (bcd_add_sub): Likewise.
2068 (BCD_TEST): Likewise.
2069 (bcd<bcd_add_sub>): Likewise.
2070 (bcd<bcd_add_sub>_test): Likewise.
2071 (bcd<bcd_add_sub>_test2): Likewise.
2072 (bcd<bcd_add_sub>_<code>): Likewise.
2073 (peephole2 for combined bcd ops): Likewise.
2075 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
2076 decimal floating point builtin functions.
2077 (UNSPEC_DENBCD): Likewise.
2078 (UNSPEC_DXEX): Likewise.
2079 (UNSPEC_DIEX): Likewise.
2080 (UNSPEC_DSCLI): Likewise.
2081 (UNSPEC_DSCRI): Likewise.
2082 (D64_D128): Likewise.
2083 (dfp_suffix): Likewise.
2084 (dfp_ddedpd_<mode>): Likewise.
2085 (dfp_denbcd_<mode>): Likewise.
2086 (dfp_dxex_<mode>): Likewise.
2087 (dfp_diex_<mode>): Likewise.
2088 (dfp_dscli_<mode>): Likewise.
2089 (dfp_dscri_<mode>): Likewise.
2091 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
2093 (UNSPEC_CDTBCD): Likewise.
2094 (UNSPEC_CBCDTD): Likewise.
2095 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
2096 (UNSPEC_DIVEO): Likewise.
2097 (UNSPEC_DIVEU): Likewise.
2098 (UNSPEC_DIVEUO): Likewise.
2099 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
2100 pack/unpack 128-bit types.
2101 (UNSPEC_PACK_128BIT): Likewise.
2102 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
2103 (udiv<mode>3): Use idiv_ldiv mode attribute.
2104 (div<mode>3): Likewise.
2105 (addg6s): Add new BCD builtin functions.
2108 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
2109 (div_extend): Likewise.
2110 (div<div_extend>_<mode>"): Likewise.
2111 (FP128_64): Add support for new builtin functions to pack/unpack
2113 (unpack<mode>): Likewise.
2114 (unpacktf_0): Likewise.
2115 (unpacktf_1): Likewise.
2116 (unpack<mode>_dm): Likewise.
2117 (unpack<mode>_nodm): Likewise.
2118 (pack<mode>): Likewise.
2119 (unpackv1ti): Likewise.
2120 (packv1ti): Likewise.
2122 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
2124 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
2127 2014-04-24 Jakub Jelinek <jakub@redhat.com>
2129 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
2130 * gimplify.c (omp_is_private): Change last argument's type to int.
2131 Only diagnose lastprivate if the simd argument is 1, only diagnose
2132 linear if the simd argument is 2.
2133 (gimplify_omp_for): Adjust omp_is_private callers. When adding
2134 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
2135 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
2136 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
2137 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
2138 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2139 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
2140 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2141 * tree-nested.c (convert_nonlocal_omp_clauses,
2142 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
2144 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
2147 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
2150 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
2152 * flag-types.h (enum ivar_visibility): Add.
2154 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
2156 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
2157 function * argument.
2159 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
2161 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
2163 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2164 Tom de Vries <tom@codesourcery.com>
2166 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
2167 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
2169 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
2170 * emit-rtl.c (try_split): Same.
2172 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2173 Tom de Vries <tom@codesourcery.com>
2175 * common.opt (fuse-caller-save): New option.
2177 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
2179 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
2180 elements for big-endian.
2182 2014-04-24 Richard Biener <rguenther@suse.de>
2184 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
2185 during TER and instead use the sepops interface for expanding
2186 non-GIMPLE_SINGLE_RHS.
2188 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2190 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
2191 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
2193 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2195 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
2196 assembler 64-bit option.
2197 * configure: Regenerate.
2199 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2201 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
2202 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
2203 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
2204 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
2205 (TARGET_CRYPTO): Take TARGET_SIMD into account.
2207 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2209 * config/aarch64/aarch64-builtins.c
2210 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
2211 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
2212 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
2213 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
2215 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
2216 (Vrevsuff): New mode attribute.
2218 2014-04-24 Terry Guo <terry.guo@arm.com>
2220 * config/arm/arm.h (machine_function): Define variable
2221 after_arm_reorg here.
2222 * config/arm/arm.c (after_arm_reorg): Remove the definition.
2223 (arm_split_constant): Update the way to access variable
2226 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
2228 2014-04-23 Tom de Vries <tom@codesourcery.com>
2230 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
2232 2014-04-23 David Malcolm <dmalcolm@redhat.com>
2234 * is-a.h: Update comments to reflect the following changes to the
2235 "pointerness" of the API, making the template parameter match the
2236 return type, allowing use of is-a.h with typedefs of pointers.
2237 (is_a_helper::cast): Return a T rather then a pointer to a T, so
2238 that the return type matches the parameter to the is_a_helper.
2240 (dyn_cast): Likewise.
2242 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
2243 pointer from the is-a.h API.
2245 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
2246 (is_a_helper <cgraph_node *>::test): ...this, matching change to
2248 (is_a_helper <varpool_node>::test): Likewise, convert to...
2249 (is_a_helper <varpool_node *>::test): ...this.
2251 (varpool_first_variable): Update for removal of implicit pointer
2252 from the is-a.h API.
2253 (varpool_next_variable): Likewise.
2254 (varpool_first_static_initializer): Likewise.
2255 (varpool_next_static_initializer): Likewise.
2256 (varpool_first_defined_variable): Likewise.
2257 (varpool_next_defined_variable): Likewise.
2258 (cgraph_first_defined_function): Likewise.
2259 (cgraph_next_defined_function): Likewise.
2260 (cgraph_first_function): Likewise.
2261 (cgraph_next_function): Likewise.
2262 (cgraph_first_function_with_gimple_body): Likewise.
2263 (cgraph_next_function_with_gimple_body): Likewise.
2264 (cgraph_alias_target): Likewise.
2265 (varpool_alias_target): Likewise.
2266 (cgraph_function_or_thunk_node): Likewise.
2267 (varpool_variable_node): Likewise.
2268 (symtab_real_symbol_p): Likewise.
2269 * cgraphunit.c (referred_to_p): Likewise.
2270 (analyze_functions): Likewise.
2271 (handle_alias_pairs): Likewise.
2272 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2273 * gimple-ssa.h (gimple_vuse_op): Likewise.
2274 (gimple_vdef_op): Likewise.
2275 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
2276 * gimple.c (gimple_build_asm_1): Likewise.
2277 (gimple_build_try): Likewise.
2278 (gimple_build_resx): Likewise.
2279 (gimple_build_eh_dispatch): Likewise.
2280 (gimple_build_omp_for): Likewise.
2281 (gimple_omp_for_set_clauses): Likewise.
2283 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
2284 (is_a_helper <gimple_statement_asm *>::test): ...this.
2285 (is_a_helper <gimple_statement_bind>::test): Convert to...
2286 (is_a_helper <gimple_statement_bind *>::test): ...this.
2287 (is_a_helper <gimple_statement_call>::test): Convert to...
2288 (is_a_helper <gimple_statement_call *>::test): ...this.
2289 (is_a_helper <gimple_statement_catch>::test): Convert to...
2290 (is_a_helper <gimple_statement_catch *>::test): ...this.
2291 (is_a_helper <gimple_statement_resx>::test): Convert to...
2292 (is_a_helper <gimple_statement_resx *>::test): ...this.
2293 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
2294 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
2295 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
2296 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
2297 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
2298 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
2299 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
2300 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
2301 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
2302 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
2303 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
2304 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
2305 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
2306 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
2307 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
2308 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
2309 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
2310 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
2311 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
2312 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
2313 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
2314 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
2315 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
2316 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
2317 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
2318 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
2319 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
2320 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
2321 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
2322 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
2323 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
2324 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
2325 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
2326 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
2327 (is_a_helper <gimple_statement_phi>::test): Convert to...
2328 (is_a_helper <gimple_statement_phi *>::test): ...this.
2329 (is_a_helper <gimple_statement_transaction>::test): Convert to...
2330 (is_a_helper <gimple_statement_transaction *>::test): ...this.
2331 (is_a_helper <gimple_statement_try>::test): Convert to...
2332 (is_a_helper <gimple_statement_try *>::test): ...this.
2333 (is_a_helper <gimple_statement_wce>::test): Convert to...
2334 (is_a_helper <gimple_statement_wce *>::test): ...this.
2335 (is_a_helper <const gimple_statement_asm>::test): Convert to...
2336 (is_a_helper <const gimple_statement_asm *>::test): ...this.
2337 (is_a_helper <const gimple_statement_bind>::test): Convert to...
2338 (is_a_helper <const gimple_statement_bind *>::test): ...this.
2339 (is_a_helper <const gimple_statement_call>::test): Convert to...
2340 (is_a_helper <const gimple_statement_call *>::test): ...this.
2341 (is_a_helper <const gimple_statement_catch>::test): Convert to...
2342 (is_a_helper <const gimple_statement_catch *>::test): ...this.
2343 (is_a_helper <const gimple_statement_resx>::test): Convert to...
2344 (is_a_helper <const gimple_statement_resx *>::test): ...this.
2345 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
2346 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
2347 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
2348 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
2349 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
2351 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
2353 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
2355 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
2357 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
2358 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
2359 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
2361 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
2362 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
2364 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
2365 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
2366 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
2367 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
2368 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
2369 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
2371 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
2372 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
2373 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
2374 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
2376 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
2377 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
2378 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
2379 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
2380 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
2381 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
2382 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
2383 (is_a_helper <const gimple_statement_phi>::test): Convert to...
2384 (is_a_helper <const gimple_statement_phi *>::test): ...this.
2385 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
2386 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
2387 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
2388 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
2389 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
2390 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
2391 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
2393 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
2395 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
2396 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
2398 (gimple_use_ops): Update for removal of implicit pointer from the
2400 (gimple_set_use_ops): Likewise.
2401 (gimple_vuse): Likewise.
2402 (gimple_vdef): Likewise.
2403 (gimple_vuse_ptr): Likewise.
2404 (gimple_vdef_ptr): Likewise.
2405 (gimple_set_vuse): Likewise.
2406 (gimple_set_vdef): Likewise.
2407 (gimple_omp_return_set_lhs): Likewise.
2408 (gimple_omp_return_lhs): Likewise.
2409 (gimple_omp_return_lhs_ptr): Likewise.
2410 (gimple_call_fntype): Likewise.
2411 (gimple_call_set_fntype): Likewise.
2412 (gimple_call_set_internal_fn): Likewise.
2413 (gimple_call_use_set): Likewise.
2414 (gimple_call_clobber_set): Likewise.
2415 (gimple_bind_vars): Likewise.
2416 (gimple_bind_set_vars): Likewise.
2417 (gimple_bind_body_ptr): Likewise.
2418 (gimple_bind_set_body): Likewise.
2419 (gimple_bind_add_stmt): Likewise.
2420 (gimple_bind_block): Likewise.
2421 (gimple_bind_set_block): Likewise.
2422 (gimple_asm_ninputs): Likewise.
2423 (gimple_asm_noutputs): Likewise.
2424 (gimple_asm_nclobbers): Likewise.
2425 (gimple_asm_nlabels): Likewise.
2426 (gimple_asm_input_op): Likewise.
2427 (gimple_asm_input_op_ptr): Likewise.
2428 (gimple_asm_output_op): Likewise.
2429 (gimple_asm_output_op_ptr): Likewise.
2430 (gimple_asm_set_output_op): Likewise.
2431 (gimple_asm_clobber_op): Likewise.
2432 (gimple_asm_set_clobber_op): Likewise.
2433 (gimple_asm_label_op): Likewise.
2434 (gimple_asm_set_label_op): Likewise.
2435 (gimple_asm_string): Likewise.
2436 (gimple_catch_types): Likewise.
2437 (gimple_catch_types_ptr): Likewise.
2438 (gimple_catch_handler_ptr): Likewise.
2439 (gimple_catch_set_types): Likewise.
2440 (gimple_catch_set_handler): Likewise.
2441 (gimple_eh_filter_types): Likewise.
2442 (gimple_eh_filter_types_ptr): Likewise.
2443 (gimple_eh_filter_failure_ptr): Likewise.
2444 (gimple_eh_filter_set_types): Likewise.
2445 (gimple_eh_filter_set_failure): Likewise.
2446 (gimple_eh_must_not_throw_fndecl): Likewise.
2447 (gimple_eh_must_not_throw_set_fndecl): Likewise.
2448 (gimple_eh_else_n_body_ptr): Likewise.
2449 (gimple_eh_else_e_body_ptr): Likewise.
2450 (gimple_eh_else_set_n_body): Likewise.
2451 (gimple_eh_else_set_e_body): Likewise.
2452 (gimple_try_eval_ptr): Likewise.
2453 (gimple_try_cleanup_ptr): Likewise.
2454 (gimple_try_set_eval): Likewise.
2455 (gimple_try_set_cleanup): Likewise.
2456 (gimple_wce_cleanup_ptr): Likewise.
2457 (gimple_wce_set_cleanup): Likewise.
2458 (gimple_phi_capacity): Likewise.
2459 (gimple_phi_num_args): Likewise.
2460 (gimple_phi_result): Likewise.
2461 (gimple_phi_result_ptr): Likewise.
2462 (gimple_phi_set_result): Likewise.
2463 (gimple_phi_arg): Likewise.
2464 (gimple_phi_set_arg): Likewise.
2465 (gimple_resx_region): Likewise.
2466 (gimple_resx_set_region): Likewise.
2467 (gimple_eh_dispatch_region): Likewise.
2468 (gimple_eh_dispatch_set_region): Likewise.
2469 (gimple_omp_critical_name): Likewise.
2470 (gimple_omp_critical_name_ptr): Likewise.
2471 (gimple_omp_critical_set_name): Likewise.
2472 (gimple_omp_for_clauses): Likewise.
2473 (gimple_omp_for_clauses_ptr): Likewise.
2474 (gimple_omp_for_set_clauses): Likewise.
2475 (gimple_omp_for_collapse): Likewise.
2476 (gimple_omp_for_index): Likewise.
2477 (gimple_omp_for_index_ptr): Likewise.
2478 (gimple_omp_for_set_index): Likewise.
2479 (gimple_omp_for_initial): Likewise.
2480 (gimple_omp_for_initial_ptr): Likewise.
2481 (gimple_omp_for_set_initial): Likewise.
2482 (gimple_omp_for_final): Likewise.
2483 (gimple_omp_for_final_ptr): Likewise.
2484 (gimple_omp_for_set_final): Likewise.
2485 (gimple_omp_for_incr): Likewise.
2486 (gimple_omp_for_incr_ptr): Likewise.
2487 (gimple_omp_for_set_incr): Likewise.
2488 (gimple_omp_for_pre_body_ptr): Likewise.
2489 (gimple_omp_for_set_pre_body): Likewise.
2490 (gimple_omp_parallel_clauses): Likewise.
2491 (gimple_omp_parallel_clauses_ptr): Likewise.
2492 (gimple_omp_parallel_set_clauses): Likewise.
2493 (gimple_omp_parallel_child_fn): Likewise.
2494 (gimple_omp_parallel_child_fn_ptr): Likewise.
2495 (gimple_omp_parallel_set_child_fn): Likewise.
2496 (gimple_omp_parallel_data_arg): Likewise.
2497 (gimple_omp_parallel_data_arg_ptr): Likewise.
2498 (gimple_omp_parallel_set_data_arg): Likewise.
2499 (gimple_omp_task_clauses): Likewise.
2500 (gimple_omp_task_clauses_ptr): Likewise.
2501 (gimple_omp_task_set_clauses): Likewise.
2502 (gimple_omp_task_child_fn): Likewise.
2503 (gimple_omp_task_child_fn_ptr): Likewise.
2504 (gimple_omp_task_set_child_fn): Likewise.
2505 (gimple_omp_task_data_arg): Likewise.
2506 (gimple_omp_task_data_arg_ptr): Likewise.
2507 (gimple_omp_task_set_data_arg): Likewise.
2508 (gimple_omp_taskreg_clauses): Likewise.
2509 (gimple_omp_taskreg_clauses_ptr): Likewise.
2510 (gimple_omp_taskreg_set_clauses): Likewise.
2511 (gimple_omp_taskreg_child_fn): Likewise.
2512 (gimple_omp_taskreg_child_fn_ptr): Likewise.
2513 (gimple_omp_taskreg_set_child_fn): Likewise.
2514 (gimple_omp_taskreg_data_arg): Likewise.
2515 (gimple_omp_taskreg_data_arg_ptr): Likewise.
2516 (gimple_omp_taskreg_set_data_arg): Likewise.
2517 (gimple_omp_task_copy_fn): Likewise.
2518 (gimple_omp_task_copy_fn_ptr): Likewise.
2519 (gimple_omp_task_set_copy_fn): Likewise.
2520 (gimple_omp_task_arg_size): Likewise.
2521 (gimple_omp_task_arg_size_ptr): Likewise.
2522 (gimple_omp_task_set_arg_size): Likewise.
2523 (gimple_omp_task_arg_align): Likewise.
2524 (gimple_omp_task_arg_align_ptr): Likewise.
2525 (gimple_omp_task_set_arg_align): Likewise.
2526 (gimple_omp_single_clauses): Likewise.
2527 (gimple_omp_single_clauses_ptr): Likewise.
2528 (gimple_omp_single_set_clauses): Likewise.
2529 (gimple_omp_target_clauses): Likewise.
2530 (gimple_omp_target_clauses_ptr): Likewise.
2531 (gimple_omp_target_set_clauses): Likewise.
2532 (gimple_omp_target_child_fn): Likewise.
2533 (gimple_omp_target_child_fn_ptr): Likewise.
2534 (gimple_omp_target_set_child_fn): Likewise.
2535 (gimple_omp_target_data_arg): Likewise.
2536 (gimple_omp_target_data_arg_ptr): Likewise.
2537 (gimple_omp_target_set_data_arg): Likewise.
2538 (gimple_omp_teams_clauses): Likewise.
2539 (gimple_omp_teams_clauses_ptr): Likewise.
2540 (gimple_omp_teams_set_clauses): Likewise.
2541 (gimple_omp_sections_clauses): Likewise.
2542 (gimple_omp_sections_clauses_ptr): Likewise.
2543 (gimple_omp_sections_set_clauses): Likewise.
2544 (gimple_omp_sections_control): Likewise.
2545 (gimple_omp_sections_control_ptr): Likewise.
2546 (gimple_omp_sections_set_control): Likewise.
2547 (gimple_omp_for_set_cond): Likewise.
2548 (gimple_omp_for_cond): Likewise.
2549 (gimple_omp_atomic_store_set_val): Likewise.
2550 (gimple_omp_atomic_store_val): Likewise.
2551 (gimple_omp_atomic_store_val_ptr): Likewise.
2552 (gimple_omp_atomic_load_set_lhs): Likewise.
2553 (gimple_omp_atomic_load_lhs): Likewise.
2554 (gimple_omp_atomic_load_lhs_ptr): Likewise.
2555 (gimple_omp_atomic_load_set_rhs): Likewise.
2556 (gimple_omp_atomic_load_rhs): Likewise.
2557 (gimple_omp_atomic_load_rhs_ptr): Likewise.
2558 (gimple_omp_continue_control_def): Likewise.
2559 (gimple_omp_continue_control_def_ptr): Likewise.
2560 (gimple_omp_continue_set_control_def): Likewise.
2561 (gimple_omp_continue_control_use): Likewise.
2562 (gimple_omp_continue_control_use_ptr): Likewise.
2563 (gimple_omp_continue_set_control_use): Likewise.
2564 (gimple_transaction_body_ptr): Likewise.
2565 (gimple_transaction_label): Likewise.
2566 (gimple_transaction_label_ptr): Likewise.
2567 (gimple_transaction_set_body): Likewise.
2568 (gimple_transaction_set_label): Likewise.
2570 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
2571 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2572 * ipa-ref.c (ipa_record_reference): Likewise.
2573 * ipa-reference.c (analyze_function): Likewise.
2574 (ipa_reference_write_optimization_summary): Likewise.
2575 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
2576 (address_taken_from_non_vtable_p): Likewise.
2577 (comdat_can_be_unshared_p_1): Likewise.
2578 * lto-cgraph.c (lto_output_ref): Likewise.
2579 (add_references): Likewise.
2580 (compute_ltrans_boundary): Likewise.
2581 (output_symtab): Likewise.
2582 (input_ref): Likewise.
2583 (input_cgraph_1): Likewise.
2584 (output_cgraph_opt_summary): Likewise.
2585 * lto-streamer-out.c (lto_output): Likewise.
2586 (output_symbol_p): Likewise.
2587 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
2588 (lsei_start_function_in_partition): Likewise.
2589 (lsei_next_variable_in_partition): Likewise.
2590 (lsei_start_variable_in_partition): Likewise.
2591 * symtab.c (insert_to_assembler_name_hash): Likewise.
2592 (unlink_from_assembler_name_hash): Likewise.
2593 (symtab_unregister_node): Likewise.
2594 (symtab_remove_node): Likewise.
2595 (dump_symtab_node): Likewise.
2596 (verify_symtab_base): Likewise.
2597 (verify_symtab_node): Likewise.
2598 (symtab_make_decl_local): Likewise.
2599 (symtab_alias_ultimate_target): Likewise.
2600 (symtab_resolve_alias): Likewise.
2601 (symtab_get_symbol_partitioning_class): Likewise.
2602 * tree-phinodes.c (allocate_phi_node): Likewise.
2603 (reserve_phi_args_for_new_edge): Likewise.
2604 (remove_phi_args): Likewise.
2605 * varpool.c (varpool_node_for_asm): Likewise.
2606 (varpool_remove_unreferenced_decls): Likewise.
2608 2014-04-23 Jeff Law <law@redhat.com>
2610 PR tree-optimization/60902
2611 * tree-ssa-threadedge.c
2612 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
2613 invalidate outputs from statements that do not produce useful
2614 outputs for threading.
2616 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2618 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
2619 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
2620 machine descriptions for Stack Smashing Protector.
2622 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
2624 * aarch64.md (<optab>_rol<mode>3): New pattern.
2625 (<optab>_rolsi3_uxtw): Likewise.
2626 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
2628 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
2630 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
2631 (arm_cortex_a12_tune): Likewise.
2633 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2635 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
2637 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2639 * config/arm/arm.md (arm_rev16si2): New pattern.
2640 (arm_rev16si2_alt): Likewise.
2641 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
2643 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2645 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
2646 (rev16<mode>2_alt): Likewise.
2647 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
2648 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
2649 (aarch_rev16_shleft_mask_imm_p): Likewise.
2650 (aarch_rev16_p_1): Likewise.
2651 (aarch_rev16_p): Likewise.
2652 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
2653 (aarch_rev16_shright_mask_imm_p): Likewise.
2654 (aarch_rev16_shleft_mask_imm_p): Likewise.
2656 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2658 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
2659 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
2661 (cortex_a53_extra_costs): Likewise.
2662 (cortex_a57_extra_costs): Likewise.
2663 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
2664 (cortexa7_extra_costs): Likewise.
2665 (cortexa8_extra_costs): Likewise.
2666 (cortexa12_extra_costs): Likewise.
2667 (cortexa15_extra_costs): Likewise.
2668 (v7m_extra_costs): Likewise.
2669 (arm_new_rtx_costs): Handle BSWAP.
2671 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2673 * config/arm/arm.c (cortexa8_extra_costs): New table.
2674 (arm_cortex_a8_tune): New tuning struct.
2675 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
2677 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2679 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
2681 2014-04-23 Richard Biener <rguenther@suse.de>
2683 * Makefile.in (OBJS): Remove loop-unswitch.o.
2684 * tree-pass.h (make_pass_rtl_unswitch): Remove.
2685 * passes.def (pass_rtl_unswitch): Likewise.
2686 * loop-init.c (gate_rtl_unswitch): Likewise.
2687 (rtl_unswitch): Likewise.
2688 (pass_data_rtl_unswitch): Likewise.
2689 (pass_rtl_unswitch): Likewise.
2690 (make_pass_rtl_unswitch): Likewise.
2691 * rtl.h (reversed_condition): Likewise.
2692 (compare_and_jump_seq): Likewise.
2693 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
2695 * loop-unroll.c (compare_and_jump_seq): Likewise.
2697 2014-04-23 Richard Biener <rguenther@suse.de>
2699 PR tree-optimization/60903
2700 * tree-ssa-loop-im.c (analyze_memory_references): Remove
2701 commented code block.
2702 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
2703 loop flags to newly created BBs and edges.
2705 2014-04-23 Nick Clifton <nickc@redhat.com>
2707 * config/msp430/msp430.c (msp430_handle_option): Move function
2709 (msp430_option_override): Simplify mcu and mcpu option handling.
2710 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
2711 support for -mhwmult command line option.
2712 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
2713 -mhwmult command line option.
2714 (msp430_hwmult_enabled): Delete.
2715 (msp43o_output_labelref): Add support for -mhwmult command line option.
2716 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
2717 (umulsidi3): Likewise.
2718 * config/msp430/msp430.opt (mmcu): Add Report attribute.
2719 (mcpu, mlarge, msmall): Likewise.
2720 (mhwmult): New option.
2721 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
2723 (msp430_is_f5_mcu): Remove prototype.
2724 (msp430_use_f5_series_hwmult): Add prototype.
2725 * config/msp430/msp430-opts.h: New file.
2726 * common/config/msp430: New directory.
2727 * common/config/msp430/msp430-common.c: New file.
2728 * config.gcc (msp430): Remove target_has_targetm_common.
2729 * doc/invoke.texi: Document -mhwmult command line option.
2731 2014-04-23 Nick Clifton <nickc@redhat.com>
2733 * config/i386/cygwin.h (ENDFILE_SPEC): Include
2734 default-manifest.o if it can be found in the search path.
2735 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
2737 2014-04-23 Terry Guo <terry.guo@arm.com>
2739 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
2741 2014-04-23 Richard Biener <rguenther@suse.de>
2744 * tree-inline.c (declare_return_variable): Use mark_addressable.
2746 2014-04-23 Richard Biener <rguenther@suse.de>
2749 * loop-init.c (loop_optimizer_init): Make sure to apply
2750 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
2752 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2755 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
2757 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
2758 if flag_sanitize_undefined_trap_on_error.
2759 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
2760 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
2761 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
2762 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
2763 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
2764 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
2765 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
2766 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
2767 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
2768 * ubsan.c (ubsan_instrument_unreachable): Return
2769 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
2770 (ubsan_expand_null_ifn): Emit __builtin_trap ()
2771 if flag_sanitize_undefined_trap_on_error and
2772 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
2773 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
2774 instrument_bool_enum_load): Emit __builtin_trap () if
2775 flag_sanitize_undefined_trap_on_error and
2776 __builtin_handle_*_abort () if !flag_sanitize_recover.
2777 * doc/invoke.texi (-fsanitize-recover,
2778 -fsanitize-undefined-trap-on-error): Document.
2780 2014-04-22 Christian Bruel <christian.bruel@st.com>
2782 * config/sh/sh.md (mov<mode>): Replace movQIHI.
2783 Force immediates to SImode.
2785 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
2787 * config/nios2/nios2.md (UNSPEC_ROUND): New.
2789 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
2790 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
2791 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
2792 (nios2_fpu_insn): Add entry for round.
2793 (N2FPU_NO_ERRNO_P): Define.
2794 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
2796 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
2798 2014-04-22 Richard Henderson <rth@redhat.com>
2800 * config/aarch64/aarch64 (addti3, subti3): New expanders.
2801 (add<GPI>3_compare0): Remove leading * from name.
2802 (add<GPI>3_carryin): Likewise.
2803 (sub<GPI>3_compare0): Likewise.
2804 (sub<GPI>3_carryin): Likewise.
2805 (<su_optab>mulditi3): New expander.
2806 (multi3): New expander.
2807 (madd<GPI>): Remove leading * from name.
2809 2014-04-22 Martin Jambor <mjambor@suse.cz>
2811 * cgraphclones.c (cgraph_function_versioning): Copy
2812 ipa_transforms_to_apply instead of asserting it is empty.
2814 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
2817 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
2818 on count_exp to get mode.
2820 2014-04-22 Andrew Pinski <apinski@cavium.com>
2822 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
2823 Handle TLS for ILP32.
2824 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
2825 (tlsie_small_<mode>): this and handle PTR.
2826 (tlsie_small_sidi): New pattern.
2827 (tlsle_small): Change to an expand to handle ILP32.
2828 (tlsle_small_<mode>): New pattern.
2829 (tlsdesc_small): Rename to ...
2830 (tlsdesc_small_<mode>): this and handle PTR.
2832 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2834 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
2836 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2838 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2839 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
2840 (aarch64_types_signed_poly_qualifiers): Likewise.
2841 (aarch64_types_unsigned_signed_qualifiers): Likewise.
2842 (aarch64_types_poly_signed_qualifiers): Likewise.
2843 (TYPES_REINTERP_SS): Type macro added.
2844 (TYPES_REINTERP_SU): Likewise.
2845 (TYPES_REINTERP_SP): Likewise.
2846 (TYPES_REINTERP_US): Likewise.
2847 (TYPES_REINTERP_PS): Likewise.
2848 (aarch64_fold_builtin): New expression folding added.
2849 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
2850 Declarations removed.
2851 (REINTERP_SS): Declarations added.
2852 (REINTERP_US): Likewise.
2853 (REINTERP_PS): Likewise.
2854 (REINTERP_SU): Likewise.
2855 (REINTERP_SP): Likewise.
2856 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
2857 (vreinterpretq_p8_f64): Likewise.
2858 (vreinterpret_p16_f64): Likewise.
2859 (vreinterpretq_p16_f64): Likewise.
2860 (vreinterpret_f32_f64): Likewise.
2861 (vreinterpretq_f32_f64): Likewise.
2862 (vreinterpret_f64_f32): Likewise.
2863 (vreinterpret_f64_p8): Likewise.
2864 (vreinterpret_f64_p16): Likewise.
2865 (vreinterpret_f64_s8): Likewise.
2866 (vreinterpret_f64_s16): Likewise.
2867 (vreinterpret_f64_s32): Likewise.
2868 (vreinterpret_f64_s64): Likewise.
2869 (vreinterpret_f64_u8): Likewise.
2870 (vreinterpret_f64_u16): Likewise.
2871 (vreinterpret_f64_u32): Likewise.
2872 (vreinterpret_f64_u64): Likewise.
2873 (vreinterpretq_f64_f32): Likewise.
2874 (vreinterpretq_f64_p8): Likewise.
2875 (vreinterpretq_f64_p16): Likewise.
2876 (vreinterpretq_f64_s8): Likewise.
2877 (vreinterpretq_f64_s16): Likewise.
2878 (vreinterpretq_f64_s32): Likewise.
2879 (vreinterpretq_f64_s64): Likewise.
2880 (vreinterpretq_f64_u8): Likewise.
2881 (vreinterpretq_f64_u16): Likewise.
2882 (vreinterpretq_f64_u32): Likewise.
2883 (vreinterpretq_f64_u64): Likewise.
2884 (vreinterpret_s64_f64): Likewise.
2885 (vreinterpretq_s64_f64): Likewise.
2886 (vreinterpret_u64_f64): Likewise.
2887 (vreinterpretq_u64_f64): Likewise.
2888 (vreinterpret_s8_f64): Likewise.
2889 (vreinterpretq_s8_f64): Likewise.
2890 (vreinterpret_s16_f64): Likewise.
2891 (vreinterpretq_s16_f64): Likewise.
2892 (vreinterpret_s32_f64): Likewise.
2893 (vreinterpretq_s32_f64): Likewise.
2894 (vreinterpret_u8_f64): Likewise.
2895 (vreinterpretq_u8_f64): Likewise.
2896 (vreinterpret_u16_f64): Likewise.
2897 (vreinterpretq_u16_f64): Likewise.
2898 (vreinterpret_u32_f64): Likewise.
2899 (vreinterpretq_u32_f64): Likewise.
2901 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2903 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2904 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
2905 (vreinterpret_p8_s8): Likewise.
2906 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
2907 (vreinterpret_p8_s16): Likewise.
2908 (vreinterpret_p8_s32): Likewise.
2909 (vreinterpret_p8_s64): Likewise.
2910 (vreinterpret_p8_f32): Likewise.
2911 (vreinterpret_p8_u8): Likewise.
2912 (vreinterpret_p8_u16): Likewise.
2913 (vreinterpret_p8_u32): Likewise.
2914 (vreinterpret_p8_u64): Likewise.
2915 (vreinterpret_p8_p16): Likewise.
2916 (vreinterpretq_p8_s8): Likewise.
2917 (vreinterpretq_p8_s16): Likewise.
2918 (vreinterpretq_p8_s32): Likewise.
2919 (vreinterpretq_p8_s64): Likewise.
2920 (vreinterpretq_p8_f32): Likewise.
2921 (vreinterpretq_p8_u8): Likewise.
2922 (vreinterpretq_p8_u16): Likewise.
2923 (vreinterpretq_p8_u32): Likewise.
2924 (vreinterpretq_p8_u64): Likewise.
2925 (vreinterpretq_p8_p16): Likewise.
2926 (vreinterpret_p16_s8): Likewise.
2927 (vreinterpret_p16_s16): Likewise.
2928 (vreinterpret_p16_s32): Likewise.
2929 (vreinterpret_p16_s64): Likewise.
2930 (vreinterpret_p16_f32): Likewise.
2931 (vreinterpret_p16_u8): Likewise.
2932 (vreinterpret_p16_u16): Likewise.
2933 (vreinterpret_p16_u32): Likewise.
2934 (vreinterpret_p16_u64): Likewise.
2935 (vreinterpret_p16_p8): Likewise.
2936 (vreinterpretq_p16_s8): Likewise.
2937 (vreinterpretq_p16_s16): Likewise.
2938 (vreinterpretq_p16_s32): Likewise.
2939 (vreinterpretq_p16_s64): Likewise.
2940 (vreinterpretq_p16_f32): Likewise.
2941 (vreinterpretq_p16_u8): Likewise.
2942 (vreinterpretq_p16_u16): Likewise.
2943 (vreinterpretq_p16_u32): Likewise.
2944 (vreinterpretq_p16_u64): Likewise.
2945 (vreinterpretq_p16_p8): Likewise.
2946 (vreinterpret_f32_s8): Likewise.
2947 (vreinterpret_f32_s16): Likewise.
2948 (vreinterpret_f32_s32): Likewise.
2949 (vreinterpret_f32_s64): Likewise.
2950 (vreinterpret_f32_u8): Likewise.
2951 (vreinterpret_f32_u16): Likewise.
2952 (vreinterpret_f32_u32): Likewise.
2953 (vreinterpret_f32_u64): Likewise.
2954 (vreinterpret_f32_p8): Likewise.
2955 (vreinterpret_f32_p16): Likewise.
2956 (vreinterpretq_f32_s8): Likewise.
2957 (vreinterpretq_f32_s16): Likewise.
2958 (vreinterpretq_f32_s32): Likewise.
2959 (vreinterpretq_f32_s64): Likewise.
2960 (vreinterpretq_f32_u8): Likewise.
2961 (vreinterpretq_f32_u16): Likewise.
2962 (vreinterpretq_f32_u32): Likewise.
2963 (vreinterpretq_f32_u64): Likewise.
2964 (vreinterpretq_f32_p8): Likewise.
2965 (vreinterpretq_f32_p16): Likewise.
2966 (vreinterpret_s64_s8): Likewise.
2967 (vreinterpret_s64_s16): Likewise.
2968 (vreinterpret_s64_s32): Likewise.
2969 (vreinterpret_s64_f32): Likewise.
2970 (vreinterpret_s64_u8): Likewise.
2971 (vreinterpret_s64_u16): Likewise.
2972 (vreinterpret_s64_u32): Likewise.
2973 (vreinterpret_s64_u64): Likewise.
2974 (vreinterpret_s64_p8): Likewise.
2975 (vreinterpret_s64_p16): Likewise.
2976 (vreinterpretq_s64_s8): Likewise.
2977 (vreinterpretq_s64_s16): Likewise.
2978 (vreinterpretq_s64_s32): Likewise.
2979 (vreinterpretq_s64_f32): Likewise.
2980 (vreinterpretq_s64_u8): Likewise.
2981 (vreinterpretq_s64_u16): Likewise.
2982 (vreinterpretq_s64_u32): Likewise.
2983 (vreinterpretq_s64_u64): Likewise.
2984 (vreinterpretq_s64_p8): Likewise.
2985 (vreinterpretq_s64_p16): Likewise.
2986 (vreinterpret_u64_s8): Likewise.
2987 (vreinterpret_u64_s16): Likewise.
2988 (vreinterpret_u64_s32): Likewise.
2989 (vreinterpret_u64_s64): Likewise.
2990 (vreinterpret_u64_f32): Likewise.
2991 (vreinterpret_u64_u8): Likewise.
2992 (vreinterpret_u64_u16): Likewise.
2993 (vreinterpret_u64_u32): Likewise.
2994 (vreinterpret_u64_p8): Likewise.
2995 (vreinterpret_u64_p16): Likewise.
2996 (vreinterpretq_u64_s8): Likewise.
2997 (vreinterpretq_u64_s16): Likewise.
2998 (vreinterpretq_u64_s32): Likewise.
2999 (vreinterpretq_u64_s64): Likewise.
3000 (vreinterpretq_u64_f32): Likewise.
3001 (vreinterpretq_u64_u8): Likewise.
3002 (vreinterpretq_u64_u16): Likewise.
3003 (vreinterpretq_u64_u32): Likewise.
3004 (vreinterpretq_u64_p8): Likewise.
3005 (vreinterpretq_u64_p16): Likewise.
3006 (vreinterpret_s8_s16): Likewise.
3007 (vreinterpret_s8_s32): Likewise.
3008 (vreinterpret_s8_s64): Likewise.
3009 (vreinterpret_s8_f32): Likewise.
3010 (vreinterpret_s8_u8): Likewise.
3011 (vreinterpret_s8_u16): Likewise.
3012 (vreinterpret_s8_u32): Likewise.
3013 (vreinterpret_s8_u64): Likewise.
3014 (vreinterpret_s8_p8): Likewise.
3015 (vreinterpret_s8_p16): Likewise.
3016 (vreinterpretq_s8_s16): Likewise.
3017 (vreinterpretq_s8_s32): Likewise.
3018 (vreinterpretq_s8_s64): Likewise.
3019 (vreinterpretq_s8_f32): Likewise.
3020 (vreinterpretq_s8_u8): Likewise.
3021 (vreinterpretq_s8_u16): Likewise.
3022 (vreinterpretq_s8_u32): Likewise.
3023 (vreinterpretq_s8_u64): Likewise.
3024 (vreinterpretq_s8_p8): Likewise.
3025 (vreinterpretq_s8_p16): Likewise.
3026 (vreinterpret_s16_s8): Likewise.
3027 (vreinterpret_s16_s32): Likewise.
3028 (vreinterpret_s16_s64): Likewise.
3029 (vreinterpret_s16_f32): Likewise.
3030 (vreinterpret_s16_u8): Likewise.
3031 (vreinterpret_s16_u16): Likewise.
3032 (vreinterpret_s16_u32): Likewise.
3033 (vreinterpret_s16_u64): Likewise.
3034 (vreinterpret_s16_p8): Likewise.
3035 (vreinterpret_s16_p16): Likewise.
3036 (vreinterpretq_s16_s8): Likewise.
3037 (vreinterpretq_s16_s32): Likewise.
3038 (vreinterpretq_s16_s64): Likewise.
3039 (vreinterpretq_s16_f32): Likewise.
3040 (vreinterpretq_s16_u8): Likewise.
3041 (vreinterpretq_s16_u16): Likewise.
3042 (vreinterpretq_s16_u32): Likewise.
3043 (vreinterpretq_s16_u64): Likewise.
3044 (vreinterpretq_s16_p8): Likewise.
3045 (vreinterpretq_s16_p16): Likewise.
3046 (vreinterpret_s32_s8): Likewise.
3047 (vreinterpret_s32_s16): Likewise.
3048 (vreinterpret_s32_s64): Likewise.
3049 (vreinterpret_s32_f32): Likewise.
3050 (vreinterpret_s32_u8): Likewise.
3051 (vreinterpret_s32_u16): Likewise.
3052 (vreinterpret_s32_u32): Likewise.
3053 (vreinterpret_s32_u64): Likewise.
3054 (vreinterpret_s32_p8): Likewise.
3055 (vreinterpret_s32_p16): Likewise.
3056 (vreinterpretq_s32_s8): Likewise.
3057 (vreinterpretq_s32_s16): Likewise.
3058 (vreinterpretq_s32_s64): Likewise.
3059 (vreinterpretq_s32_f32): Likewise.
3060 (vreinterpretq_s32_u8): Likewise.
3061 (vreinterpretq_s32_u16): Likewise.
3062 (vreinterpretq_s32_u32): Likewise.
3063 (vreinterpretq_s32_u64): Likewise.
3064 (vreinterpretq_s32_p8): Likewise.
3065 (vreinterpretq_s32_p16): Likewise.
3066 (vreinterpret_u8_s8): Likewise.
3067 (vreinterpret_u8_s16): Likewise.
3068 (vreinterpret_u8_s32): Likewise.
3069 (vreinterpret_u8_s64): Likewise.
3070 (vreinterpret_u8_f32): Likewise.
3071 (vreinterpret_u8_u16): Likewise.
3072 (vreinterpret_u8_u32): Likewise.
3073 (vreinterpret_u8_u64): Likewise.
3074 (vreinterpret_u8_p8): Likewise.
3075 (vreinterpret_u8_p16): Likewise.
3076 (vreinterpretq_u8_s8): Likewise.
3077 (vreinterpretq_u8_s16): Likewise.
3078 (vreinterpretq_u8_s32): Likewise.
3079 (vreinterpretq_u8_s64): Likewise.
3080 (vreinterpretq_u8_f32): Likewise.
3081 (vreinterpretq_u8_u16): Likewise.
3082 (vreinterpretq_u8_u32): Likewise.
3083 (vreinterpretq_u8_u64): Likewise.
3084 (vreinterpretq_u8_p8): Likewise.
3085 (vreinterpretq_u8_p16): Likewise.
3086 (vreinterpret_u16_s8): Likewise.
3087 (vreinterpret_u16_s16): Likewise.
3088 (vreinterpret_u16_s32): Likewise.
3089 (vreinterpret_u16_s64): Likewise.
3090 (vreinterpret_u16_f32): Likewise.
3091 (vreinterpret_u16_u8): Likewise.
3092 (vreinterpret_u16_u32): Likewise.
3093 (vreinterpret_u16_u64): Likewise.
3094 (vreinterpret_u16_p8): Likewise.
3095 (vreinterpret_u16_p16): Likewise.
3096 (vreinterpretq_u16_s8): Likewise.
3097 (vreinterpretq_u16_s16): Likewise.
3098 (vreinterpretq_u16_s32): Likewise.
3099 (vreinterpretq_u16_s64): Likewise.
3100 (vreinterpretq_u16_f32): Likewise.
3101 (vreinterpretq_u16_u8): Likewise.
3102 (vreinterpretq_u16_u32): Likewise.
3103 (vreinterpretq_u16_u64): Likewise.
3104 (vreinterpretq_u16_p8): Likewise.
3105 (vreinterpretq_u16_p16): Likewise.
3106 (vreinterpret_u32_s8): Likewise.
3107 (vreinterpret_u32_s16): Likewise.
3108 (vreinterpret_u32_s32): Likewise.
3109 (vreinterpret_u32_s64): Likewise.
3110 (vreinterpret_u32_f32): Likewise.
3111 (vreinterpret_u32_u8): Likewise.
3112 (vreinterpret_u32_u16): Likewise.
3113 (vreinterpret_u32_u64): Likewise.
3114 (vreinterpret_u32_p8): Likewise.
3115 (vreinterpret_u32_p16): Likewise.
3116 (vreinterpretq_u32_s8): Likewise.
3117 (vreinterpretq_u32_s16): Likewise.
3118 (vreinterpretq_u32_s32): Likewise.
3119 (vreinterpretq_u32_s64): Likewise.
3120 (vreinterpretq_u32_f32): Likewise.
3121 (vreinterpretq_u32_u8): Likewise.
3122 (vreinterpretq_u32_u16): Likewise.
3123 (vreinterpretq_u32_u64): Likewise.
3124 (vreinterpretq_u32_p8): Likewise.
3125 (vreinterpretq_u32_p16): Likewise.
3127 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3129 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
3131 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
3133 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
3134 (vqnegd_s64): Likewise.
3135 (vqabs_s64): Likewise.
3136 (vqabsd_s64): Likewise.
3138 2014-04-22 Richard Henderson <rth@redhat.com>
3140 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
3141 computation to the top of the loop.
3143 2014-04-22 Renlin <renlin.li@arm.com>
3144 Jiong Wang <jiong.wang@arm.com>
3146 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
3147 * config/aarch64/aarch64.c (aarch64_layout_frame)
3148 (aarch64_initial_elimination_offset): Likewise.
3150 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
3152 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
3155 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
3157 * machmode.h (bitwise_mode_for_mode): Declare.
3158 * stor-layout.h (bitwise_type_for_mode): Likewise.
3159 * stor-layout.c (bitwise_mode_for_mode): New function.
3160 (bitwise_type_for_mode): Likewise.
3161 * builtins.c (fold_builtin_memory_op): Use it instead of
3162 int_mode_for_mode and build_nonstandard_integer_type.
3164 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3166 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
3167 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
3168 (*-*-solaris2*): Simplify.
3169 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
3170 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
3171 *-*-solaris2.9* handling.
3173 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
3175 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
3176 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
3178 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
3179 * configure: Regenerate.
3181 * config/i386/sol2-9.h: Remove.
3183 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
3184 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
3185 Remove Solaris 9 references.
3187 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
3189 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
3190 (floatuns<GPI:mode><GPF:mode>2): Remove.
3191 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
3192 and floatuns conversions.
3193 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
3194 and floatuns conversions.
3195 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
3196 (w1,w2): New mode attributes for inequal width conversions.
3198 2014-04-22 Renlin Li <Renlin.Li@arm.com>
3200 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
3201 the output asm format.
3203 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
3205 * config/aarch64/aarch64-simd.md
3206 (aarch64_cm<optab>di): Always split.
3207 (*aarch64_cm<optab>di): New.
3208 (aarch64_cmtstdi): Always split.
3209 (*aarch64_cmtstdi): New.
3211 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3213 PR tree-optimization/60823
3214 * omp-low.c (ipa_simd_modify_function_body): Go through
3215 all SSA_NAMEs and for those refering to vector arguments
3216 which are going to be replaced adjust SSA_NAME_VAR and,
3217 if it is a default definition, change it into a non-default
3218 definition assigned at the beginning of function from new_decl.
3219 (ipa_simd_modify_stmt_ops): Rewritten.
3220 * tree-dfa.c (set_ssa_default_def): When removing default def,
3221 check for NULL loc instead of NULL *loc.
3223 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3225 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
3226 restrictions on core registers for DImode values in Thumb2.
3228 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3230 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
3231 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
3233 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3235 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
3236 (*iordi_notzesidi_di): Likewise.
3237 (*iordi_notsesidi_di): Likewise.
3239 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3241 * config/arm/arm-protos.h (tune_params): New struct members.
3242 * config/arm/arm.c: Initialise tune_params per processor.
3243 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
3244 for speed, based on new tune_params.
3246 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3248 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
3249 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
3250 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
3251 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
3252 * config/aarch64/arm_neon.h (vrnd_f64): Added.
3253 (vrnda_f64): Likewise.
3254 (vrndi_f64): Likewise.
3255 (vrndm_f64): Likewise.
3256 (vrndn_f64): Likewise.
3257 (vrndp_f64): Likewise.
3258 (vrndx_f64): Likewise.
3260 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3262 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
3263 GET_MODE_SIZE argument is enum machine_mode.
3265 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3268 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
3269 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
3271 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
3274 * asan.c (asan_emit_stack_protection): Force the base to align to
3275 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
3276 appropriate bits if STRICT_ALIGNMENT.
3277 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
3279 (expand_used_vars): Leave a space in the stack frame for alignment
3280 if STRICT_ALIGNMENT.
3282 2014-04-21 David Malcolm <dmalcolm@redhat.com>
3284 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
3286 (gimple_store_p): Likewise.
3287 (gimple_assign_load_p): Likewise.
3288 (gimple_assign_cast_p): Likewise.
3289 (gimple_clobber_p): Likewise.
3291 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
3292 rather than a gimple.
3293 (gimple_assign_cast_p): Likewise.
3295 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
3298 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
3299 If mode is DDmode and TARGET_E500_DOUBLE allow move.
3301 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
3302 more debug information for E500 if -mdebug=reg.
3304 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
3307 * config/i386/i386.c (ix86_expand_builtin)
3308 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
3309 register for target RTX.
3310 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
3312 2014-04-18 Cong Hou <congh@google.com>
3314 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
3315 the widen-mult pattern by handling two operands with different sizes,
3316 and operands whose size is smaller than half of the result type.
3318 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3320 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
3321 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
3322 (do_estimate_edge_time): Compute it.
3323 * ipa-inline.c (want_inline_small_function_p): Bypass
3324 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
3326 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3328 * ipa-inline.c (spec_rem): New static variable.
3329 (dump_overall_stats): New function.
3330 (dump_inline_stats): New function.
3332 2014-04-18 Richard Henderson <rth@redhat.com>
3334 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
3335 to GET_MODE_SIZE, not a reg_class_t.
3337 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3339 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
3340 (vsx_xxmrglw_<mode>): Likewise.
3342 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3345 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
3346 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
3347 (rs6000_init_hard_regno_mode_ok): Likewise.
3349 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
3351 * ipa-inline.c (inline_small_functions): Account only non-cold
3353 * doc/invoke.texi (inline-unit-growth): Update documentation.
3355 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
3357 * config/rs6000/rs6000.md (addti3, subti3): New.
3359 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
3362 * config/i386/i386.c (ix86_expand_clear): Remove outdated
3363 comment. Check optimize_insn_for_size_p instead of
3364 optimize_insn_for_speed_p.
3366 2014-04-17 Martin Jambor <mjambor@suse.cz>
3368 * gimple-iterator.c (gsi_start_edge): New function.
3369 * gimple-iterator.h (gsi_start_edge): Declare.
3370 * tree-sra.c (single_non_eh_succ): New function.
3371 (disqualify_ops_if_throwing_stmt): Renamed to
3372 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
3373 having one non-EH successor BB.
3374 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
3375 generate loads into replacements.
3376 (sra_modify_assign): Likewise and and also use the simple path for
3378 (sra_modify_function_body): Commit statements on edges.
3380 2014-04-17 Richard Biener <rguenther@suse.de>
3383 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
3384 comparison results and add clarifying comment.
3386 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3388 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
3389 (blank_mode): Initialize it.
3390 (emit_mode_size_inline, emit_mode_nunits_inline,
3391 emit_mode_inner_inline): New functions.
3392 (emit_insn_modes_h): Call them and surround their output with
3393 #if GCC_VERSION >= 4001 ... #endif.
3394 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
3395 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
3396 mode_* arrays if the argument is __builtin_constant_p.
3397 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
3398 is enum machine_mode.
3400 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3402 * passes.c (opt_pass::execute): Adjust.
3403 (pass_manager::execute_pass_mode_switching): Likewise.
3404 (early_local_passes::execute): Likewise.
3405 (execute_one_pass): Pass cfun to the pass's execute method.
3406 * tree-pass.h (opt_pass::execute): Add function * argument.
3407 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3408 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3409 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3410 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3411 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3412 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
3413 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
3414 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3415 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3416 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3417 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3418 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3419 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3420 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3421 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3422 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3423 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3424 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3425 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3426 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3427 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3428 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3429 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3430 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3431 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3432 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3433 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3434 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3435 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3438 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3440 * passes.c (opt_pass::gate): Take function * argument.
3441 (gate_all_early_local_passes): Merge into
3442 (early_local_passes::gate): this.
3443 (gate_all_early_optimizations): Merge into
3444 (all_early_optimizations::gate): this.
3445 (gate_all_optimizations): Mege into
3446 (all_optimizations::gate): this.
3447 (gate_all_optimizations_g): Merge into
3448 (all_optimizations_g::gate): this.
3449 (gate_rest_of_compilation): Mege into
3450 (rest_of_compilation::gate): this.
3451 (gate_postreload): Merge into
3452 (postreload::gate): this.
3453 (dump_one_pass): Pass cfun to the pass's gate method.
3454 (execute_ipa_summary_passes): Likewise.
3455 (execute_one_pass): Likewise.
3456 (ipa_write_summaries_2): Likewise.
3457 (ipa_write_optimization_summaries_1): Likewise.
3458 (ipa_read_summaries_1): Likewise.
3459 (ipa_read_optimization_summaries_1): Likewise.
3460 (execute_ipa_stmt_fixups): Likewise.
3461 * tree-pass.h (opt_pass::gate): Add function * argument.
3462 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
3463 combine-stack-adj.c, combine.c, compare-elim.c,
3464 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3465 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
3466 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
3467 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
3468 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3469 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3470 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3471 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
3472 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
3473 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
3474 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3475 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3476 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
3477 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3478 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3479 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3480 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3481 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3482 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3483 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3484 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3485 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3486 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3487 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
3488 var-tracking.c, vtable-verify.c, web.c: Adjust.
3490 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3492 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
3493 * configure: Regenerate.
3495 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3497 * passes.c (dump_one_pass): don't check pass->has_gate.
3498 (execute_ipa_summary_passes): Likewise.
3499 (execute_one_pass): Likewise.
3500 (ipa_write_summaries_2): Likewise.
3501 (ipa_write_optimization_summaries_1): Likewise.
3502 (ipa_read_optimization_summaries_1): Likewise.
3503 (execute_ipa_stmt_fixups): Likewise.
3504 * tree-pass.h (pass_data::has_gate): Remove.
3505 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3506 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3507 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3508 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3509 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3510 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
3511 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
3512 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
3513 gimple-low.c, gimple-ssa-isolate-paths.c,
3514 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
3515 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3516 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3517 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
3518 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
3519 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
3520 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
3521 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3522 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
3523 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3524 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3525 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3526 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3527 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3528 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3529 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3530 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3531 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3532 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3533 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3534 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3535 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3538 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3540 * pass_manager.h (pass_manager::register_dump_files_1): Remove
3542 * passes.c (pass_manager::register_dump_files_1): Merge into
3543 (pass_manager::register_dump_files): this, and remove its handling of
3544 properties since the pass always has the properties anyway.
3545 (pass_manager::pass_manager): Adjust.
3547 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3549 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
3550 * passes.c (pass_manager::register_dump_files_1): Remove dead code
3551 dealing with properties.
3552 (pass_manager::register_dump_files): Adjust.
3554 2014-03-20 Mark Wielaard <mjw@redhat.com>
3556 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
3557 then represent the bound as normal constant value.
3559 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3562 Forward port from 4.8 branch
3563 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
3565 * config/i386/bmiintrin.h (_blsi_u32): New.
3569 (_blsmsk_u32): Ditto.
3570 (_blsmsk_u64): Ditto.
3571 (_tzcnt_u32): Ditto.
3572 (_tzcnt_u64): Ditto.
3574 2014-04-17 Kito Cheng <kito@0xlab.org>
3576 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
3578 2014-04-17 Richard Biener <rguenther@suse.de>
3581 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
3582 boolean results for comparisons.
3584 2014-04-17 Richard Biener <rguenther@suse.de>
3586 PR tree-optimization/60836
3587 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
3588 initial PHI args to be gimple values.
3590 2014-04-17 Richard Biener <rguenther@suse.de>
3592 PR tree-optimization/60841
3593 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
3594 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
3595 of stmts to SLP build.
3596 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
3597 (vect_analyze_slp): Likewise.
3598 (vect_analyze_slp_instance): Likewise.
3599 (vect_build_slp_tree): Limit overall SLP tree growth.
3600 * tree-vectorizer.h (vect_analyze_data_refs,
3601 vect_analyze_slp): Adjust prototypes.
3603 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3605 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
3608 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3610 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
3611 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
3612 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
3613 for TARGET_SLOW_PSHUFB
3615 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3617 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
3618 * config/i386/i386.c (intel_cost): Ditto.
3620 2014-04-17 Joey Ye <joey.ye@arm.com>
3622 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
3624 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3626 * opts.c (common_handle_option): Disable -fipa-reference coorectly
3629 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3631 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
3632 (type_all_derivations_known_p): New predicate.
3633 (type_all_ctors_visible_p): New predicate.
3634 (type_possibly_instantiated_p): New predicate.
3635 (get_odr_type): Compute all_derivations_known.
3636 (dump_odr_type): Dump the flag.
3637 (maybe_record_type): Cleanup.
3638 (record_target_from_binfo): Add bases_to_consider array;
3639 record bases for types w/o instances and skip CXX destructor.
3640 (possible_polymorphic_call_targets_1): Add bases_to_consider
3641 and consider_construction parameters; check if type may have instance.
3642 (get_polymorphic_call_info): Set maybe_in_construction to true
3643 when we know nothing.
3644 (record_targets_from_bases): Skip CXX destructors; they are
3645 never called for types in construction.
3646 (possible_polymorphic_call_targets): Do not record target when
3647 type may not have instance.
3649 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3652 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
3653 external aliases alive, too.
3655 2014-04-16 Andrew Pinski <apinski@cavium.com>
3657 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
3660 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3662 * final.c (compute_alignments): Do not apply loop alignment to a block
3663 falling through to the exit.
3665 2014-04-16 Catherine Moore <clm@codesourcery.com>
3667 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
3668 Adjust constraints for microMIPS store patterns.
3670 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3672 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
3674 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3676 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
3677 (append_use): Run at -O0.
3678 (append_vdef): Likewise.
3679 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
3680 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
3682 2014-04-16 Jakub Jelinek <jakub@redhat.com>
3684 PR tree-optimization/60844
3685 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
3686 (propagate_op_to_single_use, remove_visited_stmt_chain,
3687 linearize_expr, repropagate_negates, reassociate_bb): Use it
3688 instead of gsi_remove.
3690 2014-04-16 Martin Jambor <mjambor@suse.cz>
3692 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
3693 ipa_transforms_to_apply.
3694 (cgraph_function_versioning): Assert that old_node has empty
3695 ipa_transforms_to_apply.
3696 * trans-mem.c (ipa_tm_create_version): Likewise.
3697 * tree-inline.c (tree_function_versioning): Do not duplicate
3698 ipa_transforms_to_apply.
3700 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3703 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
3705 Pass necessary as flags on 64-bit Solaris/x86.
3706 Use lowercase relocs for x86_64-*-*.
3707 * configure: Regenerate.
3709 2014-04-15 Jan Hubicka <jh@suse.cz>
3711 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
3712 (maybe_record_node, likely_target_p): Use it.
3714 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3717 Revert following patch
3719 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3722 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3723 software floating point or no floating point registers, do not
3724 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3725 in GPRs that occurs after we tested for GPRs that would never be
3728 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3729 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3730 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3731 specifically allow DDmode, since that does not use the SPE SIMD
3734 2014-03-21 Mark Wielaard <mjw@redhat.com>
3736 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
3737 as unsigned or int depending on type and value used.
3739 2014-04-15 Richard Biener <rguenther@suse.de>
3741 PR rtl-optimization/56965
3742 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
3743 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
3745 * alias.c (true_dependence_1): Do not call
3746 nonoverlapping_component_refs_p.
3747 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
3748 nonoverlapping_component_refs_p.
3749 (indirect_refs_may_alias_p): Likewise.
3751 2014-04-15 Teresa Johnson <tejohnson@google.com>
3753 * cfg.c (dump_bb_info): Fix flags check.
3754 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
3756 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3758 PR rtl-optimization/60663
3759 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
3762 2014-04-15 Richard Biener <rguenther@suse.de>
3764 * lto-streamer.h (LTO_major_version): Bump to 4.
3766 2014-04-15 Richard Biener <rguenther@suse.de>
3768 * common.opt (lto_partition_model): New enum.
3769 (flto-partition=): Merge separate options with a single with argument,
3770 add -flto-partition=one support.
3771 * flag-types.h (enum lto_partition_model): Declare.
3772 * opts.c (finish_options): Remove duplicate -flto-partition=
3774 * lto-wrapper.c (run_gcc): Adjust.
3776 2014-04-15 Richard Biener <rguenther@suse.de>
3778 * alias.c (ncr_compar): New function.
3779 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
3781 2014-04-15 Richard Biener <rguenther@suse.de>
3783 * alias.c (record_component_aliases): Do not walk BINFOs.
3785 2014-04-15 Richard Biener <rguenther@suse.de>
3787 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
3788 Add struct function argument and adjust.
3789 (find_func_aliases_for_call): Likewise.
3790 (find_func_aliases): Likewise.
3791 (find_func_clobbers): Likewise.
3792 (intra_create_variable_infos): Likewise.
3793 (compute_points_to_sets): Likewise.
3794 (ipa_pta_execute): Adjust. Do not push/pop cfun.
3796 2014-04-15 Richard Biener <rguenther@suse.de>
3798 * tree.c (iterative_hash_expr): Use enum tree_code_class
3799 to store TREE_CODE_CLASS.
3800 (tree_block): Likewise.
3801 (tree_set_block): Likewise.
3802 * tree.h (fold_build_pointer_plus_loc): Use
3803 convert_to_ptrofftype_loc.
3805 2014-04-15 Jakub Jelinek <jakub@redhat.com>
3808 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
3811 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
3813 * cfgloop.h (struct loop): Move force_vectorize down.
3814 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
3815 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
3816 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
3817 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
3818 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
3819 * tree-core.h (enum annot_expr_kind): Add new kind values.
3820 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
3821 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
3823 * tree.def (ANNOTATE_EXPR): Tweak comment.
3825 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3827 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
3830 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
3832 * tree.h (TYPE_IDENTIFIER): Declare.
3833 * tree.c (subrange_type_for_debug_p): Use it.
3834 * godump.c (go_format_type): Likewise.
3835 * dwarf2out.c (is_cxx_auto, modified_type_die,
3836 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
3837 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
3839 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3842 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
3844 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
3846 * config/i386/i386.c (examine_argument): Return bool. Return true if
3847 parameter should be passed in memory.
3848 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
3849 (construct_container): Update calls to examine_argument.
3850 (function_arg_advance_64): Ditto.
3851 (return_in_memory_32): Merge with ix86_return_in_memory.
3852 (return_in_memory_64): Ditto.
3853 (return_in_memory_ms_64): Ditto.
3855 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3857 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
3858 * coverage.c (coverage_compute_profile_id): Handle externally visible
3861 2014-04-14 Martin Jambor <mjambor@suse.cz>
3863 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
3864 DECL_DISREGARD_INLINE_LIMITS functions.
3866 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3869 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
3871 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3874 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
3875 optimize_insn_for_speed_p instead of
3876 optimize_function_for_speed_p.
3878 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
3880 * doc/invoke.texi (free): Document AArch64.
3882 2014-04-14 Richard Biener <rguenther@suse.de>
3884 PR tree-optimization/60042
3885 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
3886 (insert_into_preds_of_block): Do not prevent PHI insertion
3887 for REFERENCE exprs here ...
3888 (eliminate_dom_walker::before_dom_children): ... but prevent
3889 their use here under similar conditions when applied to the
3890 IL after PRE optimizations.
3892 2014-04-14 Richard Biener <rguenther@suse.de>
3894 * passes.def: Move early points-to after early SRA.
3896 2014-04-14 Richard Biener <rguenther@suse.de>
3898 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
3899 check for which sign-changes we allow when forwarding
3900 a converted value into a switch.
3902 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3904 * stor-layout.c (place_field): Finalize non-constant offset for the
3907 2014-04-14 Richard Biener <rguenther@suse.de>
3909 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
3911 (expand_switch_using_bit_tests_p): Likewise.
3912 (process_switch): Compute and pass on speed_p based on the
3914 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
3915 optimize_bb_for_speed_p.
3917 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3919 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
3920 * function.h (struct function): Rename has_force_vect_loops into
3921 has_force_vectorize_loops.
3922 * lto-streamer-in.c (input_cfg): Adjust for renaming.
3923 (input_struct_function_base): Likewise.
3924 * lto-streamer-out.c (output_cfg): Likewise.
3925 (output_struct_function_base): Likewise.
3926 * omp-low.c (expand_omp_simd): Likewise.
3927 * tree-cfg.c (move_sese_region_to_fn): Likewise.
3928 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
3929 (version_loop_for_if_conversion): Likewise.
3930 (tree_if_conversion): Likewise.
3931 (main_tree_if_conversion): Likewise.
3932 (gate_tree_if_conversion): Likewise.
3933 * tree-inline.c (copy_loops): Likewise.
3934 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
3935 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
3936 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
3937 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
3938 * tree-vectorizer.c (vectorize_loops): Likewise.
3939 * tree-vectorizer.h (unlimited_cost_model): Likewise.
3941 2014-04-14 Richard Biener <rguenther@suse.de>
3944 * lto-streamer-out.c (wrap_refs): New function.
3945 (lto_output): Wrap symbol references in global initializes in
3946 type-preserving MEM_REFs.
3948 2014-04-14 Christian Bruel <christian.bruel@st.com>
3950 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
3952 2014-04-14 Christian Bruel <christian.bruel@st.com>
3954 * config/sh/sh.md (setmemqi): New expand pattern.
3955 * config/sh/sh.h (CLEAR_RATIO): Define.
3956 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
3957 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
3959 2014-04-14 Richard Biener <rguenther@suse.de>
3962 * fold-const.c (negate_expr_p): Don't negate directional rounding
3964 (fold_negate_expr): Likewise.
3966 2014-04-14 Richard Biener <rguenther@suse.de>
3968 PR tree-optimization/59817
3969 PR tree-optimization/60453
3970 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
3971 recursion to catch all CHRECs in the scalar evolution and restrict
3972 the predicate for the remains appropriately.
3974 2014-04-12 Catherine Moore <clm@codesourcery.com>
3976 * config/mips/constraints.md: Add new register constraint "kb".
3977 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
3978 (*movhi_internal): Likewise.
3979 (*movqi_internal): Likewise.
3980 * config/mips/mips.h (M16_STORE_REGS): New register class.
3981 (REG_CLASS_NAMES): Add M16_STORE_REGS.
3982 (REG_CLASS_CONTENTS): Likewise.
3983 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
3985 2014-04-11 Tobias Burnus <burnus@net-b.de>
3988 * doc/invoke.texi (-Wformat-signedness): Document it.
3989 (Wformat=2): Mention that this enables -Wformat-signedness.
3991 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
3993 * common/config/epiphany/epiphany-common.c
3994 (epiphany_option_optimization_table): Enable section anchors by
3995 default at -O1 or higher.
3996 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
3997 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
3998 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
3999 carries no extra cost.
4000 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
4001 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
4002 * config/epiphany/predicates.md (memclob_operand): New predicate.
4003 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
4004 Use memclob_operand predicate and X constraint for operand 3.
4006 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4008 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
4009 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
4012 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4014 PR rtl-optimization/60651
4015 * mode-switching.c (optimize_mode_switching): Make sure to emit
4016 sets of a lower numbered entity before sets of a higher numbered
4017 entity to a mode of the same or lower priority.
4018 When creating a seginfo for a basic block that starts with a code
4019 label, move the insertion point past the code label.
4020 (new_seginfo): Document and enforce requirement that
4021 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
4022 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
4023 * doc/tm.texi: Regenerate.
4025 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
4028 * config/arc/arc.c (arc_save_restore): Fix assert typo.
4030 2013-04-11 Jakub Jelinek <jakub@redhat.com>
4032 * BASE-VER: Set to 4.10.0.
4034 2014-04-11 Tobias Burnus <burnus@net-b.de>
4037 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
4038 * doc/gcc.texi (Service): Update description in the @menu
4039 * doc/invoke.texi (Option Summary): Remove misplaced and
4042 2014-04-11 Steve Ellcey <sellcey@mips.com>
4043 Jakub Jelinek <jakub@redhat.com>
4046 * expr.c (convert_move): Use emit_store_flag_force instead of
4047 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
4050 2014-04-11 Richard Biener <rguenther@suse.de>
4053 * varasm.c (assemble_alias): Avoid endless error reporting
4054 recursion by setting TREE_ASM_WRITTEN.
4056 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4058 * config/s390/s390.md: Add a splitter for NOT rtx.
4060 2014-04-11 Jakub Jelinek <jakub@redhat.com>
4062 PR rtl-optimization/60663
4063 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
4065 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
4066 Jakub Jelinek <jakub@redhat.com>
4069 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
4070 flag from decl_node to node.
4072 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4075 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
4076 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
4077 ameliorating the cases where it can be.
4079 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
4082 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4084 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4085 (loadsync_<mode>): Change mode.
4086 (load_quadpti, store_quadpti): New.
4087 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4088 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4089 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
4091 2014-04-09 Cong Hou <congh@google.com>
4094 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
4097 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4099 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
4100 instead of vnor to exploit possible fusion opportunity in the
4102 (altivec_expand_vec_perm_const_le): Likewise.
4104 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4106 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4107 (loadsync_<mode>): Change mode.
4108 (load_quadpti, store_quadpti): New.
4109 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4110 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4112 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
4115 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
4116 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
4117 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
4119 2014-04-08 Richard Biener <rguenther@suse.de>
4122 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
4123 a 64bit widest int print double-int similar to on HWI64 hosts.
4125 2014-04-08 Richard Biener <rguenther@suse.de>
4127 PR tree-optimization/60785
4128 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
4129 default defs properly.
4131 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
4133 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
4134 (Weffc++): Likewise.
4136 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
4138 * ipa-devirt.c (maybe_record_node): When node is not recorded,
4139 set completep to false rather than true.
4141 2014-04-07 Douglas B Rupp <rupp@adacore.com>
4144 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
4145 ARM_TARGET2_DWARF_FORMAT.
4147 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
4150 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
4151 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
4154 2014-04-07 Richard Biener <rguenther@suse.de>
4156 PR tree-optimization/60766
4157 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
4158 (may_eliminate_iv): Convert cand_value_at result to desired type.
4160 2014-04-07 Jason Merrill <jason@redhat.com>
4163 * common.opt (-fno-gnu-unique): Add.
4164 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
4166 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4168 * haifa-sched.c: Fix outdated function reference and minor
4169 grammar errors in introductory comment.
4171 2014-04-07 Richard Biener <rguenther@suse.de>
4174 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
4176 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
4178 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
4181 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
4182 size accounting for thunks.
4183 (pa_asm_output_mi_thunk): Use final_start_function() and
4184 final_end_function() to output function start and end directives.
4186 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4188 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
4189 device specific ISA/ feature information. Remove short_sp and
4190 errata_skip ds. Add avr_device_specific_features enum to have device
4192 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
4193 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
4194 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
4195 updated device specific info.
4196 * config/avr/avr-mcus.def: Merge device specific details to
4197 dev_attribute field.
4198 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
4200 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
4201 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
4202 assembler if RMW isa supported by current device.
4203 * config/avr/genmultilib.awk: Update as device info structure changed.
4204 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
4206 2014-04-04 Cong Hou <congh@google.com>
4208 PR tree-optimization/60656
4209 * tree-vect-stmts.c (supportable_widening_operation):
4210 Fix a bug that elements in a vector with vect_used_by_reduction
4211 property are incorrectly reordered when the operation on it is not
4212 consistant with the one in reduction operation.
4214 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
4216 PR rtl-optimization/60155
4217 * gcse.c (record_set_data): New function.
4218 (single_set_gcse): New function.
4219 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
4220 (hoist_code): Likewise.
4221 (get_pressure_class_and_nregs): Likewise.
4223 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
4225 * explow.c (probe_stack_range): Emit a final optimization blockage.
4227 2014-04-04 Anthony Green <green@moxielogic.com>
4229 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
4232 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
4235 * lto-cgraph.c (input_overwrite_node): Check that partitioning
4236 flags are set only during streaming.
4237 * ipa.c (process_references, walk_polymorphic_call_targets,
4238 symtab_remove_unreachable_nodes): Drop bodies of always inline
4239 after early inlining.
4240 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
4242 2014-04-04 Jakub Jelinek <jakub@redhat.com>
4243 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4246 * dwarf2out.c (const_ok_for_output_1): Reject expressions
4249 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4252 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
4254 (cortex_a53_fdivd): Likewise.
4256 2014-04-04 Martin Jambor <mjambor@suse.cz>
4259 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
4261 * cgraph.c (clone_of_p): Also return true if thunks match.
4262 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
4263 cgraph_function_or_thunk_node and an obsolete comment.
4264 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
4266 (build_function_decl_skip_args): Likewise.
4267 (set_new_clone_decl_and_node_flags): New function.
4268 (duplicate_thunk_for_node): Likewise.
4269 (redirect_edge_duplicating_thunks): Likewise.
4270 (cgraph_clone_node): New parameter args_to_skip, pass it to
4271 redirect_edge_duplicating_thunks which is called instead of
4272 cgraph_redirect_edge_callee.
4273 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
4274 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
4276 2014-04-04 Jeff Law <law@redhat.com>
4279 * config/arm/predicates.md (const_int_I_operand): New predicate.
4280 (const_int_M_operand): Similarly.
4281 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
4283 (insv_t2, extv_reg, extzv_t2): Likewise.
4284 (load_multiple_with_writeback): Similarly for const_int_I_operand.
4285 (pop_multiple_with_writeback_and_return): Likewise.
4286 (vfp_pop_multiple_with_writeback): Likewise
4288 2014-04-04 Richard Biener <rguenther@suse.de>
4291 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
4292 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
4294 * gimplify.h (gimple_add_tmp_var_fn): Declare.
4295 * gimplify.c (gimple_add_tmp_var_fn): New function.
4296 * gimple-expr.h (create_tmp_reg_fn): Declare.
4297 * gimple-expr.c (create_tmp_reg_fn): New function.
4298 * gimple-low.c (record_vars_into): Don't change cfun.
4299 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
4300 code generation without cfun.
4302 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
4305 * Makefile.in (install-driver): Fix shell scripting.
4307 2014-04-03 Cong Hou <congh@google.com>
4309 PR tree-optimization/60505
4310 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
4311 threshold of number of iterations below which no vectorization
4313 * tree-vect-loop.c (new_loop_vec_info):
4314 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
4315 * tree-vect-loop.c (vect_analyze_loop_operations):
4316 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
4317 * tree-vect-loop.c (vect_transform_loop):
4318 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
4319 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
4320 of iterations of the loop and see if we should build the epilogue.
4322 2014-04-03 Richard Biener <rguenther@suse.de>
4324 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
4325 (streamer_tree_cache_create): Adjust.
4326 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
4327 to allow optional nodes array.
4328 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
4329 (streamer_tree_cache_append): Likewise.
4330 (streamer_tree_cache_create): Create nodes array optionally
4331 as specified by parameter.
4332 * lto-streamer-out.c (create_output_block): Avoid maintaining
4333 the node array in the writer cache.
4334 (DFS_write_tree): Remove assertion.
4335 (produce_asm_for_decls): Free the out decl state hash table early.
4336 * lto-streamer-in.c (lto_data_in_create): Adjust for
4337 streamer_tree_cache_create prototype change.
4339 2014-04-03 Richard Biener <rguenther@suse.de>
4341 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
4342 set TREE_CHAIN to NULL_TREE.
4344 2014-04-03 Richard Biener <rguenther@suse.de>
4346 PR tree-optimization/60740
4347 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
4348 over all GIMPLE_COND operands.
4350 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
4352 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
4353 (Weffc++): Remove Scott's numbering, merge lists and reference
4356 2014-04-03 Nick Clifton <nickc@redhat.com>
4358 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
4361 2014-04-03 Martin Jambor <mjambor@suse.cz>
4363 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
4364 mention gcc_unreachable before failing.
4365 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
4368 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
4371 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
4372 inconsistent code and instead mark the context inconsistent.
4373 (possible_polymorphic_call_targets): For inconsistent contexts
4374 return empty complete list.
4376 2014-04-02 Anthony Green <green@moxielogic.com>
4378 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
4379 (extendqisi2, extendhisi2): Define.
4380 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
4381 (WCHAR_TYPE): Change to unsigned int.
4383 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4385 PR tree-optimization/60733
4386 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
4387 insertion point for PHI candidates to be the end of the feeding
4388 block for the PHI argument.
4390 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
4392 PR rtl-optimization/60650
4393 * lra-constraints.c (process_alt_operands): Decrease reject for
4394 earlyclobber matching.
4396 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4398 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
4400 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4402 * config/spu/spu.c (pad_bb): Do not crash when the last
4403 insn is CODE_FOR_blockage.
4405 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4407 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
4408 lies outside the target mode.
4410 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4413 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4414 software floating point or no floating point registers, do not
4415 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4416 in GPRs that occurs after we tested for GPRs that would never be
4419 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4420 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4421 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4422 specifically allow DDmode, since that does not use the SPE SIMD
4425 2014-04-02 Richard Biener <rguenther@suse.de>
4428 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4429 MODE_INTs. Properly use negv_optab.
4430 (expand_abs): Likewise.
4432 2014-04-02 Richard Biener <rguenther@suse.de>
4435 * Makefile.in (install-driver): Guard extra installs with special
4438 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4440 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4443 2014-04-01 Richard Henderson <rth@redhat.com>
4446 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
4447 alternative enabled before register allocation.
4449 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
4451 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
4452 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
4454 (nios2_large_got_address): Remove unneeded 'sym' parameter.
4455 (nios2_got_address): Update nios2_large_got_address call site.
4456 (nios2_delegitimize_address): New function.
4457 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
4458 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
4459 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
4461 2014-04-01 Martin Husemann <martin@duskware.de>
4463 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
4466 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
4468 PR rtl-optimization/60604
4469 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
4470 check from register_operand.
4471 (register_operand): Redefine in terms of general_operand.
4472 (nonmemory_operand): Use register_operand for the non-constant cases.
4474 2014-04-01 Richard Biener <rguenther@suse.de>
4476 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
4478 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
4480 * doc/invoke.texi (mapp-regs): Clarify.
4482 2014-03-31 Ulrich Drepper <drepper@gmail.com>
4484 * config/i386/avx512fintrin.h (__v32hi): Define type.
4485 (__v64qi): Likewise.
4486 (_mm512_set1_epi8): Define.
4487 (_mm512_set1_epi16): Define.
4488 (_mm512_set4_epi32): Define.
4489 (_mm512_set4_epi64): Define.
4490 (_mm512_set4_pd): Define.
4491 (_mm512_set4_ps): Define.
4492 (_mm512_setr4_epi64): Define.
4493 (_mm512_setr4_epi32): Define.
4494 (_mm512_setr4_pd): Define.
4495 (_mm512_setr4_ps): Define.
4496 (_mm512_setzero_epi32): Define.
4498 2014-03-31 Martin Jambor <mjambor@suse.cz>
4501 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
4502 callsite_arguments_match_p. Updated all callers. Also check types of
4503 corresponding formal parameters and actual arguments.
4504 (not_all_callers_have_enough_arguments_p) Renamed to
4505 some_callers_have_mismatched_arguments_p.
4507 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
4509 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
4511 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
4514 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
4517 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
4519 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
4520 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
4522 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
4523 Use FMAMODE_NOVF512 mode iterator.
4524 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
4525 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
4526 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
4527 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
4529 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
4530 Use VF_128_256 mode iterator.
4531 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
4534 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4536 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
4537 static chain if needed.
4539 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4542 * lra-constraints.c (index_part_to_reg): New.
4543 (process_address): Use it.
4545 2014-03-27 Jeff Law <law@redhat.com>
4546 Jakub Jelinek <jakub@redhat.com>
4549 * expr.c (do_tablejump): Use simplify_gen_binary rather than
4550 gen_rtx_{PLUS,MULT} to build up the address expression.
4552 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
4553 creating non-canonical RTL.
4555 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4558 * ipa-inline.c (want_inline_small_function_p): Short circuit large
4559 functions; reorganize to make cheap checks first.
4560 (inline_small_functions): Do not estimate growth when dumping;
4562 * ipa-inline.h (inline_summary): Add min_size.
4563 (growth_likely_positive): New function.
4564 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
4565 (set_cond_stmt_execution_predicate): Cleanup.
4566 (estimate_edge_size_and_time): Compute min_size.
4567 (estimate_calls_size_and_time): Likewise.
4568 (estimate_node_size_and_time): Likewise.
4569 (inline_update_overall_summary): Update min_size.
4570 (do_estimate_edge_time): Likewise.
4571 (do_estimate_edge_size): Update.
4572 (do_estimate_edge_hints): Update.
4573 (growth_likely_positive): New function.
4575 2014-03-28 Jakub Jelinek <jakub@redhat.com>
4578 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
4579 also if addr has VOIDmode.
4581 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4583 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
4584 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
4586 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
4587 instructions as well as AdvancedSIMD loads.
4589 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4591 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
4592 Use crypto_aese type.
4593 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
4594 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
4595 crypto_aese, crypto_aesmc. Move to types.md.
4596 * config/arm/types.md (crypto_aes): Split into crypto_aese,
4598 * config/arm/iterators.md (crypto_type): Likewise.
4600 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4602 * cgraph.c: Include expr.h and tree-dfa.h.
4603 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
4606 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4609 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
4610 regs from checking multi-reg pseudos.
4612 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4614 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
4616 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4618 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
4619 if it would clobber the stack pointer, even temporarily.
4621 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
4623 * mode-switching.c: Make small adjustments to the top comment.
4625 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4627 * config/rs6000/constraints.md (wD constraint): New constraint to
4628 match the constant integer to get the top DImode/DFmode out of a
4629 vector in a VSX register.
4631 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
4632 match the constant integer to get the top DImode/DFmode out of a
4633 vector in a VSX register.
4635 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
4638 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
4641 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
4642 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
4644 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
4645 Optimize vec_extract of 64-bit values, where the value being
4646 extracted is in the top word, where we can use scalar
4647 instructions. Add direct move and store support. Combine the big
4648 endian/little endian vector select load support into a single insn.
4649 (vsx_extract_<mode>_internal1): Likewise.
4650 (vsx_extract_<mode>_internal2): Likewise.
4651 (vsx_extract_<mode>_load): Likewise.
4652 (vsx_extract_<mode>_store): Likewise.
4653 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
4654 combined into vsx_extract_<mode>_load.
4655 (vsx_extract_<mode>_one_le): Likewise.
4657 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
4658 define the top 64-bit vector element.
4660 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
4663 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4664 Document vec_vbpermq builtin.
4667 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
4668 enable use of xxsldwi and xxpermdi builtin functions.
4669 (vec_xxpermdi): Likewise.
4671 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4672 Document use of vec_xxsldwi and vec_xxpermdi builtins.
4674 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
4676 PR rtl-optimization/60650
4677 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
4679 (find_spills_for): New.
4680 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
4681 Spill all pseudos on the second iteration.
4683 2014-03-27 Marek Polacek <polacek@redhat.com>
4686 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
4689 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4691 * config/s390/s390.c (s390_can_use_return_insn): Check for
4692 call-saved FPRs on 31 bit.
4694 2014-03-27 Jakub Jelinek <jakub@redhat.com>
4697 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
4698 if they need regimplification, just drop them instead of
4699 calling gimple_regimplify_operands on them.
4701 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
4704 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
4705 (aarch64_frame_pointer_required): Adjust logic.
4706 (aarch64_can_eliminate): Adjust logic.
4707 (aarch64_override_options_after_change): Adjust logic.
4709 2014-03-27 Dehao Chen <dehao@google.com>
4711 * ipa-inline.c (early_inliner): Update node's inline info.
4713 2014-03-26 Dehao Chen <dehao@google.com>
4715 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
4716 compiler inserted conditional jumps for NAN float check.
4718 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4720 * ubsan.h (ubsan_create_data): Change second argument's type
4721 to const location_t *.
4722 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
4724 (ubsan_create_data): Change second argument to const location_t *PLOC.
4725 Create Loc field whenever PLOC is non-NULL.
4726 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
4727 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
4731 * real.c (real_to_integer2): Change type of low to UHWI.
4733 2014-03-26 Tobias Burnus <burnus@net-b.de>
4735 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
4736 (CILK_SELF_SPECS): New define.
4737 (driver_self_specs): Use it.
4739 2014-03-26 Richard Biener <rguenther@suse.de>
4741 * tree-pretty-print.c (percent_K_format): Implement special
4742 case for LTO and its stripped down BLOCK tree.
4744 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4747 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
4749 * tree-vrp.c (simplify_internal_call_using_ranges): If only
4750 one range is range_int_cst_p, but not both, at least optimize
4751 addition/subtraction of 0 and multiplication by 0 or 1.
4752 * gimple-fold.c (gimple_fold_call): Fold
4753 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
4754 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
4755 INTEGER_CSTs, try to fold at least x * 0 and y - y.
4757 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
4759 PR rtl-optimization/60452
4760 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
4761 <case REG>: Return 1 for invalid offsets from the frame pointer.
4763 2014-03-26 Marek Polacek <polacek@redhat.com>
4766 * doc/extend.texi (C Extensions): Mention variable-length arrays in
4769 2014-03-26 Marek Polacek <polacek@redhat.com>
4772 * doc/extend.texi (Designated Inits): Describe what happens to omitted
4775 2014-03-26 Marek Polacek <polacek@redhat.com>
4778 * ira-color.c (update_conflict_hard_regno_costs): Perform the
4779 multiplication in unsigned type.
4781 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4783 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
4785 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4787 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
4789 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4792 * cif-code.def (UNREACHABLE) New code.
4793 * ipa-inline.c (inline_small_functions): Skip edges to
4794 __builtlin_unreachable.
4795 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
4796 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
4797 predicate to __bulitin_unreachable.
4798 (set_cond_stmt_execution_predicate): Fix issue when
4799 invert_tree_comparison returns ERROR_MARK.
4800 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
4801 propagate to inline clones.
4802 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
4804 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
4805 * cgraphclones.c (cgraph_clone_node): If call destination is already
4806 ureachable, do not redirect it back.
4807 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
4810 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4812 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
4813 Do not modify inline clones.
4815 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4817 * config/i386/i386.md (general_sext_operand): New mode attr.
4818 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
4819 don't generate (sign_extend (const_int)).
4820 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
4821 operands[2]. Use We constraint instead of <i> and
4822 <general_sext_operand> predicate instead of <general_operand>.
4823 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
4824 * config/i386/constraints.md (We): New constraint.
4825 * config/i386/predicates.md (x86_64_sext_operand,
4826 sext_operand): New predicates.
4828 2014-03-25 Martin Jambor <mjambor@suse.cz>
4831 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
4832 inconsistent devirtualizations to __builtin_unreachable.
4834 2014-03-25 Marek Polacek <polacek@redhat.com>
4837 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
4839 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
4841 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
4842 order of elements for big-endian.
4844 2014-03-25 Richard Biener <rguenther@suse.de>
4847 * gimplify-me.c (gimple_regimplify_operands): Update the
4850 2014-03-25 Martin Jambor <mjambor@suse.cz>
4853 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
4854 (lto_output_varpool_node): Likewise.
4855 (input_overwrite_node): Likewise.
4856 (input_varpool_node): Likewise.
4858 2014-03-25 Richard Biener <rguenther@suse.de>
4860 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
4861 (run_gcc): Likewise.
4863 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4865 * combine.c (simplify_compare_const): Add MODE argument.
4866 Handle mode_width 0 as very large mode_width.
4867 (try_combine, simplify_comparison): Adjust callers.
4869 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
4870 type to avoid signed integer overflow.
4871 * explow.c (plus_constant): Likewise.
4873 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
4875 * doc/generic.texi: Correct typos.
4877 2014-03-24 Tobias Burnus <burnus@net-b.de>
4879 * doc/invoke.texi (-flto): Expand section about
4880 using static libraries with LTO.
4882 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4884 PR rtl-optimization/60501
4885 * optabs.def (addptr3_optab): New optab.
4886 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
4887 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
4888 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
4890 * lra.c (emit_add3_insn): Use the addptr pattern if available.
4892 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
4894 2014-03-24 Ulrich Drepper <drepper@gmail.com>
4896 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
4899 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
4900 (_mm256_undefined_ps): Define.
4901 (_mm256_undefined_pd): Define.
4902 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
4903 (_mm_undefined_pd): Define.
4904 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
4905 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
4906 (_mm512_undefined_ps): Define.
4907 (_mm512_undefined_pd): Define.
4908 Use _mm*_undefined_*.
4909 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
4911 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
4913 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
4914 (lshr_simd): DI mode added.
4915 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
4916 (aarch64_ushr_simddi): Likewise.
4917 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
4918 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
4919 (vshrd_n_u64): Likewise.
4921 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4923 * Makefile.in (s-macro_list): Depend on cc1.
4925 2014-03-23 Teresa Johnson <tejohnson@google.com>
4927 * ipa-utils.c (ipa_print_order): Use specified dump file.
4929 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
4931 PR rtl-optimization/60601
4932 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
4934 * gcc.c (eval_spec_function): Initialize save_growing_value.
4936 2014-03-22 Jakub Jelinek <jakub@redhat.com>
4939 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
4940 code == MINUS_EXPR, never swap op0 with op1.
4942 * toplev.c (init_local_tick): Avoid signed integer multiplication
4944 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
4945 shift by first operand's bitsize.
4947 2014-03-21 Jakub Jelinek <jakub@redhat.com>
4950 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
4952 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
4953 TARGET_ISA_64BIT_P(x).
4955 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4957 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
4958 pattern for vector nor instead of subtract from splat(-1).
4959 (altivec_expand_vec_perm_const_le): Likewise.
4961 2014-03-21 Richard Henderson <rth@twiddle.net>
4964 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
4965 related insns after epilogue_completed.
4967 2014-03-21 Martin Jambor <mjambor@suse.cz>
4970 * cgraph.h (symtab_node): New flag body_removed.
4971 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
4972 when removing bodies.
4973 * symtab.c (dump_symtab_base): Dump body_removed flag.
4974 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
4975 had their bodies removed.
4977 2014-03-21 Martin Jambor <mjambor@suse.cz>
4980 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
4983 2014-03-21 Richard Biener <rguenther@suse.de>
4985 PR tree-optimization/60577
4986 * tree-core.h (struct tree_base): Document nothrow_flag use
4987 in VAR_DECL_NONALIASED.
4988 * tree.h (VAR_DECL_NONALIASED): New.
4989 (may_be_aliased): Adjust.
4990 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
4992 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
4994 * expr.c (expand_expr_real_1): Remove outdated comment.
4996 2014-03-20 Jakub Jelinek <jakub@redhat.com>
4999 * ira.c (adjust_cleared_regs): Call copy_rtx on
5000 *reg_equiv[REGNO (loc)].src_p before passing it to
5001 simplify_replace_fn_rtx.
5004 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
5005 into CONST, put pic register as first operand of PLUS. Use
5006 gen_const_mem for both 32-bit and 64-bit PIC got loads.
5008 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5010 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
5012 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5014 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
5015 around for store forwarding issue in the FPU on the UT699.
5016 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
5017 loads and operations if -mfix-ut699 is specified.
5018 (divtf3_hq): Tweak attribute.
5019 (sqrttf2_hq): Likewise.
5021 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5023 * calls.c (store_one_arg): Remove incorrect const qualification on the
5024 type of the temporary.
5025 * cfgexpand.c (expand_return): Likewise.
5026 * expr.c (expand_constructor): Likewise.
5027 (expand_expr_real_1): Likewise.
5029 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5031 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
5034 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
5037 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
5039 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
5041 * config/arm/aarch-common-protos.h
5042 (alu_cost_table): Fix spelling of "extend".
5043 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
5045 2014-03-19 Richard Biener <rguenther@suse.de>
5048 * tree-core.h (tree_type_common): Re-order pointer members
5049 to reduce recursion depth during GC walks.
5051 2014-03-19 Marek Polacek <polacek@redhat.com>
5054 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
5055 before accessing it.
5057 2014-03-19 Richard Biener <rguenther@suse.de>
5060 * lto-streamer-in.c (input_function): In WPA stage do not drop
5063 2014-03-19 Jakub Jelinek <jakub@redhat.com>
5065 PR tree-optimization/60559
5066 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
5067 with build_zero_cst assignment.
5069 2014-03-18 Kai Tietz <ktietz@redhat.com>
5071 PR rtl-optimization/56356
5072 * sdbout.c (sdbout_parms): Verify that parms'
5073 incoming argument is valid.
5074 (sdbout_reg_parms): Likewise.
5076 2014-03-18 Richard Henderson <rth@redhat.com>
5079 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
5080 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
5081 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
5083 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
5085 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
5086 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
5087 Italicize plugin event names in description. Explain that
5088 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
5089 Remind that no GCC functions should be called after PLUGIN_FINISH.
5090 Explain what pragmas with expansion are.
5092 2014-03-18 Martin Liska <mliska@suse.cz>
5094 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
5095 gimple call statement is update.
5096 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
5097 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
5099 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5102 * ubsan.c (ubsan_instrument_unreachable): Call
5103 initialize_sanitizer_builtins.
5104 (ubsan_pass): Likewise.
5107 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
5108 varpool_finalize_decl instead of rest_of_decl_compilation.
5110 2014-03-18 Richard Biener <rguenther@suse.de>
5112 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
5113 by using bitmap_and_compl instead of bitmap_and_compl_into.
5114 (df_rd_transfer_function): Likewise.
5116 2014-03-18 Richard Biener <rguenther@suse.de>
5118 * doc/lto.texi (fresolution): Fix typo.
5120 2014-03-18 Richard Biener <rguenther@suse.de>
5122 * doc/invoke.texi (flto): Update for changes in 4.9.
5124 2014-03-18 Richard Biener <rguenther@suse.de>
5126 * doc/loop.texi: Remove section on the removed lambda framework.
5127 Update loop docs with recent changes in preserving loop structure.
5129 2014-03-18 Richard Biener <rguenther@suse.de>
5131 * doc/lto.texi (-fresolution): Document.
5133 2014-03-18 Richard Biener <rguenther@suse.de>
5135 * doc/contrib.texi: Adjust my name.
5137 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5140 * internal-fn.c: Include diagnostic-core.h.
5141 (expand_BUILTIN_EXPECT): New function.
5142 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
5143 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
5144 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
5145 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
5147 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
5148 Revert 3 argument __builtin_expect code.
5149 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
5150 * gimple-fold.c (gimple_fold_call): Likewise.
5151 * tree.h (fold_builtin_expect): New prototype.
5152 * builtins.c (build_builtin_expect_predicate): Add predictor
5153 argument, if non-NULL, create 3 argument __builtin_expect.
5154 (fold_builtin_expect): No longer static. Add ARG2 argument,
5155 pass it through to build_builtin_expect_predicate.
5156 (fold_builtin_2): Adjust caller.
5157 (fold_builtin_3): Handle BUILT_IN_EXPECT.
5158 * internal-fn.def (BUILTIN_EXPECT): New.
5160 2014-03-18 Tobias Burnus <burnus@net-b.de>
5163 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
5164 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
5165 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
5167 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
5170 * predict.c (combine_predictions_for_bb): Fix up formatting.
5171 (expr_expected_value_1, expr_expected_value): Add predictor argument,
5172 fill what it points to if non-NULL.
5173 (tree_predict_by_opcode): Adjust caller, use the predictor.
5174 * predict.def (PRED_COMPARE_AND_SWAP): Add.
5176 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
5178 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
5179 proper constant for the store mode.
5181 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
5183 * symtab.c (change_decl_assembler_name): Fix transparent alias
5186 2014-03-16 Renlin Li <Renlin.Li@arm.com>
5188 * config/aarch64/aarch64.c: Correct the comments about the
5189 aarch64 stack layout.
5191 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
5193 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
5194 check for GF_OMP_FOR_KIND_FOR.
5196 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
5198 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
5199 ymm and zmm register names.
5201 2014-03-17 Jakub Jelinek <jakub@redhat.com>
5204 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
5205 note creation for the 2010-08-31 changes.
5207 2014-03-17 Marek Polacek <polacek@redhat.com>
5210 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
5211 as -fno-tree-loop-vectorize.
5212 (expand_omp_simd): Likewise.
5214 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
5216 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
5217 (eligible_for_call_delay): New prototype.
5218 * config/sparc/sparc.c (tls_call_delay): Rename into...
5219 (eligible_for_call_delay): ...this. Return false if the instruction
5220 cannot be put in the delay slot of a branch.
5221 (eligible_for_restore_insn): Simplify.
5222 (eligible_for_return_delay): Return false if the instruction cannot be
5223 put in the delay slot of a branch and simplify.
5224 (eligible_for_sibcall_delay): Return false if the instruction cannot be
5225 put in the delay slot of a branch.
5226 * config/sparc/sparc.md (fix_ut699): New attribute.
5227 (tls_call_delay): Delete.
5228 (in_call_delay): Reimplement.
5229 (eligible_for_sibcall_delay): Rename into...
5230 (in_sibcall_delay): ...this.
5231 (eligible_for_return_delay): Rename into...
5232 (in_return_delay): ...this.
5233 (in_branch_delay): Reimplement.
5234 (in_uncond_branch_delay): Delete.
5235 (in_annul_branch_delay): Delete.
5237 2014-03-14 Richard Henderson <rth@redhat.com>
5240 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
5241 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
5242 (*floathi<X87MODEF>2_i387_with_temp): Remove.
5243 (floathi splitters): Remove.
5244 (float<SWI48x>xf2): New pattern.
5245 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
5246 code that tried to handle DImode for 32-bit, but which was excluded
5247 by the pattern's condition. Drop allocation of stack temporary.
5248 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
5249 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
5250 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
5251 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
5252 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
5253 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
5254 (*float<SWI48><MODEF>2_sse_interunit): Remove.
5255 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
5256 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
5257 (*float<SWI48x><X87MODEF>2_i387): Remove.
5258 (all float _with_temp splitters): Remove.
5259 (*float<SWI48x><MODEF>2_i387): New pattern.
5260 (*float<SWI48><MODEF>2_sse): New pattern.
5261 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
5262 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
5264 2014-03-14 Jakub Jelinek <jakub@redhat.com>
5265 Marek Polacek <polacek@redhat.com>
5268 * common.opt (dump_base_name_prefixed): New Variable.
5269 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
5270 if x_dump_base_name_prefixed is already set, set it at the end.
5272 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
5274 PR rtl-optimization/60508
5275 * lra-constraints.c (get_reload_reg): Add new parameter
5277 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
5278 Pass the new parameter values.
5280 2014-03-14 Richard Biener <rguenther@suse.de>
5282 * common.opt: Revert unintented changes from r205065.
5285 2014-03-14 Richard Biener <rguenther@suse.de>
5288 * cfghooks.c (split_block): Properly adjust all loops the
5289 block was a latch of.
5291 2014-03-14 Martin Jambor <mjambor@suse.cz>
5294 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
5297 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
5300 * config/avr/avr.c (avr_set_current_function): Pass function name
5301 through default_strip_name_encoding before sanity checking instead
5302 of skipping the first char of the assembler name.
5304 2014-03-13 Richard Henderson <rth@redhat.com>
5307 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
5308 (ix86_force_to_memory, ix86_free_from_memory): Remove.
5309 * config/i386/i386-protos.h: Likewise.
5310 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
5311 in the expander instead of a splitter.
5312 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
5313 any possibility of requiring a memory.
5314 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
5315 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
5316 (fp branch splitters): Update for ix86_split_fp_branch.
5317 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
5318 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
5319 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
5320 (*fop_<MODEF>_2_i387): Remove f/r alternative.
5321 (*fop_<MODEF>_3_i387): Likewise.
5322 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
5323 (splitters for the fop_* register patterns): Remove.
5324 (fscalexf4_i387): Rename from *fscalexf4_i387.
5325 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
5327 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5329 PR tree-optimization/59779
5330 * tree-dfa.c (get_ref_base_and_extent): Use double_int
5331 type for bitsize and maxsize instead of HOST_WIDE_INT.
5333 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
5335 PR rtl-optimization/57320
5336 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
5337 the CFG after thread_prologue_and_epilogue_insns.
5339 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
5341 PR rtl-optimization/57189
5342 * lra-constraints.c (process_alt_operands): Disfavor spilling
5345 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
5347 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
5349 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5351 PR tree-optimization/59025
5353 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
5354 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
5356 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
5359 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
5360 calls of avr_out_plus_1.
5362 2014-03-13 Bin Cheng <bin.cheng@arm.com>
5364 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
5365 BB's single pred and update the father loop's latch info later.
5367 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5369 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
5373 (VEC_base): Likewise.
5374 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
5375 registers, we need to swap double words in little endian mode.
5377 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
5378 to be a container mode for 128-bit integer operations added in ISA
5379 2.07. Unlike TImode and PTImode, the preferred register set is
5380 the Altivec/VMX registers for the 128-bit operations.
5382 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
5384 (rs6000_split_128bit_ok_p): Likewise.
5386 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
5387 macros for creating ISA 2.07 normal and overloaded builtin
5388 functions with 3 arguments.
5389 (BU_P8V_OVERLOAD_3): Likewise.
5390 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
5391 for use as overloaded functions.
5392 (VPERM_1TI_UNS): Likewise.
5393 (VSEL_1TI): Likewise.
5394 (VSEL_1TI_UNS): Likewise.
5395 (ST_INTERNAL_1ti): Likewise.
5396 (LD_INTERNAL_1ti): Likewise.
5397 (XXSEL_1TI): Likewise.
5398 (XXSEL_1TI_UNS): Likewise.
5399 (VPERM_1TI): Likewise.
5400 (VPERM_1TI_UNS): Likewise.
5401 (XXPERMDI_1TI): Likewise.
5402 (SET_1TI): Likewise.
5403 (LXVD2X_V1TI): Likewise.
5404 (STXVD2X_V1TI): Likewise.
5405 (VEC_INIT_V1TI): Likewise.
5406 (VEC_SET_V1TI): Likewise.
5407 (VEC_EXT_V1TI): Likewise.
5408 (EQV_V1TI): Likewise.
5409 (NAND_V1TI): Likewise.
5410 (ORC_V1TI): Likewise.
5411 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
5412 added in ISA 2.07. Add both normal 'altivec' builtins, and the
5414 (VADDUQM): Likewise.
5415 (VSUBCUQ): Likewise.
5416 (VADDEUQM): Likewise.
5417 (VADDECUQ): Likewise.
5418 (VSUBEUQM): Likewise.
5419 (VSUBECUQ): Likewise.
5421 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5422 __int128_t and __uint128_t types.
5423 (__uint128_type): Likewise.
5424 (altivec_categorize_keyword): Add support for vector __int128_t,
5425 vector __uint128_t, vector __int128, and vector unsigned __int128
5426 as a container type for TImode operations that need to be done in
5427 VSX/Altivec registers.
5428 (rs6000_macro_to_expand): Likewise.
5429 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5430 to support 128-bit integer instructions vaddcuq, vadduqm,
5431 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5432 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5434 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5435 for V1TImode, and set up preferences to use VSX/Altivec registers.
5436 Setup VSX reload handlers.
5437 (rs6000_debug_reg_global): Likewise.
5438 (rs6000_init_hard_regno_mode_ok): Likewise.
5439 (rs6000_preferred_simd_mode): Likewise.
5440 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5441 (easy_altivec_constant): Likewise.
5442 (output_vec_const_move): Likewise.
5443 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5445 (rs6000_expand_vector_extract): Likewise.
5446 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
5448 (rs6000_const_vec): Add support for V1TImode.
5449 (rs6000_emit_le_vsx_load): Swap double words when loading or
5450 storing TImode/V1TImode.
5451 (rs6000_emit_le_vsx_store): Likewise.
5452 (rs6000_emit_le_vsx_move): Likewise.
5453 (rs6000_emit_move): Add support for V1TImode.
5454 (altivec_expand_ld_builtin): Likewise.
5455 (altivec_expand_st_builtin): Likewise.
5456 (altivec_expand_vec_init_builtin): Likewise.
5457 (altivec_expand_builtin): Likewise.
5458 (rs6000_init_builtins): Add support for V1TImode type. Add
5459 support for ISA 2.07 128-bit integer builtins. Define type names
5460 for the VSX/Altivec vector types.
5461 (altivec_init_builtins): Add support for overloaded vector
5462 functions with V1TImode type.
5463 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
5464 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
5466 (rs6000_split_128bit_ok_p): Likewise.
5467 (rs6000_handle_altivec_attribute): Create V1TImode from vector
5468 __int128_t and vector __uint128_t.
5470 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
5471 and mode attributes.
5478 (VS_scalar): Likewise.
5479 (VS_double): Likewise.
5480 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
5482 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
5483 we support the ISA 2.07 128-bit integer arithmetic instructions.
5484 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
5485 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
5486 and TImode types for use with the builtin functions.
5487 (V1TI_type_node): Likewise.
5488 (unsigned_V1TI_type_node): Likewise.
5489 (intTI_type_internal_node): Likewise.
5490 (uintTI_type_internal_node): Likewise.
5492 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
5493 128-bit builtin functions.
5494 (UNSPEC_VADDEUQM): Likewise.
5495 (UNSPEC_VADDECUQ): Likewise.
5496 (UNSPEC_VSUBCUQ): Likewise.
5497 (UNSPEC_VSUBEUQM): Likewise.
5498 (UNSPEC_VSUBECUQ): Likewise.
5499 (VM): Add V1TImode to vector mode iterators.
5501 (VI_unit): Likewise.
5502 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
5503 (altivec_vaddcuq): Likewise.
5504 (altivec_vsubuqm): Likewise.
5505 (altivec_vsubcuq): Likewise.
5506 (altivec_vaddeuqm): Likewise.
5507 (altivec_vaddecuq): Likewise.
5508 (altivec_vsubeuqm): Likewise.
5509 (altivec_vsubecuq): Likewise.
5511 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
5513 (BOOL_128): Likewise.
5514 (BOOL_REGS_OUTPUT): Likewise.
5515 (BOOL_REGS_OP1): Likewise.
5516 (BOOL_REGS_OP2): Likewise.
5517 (BOOL_REGS_UNARY): Likewise.
5518 (BOOL_REGS_AND_CR0): Likewise.
5520 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
5521 128-bit integer builtin support.
5522 (vec_vadduqm): Likewise.
5523 (vec_vaddecuq): Likewise.
5524 (vec_vaddeuqm): Likewise.
5525 (vec_vsubecuq): Likewise.
5526 (vec_vsubeuqm): Likewise.
5527 (vec_vsubcuq): Likewise.
5528 (vec_vsubuqm): Likewise.
5530 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5531 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
5532 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
5533 128-bit integer add/subtract to ISA 2.07.
5535 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
5537 * config/arc/arc.c (arc_predicate_delay_insns):
5538 Fix third argument passed to conditionalize_nonjump.
5540 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
5542 * config/aarch64/aarch64-builtins.c
5543 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
5544 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
5545 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
5546 instead of __builtin_lfloor.
5547 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
5549 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5551 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
5552 (tree_ssa_ifcombine_bb_1): New function.
5553 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
5554 is an empty forwarder block to then_bb or vice versa and then_bb
5555 and else_bb are effectively swapped.
5557 2014-03-12 Christian Bruel <christian.bruel@st.com>
5560 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
5561 REG_CFA_DEF_CFA note.
5562 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
5563 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
5565 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
5567 PR tree-optimization/60454
5568 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
5570 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5572 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
5573 Do not define target_cpu_default2 to generic.
5574 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
5575 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
5576 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
5578 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5579 Marc Glisse <marc.glisse@inria.fr>
5581 PR tree-optimization/60502
5582 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
5583 instead of build_low_bits_mask.
5585 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5588 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
5589 if there are multiple uses, but op doesn't live on E edge.
5590 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
5591 clobber stmts before __builtin_unreachable.
5593 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
5595 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
5596 hard_frame_pointer_rtx.
5597 * cse.c (cse_insn): Remove volatile check.
5598 * cselib.c (cselib_process_insn): Likewise.
5599 * dse.c (scan_insn): Likewise.
5601 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5603 * config/arc/arc.c (conditionalize_nonjump): New function,
5605 (arc_ifcvt): ... this.
5606 (arc_predicate_delay_insns): Use it.
5608 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5610 * config/arc/predicates.md (extend_operand): During/after reload,
5611 allow const_int_operand.
5612 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
5613 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
5614 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
5616 (umulsi3_highpart_i): Likewise.
5618 2014-03-11 Richard Biener <rguenther@suse.de>
5620 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
5621 Add asserts to guard possible wrong-code bugs.
5623 2014-03-11 Richard Biener <rguenther@suse.de>
5625 PR tree-optimization/60429
5626 PR tree-optimization/60485
5627 * tree-ssa-structalias.c (set_union_with_increment): Properly
5628 take into account all fields that overlap the shifted vars.
5629 (do_sd_constraint): Likewise.
5630 (do_ds_constraint): Likewise.
5631 (get_constraint_for_ptr_offset): Likewise.
5633 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
5635 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
5636 (nios2_compute_frame_layout):
5637 Add calculation of cfun->machine->fp_save_offset.
5638 (nios2_expand_prologue): Correct setting of frame pointer register
5640 (nios2_expand_epilogue): Update recovery of stack pointer from
5641 frame pointer accordingly.
5642 (nios2_initial_elimination_offset): Update calculation of offset
5643 for eliminating to HARD_FRAME_POINTER_REGNUM.
5645 2014-03-10 Jakub Jelinek <jakub@redhat.com>
5648 * ipa.c (symtab_remove_unreachable_nodes): Don't call
5649 cgraph_get_create_node on VAR_DECLs.
5651 2014-03-10 Richard Biener <rguenther@suse.de>
5654 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
5656 2014-03-08 Douglas B Rupp <rupp@gnat.com>
5658 * config/vms/vms.opt (vms_float_format): New variable.
5660 2014-03-08 Tobias Burnus <burnus@net-b.de>
5662 * doc/invoke.texi (-fcilkplus): Update implementation status.
5664 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
5665 Richard Biener <rguenther@suse.de>
5667 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
5668 consistently accross all TUs.
5669 (run_gcc): Enable -fshort-double automatically at link at link-time
5670 and disallow override.
5672 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
5675 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
5676 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
5677 if they can't be used.
5679 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5681 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
5682 for Solaris 11/x86 ld.
5683 * configure: Regenerate.
5685 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5687 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
5688 (LIB_TLS_SPEC): Save as ld_tls_libs.
5689 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
5690 (HAVE_AS_IX86_TLSLDM): New test.
5691 * configure, config.in: Regenerate.
5692 * config/i386/i386.c (legitimize_tls_address): Fall back to
5693 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
5694 cannot support TLS_MODEL_LOCAL_DYNAMIC.
5695 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
5696 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
5698 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
5700 * common.opt (fira-loop-pressure): Mark as optimization.
5702 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
5704 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
5705 an OpenMP mappable type.
5707 2014-03-06 Matthias Klose <doko@ubuntu.com>
5709 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
5710 MULTILIB_OSDIRNAMES is not defined.
5712 2014-03-06 Jakub Jelinek <jakub@redhat.com>
5713 Meador Inge <meadori@codesourcery.com>
5716 * config/arm/arm.c (arm_tls_symbol_p): Remove.
5717 (arm_legitimize_address): Call legitimize_tls_address for any
5718 arm_tls_referenced_p expression, handle constant addend. Call it
5719 before testing for !TARGET_ARM.
5720 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
5722 2014-03-06 Richard Biener <rguenther@suse.de>
5728 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5730 * tree-streamer.c (record_common_node): Assert we don't record
5731 nodes with type double.
5732 (preload_common_node): Skip type double, complex double and double
5733 pointer since it is now frontend dependent due to fshort-double option.
5735 2014-03-06 Richard Biener <rguenther@suse.de>
5737 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
5738 or -fno-lto is specified and the linker has full plugin support.
5739 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
5740 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
5741 * lto-wrapper.c (merge_and_complain): Merge compile-time
5742 optimization levels.
5743 (run_gcc): And pass it through to the link options.
5745 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
5749 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5751 * cselib.c (remove_useless_values): Skip to avoid quadratic
5752 behavior if the condition moved from...
5753 (cselib_process_insn): ... here holds.
5755 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5758 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
5759 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
5762 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
5763 (TM_H): Add x86-tune.def.
5765 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5767 * config/aarch64/aarch64.c (generic_tunings):
5768 Use cortexa57_extra_costs.
5770 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5773 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
5774 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
5775 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
5778 2014-03-04 Heiher <r@hev.cc>
5780 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
5781 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
5783 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
5785 * config/i386/predicates.md (const2356_operand): Change to ...
5786 (const2367_operand): ... this.
5787 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
5789 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5790 (*avx512pf_scatterpf<mode>sf): Ditto.
5791 (avx512pf_scatterpf<mode>df): Ditto.
5792 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5793 (*avx512pf_scatterpf<mode>df): Ditto.
5794 * config/i386/i386.c (ix86_expand_builtin): Update
5795 incorrect hint operand error message.
5797 2014-03-04 Richard Biener <rguenther@suse.de>
5799 * lto-section-in.c (lto_get_section_data): Fix const cast.
5801 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5803 * tree-streamer.c (record_common_node): Assert we don't record
5804 nodes with type double.
5805 (preload_common_node): Skip type double, complex double and double
5806 pointer since it is now frontend dependent due to fshort-double option.
5808 2014-03-04 Richard Biener <rguenther@suse.de>
5811 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
5812 (lto_input_toplevel_asms): Likewise.
5813 * lto-section-in.c (lto_get_section_data): Instead do it here
5816 2014-03-04 Richard Biener <rguenther@suse.de>
5818 PR tree-optimization/60382
5819 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
5820 dead PHIs a reduction.
5822 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
5824 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
5826 (_mm_prefetch): Move out of GCC target("sse") pragma.
5827 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
5828 GCC target("prfchw") pragma.
5829 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
5831 * config/i386/i386.c (ix86_option_override_internal): Enable
5832 -mprfchw with -mprefetchwt1.
5834 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5836 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
5839 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5841 * opts.h (CL_PCH_IGNORE): Define.
5842 * targhooks.c (option_affects_pch_p):
5843 Return false for options that have CL_PCH_IGNORE set.
5844 * opt-functions.awk: Process PchIgnore.
5845 * doc/options.texi: Document PchIgnore.
5847 * config/arc/arc.opt (misize): Add PchIgnore property.
5849 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5851 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
5852 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
5853 constraint on constants to permit them being loaded into
5854 GENERAL_REGS or BASE_REGS.
5856 2014-03-03 Nick Clifton <nickc@redhat.com>
5858 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
5859 anti-cacnonical alternatives.
5860 (negandhi3_real): New pattern.
5861 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
5863 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5865 * config/avr/avr-mcus.def: Remove atxmega16x1.
5866 * config/avr/avr-tables.opt: Regenerate.
5867 * config/avr/t-multilib: Regenerate.
5868 * doc/avr-mmcu.texi: Regenerate.
5870 2014-03-03 Tobias Grosser <tobias@grosser.es>
5871 Mircea Namolaru <mircea.namolaru@inria.fr>
5873 PR tree-optimization/58028
5874 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
5877 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5879 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
5880 not handled by recognizers.
5882 2014-03-03 Jakub Jelinek <jakub@redhat.com>
5885 * function.c (expand_function_end): Don't emit
5886 clobber_return_register sequence if clobber_after is a BARRIER.
5887 * cfgexpand.c (construct_exit_block): Append instructions before
5888 return_label to prev_bb.
5890 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5892 * config/rs6000/constraints.md: Document reserved use of "wc".
5894 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5897 * ipa.c (function_and_variable_visibility): When dissolving comdat
5898 group, also set all symbols to local.
5900 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5905 2013-12-14 Jan Hubicka <jh@suse.cz>
5907 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
5909 2014-03-02 Jon Beniston <jon@beniston.com>
5915 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
5916 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
5917 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
5918 (simple_return, *simple_return): New patterns
5919 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
5920 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
5922 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
5924 * dwarf2out.c (gen_subprogram_die): Tidy.
5926 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
5929 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
5930 (*mov_t_msb_neg_negc): ... this new insn.
5932 2014-02-28 Jason Merrill <jason@redhat.com>
5935 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
5938 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
5941 * dwarf2out.c (decltype_auto_die): New static.
5942 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
5943 (gen_type_die_with_usage): Handle 'decltype(auto)'.
5944 (is_cxx_auto): Likewise.
5946 2014-02-28 Ian Bolton <ian.bolton@arm.com>
5948 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
5949 we are not using general regs only.
5951 2014-02-28 Richard Biener <rguenther@suse.de>
5954 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
5955 previous fix and only allow to remove trivial pre-headers
5956 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
5957 (remove_forwarder_block): Properly update the latch of a loop.
5959 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5962 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
5963 (cselib_preserved_hash_table): New.
5964 (preserve_constants_and_equivs): Move preserved vals to it.
5965 (cselib_find_slot): Look it up first.
5966 (cselib_init): Initialize it.
5967 (cselib_finish): Release it.
5968 (dump_cselib_table): Dump it.
5970 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5973 * cselib.c (remove_useless_values): Skip to avoid quadratic
5974 behavior if the condition moved from...
5975 (cselib_process_insn): ... here holds.
5977 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5980 * var-tracking.c (vt_initialize): Apply the same condition to
5981 preserve the CFA base value.
5983 2014-02-28 Joey Ye <joey.ye@arm.com>
5986 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
5987 if reload in progress or completed.
5989 2014-02-28 Tobias Burnus <burnus@net-b.de>
5992 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
5995 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
5997 * doc/tm.texi.in (Condition Code Status): Update documention for
5998 relative locations of cc0-setter and cc0-user.
6000 2014-02-27 Jeff Law <law@redhat.com>
6002 PR rtl-optimization/52714
6003 * combine.c (try_combine): When splitting an unrecognized PARALLEL
6004 into two independent simple sets, if I3 is a jump, ensure the
6005 pattern we place into I3 is a (set (pc) ...).
6007 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
6008 Jeff Law <law@redhat.com>
6010 PR rtl-optimization/49847
6011 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
6012 are in different blocks.
6013 * doc/tm.texi (Condition Code Status): Update documention for
6014 relative locations of cc0-setter and cc0-user.
6016 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
6019 * lra.c (lra_emit_add): Check SUBREG too.
6021 2014-02-27 Andreas Schwab <schwab@suse.de>
6023 * config/m68k/m68k.c (m68k_option_override): Disable
6024 -flive-range-shrinkage for classic m68k.
6025 (m68k_override_options_after_change): Likewise.
6027 2014-02-27 Marek Polacek <polacek@redhat.com>
6030 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
6031 -Wmaybe-uninitialized.
6033 2014-02-27 Alan Modra <amodra@gmail.com>
6036 * reload1.c (emit_input_reload_insns): When reload_override_in,
6037 set old to rl->in_reg when rl->in_reg is a subreg.
6039 2014-02-26 Richard Biener <rguenther@suse.de>
6042 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
6044 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6046 * common/config/i386/predicates.md (const1256_operand): Remove.
6047 (const2356_operand): New.
6048 (const_1_to_2_operand): Remove.
6049 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6050 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6051 (*avx512pf_gatherpf<mode>sf): Ditto.
6052 (avx512pf_gatherpf<mode>df): Ditto.
6053 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6054 (*avx512pf_gatherpf<mode>df): Ditto.
6055 (avx512pf_scatterpf<mode>sf): Ditto.
6056 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6057 (*avx512pf_scatterpf<mode>sf): Ditto.
6058 (avx512pf_scatterpf<mode>df): Ditto.
6059 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6060 (*avx512pf_scatterpf<mode>df): Ditto.
6061 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
6063 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
6065 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
6066 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
6067 (_mm512_mask_testn_epi64_mask): Move to ...
6068 * config/i386/avx512cdintrin.h: Here.
6069 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
6070 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
6071 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
6072 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
6073 TARGET_AVX512F from TARGET_AVX512CD.
6075 2014-02-26 Richard Biener <rguenther@suse.de>
6078 * ipa.c (walk_polymorphic_call_targets): Properly guard
6079 call to inline_update_overall_summary.
6081 2014-02-26 Bin Cheng <bin.cheng@arm.com>
6084 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
6085 and latches only if requested. Fix latch if it is removed.
6086 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
6087 LOOPS_HAVE_PREHEADERS.
6089 2014-02-25 Andrew Pinski <apinski@cavium.com>
6091 * builtins.c (expand_builtin_thread_pointer): Create a new target
6092 when the target is NULL.
6094 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
6096 PR rtl-optimization/60317
6097 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6098 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6099 * lra-assigns.c: Include params.h.
6100 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
6101 other reload pseudos considerations.
6103 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6105 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
6106 to use canonical form for nor<mode>3.
6108 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6111 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
6114 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6116 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
6117 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
6118 (ix86_handle_option): Handle OPT_mprefetchwt1.
6119 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
6120 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
6122 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
6123 OPTION_MASK_ISA_PREFETCHWT1.
6124 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
6125 (PTA_PREFETCHWT1): New.
6126 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
6127 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
6128 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
6129 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
6130 (*prefetch_avx512pf_<mode>_: Change into ...
6131 (*prefetch_prefetchwt1_<mode>: This.
6132 * config/i386/i386.opt (mprefetchwt1): New.
6133 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
6134 (_mm_prefetch): Handle intent to write.
6135 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
6137 2014-02-25 Richard Biener <rguenther@suse.de>
6140 * emit-rtl.c (mem_attrs_htab): Remove.
6141 (mem_attrs_htab_hash): Likewise.
6142 (mem_attrs_htab_eq): Likewise.
6143 (set_mem_attrs): Always allocate new mem-attrs when something changed.
6144 (init_emit_once): Do not allocate mem_attrs_htab.
6146 2014-02-25 Richard Biener <rguenther@suse.de>
6149 * lto-opts.c (lto_write_options): Output non-explicit conservative
6150 -fwrapv, -fno-trapv and -fno-strict-overflow.
6151 * lto-wrapper.c (merge_and_complain): Handle merging those options.
6152 (run_gcc): And pass them through.
6154 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6156 * sel-sched.c (calculate_new_fences): New parameter ptime.
6157 Calculate it as a maximum over all fence cycles.
6158 (sel_sched_region_2): Adjust the call to calculate_new_fences.
6159 Print the final schedule timing when sched_verbose.
6161 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6163 PR rtl-optimization/60292
6164 * sel-sched.c (fill_vec_av_set): Do not reset target availability
6165 bit fot the fence instruction.
6167 2014-02-24 Alangi Derick <alangiderick@gmail.com>
6169 * calls.h: Fix typo in comment.
6171 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
6173 * config/pa/pa.c (pa_output_move_double): Don't valididate when
6174 adjusting offsetable addresses.
6176 2014-02-24 Guozhi Wei <carrot@google.com>
6178 * sparseset.h (sparseset_pop): Fix the wrong index.
6180 2014-02-24 Walter Lee <walt@tilera.com>
6182 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
6183 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
6185 * common/config/tilegx/tilegx-common.c
6186 (TARGET_DEFAULT_TARGET_FLAGS): Define.
6187 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
6189 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
6190 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
6191 (tilegx_gimplify_va_arg_expr): Handle big endian.
6192 (tilegx_expand_unaligned_load): Ditto.
6193 (tilegx_expand_unaligned_store): Ditto.
6194 (TARGET_RETURN_IN_MSB): New.
6195 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
6196 (TARGET_ENDIAN_DEFAULT): New.
6197 (TARGET_BIG_ENDIAN): Handle big endian.
6198 (BYTES_BIG_ENDIAN): Ditto.
6199 (WORDS_BIG_ENDIAN): Ditto.
6200 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
6203 * config/tilegx/tilegx.md (extv): Handle big endian.
6205 (insn_st<n>): Ditto.
6206 (insn_st<n>_add<bitsuffix>): Ditto.
6207 (insn_stnt<n>): Ditto.
6208 (insn_stnt<n>_add<bitsuffix>):Ditto.
6209 (vec_interleave_highv8qi): Handle big endian.
6210 (vec_interleave_highv8qi_be): New.
6211 (vec_interleave_highv8qi_le): New.
6212 (insn_v1int_h): Handle big endian.
6213 (vec_interleave_lowv8qi): Handle big endian.
6214 (vec_interleave_lowv8qi_be): New.
6215 (vec_interleave_lowv8qi_le): New.
6216 (insn_v1int_l): Handle big endian.
6217 (vec_interleave_highv4hi): Handle big endian.
6218 (vec_interleave_highv4hi_be): New.
6219 (vec_interleave_highv4hi_le): New.
6220 (insn_v2int_h): Handle big endian.
6221 (vec_interleave_lowv4hi): Handle big endian.
6222 (vec_interleave_lowv4hi_be): New.
6223 (vec_interleave_lowv4hi_le): New.
6224 (insn_v2int_l): Handle big endian.
6225 (vec_interleave_highv2si): Handle big endian.
6226 (vec_interleave_highv2si_be): New.
6227 (vec_interleave_highv2si_le): New.
6228 (insn_v4int_h): Handle big endian.
6229 (vec_interleave_lowv2si): Handle big endian.
6230 (vec_interleave_lowv2si_be): New.
6231 (vec_interleave_lowv2si_le): New.
6232 (insn_v4int_l): Handle big endian.
6233 * config/tilegx/tilegx.opt (mbig-endian): New option.
6234 (mlittle-endian): New option.
6235 * doc/install.texi: Document tilegxbe-linux.
6236 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
6238 2014-02-24 Martin Jambor <mjambor@suse.cz>
6241 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
6242 there are no parameter descriptors.
6244 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
6246 PR rtl-optimization/60268
6247 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
6248 initialization to ...
6249 (sched_rgn_init): ... here.
6250 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
6252 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6254 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
6257 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
6259 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
6262 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6264 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
6265 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
6267 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6269 * config/microblaze/predicates.md: Add cmp_op predicate.
6270 * config/microblaze/microblaze.md: Add branch_compare instruction
6271 which uses cmp_op predicate and emits cmp insn before branch.
6272 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
6273 to microblaze_expand_conditional_branch and consolidate logic.
6274 (microblaze_expand_conditional_branch): emit branch_compare
6275 insn instead of handling cmp op separate from branch insn.
6277 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6279 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
6282 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6284 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
6285 define_insn with define_expand and new define_insn
6286 *altivec_lve<VI_char>x_internal.
6287 (altivec_stve<VI_char>x): Replace define_insn with define_expand
6288 and new define_insn *altivec_stve<VI_char>x_internal.
6289 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
6291 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
6293 (altivec_expand_stvex_be): New function.
6295 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
6297 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
6298 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
6299 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
6300 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
6302 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
6305 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
6306 instead of emit_move_insn.
6308 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6310 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
6312 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
6313 gen_altivec_vsumsws.
6315 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6317 * config/rs6000/altivec.md (altivec_lvxl): Rename as
6318 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
6319 (altivec_lvxl_<mode>): New define_expand incorporating
6320 -maltivec=be semantics where needed.
6321 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
6322 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
6323 semantics where needed.
6324 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
6325 (altivec_stvx_<mode>): New define_expand incorporating
6326 -maltivec=be semantics where needed.
6327 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
6328 VM2 iterator instead of V4SI.
6329 (altivec_stvxl_<mode>): New define_expand incorporating
6330 -maltivec=be semantics where needed.
6331 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
6332 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
6333 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
6334 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
6335 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
6336 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
6337 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
6338 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
6339 ALTIVEC_BUILTIN_STVXL.
6340 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
6341 (altivec_expand_stvx_be): Likewise.
6342 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
6343 (altivec_expand_lvx_be): Likewise.
6344 (altivec_expand_stvx_be): Likewise.
6345 (altivec_expand_builtin): Add cases for
6346 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
6347 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
6348 (altivec_init_builtins): Add definitions for
6349 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
6350 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
6352 2014-02-21 Catherine Moore <clm@codesourcery.com>
6354 * doc/invoke.texi (mvirt, mno-virt): Document.
6355 * config/mips/mips.opt (mvirt): New option.
6356 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
6358 2014-02-21 Richard Biener <rguenther@suse.de>
6360 PR tree-optimization/60276
6361 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
6362 (STMT_VINFO_MIN_NEG_DIST): New macro.
6363 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
6364 STMT_VINFO_MIN_NEG_DIST.
6365 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
6366 made for negative dependence distances still hold.
6368 2014-02-21 Richard Biener <rguenther@suse.de>
6371 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
6372 DECL_INITIAL for globals not in the current function context.
6374 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6376 PR tree-optimization/56490
6377 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
6378 * tree-ssa-uninit.c: Include params.h.
6379 (compute_control_dep_chain): Add num_calls argument, return false
6380 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
6381 num_calls to recursive call.
6382 (find_predicates): Change dep_chain into normal array,
6383 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
6384 variable and adjust compute_control_dep_chain caller.
6385 (find_def_preds): Likewise.
6387 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
6389 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
6390 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
6392 2014-02-21 Nick Clifton <nickc@redhat.com>
6394 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
6395 (pushhi1): Likewise.
6396 (popqi1): Add mode to pre_dec.
6399 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6401 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
6402 mode for mask of V8SFmode permutation.
6404 2014-02-20 Richard Henderson <rth@redhat.com>
6407 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
6408 a new pseudo for OLDVAL.
6410 2014-02-20 Jakub Jelinek <jakub@redhat.com>
6413 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
6414 gen_reg_rtx if d->testing_p.
6415 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
6416 if d->testing_p and we will certainly return true.
6417 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6420 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6422 * emit-rtl.c (gen_reg_rtx): Assert that
6423 crtl->emit.regno_pointer_align_length is non-zero.
6425 2014-02-20 Richard Henderson <rth@redhat.com>
6428 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6429 on failure the store back into EXPECT.
6431 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6432 Sandra Loosemore <sandra@codesourcery.com>
6434 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6435 * config/nios2/nios2.c (nios2_function_profiler): Add
6436 -fPIC (flag_pic == 2) support.
6437 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6438 (nios2_large_offset_p): New function.
6439 (nios2_unspec_reloc_p): Move up position, update to use
6440 nios2_large_offset_p.
6441 (nios2_unspec_address): Remove function.
6442 (nios2_unspec_offset): New function.
6443 (nios2_large_got_address): New function.
6444 (nios2_got_address): Add large offset support.
6445 (nios2_legitimize_tls_address): Update usage of removed and new
6447 (nios2_symbol_binds_local_p): New function.
6448 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
6449 (nios2_legitimize_address): Update to use nios2_large_offset_p.
6450 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
6451 (nios2_print_operand): Merge H/L processing, add hiadj/lo
6452 processing for (const (unspec ...)).
6453 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
6455 2014-02-20 Richard Biener <rguenther@suse.de>
6457 * tree-cfg.c (replace_uses_by): Mark altered BBs before
6458 doing the substitution.
6459 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
6461 2014-02-20 Martin Jambor <mjambor@suse.cz>
6464 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
6465 info when checking whether lattices are bottom.
6467 2014-02-20 Richard Biener <rguenther@suse.de>
6470 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
6473 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
6476 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
6477 parameter specifying the scaling.
6478 (inline_call): Update.
6479 (want_inline_recursively): Guard division by zero.
6480 (recursive_inlining): Update.
6481 * ipa-inline.h (clone_inlined_nodes): Update.
6483 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6486 * config/i386/i386.c (classify_argument): Pass structures of size
6487 64 bytes or less in register.
6489 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6490 Kirill Yukhin <kirill.yukhin@intel.com>
6492 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
6493 (_mm_rcp28_round_ss): Ditto.
6494 (_mm_rsqrt28_round_sd): Ditto.
6495 (_mm_rsqrt28_round_ss): Ditto.
6496 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
6497 (_mm_rcp14_round_ss): Ditto.
6498 (_mm_rsqrt14_round_sd): Ditto.
6499 (_mm_rsqrt14_round_ss): Ditto.
6500 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
6501 the first input operand, get rid of match_dup.
6502 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
6504 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
6506 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
6507 operand as the first input operand, set type attribute.
6508 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
6510 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
6511 operand as the first input operand, set type attribute.
6513 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6515 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
6518 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
6521 * config/i386/i386.c (construct_container): Remove TFmode check
6522 for X86_64_INTEGER_CLASS.
6524 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
6527 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
6528 only when -Wpsabi is enabled.
6530 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
6533 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
6534 passing arrays in registers are the same as for structs, so remove the
6535 special case for them.
6537 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
6539 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
6540 destination type, extract only the valid bits if the source type is not
6541 integral and has a different mode.
6543 2014-02-19 Richard Biener <rguenther@suse.de>
6546 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
6549 2014-02-19 Richard Biener <rguenther@suse.de>
6552 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
6553 (ipa_modify_call_arguments): Emit an argument load explicitely and
6554 preserve virtual SSA form there and for the replacement call.
6555 Do not update SSA form nor free dominance info.
6557 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6559 * ipa.c (function_and_variable_visibility): Also clear WEAK
6560 flag when disolving COMDAT_GROUP.
6562 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6564 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
6565 * ipa-prop.c (ipa_set_jf_known_type): Return early when
6567 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
6568 do more sanity checks.
6569 (detect_type_change): Return true when giving up early.
6570 (compute_complex_assign_jump_func): Fix type parameter of
6571 ipa_set_ancestor_jf.
6572 (compute_complex_ancestor_jump_func): Likewise.
6573 (update_jump_functions_after_inlining): Fix updating of
6575 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
6577 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6579 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
6580 inline clones when edge disappears.
6582 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6585 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
6586 Split 64-bit moves into 2 patterns. Do not allow the use of
6587 direct move for TDmode in little endian, since the decimal value
6588 has little endian bytes within a word, but the 64-bit pieces are
6589 ordered in a big endian fashion, and normal subreg's of TDmode are
6591 (mov<mode>_64bit_dm): Likewise.
6592 (movtd_64bit_nodm): Likewise.
6594 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6596 PR tree-optimization/60174
6597 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
6598 statement of an SSA_NAME that occurs in an abnormal PHI node.
6600 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6603 * final.c (SEEN_BB): Remove.
6604 (SEEN_NOTE, SEEN_EMITTED): Renumber.
6605 (final_scan_insn): Don't force_source_line on second
6606 NOTE_INSN_BASIC_BLOCK.
6608 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
6611 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
6612 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
6613 (type_natural_mode): Warn ABI change when %zmm register is not
6614 available for AVX512F vector value passing.
6616 2014-02-18 Kai Tietz <ktietz@redhat.com>
6619 * config/i386/i386.c (ix86_expand_prologue): Use value in
6620 rax register as displacement when restoring %r10 or %rax.
6621 Fix wrong offset when restoring both registers.
6623 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6625 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
6626 assertion with conditional return.
6628 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6629 Uros Bizjak <ubizjak@gmail.com>
6632 * config/i386/driver-i386.c (host_detect_local_cpu): If
6633 YMM state is not saved by the OS, also clear has_f16c. Move
6634 CPUID 0x80000001 handling before YMM state saving checking.
6636 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
6638 PR rtl-optimization/58960
6639 * haifa-sched.c (alloc_global_sched_pressure_data): New,
6640 factored out from ...
6641 (sched_init): ... here.
6642 (free_global_sched_pressure_data): New, factored out from ...
6643 (sched_finish): ... here.
6644 * sched-int.h (free_global_sched_pressure_data): Declare.
6645 * sched-rgn.c (nr_regions_initial): New static global.
6646 (haifa_find_rgns): Initialize it.
6647 (schedule_region): Disable sched-pressure for the newly
6650 2014-02-17 Richard Biener <rguenther@suse.de>
6652 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
6653 release SSA defs of pattern stmts.
6655 2014-02-17 Richard Biener <rguenther@suse.de>
6657 * tree-inline.c (expand_call_inline): Release the virtual
6658 operand defined by the call we are about to inline.
6660 2014-02-17 Richard Biener <rguenther@suse.de>
6662 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
6664 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
6665 Ilya Tocar <ilya.tocar@intel.com>
6667 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
6668 arguments order in builtin.
6669 (_mm512_permutexvar_epi64): Ditto.
6670 (_mm512_mask_permutexvar_epi64): Ditto
6671 (_mm512_maskz_permutexvar_epi32): Ditto
6672 (_mm512_permutexvar_epi32): Ditto
6673 (_mm512_mask_permutexvar_epi32): Ditto
6675 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6677 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
6678 (p8_vmrgow): Likewise.
6680 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6682 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
6685 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
6688 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
6689 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
6690 into 64-bit and 32-bit moves. On 64-bit moves, add support for
6691 using direct move instructions on ISA 2.07. Also adjust
6692 instruction length for 64-bit.
6693 (mov<mode>_64bit, TFmode/TDmode): Likewise.
6694 (mov<mode>_32bit, TFmode/TDmode): Likewise.
6696 2014-02-15 Alan Modra <amodra@gmail.com>
6700 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
6701 find_replacement on parts of insn rtl that might be reloaded.
6703 2014-02-15 Richard Biener <rguenther@suse.de>
6705 PR tree-optimization/60183
6706 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
6707 (tree_ssa_phiprop): Calculate and free post-dominators.
6709 2014-02-14 Jeff Law <law@redhat.com>
6711 PR rtl-optimization/60131
6712 * ree.c (get_extended_src_reg): New function.
6713 (combine_reaching_defs): Use it rather than assuming location of REG.
6714 (find_and_remove_re): Verify first operand of extension is
6715 a REG before adding the insns to the copy list.
6717 2014-02-14 Roland McGrath <mcgrathr@google.com>
6719 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
6720 * configure: Regenerated.
6721 * config.in: Regenerated.
6722 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
6723 instead of ASM_SHORT.
6725 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
6726 Richard Earnshaw <rearnsha@arm.com>
6728 PR rtl-optimization/59535
6729 * lra-constraints.c (process_alt_operands): Encourage alternative
6730 when unassigned pseudo class is superset of the alternative class.
6731 (inherit_reload_reg): Don't inherit when optimizing for code size.
6732 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
6733 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
6734 modes not less than 4 for Thumb1.
6736 2014-02-14 Kyle McMartin <kyle@redhat.com>
6739 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
6741 2014-02-14 Richard Biener <rguenther@suse.de>
6743 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
6744 (get_frame_arg): Drop the assert with langhook types_compatible_p.
6745 Do not strip INDIRECT_REFs.
6747 2014-02-14 Richard Biener <rguenther@suse.de>
6750 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
6751 DECL_FUNCTION_SPECIFIC_TARGET.
6752 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
6753 * tree-streamer-out.c (pack_ts_target_option): Remove.
6754 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
6755 (write_ts_function_decl_tree_pointers): Do not stream
6756 DECL_FUNCTION_SPECIFIC_TARGET.
6757 * tree-streamer-in.c (unpack_ts_target_option): Remove.
6758 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
6759 (lto_input_ts_function_decl_tree_pointers): Do not stream
6760 DECL_FUNCTION_SPECIFIC_TARGET.
6762 2014-02-14 Jakub Jelinek <jakub@redhat.com>
6764 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
6765 (get_initial_def_for_induction, vectorizable_induction): Ignore
6766 debug stmts when looking for exit_phi.
6767 (vectorizable_live_operation): Fix up condition.
6769 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6771 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
6772 nreverse() because it changes the content of original tree list.
6774 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6776 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
6777 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
6779 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6781 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
6782 GNU coding standards.
6784 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6787 * dwarf2out.c (gen_subprogram_die): Don't call
6788 add_calling_convention_attribute if subr_die is old_die.
6790 2014-02-13 Sharad Singhai <singhai@google.com>
6792 * doc/optinfo.texi: Fix order of nodes.
6794 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
6796 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
6797 operands[2], not operands[3].
6799 2014-02-13 Richard Biener <rguenther@suse.de>
6802 * doc/install.texi (ISL): Update recommended version to 0.12.2,
6803 mention the possibility of an in-tree build.
6804 (CLooG): Update recommended version to 0.18.1, mention the
6805 possibility of an in-tree build and clarify that the ISL
6806 bundled with CLooG does not work.
6808 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6811 * expr.c (compress_float_constant): If x is a hard register,
6812 extend into a pseudo and then move to x.
6814 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
6816 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
6817 caused by bad second argument to warning_at() with -mhotpatch and
6818 nested functions (e.g. with gfortran).
6820 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
6822 * opts.c (option_name): Remove "enabled by default" rider.
6824 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
6826 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
6828 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
6829 Uros Bizjak <ubizjak@gmail.com>
6832 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
6833 * configure: Regenerated.
6835 2014-02-12 Richard Biener <rguenther@suse.de>
6837 * vec.c (vec_prefix::calculate_allocation): Move as
6838 inline variant to vec.h.
6839 (vec_prefix::calculate_allocation_1): New out-of-line version.
6840 * vec.h (vec_prefix::calculate_allocation_1): Declare.
6841 (vec_prefix::m_has_auto_buf): Rename to ...
6842 (vec_prefix::m_using_auto_storage): ... this.
6843 (vec_prefix::calculate_allocation): Inline the easy cases
6844 and dispatch to calculate_allocation_1 which doesn't need the
6846 (va_heap::reserve): Use gcc_checking_assert.
6847 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
6848 m_using_auto_storage.
6849 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
6851 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
6852 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
6853 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
6855 2014-02-12 Richard Biener <rguenther@suse.de>
6857 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
6858 when we found a dependence.
6860 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
6862 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
6864 (maybe_fold_stmt): ... into this new function.
6865 * omp-low.c (lower_omp): Update comment.
6867 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
6870 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
6873 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
6875 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
6876 identifiers in comments.
6877 (cortexa53_extra_costs): Likewise.
6878 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
6879 (cortexa7_extra_costs): Likewise.
6880 (cortexa12_extra_costs): Likewise.
6881 (cortexa15_extra_costs): Likewise.
6882 (v7m_extra_costs): Likewise.
6884 2014-02-12 Richard Biener <rguenther@suse.de>
6887 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
6888 of posix_memalign being successful.
6889 (lower_stmt): Restrict lowering of posix_memalign to when
6890 -ftree-bit-ccp is enabled.
6892 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6894 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
6896 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
6898 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
6900 PR rtl-optimization/60116
6901 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
6902 other_insn once the combination has been validated.
6904 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
6907 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
6909 * ipa-devirt.c: Include demangle.h
6910 (odr_violation_reported): New static variable.
6911 (add_type_duplicate): Update odr_violations.
6912 (maybe_record_node): Add completep parameter; update it.
6913 (record_target_from_binfo): Add COMPLETEP parameter;
6914 update it as needed.
6915 (possible_polymorphic_call_targets_1): Likewise.
6916 (struct polymorphic_call_target_d): Add nonconstruction_targets;
6917 rename FINAL to COMPLETE.
6918 (record_targets_from_bases): Sanity check we found the binfo;
6919 fix COMPLETEP updating.
6920 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
6921 parameter, fix computing of COMPLETEP.
6922 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
6923 at LTO time do demangling.
6924 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
6925 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
6927 (gimple_get_virt_method_for_binfo): Likewise.
6928 * gimple-fold.h (gimple_get_virt_method_for_binfo,
6929 gimple_get_virt_method_for_vtable): Update prototypes.
6931 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
6934 * genautomata.c (add_presence_absence): Fix typo with
6935 {final_}presence_list.
6937 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
6940 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
6941 for VSX/Altivec vectors that land in GPR registers.
6943 2014-02-11 Richard Henderson <rth@redhat.com>
6944 Jakub Jelinek <jakub@redhat.com>
6947 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
6948 around drhs if type conversion to lacc->type is not useless.
6950 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6952 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
6954 (cortex-a57.cortex-a53): Likewise.
6955 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
6957 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6959 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
6962 2014-02-11 Renlin Li <Renlin.Li@arm.com>
6964 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
6965 add_options_for_arm_vfp3.
6967 2014-02-11 Jeff Law <law@redhat.com>
6970 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
6971 object with an undesirable mode.
6973 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6976 * config/i386/sol2-9.h: New file.
6977 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
6978 *-*-solaris2.9*): Use it.
6980 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6982 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
6983 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
6985 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6987 * config/microblaze/microblaze.c: Extend mcpu version format
6989 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
6991 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
6993 2014-02-10 Richard Henderson <rth@redhat.com>
6996 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
6997 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
6998 ms-abi vs -mno-accumulate-outgoing-args.
6999 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
7000 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
7003 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
7006 * cfgexpand.c (expand_asm_operands): Attach source location to
7007 ASM_INPUT rtx objects.
7008 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
7010 2014-02-10 Nick Clifton <nickc@redhat.com>
7012 * config/mn10300/mn10300.c (popcount): New function.
7013 (mn10300_expand_prologue): Include saved registers in stack usage
7016 2014-02-10 Jeff Law <law@redhat.com>
7019 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
7020 when changing the SET_DEST of a prior insn to avoid an input reload.
7022 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7024 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
7025 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
7026 -mcall-openbsd, or -mcall-linux.
7027 (CC1_ENDIAN_BIG_SPEC): Remove.
7028 (CC1_ENDIAN_LITTLE_SPEC): Remove.
7029 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7030 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
7031 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
7032 and %cc1_endian_default.
7033 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7035 2014-02-10 Richard Biener <rguenther@suse.de>
7037 PR tree-optimization/60115
7038 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
7039 MEM_REF handling. Properly verify that the accesses are not
7040 out of the objects bound.
7042 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7044 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
7047 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
7049 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
7050 proper constants and fix formatting.
7051 (possible_polymorphic_call_targets): Fix formatting.
7053 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
7054 Ilya Tocar <ilya.tocar@intel.com>
7056 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
7057 (_mm512_loadu_epi32): Renamed into...
7058 (_mm512_loadu_si512): This.
7059 (_mm512_storeu_epi32): Renamed into...
7060 (_mm512_storeu_si512): This.
7061 (_mm512_maskz_ceil_ps): Removed.
7062 (_mm512_maskz_ceil_pd): Ditto.
7063 (_mm512_maskz_floor_ps): Ditto.
7064 (_mm512_maskz_floor_pd): Ditto.
7065 (_mm512_floor_round_ps): Ditto.
7066 (_mm512_floor_round_pd): Ditto.
7067 (_mm512_ceil_round_ps): Ditto.
7068 (_mm512_ceil_round_pd): Ditto.
7069 (_mm512_mask_floor_round_ps): Ditto.
7070 (_mm512_mask_floor_round_pd): Ditto.
7071 (_mm512_mask_ceil_round_ps): Ditto.
7072 (_mm512_mask_ceil_round_pd): Ditto.
7073 (_mm512_maskz_floor_round_ps): Ditto.
7074 (_mm512_maskz_floor_round_pd): Ditto.
7075 (_mm512_maskz_ceil_round_ps): Ditto.
7076 (_mm512_maskz_ceil_round_pd): Ditto.
7077 (_mm512_expand_pd): Ditto.
7078 (_mm512_expand_ps): Ditto.
7079 * config/i386/i386.c (ix86_builtins): Remove
7080 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
7081 (bdesc_args): Ditto.
7082 * config/i386/predicates.md (const1256_operand): New.
7083 (const_1_to_2_operand): Ditto.
7084 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7085 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7086 (*avx512pf_gatherpf<mode>sf): Ditto.
7087 (avx512pf_gatherpf<mode>df): Ditto.
7088 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7089 (*avx512pf_gatherpf<mode>df): Ditto.
7090 (avx512pf_scatterpf<mode>sf): Ditto.
7091 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7092 (*avx512pf_scatterpf<mode>sf): Ditto.
7093 (avx512pf_scatterpf<mode>df): Ditto.
7094 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7095 (*avx512pf_scatterpf<mode>df): Ditto.
7096 (avx512f_expand<mode>): Removed.
7097 (<shift_insn><mode>3<mask_name>): Change predicate type.
7099 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7101 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
7102 not at the end of datarefs vector use ordered_remove to avoid
7103 reordering datarefs vector.
7106 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
7107 mark local addressable non-static vars as GOVD_PRIVATE
7108 instead of GOVD_LOCAL.
7109 * omp-low.c (lower_omp_for): Move gimple_bind_vars
7110 and BLOCK_VARS of gimple_bind_block to new_stmt rather
7114 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
7115 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
7116 assume_aligned or alloc_align attributes.
7117 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
7118 arguments. Handle also assume_aligned and alloc_align attributes.
7119 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
7120 calls to functions with assume_aligned or alloc_align attributes.
7121 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
7123 2014-02-08 Terry Guo <terry.guo@arm.com>
7125 * doc/invoke.texi: Document ARM -march=armv7e-m.
7127 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7129 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
7130 flag on __cilkrts_rethrow builtin.
7133 * ipa-cp.c (determine_versionability): Fail at -O0
7134 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
7135 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
7138 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7141 * tree-inline.c (copy_forbidden): Fail for
7142 __attribute__((optimize (0))) functions.
7144 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7146 * varpool.c: Include pointer-set.h.
7147 (varpool_remove_unreferenced_decls): Variables in other partitions
7148 will not be output; be however careful to not lose information
7151 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7153 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
7154 lookup in the vtable constructor.
7156 2014-02-07 Jeff Law <law@redhat.com>
7159 * config/m68k/m68k.md (ashldi_extsi): Turn into a
7160 define_insn_and_split.
7162 * ipa-inline.c (inline_small_functions): Fix typos.
7164 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7166 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
7167 (s390_can_use_return_insn): Declare.
7168 * config/s390/s390.h (EPILOGUE_USES): Define.
7169 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
7171 (s390_chunkify_start): Handle return JUMP_LABELs.
7172 (s390_early_mach): Emit a main_pool instruction on the entry edge.
7173 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
7174 (s390_can_use_return_insn): New functions.
7175 (s390_fix_long_loop_prediction): Handle conditional returns.
7176 (TARGET_SET_UP_BY_PROLOGUE): Define.
7177 * config/s390/s390.md (ANY_RETURN): New code iterator.
7178 (*creturn, *csimple_return, return, simple_return): New patterns.
7180 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7182 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
7183 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
7184 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
7185 REG_CFA_RESTORE list when deciding not to restore a register.
7187 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7189 * config/s390/s390.c: Include tree-pass.h and context.h.
7190 (s390_early_mach): New function, split out from...
7191 (s390_emit_prologue): ...here.
7192 (pass_data_s390_early_mach): New pass structure.
7193 (pass_s390_early_mach): New class.
7194 (s390_option_override): Create and register early_mach pass.
7195 Move to end of file.
7197 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7199 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
7200 to match for the exit block.
7202 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7204 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
7205 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
7206 Reject misaligned operands.
7208 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7210 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
7212 2014-02-07 Richard Biener <rguenther@suse.de>
7215 * gimple-low.c (lower_builtin_posix_memalign): New function.
7216 (lower_stmt): Call it to lower posix_memalign in a way
7217 to make alignment info accessible.
7219 2014-02-07 Jakub Jelinek <jakub@redhat.com>
7222 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
7223 __builtin_setjmp_receiver.
7225 2014-02-07 Richard Biener <rguenther@suse.de>
7228 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
7229 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
7230 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7231 Handle BUILT_IN_POSIX_MEMALIGN.
7232 (find_func_clobbers): Likewise.
7233 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
7234 (call_may_clobber_ref_p_1): Likewise.
7236 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7239 * ipa-devirt.c (record_target_from_binfo): Remove overactive
7242 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7245 * lto-cgraph.c (lto_output_node): Use
7246 symtab_get_symbol_partitioning_class.
7247 (lto_output_varpool_node): likewise.
7248 (symtab_get_symbol_partitioning_class): Move here from
7250 * cgraph.h (symbol_partitioning_class): Likewise.
7251 (symtab_get_symbol_partitioning_class): Declare.
7253 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7255 * ggc.h (ggc_internal_cleared_alloc): New macro.
7256 * vec.h (vec_safe_copy): Handle memory stats.
7257 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
7258 * target-globals.c (save_target_globals): Likewise.
7260 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7263 * expr.c (emit_move_resolve_push): Export; be bit more selective
7264 on when to clear alias set.
7265 * expr.h (emit_move_resolve_push): Declare.
7266 * function.h (struct function): Add tail_call_marked.
7267 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
7268 * config/i386/i386-protos.h (ix86_expand_push): Remove.
7269 * config/i386/i386.md (TImode move expander): De not call
7271 (FP push expanders): Preserve memory attributes.
7272 * config/i386/sse.md (push<mode>1): Remove.
7273 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
7274 (ix86_expand_push): Remove.
7275 * config/i386/mmx.md (push<mode>1): Remove.
7277 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7279 PR rtl-optimization/60030
7280 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
7281 lopart with paradoxical subreg before shifting it up by hprec.
7283 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7285 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
7286 Remove extra newline at end of file.
7287 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
7288 (arm_issue_rate): Handle cortexa57.
7289 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
7290 (cortex-a57.cortex-a53): Likewise.
7292 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7295 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
7296 don't record in REG_FRAME_RELATED_EXPR registers not set in that
7298 (arm_expand_prologue): Adjust all callers.
7299 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
7300 info, registers also at the lowest numbered registers side. Use
7301 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
7305 * var-tracking.c (adjust_mems): Before adding a SET to
7306 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
7308 2014-02-06 Alan Modra <amodra@gmail.com>
7311 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
7312 change SDmode to DDmode when lra_in_progress.
7314 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7317 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
7318 free_data_ref on the dr first, and before goto again also set dr
7319 to the next dr. For simd_lane_access, free old datarefs[i] before
7320 overwriting it. For get_vectype_for_scalar_type failure, don't
7321 free_data_ref if simd_lane_access.
7323 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
7326 * tree.h (opts_for_fn): New inline function.
7327 (opt_for_fn): Define.
7328 * config/i386/i386.c (ix86_function_regparm): Use
7329 opt_for_fn (decl, optimize) instead of optimize.
7331 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
7333 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
7334 for SYMBOL_REF in large memory model.
7336 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7338 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
7340 (cortex-a57): Likewise.
7341 (cortex-a57.cortex-a53): Likewise.
7343 2014-02-06 Yury Gribov <y.gribov@samsung.com>
7344 Kugan Vivekanandarajah <kuganv@linaro.org>
7346 * config/arm/arm.c (arm_vector_alignment_reachable): Check
7348 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
7350 2014-02-06 Richard Biener <rguenther@suse.de>
7352 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
7353 set_loop_copy and initialize_original_copy_tables.
7355 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
7357 * config/aarch64/aarch64-simd.md
7358 (aarch64_ashr_simddi): Change QI to SI.
7360 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7361 Jakub Jelinek <jakub@redhat.com>
7364 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
7367 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7369 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
7370 CODE_FOR_altivec_vpku[hw]um to
7371 CODE_FOR_altivec_vpku[hw]um_direct.
7372 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
7373 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
7374 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
7375 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
7377 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7379 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
7380 generation for -maltivec=be.
7381 (altivec_vsumsws): Simplify redundant test.
7383 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7385 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
7386 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
7387 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
7388 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
7389 gen_altivec_vpkuwum.
7390 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
7392 (altivec_vpks<VI_char>ss): Likewise.
7393 (altivec_vpks<VI_char>us): Likewise.
7394 (altivec_vpku<VI_char>us): Likewise.
7395 (altivec_vpku<VI_char>um): Likewise.
7396 (altivec_vpku<VI_char>um_direct): New (copy of
7397 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
7399 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
7400 target is little endian and -maltivec=be is not specified.
7401 (*altivec_vupkhs<VU_char>_direct): New (copy of
7402 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
7403 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
7404 target is little endian and -maltivec=be is not specified.
7405 (*altivec_vupkls<VU_char>_direct): New (copy of
7406 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
7407 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
7408 little endian and -maltivec=be is not specified.
7409 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
7410 little endian and -maltivec=be is not specified.
7412 2014-02-05 Richard Henderson <rth@redhat.com>
7415 * combine-stack-adj.c: Revert r206943.
7416 * sched-int.h (struct deps_desc): Add last_args_size.
7417 * sched-deps.c (init_deps): Initialize it.
7418 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7419 contain REG_ARGS_SIZE notes.
7421 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7423 * lto-cgraph.c (asm_nodes_output): Make global.
7424 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7425 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7426 (driver_handle_option): Handle OPT_fwpa.
7428 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7431 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7432 a comment typo and formatting issue. If odr_hash hasn't been
7433 created, return vNULL and set *completep to false.
7436 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7437 bb with no successors.
7439 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7442 * doc/invoke.texi (-march): Clarify documentation for ARM.
7446 2014-02-05 Richard Biener <rguenther@suse.de>
7448 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
7449 when not vectorizing because of too many alias checks.
7450 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
7451 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
7453 2014-02-05 Nick Clifton <nickc@redhat.com>
7455 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
7456 accept extended registers in any mode when compiling for the MN10300.
7458 2014-02-05 Yury Gribov <y.gribov@samsung.com>
7460 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
7461 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
7462 sanitization attributes.
7463 (can_inline_edge_p): Likewise.
7464 (sanitize_attrs_match_for_inline_p): New function.
7466 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7468 * ipa-prop.c (detect_type_change): Shor circuit testing of
7469 type changes on THIS pointer.
7471 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
7474 * config/pa/pa.c (legitimize_tls_address): Return original address
7475 if not passed a SYMBOL_REF rtx.
7476 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
7478 (pa_emit_move_sequence): Simplify TLS source operands.
7479 (pa_legitimate_constant_p): Reject all TLS constants.
7480 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
7481 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
7483 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7485 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
7486 groups when we know they are controlled by LTO.
7487 * varasm.c (default_binds_local_p_1): If object is in other partition,
7488 it will be resolved locally.
7490 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7492 * config/host-linux.c (linux_gt_pch_use_address): Don't
7493 use SSIZE_MAX because it is not always defined.
7495 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
7498 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
7499 threshold for pseudo splitting.
7500 (update_ebb_live_info): Process call argument hard registers and
7501 hard registers from insn definition too.
7502 (max_small_class_regs_num): New constant.
7503 (inherit_in_ebb): Update live hard regs through EBBs. Update
7504 reloads_num only for small register classes. Don't split for
7507 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
7510 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
7513 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7515 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
7518 2014-02-04 Marek Polacek <polacek@redhat.com>
7520 * gdbinit.in (pel): Define.
7522 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7524 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
7527 2014-02-04 Richard Biener <rguenther@suse.de>
7530 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
7531 in function context local.
7532 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
7533 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
7534 similar to LTO_imported_decl_ref.
7536 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7538 PR tree-optimization/60002
7539 * cgraphclones.c (build_function_decl_skip_args): Clear
7542 PR tree-optimization/60023
7543 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
7544 false to gsi_replace.
7545 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
7546 has been in some EH region and vec_stmt could throw, add
7547 vec_stmt into the same EH region.
7548 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
7549 has no lhs, ignore it.
7550 * internal-fn.c (expand_MASK_LOAD): Likewise.
7553 * tree-inline.c (copy_forbidden): Fail for
7554 __attribute__((optimize (0))) functions.
7557 * omp-low.c (simd_clone_struct_copy): If from->inbranch
7558 is set, copy one less argument.
7559 (expand_simd_clones): Don't subtract clone_info->inbranch
7560 from simd_clone_struct_alloc argument.
7562 PR rtl-optimization/57915
7563 * recog.c (simplify_while_replacing): If all unary/binary/relational
7564 operation arguments are constant, attempt to simplify those.
7567 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
7568 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
7570 2014-02-04 Richard Biener <rguenther@suse.de>
7572 PR tree-optimization/60012
7573 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
7574 TBAA disambiguation to all DDRs.
7576 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7579 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
7580 (LINK_SPEC): Use it for -shared, -shared-libgcc.
7582 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7585 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
7587 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7589 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
7590 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
7592 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7595 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
7596 to figure out targets of polymorphic calls with known decl.
7597 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7598 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
7599 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
7600 (get_polymorphic_call_info): ... here.
7601 (get_polymorphic_call_info_from_invariant): New function.
7603 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7605 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
7606 lookup via vtable pointer; check for type consistency
7607 and turn inconsitent facts into UNREACHABLE.
7608 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7609 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
7610 type inconsistent querries; return UNREACHABLE instead.
7612 2014-02-03 Richard Henderson <rth@twiddle.net>
7615 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
7616 already processed this node.
7617 (normalize_one_pred_1): Pass along mark_set.
7618 (normalize_one_pred): Create and destroy a pointer_set_t.
7619 (normalize_one_pred_chain): Likewise.
7621 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
7623 PR gcov-profile/58602
7624 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
7626 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7629 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
7630 -fno-devirtualize; try to devirtualize by the knowledge of
7631 virtual table pointer given by aggregate propagation.
7632 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7633 (ipa_print_node_jump_functions): Dump also offset that
7634 is relevant for polymorphic calls.
7635 (determine_known_aggregate_parts): Add arg_type parameter; use it
7636 instead of determining the type from pointer type.
7637 (ipa_compute_jump_functions_for_edge): Update call of
7638 determine_known_aggregate_parts.
7639 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
7640 (gimple_get_virt_method_for_binfo): ... here; simplify using
7641 vtable_pointer_value_to_vtable.
7642 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
7643 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
7644 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
7645 (vtable_pointer_value_to_vtable): Break out from ...; handle also
7647 (vtable_pointer_value_to_binfo): ... here.
7648 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
7650 2014-02-03 Teresa Johnson <tejohnson@google.com>
7652 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
7653 redef of outer loop index variable.
7655 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
7659 * doc/extend.texi (Function Attributes): Typo.
7661 2014-02-03 Cong Hou <congh@google.com>
7663 PR tree-optimization/60000
7664 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
7665 if the vectorized statement is a store. A store statement can only
7666 appear at the end of pattern statements.
7668 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7670 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
7671 (ix86_option_override_internal): Default long double to 64-bit for
7672 32-bit Bionic and to 128-bit for 64-bit Bionic.
7674 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
7675 TARGET_LONG_DOUBLE_128 is true.
7676 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
7678 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
7679 (mlong-double-64): Negate -mlong-double-128.
7680 (mlong-double-128): New option.
7682 * config/i386/i386-c.c (ix86_target_macros): Define
7683 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
7685 * doc/invoke.texi: Document -mlong-double-128.
7687 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7689 PR rtl-optimization/60024
7690 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
7692 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
7694 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
7696 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7698 PR rtl-optimization/57662
7699 * sel-sched.c (code_motion_path_driver): Do not mark already not
7700 existing blocks in the visiting bitmap.
7702 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7704 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
7705 on the insn being emitted.
7707 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
7708 Will Deacon <will.deacon@arm.com>
7710 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
7712 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7714 * config/arm/arm-tables.opt: Regenerate.
7716 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7718 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
7719 for vector types other than V16QImode.
7720 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
7721 define_expand, and call altivec_expand_vec_perm_le when producing
7722 code with little endian element order.
7723 (*altivec_vperm_<mode>_internal): New insn having previous
7724 behavior of altivec_vperm_<mode>.
7725 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
7726 altivec_expand_vec_perm_le when producing code with little endian
7728 (*altivec_vperm_<mode>_uns_internal): New insn having previous
7729 behavior of altivec_vperm_<mode>_uns.
7731 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7733 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
7734 (altivec_vsumsws): Add handling for -maltivec=be with a little
7736 (altivec_vsumsws_direct): New.
7737 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
7738 gen_altivec_vsumsws.
7740 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7742 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
7743 vtable_pointer_value_to_binfo): New functions.
7744 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
7745 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
7747 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
7749 * config/nios2/nios2.md (load_got_register): Initialize GOT
7750 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
7751 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
7753 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7755 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
7756 preserverd by passthrough, do not propagate the type.
7758 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7760 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
7761 (mips_atomic_assign_expand_fenv): New function.
7762 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
7764 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7766 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
7767 (__builtin_mips_set_fcsr): Likewise.
7768 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
7769 MIPS_USI_FTYPE_VOID.
7770 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
7771 (mips16_expand_set_fcsr): Likewise.
7772 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
7773 (mips16_set_fcsr_stub): Likewise.
7774 (mips16_get_fcsr_one_only_stub): New class.
7775 (mips16_set_fcsr_one_only_stub): Likewise.
7776 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
7777 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
7778 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
7779 (hard_float): New availability predicate.
7780 (mips_builtins): Add get_fcsr and set_fcsr.
7781 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
7782 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
7783 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
7784 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
7785 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
7788 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7790 * config/mips/mips.c (mips_one_only_stub): New class.
7791 (mips_need_mips16_rdhwr_p): Replace with...
7792 (mips16_rdhwr_stub): ...this new variable.
7793 (mips16_stub_call_address): New function.
7794 (mips16_rdhwr_one_only_stub): New class.
7795 (mips_expand_thread_pointer): Use mips16_stub_call_address.
7796 (mips_output_mips16_rdhwr): Delete.
7797 (mips_finish_stub): New function.
7798 (mips_code_end): Use it to handle rdhwr stubs.
7800 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
7803 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
7804 when calculating size of integer atomic types.
7806 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
7808 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
7810 2014-02-01 Jakub Jelinek <jakub@redhat.com>
7812 PR tree-optimization/60003
7813 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
7814 * profile.c (branch_prob): Use gimple_call_builtin_p
7815 to check for BUILT_IN_SETJMP_RECEIVER.
7816 * tree-inline.c (copy_bb): Call notice_special_calls.
7818 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
7821 * lra-constraints.c (process_alt_operands): Update reload_sum only
7824 2014-01-31 Richard Henderson <rth@redhat.com>
7827 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
7828 until after else_eh is processed.
7830 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7832 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
7833 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
7834 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
7835 in smmintrin.h, remove them.
7836 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
7837 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
7838 * config/i386/i386.md (ROUND_SAE): Fix value.
7839 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
7840 (const48_operand): New.
7841 * config/i386/subst.md (round), (round_expand): Use
7842 const_4_or_8_to_11_operand.
7843 (round_saeonly), (round_saeonly_expand): Use const48_operand.
7845 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7847 * config/i386/constraints.md (Yk): Swap meaning with k.
7848 * config/i386/i386.md (movhi_internal): Change Yk to k.
7849 (movqi_internal): Ditto.
7850 (*k<logic><mode>): Ditto.
7853 (kandn<mode>): Ditto.
7854 (*<code>hi_1): Ditto.
7855 (*<code>qi_1): Ditto.
7856 (kxnor<mode>): Ditto.
7857 (kortestzhi): Ditto.
7858 (kortestchi): Ditto.
7860 (*one_cmplhi2_1): Ditto.
7861 (*one_cmplqi2_1): Ditto.
7862 * config/i386/sse.md (): Change k to Yk.
7863 (avx512f_load<mode>_mask): Ditto.
7864 (avx512f_blendm<mode>): Ditto.
7865 (avx512f_store<mode>_mask): Ditto.
7866 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
7867 (avx512f_storedqu<mode>_mask): Ditto.
7868 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
7870 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
7871 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
7872 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
7873 (avx512f_maskcmp<mode>3): Ditto.
7874 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
7875 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
7876 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
7877 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
7878 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
7879 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
7880 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
7881 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
7882 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
7883 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
7884 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
7885 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
7886 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
7887 (vec_extract_lo_<mode>_maskm): Ditto.
7888 (vec_extract_hi_<mode>_maskm): Ditto.
7889 (avx512f_vternlog<mode>_mask): Ditto.
7890 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
7891 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
7892 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
7893 (avx512f_<code>v8div16qi2_mask): Ditto.
7894 (avx512f_<code>v8div16qi2_mask_store): Ditto.
7895 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
7896 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
7897 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
7898 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
7899 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7900 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7901 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7902 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7903 (avx512cd_maskb_vec_dupv8di): Ditto.
7904 (avx512cd_maskw_vec_dupv16si): Ditto.
7905 (avx512f_vpermi2var<mode>3_maskz): Ditto.
7906 (avx512f_vpermi2var<mode>3_mask): Ditto.
7907 (avx512f_vpermi2var<mode>3_mask): Ditto.
7908 (avx512f_vpermt2var<mode>3_maskz): Ditto.
7909 (*avx512f_gathersi<mode>): Ditto.
7910 (*avx512f_gathersi<mode>_2): Ditto.
7911 (*avx512f_gatherdi<mode>): Ditto.
7912 (*avx512f_gatherdi<mode>_2): Ditto.
7913 (*avx512f_scattersi<mode>): Ditto.
7914 (*avx512f_scatterdi<mode>): Ditto.
7915 (avx512f_compress<mode>_mask): Ditto.
7916 (avx512f_compressstore<mode>_mask): Ditto.
7917 (avx512f_expand<mode>_mask): Ditto.
7918 * config/i386/subst.md (mask): Change k to Yk.
7919 (mask_scalar_merge): Ditto.
7922 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
7924 * doc/extend.texi (Vector Extensions): Document ?: in C++.
7926 2014-01-31 Richard Biener <rguenther@suse.de>
7929 * builtins.c (fold_builtin_memory_op): Make sure to not
7930 use a floating-point mode or a boolean or enumeral type for
7933 2014-01-30 DJ Delorie <dj@redhat.com>
7935 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
7936 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
7937 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
7938 whenever main() has an epilogue.
7940 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7942 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
7943 unused variable "field".
7944 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
7945 (vsx_mergeh_<mode>): Likewise.
7946 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
7947 (altivec_vmrghh): Likewise.
7948 (altivec_vmrghw): Likewise.
7949 (altivec_vmrglb): Likewise.
7950 (altivec_vmrglh): Likewise.
7951 (altivec_vmrglw): Likewise.
7952 (altivec_vspltb): Add missing uses.
7953 (altivec_vsplth): Likewise.
7954 (altivec_vspltw): Likewise.
7955 (altivec_vspltsf): Likewise.
7957 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7960 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
7961 frame related instructions.
7963 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
7965 PR rtl-optimization/59959
7966 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
7967 any reload of register whose subreg is invalid.
7969 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7971 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
7972 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
7973 Add missing return type - void.
7975 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7977 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
7978 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
7979 remove element index adjustment for endian (now handled in vsx.md
7981 (altivec_expand_vec_perm_const): Use
7982 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
7983 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
7984 (vsx_xxspltw_<mode>): Adjust element index for little endian.
7985 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
7986 define_expand and a new define_insn *altivec_vspltb_internal;
7987 adjust for -maltivec=be on a little endian target.
7988 (altivec_vspltb_direct): New.
7989 (altivec_vsplth): Divide into a define_expand and a new
7990 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
7991 little endian target.
7992 (altivec_vsplth_direct): New.
7993 (altivec_vspltw): Divide into a define_expand and a new
7994 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
7995 little endian target.
7996 (altivec_vspltw_direct): New.
7997 (altivec_vspltsf): Divide into a define_expand and a new
7998 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
7999 a little endian target.
8001 2014-01-30 Richard Biener <rguenther@suse.de>
8003 PR tree-optimization/59993
8004 * tree-ssa-forwprop.c (associate_pointerplus): Check we
8005 can propagate form the earlier stmt and avoid the transform
8006 when the intermediate result is needed.
8008 2014-01-30 Alangi Derick <alangiderick@gmail.com>
8010 * README.Portability: Fix typo.
8012 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
8014 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
8015 comparison_operator with ordered_comparison_operator.
8017 2014-01-30 Nick Clifton <nickc@redhat.com>
8019 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
8020 Rename to mn10300_store_multiple_regs.
8021 * config/mn10300/mn10300.c: Likewise.
8022 * config/mn10300/mn10300.md (store_movm): Fix typo: call
8023 store_multiple_regs.
8024 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
8025 Call mn10300_store_multiple_regs.
8027 2014-01-30 Nick Clifton <nickc@redhat.com>
8028 DJ Delorie <dj@redhat.com>
8030 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
8031 %fp 2 to keep registers after it properly word-aligned.
8032 (rl78_alloc_physical_registers_umul): Handle the case where both
8033 input operands are the same.
8035 2014-01-30 Richard Biener <rguenther@suse.de>
8037 PR tree-optimization/59903
8038 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
8041 2014-01-30 Jason Merrill <jason@redhat.com>
8044 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
8047 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
8049 2014-01-30 Richard Biener <rguenther@suse.de>
8051 PR tree-optimization/59951
8052 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
8054 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
8057 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
8058 SFmode to DFmode case.
8060 2014-01-29 DJ Delorie <dj@redhat.com>
8062 * config/msp430/msp430.opt (-minrt): New.
8063 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
8065 (ENDFILE_SPEC): Likewise.
8067 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
8069 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
8070 (estimate_function_body_sizes): Use it.
8072 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
8075 * dwarf2out.c (is_cxx_auto): New.
8076 (is_base_type): Use it.
8077 (gen_type_die_with_usage): Likewise.
8079 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8081 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
8082 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
8083 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
8084 -maltivec=be with LE targets.
8085 (vsx_mergeh_<mode>): Likewise.
8086 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
8087 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
8088 (altivec_vmrghb): Replace with define_expand and new
8089 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
8090 (altivec_vmrghb_direct): New define_insn.
8091 (altivec_vmrghh): Replace with define_expand and new
8092 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
8093 (altivec_vmrghh_direct): New define_insn.
8094 (altivec_vmrghw): Replace with define_expand and new
8095 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
8096 (altivec_vmrghw_direct): New define_insn.
8097 (*altivec_vmrghsf): Adjust for endianness.
8098 (altivec_vmrglb): Replace with define_expand and new
8099 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
8100 (altivec_vmrglb_direct): New define_insn.
8101 (altivec_vmrglh): Replace with define_expand and new
8102 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
8103 (altivec_vmrglh_direct): New define_insn.
8104 (altivec_vmrglw): Replace with define_expand and new
8105 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
8106 (altivec_vmrglw_direct): New define_insn.
8107 (*altivec_vmrglsf): Adjust for endianness.
8108 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8109 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8110 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8111 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8112 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8113 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8114 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8115 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8117 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
8119 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
8120 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
8123 2014-01-29 Richard Biener <rguenther@suse.de>
8125 PR tree-optimization/58742
8126 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
8127 associate_pointerplus_align.
8128 (associate_pointerplus_diff): New function.
8129 (associate_pointerplus): Likewise. Call associate_pointerplus_align
8130 and associate_pointerplus_diff.
8132 2014-01-29 Richard Biener <rguenther@suse.de>
8134 * lto-streamer.h (LTO_major_version): Bump to 3.
8135 (LTO_minor_version): Reset to 0.
8137 2014-01-29 Renlin Li <Renlin.Li@arm.com>
8139 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
8140 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
8141 (arm_file_start): Generate correct asm header for armv7ve.
8142 * config/arm/bpabi.h: Add multilib support for armv7ve.
8143 * config/arm/driver-arm.c: Change the architectures of cortex-a7
8144 and cortex-a15 to armv7ve.
8145 * config/arm/t-aprofile: Add multilib support for armv7ve.
8146 * doc/invoke.texi: Document -march=armv7ve.
8148 2014-01-29 Richard Biener <rguenther@suse.de>
8150 PR tree-optimization/58742
8151 * tree-ssa-forwprop.c (associate_plusminus): Return true
8152 if we changed sth, defer EH cleanup to ...
8153 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
8154 (simplify_mult): New function.
8156 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8159 PR tree-optimization/59920
8160 * tree.c (build_common_builtin_nodes): Remove
8161 __builtin_setjmp_dispatcher initialization.
8162 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
8163 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
8164 instead of gsi_after_labels + manually skipping debug stmts.
8165 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
8166 ignore bbs with IFN_ABNORMAL_DISPATCHER.
8167 * tree-inline.c (copy_edges_for_bb): Remove
8168 can_make_abnormal_goto argument, instead add abnormal_goto_dest
8169 argument. Ignore computed_goto_p stmts. Don't call
8170 make_abnormal_goto_edges. If a call might need abnormal edges
8171 for non-local gotos, see if it already has an edge to
8172 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
8173 with true argument, don't do anything then, otherwise add
8174 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
8175 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
8177 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
8178 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
8179 (lower_stmt): Don't set data->calls_builtin_setjmp.
8180 (lower_builtin_setjmp): Adjust comment.
8181 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
8182 * tree-cfg.c (found_computed_goto): Remove.
8183 (factor_computed_gotos): Remove.
8184 (make_goto_expr_edges): Return bool, true for computed gotos.
8185 Don't call make_abnormal_goto_edges.
8186 (build_gimple_cfg): Don't set found_computed_goto, don't call
8187 factor_computed_gotos.
8188 (computed_goto_p): No longer static.
8189 (make_blocks): Don't set found_computed_goto.
8190 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
8191 (make_edges): If make_goto_expr_edges returns true, push bb
8192 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
8193 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
8194 vector. Record mapping between bbs and OpenMP regions if there
8195 are any, adjust make_gimple_omp_edges caller. Call
8196 handle_abnormal_edges.
8197 (make_abnormal_goto_edges): Remove.
8198 * tree-cfg.h (make_abnormal_goto_edges): Remove.
8199 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
8200 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
8201 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
8202 * internal-fn.def (ABNORMAL_DISPATCHER): New.
8203 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
8204 filling *region also set *region_idx to (*region)->entry->index.
8207 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
8208 For REGs set ORIGINAL_REGNO.
8210 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
8212 * doc/md.texi: Mention that a target shouldn't implement
8213 vec_widen_(s|u)mul_even/odd pair if it is less efficient
8216 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8218 PR tree-optimization/59594
8219 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
8220 a copy of the datarefs vector rather than the vector itself.
8222 2014-01-28 Jason Merrill <jason@redhat.com>
8225 * dwarf2out.c (auto_die): New static.
8226 (gen_type_die_with_usage): Handle C++1y 'auto'.
8227 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
8230 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
8233 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
8234 (SPEC_X32): Likewise.
8235 (SPEC_64): Likewise.
8236 * config/i386/i386.c (ix86_option_override_internal): Turn off
8237 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
8239 (x86_file_start): Output .code16gcc for TARGET_16BIT.
8240 * config/i386/i386.h (TARGET_16BIT): New macro.
8241 (TARGET_16BIT_P): Likewise.
8242 * config/i386/i386.opt: Add m16.
8243 * doc/invoke.texi: Document -m16.
8245 2014-01-28 Jakub Jelinek <jakub@redhat.com>
8247 PR preprocessor/59935
8248 * input.c (location_get_source_line): Bail out on when line number
8249 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
8251 2014-01-28 Richard Biener <rguenther@suse.de>
8253 PR tree-optimization/58742
8254 * tree-ssa-forwprop.c (associate_plusminus): Handle
8255 pointer subtraction of the form (T)(P + A) - (T)P.
8257 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8259 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
8262 2014-01-28 Richard Biener <rguenther@suse.de>
8265 2014-01-28 Richard Biener <rguenther@suse.de>
8267 PR rtl-optimization/45364
8268 PR rtl-optimization/59890
8269 * var-tracking.c (local_get_addr_clear_given_value): Handle
8270 already cleared slot.
8271 (val_reset): Handle not allocated local_get_addr_cache.
8272 (vt_find_locations): Use post-order on the inverted CFG.
8274 2014-01-28 Richard Biener <rguenther@suse.de>
8276 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
8278 2014-01-28 Richard Biener <rguenther@suse.de>
8280 PR rtl-optimization/45364
8281 PR rtl-optimization/59890
8282 * var-tracking.c (local_get_addr_clear_given_value): Handle
8283 already cleared slot.
8284 (val_reset): Handle not allocated local_get_addr_cache.
8285 (vt_find_locations): Use post-order on the inverted CFG.
8287 2014-01-28 Alan Modra <amodra@gmail.com>
8289 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
8290 * configure.ac <recursive call for build != host>: Define
8291 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
8292 and LD_FOR_BUILD too.
8293 * configure: Regenerate.
8295 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
8297 * config/i386/i386.c (get_builtin_code_for_version): Separate
8298 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
8299 Broadwell from Haswell.
8301 2014-01-27 Steve Ellcey <sellcey@mips.com>
8303 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
8304 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
8305 * config/mips/mips.c (mips_option_override): Change setting
8307 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
8308 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
8309 Change from Mask to Var.
8311 2014-01-27 Jeff Law <law@redhat.com>
8313 * ipa-inline.c (inline_small_functions): Fix typo.
8315 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
8317 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
8318 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
8319 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
8320 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
8321 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
8322 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
8323 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
8324 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
8325 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
8326 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
8327 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
8328 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
8329 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
8330 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
8331 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
8332 (_mm512_storeu_epi64): Ditto.
8333 (_mm512_cmpge_epi32_mask): Ditto.
8334 (_mm512_cmpge_epu32_mask): Ditto.
8335 (_mm512_cmpge_epi64_mask): Ditto.
8336 (_mm512_cmpge_epu64_mask): Ditto.
8337 (_mm512_cmple_epi32_mask): Ditto.
8338 (_mm512_cmple_epu32_mask): Ditto.
8339 (_mm512_cmple_epi64_mask): Ditto.
8340 (_mm512_cmple_epu64_mask): Ditto.
8341 (_mm512_cmplt_epi32_mask): Ditto.
8342 (_mm512_cmplt_epu32_mask): Ditto.
8343 (_mm512_cmplt_epi64_mask): Ditto.
8344 (_mm512_cmplt_epu64_mask): Ditto.
8345 (_mm512_cmpneq_epi32_mask): Ditto.
8346 (_mm512_cmpneq_epu32_mask): Ditto.
8347 (_mm512_cmpneq_epi64_mask): Ditto.
8348 (_mm512_cmpneq_epu64_mask): Ditto.
8349 (_mm512_expand_pd): Ditto.
8350 (_mm512_expand_ps): Ditto.
8351 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
8352 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
8353 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
8354 * config/i386/i386.c (ix86_builtins): Add
8355 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
8356 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
8357 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
8358 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
8359 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
8360 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
8361 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
8362 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
8363 IX86_BUILTIN_PMOVUSQW512_MEM.
8364 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
8365 __builtin_ia32_pmovsqd512mem_mask,
8366 __builtin_ia32_pmovqd512mem_mask,
8367 __builtin_ia32_pmovusqw512mem_mask,
8368 __builtin_ia32_pmovsqw512mem_mask,
8369 __builtin_ia32_pmovqw512mem_mask,
8370 __builtin_ia32_pmovusdw512mem_mask,
8371 __builtin_ia32_pmovsdw512mem_mask,
8372 __builtin_ia32_pmovdw512mem_mask,
8373 __builtin_ia32_pmovqb512mem_mask,
8374 __builtin_ia32_pmovusqb512mem_mask,
8375 __builtin_ia32_pmovsqb512mem_mask,
8376 __builtin_ia32_pmovusdb512mem_mask,
8377 __builtin_ia32_pmovsdb512mem_mask,
8378 __builtin_ia32_pmovdb512mem_mask.
8379 (bdesc_args): Add __builtin_ia32_expanddf512,
8380 __builtin_ia32_expandsf512.
8381 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
8382 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
8383 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
8384 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
8385 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
8386 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
8387 (avx512f_<code>v8div16qi2_mask_store): This.
8388 (avx512f_expand<mode>): New.
8390 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
8392 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
8394 (_mm512_mask_prefetch_i64gather_pd): Ditto.
8395 (_mm512_prefetch_i32scatter_pd): Ditto.
8396 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
8397 (_mm512_prefetch_i64scatter_pd): Ditto.
8398 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
8399 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
8400 (_mm512_mask_prefetch_i64gather_ps): Ditto.
8401 (_mm512_prefetch_i32scatter_ps): Ditto.
8402 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
8403 (_mm512_prefetch_i64scatter_ps): Ditto.
8404 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
8405 * config/i386/i386-builtin-types.def: Define
8406 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
8407 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
8408 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
8409 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
8410 IX86_BUILTIN_SCATTERPFQPD.
8411 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
8412 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
8413 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
8414 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
8415 __builtin_ia32_scatterpfqps.
8416 (ix86_expand_builtin): Expand new built-ins.
8417 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8418 fix memory access data type.
8419 (*avx512pf_gatherpf<mode>_mask): Ditto.
8420 (*avx512pf_gatherpf<mode>): Ditto.
8421 (avx512pf_scatterpf<mode>): Ditto.
8422 (*avx512pf_scatterpf<mode>_mask): Ditto.
8423 (*avx512pf_scatterpf<mode>): Ditto.
8424 (GATHER_SCATTER_SF_MEM_MODE): New.
8425 (avx512pf_gatherpf<mode>df): Ditto.
8426 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8427 (*avx512pf_scatterpf<mode>df): Ditto.
8429 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8432 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8433 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8436 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8438 * common/config/arm/arm-common.c
8439 (arm_rewrite_mcpu): Handle multiple names.
8441 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8443 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8445 * gimple-builder.h (create_gimple_tmp): Delete.
8447 2014-01-27 Christian Bruel <christian.bruel@st.com>
8449 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
8452 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
8454 * config/pa/pa.md (call): Generate indirect long calls to non-local
8455 functions when outputing 32-bit code.
8456 (call_value): Likewise except for special call to buggy powf function.
8458 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
8459 portable runtime and PIC indirect calls.
8460 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
8461 and PIC call sequences. Use ldo instead of blr to set return register
8462 in PIC call sequence.
8464 2014-01-25 Walter Lee <walt@tilera.com>
8466 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
8467 avoid clobbering a live register.
8469 2014-01-25 Walter Lee <walt@tilera.com>
8471 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
8472 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
8473 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
8474 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
8476 2014-01-25 Walter Lee <walt@tilera.com>
8478 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
8479 arguments on even registers.
8480 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
8482 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
8483 (BIGGEST_ALIGNMENT): Ditto.
8484 (BIGGEST_FIELD_ALIGNMENT): Ditto.
8486 2014-01-25 Walter Lee <walt@tilera.com>
8488 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
8489 insns before bundling.
8490 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8492 2014-01-25 Walter Lee <walt@tilera.com>
8494 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
8495 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
8496 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
8498 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8500 * config/mips/constraints.md (kl): Delete.
8501 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
8502 define expands, using...
8503 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
8504 instructions for MIPS16.
8505 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
8506 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
8508 2014-01-25 Walter Lee <walt@tilera.com>
8510 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
8514 2014-01-25 Walter Lee <walt@tilera.com>
8516 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
8517 (TARGET_EXPAND_TO_RTL_HOOK): Define.
8519 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8521 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
8524 2014-01-25 Jakub Jelinek <jakub@redhat.com>
8526 * print-rtl.c (in_call_function_usage): New var.
8527 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
8528 EXPR_LIST mode as mode and not as reg note name.
8531 * cfgloopmanip.c (copy_loop_info): If
8532 loop->warned_aggressive_loop_optimizations, make sure
8533 the flag is set in target loop too.
8535 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
8537 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
8539 * builtins.def: Likewise.
8540 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
8541 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
8542 * ira.c (ira_setup_eliminable_regset): Likewise.
8543 * omp-low.c (gate_expand_omp): Likewise.
8544 (execute_lower_omp): Likewise.
8545 (diagnose_sb_0): Likewise.
8546 (gate_diagnose_omp_blocks): Likewise.
8547 (simd_clone_clauses_extract): Likewise.
8550 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8552 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
8553 correction for little endian...
8554 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
8557 2014-01-24 Jeff Law <law@redhat.com>
8559 PR tree-optimization/59919
8560 * tree-vrp.c (find_assert_locations_1): Do not register asserts
8561 for non-returning calls.
8563 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
8565 * common/config/aarch64/aarch64-common.c
8566 (aarch64_rewrite_mcpu): Handle multiple names.
8567 * config/aarch64/aarch64.h
8568 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8570 2014-01-24 Dodji Seketeli <dodji@redhat.com>
8572 * input.c (add_file_to_cache_tab): Handle the case where fopen
8575 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
8578 * config/i386/i386.md (pushsf splitter): Get stack adjustment
8579 from push operand if code of push isn't PRE_DEC.
8581 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8584 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
8585 -mquad-memory-atomic. Update -mquad-memory documentation to say
8586 it is only used for non-atomic loads/stores.
8588 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
8589 -mquad-memory or -mquad-memory-atomic switches.
8591 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
8592 -mquad-memory-atomic to ISA 2.07 support.
8594 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
8595 to separate support of normal quad word memory operations (ldq, stq)
8596 from the atomic quad word memory operations.
8598 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
8599 support to separate non-atomic quad word operations from atomic
8600 quad word operations. Disable non-atomic quad word operations in
8601 little endian mode so that we don't have to swap words after the
8602 load and before the store.
8603 (quad_load_store_p): Add comment about atomic quad word support.
8604 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
8605 options printed with -mdebug=reg.
8607 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
8608 -mquad-memory-atomic as the test for whether we have quad word
8609 atomic instructions.
8610 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
8611 or -mp8-vector are used, allow byte/half-word atomic operations.
8613 * config/rs6000/sync.md (load_lockedti): Insure that the address
8614 is a proper indexed or indirect address for the lqarx instruction.
8615 On little endian systems, swap the hi/lo registers after the lqarx
8617 (load_lockedpti): Use indexed_or_indirect_operand predicate to
8618 insure the address is valid for the lqarx instruction.
8619 (store_conditionalti): Insure that the address is a proper indexed
8620 or indirect address for the stqcrx. instruction. On little endian
8621 systems, swap the hi/lo registers before doing the stqcrx.
8623 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
8624 insure the address is valid for the stqcrx. instruction.
8626 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8627 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
8628 type of quad memory support is available.
8630 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
8633 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
8634 there is a danger of looping.
8636 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
8638 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8639 force flag_ira_loop_pressure if set via command line.
8641 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8643 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
8644 (ashr_simd): New builtin handling DI mode.
8645 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
8646 (aarch64_sshr_simddi): New match pattern.
8647 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
8648 (vshrd_n_s64): Likewise.
8649 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
8651 2014-01-23 Nick Clifton <nickc@redhat.com>
8653 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
8654 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
8655 favour of mcu specific scripts.
8656 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
8659 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8660 Alex Velenko <Alex.Velenko@arm.com>
8662 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
8663 (vaddv_s16): Likewise.
8664 (vaddv_s32): Likewise.
8665 (vaddv_u8): Likewise.
8666 (vaddv_u16): Likewise.
8667 (vaddv_u32): Likewise.
8668 (vaddvq_s8): Likewise.
8669 (vaddvq_s16): Likewise.
8670 (vaddvq_s32): Likewise.
8671 (vaddvq_s64): Likewise.
8672 (vaddvq_u8): Likewise.
8673 (vaddvq_u16): Likewise.
8674 (vaddvq_u32): Likewise.
8675 (vaddvq_u64): Likewise.
8676 (vaddv_f32): Likewise.
8677 (vaddvq_f32): Likewise.
8678 (vaddvq_f64): Likewise.
8679 (vmaxv_f32): Likewise.
8680 (vmaxv_s8): Likewise.
8681 (vmaxv_s16): Likewise.
8682 (vmaxv_s32): Likewise.
8683 (vmaxv_u8): Likewise.
8684 (vmaxv_u16): Likewise.
8685 (vmaxv_u32): Likewise.
8686 (vmaxvq_f32): Likewise.
8687 (vmaxvq_f64): Likewise.
8688 (vmaxvq_s8): Likewise.
8689 (vmaxvq_s16): Likewise.
8690 (vmaxvq_s32): Likewise.
8691 (vmaxvq_u8): Likewise.
8692 (vmaxvq_u16): Likewise.
8693 (vmaxvq_u32): Likewise.
8694 (vmaxnmv_f32): Likewise.
8695 (vmaxnmvq_f32): Likewise.
8696 (vmaxnmvq_f64): Likewise.
8697 (vminv_f32): Likewise.
8698 (vminv_s8): Likewise.
8699 (vminv_s16): Likewise.
8700 (vminv_s32): Likewise.
8701 (vminv_u8): Likewise.
8702 (vminv_u16): Likewise.
8703 (vminv_u32): Likewise.
8704 (vminvq_f32): Likewise.
8705 (vminvq_f64): Likewise.
8706 (vminvq_s8): Likewise.
8707 (vminvq_s16): Likewise.
8708 (vminvq_s32): Likewise.
8709 (vminvq_u8): Likewise.
8710 (vminvq_u16): Likewise.
8711 (vminvq_u32): Likewise.
8712 (vminnmv_f32): Likewise.
8713 (vminnmvq_f32): Likewise.
8714 (vminnmvq_f64): Likewise.
8716 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8718 * config/aarch64/aarch64-simd.md
8719 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
8720 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
8721 (*aarch64_mul3_elt<mode>): Likewise.
8722 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
8723 (*aarch64_mul3_elt_to_64v2df): Likewise.
8724 (*aarch64_mla_elt<mode>): Likewise.
8725 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8726 (*aarch64_mls_elt<mode>): Likewise.
8727 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8728 (*aarch64_fma4_elt<mode>): Likewise.
8729 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8730 (*aarch64_fma4_elt_to_64v2df): Likewise.
8731 (*aarch64_fnma4_elt<mode>): Likewise.
8732 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8733 (*aarch64_fnma4_elt_to_64v2df): Likewise.
8734 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
8735 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
8736 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8737 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8738 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
8739 (aarch64_sqdmull_lane<mode>_internal): Likewise.
8740 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
8742 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
8744 * config/aarch64/aarch64-simd.md
8745 (aarch64_be_checked_get_lane<mode>): New define_expand.
8746 * config/aarch64/aarch64-simd-builtins.def
8747 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
8748 New builtin definition.
8749 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
8750 Use new safe be builtin.
8752 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8754 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
8756 (aarch64_be_st1<mode>): Likewise.
8757 (aarch_ld1<VALL:mode>): Define_expand modified.
8758 (aarch_st1<VALL:mode>): Likewise.
8759 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
8760 (UNSPEC_ST1): Likewise.
8762 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
8764 * config/microblaze/microblaze.md: Add trap insn and attribute
8766 2014-01-23 Dodji Seketeli <dodji@redhat.com>
8768 PR preprocessor/58580
8769 * input.h (location_get_source_line): Take an additional line_size
8771 (void diagnostics_file_cache_fini): Declare new function.
8772 * input.c (struct fcache): New type.
8773 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
8774 New static constants.
8775 (diagnostic_file_cache_init, total_lines_num)
8776 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
8777 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
8778 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
8779 (get_next_line, read_next_line, goto_next_line, read_line_num):
8780 New static function definitions.
8781 (diagnostic_file_cache_fini): New function.
8782 (location_get_source_line): Take an additional output line_len
8783 parameter. Re-write using lookup_or_add_file_to_cache_tab and
8785 * diagnostic.c (diagnostic_finish): Call
8786 diagnostic_file_cache_fini.
8787 (adjust_line): Take an additional input parameter for the length
8788 of the line, rather than calculating it with strlen.
8789 (diagnostic_show_locus): Adjust the use of
8790 location_get_source_line and adjust_line with respect to their new
8791 signature. While displaying a line now, do not stop at the first
8792 null byte. Rather, display the zero byte as a space and keep
8793 going until we reach the size of the line.
8794 * Makefile.in: Add vec.o to OBJS-libcommon
8796 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8797 Ilya Tocar <ilya.tocar@intel.com>
8799 * config/i386/avx512fintrin.h (_mm512_kmov): New.
8800 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
8801 (__builtin_ia32_kmov16): Ditto.
8802 * config/i386/i386.md (UNSPEC_KMOV): New.
8805 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8807 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
8808 (_mm512_storeu_si512): Ditto.
8810 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
8813 * rtl.h (get_referenced_operands): Declare.
8814 * recog.c (get_referenced_operands): New function.
8815 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
8816 operands have been referenced when recording LO_SUM references.
8818 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
8820 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
8822 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8824 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
8825 Enable for generic and recent AMD targets.
8827 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8829 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
8830 ARG_SIZE note when adjustment was eliminated.
8832 2014-01-22 Jeff Law <law@redhat.com>
8834 PR tree-optimization/59597
8835 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
8836 in file. Accept new argument REGISTERING and use it to modify
8837 dump output appropriately.
8838 (register_jump_thread): Corresponding changes.
8839 (mark_threaded_blocks): Reinstate code to cancel unprofitable
8840 thread paths involving joiner blocks. Add code to dump cancelled
8841 jump threading paths.
8843 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
8845 PR rtl-optimization/59477
8846 * lra-constraints.c (inherit_in_ebb): Process call for living hard
8847 regs. Update reloads_num and potential_reload_hard_regs for all insns.
8849 2014-01-22 Tom Tromey <tromey@redhat.com>
8851 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
8853 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
8855 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8857 PR rtl-optimization/59896
8858 * lra-constraints.c (process_alt_operands): Check unused note for
8859 matched operands of insn with no output reloads.
8861 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
8863 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
8864 (mips_move_from_gpr_cost): Likewise.
8866 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8868 PR rtl-optimization/59858
8869 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
8870 ira_class_hard_regs_num.
8871 (process_alt_operands): Increase reject for dying matched operand.
8873 2014-01-21 Jakub Jelinek <jakub@redhat.com>
8876 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
8877 smaller than size, perform several stores or loads and stores
8878 at dst + count - size to store or copy all of size bytes, rather
8879 than just last modesize bytes.
8881 2014-01-20 DJ Delorie <dj@redhat.com>
8883 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
8884 that CLOBBERs are REGs before propogating their values.
8886 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
8889 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
8890 (cgraph_inline_failed_type): New function.
8891 * cgraph.h (DEFCIFCODE): Add type.
8892 (cgraph_inline_failed_type_t): New enum.
8893 (cgraph_inline_failed_type): New prototype.
8894 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
8895 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
8896 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
8897 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
8898 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
8899 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
8900 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
8901 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
8902 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
8903 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
8904 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
8905 OPTIMIZATION_MISMATCH.
8906 * tree-inline.c (expand_call_inline): Emit errors during
8907 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
8909 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8912 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
8913 mode attribute in insn output.
8915 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
8917 * output.h (output_constant): Delete.
8918 * varasm.c (output_constant): Make private.
8920 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
8922 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
8924 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8927 * tree.h (fold_builtin_strcat): New prototype.
8928 * builtins.c (fold_builtin_strcat): No longer static. Add len
8929 argument, if non-NULL, don't call c_strlen. Optimize
8930 directly into __builtin_memcpy instead of __builtin_strcpy.
8931 (fold_builtin_2): Adjust fold_builtin_strcat caller.
8932 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
8934 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8936 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8937 for SImode_address_operand operands, having only a REG argument.
8939 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
8941 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
8942 loader name using mbig-endian.
8943 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
8945 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8947 * doc/invoke.texi (-march): Clarify documentation for AArch64.
8951 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
8953 * config/aarch64/aarch64-protos.h
8954 (aarch64_cannot_change_mode_class_ptr): Declare.
8955 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
8956 aarch64_cannot_change_mode_class_ptr): New.
8957 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
8958 backend hook aarch64_cannot_change_mode_class.
8960 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8962 * common/config/aarch64/aarch64-common.c
8963 (aarch64_handle_option): Don't handle any option order logic here.
8964 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
8965 selected_cpu, warn on architecture version mismatch.
8966 (aarch64_override_options): Fix parsing order for option strings.
8968 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8969 Iain Sandoe <iain@codesourcery.com>
8972 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
8973 warning. Amend comment to reflect current functionality.
8975 2014-01-20 Richard Biener <rguenther@suse.de>
8978 * builtins.c (fold_builtin_strcat): Remove case better handled
8979 by tree-ssa-strlen.c.
8981 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
8983 * config/aarch64/aarch64.opt
8984 (mcpu, march, mtune): Make case-insensitive.
8986 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8989 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8990 if operands[1] is a REG or ZERO_EXTEND of a REG.
8992 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
8994 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
8996 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
8998 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
8999 long non-pic millicode calls.
9001 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9003 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
9005 2014-01-19 Kito Cheng <kito@0xlab.org>
9007 * builtins.c (expand_movstr): Check movstr expand done or fail.
9009 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9010 H.J. Lu <hongjiu.lu@intel.com>
9013 * config/i386/i386.md (*lea<mode>): Zero-extend return register
9014 to DImode for zero-extended addresses.
9016 2014-01-19 Jakub Jelinek <jakub@redhat.com>
9018 PR rtl-optimization/57763
9019 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
9020 on the new indirect jump_insn and increment LABEL_NUSES (label).
9022 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
9026 * config.gcc (x86_archs): New variable.
9027 (x86_64_archs): Likewise.
9028 (x86_cpus): Likewise.
9029 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
9030 --with-arch/--with-cpu= options.
9031 Support --with-arch=/--with-cpu={nehalem,westmere,
9032 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
9034 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9036 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
9037 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
9039 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9041 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
9043 2014-01-18 Jakub Jelinek <jakub@redhat.com>
9046 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
9047 clear cpp_get_options (parse_in)->warn_unused_macros for
9048 ix86_target_macros_internal with cpp_define.
9050 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
9052 * jump.c (delete_related_insns): Keep (use (insn))s.
9053 * reorg.c (redundant_insn): Check for barriers too.
9055 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9057 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
9059 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
9061 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
9062 call to $$dyncall when TARGET_LONG_CALLS is true.
9064 2014-01-17 Jeff Law <law@redhat.com>
9066 * ree.c (combine_set_extension): Temporarily disable test for
9067 changing number of hard registers.
9069 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9072 * ipa-inline-analysis.c (inline_free_summary):
9073 Do not free summary of aliases.
9075 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9078 * gimplify.c (gimplify_expr): Use create_tmp_var
9079 instead of create_tmp_var_raw. If cond doesn't have
9080 integral type, don't add the IFN_ANNOTATE builtin at all.
9082 2014-01-17 Martin Jambor <mjambor@suse.cz>
9085 * ipa-cp.c (prev_edge_clone): New variable.
9086 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
9087 Also resize prev_edge_clone vector.
9088 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
9089 (ipcp_edge_removal_hook): New function.
9090 (ipcp_driver): Register ipcp_edge_removal_hook.
9092 2014-01-17 Andrew Pinski <apinski@cavium.com>
9093 Steve Ellcey <sellcey@mips.com>
9096 * config/mips/mips.c (mips_print_operand): Check operand mode instead
9099 2014-01-17 Jeff Law <law@redhat.com>
9102 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
9103 so that pass_ccp runs first.
9105 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9107 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
9108 (ix86_adjust_cost): Use !TARGET_XXX.
9109 (do_reorder_for_imul): Likewise.
9110 (swap_top_of_ready_list): Likewise.
9111 (ix86_sched_reorder): Likewise.
9113 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9115 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9116 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
9117 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
9118 (intel_memset): New. Duplicate slm_memset.
9119 (intel_cost): New. Duplicate slm_cost.
9120 (m_INTEL): New macro.
9121 (processor_target_table): Add "intel".
9122 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
9123 with PROCESSOR_INTEL for "intel".
9124 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
9125 PROCESSOR_SILVERMONT.
9126 (ix86_issue_rate): Likewise.
9127 (ix86_adjust_cost): Likewise.
9128 (ia32_multipass_dfa_lookahead): Likewise.
9129 (swap_top_of_ready_list): Likewise.
9130 (ix86_sched_reorder): Likewise.
9131 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
9132 instead of TARGET_OPT_AGU.
9133 * config/i386/i386.h (TARGET_INTEL): New.
9134 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
9135 (processor_type): Add PROCESSOR_INTEL.
9136 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
9137 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
9139 2014-01-17 Marek Polacek <polacek@redhat.com>
9142 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
9145 2014-01-17 Richard Biener <rguenther@suse.de>
9147 PR tree-optimization/46590
9148 * opts.c (default_options_table): Add entries for
9149 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
9150 all enabled at -O1 but not for -Og.
9151 * common.opt (fbranch-count-reg): Remove Init(1).
9152 (fmove-loop-invariants): Likewise.
9153 (ftree-pta): Likewise.
9155 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9157 * config/i386/i386.c (ix86_data_alignment): For compatibility with
9158 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
9159 decls to at least the GCC 4.8 used alignments.
9162 * tree-nested.c (convert_nonlocal_reference_stmt,
9163 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
9164 of GIMPLE_BIND stmts, adjust associated decls.
9166 2014-01-17 Richard Biener <rguenther@suse.de>
9168 PR tree-optimization/46590
9169 * vec.h (vec<>::bseach): New member function implementing
9170 binary search according to C89 bsearch.
9171 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
9172 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
9173 bitmap pointer again. Make accesses_in_loop a flat array.
9174 (mem_ref_obstack): New global.
9175 (outermost_indep_loop): Adjust for mem_ref->stored changes.
9176 (mark_ref_stored): Likewise.
9177 (ref_indep_loop_p_2): Likewise.
9178 (set_ref_stored_in_loop): New helper function.
9179 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
9180 (memref_free): Adjust.
9181 (record_mem_ref_loc): Simplify.
9182 (gather_mem_refs_stmt): Adjust.
9183 (sort_locs_in_loop_postorder_cmp): New function.
9184 (analyze_memory_references): Sort accesses_in_loop after
9185 loop postorder number.
9186 (find_ref_loc_in_loop_cmp): New function.
9187 (for_all_locs_in_loop): Find relevant cluster of locs in
9188 accesses_in_loop and iterate without recursion.
9189 (execute_sm): Avoid uninit warning.
9190 (struct ref_always_accessed): Simplify.
9191 (ref_always_accessed::operator ()): Likewise.
9192 (ref_always_accessed_p): Likewise.
9193 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
9194 loop postorder numbers here.
9195 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
9198 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9201 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
9202 on decls for which assemble_alias has been called.
9204 2014-01-17 Nick Clifton <nickc@redhat.com>
9206 * config/msp430/msp430.opt: (mcpu): New option.
9207 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
9208 (msp430_option_override): Parse target_cpu. If the MCU name
9209 matches a generic string, clear target_mcu.
9210 (msp430_attr): Allow numeric interrupt values up to 63.
9211 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
9212 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
9214 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
9216 * config/msp430/msp430.md (popm): Use %J rather than %I.
9217 (addsi3): Use msp430_nonimmediate_operand for operand 2.
9218 (addhi_cy_i): Use immediate_operand for operand 2.
9219 * doc/invoke.texi: Document -mcpu option.
9221 2014-01-17 Richard Biener <rguenther@suse.de>
9223 PR rtl-optimization/38518
9224 * df.h (df_analyze_loop): Declare.
9225 * df-core.c: Include cfgloop.h.
9226 (df_analyze_1): Split out main part of df_analyze.
9227 (df_analyze): Adjust.
9228 (loop_inverted_post_order_compute): New function.
9229 (loop_post_order_compute): Likewise.
9230 (df_analyze_loop): New function avoiding whole-function
9232 * loop-invariant.c (find_defs): Use df_analyze_loop.
9233 (find_invariants): Adjust.
9234 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
9236 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
9238 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
9239 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
9241 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
9243 * ipa-ref.c (ipa_remove_stmt_references): Fix references
9244 traversal when removing references.
9246 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
9249 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
9251 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
9254 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
9255 pushing a reload for an autoinc when we had previously reloaded an
9256 inner part of the address.
9258 2014-01-16 Jakub Jelinek <jakub@redhat.com>
9260 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
9262 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
9263 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
9264 when not giving up or versioning for alias only because of
9266 (vect_analyze_data_ref_dependences): Set to true.
9267 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
9269 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
9270 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
9274 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
9277 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
9278 operand 0 predicate for gathers, use a new pseudo as subtarget.
9280 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9283 * lra-constraints.c (process_alt_operands): Add printing debug info.
9284 Check absence of input/output reloads for matched operands too.
9286 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9288 PR rtl-optimization/59835
9289 * ira.c (ira_init_register_move_cost): Increase cost for
9292 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
9294 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
9296 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
9299 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
9300 non-register objects. Use gen_(high/low)part more consistently.
9303 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9306 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
9307 endian support, remove tests for WORDS_BIG_ENDIAN.
9308 (p8_mfvsrd_3_<mode>): Likewise.
9309 (reload_gpr_from_vsx<mode>): Likewise.
9310 (reload_gpr_from_vsxsf): Likewise.
9311 (p8_mfvsrd_4_disf): Likewise.
9313 2014-01-16 Richard Biener <rguenther@suse.de>
9315 PR rtl-optimization/46590
9316 * lcm.c (compute_antinout_edge): Use postorder iteration.
9317 (compute_laterin): Use inverted postorder iteration.
9319 2014-01-16 Nick Clifton <nickc@redhat.com>
9322 * varasm.c (output_constant): Return the number of bytes actually
9324 (output_constructor_array_range): Update the field size with the
9325 number of bytes emitted by output_constant.
9326 (output_constructor_regular_field): Likewise. Also do not
9327 complain if the total number of bytes emitted is now greater
9328 than the expected fieldpos.
9329 * output.h (output_constant): Update prototype and descriptive comment.
9331 2014-01-16 Marek Polacek <polacek@redhat.com>
9334 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
9335 it is error_mark_node.
9337 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
9339 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
9340 VALID_AVX256_REG_OR_OI_MODE.
9342 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
9344 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
9345 current procedure should be profiled.
9347 2014-01-15 Andrew Pinski <apinski@cavium.com>
9349 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
9350 of moving from/to the STACK_REG register class.
9352 2014-01-15 Richard Henderson <rth@redhat.com>
9355 * reginfo.c (global_regs_decl): Globalize.
9356 * rtl.h (global_regs_decl): Declare.
9357 * ira.c (do_reload): Diagnose frame_pointer_needed and it
9358 reserved via global_regs.
9360 2014-01-15 Teresa Johnson <tejohnson@google.com>
9362 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
9364 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9366 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
9367 and vmulosh rather than call gen_vec_widen_smult_*.
9368 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
9369 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
9370 (vec_widen_smult_even_v16qi): Likewise.
9371 (vec_widen_umult_even_v8hi): Likewise.
9372 (vec_widen_smult_even_v8hi): Likewise.
9373 (vec_widen_umult_odd_v16qi): Likewise.
9374 (vec_widen_smult_odd_v16qi): Likewise.
9375 (vec_widen_umult_odd_v8hi): Likewise.
9376 (vec_widen_smult_odd_v8hi): Likewise.
9377 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
9378 vmuloub rather than call gen_vec_widen_umult_*.
9379 (vec_widen_umult_lo_v16qi): Likewise.
9380 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
9381 vmulosb rather than call gen_vec_widen_smult_*.
9382 (vec_widen_smult_lo_v16qi): Likewise.
9383 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
9384 rather than call gen_vec_widen_umult_*.
9385 (vec_widen_umult_lo_v8hi): Likewise.
9386 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
9387 rather than call gen_vec_widen_smult_*.
9388 (vec_widen_smult_lo_v8hi): Likewise.
9390 2014-01-15 Jeff Law <law@redhat.com>
9392 PR tree-optimization/59747
9393 * ree.c (find_and_remove_re): Properly handle case where a second
9394 eliminated extension requires widening a copy created for elimination
9395 of a prior extension.
9396 (combine_set_extension): Ensure that the number of hard regs needed
9397 for a destination register does not change when we widen it.
9399 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
9401 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
9402 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
9403 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
9404 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
9405 (avr-*-rtems*): Likewise.
9406 (bfin*-rtems*): Likewise.
9407 (moxie-*-rtems*): Likewise.
9408 (h8300-*-rtems*): Likewise.
9409 (i[34567]86-*-rtems*): Likewise.
9410 (lm32-*-rtems*): Likewise.
9411 (m32r-*-rtems*): Likewise.
9412 (m68k-*-rtems*): Likewise.
9413 (microblaze*-*-rtems*): Likewise.
9414 (mips*-*-rtems*): Likewise.
9415 (powerpc-*-rtems*): Likewise.
9416 (sh-*-rtems*): Likewise.
9417 (sparc-*-rtems*): Likewise.
9418 (sparc64-*-rtems*): Likewise.
9419 (v850-*-rtems*): Likewise.
9420 (m32c-*-rtems*): Likewise.
9422 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9424 PR rtl-optimization/59511
9425 * ira.c (ira_init_register_move_cost): Use memory costs for some
9426 cases of register move cost calculations.
9427 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9428 instead of BB frequency.
9429 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9430 * lra-assigns.c (find_hard_regno_for): Ditto.
9432 2014-01-15 Richard Biener <rguenther@suse.de>
9434 PR tree-optimization/59822
9435 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9436 (vectorizable_load): Use it to hoist defs of uses of invariant
9437 loads out of the loop.
9439 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9440 Kugan Vivekanandarajah <kuganv@linaro.org>
9443 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9446 2014-01-15 Richard Biener <rguenther@suse.de>
9448 PR rtl-optimization/59802
9449 * lcm.c (compute_available): Use inverted postorder to seed
9450 the initial worklist.
9452 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9455 * config/s390/s390.c (s390_preferred_reload_class): Don't return
9456 ADDR_REGS for invalid symrefs in non-PIC code.
9458 2014-01-15 Jakub Jelinek <jakub@redhat.com>
9461 * builtins.c (determine_block_size): Initialize *probable_max_size
9462 even if len_rtx is CONST_INT.
9464 2014-01-14 Andrew Pinski <apinski@cavium.com>
9466 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
9467 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
9468 (cortexa53_tunings): Likewise.
9469 (aarch64_sched_issue_rate): New function.
9470 (TARGET_SCHED_ISSUE_RATE): Define.
9472 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9474 * ira-costs.c (find_costs_and_classes): Add missed
9475 ira_init_register_move_cost_if_necessary.
9477 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9480 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
9482 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
9485 * config/i386/i386.c (type_natural_mode): Add a bool parameter
9486 to indicate if type is used for function return value. Warn ABI
9487 change if the vector mode isn't available for function return value.
9488 (ix86_function_arg_advance): Pass false to type_natural_mode.
9489 (ix86_function_arg): Likewise.
9490 (ix86_gimplify_va_arg): Likewise.
9491 (function_arg_32): Don't warn ABI change.
9492 (ix86_function_value): Pass true to type_natural_mode.
9493 (ix86_return_in_memory): Likewise.
9494 (ix86_struct_value_rtx): Removed.
9495 (TARGET_STRUCT_VALUE_RTX): Likewise.
9497 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9499 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
9500 converting a conditional jump into a conditional return.
9502 2014-01-14 Richard Biener <rguenther@suse.de>
9504 PR tree-optimization/58921
9505 PR tree-optimization/59006
9506 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
9507 hoisting invariant stmts.
9508 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
9509 invariant loads on the preheader edge if possible.
9511 2014-01-14 Joey Ye <joey.ye@arm.com>
9513 * doc/plugin.texi (Building GCC plugins): Update to C++.
9515 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
9517 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
9518 (_mm_rcp28_round_ss): Ditto.
9519 (_mm_rsqrt28_round_sd): Ditto.
9520 (_mm_rsqrt28_round_ss): Ditto.
9521 (_mm_rcp28_sd): Ditto.
9522 (_mm_rcp28_ss): Ditto.
9523 (_mm_rsqrt28_sd): Ditto.
9524 (_mm_rsqrt28_ss): Ditto.
9525 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
9526 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
9527 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
9528 (IX86_BUILTIN_RCP28SD): Ditto.
9529 (IX86_BUILTIN_RCP28SS): Ditto.
9530 (IX86_BUILTIN_RSQRT28SD): Ditto.
9531 (IX86_BUILTIN_RSQRT28SS): Ditto.
9532 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
9533 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
9534 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
9535 (ix86_expand_special_args_builtin): Expand new FTYPE.
9536 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
9537 (srcp14<mode>): Make insn unary.
9538 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
9539 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
9540 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
9541 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
9542 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
9543 Fix rounding: make it SAE only.
9544 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9546 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9548 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
9549 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
9550 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
9551 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
9552 (round_saeonly_mask_scalar_operand4): Ditto.
9553 (round_saeonly_mask_scalar_op3): Ditto.
9554 (round_saeonly_mask_scalar_op4): Ditto.
9556 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9558 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9559 Implement -maltivec=be for vec_insert and vec_extract.
9561 2014-01-10 DJ Delorie <dj@redhat.com>
9563 * config/msp430/msp430.md (call_internal): Don't allow memory
9564 references with SP as the base register.
9565 (call_value_internal): Likewise.
9566 * config/msp430/constraints.md (Yc): New. For memory references
9567 that don't use SP as a base register.
9569 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
9570 "an integer without a # prefix"
9571 * config/msp430/msp430.md (epilogue_helper): Use it.
9573 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9576 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
9577 AVX512F gather builtins.
9578 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
9579 on gather decls with INTEGER_TYPE masktype.
9580 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
9581 directly into the builtin rather than hoisting it before loop.
9583 PR tree-optimization/59387
9584 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
9585 (scev_const_prop): If folded_casts and type has undefined overflow,
9586 use force_gimple_operand instead of force_gimple_operand_gsi and
9587 for each added stmt if it is assign with
9588 arith_code_with_undefined_signed_overflow, call
9589 rewrite_to_defined_overflow.
9590 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
9591 gimple-fold.h instead.
9592 (arith_code_with_undefined_signed_overflow,
9593 rewrite_to_defined_overflow): Moved to ...
9594 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
9595 rewrite_to_defined_overflow): ... here. No longer static.
9596 Include gimplify-me.h.
9597 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
9598 rewrite_to_defined_overflow): New prototypes.
9600 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9602 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
9604 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
9606 * builtins.c (get_object_alignment_2): Minor tweak.
9607 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
9609 2014-01-13 Christian Bruel <christian.bruel@st.com>
9611 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
9612 optimized non constant lengths.
9614 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9617 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
9618 load as __atomic_load_N if possible.
9620 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9622 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
9624 (rs6000_expand_builtin): Adjust call.
9626 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9629 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
9630 * config/rs6000/rs6000.c: Include target-globals.h.
9631 (rs6000_set_current_function): Instead of doing target_reinit
9632 unconditionally, use save_target_globals_default_opts and
9633 restore_target_globals.
9635 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
9637 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
9638 (rs6000_expand_builtin): Handle mffs and mtfsf.
9639 (rs6000_init_builtins): Define mffs and mtfsf.
9640 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
9641 (rs6000_mffs): New pattern.
9642 (rs6000_mtfsf): New pattern.
9644 2014-01-11 Bin Cheng <bin.cheng@arm.com>
9646 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
9647 Start narrowing with START. Apply candidate-use pair
9648 and check overall cost in narrowing.
9649 (iv_ca_prune): Pass new argument.
9651 2014-01-10 Jeff Law <law@redhat.com>
9654 * ree.c (combine_reaching_defs): Ensure the defining statement
9655 occurs before the extension when optimizing extensions with
9656 different source and destination hard registers.
9658 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9661 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
9662 vtables into the type inheritance graph.
9664 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9666 PR rtl-optimization/59754
9667 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
9668 modes in the REGNO != REGNO case.
9670 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9672 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
9674 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9676 PR tree-optimization/59745
9677 * tree-predcom.c (tree_predictive_commoning_loop): Call
9678 free_affine_expand_cache if giving up because components is NULL.
9680 * target-globals.c (save_target_globals): Allocate < 4KB structs using
9681 GC in payload of target_globals struct instead of allocating them on
9682 the heap and the larger structs separately using GC.
9683 * target-globals.h (struct target_globals): Make regs, hard_regs,
9684 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
9685 of GTY((skip)) and change type to void *.
9686 (reset_target_globals): Cast loads from those fields to corresponding
9689 2014-01-10 Steve Ellcey <sellcey@mips.com>
9692 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
9693 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
9694 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
9696 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
9699 * aarch64-modes.def (CC_Zmode): New flags mode.
9700 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
9701 represents an equality.
9702 (aarch64_get_condition_code): Handle CC_Zmode.
9703 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
9705 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9707 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
9708 extraction in good case.
9710 2014-01-10 Richard Biener <rguenther@suse.de>
9712 PR tree-optimization/59374
9713 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
9714 checking after SLP discovery. Mark stmts not participating
9715 in any SLP instance properly.
9717 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9719 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
9720 when handling a SET rtx.
9722 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9724 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
9725 (cortex-a57): Likewise.
9726 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
9728 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9730 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
9731 non-iwmmxt builtins.
9733 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9737 * ipa-devirt.c record_target_from_binfo): Take as argument
9738 stack of binfos and lookup matching one for virtual inheritance.
9739 (possible_polymorphic_call_targets_1): Update.
9741 2014-01-10 Huacai Chen <chenhc@lemote.com>
9743 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
9744 kernel strings for Loongson-2E/2F/3A.
9746 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9749 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
9750 is_gimple_call before calling gimple_call_internal_p.
9752 2014-01-09 Steve Ellcey <sellcey@mips.com>
9754 * Makefile.in (TREE_FLOW_H): Remove.
9755 (TREE_SSA_H): Add file names from tree-flow.h.
9756 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
9757 * tree.h: Remove tree-flow.h reference.
9758 * hash-table.h: Remove tree-flow.h reference.
9759 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
9760 reference with tree-ssa-loop.h.
9762 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9764 * doc/invoke.texi: Add -maltivec={be,le} options, and document
9765 default element-order behavior for -maltivec.
9766 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
9767 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
9768 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
9769 when targeting big endian, at least for now.
9770 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
9772 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9775 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
9776 var satisfies use_register_for_decl, just take into account type
9777 alignment, rather than decl alignment.
9779 PR tree-optimization/59622
9780 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
9781 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
9782 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
9783 Don't devirtualize for inplace at all. For targets.length () == 1,
9784 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
9786 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9788 * config/i386/i386.md (cpu): Remove the unused btver1.
9790 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9792 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
9794 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9797 * tree-core.h (struct target_globals): New forward declaration.
9798 (struct tree_target_option): Add globals field.
9799 * tree.h (TREE_TARGET_GLOBALS): Define.
9800 (prepare_target_option_nodes_for_pch): New prototype.
9801 * target-globals.h (struct target_globals): Define even if
9803 * tree.c (prepare_target_option_node_for_pch,
9804 prepare_target_option_nodes_for_pch): New functions.
9805 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
9806 * config/i386/i386.c: Include target-globals.h.
9807 (ix86_set_current_function): Instead of doing target_reinit
9808 unconditionally, use save_target_globals_default_opts and
9809 restore_target_globals.
9811 2014-01-09 Richard Biener <rguenther@suse.de>
9813 PR tree-optimization/59715
9814 * tree-cfg.h (split_critical_edges): Declare.
9815 * tree-cfg.c (split_critical_edges): Export.
9816 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
9818 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
9820 * cfgexpand.c (expand_stack_vars): Optionally disable
9821 asan stack protection.
9822 (expand_used_vars): Likewise.
9823 (partition_stack_vars): Likewise.
9824 * asan.c (asan_emit_stack_protection): Optionally disable
9825 after return stack usage.
9826 (instrument_derefs): Optionally disable memory access instrumentation.
9827 (instrument_builtin_call): Likewise.
9828 (instrument_strlen_call): Likewise.
9829 (asan_protect_global): Optionally disable global variables protection.
9830 * doc/invoke.texi: Added doc for new options.
9831 * params.def: Added new options.
9832 * params.h: Likewise.
9834 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9836 PR rtl-optimization/59724
9837 * ifcvt.c (cond_exec_process_if_block): Don't call
9838 flow_find_head_matching_sequence with 0 longest_match.
9839 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
9840 non-active insns if !stop_after.
9841 (try_head_merge_bb): Revert 2014-01-07 changes.
9843 2014-01-08 Jeff Law <law@redhat.com>
9845 * ree.c (get_sub_rtx): New function, extracted from...
9846 (merge_def_and_ext): Here.
9847 (combine_reaching_defs): Use get_sub_rtx.
9849 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
9851 * cgraph.h (varpool_variable_node): Do not choke on null node.
9853 2014-01-08 Catherine Moore <clm@codesourcery.com>
9855 * config/mips/mips.md (simple_return): Attempt to use JRC
9857 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
9859 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9861 PR rtl-optimization/59137
9862 * reorg.c (steal_delay_list_from_target): Call update_block for
9864 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
9866 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9868 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
9869 two duplicate entries.
9871 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9874 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
9876 * config/mips/mips.c (mips_truncated_op_cost): New function.
9877 (mips_rtx_costs): Adjust test for BADDU.
9878 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
9880 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
9882 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
9883 (*baddu_si): ...this new pattern.
9885 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9888 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
9890 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9893 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
9895 (expand_expr, expand_normal): Adjust.
9896 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
9897 inner_reference_p. Use inner_reference_p to expand inner references.
9898 (store_expr): Adjust.
9899 * cfgexpand.c (expand_call_stmt): Adjust.
9901 2014-01-08 Rong Xu <xur@google.com>
9903 * gcov-io.c (gcov_var): Move from gcov-io.h.
9904 (gcov_position): Ditto.
9905 (gcov_is_error): Ditto.
9906 (gcov_rewrite): Ditto.
9907 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
9908 only part to libgcc/libgcov.h.
9910 2014-01-08 Marek Polacek <polacek@redhat.com>
9913 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
9915 2014-01-08 Marek Polacek <polacek@redhat.com>
9918 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
9920 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9922 PR rtl-optimization/59649
9923 * stor-layout.c (get_mode_bounds): For BImode return
9924 0 and STORE_FLAG_VALUE.
9926 2014-01-08 Richard Biener <rguenther@suse.de>
9929 * gimple.h (is_gimple_builtin_call): Remove.
9930 (gimple_builtin_call_types_compatible_p): New.
9931 (gimple_call_builtin_p): New overload.
9932 * gimple.c (is_gimple_builtin_call): Remove.
9933 (validate_call): Rename to ...
9934 (gimple_builtin_call_types_compatible_p): ... this and export. Also
9936 (validate_type): New static function.
9937 (gimple_call_builtin_p): New overload and adjust.
9938 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
9939 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
9940 (gimple_fold_stmt_to_constant_1): Likewise.
9941 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
9943 2014-01-08 Richard Biener <rguenther@suse.de>
9946 * gimplify.c (gimplify_expr): Gimplify register-register type
9947 VIEW_CONVERT_EXPRs to separate stmts.
9949 2014-01-07 Jeff Law <law@redhat.com>
9952 * ree.c (combine_set_extension): Handle case where source
9953 and destination registers in an extension insn are different.
9954 (combine_reaching_defs): Allow source and destination registers
9955 in extension to be different under limited circumstances.
9956 (add_removable_extension): Remove restriction that the
9957 source and destination registers in the extension are the same.
9958 (find_and_remove_re): Emit a copy from the extension's
9959 destination to its source after the defining insn if
9960 the source and destination registers are different.
9963 * ifcvt.c (merge_if_block): If we are merging a block with more than
9964 one successor with a block with no successors, remove any BARRIER
9965 after the second block.
9967 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
9969 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
9971 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
9974 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
9975 for 14-bit register offsets when INT14_OK_STRICT is false.
9977 2014-01-07 Roland Stigge <stigge@antcom.de>
9978 Michael Meissner <meissner@linux.vnet.ibm.com>
9981 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
9982 Only check TFmode for SPE constants. Don't check TImode or TDmode.
9984 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
9986 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
9989 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
9991 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
9992 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
9993 rtx is const0_rtx or not.
9995 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
9998 * target-globals.c (save_target_globals): Remove this_fn_optab
10000 * toplev.c: Include optabs.h.
10001 (target_reinit): Temporarily restore the global options if another
10002 set of options are in force.
10004 2014-01-07 Jakub Jelinek <jakub@redhat.com>
10006 PR rtl-optimization/58668
10007 * cfgcleanup.c (flow_find_cross_jump): Don't count
10008 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
10009 to determine what is counted.
10010 (flow_find_head_matching_sequence): Use active_insn_p to determine
10012 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
10014 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
10015 determine what is counted.
10017 PR tree-optimization/59643
10018 * tree-predcom.c (split_data_refs_to_components): If one dr is
10019 read and one write, determine_offset fails and the write isn't
10020 in the bad component, just put the read into the bad component.
10022 2014-01-07 Mike Stump <mikestump@comcast.net>
10023 Jakub Jelinek <jakub@redhat.com>
10026 * tree-core.h (struct tree_optimization_option): Change optabs
10027 type from unsigned char * to void *.
10028 * optabs.c (init_tree_optimization_optabs): Adjust
10029 TREE_OPTIMIZATION_OPTABS initialization.
10031 2014-01-06 Jakub Jelinek <jakub@redhat.com>
10034 * config/i386/i386.h (struct machine_function): Add
10035 no_drap_save_restore field.
10036 * config/i386/i386.c (ix86_save_reg): Use
10037 !cfun->machine->no_drap_save_restore instead of
10038 crtl->stack_realign_needed.
10039 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
10040 this function clears frame_pointer_needed. Set
10041 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
10042 and DRAP reg is needed.
10044 2014-01-06 Marek Polacek <polacek@redhat.com>
10047 * doc/implement-c.texi: Mention that other integer types are
10048 permitted as bit-field types in strictly conforming mode.
10050 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
10052 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
10053 is newly allocated.
10055 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
10057 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
10059 2014-01-06 Martin Jambor <mjambor@suse.cz>
10062 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
10064 * ipa-prop.c (ipa_print_node_params): Fix indentation.
10066 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
10070 * var-tracking.c (add_stores): Preserve the value of the source even if
10071 we don't record the store.
10073 2014-01-06 Terry Guo <terry.guo@arm.com>
10075 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
10077 2014-01-05 Iain Sandoe <iain@codesourcery.com>
10080 * config/darwin.c (darwin_function_section): Adjust return values to
10081 correspond to optimisation changes made in r206070.
10083 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
10085 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
10086 from prefetch_block tune setting.
10087 (nocona_cost): Correct size of prefetch block to 64.
10089 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
10091 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
10092 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
10093 used to save the static chain register in the computation of the offset
10094 from which the FP registers need to be restored.
10096 2014-01-04 Jakub Jelinek <jakub@redhat.com>
10098 PR tree-optimization/59519
10099 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
10100 ICE if get_current_def (current_new_name) is already non-NULL, as long
10101 as it is a phi result of some other phi in *new_exit_bb that has
10104 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
10105 or vmovdqu* for misaligned_operand.
10106 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
10107 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
10108 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
10109 aligned_mem for AVX512F masked aligned load and store builtins and for
10110 non-temporal moves.
10112 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
10114 PR tree-optimization/59651
10115 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
10116 Address range for negative step should be added by TYPE_SIZE_UNIT.
10118 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
10120 * config/m68k/m68k.c (handle_move_double): Handle pushes with
10121 overlapping registers also for registers other than the stack pointer.
10123 2014-01-03 Marek Polacek <polacek@redhat.com>
10126 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
10129 2014-01-03 Jakub Jelinek <jakub@redhat.com>
10132 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
10135 * config/i386/i386.md (MODE_SIZE): New mode attribute.
10136 (push splitter): Use <P:MODE_SIZE> instead of
10137 GET_MODE_SIZE (<P:MODE>mode).
10138 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
10139 (mov -1, reg peephole2): Likewise.
10140 * config/i386/sse.md (*mov<mode>_internal,
10141 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
10142 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
10143 *<code><mode>3, *andnot<mode>3<mask_name>,
10144 <mask_codefor><code><mode>3<mask_name>): Likewise.
10145 * config/i386/subst.md (mask_mode512bit_condition,
10146 sd_mask_mode512bit_condition): Likewise.
10148 2014-01-02 Xinliang David Li <davidxl@google.com>
10150 PR tree-optimization/59303
10151 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
10152 (dump_predicates): Better output format.
10153 (pred_equal_p): New function.
10154 (is_neq_relop_p): Ditto.
10155 (is_neq_zero_form_p): Ditto.
10156 (pred_expr_equal_p): Ditto.
10157 (pred_neg_p): Ditto.
10158 (simplify_pred): Ditto.
10159 (simplify_preds_2): Ditto.
10160 (simplify_preds_3): Ditto.
10161 (simplify_preds_4): Ditto.
10162 (simplify_preds): Ditto.
10163 (push_pred): Ditto.
10164 (push_to_worklist): Ditto.
10165 (get_pred_info_from_cmp): Ditto.
10166 (is_degenerated_phi): Ditto.
10167 (normalize_one_pred_1): Ditto.
10168 (normalize_one_pred): Ditto.
10169 (normalize_one_pred_chain): Ditto.
10170 (normalize_preds): Ditto.
10171 (normalize_cond_1): Remove function.
10172 (normalize_cond): Ditto.
10173 (is_gcond_subset_of): Ditto.
10174 (is_subset_of_any): Ditto.
10175 (is_or_set_subset_of): Ditto.
10176 (is_and_set_subset_of): Ditto.
10177 (is_norm_cond_subset_of): Ditto.
10178 (pred_chain_length_cmp): Ditto.
10179 (convert_control_dep_chain_into_preds): Type change.
10180 (find_predicates): Ditto.
10181 (find_def_preds): Ditto.
10182 (destroy_predicates_vecs): Ditto.
10183 (find_matching_predicates_in_rest_chains): Ditto.
10184 (use_pred_not_overlap_with_undef_path_pred): Ditto.
10185 (is_pred_expr_subset): Ditto.
10186 (is_pred_chain_subset_of): Ditto.
10187 (is_included_in): Ditto.
10188 (is_superset_of): Ditto.
10190 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10192 Update copyright years.
10194 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10196 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
10197 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
10198 config/arc/arc.md, config/arc/arc.opt,
10199 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
10200 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
10201 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
10202 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
10203 config/linux-protos.h, config/linux.c, config/winnt-c.c,
10204 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
10205 vtable-verify.c, vtable-verify.h: Use the standard form for the
10208 2014-01-02 Tobias Burnus <burnus@net-b.de>
10210 * gcc.c (process_command): Update copyright notice dates.
10211 * gcov-dump.c: Ditto.
10213 * doc/cpp.texi: Bump @copying's copyright year.
10214 * doc/cppinternals.texi: Ditto.
10215 * doc/gcc.texi: Ditto.
10216 * doc/gccint.texi: Ditto.
10217 * doc/gcov.texi: Ditto.
10218 * doc/install.texi: Ditto.
10219 * doc/invoke.texi: Ditto.
10221 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10223 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
10225 2014-01-01 Jakub Jelinek <jakub@redhat.com>
10227 * config/i386/sse.md (*mov<mode>_internal): Guard
10228 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
10230 PR rtl-optimization/59647
10231 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
10232 new_rtx into UNSIGNED_FLOAT rtxes.
10234 Copyright (C) 2014 Free Software Foundation, Inc.
10236 Copying and distribution of this file, with or without modification,
10237 are permitted in any medium without royalty provided the copyright
10238 notice and this notice are preserved.