1 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
2 Matthias Klose <doko@ubuntu.com>
5 * optc-gen.awk: Fix option handling for -Wunused-parameter.
7 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
10 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
12 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
15 * config/alpha/alpha.c: Include gimple-iterator.h.
16 (alpha_gimple_fold_builtin): New function. Move
17 ALPHA_BUILTIN_UMULH folding from ...
18 (alpha_fold_builtin): ... here.
19 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
21 2014-05-08 Wei Mi <wmi@google.com>
24 * config/i386/i386.c (ix86_compute_frame_layout): Update
25 preferred_stack_boundary for call, expanded from tls descriptor.
26 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
27 to depend on SP register.
28 (*tls_local_dynamic_base_32_gnu): Ditto.
29 (*tls_local_dynamic_32_once): Ditto.
30 (tls_global_dynamic_64_<mode>): Set
31 ix86_tls_descriptor_calls_expanded_in_cfun.
32 (tls_local_dynamic_base_64_<mode>): Ditto.
33 (tls_global_dynamic_32): Set
34 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
35 to depend on SP register.
36 (tls_local_dynamic_base_32): Ditto.
38 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
40 * config/arm/arm_neon.h: Update comment.
41 * config/arm/neon-docgen.ml: Delete.
42 * config/arm/neon-gen.ml: Delete.
43 * doc/arm-neon-intrinsics.texi: Update comment.
45 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
47 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
49 (vand, vorr, veor, vorn, vbic): Remove.
50 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
52 (neon_vsub_unspec): Likewise.
53 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
55 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
57 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
67 (vaddq_s16): Likewise.
68 (vaddq_s32): Likewise.
69 (vaddq_s64): Likewise.
70 (vaddq_f32): Likewise.
72 (vaddq_u16): Likewise.
73 (vaddq_u32): Likewise.
74 (vaddq_u64): Likewise.
84 (vmulq_s16): Likewise.
85 (vmulq_s32): Likewise.
86 (vmulq_f32): Likewise.
88 (vmulq_u16): Likewise.
89 (vmulq_u32): Likewise.
100 (vsubq_s16): Likewise.
101 (vsubq_s32): Likewise.
102 (vsubq_s64): Likewise.
103 (vsubq_f32): Likewise.
104 (vsubq_u8): Likewise.
105 (vsubq_u16): Likewise.
106 (vsubq_u32): Likewise.
107 (vsubq_u64): Likewise.
109 (vand_s16): Likewise.
110 (vand_s32): Likewise.
112 (vand_u16): Likewise.
113 (vand_u32): Likewise.
114 (vand_s64): Likewise.
115 (vand_u64): Likewise.
116 (vandq_s8): Likewise.
117 (vandq_s16): Likewise.
118 (vandq_s32): Likewise.
119 (vandq_s64): Likewise.
120 (vandq_u8): Likewise.
121 (vandq_u16): Likewise.
122 (vandq_u32): Likewise.
123 (vandq_u64): Likewise.
125 (vorr_s16): Likewise.
126 (vorr_s32): Likewise.
128 (vorr_u16): Likewise.
129 (vorr_u32): Likewise.
130 (vorr_s64): Likewise.
131 (vorr_u64): Likewise.
132 (vorrq_s8): Likewise.
133 (vorrq_s16): Likewise.
134 (vorrq_s32): Likewise.
135 (vorrq_s64): Likewise.
136 (vorrq_u8): Likewise.
137 (vorrq_u16): Likewise.
138 (vorrq_u32): Likewise.
139 (vorrq_u64): Likewise.
141 (veor_s16): Likewise.
142 (veor_s32): Likewise.
144 (veor_u16): Likewise.
145 (veor_u32): Likewise.
146 (veor_s64): Likewise.
147 (veor_u64): Likewise.
148 (veorq_s8): Likewise.
149 (veorq_s16): Likewise.
150 (veorq_s32): Likewise.
151 (veorq_s64): Likewise.
152 (veorq_u8): Likewise.
153 (veorq_u16): Likewise.
154 (veorq_u32): Likewise.
155 (veorq_u64): Likewise.
157 (vbic_s16): Likewise.
158 (vbic_s32): Likewise.
160 (vbic_u16): Likewise.
161 (vbic_u32): Likewise.
162 (vbic_s64): Likewise.
163 (vbic_u64): Likewise.
164 (vbicq_s8): Likewise.
165 (vbicq_s16): Likewise.
166 (vbicq_s32): Likewise.
167 (vbicq_s64): Likewise.
168 (vbicq_u8): Likewise.
169 (vbicq_u16): Likewise.
170 (vbicq_u32): Likewise.
171 (vbicq_u64): Likewise.
173 (vorn_s16): Likewise.
174 (vorn_s32): Likewise.
176 (vorn_u16): Likewise.
177 (vorn_u32): Likewise.
178 (vorn_s64): Likewise.
179 (vorn_u64): Likewise.
180 (vornq_s8): Likewise.
181 (vornq_s16): Likewise.
182 (vornq_s32): Likewise.
183 (vornq_s64): Likewise.
184 (vornq_u8): Likewise.
185 (vornq_u16): Likewise.
186 (vornq_u32): Likewise.
187 (vornq_u64): Likewise.
189 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
191 * wide-int.cc (UTItype): Define.
192 (UDWtype): Define for appropriate W_TYPE_SIZE.
194 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
196 PR tree-optimization/59100
197 * tree-ssa-phiopt.c: Include tree-inline.h.
198 (neutral_element_p, absorbing_element_p): New functions.
199 (value_replacement): Handle conditional binary operations with a
200 neutral or absorbing element.
202 2014-05-08 Richard Biener <rguenther@suse.de>
204 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
205 folding the expression.
206 (valueize_expr): Remove.
207 (visit_reference_op_load): Do not valueize the result of
209 (simplify_binary_expression): Likewise.
210 (simplify_unary_expression): Likewise.
212 2014-05-08 Richard Biener <rguenther@suse.de>
214 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
215 looking at TYPE_ARG_TYPES.
217 2014-05-08 Richard Biener <rguenther@suse.de>
219 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
220 pointer propagation special-case.
222 2014-05-08 Bin Cheng <bin.cheng@arm.com>
224 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
225 core part of address expressions.
227 2014-05-08 Alan Modra <amodra@gmail.com>
230 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
231 loads and stores when -mno-strict-align at any alignment.
232 (expand_block_clear): Similarly. Also correct calculation of
235 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
238 * tree-complex.c (expand_complex_move): Keep line info when expanding
240 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
241 of complex expression. Use new argument to display correct location
242 for values coming from phi statement.
243 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
244 (warn_uninitialized_phi): Pass location of phi argument to
246 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
247 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
249 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
251 * config/rs6000/predicates.md (indexed_address_mem): New.
252 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
253 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
254 fpstore_ux, fpstore_u.
255 (sign_extend, indexed, update): New.
256 (cell_micro): Adjust.
257 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
258 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
259 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
260 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
261 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
262 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
263 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
264 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
265 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
266 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
267 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
268 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
269 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
270 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
271 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
273 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
274 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
275 *vsx_extract_<mode>_store): Adjust.
276 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
277 is_cracked_insn, insn_must_be_first_in_group,
278 insn_must_be_last_in_group): Adjust.
280 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
282 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
283 ppc440-fpstore): Adjust.
284 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
285 ppc476-fpstore): Adjust.
286 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
287 ppc601-fpstore): Adjust.
288 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
290 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
292 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
293 ppc7450-fpstore): Adjust.
294 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
295 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
296 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
298 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
299 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
300 cell-fpstore, cell-fpstore-update): Adjust.
301 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
302 ppce300c3_store, ppce300c3_fpstore): Adjust.
303 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
304 e500mc_fpstore): Adjust.
305 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
306 e500mc64_store, e500mc64_fpstore): Adjust.
307 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
308 e5500_fpstore): Adjust.
309 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
310 e6500_fpstore): Adjust.
311 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
313 * config/rs6000/power4.md (power4-load, power4-load-ext,
314 power4-load-ext-update, power4-load-ext-update-indexed,
315 power4-load-update-indexed, power4-load-update, power4-fpload,
316 power4-fpload-update, power4-store, power4-store-update,
317 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
319 * config/rs6000/power5.md (power5-load, power5-load-ext,
320 power5-load-ext-update, power5-load-ext-update-indexed,
321 power5-load-update-indexed, power5-load-update, power5-fpload,
322 power5-fpload-update, power5-store, power5-store-update,
323 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
325 * config/rs6000/power6.md (power6-load, power6-load-ext,
326 power6-load-update, power6-load-update-indexed,
327 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
328 power6-fpload-update, power6-store, power6-store-update,
329 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
331 * config/rs6000/power7.md (power7-load, power7-load-ext,
332 power7-load-update, power7-load-update-indexed,
333 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
334 power7-fpload-update, power7-store, power7-store-update,
335 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
337 * config/rs6000/power8.md (power8-load, power8-load-update,
338 power8-load-ext, power8-load-ext-update, power8-fpload,
339 power8-fpload-update, power8-store, power8-store-update-indexed,
340 power8-fpstore, power8-fpstore-update): Adjust.
341 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
343 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
344 titan_lsu_store, titan_lsu_fpstore): Adjust.
345 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
347 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
350 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
351 unrolled byte insns. Emit address increments after move insns.
353 2014-05-07 David Malcolm <dmalcolm@redhat.com>
355 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
356 const_gimple, rather than a gimple.
357 (gimple_call_builtin_p): Likewise, for the three variants.
359 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
360 (gimple_call_builtin_p): Likewise, for the three variants.
362 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
364 PR tree-optimization/61095
365 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
367 2014-05-07 Richard Biener <rguenther@suse.de>
369 PR tree-optimization/61034
370 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
371 (maybe_skip_until): Use translate to take into account
372 lattices when trying to do disambiguations.
373 (get_continuation_for_phi_1): Likewise.
374 (get_continuation_for_phi): Adjust for added translate arguments.
375 (walk_non_aliased_vuses): Likewise.
376 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
377 (walk_non_aliased_vuses): Likewise.
378 (call_may_clobber_ref_p_1): Declare.
379 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
380 calls. Stop early if we are only supposed to disambiguate.
381 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
383 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
385 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
386 Emit an error when the function has arguments.
388 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
390 * cfgloop.h (unswitch_loops): Remove.
391 * doc/passes.texi: Remove references to loop-unswitch.c
392 * timevar.def (TV_LOOP_UNSWITCH): Remove.
394 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
396 * tree-vect-data-refs.c (vect_grouped_load_supported): New
397 check for loads group of length 3.
398 (vect_permute_load_chain): New permutations for loads group of
400 * tree-vect-stmts.c (vect_model_load_cost): Change cost
401 of vec_perm_shuffle for the new permutations.
403 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
405 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
406 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
407 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
408 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
409 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
410 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
411 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
412 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
414 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
416 * loop-unswitch.c: Delete.
418 2014-05-07 Richard Biener <rguenther@suse.de>
420 * config.gcc: Always set need_64bit_hwint to yes.
422 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
424 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
425 of using optimize_size.
427 2014-05-06 Mike Stump <mikestump@comcast.net>
429 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
431 2014-05-06 Joseph Myers <joseph@codesourcery.com>
433 * config/i386/sse.md (*mov<mode>_internal)
434 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
435 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
436 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
437 (*<code><mode>3, *andnot<mode>3<mask_name>)
438 (<mask_codefor><code><mode>3<mask_name>): Only consider
439 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
441 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
444 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
446 * lra-constraints.c (valid_address_p): Move earlier in file.
447 Add a constraint argument to the address_info version.
448 (satisfies_memory_constraint_p): New function.
449 (satisfies_address_constraint_p): Likewise.
450 (process_alt_operands, curr_insn_transform): Use them.
451 (process_address): Pass the constraint to valid_address_p when
452 checking address operands.
454 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
456 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
457 to their respective blocks. Fix inadvertent use of "node".
459 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
461 * emit-rtl.c (init_derived_machine_modes): New functionm, split
463 (init_emit_once): ...here.
464 * rtl.h (init_derived_machine_modes): Declare.
465 * toplev.c (do_compile): Call it even if no_backend.
467 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
468 Mike Stump <mikestump@comcast.net>
469 Richard Sandiford <rdsandiford@googlemail.com>
470 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
472 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
473 (rtx_equal_for_memref_p): Update comment.
474 (adjust_offset_for_component_ref): Use wide-int interfaces.
475 * builtins.c (get_object_alignment_2): Likewise.
476 (c_readstr): Likewise.
477 (target_char_cast): Add comment.
478 (determine_block_size): Use wide-int interfaces.
479 (expand_builtin_signbit): Likewise.
480 (fold_builtin_int_roundingfn): Likewise.
481 (fold_builtin_bitop): Likewise.
482 (fold_builtin_bswap): Likewise.
483 (fold_builtin_logarithm): Use signop.
484 (fold_builtin_pow): Likewise.
485 (fold_builtin_memory_op): Use wide-int interfaces.
486 (fold_builtin_object_size): Likewise.
487 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
488 nb_iterations_estimate.
489 (record_niter_bound): Use wide-int interfaces.
490 (get_estimated_loop_iterations_int): Likewise.
491 (get_estimated_loop_iterations): Likewise.
492 (get_max_loop_iterations): Likewise.
493 * cfgloop.h: Include wide-int.h.
494 (struct nb_iter_bound): Change bound to widest_int.
495 (struct loop): Change nb_iterations_upper_bound and
496 nb_iterations_estimate to widest_int.
497 (record_niter_bound): Switch to use widest_int.
498 (get_estimated_loop_iterations): Likewise.
499 (get_max_loop_iterations): Likewise.
500 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
502 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
503 * combine.c (try_combine): Likewise.
504 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
505 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
507 (aarch64_float_const_representable_p): Likewise.
508 * config/arc/arc.c: Include wide-int.h.
509 (arc_can_use_doloop_p): Use wide-int interfaces.
510 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
511 (vfp3_const_double_index): Likewise.
512 * config/avr/avr.c (avr_out_round): Likewise.
513 (avr_fold_builtin): Likewise.
514 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
515 (bfin_can_use_doloop_p): Likewise.
516 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
517 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
518 * config/i386/i386.c: Include wide-int.h.
519 (ix86_data_alignment): Use wide-int interfaces.
520 (ix86_local_alignment): Likewise.
521 (ix86_emit_swsqrtsf): Update real_from_integer.
522 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
523 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
524 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
525 (zero_constant): Likewise.
526 (input_operand): Likewise.
527 (splat_input_operand): Likewise.
528 (non_logical_cint_operand): Change const_double to const_wide_int.
529 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
530 (easy_altivec_constant): Remove comment.
531 (paired_expand_vector_init): Use CONSTANT_P.
532 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
533 (rs6000_emit_move): Update checks.
534 (rs6000_aggregate_candidate): Use wide-int interfaces.
535 (rs6000_expand_ternop_builtin): Likewise.
536 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
537 (rs6000_assemble_integer): Likewise.
538 (rs6000_hash_constant): Likewise.
539 (output_toc): Likewise.
540 (rs6000_rtx_costs): Likewise.
541 (rs6000_emit_swrsqrt); Update call to real_from_integer.
542 * config/rs6000/rs6000-c.c: Include wide-int.h.
543 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
544 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
545 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
546 Handle CONST_WIDE_INT.
547 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
548 Use tree_fits_uhwi_p.
549 * config/sparc/sparc.c: Include wide-int.h.
550 (sparc_fold_builtin): Use wide-int interfaces.
551 * config/vax/vax.c: Include wide-int.h.
552 (vax_float_literal): Use real_from_integer.
553 * coretypes.h (struct hwivec_def): New.
556 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
557 (equiv_constant): Handle CONST_WIDE_INT.
558 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
559 (cselib_hash_rtx): Handle CONST_WIDE_INT.
560 * dbxout.c (stabstr_U): Use wide-int interfaces.
561 (dbxout_type): Update to use cst_fits_shwi_p.
562 * defaults.h (LOG2_BITS_PER_UNIT): Define.
563 (TARGET_SUPPORTS_WIDE_INT): Add default.
564 * dfp.c: Include wide-int.h.
565 (decimal_real_to_integer2): Use wide-int interfaces and rename to
566 decimal_real_to_integer.
567 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
568 decimal_real_to_integer.
569 * doc/generic.texi (Constant expressions): Update for wide_int.
570 * doc/rtl.texi (const_double): Likewise.
571 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
572 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
573 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
574 (REAL_VALUE_FROM_INT): Remove.
575 (TARGET_SUPPORTS_WIDE_INT): New.
576 * doc/tm.texi: Regenerate.
577 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
578 * double-int.h: Include wide-int.h.
579 (struct wi::int_traits): New.
580 * dwarf2out.c (get_full_len): New.
581 (dw_val_equal_p): Add case dw_val_class_wide_int.
582 (size_of_loc_descr): Likewise.
583 (output_loc_operands): Likewise.
584 (insert_double): Remove.
585 (insert_wide_int): New.
587 (print_die): Add case dw_val_class_wide_int.
588 (attr_checksum): Likewise.
589 (attr_checksum_ordered): Likewise.
590 (same_dw_val_p): Likewise.
591 (size_of_die): Likewise.
592 (value_format): Likewise.
593 (output_die): Likewise.
594 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
596 (clz_loc_descriptor): Use wide-int interfaces.
597 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
598 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
599 (round_up_to_align): Use wide-int interfaces.
600 (field_byte_offset): Likewise.
601 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
602 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
603 CONST_DOUBLE handling. Use wide-int interfaces.
604 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
605 (gen_enumeration_type_die): Use add_AT_wide.
606 (hash_loc_operands): Add case dw_val_class_wide_int.
607 (compare_loc_operands): Likewise.
608 * dwarf2out.h: Include wide-int.h.
610 (enum dw_val_class): Add dw_val_class_wide_int.
611 (struct dw_val_struct): Add val_wide.
612 * emit-rtl.c (const_wide_int_htab): New.
613 (const_wide_int_htab_hash): New.
614 (const_wide_int_htab_eq): New.
615 (lookup_const_wide_int): New.
616 (const_double_htab_hash): Use wide-int interfaces.
617 (const_double_htab_eq): Likewise.
618 (rtx_to_double_int): Conditionally compile for wide-int.
619 (immed_double_int_const): Rename to immed_wide_int_const and
621 (immed_double_const): Conditionally compile for wide-int.
622 (init_emit_once): Use wide-int interfaces.
623 * explow.c (plus_constant): Likewise.
624 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
625 (lshift_value): Use wide-int interfaces.
626 (expand_mult): Likewise.
627 (choose_multiplier): Likewise.
628 (expand_smod_pow2): Likewise.
629 (make_tree): Likewise.
630 * expr.c (convert_modes): Consolidate handling of constants.
631 Use wide-int interfaces.
632 (emit_group_load_1): Add note.
633 (store_expr): Update comment.
634 (get_inner_reference): Use wide-int interfaces.
635 (expand_constructor): Update comment.
636 (expand_expr_real_2): Use wide-int interfaces.
637 (expand_expr_real_1): Likewise.
638 (reduce_to_bit_field_precision): Likewise.
639 (const_vector_from_tree): Likewise.
640 * final.c: Include wide-int-print.h.
641 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
642 * fixed-value.c: Include wide-int.h.
643 (fixed_from_string): Use wide-int interfaces.
644 (fixed_to_decimal): Likewise.
645 (fixed_convert_from_real): Likewise.
646 (real_convert_from_fixed): Likewise.
647 * fold-const.h (mem_ref_offset): Return an offset_int.
648 (div_if_zero_remainder): Remove code parameter.
649 * fold-const.c (div_if_zero_remainder): Remove code parameter.
650 Use wide-int interfaces.
651 (may_negate_without_overflow_p): Use wide-int interfaces.
652 (negate_expr_p): Likewise.
653 (fold_negate_expr): Likewise.
654 (int_const_binop_1): Likewise.
655 (const_binop): Likewise.
656 (fold_convert_const_int_from_int): Likewise.
657 (fold_convert_const_int_from_real): Likewise.
658 (fold_convert_const_int_from_fixed): Likewise.
659 (fold_convert_const_fixed_from_int): Likewise.
660 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
661 (sign_bit_p): Use wide-int interfaces.
662 (make_range_step): Likewise.
663 (build_range_check): Likewise. Pass an integer of the correct type
664 instead of using integer_one_node.
665 (range_predecessor): Pass an integer of the correct type instead
666 of using integer_one_node.
667 (range_successor): Likewise.
668 (merge_ranges): Likewise.
669 (unextend): Use wide-int interfaces.
670 (extract_muldiv_1): Likewise.
671 (fold_div_compare): Likewise.
672 (fold_single_bit_test): Likewise.
673 (fold_sign_changed_comparison): Likewise.
674 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
675 (fold_plusminus_mult_expr): Use wide-int interfaces.
676 (native_encode_int): Likewise.
677 (native_interpret_int): Likewise.
678 (fold_unary_loc): Likewise.
679 (pointer_may_wrap_p): Likewise.
680 (size_low_cst): Likewise.
681 (mask_with_tz): Likewise.
682 (fold_binary_loc): Likewise.
683 (fold_ternary_loc): Likewise.
684 (multiple_of_p): Likewise.
685 (tree_call_nonnegative_warnv_p): Update calls to
686 tree_int_cst_min_precision and real_from_integer.
687 (fold_negate_const): Use wide-int interfaces.
688 (fold_abs_const): Likewise.
689 (fold_relational_const): Use tree_int_cst_lt.
690 (round_up_loc): Use wide-int interfaces.
691 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
692 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
693 * gengtype.c: Remove include of double-int.h.
694 (do_typedef): Use wide-int interfaces.
695 (open_base_files): Add wide-int.h.
696 (main): Add offset_int and widest_int typedefs.
697 * gengtype-lex.l: Handle "^".
698 (CXX_KEYWORD): Add "static".
699 * gengtype-parse.c (require3): New.
700 (require_template_declaration): Handle constant template arguments
701 and nested templates.
702 * gengtype-state.c: Don't include "double-int.h".
703 * genpreds.c (write_one_predicate_function): Update comment.
704 (write_tm_constrs_h): Add check for hval and lval use in
706 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
707 (add_to_sequence): Likewise.
708 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
709 and const_double_operand.
710 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
712 * gimple-fold.c (get_base_constructor): Likewise.
713 (fold_array_ctor_reference): Likewise.
714 (fold_nonarray_ctor_reference): Likewise.
715 (fold_const_aggregate_ref_1): Likewise.
716 (gimple_val_nonnegative_real_p): Likewise.
717 (gimple_fold_indirect_ref): Likewise.
718 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
719 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
720 (struct slsr_cand_d): Change index to be widest_int.
721 (struct incr_info_d): Change incr to be widest_int.
722 (alloc_cand_and_find_basis): Use wide-int interfaces.
723 (slsr_process_phi): Likewise.
724 (backtrace_base_for_ref): Likewise. Return a widest_int.
725 (restructure_reference): Take a widest_int instead of a double_int.
726 (slsr_process_ref): Use wide-int interfaces.
727 (create_mul_ssa_cand): Likewise.
728 (create_mul_imm_cand): Likewise.
729 (create_add_ssa_cand): Likewise.
730 (create_add_imm_cand): Take a widest_int instead of a double_int.
731 (slsr_process_add): Use wide-int interfaces.
732 (slsr_process_cast): Likewise.
733 (slsr_process_copy): Likewise.
734 (dump_candidate): Likewise.
735 (dump_incr_vec): Likewise.
736 (replace_ref): Likewise.
737 (cand_increment): Likewise. Return a widest_int.
738 (cand_abs_increment): Likewise.
739 (replace_mult_candidate): Take a widest_int instead of a double_int.
740 (replace_unconditional_candidate): Use wide-int interfaces.
741 (incr_vec_index): Take a widest_int instead of a double_int.
742 (create_add_on_incoming_edge): Likewise.
743 (create_phi_basis): Use wide-int interfaces.
744 (replace_conditional_candidate): Likewise.
745 (record_increment): Take a widest_int instead of a double_int.
746 (record_phi_increments): Use wide-int interfaces.
747 (phi_incr_cost): Take a widest_int instead of a double_int.
748 (lowest_cost_path): Likewise.
749 (total_savings): Likewise.
750 (analyze_increments): Use wide-int interfaces.
751 (ncd_with_phi): Take a widest_int instead of a double_int.
752 (ncd_of_cand_and_phis): Likewise.
753 (nearest_common_dominator_for_cands): Likewise.
754 (insert_initializers): Use wide-int interfaces.
755 (all_phi_incrs_profitable): Likewise.
756 (replace_one_candidate): Likewise.
757 (replace_profitable_candidates): Likewise.
758 * godump.c: Include wide-int-print.h.
759 (go_output_typedef): Use wide-int interfaces.
760 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
761 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
762 (build_loop_iteration_domains): Likewise.
763 * hooks.h: Include wide-int.h rather than double-int.h.
764 (hook_bool_dint_dint_uint_bool_true): Delete.
765 (hook_bool_wint_wint_uint_bool_true): Declare.
766 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
767 (hook_bool_wint_wint_uint_bool_true): New.
768 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
770 (ubsan_expand_si_overflow_mul_check): Likewise.
771 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
772 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
773 (get_ancestor_addr_info): Likewise.
774 (ipa_modify_call_arguments): Likewise.
775 * loop-doloop.c (doloop_modify): Likewise.
776 (doloop_optimize): Likewise.
777 * loop-iv.c (iv_number_of_iterations): Likewise.
778 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
779 (unroll_loop_constant_iterations): Likewise.
780 (decide_unroll_runtime_iterations): Likewise.
781 (unroll_loop_runtime_iterations): Likewise.
782 (decide_peel_simple): Likewise.
783 (decide_unroll_stupid): Likewise.
784 * lto-streamer-in.c (streamer_read_wi): Add.
785 (input_cfg): Use wide-int interfaces.
786 (lto_input_tree_1): Likewise.
787 * lto-streamer-out.c (streamer_write_wi): Add.
788 (hash_tree): Use wide-int interfaces.
789 (output_cfg): Likewise.
790 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
791 (GTFILES): Add wide-int.h and signop.h.
792 (TAGS): Look for .cc files too.
793 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
794 * optabs.c (expand_subword_shift): Likewise.
795 (expand_doubleword_shift): Likewise.
796 (expand_absneg_bit): Likewise.
797 (expand_copysign_absneg): Likewise.
798 (expand_copysign_bit): Likewise.
799 * postreload.c (reload_cse_simplify_set): Likewise.
800 * predict.c (predict_iv_comparison): Likewise.
801 * pretty-print.h: Include wide-int-print.h.
803 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
804 * print-tree.c: Include wide-int-print.h.
805 (print_node_brief): Use wide-int interfaces.
806 (print_node): Likewise.
807 * read-rtl.c (validate_const_wide_int): New.
808 (read_rtx_code): Add CONST_WIDE_INT case.
809 * real.c: Include wide-int.h.
810 (real_to_integer2): Delete.
811 (real_to_integer): New function, returning a wide_int.
812 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
813 (ten_to_ptwo): Update call to real_from_integer.
814 (real_digit): Likewise.
815 * real.h: Include signop.h, wide-int.h and insn-modes.h.
816 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
817 (REAL_VALUE_TO_INT): Delete.
818 (real_to_integer): Declare a wide-int form.
819 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
820 * recog.c (const_int_operand): Improve comment.
821 (const_scalar_int_operand): New.
822 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
823 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
824 (split_double): Likewise.
825 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
826 (rtx_size): Likewise.
827 (rtx_alloc_stat_v): New.
828 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
829 (cwi_output_hex): New.
830 (iterative_hash_rtx): Handle CONST_WIDE_INT.
831 (cwi_check_failed_bounds): New.
832 * rtl.def (CONST_WIDE_INT): New.
833 * rtl.h: Include <utility> and wide-int.h.
834 (struct hwivec_def): New.
835 (CWI_GET_NUM_ELEM): New.
836 (CWI_PUT_NUM_ELEM): New.
837 (struct rtx_def): Add num_elem and hwiv.
838 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
839 (CASE_CONST_UNIQUE): Likewise.
840 (CASE_CONST_ANY): Likewise.
841 (CONST_SCALAR_INT_P): Likewise.
842 (CONST_WIDE_INT_P): New.
845 (cwi_check_failed_bounds): New.
848 (CONST_WIDE_INT_VEC) New.
849 (CONST_WIDE_INT_NUNITS) New.
850 (CONST_WIDE_INT_ELT) New.
851 (rtx_mode_t): New type.
852 (wi::int_traits <rtx_mode_t>): New.
854 (wi::min_value): New.
855 (wi::max_value): New.
857 (const_wide_int_alloc): New.
858 (immed_wide_int_const): New.
859 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
860 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
861 * signop.h: New file.
862 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
863 (simplify_const_unary_operation): Use wide-int interfaces.
864 (simplify_binary_operation_1): Likewise.
865 (simplify_const_binary_operation): Likewise.
866 (simplify_const_relational_operation): Likewise.
867 (simplify_immed_subreg): Likewise.
868 * stmt.c (expand_case): Likewise.
869 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
870 signop rather than a bool.
871 * stor-layout.c (layout_type): Use wide-int interfaces.
872 (initialize_sizetypes): Update calls to
873 set_min_and_max_values_for_integral_type.
874 (set_min_and_max_values_for_integral_type): Take a signop rather
875 than a bool. Use wide-int interfaces.
876 (fixup_signed_type): Update accordingly. Remove
877 HOST_BITS_PER_DOUBLE_INT limit.
878 (fixup_unsigned_type): Likewise.
879 * system.h (STATIC_CONSTANT_P): New.
880 (STATIC_ASSERT): New.
881 * target.def (can_use_doloop_p): Take widest_ints rather than
883 * target.h: Include wide-int.h rather than double-int.h.
884 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
886 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
887 rather than INT_CST_LT_UNSIGNED.
888 (can_use_doloop_if_innermost): Take widest_ints rather than
890 * tree-affine.c: Include wide-int-print.h.
891 (double_int_ext_for_comb): Delete.
892 (wide_int_ext_for_comb): New.
893 (aff_combination_zero): Use wide-int interfaces.
894 (aff_combination_const): Take a widest_int instead of a double_int.
895 (aff_combination_elt): Use wide-int interfaces.
896 (aff_combination_scale): Take a widest_int instead of a double_int.
897 (aff_combination_add_elt): Likewise.
898 (aff_combination_add_cst): Likewise.
899 (aff_combination_add): Use wide-int interfaces.
900 (aff_combination_convert): Likewise.
901 (tree_to_aff_combination): Likewise.
902 (add_elt_to_tree): Take a widest_int instead of a double_int.
903 (aff_combination_to_tree): Use wide-int interfaces.
904 (aff_combination_remove_elt): Likewise.
905 (aff_combination_add_product): Take a widest_int instead of
907 (aff_combination_mult): Use wide-int interfaces.
908 (aff_combination_expand): Likewise.
909 (double_int_constant_multiple_p): Delete.
910 (wide_int_constant_multiple_p): New.
911 (aff_combination_constant_multiple_p): Take a widest_int pointer
912 instead of a double_int pointer.
913 (print_aff): Use wide-int interfaces.
914 (get_inner_reference_aff): Take a widest_int pointer
915 instead of a double_int pointer.
916 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
917 * tree-affine.h: Include wide-int.h.
918 (struct aff_comb_elt): Change type of coef to widest_int.
919 (struct affine_tree_combination): Change type of offset to widest_int.
920 (double_int_ext_for_comb): Delete.
921 (wide_int_ext_for_comb): New.
922 (aff_combination_const): Use widest_int instead of double_int.
923 (aff_combination_scale): Likewise.
924 (aff_combination_add_elt): Likewise.
925 (aff_combination_constant_multiple_p): Likewise.
926 (get_inner_reference_aff): Likewise.
927 (aff_comb_cannot_overlap_p): Likewise.
928 (aff_combination_zero_p): Use wide-int interfaces.
929 * tree.c: Include tree.h.
930 (init_ttree): Use make_int_cst.
931 (tree_code_size): Removed code for INTEGER_CST case.
932 (tree_size): Add INTEGER_CST case.
933 (make_node_stat): Update comment.
934 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
935 (build_int_cst_type): Use wide-int interfaces.
936 (double_int_to_tree): Likewise.
937 (double_int_fits_to_tree_p): Delete.
938 (force_fit_type_double): Delete.
939 (force_fit_type): New.
940 (int_cst_hash_hash): Use wide-int interfaces.
941 (int_cst_hash_eq): Likewise.
942 (build_int_cst_wide): Delete.
943 (wide_int_to_tree): New.
944 (cache_integer_cst): Use wide-int interfaces.
945 (build_low_bits_mask): Likewise.
946 (cst_and_fits_in_hwi): Likewise.
947 (real_value_from_int_cst): Likewise.
948 (make_int_cst_stat): New.
949 (integer_zerop): Use wide_int interfaces.
950 (integer_onep): Likewise.
951 (integer_all_onesp): Likewise.
952 (integer_pow2p): Likewise.
953 (integer_nonzerop): Likewise.
954 (tree_log2): Likewise.
955 (tree_floor_log2): Likewise.
956 (tree_ctz): Likewise.
957 (int_size_in_bytes): Likewise.
958 (mem_ref_offset): Return an offset_int rather than a double_int.
959 (build_type_attribute_qual_variant): Use wide_int interfaces.
960 (type_hash_eq): Likewise
961 (tree_int_cst_equal): Likewise.
962 (tree_int_cst_lt): Delete.
963 (tree_int_cst_compare): Likewise.
964 (tree_fits_shwi_p): Use wide_int interfaces.
965 (tree_fits_uhwi_p): Likewise.
966 (tree_int_cst_sign_bit): Likewise.
967 (tree_int_cst_sgn): Likewise.
968 (tree_int_cst_min_precision): Take a signop rather than a bool.
969 (simple_cst_equal): Use wide_int interfaces.
970 (compare_tree_int): Likewise.
971 (iterative_hash_expr): Likewise.
972 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
974 (get_type_static_bounds): Use wide_int interfaces.
975 (tree_int_cst_elt_check_failed): New.
976 (build_common_tree_nodes): Reordered to set prec before filling in
978 (int_cst_value): Check cst_and_fits_in_hwi.
979 (widest_int_cst_value): Use wide_int interfaces.
980 (upper_bound_in_type): Likewise.
981 (lower_bound_in_type): Likewise.
982 (num_ending_zeros): Likewise.
983 (drop_tree_overflow): Likewise.
984 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
985 (gen_conditions_for_pow_cst_base): Likewise.
986 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
987 (group_case_labels_stmt): Use wide-int interfaces.
988 (verify_gimple_assign_binary): Likewise.
989 (print_loop): Likewise.
990 * tree-chrec.c (tree_fold_binomial): Likewise.
991 * tree-core.h (struct tree_base): Add int_length.
992 (struct tree_int_cst): Change rep of value.
993 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
994 (dr_may_alias_p): Likewise.
995 (max_stmt_executions_tree): Likewise.
996 * tree.def (INTEGER_CST): Update comment.
997 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
998 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
999 * tree-dump.c: Include wide-int.h and wide-int-print.h.
1000 (dequeue_and_dump): Use wide-int interfaces.
1001 * tree.h: Include wide-int.h.
1002 (NULL_TREE): Moved to earlier loc in file.
1003 (TREE_INT_CST_ELT_CHECK): New.
1004 (tree_int_cst_elt_check_failed): New.
1006 (TREE_INT_CST): Delete.
1007 (TREE_INT_CST_LOW): Use wide-int interfaces.
1008 (TREE_INT_CST_HIGH): Delete.
1009 (TREE_INT_CST_NUNITS): New.
1010 (TREE_INT_CST_EXT_NUNITS): Likewise.
1011 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
1012 (TREE_INT_CST_ELT): Likewise.
1013 (INT_CST_LT): Delete.
1014 (tree_int_cst_elt_check): New (two forms).
1015 (type_code_size): Update comment.
1016 (make_int_cst_stat, make_int_cst): New.
1017 (tree_to_double_int): Delete.
1018 (double_int_fits_to_tree_p): Delete.
1019 (force_fit_type_double): Delete.
1020 (build_int_cstu): Replace with out-of-line function.
1021 (build_int_cst_wide): Delete.
1022 (tree_int_cst_lt): Define inline.
1023 (tree_int_cst_le): New.
1024 (tree_int_cst_compare): Define inline.
1025 (tree_int_cst_min_precision): Take a signop rather than a bool.
1026 (wi::int_traits <const_tree>): New.
1027 (wi::int_traits <tree>): New.
1028 (wi::extended_tree): New.
1029 (wi::int_traits <wi::extended_tree>): New.
1030 (wi::to_widest): New.
1031 (wi::to_offset): New.
1032 (wi::fits_to_tree_p): New.
1033 (wi::min_value): New.
1034 (wi::max_value): New.
1035 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
1036 (copy_tree_body_r): Likewise.
1037 * tree-object-size.c (compute_object_offset): Likewise.
1038 (addr_object_size): Likewise.
1039 * tree-predcom.c: Include wide-int-print.h.
1040 (struct dref_d): Change type of offset to widest_int.
1041 (dump_dref): Call wide-int printer.
1042 (aff_combination_dr_offset): Use wide-int interfaces.
1043 (determine_offset): Take a widest_int pointer rather than a
1045 (split_data_refs_to_components): Use wide-int interfaces.
1046 (suitable_component_p): Likewise.
1047 (order_drefs): Likewise.
1048 (add_ref_to_chain): Likewise.
1049 (valid_initializer_p): Likewise.
1050 (determine_roots_comp): Likewise.
1051 * tree-pretty-print.c: Include wide-int-print.h.
1052 (dump_generic_node): Use wide-int interfaces.
1053 * tree-sra.c (sra_ipa_modify_expr): Likewise.
1054 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
1055 (move_fixed_address_to_symbol): Likewise.
1056 (move_hint_to_base): Likewise.
1057 (move_pointer_to_base): Likewise.
1058 (move_variant_to_index): Likewise.
1059 (most_expensive_mult_to_index): Likewise.
1060 (addr_to_parts): Likewise.
1061 (copy_ref_info): Likewise.
1062 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
1063 (indirect_refs_may_alias_p): Likewise.
1064 (stmt_kills_ref_p_1): Likewise.
1065 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
1066 * tree-ssa-ccp.c: Update comment at top of file. Include
1068 (struct prop_value_d): Change type of mask to widest_int.
1069 (extend_mask): New function.
1070 (dump_lattice_value): Use wide-int interfaces.
1071 (get_default_value): Likewise.
1072 (set_constant_value): Likewise.
1073 (set_value_varying): Likewise.
1074 (valid_lattice_transition): Likewise.
1075 (set_lattice_value): Likewise.
1076 (value_to_double_int): Delete.
1077 (value_to_wide_int): New.
1078 (get_value_from_alignment): Use wide-int interfaces.
1079 (get_value_for_expr): Likewise.
1080 (do_dbg_cnt): Likewise.
1081 (ccp_finalize): Likewise.
1082 (ccp_lattice_meet): Likewise.
1083 (bit_value_unop_1): Use widest_ints rather than double_ints.
1084 (bit_value_binop_1): Likewise.
1085 (bit_value_unop): Use wide-int interfaces.
1086 (bit_value_binop): Likewise.
1087 (bit_value_assume_aligned): Likewise.
1088 (evaluate_stmt): Likewise.
1089 (ccp_fold_stmt): Likewise.
1090 (visit_cond_stmt): Likewise.
1091 (ccp_visit_stmt): Likewise.
1092 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
1093 (constant_pointer_difference): Likewise.
1094 (associate_pointerplus): Likewise.
1095 (combine_conversions): Likewise.
1096 * tree-ssa-loop.h: Include wide-int.h.
1097 (struct tree_niter_desc): Change type of max to widest_int.
1098 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
1099 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
1100 (remove_redundant_iv_tests): Likewise.
1101 (canonicalize_loop_induction_variables): Likewise.
1102 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
1103 (constant_multiple_of): Take a widest_int pointer instead of
1104 a double_int pointer.
1105 (get_computation_aff): Use wide-int interfaces.
1106 (ptr_difference_cost): Likewise.
1107 (difference_cost): Likewise.
1108 (get_loop_invariant_expr_id): Likewise.
1109 (get_computation_cost_at): Likewise.
1110 (iv_elimination_compare_lt): Likewise.
1111 (may_eliminate_iv): Likewise.
1112 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
1113 instead of double_int.
1114 (max_loop_iterations): Likewise.
1115 (max_stmt_executions): Likewise.
1116 (estimated_stmt_executions): Likewise.
1117 * tree-ssa-loop-niter.c: Include wide-int-print.h.
1118 (split_to_var_and_offset): Use wide-int interfaces.
1119 (determine_value_range): Likewise.
1120 (bound_difference_of_offsetted_base): Likewise.
1121 (bounds_add): Take a widest_int instead of a double_int.
1122 (number_of_iterations_ne_max): Use wide-int interfaces.
1123 (number_of_iterations_ne): Likewise.
1124 (number_of_iterations_lt_to_ne): Likewise.
1125 (assert_loop_rolls_lt): Likewise.
1126 (number_of_iterations_lt): Likewise.
1127 (number_of_iterations_le): Likewise.
1128 (number_of_iterations_cond): Likewise.
1129 (number_of_iterations_exit): Likewise.
1130 (finite_loop_p): Likewise.
1131 (derive_constant_upper_bound_assign): Likewise.
1132 (derive_constant_upper_bound): Return a widest_int.
1133 (derive_constant_upper_bound_ops): Likewise.
1134 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
1135 (record_estimate): Take a widest_int rather than a double_int.
1136 (record_nonwrapping_iv): Use wide-int interfaces.
1137 (double_int_cmp): Delete.
1138 (wide_int_cmp): New.
1139 (bound_index): Take a widest_int rather than a double_int.
1140 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
1141 (maybe_lower_iteration_bound): Likewise.
1142 (estimate_numbers_of_iterations_loop): Likewise.
1143 (estimated_loop_iterations): Take a widest_int pointer than than
1144 a double_int pointer.
1145 (estimated_loop_iterations_int): Use wide-int interfaces.
1146 (max_loop_iterations): Take a widest_int pointer than than
1147 a double_int pointer.
1148 (max_loop_iterations_int): Use wide-int interfaces.
1149 (max_stmt_executions): Take a widest_int pointer than than
1150 a double_int pointer.
1151 (estimated_stmt_executions): Likewise.
1152 (n_of_executions_at_most): Use wide-int interfaces.
1153 (scev_probably_wraps_p): Likewise.
1154 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
1156 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
1158 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
1159 double_ints. Adjust for trailing_wide_ints <3> representation.
1160 (set_nonzero_bits): Likewise.
1161 (get_range_info): Return wide_ints rather than double_ints.
1162 Adjust for trailing_wide_ints <3> representation.
1163 (get_nonzero_bits): Likewise.
1164 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
1166 * tree-ssanames.h (struct range_info_def): Replace min, max and
1167 nonzero_bits with a trailing_wide_ints <3>.
1168 (set_range_info): Use wide_int_refs rather than double_ints.
1169 (set_nonzero_bits): Likewise.
1170 (get_range_info): Return wide_ints rather than double_ints.
1171 (get_nonzero_bits): Likewise.
1172 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
1173 * tree-ssa-pre.c (phi_translate_1): Likewise.
1174 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
1175 (acceptable_pow_call): Likewise.
1176 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
1178 (vn_reference_fold_indirect): Likewise.
1179 (vn_reference_maybe_forwprop_address): Likewise.
1180 (valueize_refs_1): Likewise.
1181 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
1182 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
1183 tree_int_cst_lt and tree_int_cst_le.
1184 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
1186 (streamer_alloc_tree): Likewise.
1187 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
1188 (streamer_write_tree_header): Likewise.
1189 (streamer_write_integer_cst): Likewise.
1190 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
1191 (build_constructors): Likewise.
1192 (array_value_type): Likewise.
1193 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
1194 (vect_check_gather): Likewise.
1195 * tree-vect-generic.c (build_replicated_const): Likewise.
1196 (expand_vector_divmod): Likewise.
1197 * tree-vect-loop.c (vect_transform_loop): Likewise.
1198 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
1199 (vect_do_peeling_for_alignment): Likewise.
1200 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
1201 * tree-vrp.c: Include wide-int.h.
1202 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
1203 (extract_range_from_assert): Use wide-int interfaces.
1204 (vrp_int_const_binop): Likewise.
1205 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
1206 double_int pointers.
1207 (ranges_from_anti_range): Use wide-int interfaces.
1208 (quad_int_cmp): Delete.
1209 (quad_int_pair_sort): Likewise.
1210 (extract_range_from_binary_expr_1): Use wide-int interfaces.
1211 (extract_range_from_unary_expr_1): Likewise.
1212 (adjust_range_with_scev): Likewise.
1213 (masked_increment): Take and return wide_ints rather than double_ints.
1214 (register_edge_assert_for_2): Use wide-int interfaces.
1215 (check_array_ref): Likewise.
1216 (search_for_addr_array): Likewise.
1217 (maybe_set_nonzero_bits): Likewise.
1218 (union_ranges): Pass an integer of the correct type instead of
1219 using integer_one_node.
1220 (intersect_ranges): Likewise.
1221 (simplify_truth_ops_using_ranges): Likewise.
1222 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
1223 (range_fits_type_p): Likewise.
1224 (simplify_cond_using_ranges): Likewise. Take a signop rather than
1226 (simplify_conversion_using_ranges): Use wide-int interfaces.
1227 (simplify_float_conversion_using_ranges): Likewise.
1228 (vrp_finalize): Likewise.
1229 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
1230 (gimple_stringops_transform): Likewise.
1231 * varasm.c (decode_addr_const): Likewise.
1232 (const_hash_1): Likewise.
1233 (const_rtx_hash_1): Likewise
1234 (output_constant): Likewise.
1235 (array_size_for_constructor): Likewise.
1236 (output_constructor_regular_field): Likewise.
1237 (output_constructor_bitfield): Likewise.
1238 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
1239 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
1241 * gencheck.c: Define BITS_PER_UNIT.
1244 * wide-int-print.cc: New.
1245 * wide-int-print.h: New.
1247 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1249 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
1251 2014-05-06 Richard Biener <rguenther@suse.de>
1253 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
1254 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
1255 (TODO_verify_all): Adjust.
1256 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
1257 TODO_verify_stmts and TODO_verify_rtl_sharing.
1258 * bb-reorder.c: Likewise.
1259 * cfgexpand.c: Likewise.
1260 * cprop.c: Likewise.
1262 * function.c: Likewise.
1263 * fwprop.c: Likewise.
1265 * gimple-ssa-isolate-paths.c: Likewise.
1266 * gimple-ssa-strength-reduction.c: Likewise.
1267 * ipa-split.c: Likewise.
1268 * loop-init.c: Likewise.
1269 * loop-unroll.c: Likewise.
1270 * lower-subreg.c: Likewise.
1271 * modulo-sched.c: Likewise.
1272 * postreload-gcse.c: Likewise.
1273 * predict.c: Likewise.
1274 * recog.c: Likewise.
1275 * sched-rgn.c: Likewise.
1276 * store-motion.c: Likewise.
1277 * tracer.c: Likewise.
1278 * trans-mem.c: Likewise.
1279 * tree-call-cdce.c: Likewise.
1280 * tree-cfg.c: Likewise.
1281 * tree-cfgcleanup.c: Likewise.
1282 * tree-complex.c: Likewise.
1283 * tree-eh.c: Likewise.
1284 * tree-emutls.c: Likewise.
1285 * tree-if-conv.c: Likewise.
1286 * tree-into-ssa.c: Likewise.
1287 * tree-loop-distribution.c: Likewise.
1288 * tree-object-size.c: Likewise.
1289 * tree-parloops.c: Likewise.
1290 * tree-pass.h: Likewise.
1291 * tree-sra.c: Likewise.
1292 * tree-ssa-ccp.c: Likewise.
1293 * tree-ssa-copy.c: Likewise.
1294 * tree-ssa-copyrename.c: Likewise.
1295 * tree-ssa-dce.c: Likewise.
1296 * tree-ssa-dom.c: Likewise.
1297 * tree-ssa-dse.c: Likewise.
1298 * tree-ssa-forwprop.c: Likewise.
1299 * tree-ssa-ifcombine.c: Likewise.
1300 * tree-ssa-loop-ch.c: Likewise.
1301 * tree-ssa-loop-ivcanon.c: Likewise.
1302 * tree-ssa-loop.c: Likewise.
1303 * tree-ssa-math-opts.c: Likewise.
1304 * tree-ssa-phiopt.c: Likewise.
1305 * tree-ssa-phiprop.c: Likewise.
1306 * tree-ssa-pre.c: Likewise.
1307 * tree-ssa-reassoc.c: Likewise.
1308 * tree-ssa-sink.c: Likewise.
1309 * tree-ssa-strlen.c: Likewise.
1310 * tree-ssa-tail-merge.c: Likewise.
1311 * tree-ssa-uncprop.c: Likewise.
1312 * tree-switch-conversion.c: Likewise.
1313 * tree-tailcall.c: Likewise.
1314 * tree-vect-generic.c: Likewise.
1315 * tree-vectorizer.c: Likewise.
1316 * tree-vrp.c: Likewise.
1318 * var-tracking.c: Likewise.
1319 * bt-load.c: Likewise.
1320 * cfgcleanup.c: Likewise.
1321 * combine-stack-adj.c: Likewise.
1322 * combine.c: Likewise.
1323 * compare-elim.c: Likewise.
1324 * config/epiphany/resolve-sw-modes.c: Likewise.
1325 * config/i386/i386.c: Likewise.
1326 * config/mips/mips.c: Likewise.
1327 * config/s390/s390.c: Likewise.
1328 * config/sh/sh_treg_combine.cc: Likewise.
1329 * config/sparc/sparc.c: Likewise.
1332 * final.c: Likewise.
1333 * ifcvt.c: Likewise.
1334 * mode-switching.c: Likewise.
1335 * passes.c: Likewise.
1336 * postreload.c: Likewise.
1338 * reg-stack.c: Likewise.
1339 * regcprop.c: Likewise.
1340 * regrename.c: Likewise.
1343 2014-05-06 Richard Biener <rguenther@suse.de>
1346 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
1347 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
1349 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
1352 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
1354 2014-05-05 Radovan Obradovic <robradovic@mips.com>
1355 Tom de Vries <tom@codesourcery.com>
1357 * target.def (call_fusage_contains_non_callee_clobbers): New
1359 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
1361 (@node Miscellaneous Register Hooks): New node.
1362 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
1363 * doc/tm.texi: Regenerate.
1365 2014-05-05 Marek Polacek <polacek@redhat.com>
1368 * opts.c (common_handle_option): Call error_at instead of warning_at.
1370 2014-05-05 Richard Biener <rguenther@suse.de>
1372 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
1373 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
1374 under the TODO_verify_il umbrella.
1376 2014-05-05 Richard Biener <rguenther@suse.de>
1378 * passes.c (execute_function_todo): Move TODO_verify_flow under
1379 the TODO_verify_ul umbrella.
1381 2014-05-05 Richard Biener <rguenther@suse.de>
1384 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
1385 X & CST away from a CST that is the mask of a mode.
1387 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1389 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
1390 int argument to enum machine_mode.
1391 (picochip_class_max_nregs): Ditto.
1392 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
1393 (picochip_class_max_nregs): Ditto.
1395 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1397 * target.def: Add new target hook.
1398 * doc/tm.texi: Regenerate.
1399 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
1400 * targhooks.c (default_keep_leaf_when_profiled): New function.
1402 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
1403 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
1405 2014-05-05 Bin Cheng <bin.cheng@arm.com>
1407 PR tree-optimization/60363
1408 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
1409 (copy_phi_args): New parameters. Call get_value_locus_in_path.
1410 (update_destination_phis): New parameter.
1411 (create_edge_and_update_destination_phis): Ditto.
1412 (ssa_fix_duplicate_block_edges): Pass new arguments.
1413 (thread_single_edge): Ditto.
1415 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
1417 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
1418 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
1419 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
1420 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
1421 Use RS6000_BTM_HARD_FLOAT.
1422 (BU_MISC_2): Likewise.
1423 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
1424 RS6000_BTM_HARD_FLOAT.
1425 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
1427 (rs6000_invalid_builtin): Add hard floating builtin support.
1428 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
1429 hard float builtins.
1430 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
1432 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1434 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
1435 Add missing function* argument.
1437 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1439 * lra-constraints.c (valid_address_p): Move earlier in file.
1440 Add a constraint argument to the address_info version.
1441 (satisfies_memory_constraint_p): New function.
1442 (satisfies_address_constraint_p): Likewise.
1443 (process_alt_operands, curr_insn_transform): Use them.
1444 (process_address): Pass the constraint to valid_address_p when
1445 checking address operands.
1447 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1449 * config/mips/mips.c (mips_isa_rev): New variable.
1450 (mips_set_architecture): Set it.
1451 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1453 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1454 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1455 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1456 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1457 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
1458 conditions in terms of mips_isa_rev.
1459 (mips_isa_rev): Declare.
1461 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1463 * config/sh/sh-mem.cc: Use tabs instead of spaces.
1464 (prob_unlikely, prob_likely): Make variables const.
1466 2014-05-03 Denis Chertykov <chertykov@gmail.com>
1468 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
1470 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1472 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
1474 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1476 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
1477 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
1478 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
1480 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
1481 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
1483 Replace usage of ROUND_REG with sh_round_reg.
1484 Use CEIL instead of ROUND_ADVANCE.
1486 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1489 * config/sh/sh.c: Include stdlib headers before everything else.
1491 2014-05-02 Jakub Jelinek <jakub@redhat.com>
1493 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
1494 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
1495 (gimplify_adjust_omp_clauses): Simd region is never
1496 directly nested in combined parallel. Instead, for linear
1497 with copyin/copyout, if in combined for simd loop, make decl
1498 firstprivate/lastprivate on OMP_FOR.
1499 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
1500 expand_omp_for_static_chunk): When setting endvar, also set
1501 fd->loop.v to the same value.
1503 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1505 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
1507 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
1509 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
1512 2014-05-02 Marek Polacek <polacek@redhat.com>
1514 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
1516 2014-05-02 Kito Cheng <kito@0xlab.org>
1518 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
1519 to a C expression marco.
1520 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
1521 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
1522 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
1523 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
1524 HONOR_REG_ALLOC_ORDER.
1525 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
1527 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1529 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
1531 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1533 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
1535 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
1537 * tree-if-conv.c (is_cond_scalar_reduction): New function.
1538 (convert_scalar_cond_reduction): Likewise.
1539 (predicate_scalar_phi): Add recognition and transformation
1540 of simple conditioanl reduction to be vectorizable.
1542 2014-05-01 Marek Polacek <polacek@redhat.com>
1545 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
1547 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
1549 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
1550 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
1551 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
1552 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
1553 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
1554 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
1555 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
1556 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
1558 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
1560 * config/arc/arc.opt (mlra): Move comment above option name
1561 to avoid mis-parsing as language options.
1563 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1565 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
1566 * config/sol2.h: ... here.
1567 * config/sol2-10.h: Remove.
1569 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
1570 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
1571 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
1572 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
1573 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
1574 * config/sol2.h: ... here.
1575 (SECTION_NAME_FORMAT): Don't redefine.
1576 (STARTFILE_ARCH32_SPEC): Rename to ...
1577 (STARTFILE_ARCH_SPEC): ... this.
1578 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
1579 * config/sparc/sol2.h: ... here.
1580 (SECTION_NAME_FORMAT): Don't undef.
1581 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
1582 (SUBTARGET_EXTRA_SPECS): Remove.
1583 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
1585 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
1586 (MD_STARTFILE_PREFIX): Remove.
1587 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
1588 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
1589 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
1590 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
1591 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
1592 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
1593 * config/i386/sol2.h: ... here.
1594 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
1595 * config/i386/sol2-bi.h: Remove.
1596 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
1597 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
1599 * config/i386/t-sol2-64: Rename to ...
1600 * config/i386/t-sol2: ... this.
1601 * config/sparc/t-sol2-64: Rename to ...
1602 * config/sparc/t-sol2: ... this.
1604 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
1605 sol2_tm_file_head, sol2_tm_file_tail.
1606 Include ${cpu_type}/sol2.h before sol2.h.
1608 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
1609 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
1610 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
1611 Reflect i386/t-sol2-64 renaming.
1612 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
1613 Reflect sparc/t-sol2-64 renaming.
1615 2014-04-30 Richard Biener <rguenther@suse.de>
1617 * passes.c (execute_function_todo): Move TODO_verify_stmts
1618 and TODO_verify_ssa under the TODO_verify_il umbrella.
1619 * tree-ssa.h (verify_ssa): Adjust prototype.
1620 * tree-ssa.c (verify_ssa): Add parameter to tell whether
1621 we should verify SSA operands.
1622 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
1623 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
1624 whether we should verify whether not throwing stmts have EH info.
1625 * graphite-scop-detection.c (create_sese_edges): Adjust.
1626 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
1627 * tree-eh.c (lower_try_finally_switch): Do not add the
1628 default case label twice.
1630 2014-04-30 Marek Polacek <polacek@redhat.com>
1632 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
1633 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
1634 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
1635 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
1637 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
1639 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
1640 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
1641 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
1642 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
1643 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
1644 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
1645 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
1646 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
1648 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1650 * tree-cfg.c (dump_function_to_file): Dump the return type of
1651 functions, in a line to itself before the function body, mimicking
1652 the layout of a C function.
1654 2014-04-29 Jakub Jelinek <jakub@redhat.com>
1656 PR tree-optimization/60971
1657 * tree-tailcall.c (process_assignment): Reject conversions which
1660 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
1662 * calls.c (initialize_argument_information): Always treat
1663 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
1664 (expand_call): Likewise.
1665 (emit_library_call_calue_1): Likewise.
1666 * expr.c (PUSH_ARGS_REVERSED): Do not define.
1667 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
1670 2014-04-29 Nick Clifton <nickc@redhat.com>
1672 * config/msp430/msp430.md (umulsidi): Fix typo.
1673 (mulhisi3): Enable even inside interrupt handlers.
1674 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
1675 bigger return address pushed in large mode.
1677 2014-04-29 Nick Clifton <nickc@redhat.com>
1679 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
1680 (arc_init_reg_tables): Use a machine_mode enum to iterate over
1682 * config/m32r/m32r.c (init_reg_tables): Likewise.
1683 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
1684 enum to hold the modes.
1686 2014-04-29 Richard Biener <rguenther@suse.de>
1688 * dominance.c (free_dominance_info): Add overload with
1690 (dom_info_state): Likewise.
1691 (dom_info_available_p): Likewise.
1692 * basic-block.h (free_dominance_info, dom_info_state,
1693 dom_info_available_p): Declare overloads.
1694 * passes.c (execute_function_todo): Verify that verifiers
1695 don't change dominator info state. Drop dominator info
1696 for IPA pass invocations.
1697 * cgraph.c (release_function_body): Restore asserts that
1698 dominator information is released.
1700 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
1702 * doc/invoke.texi: Fix typo.
1703 * tree-vrp.c: Fix typos.
1704 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
1706 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1708 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
1710 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1712 * config/aarch64/aarch64-builtins.c
1713 (aarch64_types_storestruct_lane_qualifiers): New.
1714 (TYPES_STORESTRUCT_LANE): Likewise.
1715 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
1716 (st3_lane): Likewise.
1717 (st4_lane): Likewise.
1718 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
1719 (vec_store_lanesci_lane<mode>): Likewise.
1720 (vec_store_lanesxi_lane<mode>): Likewise.
1721 (aarch64_st2_lane<VQ:mode>): Likewise.
1722 (aarch64_st3_lane<VQ:mode>): Likewise.
1723 (aarch64_st4_lane<VQ:mode>): Likewise.
1724 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
1725 * config/aarch64/arm_neon.h
1726 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
1727 use new macro arguments.
1728 (__ST3_LANE_FUNC): Likewise.
1729 (__ST4_LANE_FUNC): Likewise.
1730 * config/aarch64/iterators.md (V_TWO_ELEM): New.
1731 (V_THREE_ELEM): Likewise.
1732 (V_FOUR_ELEM): Likewise.
1734 2014-04-28 David Malcolm <dmalcolm@redhat.com>
1736 * doc/gimple.texi: Replace the description of the now-defunct
1737 union gimple_statement_d with a diagram showing the
1738 gimple_statement_base class hierarchy and its relationships to
1739 the GSS_ and GIMPLE_ enums.
1741 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1743 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
1744 * config/aarch64/aarch64.c
1745 (aarch64_cannot_change_mode_class): Weaken conditions.
1746 (aarch64_modes_tieable_p): New.
1747 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
1749 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
1751 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
1752 (loadsync_<mode>): Change mode.
1753 (load_quadpti, store_quadpti): New.
1754 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
1755 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
1757 2014-04-28 Martin Jambor <mjambor@suse.cz>
1759 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
1760 same alias type as the original statement.
1761 (subreplacement_assignment_data): New type.
1762 (handle_unscalarized_data_in_subtree): New type of parameter,
1763 generate new memory accesses with same alias type as the original
1765 (load_assign_lhs_subreplacements): Likewise.
1766 (sra_modify_constructor_assign): Generate new memory accesses with
1767 same alias type as the original statement.
1769 2014-04-28 Richard Biener <rguenther@suse.de>
1771 * tree-pass.h (TODO_verify_il): Define.
1772 (TODO_verify_all): Complete properly.
1773 * passes.c (execute_function_todo): Move existing loop-closed
1774 SSA verification under TODO_verify_il.
1775 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
1776 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
1777 Fix tree sharing issue.
1779 2014-04-28 Richard Biener <rguenther@suse.de>
1782 * builtins.def (DEF_C11_BUILTIN): Add.
1783 (BUILT_IN_ALIGNED_ALLOC): Likewise.
1784 * coretypes.h (enum function_class): Add function_c11_misc.
1785 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1786 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
1787 (call_may_clobber_ref_p_1): Likewise.
1788 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
1789 (mark_all_reaching_defs_necessary_1): Likewise.
1790 (propagate_necessity): Likewise.
1791 (eliminate_unnecessary_stmts): Likewise.
1792 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
1794 2014-04-28 Richard Biener <rguenther@suse.de>
1796 * tree-vrp.c (vrp_var_may_overflow): Remove.
1797 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
1798 with overflow immediately bump to one before that value and
1799 let iteration figure out overflow status.
1801 2014-04-28 Richard Biener <rguenther@suse.de>
1803 * configure.ac: Do valgrind header checks unconditionally.
1804 Add --enable-valgrind-annotations.
1805 * system.h: Guard valgrind header inclusion with
1806 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
1807 * alloc-pool.c (pool_alloc, pool_free): Use
1808 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
1809 to guard possibly dead code.
1810 * config.in: Regenerated.
1811 * configure: Likewise.
1813 2014-04-28 Jeff Law <law@redhat.com>
1815 PR tree-optimization/60902
1816 * tree-ssa-threadedge.c
1817 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
1818 over real defs when invalidating outputs from statements that do not
1819 produce useful outputs for threading.
1821 2014-04-28 Richard Biener <rguenther@suse.de>
1823 PR tree-optimization/60979
1824 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
1825 SCOPs that end in a block with a successor with abnormal
1828 2014-04-28 Richard Biener <rguenther@suse.de>
1830 * tree-pass.h (execute_pass_list): Adjust prototype.
1831 * passes.c (pass_manager::execute_early_local_passes): Adjust.
1832 (do_per_function): Change callback signature, push all actual
1833 work to the callbals.
1834 (do_per_function_toporder): Likewise.
1835 (execute_function_dump): Adjust.
1836 (execute_function_todo): Likewise.
1837 (clear_last_verified): Likewise.
1838 (verify_curr_properties): Likewise.
1839 (update_properties_after_pass): Likewise.
1840 (execute_pass_list_1): Split out from ...
1841 (execute_pass_list): ... here. Adjust.
1842 (execute_ipa_pass_list): Likewise.
1843 * cgraphunit.c (cgraph_add_new_function): Adjust.
1844 (analyze_function): Likewise.
1845 (expand_function): Likewise.
1846 * cgraph.c (release_function_body): Free dominance info
1847 here instead of asserting it was magically freed elsewhere.
1849 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
1851 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
1852 * configure: Regenerate.
1853 * config/sparc/sparc.opt (muser-mode): New option.
1854 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
1856 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
1857 * doc/invoke.texi (SPARC options): Document -muser-mode.
1859 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
1861 * cselib.c (find_slot_memmode): Delete.
1862 (cselib_hasher): Change compare_type to a struct.
1863 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
1865 (preserve_constants_and_equivs): Adjust for new compare_type.
1866 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
1867 (wrap_constant): Delete.
1868 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
1870 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
1872 * doc/install.texi (Building with profile feedback): Remove
1875 2014-04-26 Tom de Vries <tom@codesourcery.com>
1877 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
1880 2014-04-25 Cary Coutant <ccoutant@google.com>
1883 * dwarf2out.c (should_move_die_to_comdat): A type definition
1884 can contain a subprogram definition, but don't move it to a
1886 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
1887 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
1889 (clone_tree_hash): Rename to...
1890 (clone_tree_partial): ...this; change callers. Copy
1891 DW_TAG_subprogram DIEs as declarations.
1892 (copy_decls_walk): Don't copy children of a declaration into a
1895 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
1898 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
1901 2014-04-25 Jiong Wang <jiong.wang@arm.com>
1903 * config/arm/predicates.md (call_insn_operand): Add long_call check.
1904 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
1906 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
1909 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1911 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
1913 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1915 PR tree-optimization/60930
1916 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
1917 creating a multiply candidate by folding two constant
1918 multiplicands when the result overflows.
1920 2014-04-25 Jakub Jelinek <jakub@redhat.com>
1922 PR tree-optimization/60960
1923 * tree-vect-generic.c (expand_vector_operation): Only call
1924 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
1926 2014-04-25 Tom de Vries <tom@codesourcery.com>
1928 * expr.c (clobber_reg_mode): New function.
1929 * expr.h (clobber_reg): New function.
1931 2014-04-25 Tom de Vries <tom@codesourcery.com>
1933 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
1936 2014-04-25 Radovan Obradovic <robradovic@mips.com>
1937 Tom de Vries <tom@codesourcery.com>
1939 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
1941 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
1942 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
1943 new argument to find_all_hard_reg_sets call.
1945 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1947 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
1948 Use HOST_WIDE_INT_C for mask literal.
1949 (aarch_rev16_shleft_mask_imm_p): Likewise.
1951 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
1954 * config/sparc/sparc.md (ashlsi3_extend): Delete.
1956 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
1958 PR preprocessor/56540
1959 * config/i386/i386-c.c (ix86_target_macros): Define
1960 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
1962 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1964 * configure.ac (tga_func): Remove.
1965 (LIB_TLS_SPEC): Remove.
1966 * configure: Regenerate.
1967 * config.in: Regenerate.
1968 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
1970 2014-04-25 Richard Biener <rguenther@suse.de>
1973 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
1974 call stmt use/clobber sets during stmt walk instead of
1975 walking the possibly incomplete set of caller edges.
1977 2014-04-25 Richard Biener <rguenther@suse.de>
1980 * passes.c (apply_ipa_transforms): Inline into only caller ...
1981 (execute_one_pass): ... here. Properly bring in function
1982 bodies for nodes we want to apply IPA transforms to.
1984 2014-04-24 Cong Hou <congh@google.com>
1986 PR tree-optimization/60896
1987 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
1988 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
1989 (vect_mark_pattern_stmts): Set the def type of all statements in
1990 PATTERN_DEF_SEQ as vect_internal_def.
1992 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
1994 * doc/extend.texi (PowerPC Built-in Functions): Document new
1995 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
1996 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
1998 * config/rs6000/predicates.md (const_0_to_3_operand): New
1999 predicate to match 0..3 integer constants.
2001 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
2002 to support adding miscellaneous builtin functions.
2003 (BU_DFP_MISC_2): Likewise.
2004 (BU_P7_MISC_1): Likewise.
2005 (BU_P7_MISC_2): Likewise.
2006 (BU_P8V_MISC_3): Likewise.
2007 (BU_MISC_1): Likewise.
2008 (BU_MISC_2): Likewise.
2009 (DIVWE): Add extended divide builtin functions.
2012 (DIVWEUO): Likewise.
2016 (DIVDEUO): Likewise.
2017 (DXEX): Add decimal floating-point builtin functions.
2020 (DDEDPDQ): Likewise.
2022 (DENBCDQ): Likewise.
2029 (CDTBCD): Add new BCD builtin functions.
2033 (BCDADD_LT): Likewise.
2034 (BCDADD_EQ): Likewise.
2035 (BCDADD_GT): Likewise.
2036 (BCDADD_OV): Likewise.
2038 (BCDSUB_LT): Likewise.
2039 (BCDSUB_EQ): Likewise.
2040 (BCDSUB_GT): Likewise.
2041 (BCDSUB_OV): Likewise.
2042 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
2043 (UNPACK_TD): Likewise.
2044 (PACK_TF): Likewise.
2045 (UNPACK_TF): Likewise.
2046 (UNPACK_TF_0): Likewise.
2047 (UNPACK_TF_1): Likewise.
2048 (PACK_V1TI): Likewise.
2049 (UNPACK_V1TI): Likewise.
2051 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
2052 support for decimal floating point builtin functions.
2053 (rs6000_expand_ternop_builtin): Add checks for the new builtin
2054 functions that take constant arguments.
2055 (rs6000_invalid_builtin): Add decimal floating point builtin support.
2056 (rs6000_init_builtins): Setup long double, _Decimal64, and
2057 _Decimal128 types for new builtin functions.
2058 (builtin_function_type): Set the unsigned flags appropriately for
2059 the new builtin functions.
2060 (rs6000_opt_masks): Add support for decimal floating point builtin
2063 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
2064 floating point builtin functions.
2065 (RS6000_BTM_COMMON): Likewise.
2066 (RS6000_BTI_long_double): Likewise.
2067 (RS6000_BTI_dfloat64): Likewise.
2068 (RS6000_BTI_dfloat128): Likewise.
2069 (long_double_type_internal_node): Likewise.
2070 (dfloat64_type_internal_node): Likewise.
2071 (dfloat128_type_internal_node): Likewise.
2073 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
2074 2.07 bcd arithmetic instructions.
2075 (UNSPEC_BCDSUB): Likewise.
2076 (UNSPEC_BCD_OVERFLOW): Likewise.
2077 (UNSPEC_BCD_ADD_SUB): Likewise.
2078 (bcd_add_sub): Likewise.
2079 (BCD_TEST): Likewise.
2080 (bcd<bcd_add_sub>): Likewise.
2081 (bcd<bcd_add_sub>_test): Likewise.
2082 (bcd<bcd_add_sub>_test2): Likewise.
2083 (bcd<bcd_add_sub>_<code>): Likewise.
2084 (peephole2 for combined bcd ops): Likewise.
2086 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
2087 decimal floating point builtin functions.
2088 (UNSPEC_DENBCD): Likewise.
2089 (UNSPEC_DXEX): Likewise.
2090 (UNSPEC_DIEX): Likewise.
2091 (UNSPEC_DSCLI): Likewise.
2092 (UNSPEC_DSCRI): Likewise.
2093 (D64_D128): Likewise.
2094 (dfp_suffix): Likewise.
2095 (dfp_ddedpd_<mode>): Likewise.
2096 (dfp_denbcd_<mode>): Likewise.
2097 (dfp_dxex_<mode>): Likewise.
2098 (dfp_diex_<mode>): Likewise.
2099 (dfp_dscli_<mode>): Likewise.
2100 (dfp_dscri_<mode>): Likewise.
2102 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
2104 (UNSPEC_CDTBCD): Likewise.
2105 (UNSPEC_CBCDTD): Likewise.
2106 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
2107 (UNSPEC_DIVEO): Likewise.
2108 (UNSPEC_DIVEU): Likewise.
2109 (UNSPEC_DIVEUO): Likewise.
2110 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
2111 pack/unpack 128-bit types.
2112 (UNSPEC_PACK_128BIT): Likewise.
2113 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
2114 (udiv<mode>3): Use idiv_ldiv mode attribute.
2115 (div<mode>3): Likewise.
2116 (addg6s): Add new BCD builtin functions.
2119 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
2120 (div_extend): Likewise.
2121 (div<div_extend>_<mode>"): Likewise.
2122 (FP128_64): Add support for new builtin functions to pack/unpack
2124 (unpack<mode>): Likewise.
2125 (unpacktf_0): Likewise.
2126 (unpacktf_1): Likewise.
2127 (unpack<mode>_dm): Likewise.
2128 (unpack<mode>_nodm): Likewise.
2129 (pack<mode>): Likewise.
2130 (unpackv1ti): Likewise.
2131 (packv1ti): Likewise.
2133 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
2135 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
2138 2014-04-24 Jakub Jelinek <jakub@redhat.com>
2140 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
2141 * gimplify.c (omp_is_private): Change last argument's type to int.
2142 Only diagnose lastprivate if the simd argument is 1, only diagnose
2143 linear if the simd argument is 2.
2144 (gimplify_omp_for): Adjust omp_is_private callers. When adding
2145 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
2146 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
2147 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
2148 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
2149 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2150 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
2151 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2152 * tree-nested.c (convert_nonlocal_omp_clauses,
2153 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
2155 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
2158 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
2161 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
2163 * flag-types.h (enum ivar_visibility): Add.
2165 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
2167 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
2168 function * argument.
2170 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
2172 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
2174 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2175 Tom de Vries <tom@codesourcery.com>
2177 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
2178 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
2180 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
2181 * emit-rtl.c (try_split): Same.
2183 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2184 Tom de Vries <tom@codesourcery.com>
2186 * common.opt (fuse-caller-save): New option.
2188 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
2190 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
2191 elements for big-endian.
2193 2014-04-24 Richard Biener <rguenther@suse.de>
2195 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
2196 during TER and instead use the sepops interface for expanding
2197 non-GIMPLE_SINGLE_RHS.
2199 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2201 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
2202 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
2204 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2206 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
2207 assembler 64-bit option.
2208 * configure: Regenerate.
2210 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2212 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
2213 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
2214 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
2215 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
2216 (TARGET_CRYPTO): Take TARGET_SIMD into account.
2218 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2220 * config/aarch64/aarch64-builtins.c
2221 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
2222 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
2223 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
2224 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
2226 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
2227 (Vrevsuff): New mode attribute.
2229 2014-04-24 Terry Guo <terry.guo@arm.com>
2231 * config/arm/arm.h (machine_function): Define variable
2232 after_arm_reorg here.
2233 * config/arm/arm.c (after_arm_reorg): Remove the definition.
2234 (arm_split_constant): Update the way to access variable
2237 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
2239 2014-04-23 Tom de Vries <tom@codesourcery.com>
2241 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
2243 2014-04-23 David Malcolm <dmalcolm@redhat.com>
2245 * is-a.h: Update comments to reflect the following changes to the
2246 "pointerness" of the API, making the template parameter match the
2247 return type, allowing use of is-a.h with typedefs of pointers.
2248 (is_a_helper::cast): Return a T rather then a pointer to a T, so
2249 that the return type matches the parameter to the is_a_helper.
2251 (dyn_cast): Likewise.
2253 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
2254 pointer from the is-a.h API.
2256 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
2257 (is_a_helper <cgraph_node *>::test): ...this, matching change to
2259 (is_a_helper <varpool_node>::test): Likewise, convert to...
2260 (is_a_helper <varpool_node *>::test): ...this.
2262 (varpool_first_variable): Update for removal of implicit pointer
2263 from the is-a.h API.
2264 (varpool_next_variable): Likewise.
2265 (varpool_first_static_initializer): Likewise.
2266 (varpool_next_static_initializer): Likewise.
2267 (varpool_first_defined_variable): Likewise.
2268 (varpool_next_defined_variable): Likewise.
2269 (cgraph_first_defined_function): Likewise.
2270 (cgraph_next_defined_function): Likewise.
2271 (cgraph_first_function): Likewise.
2272 (cgraph_next_function): Likewise.
2273 (cgraph_first_function_with_gimple_body): Likewise.
2274 (cgraph_next_function_with_gimple_body): Likewise.
2275 (cgraph_alias_target): Likewise.
2276 (varpool_alias_target): Likewise.
2277 (cgraph_function_or_thunk_node): Likewise.
2278 (varpool_variable_node): Likewise.
2279 (symtab_real_symbol_p): Likewise.
2280 * cgraphunit.c (referred_to_p): Likewise.
2281 (analyze_functions): Likewise.
2282 (handle_alias_pairs): Likewise.
2283 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2284 * gimple-ssa.h (gimple_vuse_op): Likewise.
2285 (gimple_vdef_op): Likewise.
2286 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
2287 * gimple.c (gimple_build_asm_1): Likewise.
2288 (gimple_build_try): Likewise.
2289 (gimple_build_resx): Likewise.
2290 (gimple_build_eh_dispatch): Likewise.
2291 (gimple_build_omp_for): Likewise.
2292 (gimple_omp_for_set_clauses): Likewise.
2294 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
2295 (is_a_helper <gimple_statement_asm *>::test): ...this.
2296 (is_a_helper <gimple_statement_bind>::test): Convert to...
2297 (is_a_helper <gimple_statement_bind *>::test): ...this.
2298 (is_a_helper <gimple_statement_call>::test): Convert to...
2299 (is_a_helper <gimple_statement_call *>::test): ...this.
2300 (is_a_helper <gimple_statement_catch>::test): Convert to...
2301 (is_a_helper <gimple_statement_catch *>::test): ...this.
2302 (is_a_helper <gimple_statement_resx>::test): Convert to...
2303 (is_a_helper <gimple_statement_resx *>::test): ...this.
2304 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
2305 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
2306 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
2307 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
2308 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
2309 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
2310 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
2311 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
2312 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
2313 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
2314 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
2315 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
2316 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
2317 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
2318 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
2319 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
2320 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
2321 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
2322 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
2323 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
2324 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
2325 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
2326 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
2327 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
2328 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
2329 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
2330 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
2331 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
2332 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
2333 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
2334 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
2335 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
2336 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
2337 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
2338 (is_a_helper <gimple_statement_phi>::test): Convert to...
2339 (is_a_helper <gimple_statement_phi *>::test): ...this.
2340 (is_a_helper <gimple_statement_transaction>::test): Convert to...
2341 (is_a_helper <gimple_statement_transaction *>::test): ...this.
2342 (is_a_helper <gimple_statement_try>::test): Convert to...
2343 (is_a_helper <gimple_statement_try *>::test): ...this.
2344 (is_a_helper <gimple_statement_wce>::test): Convert to...
2345 (is_a_helper <gimple_statement_wce *>::test): ...this.
2346 (is_a_helper <const gimple_statement_asm>::test): Convert to...
2347 (is_a_helper <const gimple_statement_asm *>::test): ...this.
2348 (is_a_helper <const gimple_statement_bind>::test): Convert to...
2349 (is_a_helper <const gimple_statement_bind *>::test): ...this.
2350 (is_a_helper <const gimple_statement_call>::test): Convert to...
2351 (is_a_helper <const gimple_statement_call *>::test): ...this.
2352 (is_a_helper <const gimple_statement_catch>::test): Convert to...
2353 (is_a_helper <const gimple_statement_catch *>::test): ...this.
2354 (is_a_helper <const gimple_statement_resx>::test): Convert to...
2355 (is_a_helper <const gimple_statement_resx *>::test): ...this.
2356 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
2357 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
2358 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
2359 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
2360 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
2362 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
2364 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
2366 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
2368 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
2369 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
2370 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
2372 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
2373 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
2375 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
2376 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
2377 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
2378 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
2379 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
2380 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
2382 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
2383 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
2384 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
2385 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
2387 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
2388 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
2389 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
2390 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
2391 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
2392 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
2393 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
2394 (is_a_helper <const gimple_statement_phi>::test): Convert to...
2395 (is_a_helper <const gimple_statement_phi *>::test): ...this.
2396 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
2397 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
2398 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
2399 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
2400 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
2401 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
2402 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
2404 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
2406 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
2407 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
2409 (gimple_use_ops): Update for removal of implicit pointer from the
2411 (gimple_set_use_ops): Likewise.
2412 (gimple_vuse): Likewise.
2413 (gimple_vdef): Likewise.
2414 (gimple_vuse_ptr): Likewise.
2415 (gimple_vdef_ptr): Likewise.
2416 (gimple_set_vuse): Likewise.
2417 (gimple_set_vdef): Likewise.
2418 (gimple_omp_return_set_lhs): Likewise.
2419 (gimple_omp_return_lhs): Likewise.
2420 (gimple_omp_return_lhs_ptr): Likewise.
2421 (gimple_call_fntype): Likewise.
2422 (gimple_call_set_fntype): Likewise.
2423 (gimple_call_set_internal_fn): Likewise.
2424 (gimple_call_use_set): Likewise.
2425 (gimple_call_clobber_set): Likewise.
2426 (gimple_bind_vars): Likewise.
2427 (gimple_bind_set_vars): Likewise.
2428 (gimple_bind_body_ptr): Likewise.
2429 (gimple_bind_set_body): Likewise.
2430 (gimple_bind_add_stmt): Likewise.
2431 (gimple_bind_block): Likewise.
2432 (gimple_bind_set_block): Likewise.
2433 (gimple_asm_ninputs): Likewise.
2434 (gimple_asm_noutputs): Likewise.
2435 (gimple_asm_nclobbers): Likewise.
2436 (gimple_asm_nlabels): Likewise.
2437 (gimple_asm_input_op): Likewise.
2438 (gimple_asm_input_op_ptr): Likewise.
2439 (gimple_asm_output_op): Likewise.
2440 (gimple_asm_output_op_ptr): Likewise.
2441 (gimple_asm_set_output_op): Likewise.
2442 (gimple_asm_clobber_op): Likewise.
2443 (gimple_asm_set_clobber_op): Likewise.
2444 (gimple_asm_label_op): Likewise.
2445 (gimple_asm_set_label_op): Likewise.
2446 (gimple_asm_string): Likewise.
2447 (gimple_catch_types): Likewise.
2448 (gimple_catch_types_ptr): Likewise.
2449 (gimple_catch_handler_ptr): Likewise.
2450 (gimple_catch_set_types): Likewise.
2451 (gimple_catch_set_handler): Likewise.
2452 (gimple_eh_filter_types): Likewise.
2453 (gimple_eh_filter_types_ptr): Likewise.
2454 (gimple_eh_filter_failure_ptr): Likewise.
2455 (gimple_eh_filter_set_types): Likewise.
2456 (gimple_eh_filter_set_failure): Likewise.
2457 (gimple_eh_must_not_throw_fndecl): Likewise.
2458 (gimple_eh_must_not_throw_set_fndecl): Likewise.
2459 (gimple_eh_else_n_body_ptr): Likewise.
2460 (gimple_eh_else_e_body_ptr): Likewise.
2461 (gimple_eh_else_set_n_body): Likewise.
2462 (gimple_eh_else_set_e_body): Likewise.
2463 (gimple_try_eval_ptr): Likewise.
2464 (gimple_try_cleanup_ptr): Likewise.
2465 (gimple_try_set_eval): Likewise.
2466 (gimple_try_set_cleanup): Likewise.
2467 (gimple_wce_cleanup_ptr): Likewise.
2468 (gimple_wce_set_cleanup): Likewise.
2469 (gimple_phi_capacity): Likewise.
2470 (gimple_phi_num_args): Likewise.
2471 (gimple_phi_result): Likewise.
2472 (gimple_phi_result_ptr): Likewise.
2473 (gimple_phi_set_result): Likewise.
2474 (gimple_phi_arg): Likewise.
2475 (gimple_phi_set_arg): Likewise.
2476 (gimple_resx_region): Likewise.
2477 (gimple_resx_set_region): Likewise.
2478 (gimple_eh_dispatch_region): Likewise.
2479 (gimple_eh_dispatch_set_region): Likewise.
2480 (gimple_omp_critical_name): Likewise.
2481 (gimple_omp_critical_name_ptr): Likewise.
2482 (gimple_omp_critical_set_name): Likewise.
2483 (gimple_omp_for_clauses): Likewise.
2484 (gimple_omp_for_clauses_ptr): Likewise.
2485 (gimple_omp_for_set_clauses): Likewise.
2486 (gimple_omp_for_collapse): Likewise.
2487 (gimple_omp_for_index): Likewise.
2488 (gimple_omp_for_index_ptr): Likewise.
2489 (gimple_omp_for_set_index): Likewise.
2490 (gimple_omp_for_initial): Likewise.
2491 (gimple_omp_for_initial_ptr): Likewise.
2492 (gimple_omp_for_set_initial): Likewise.
2493 (gimple_omp_for_final): Likewise.
2494 (gimple_omp_for_final_ptr): Likewise.
2495 (gimple_omp_for_set_final): Likewise.
2496 (gimple_omp_for_incr): Likewise.
2497 (gimple_omp_for_incr_ptr): Likewise.
2498 (gimple_omp_for_set_incr): Likewise.
2499 (gimple_omp_for_pre_body_ptr): Likewise.
2500 (gimple_omp_for_set_pre_body): Likewise.
2501 (gimple_omp_parallel_clauses): Likewise.
2502 (gimple_omp_parallel_clauses_ptr): Likewise.
2503 (gimple_omp_parallel_set_clauses): Likewise.
2504 (gimple_omp_parallel_child_fn): Likewise.
2505 (gimple_omp_parallel_child_fn_ptr): Likewise.
2506 (gimple_omp_parallel_set_child_fn): Likewise.
2507 (gimple_omp_parallel_data_arg): Likewise.
2508 (gimple_omp_parallel_data_arg_ptr): Likewise.
2509 (gimple_omp_parallel_set_data_arg): Likewise.
2510 (gimple_omp_task_clauses): Likewise.
2511 (gimple_omp_task_clauses_ptr): Likewise.
2512 (gimple_omp_task_set_clauses): Likewise.
2513 (gimple_omp_task_child_fn): Likewise.
2514 (gimple_omp_task_child_fn_ptr): Likewise.
2515 (gimple_omp_task_set_child_fn): Likewise.
2516 (gimple_omp_task_data_arg): Likewise.
2517 (gimple_omp_task_data_arg_ptr): Likewise.
2518 (gimple_omp_task_set_data_arg): Likewise.
2519 (gimple_omp_taskreg_clauses): Likewise.
2520 (gimple_omp_taskreg_clauses_ptr): Likewise.
2521 (gimple_omp_taskreg_set_clauses): Likewise.
2522 (gimple_omp_taskreg_child_fn): Likewise.
2523 (gimple_omp_taskreg_child_fn_ptr): Likewise.
2524 (gimple_omp_taskreg_set_child_fn): Likewise.
2525 (gimple_omp_taskreg_data_arg): Likewise.
2526 (gimple_omp_taskreg_data_arg_ptr): Likewise.
2527 (gimple_omp_taskreg_set_data_arg): Likewise.
2528 (gimple_omp_task_copy_fn): Likewise.
2529 (gimple_omp_task_copy_fn_ptr): Likewise.
2530 (gimple_omp_task_set_copy_fn): Likewise.
2531 (gimple_omp_task_arg_size): Likewise.
2532 (gimple_omp_task_arg_size_ptr): Likewise.
2533 (gimple_omp_task_set_arg_size): Likewise.
2534 (gimple_omp_task_arg_align): Likewise.
2535 (gimple_omp_task_arg_align_ptr): Likewise.
2536 (gimple_omp_task_set_arg_align): Likewise.
2537 (gimple_omp_single_clauses): Likewise.
2538 (gimple_omp_single_clauses_ptr): Likewise.
2539 (gimple_omp_single_set_clauses): Likewise.
2540 (gimple_omp_target_clauses): Likewise.
2541 (gimple_omp_target_clauses_ptr): Likewise.
2542 (gimple_omp_target_set_clauses): Likewise.
2543 (gimple_omp_target_child_fn): Likewise.
2544 (gimple_omp_target_child_fn_ptr): Likewise.
2545 (gimple_omp_target_set_child_fn): Likewise.
2546 (gimple_omp_target_data_arg): Likewise.
2547 (gimple_omp_target_data_arg_ptr): Likewise.
2548 (gimple_omp_target_set_data_arg): Likewise.
2549 (gimple_omp_teams_clauses): Likewise.
2550 (gimple_omp_teams_clauses_ptr): Likewise.
2551 (gimple_omp_teams_set_clauses): Likewise.
2552 (gimple_omp_sections_clauses): Likewise.
2553 (gimple_omp_sections_clauses_ptr): Likewise.
2554 (gimple_omp_sections_set_clauses): Likewise.
2555 (gimple_omp_sections_control): Likewise.
2556 (gimple_omp_sections_control_ptr): Likewise.
2557 (gimple_omp_sections_set_control): Likewise.
2558 (gimple_omp_for_set_cond): Likewise.
2559 (gimple_omp_for_cond): Likewise.
2560 (gimple_omp_atomic_store_set_val): Likewise.
2561 (gimple_omp_atomic_store_val): Likewise.
2562 (gimple_omp_atomic_store_val_ptr): Likewise.
2563 (gimple_omp_atomic_load_set_lhs): Likewise.
2564 (gimple_omp_atomic_load_lhs): Likewise.
2565 (gimple_omp_atomic_load_lhs_ptr): Likewise.
2566 (gimple_omp_atomic_load_set_rhs): Likewise.
2567 (gimple_omp_atomic_load_rhs): Likewise.
2568 (gimple_omp_atomic_load_rhs_ptr): Likewise.
2569 (gimple_omp_continue_control_def): Likewise.
2570 (gimple_omp_continue_control_def_ptr): Likewise.
2571 (gimple_omp_continue_set_control_def): Likewise.
2572 (gimple_omp_continue_control_use): Likewise.
2573 (gimple_omp_continue_control_use_ptr): Likewise.
2574 (gimple_omp_continue_set_control_use): Likewise.
2575 (gimple_transaction_body_ptr): Likewise.
2576 (gimple_transaction_label): Likewise.
2577 (gimple_transaction_label_ptr): Likewise.
2578 (gimple_transaction_set_body): Likewise.
2579 (gimple_transaction_set_label): Likewise.
2581 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
2582 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2583 * ipa-ref.c (ipa_record_reference): Likewise.
2584 * ipa-reference.c (analyze_function): Likewise.
2585 (ipa_reference_write_optimization_summary): Likewise.
2586 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
2587 (address_taken_from_non_vtable_p): Likewise.
2588 (comdat_can_be_unshared_p_1): Likewise.
2589 * lto-cgraph.c (lto_output_ref): Likewise.
2590 (add_references): Likewise.
2591 (compute_ltrans_boundary): Likewise.
2592 (output_symtab): Likewise.
2593 (input_ref): Likewise.
2594 (input_cgraph_1): Likewise.
2595 (output_cgraph_opt_summary): Likewise.
2596 * lto-streamer-out.c (lto_output): Likewise.
2597 (output_symbol_p): Likewise.
2598 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
2599 (lsei_start_function_in_partition): Likewise.
2600 (lsei_next_variable_in_partition): Likewise.
2601 (lsei_start_variable_in_partition): Likewise.
2602 * symtab.c (insert_to_assembler_name_hash): Likewise.
2603 (unlink_from_assembler_name_hash): Likewise.
2604 (symtab_unregister_node): Likewise.
2605 (symtab_remove_node): Likewise.
2606 (dump_symtab_node): Likewise.
2607 (verify_symtab_base): Likewise.
2608 (verify_symtab_node): Likewise.
2609 (symtab_make_decl_local): Likewise.
2610 (symtab_alias_ultimate_target): Likewise.
2611 (symtab_resolve_alias): Likewise.
2612 (symtab_get_symbol_partitioning_class): Likewise.
2613 * tree-phinodes.c (allocate_phi_node): Likewise.
2614 (reserve_phi_args_for_new_edge): Likewise.
2615 (remove_phi_args): Likewise.
2616 * varpool.c (varpool_node_for_asm): Likewise.
2617 (varpool_remove_unreferenced_decls): Likewise.
2619 2014-04-23 Jeff Law <law@redhat.com>
2621 PR tree-optimization/60902
2622 * tree-ssa-threadedge.c
2623 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
2624 invalidate outputs from statements that do not produce useful
2625 outputs for threading.
2627 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2629 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
2630 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
2631 machine descriptions for Stack Smashing Protector.
2633 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
2635 * aarch64.md (<optab>_rol<mode>3): New pattern.
2636 (<optab>_rolsi3_uxtw): Likewise.
2637 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
2639 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
2641 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
2642 (arm_cortex_a12_tune): Likewise.
2644 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2646 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
2648 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2650 * config/arm/arm.md (arm_rev16si2): New pattern.
2651 (arm_rev16si2_alt): Likewise.
2652 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
2654 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2656 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
2657 (rev16<mode>2_alt): Likewise.
2658 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
2659 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
2660 (aarch_rev16_shleft_mask_imm_p): Likewise.
2661 (aarch_rev16_p_1): Likewise.
2662 (aarch_rev16_p): Likewise.
2663 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
2664 (aarch_rev16_shright_mask_imm_p): Likewise.
2665 (aarch_rev16_shleft_mask_imm_p): Likewise.
2667 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2669 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
2670 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
2672 (cortex_a53_extra_costs): Likewise.
2673 (cortex_a57_extra_costs): Likewise.
2674 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
2675 (cortexa7_extra_costs): Likewise.
2676 (cortexa8_extra_costs): Likewise.
2677 (cortexa12_extra_costs): Likewise.
2678 (cortexa15_extra_costs): Likewise.
2679 (v7m_extra_costs): Likewise.
2680 (arm_new_rtx_costs): Handle BSWAP.
2682 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2684 * config/arm/arm.c (cortexa8_extra_costs): New table.
2685 (arm_cortex_a8_tune): New tuning struct.
2686 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
2688 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2690 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
2692 2014-04-23 Richard Biener <rguenther@suse.de>
2694 * Makefile.in (OBJS): Remove loop-unswitch.o.
2695 * tree-pass.h (make_pass_rtl_unswitch): Remove.
2696 * passes.def (pass_rtl_unswitch): Likewise.
2697 * loop-init.c (gate_rtl_unswitch): Likewise.
2698 (rtl_unswitch): Likewise.
2699 (pass_data_rtl_unswitch): Likewise.
2700 (pass_rtl_unswitch): Likewise.
2701 (make_pass_rtl_unswitch): Likewise.
2702 * rtl.h (reversed_condition): Likewise.
2703 (compare_and_jump_seq): Likewise.
2704 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
2706 * loop-unroll.c (compare_and_jump_seq): Likewise.
2708 2014-04-23 Richard Biener <rguenther@suse.de>
2710 PR tree-optimization/60903
2711 * tree-ssa-loop-im.c (analyze_memory_references): Remove
2712 commented code block.
2713 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
2714 loop flags to newly created BBs and edges.
2716 2014-04-23 Nick Clifton <nickc@redhat.com>
2718 * config/msp430/msp430.c (msp430_handle_option): Move function
2720 (msp430_option_override): Simplify mcu and mcpu option handling.
2721 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
2722 support for -mhwmult command line option.
2723 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
2724 -mhwmult command line option.
2725 (msp430_hwmult_enabled): Delete.
2726 (msp43o_output_labelref): Add support for -mhwmult command line option.
2727 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
2728 (umulsidi3): Likewise.
2729 * config/msp430/msp430.opt (mmcu): Add Report attribute.
2730 (mcpu, mlarge, msmall): Likewise.
2731 (mhwmult): New option.
2732 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
2734 (msp430_is_f5_mcu): Remove prototype.
2735 (msp430_use_f5_series_hwmult): Add prototype.
2736 * config/msp430/msp430-opts.h: New file.
2737 * common/config/msp430: New directory.
2738 * common/config/msp430/msp430-common.c: New file.
2739 * config.gcc (msp430): Remove target_has_targetm_common.
2740 * doc/invoke.texi: Document -mhwmult command line option.
2742 2014-04-23 Nick Clifton <nickc@redhat.com>
2744 * config/i386/cygwin.h (ENDFILE_SPEC): Include
2745 default-manifest.o if it can be found in the search path.
2746 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
2748 2014-04-23 Terry Guo <terry.guo@arm.com>
2750 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
2752 2014-04-23 Richard Biener <rguenther@suse.de>
2755 * tree-inline.c (declare_return_variable): Use mark_addressable.
2757 2014-04-23 Richard Biener <rguenther@suse.de>
2760 * loop-init.c (loop_optimizer_init): Make sure to apply
2761 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
2763 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2766 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
2768 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
2769 if flag_sanitize_undefined_trap_on_error.
2770 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
2771 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
2772 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
2773 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
2774 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
2775 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
2776 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
2777 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
2778 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
2779 * ubsan.c (ubsan_instrument_unreachable): Return
2780 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
2781 (ubsan_expand_null_ifn): Emit __builtin_trap ()
2782 if flag_sanitize_undefined_trap_on_error and
2783 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
2784 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
2785 instrument_bool_enum_load): Emit __builtin_trap () if
2786 flag_sanitize_undefined_trap_on_error and
2787 __builtin_handle_*_abort () if !flag_sanitize_recover.
2788 * doc/invoke.texi (-fsanitize-recover,
2789 -fsanitize-undefined-trap-on-error): Document.
2791 2014-04-22 Christian Bruel <christian.bruel@st.com>
2793 * config/sh/sh.md (mov<mode>): Replace movQIHI.
2794 Force immediates to SImode.
2796 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
2798 * config/nios2/nios2.md (UNSPEC_ROUND): New.
2800 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
2801 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
2802 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
2803 (nios2_fpu_insn): Add entry for round.
2804 (N2FPU_NO_ERRNO_P): Define.
2805 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
2807 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
2809 2014-04-22 Richard Henderson <rth@redhat.com>
2811 * config/aarch64/aarch64 (addti3, subti3): New expanders.
2812 (add<GPI>3_compare0): Remove leading * from name.
2813 (add<GPI>3_carryin): Likewise.
2814 (sub<GPI>3_compare0): Likewise.
2815 (sub<GPI>3_carryin): Likewise.
2816 (<su_optab>mulditi3): New expander.
2817 (multi3): New expander.
2818 (madd<GPI>): Remove leading * from name.
2820 2014-04-22 Martin Jambor <mjambor@suse.cz>
2822 * cgraphclones.c (cgraph_function_versioning): Copy
2823 ipa_transforms_to_apply instead of asserting it is empty.
2825 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
2828 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
2829 on count_exp to get mode.
2831 2014-04-22 Andrew Pinski <apinski@cavium.com>
2833 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
2834 Handle TLS for ILP32.
2835 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
2836 (tlsie_small_<mode>): this and handle PTR.
2837 (tlsie_small_sidi): New pattern.
2838 (tlsle_small): Change to an expand to handle ILP32.
2839 (tlsle_small_<mode>): New pattern.
2840 (tlsdesc_small): Rename to ...
2841 (tlsdesc_small_<mode>): this and handle PTR.
2843 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2845 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
2847 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2849 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2850 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
2851 (aarch64_types_signed_poly_qualifiers): Likewise.
2852 (aarch64_types_unsigned_signed_qualifiers): Likewise.
2853 (aarch64_types_poly_signed_qualifiers): Likewise.
2854 (TYPES_REINTERP_SS): Type macro added.
2855 (TYPES_REINTERP_SU): Likewise.
2856 (TYPES_REINTERP_SP): Likewise.
2857 (TYPES_REINTERP_US): Likewise.
2858 (TYPES_REINTERP_PS): Likewise.
2859 (aarch64_fold_builtin): New expression folding added.
2860 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
2861 Declarations removed.
2862 (REINTERP_SS): Declarations added.
2863 (REINTERP_US): Likewise.
2864 (REINTERP_PS): Likewise.
2865 (REINTERP_SU): Likewise.
2866 (REINTERP_SP): Likewise.
2867 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
2868 (vreinterpretq_p8_f64): Likewise.
2869 (vreinterpret_p16_f64): Likewise.
2870 (vreinterpretq_p16_f64): Likewise.
2871 (vreinterpret_f32_f64): Likewise.
2872 (vreinterpretq_f32_f64): Likewise.
2873 (vreinterpret_f64_f32): Likewise.
2874 (vreinterpret_f64_p8): Likewise.
2875 (vreinterpret_f64_p16): Likewise.
2876 (vreinterpret_f64_s8): Likewise.
2877 (vreinterpret_f64_s16): Likewise.
2878 (vreinterpret_f64_s32): Likewise.
2879 (vreinterpret_f64_s64): Likewise.
2880 (vreinterpret_f64_u8): Likewise.
2881 (vreinterpret_f64_u16): Likewise.
2882 (vreinterpret_f64_u32): Likewise.
2883 (vreinterpret_f64_u64): Likewise.
2884 (vreinterpretq_f64_f32): Likewise.
2885 (vreinterpretq_f64_p8): Likewise.
2886 (vreinterpretq_f64_p16): Likewise.
2887 (vreinterpretq_f64_s8): Likewise.
2888 (vreinterpretq_f64_s16): Likewise.
2889 (vreinterpretq_f64_s32): Likewise.
2890 (vreinterpretq_f64_s64): Likewise.
2891 (vreinterpretq_f64_u8): Likewise.
2892 (vreinterpretq_f64_u16): Likewise.
2893 (vreinterpretq_f64_u32): Likewise.
2894 (vreinterpretq_f64_u64): Likewise.
2895 (vreinterpret_s64_f64): Likewise.
2896 (vreinterpretq_s64_f64): Likewise.
2897 (vreinterpret_u64_f64): Likewise.
2898 (vreinterpretq_u64_f64): Likewise.
2899 (vreinterpret_s8_f64): Likewise.
2900 (vreinterpretq_s8_f64): Likewise.
2901 (vreinterpret_s16_f64): Likewise.
2902 (vreinterpretq_s16_f64): Likewise.
2903 (vreinterpret_s32_f64): Likewise.
2904 (vreinterpretq_s32_f64): Likewise.
2905 (vreinterpret_u8_f64): Likewise.
2906 (vreinterpretq_u8_f64): Likewise.
2907 (vreinterpret_u16_f64): Likewise.
2908 (vreinterpretq_u16_f64): Likewise.
2909 (vreinterpret_u32_f64): Likewise.
2910 (vreinterpretq_u32_f64): Likewise.
2912 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2914 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2915 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
2916 (vreinterpret_p8_s8): Likewise.
2917 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
2918 (vreinterpret_p8_s16): Likewise.
2919 (vreinterpret_p8_s32): Likewise.
2920 (vreinterpret_p8_s64): Likewise.
2921 (vreinterpret_p8_f32): Likewise.
2922 (vreinterpret_p8_u8): Likewise.
2923 (vreinterpret_p8_u16): Likewise.
2924 (vreinterpret_p8_u32): Likewise.
2925 (vreinterpret_p8_u64): Likewise.
2926 (vreinterpret_p8_p16): Likewise.
2927 (vreinterpretq_p8_s8): Likewise.
2928 (vreinterpretq_p8_s16): Likewise.
2929 (vreinterpretq_p8_s32): Likewise.
2930 (vreinterpretq_p8_s64): Likewise.
2931 (vreinterpretq_p8_f32): Likewise.
2932 (vreinterpretq_p8_u8): Likewise.
2933 (vreinterpretq_p8_u16): Likewise.
2934 (vreinterpretq_p8_u32): Likewise.
2935 (vreinterpretq_p8_u64): Likewise.
2936 (vreinterpretq_p8_p16): Likewise.
2937 (vreinterpret_p16_s8): Likewise.
2938 (vreinterpret_p16_s16): Likewise.
2939 (vreinterpret_p16_s32): Likewise.
2940 (vreinterpret_p16_s64): Likewise.
2941 (vreinterpret_p16_f32): Likewise.
2942 (vreinterpret_p16_u8): Likewise.
2943 (vreinterpret_p16_u16): Likewise.
2944 (vreinterpret_p16_u32): Likewise.
2945 (vreinterpret_p16_u64): Likewise.
2946 (vreinterpret_p16_p8): Likewise.
2947 (vreinterpretq_p16_s8): Likewise.
2948 (vreinterpretq_p16_s16): Likewise.
2949 (vreinterpretq_p16_s32): Likewise.
2950 (vreinterpretq_p16_s64): Likewise.
2951 (vreinterpretq_p16_f32): Likewise.
2952 (vreinterpretq_p16_u8): Likewise.
2953 (vreinterpretq_p16_u16): Likewise.
2954 (vreinterpretq_p16_u32): Likewise.
2955 (vreinterpretq_p16_u64): Likewise.
2956 (vreinterpretq_p16_p8): Likewise.
2957 (vreinterpret_f32_s8): Likewise.
2958 (vreinterpret_f32_s16): Likewise.
2959 (vreinterpret_f32_s32): Likewise.
2960 (vreinterpret_f32_s64): Likewise.
2961 (vreinterpret_f32_u8): Likewise.
2962 (vreinterpret_f32_u16): Likewise.
2963 (vreinterpret_f32_u32): Likewise.
2964 (vreinterpret_f32_u64): Likewise.
2965 (vreinterpret_f32_p8): Likewise.
2966 (vreinterpret_f32_p16): Likewise.
2967 (vreinterpretq_f32_s8): Likewise.
2968 (vreinterpretq_f32_s16): Likewise.
2969 (vreinterpretq_f32_s32): Likewise.
2970 (vreinterpretq_f32_s64): Likewise.
2971 (vreinterpretq_f32_u8): Likewise.
2972 (vreinterpretq_f32_u16): Likewise.
2973 (vreinterpretq_f32_u32): Likewise.
2974 (vreinterpretq_f32_u64): Likewise.
2975 (vreinterpretq_f32_p8): Likewise.
2976 (vreinterpretq_f32_p16): Likewise.
2977 (vreinterpret_s64_s8): Likewise.
2978 (vreinterpret_s64_s16): Likewise.
2979 (vreinterpret_s64_s32): Likewise.
2980 (vreinterpret_s64_f32): Likewise.
2981 (vreinterpret_s64_u8): Likewise.
2982 (vreinterpret_s64_u16): Likewise.
2983 (vreinterpret_s64_u32): Likewise.
2984 (vreinterpret_s64_u64): Likewise.
2985 (vreinterpret_s64_p8): Likewise.
2986 (vreinterpret_s64_p16): Likewise.
2987 (vreinterpretq_s64_s8): Likewise.
2988 (vreinterpretq_s64_s16): Likewise.
2989 (vreinterpretq_s64_s32): Likewise.
2990 (vreinterpretq_s64_f32): Likewise.
2991 (vreinterpretq_s64_u8): Likewise.
2992 (vreinterpretq_s64_u16): Likewise.
2993 (vreinterpretq_s64_u32): Likewise.
2994 (vreinterpretq_s64_u64): Likewise.
2995 (vreinterpretq_s64_p8): Likewise.
2996 (vreinterpretq_s64_p16): Likewise.
2997 (vreinterpret_u64_s8): Likewise.
2998 (vreinterpret_u64_s16): Likewise.
2999 (vreinterpret_u64_s32): Likewise.
3000 (vreinterpret_u64_s64): Likewise.
3001 (vreinterpret_u64_f32): Likewise.
3002 (vreinterpret_u64_u8): Likewise.
3003 (vreinterpret_u64_u16): Likewise.
3004 (vreinterpret_u64_u32): Likewise.
3005 (vreinterpret_u64_p8): Likewise.
3006 (vreinterpret_u64_p16): Likewise.
3007 (vreinterpretq_u64_s8): Likewise.
3008 (vreinterpretq_u64_s16): Likewise.
3009 (vreinterpretq_u64_s32): Likewise.
3010 (vreinterpretq_u64_s64): Likewise.
3011 (vreinterpretq_u64_f32): Likewise.
3012 (vreinterpretq_u64_u8): Likewise.
3013 (vreinterpretq_u64_u16): Likewise.
3014 (vreinterpretq_u64_u32): Likewise.
3015 (vreinterpretq_u64_p8): Likewise.
3016 (vreinterpretq_u64_p16): Likewise.
3017 (vreinterpret_s8_s16): Likewise.
3018 (vreinterpret_s8_s32): Likewise.
3019 (vreinterpret_s8_s64): Likewise.
3020 (vreinterpret_s8_f32): Likewise.
3021 (vreinterpret_s8_u8): Likewise.
3022 (vreinterpret_s8_u16): Likewise.
3023 (vreinterpret_s8_u32): Likewise.
3024 (vreinterpret_s8_u64): Likewise.
3025 (vreinterpret_s8_p8): Likewise.
3026 (vreinterpret_s8_p16): Likewise.
3027 (vreinterpretq_s8_s16): Likewise.
3028 (vreinterpretq_s8_s32): Likewise.
3029 (vreinterpretq_s8_s64): Likewise.
3030 (vreinterpretq_s8_f32): Likewise.
3031 (vreinterpretq_s8_u8): Likewise.
3032 (vreinterpretq_s8_u16): Likewise.
3033 (vreinterpretq_s8_u32): Likewise.
3034 (vreinterpretq_s8_u64): Likewise.
3035 (vreinterpretq_s8_p8): Likewise.
3036 (vreinterpretq_s8_p16): Likewise.
3037 (vreinterpret_s16_s8): Likewise.
3038 (vreinterpret_s16_s32): Likewise.
3039 (vreinterpret_s16_s64): Likewise.
3040 (vreinterpret_s16_f32): Likewise.
3041 (vreinterpret_s16_u8): Likewise.
3042 (vreinterpret_s16_u16): Likewise.
3043 (vreinterpret_s16_u32): Likewise.
3044 (vreinterpret_s16_u64): Likewise.
3045 (vreinterpret_s16_p8): Likewise.
3046 (vreinterpret_s16_p16): Likewise.
3047 (vreinterpretq_s16_s8): Likewise.
3048 (vreinterpretq_s16_s32): Likewise.
3049 (vreinterpretq_s16_s64): Likewise.
3050 (vreinterpretq_s16_f32): Likewise.
3051 (vreinterpretq_s16_u8): Likewise.
3052 (vreinterpretq_s16_u16): Likewise.
3053 (vreinterpretq_s16_u32): Likewise.
3054 (vreinterpretq_s16_u64): Likewise.
3055 (vreinterpretq_s16_p8): Likewise.
3056 (vreinterpretq_s16_p16): Likewise.
3057 (vreinterpret_s32_s8): Likewise.
3058 (vreinterpret_s32_s16): Likewise.
3059 (vreinterpret_s32_s64): Likewise.
3060 (vreinterpret_s32_f32): Likewise.
3061 (vreinterpret_s32_u8): Likewise.
3062 (vreinterpret_s32_u16): Likewise.
3063 (vreinterpret_s32_u32): Likewise.
3064 (vreinterpret_s32_u64): Likewise.
3065 (vreinterpret_s32_p8): Likewise.
3066 (vreinterpret_s32_p16): Likewise.
3067 (vreinterpretq_s32_s8): Likewise.
3068 (vreinterpretq_s32_s16): Likewise.
3069 (vreinterpretq_s32_s64): Likewise.
3070 (vreinterpretq_s32_f32): Likewise.
3071 (vreinterpretq_s32_u8): Likewise.
3072 (vreinterpretq_s32_u16): Likewise.
3073 (vreinterpretq_s32_u32): Likewise.
3074 (vreinterpretq_s32_u64): Likewise.
3075 (vreinterpretq_s32_p8): Likewise.
3076 (vreinterpretq_s32_p16): Likewise.
3077 (vreinterpret_u8_s8): Likewise.
3078 (vreinterpret_u8_s16): Likewise.
3079 (vreinterpret_u8_s32): Likewise.
3080 (vreinterpret_u8_s64): Likewise.
3081 (vreinterpret_u8_f32): Likewise.
3082 (vreinterpret_u8_u16): Likewise.
3083 (vreinterpret_u8_u32): Likewise.
3084 (vreinterpret_u8_u64): Likewise.
3085 (vreinterpret_u8_p8): Likewise.
3086 (vreinterpret_u8_p16): Likewise.
3087 (vreinterpretq_u8_s8): Likewise.
3088 (vreinterpretq_u8_s16): Likewise.
3089 (vreinterpretq_u8_s32): Likewise.
3090 (vreinterpretq_u8_s64): Likewise.
3091 (vreinterpretq_u8_f32): Likewise.
3092 (vreinterpretq_u8_u16): Likewise.
3093 (vreinterpretq_u8_u32): Likewise.
3094 (vreinterpretq_u8_u64): Likewise.
3095 (vreinterpretq_u8_p8): Likewise.
3096 (vreinterpretq_u8_p16): Likewise.
3097 (vreinterpret_u16_s8): Likewise.
3098 (vreinterpret_u16_s16): Likewise.
3099 (vreinterpret_u16_s32): Likewise.
3100 (vreinterpret_u16_s64): Likewise.
3101 (vreinterpret_u16_f32): Likewise.
3102 (vreinterpret_u16_u8): Likewise.
3103 (vreinterpret_u16_u32): Likewise.
3104 (vreinterpret_u16_u64): Likewise.
3105 (vreinterpret_u16_p8): Likewise.
3106 (vreinterpret_u16_p16): Likewise.
3107 (vreinterpretq_u16_s8): Likewise.
3108 (vreinterpretq_u16_s16): Likewise.
3109 (vreinterpretq_u16_s32): Likewise.
3110 (vreinterpretq_u16_s64): Likewise.
3111 (vreinterpretq_u16_f32): Likewise.
3112 (vreinterpretq_u16_u8): Likewise.
3113 (vreinterpretq_u16_u32): Likewise.
3114 (vreinterpretq_u16_u64): Likewise.
3115 (vreinterpretq_u16_p8): Likewise.
3116 (vreinterpretq_u16_p16): Likewise.
3117 (vreinterpret_u32_s8): Likewise.
3118 (vreinterpret_u32_s16): Likewise.
3119 (vreinterpret_u32_s32): Likewise.
3120 (vreinterpret_u32_s64): Likewise.
3121 (vreinterpret_u32_f32): Likewise.
3122 (vreinterpret_u32_u8): Likewise.
3123 (vreinterpret_u32_u16): Likewise.
3124 (vreinterpret_u32_u64): Likewise.
3125 (vreinterpret_u32_p8): Likewise.
3126 (vreinterpret_u32_p16): Likewise.
3127 (vreinterpretq_u32_s8): Likewise.
3128 (vreinterpretq_u32_s16): Likewise.
3129 (vreinterpretq_u32_s32): Likewise.
3130 (vreinterpretq_u32_s64): Likewise.
3131 (vreinterpretq_u32_f32): Likewise.
3132 (vreinterpretq_u32_u8): Likewise.
3133 (vreinterpretq_u32_u16): Likewise.
3134 (vreinterpretq_u32_u64): Likewise.
3135 (vreinterpretq_u32_p8): Likewise.
3136 (vreinterpretq_u32_p16): Likewise.
3138 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3140 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
3142 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
3144 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
3145 (vqnegd_s64): Likewise.
3146 (vqabs_s64): Likewise.
3147 (vqabsd_s64): Likewise.
3149 2014-04-22 Richard Henderson <rth@redhat.com>
3151 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
3152 computation to the top of the loop.
3154 2014-04-22 Renlin <renlin.li@arm.com>
3155 Jiong Wang <jiong.wang@arm.com>
3157 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
3158 * config/aarch64/aarch64.c (aarch64_layout_frame)
3159 (aarch64_initial_elimination_offset): Likewise.
3161 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
3163 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
3166 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
3168 * machmode.h (bitwise_mode_for_mode): Declare.
3169 * stor-layout.h (bitwise_type_for_mode): Likewise.
3170 * stor-layout.c (bitwise_mode_for_mode): New function.
3171 (bitwise_type_for_mode): Likewise.
3172 * builtins.c (fold_builtin_memory_op): Use it instead of
3173 int_mode_for_mode and build_nonstandard_integer_type.
3175 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3177 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
3178 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
3179 (*-*-solaris2*): Simplify.
3180 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
3181 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
3182 *-*-solaris2.9* handling.
3184 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
3186 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
3187 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
3189 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
3190 * configure: Regenerate.
3192 * config/i386/sol2-9.h: Remove.
3194 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
3195 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
3196 Remove Solaris 9 references.
3198 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
3200 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
3201 (floatuns<GPI:mode><GPF:mode>2): Remove.
3202 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
3203 and floatuns conversions.
3204 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
3205 and floatuns conversions.
3206 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
3207 (w1,w2): New mode attributes for inequal width conversions.
3209 2014-04-22 Renlin Li <Renlin.Li@arm.com>
3211 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
3212 the output asm format.
3214 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
3216 * config/aarch64/aarch64-simd.md
3217 (aarch64_cm<optab>di): Always split.
3218 (*aarch64_cm<optab>di): New.
3219 (aarch64_cmtstdi): Always split.
3220 (*aarch64_cmtstdi): New.
3222 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3224 PR tree-optimization/60823
3225 * omp-low.c (ipa_simd_modify_function_body): Go through
3226 all SSA_NAMEs and for those refering to vector arguments
3227 which are going to be replaced adjust SSA_NAME_VAR and,
3228 if it is a default definition, change it into a non-default
3229 definition assigned at the beginning of function from new_decl.
3230 (ipa_simd_modify_stmt_ops): Rewritten.
3231 * tree-dfa.c (set_ssa_default_def): When removing default def,
3232 check for NULL loc instead of NULL *loc.
3234 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3236 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
3237 restrictions on core registers for DImode values in Thumb2.
3239 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3241 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
3242 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
3244 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3246 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
3247 (*iordi_notzesidi_di): Likewise.
3248 (*iordi_notsesidi_di): Likewise.
3250 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3252 * config/arm/arm-protos.h (tune_params): New struct members.
3253 * config/arm/arm.c: Initialise tune_params per processor.
3254 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
3255 for speed, based on new tune_params.
3257 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3259 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
3260 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
3261 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
3262 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
3263 * config/aarch64/arm_neon.h (vrnd_f64): Added.
3264 (vrnda_f64): Likewise.
3265 (vrndi_f64): Likewise.
3266 (vrndm_f64): Likewise.
3267 (vrndn_f64): Likewise.
3268 (vrndp_f64): Likewise.
3269 (vrndx_f64): Likewise.
3271 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3273 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
3274 GET_MODE_SIZE argument is enum machine_mode.
3276 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3279 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
3280 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
3282 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
3285 * asan.c (asan_emit_stack_protection): Force the base to align to
3286 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
3287 appropriate bits if STRICT_ALIGNMENT.
3288 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
3290 (expand_used_vars): Leave a space in the stack frame for alignment
3291 if STRICT_ALIGNMENT.
3293 2014-04-21 David Malcolm <dmalcolm@redhat.com>
3295 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
3297 (gimple_store_p): Likewise.
3298 (gimple_assign_load_p): Likewise.
3299 (gimple_assign_cast_p): Likewise.
3300 (gimple_clobber_p): Likewise.
3302 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
3303 rather than a gimple.
3304 (gimple_assign_cast_p): Likewise.
3306 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
3309 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
3310 If mode is DDmode and TARGET_E500_DOUBLE allow move.
3312 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
3313 more debug information for E500 if -mdebug=reg.
3315 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
3318 * config/i386/i386.c (ix86_expand_builtin)
3319 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
3320 register for target RTX.
3321 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
3323 2014-04-18 Cong Hou <congh@google.com>
3325 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
3326 the widen-mult pattern by handling two operands with different sizes,
3327 and operands whose size is smaller than half of the result type.
3329 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3331 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
3332 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
3333 (do_estimate_edge_time): Compute it.
3334 * ipa-inline.c (want_inline_small_function_p): Bypass
3335 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
3337 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3339 * ipa-inline.c (spec_rem): New static variable.
3340 (dump_overall_stats): New function.
3341 (dump_inline_stats): New function.
3343 2014-04-18 Richard Henderson <rth@redhat.com>
3345 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
3346 to GET_MODE_SIZE, not a reg_class_t.
3348 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3350 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
3351 (vsx_xxmrglw_<mode>): Likewise.
3353 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3356 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
3357 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
3358 (rs6000_init_hard_regno_mode_ok): Likewise.
3360 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
3362 * ipa-inline.c (inline_small_functions): Account only non-cold
3364 * doc/invoke.texi (inline-unit-growth): Update documentation.
3366 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
3368 * config/rs6000/rs6000.md (addti3, subti3): New.
3370 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
3373 * config/i386/i386.c (ix86_expand_clear): Remove outdated
3374 comment. Check optimize_insn_for_size_p instead of
3375 optimize_insn_for_speed_p.
3377 2014-04-17 Martin Jambor <mjambor@suse.cz>
3379 * gimple-iterator.c (gsi_start_edge): New function.
3380 * gimple-iterator.h (gsi_start_edge): Declare.
3381 * tree-sra.c (single_non_eh_succ): New function.
3382 (disqualify_ops_if_throwing_stmt): Renamed to
3383 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
3384 having one non-EH successor BB.
3385 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
3386 generate loads into replacements.
3387 (sra_modify_assign): Likewise and and also use the simple path for
3389 (sra_modify_function_body): Commit statements on edges.
3391 2014-04-17 Richard Biener <rguenther@suse.de>
3394 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
3395 comparison results and add clarifying comment.
3397 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3399 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
3400 (blank_mode): Initialize it.
3401 (emit_mode_size_inline, emit_mode_nunits_inline,
3402 emit_mode_inner_inline): New functions.
3403 (emit_insn_modes_h): Call them and surround their output with
3404 #if GCC_VERSION >= 4001 ... #endif.
3405 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
3406 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
3407 mode_* arrays if the argument is __builtin_constant_p.
3408 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
3409 is enum machine_mode.
3411 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3413 * passes.c (opt_pass::execute): Adjust.
3414 (pass_manager::execute_pass_mode_switching): Likewise.
3415 (early_local_passes::execute): Likewise.
3416 (execute_one_pass): Pass cfun to the pass's execute method.
3417 * tree-pass.h (opt_pass::execute): Add function * argument.
3418 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3419 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3420 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3421 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3422 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3423 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
3424 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
3425 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3426 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3427 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3428 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3429 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3430 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3431 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3432 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3433 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3434 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3435 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3436 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3437 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3438 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3439 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3440 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3441 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3442 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3443 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3444 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3445 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3446 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3449 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3451 * passes.c (opt_pass::gate): Take function * argument.
3452 (gate_all_early_local_passes): Merge into
3453 (early_local_passes::gate): this.
3454 (gate_all_early_optimizations): Merge into
3455 (all_early_optimizations::gate): this.
3456 (gate_all_optimizations): Mege into
3457 (all_optimizations::gate): this.
3458 (gate_all_optimizations_g): Merge into
3459 (all_optimizations_g::gate): this.
3460 (gate_rest_of_compilation): Mege into
3461 (rest_of_compilation::gate): this.
3462 (gate_postreload): Merge into
3463 (postreload::gate): this.
3464 (dump_one_pass): Pass cfun to the pass's gate method.
3465 (execute_ipa_summary_passes): Likewise.
3466 (execute_one_pass): Likewise.
3467 (ipa_write_summaries_2): Likewise.
3468 (ipa_write_optimization_summaries_1): Likewise.
3469 (ipa_read_summaries_1): Likewise.
3470 (ipa_read_optimization_summaries_1): Likewise.
3471 (execute_ipa_stmt_fixups): Likewise.
3472 * tree-pass.h (opt_pass::gate): Add function * argument.
3473 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
3474 combine-stack-adj.c, combine.c, compare-elim.c,
3475 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3476 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
3477 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
3478 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
3479 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3480 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3481 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3482 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
3483 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
3484 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
3485 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3486 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3487 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
3488 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3489 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3490 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3491 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3492 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3493 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3494 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3495 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3496 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3497 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3498 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
3499 var-tracking.c, vtable-verify.c, web.c: Adjust.
3501 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3503 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
3504 * configure: Regenerate.
3506 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3508 * passes.c (dump_one_pass): don't check pass->has_gate.
3509 (execute_ipa_summary_passes): Likewise.
3510 (execute_one_pass): Likewise.
3511 (ipa_write_summaries_2): Likewise.
3512 (ipa_write_optimization_summaries_1): Likewise.
3513 (ipa_read_optimization_summaries_1): Likewise.
3514 (execute_ipa_stmt_fixups): Likewise.
3515 * tree-pass.h (pass_data::has_gate): Remove.
3516 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3517 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3518 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3519 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3520 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3521 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
3522 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
3523 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
3524 gimple-low.c, gimple-ssa-isolate-paths.c,
3525 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
3526 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3527 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3528 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
3529 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
3530 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
3531 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
3532 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3533 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
3534 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3535 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3536 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3537 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3538 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3539 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3540 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3541 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3542 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3543 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3544 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3545 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3546 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3549 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3551 * pass_manager.h (pass_manager::register_dump_files_1): Remove
3553 * passes.c (pass_manager::register_dump_files_1): Merge into
3554 (pass_manager::register_dump_files): this, and remove its handling of
3555 properties since the pass always has the properties anyway.
3556 (pass_manager::pass_manager): Adjust.
3558 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3560 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
3561 * passes.c (pass_manager::register_dump_files_1): Remove dead code
3562 dealing with properties.
3563 (pass_manager::register_dump_files): Adjust.
3565 2014-03-20 Mark Wielaard <mjw@redhat.com>
3567 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
3568 then represent the bound as normal constant value.
3570 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3573 Forward port from 4.8 branch
3574 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
3576 * config/i386/bmiintrin.h (_blsi_u32): New.
3580 (_blsmsk_u32): Ditto.
3581 (_blsmsk_u64): Ditto.
3582 (_tzcnt_u32): Ditto.
3583 (_tzcnt_u64): Ditto.
3585 2014-04-17 Kito Cheng <kito@0xlab.org>
3587 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
3589 2014-04-17 Richard Biener <rguenther@suse.de>
3592 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
3593 boolean results for comparisons.
3595 2014-04-17 Richard Biener <rguenther@suse.de>
3597 PR tree-optimization/60836
3598 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
3599 initial PHI args to be gimple values.
3601 2014-04-17 Richard Biener <rguenther@suse.de>
3603 PR tree-optimization/60841
3604 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
3605 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
3606 of stmts to SLP build.
3607 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
3608 (vect_analyze_slp): Likewise.
3609 (vect_analyze_slp_instance): Likewise.
3610 (vect_build_slp_tree): Limit overall SLP tree growth.
3611 * tree-vectorizer.h (vect_analyze_data_refs,
3612 vect_analyze_slp): Adjust prototypes.
3614 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3616 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
3619 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3621 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
3622 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
3623 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
3624 for TARGET_SLOW_PSHUFB
3626 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3628 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
3629 * config/i386/i386.c (intel_cost): Ditto.
3631 2014-04-17 Joey Ye <joey.ye@arm.com>
3633 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
3635 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3637 * opts.c (common_handle_option): Disable -fipa-reference coorectly
3640 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3642 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
3643 (type_all_derivations_known_p): New predicate.
3644 (type_all_ctors_visible_p): New predicate.
3645 (type_possibly_instantiated_p): New predicate.
3646 (get_odr_type): Compute all_derivations_known.
3647 (dump_odr_type): Dump the flag.
3648 (maybe_record_type): Cleanup.
3649 (record_target_from_binfo): Add bases_to_consider array;
3650 record bases for types w/o instances and skip CXX destructor.
3651 (possible_polymorphic_call_targets_1): Add bases_to_consider
3652 and consider_construction parameters; check if type may have instance.
3653 (get_polymorphic_call_info): Set maybe_in_construction to true
3654 when we know nothing.
3655 (record_targets_from_bases): Skip CXX destructors; they are
3656 never called for types in construction.
3657 (possible_polymorphic_call_targets): Do not record target when
3658 type may not have instance.
3660 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3663 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
3664 external aliases alive, too.
3666 2014-04-16 Andrew Pinski <apinski@cavium.com>
3668 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
3671 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3673 * final.c (compute_alignments): Do not apply loop alignment to a block
3674 falling through to the exit.
3676 2014-04-16 Catherine Moore <clm@codesourcery.com>
3678 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
3679 Adjust constraints for microMIPS store patterns.
3681 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3683 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
3685 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3687 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
3688 (append_use): Run at -O0.
3689 (append_vdef): Likewise.
3690 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
3691 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
3693 2014-04-16 Jakub Jelinek <jakub@redhat.com>
3695 PR tree-optimization/60844
3696 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
3697 (propagate_op_to_single_use, remove_visited_stmt_chain,
3698 linearize_expr, repropagate_negates, reassociate_bb): Use it
3699 instead of gsi_remove.
3701 2014-04-16 Martin Jambor <mjambor@suse.cz>
3703 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
3704 ipa_transforms_to_apply.
3705 (cgraph_function_versioning): Assert that old_node has empty
3706 ipa_transforms_to_apply.
3707 * trans-mem.c (ipa_tm_create_version): Likewise.
3708 * tree-inline.c (tree_function_versioning): Do not duplicate
3709 ipa_transforms_to_apply.
3711 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3714 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
3716 Pass necessary as flags on 64-bit Solaris/x86.
3717 Use lowercase relocs for x86_64-*-*.
3718 * configure: Regenerate.
3720 2014-04-15 Jan Hubicka <jh@suse.cz>
3722 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
3723 (maybe_record_node, likely_target_p): Use it.
3725 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3728 Revert following patch
3730 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3733 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3734 software floating point or no floating point registers, do not
3735 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3736 in GPRs that occurs after we tested for GPRs that would never be
3739 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3740 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3741 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3742 specifically allow DDmode, since that does not use the SPE SIMD
3745 2014-03-21 Mark Wielaard <mjw@redhat.com>
3747 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
3748 as unsigned or int depending on type and value used.
3750 2014-04-15 Richard Biener <rguenther@suse.de>
3752 PR rtl-optimization/56965
3753 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
3754 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
3756 * alias.c (true_dependence_1): Do not call
3757 nonoverlapping_component_refs_p.
3758 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
3759 nonoverlapping_component_refs_p.
3760 (indirect_refs_may_alias_p): Likewise.
3762 2014-04-15 Teresa Johnson <tejohnson@google.com>
3764 * cfg.c (dump_bb_info): Fix flags check.
3765 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
3767 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3769 PR rtl-optimization/60663
3770 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
3773 2014-04-15 Richard Biener <rguenther@suse.de>
3775 * lto-streamer.h (LTO_major_version): Bump to 4.
3777 2014-04-15 Richard Biener <rguenther@suse.de>
3779 * common.opt (lto_partition_model): New enum.
3780 (flto-partition=): Merge separate options with a single with argument,
3781 add -flto-partition=one support.
3782 * flag-types.h (enum lto_partition_model): Declare.
3783 * opts.c (finish_options): Remove duplicate -flto-partition=
3785 * lto-wrapper.c (run_gcc): Adjust.
3787 2014-04-15 Richard Biener <rguenther@suse.de>
3789 * alias.c (ncr_compar): New function.
3790 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
3792 2014-04-15 Richard Biener <rguenther@suse.de>
3794 * alias.c (record_component_aliases): Do not walk BINFOs.
3796 2014-04-15 Richard Biener <rguenther@suse.de>
3798 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
3799 Add struct function argument and adjust.
3800 (find_func_aliases_for_call): Likewise.
3801 (find_func_aliases): Likewise.
3802 (find_func_clobbers): Likewise.
3803 (intra_create_variable_infos): Likewise.
3804 (compute_points_to_sets): Likewise.
3805 (ipa_pta_execute): Adjust. Do not push/pop cfun.
3807 2014-04-15 Richard Biener <rguenther@suse.de>
3809 * tree.c (iterative_hash_expr): Use enum tree_code_class
3810 to store TREE_CODE_CLASS.
3811 (tree_block): Likewise.
3812 (tree_set_block): Likewise.
3813 * tree.h (fold_build_pointer_plus_loc): Use
3814 convert_to_ptrofftype_loc.
3816 2014-04-15 Jakub Jelinek <jakub@redhat.com>
3819 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
3822 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
3824 * cfgloop.h (struct loop): Move force_vectorize down.
3825 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
3826 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
3827 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
3828 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
3829 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
3830 * tree-core.h (enum annot_expr_kind): Add new kind values.
3831 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
3832 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
3834 * tree.def (ANNOTATE_EXPR): Tweak comment.
3836 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3838 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
3841 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
3843 * tree.h (TYPE_IDENTIFIER): Declare.
3844 * tree.c (subrange_type_for_debug_p): Use it.
3845 * godump.c (go_format_type): Likewise.
3846 * dwarf2out.c (is_cxx_auto, modified_type_die,
3847 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
3848 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
3850 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3853 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
3855 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
3857 * config/i386/i386.c (examine_argument): Return bool. Return true if
3858 parameter should be passed in memory.
3859 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
3860 (construct_container): Update calls to examine_argument.
3861 (function_arg_advance_64): Ditto.
3862 (return_in_memory_32): Merge with ix86_return_in_memory.
3863 (return_in_memory_64): Ditto.
3864 (return_in_memory_ms_64): Ditto.
3866 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3868 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
3869 * coverage.c (coverage_compute_profile_id): Handle externally visible
3872 2014-04-14 Martin Jambor <mjambor@suse.cz>
3874 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
3875 DECL_DISREGARD_INLINE_LIMITS functions.
3877 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3880 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
3882 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3885 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
3886 optimize_insn_for_speed_p instead of
3887 optimize_function_for_speed_p.
3889 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
3891 * doc/invoke.texi (free): Document AArch64.
3893 2014-04-14 Richard Biener <rguenther@suse.de>
3895 PR tree-optimization/60042
3896 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
3897 (insert_into_preds_of_block): Do not prevent PHI insertion
3898 for REFERENCE exprs here ...
3899 (eliminate_dom_walker::before_dom_children): ... but prevent
3900 their use here under similar conditions when applied to the
3901 IL after PRE optimizations.
3903 2014-04-14 Richard Biener <rguenther@suse.de>
3905 * passes.def: Move early points-to after early SRA.
3907 2014-04-14 Richard Biener <rguenther@suse.de>
3909 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
3910 check for which sign-changes we allow when forwarding
3911 a converted value into a switch.
3913 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3915 * stor-layout.c (place_field): Finalize non-constant offset for the
3918 2014-04-14 Richard Biener <rguenther@suse.de>
3920 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
3922 (expand_switch_using_bit_tests_p): Likewise.
3923 (process_switch): Compute and pass on speed_p based on the
3925 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
3926 optimize_bb_for_speed_p.
3928 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3930 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
3931 * function.h (struct function): Rename has_force_vect_loops into
3932 has_force_vectorize_loops.
3933 * lto-streamer-in.c (input_cfg): Adjust for renaming.
3934 (input_struct_function_base): Likewise.
3935 * lto-streamer-out.c (output_cfg): Likewise.
3936 (output_struct_function_base): Likewise.
3937 * omp-low.c (expand_omp_simd): Likewise.
3938 * tree-cfg.c (move_sese_region_to_fn): Likewise.
3939 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
3940 (version_loop_for_if_conversion): Likewise.
3941 (tree_if_conversion): Likewise.
3942 (main_tree_if_conversion): Likewise.
3943 (gate_tree_if_conversion): Likewise.
3944 * tree-inline.c (copy_loops): Likewise.
3945 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
3946 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
3947 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
3948 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
3949 * tree-vectorizer.c (vectorize_loops): Likewise.
3950 * tree-vectorizer.h (unlimited_cost_model): Likewise.
3952 2014-04-14 Richard Biener <rguenther@suse.de>
3955 * lto-streamer-out.c (wrap_refs): New function.
3956 (lto_output): Wrap symbol references in global initializes in
3957 type-preserving MEM_REFs.
3959 2014-04-14 Christian Bruel <christian.bruel@st.com>
3961 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
3963 2014-04-14 Christian Bruel <christian.bruel@st.com>
3965 * config/sh/sh.md (setmemqi): New expand pattern.
3966 * config/sh/sh.h (CLEAR_RATIO): Define.
3967 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
3968 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
3970 2014-04-14 Richard Biener <rguenther@suse.de>
3973 * fold-const.c (negate_expr_p): Don't negate directional rounding
3975 (fold_negate_expr): Likewise.
3977 2014-04-14 Richard Biener <rguenther@suse.de>
3979 PR tree-optimization/59817
3980 PR tree-optimization/60453
3981 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
3982 recursion to catch all CHRECs in the scalar evolution and restrict
3983 the predicate for the remains appropriately.
3985 2014-04-12 Catherine Moore <clm@codesourcery.com>
3987 * config/mips/constraints.md: Add new register constraint "kb".
3988 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
3989 (*movhi_internal): Likewise.
3990 (*movqi_internal): Likewise.
3991 * config/mips/mips.h (M16_STORE_REGS): New register class.
3992 (REG_CLASS_NAMES): Add M16_STORE_REGS.
3993 (REG_CLASS_CONTENTS): Likewise.
3994 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
3996 2014-04-11 Tobias Burnus <burnus@net-b.de>
3999 * doc/invoke.texi (-Wformat-signedness): Document it.
4000 (Wformat=2): Mention that this enables -Wformat-signedness.
4002 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4004 * common/config/epiphany/epiphany-common.c
4005 (epiphany_option_optimization_table): Enable section anchors by
4006 default at -O1 or higher.
4007 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
4008 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
4009 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
4010 carries no extra cost.
4011 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
4012 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
4013 * config/epiphany/predicates.md (memclob_operand): New predicate.
4014 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
4015 Use memclob_operand predicate and X constraint for operand 3.
4017 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4019 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
4020 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
4023 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4025 PR rtl-optimization/60651
4026 * mode-switching.c (optimize_mode_switching): Make sure to emit
4027 sets of a lower numbered entity before sets of a higher numbered
4028 entity to a mode of the same or lower priority.
4029 When creating a seginfo for a basic block that starts with a code
4030 label, move the insertion point past the code label.
4031 (new_seginfo): Document and enforce requirement that
4032 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
4033 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
4034 * doc/tm.texi: Regenerate.
4036 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
4039 * config/arc/arc.c (arc_save_restore): Fix assert typo.
4041 2013-04-11 Jakub Jelinek <jakub@redhat.com>
4043 * BASE-VER: Set to 4.10.0.
4045 2014-04-11 Tobias Burnus <burnus@net-b.de>
4048 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
4049 * doc/gcc.texi (Service): Update description in the @menu
4050 * doc/invoke.texi (Option Summary): Remove misplaced and
4053 2014-04-11 Steve Ellcey <sellcey@mips.com>
4054 Jakub Jelinek <jakub@redhat.com>
4057 * expr.c (convert_move): Use emit_store_flag_force instead of
4058 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
4061 2014-04-11 Richard Biener <rguenther@suse.de>
4064 * varasm.c (assemble_alias): Avoid endless error reporting
4065 recursion by setting TREE_ASM_WRITTEN.
4067 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4069 * config/s390/s390.md: Add a splitter for NOT rtx.
4071 2014-04-11 Jakub Jelinek <jakub@redhat.com>
4073 PR rtl-optimization/60663
4074 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
4076 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
4077 Jakub Jelinek <jakub@redhat.com>
4080 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
4081 flag from decl_node to node.
4083 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4086 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
4087 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
4088 ameliorating the cases where it can be.
4090 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
4093 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4095 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4096 (loadsync_<mode>): Change mode.
4097 (load_quadpti, store_quadpti): New.
4098 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4099 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4100 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
4102 2014-04-09 Cong Hou <congh@google.com>
4105 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
4108 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4110 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
4111 instead of vnor to exploit possible fusion opportunity in the
4113 (altivec_expand_vec_perm_const_le): Likewise.
4115 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4117 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4118 (loadsync_<mode>): Change mode.
4119 (load_quadpti, store_quadpti): New.
4120 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4121 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4123 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
4126 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
4127 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
4128 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
4130 2014-04-08 Richard Biener <rguenther@suse.de>
4133 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
4134 a 64bit widest int print double-int similar to on HWI64 hosts.
4136 2014-04-08 Richard Biener <rguenther@suse.de>
4138 PR tree-optimization/60785
4139 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
4140 default defs properly.
4142 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
4144 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
4145 (Weffc++): Likewise.
4147 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
4149 * ipa-devirt.c (maybe_record_node): When node is not recorded,
4150 set completep to false rather than true.
4152 2014-04-07 Douglas B Rupp <rupp@adacore.com>
4155 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
4156 ARM_TARGET2_DWARF_FORMAT.
4158 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
4161 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
4162 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
4165 2014-04-07 Richard Biener <rguenther@suse.de>
4167 PR tree-optimization/60766
4168 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
4169 (may_eliminate_iv): Convert cand_value_at result to desired type.
4171 2014-04-07 Jason Merrill <jason@redhat.com>
4174 * common.opt (-fno-gnu-unique): Add.
4175 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
4177 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4179 * haifa-sched.c: Fix outdated function reference and minor
4180 grammar errors in introductory comment.
4182 2014-04-07 Richard Biener <rguenther@suse.de>
4185 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
4187 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
4189 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
4192 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
4193 size accounting for thunks.
4194 (pa_asm_output_mi_thunk): Use final_start_function() and
4195 final_end_function() to output function start and end directives.
4197 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4199 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
4200 device specific ISA/ feature information. Remove short_sp and
4201 errata_skip ds. Add avr_device_specific_features enum to have device
4203 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
4204 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
4205 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
4206 updated device specific info.
4207 * config/avr/avr-mcus.def: Merge device specific details to
4208 dev_attribute field.
4209 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
4211 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
4212 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
4213 assembler if RMW isa supported by current device.
4214 * config/avr/genmultilib.awk: Update as device info structure changed.
4215 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
4217 2014-04-04 Cong Hou <congh@google.com>
4219 PR tree-optimization/60656
4220 * tree-vect-stmts.c (supportable_widening_operation):
4221 Fix a bug that elements in a vector with vect_used_by_reduction
4222 property are incorrectly reordered when the operation on it is not
4223 consistant with the one in reduction operation.
4225 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
4227 PR rtl-optimization/60155
4228 * gcse.c (record_set_data): New function.
4229 (single_set_gcse): New function.
4230 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
4231 (hoist_code): Likewise.
4232 (get_pressure_class_and_nregs): Likewise.
4234 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
4236 * explow.c (probe_stack_range): Emit a final optimization blockage.
4238 2014-04-04 Anthony Green <green@moxielogic.com>
4240 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
4243 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
4246 * lto-cgraph.c (input_overwrite_node): Check that partitioning
4247 flags are set only during streaming.
4248 * ipa.c (process_references, walk_polymorphic_call_targets,
4249 symtab_remove_unreachable_nodes): Drop bodies of always inline
4250 after early inlining.
4251 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
4253 2014-04-04 Jakub Jelinek <jakub@redhat.com>
4254 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4257 * dwarf2out.c (const_ok_for_output_1): Reject expressions
4260 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4263 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
4265 (cortex_a53_fdivd): Likewise.
4267 2014-04-04 Martin Jambor <mjambor@suse.cz>
4270 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
4272 * cgraph.c (clone_of_p): Also return true if thunks match.
4273 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
4274 cgraph_function_or_thunk_node and an obsolete comment.
4275 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
4277 (build_function_decl_skip_args): Likewise.
4278 (set_new_clone_decl_and_node_flags): New function.
4279 (duplicate_thunk_for_node): Likewise.
4280 (redirect_edge_duplicating_thunks): Likewise.
4281 (cgraph_clone_node): New parameter args_to_skip, pass it to
4282 redirect_edge_duplicating_thunks which is called instead of
4283 cgraph_redirect_edge_callee.
4284 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
4285 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
4287 2014-04-04 Jeff Law <law@redhat.com>
4290 * config/arm/predicates.md (const_int_I_operand): New predicate.
4291 (const_int_M_operand): Similarly.
4292 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
4294 (insv_t2, extv_reg, extzv_t2): Likewise.
4295 (load_multiple_with_writeback): Similarly for const_int_I_operand.
4296 (pop_multiple_with_writeback_and_return): Likewise.
4297 (vfp_pop_multiple_with_writeback): Likewise
4299 2014-04-04 Richard Biener <rguenther@suse.de>
4302 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
4303 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
4305 * gimplify.h (gimple_add_tmp_var_fn): Declare.
4306 * gimplify.c (gimple_add_tmp_var_fn): New function.
4307 * gimple-expr.h (create_tmp_reg_fn): Declare.
4308 * gimple-expr.c (create_tmp_reg_fn): New function.
4309 * gimple-low.c (record_vars_into): Don't change cfun.
4310 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
4311 code generation without cfun.
4313 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
4316 * Makefile.in (install-driver): Fix shell scripting.
4318 2014-04-03 Cong Hou <congh@google.com>
4320 PR tree-optimization/60505
4321 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
4322 threshold of number of iterations below which no vectorization
4324 * tree-vect-loop.c (new_loop_vec_info):
4325 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
4326 * tree-vect-loop.c (vect_analyze_loop_operations):
4327 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
4328 * tree-vect-loop.c (vect_transform_loop):
4329 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
4330 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
4331 of iterations of the loop and see if we should build the epilogue.
4333 2014-04-03 Richard Biener <rguenther@suse.de>
4335 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
4336 (streamer_tree_cache_create): Adjust.
4337 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
4338 to allow optional nodes array.
4339 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
4340 (streamer_tree_cache_append): Likewise.
4341 (streamer_tree_cache_create): Create nodes array optionally
4342 as specified by parameter.
4343 * lto-streamer-out.c (create_output_block): Avoid maintaining
4344 the node array in the writer cache.
4345 (DFS_write_tree): Remove assertion.
4346 (produce_asm_for_decls): Free the out decl state hash table early.
4347 * lto-streamer-in.c (lto_data_in_create): Adjust for
4348 streamer_tree_cache_create prototype change.
4350 2014-04-03 Richard Biener <rguenther@suse.de>
4352 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
4353 set TREE_CHAIN to NULL_TREE.
4355 2014-04-03 Richard Biener <rguenther@suse.de>
4357 PR tree-optimization/60740
4358 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
4359 over all GIMPLE_COND operands.
4361 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
4363 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
4364 (Weffc++): Remove Scott's numbering, merge lists and reference
4367 2014-04-03 Nick Clifton <nickc@redhat.com>
4369 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
4372 2014-04-03 Martin Jambor <mjambor@suse.cz>
4374 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
4375 mention gcc_unreachable before failing.
4376 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
4379 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
4382 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
4383 inconsistent code and instead mark the context inconsistent.
4384 (possible_polymorphic_call_targets): For inconsistent contexts
4385 return empty complete list.
4387 2014-04-02 Anthony Green <green@moxielogic.com>
4389 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
4390 (extendqisi2, extendhisi2): Define.
4391 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
4392 (WCHAR_TYPE): Change to unsigned int.
4394 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4396 PR tree-optimization/60733
4397 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
4398 insertion point for PHI candidates to be the end of the feeding
4399 block for the PHI argument.
4401 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
4403 PR rtl-optimization/60650
4404 * lra-constraints.c (process_alt_operands): Decrease reject for
4405 earlyclobber matching.
4407 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4409 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
4411 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4413 * config/spu/spu.c (pad_bb): Do not crash when the last
4414 insn is CODE_FOR_blockage.
4416 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4418 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
4419 lies outside the target mode.
4421 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4424 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4425 software floating point or no floating point registers, do not
4426 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4427 in GPRs that occurs after we tested for GPRs that would never be
4430 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4431 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4432 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4433 specifically allow DDmode, since that does not use the SPE SIMD
4436 2014-04-02 Richard Biener <rguenther@suse.de>
4439 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4440 MODE_INTs. Properly use negv_optab.
4441 (expand_abs): Likewise.
4443 2014-04-02 Richard Biener <rguenther@suse.de>
4446 * Makefile.in (install-driver): Guard extra installs with special
4449 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4451 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4454 2014-04-01 Richard Henderson <rth@redhat.com>
4457 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
4458 alternative enabled before register allocation.
4460 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
4462 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
4463 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
4465 (nios2_large_got_address): Remove unneeded 'sym' parameter.
4466 (nios2_got_address): Update nios2_large_got_address call site.
4467 (nios2_delegitimize_address): New function.
4468 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
4469 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
4470 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
4472 2014-04-01 Martin Husemann <martin@duskware.de>
4474 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
4477 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
4479 PR rtl-optimization/60604
4480 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
4481 check from register_operand.
4482 (register_operand): Redefine in terms of general_operand.
4483 (nonmemory_operand): Use register_operand for the non-constant cases.
4485 2014-04-01 Richard Biener <rguenther@suse.de>
4487 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
4489 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
4491 * doc/invoke.texi (mapp-regs): Clarify.
4493 2014-03-31 Ulrich Drepper <drepper@gmail.com>
4495 * config/i386/avx512fintrin.h (__v32hi): Define type.
4496 (__v64qi): Likewise.
4497 (_mm512_set1_epi8): Define.
4498 (_mm512_set1_epi16): Define.
4499 (_mm512_set4_epi32): Define.
4500 (_mm512_set4_epi64): Define.
4501 (_mm512_set4_pd): Define.
4502 (_mm512_set4_ps): Define.
4503 (_mm512_setr4_epi64): Define.
4504 (_mm512_setr4_epi32): Define.
4505 (_mm512_setr4_pd): Define.
4506 (_mm512_setr4_ps): Define.
4507 (_mm512_setzero_epi32): Define.
4509 2014-03-31 Martin Jambor <mjambor@suse.cz>
4512 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
4513 callsite_arguments_match_p. Updated all callers. Also check types of
4514 corresponding formal parameters and actual arguments.
4515 (not_all_callers_have_enough_arguments_p) Renamed to
4516 some_callers_have_mismatched_arguments_p.
4518 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
4520 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
4522 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
4525 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
4528 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
4530 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
4531 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
4533 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
4534 Use FMAMODE_NOVF512 mode iterator.
4535 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
4536 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
4537 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
4538 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
4540 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
4541 Use VF_128_256 mode iterator.
4542 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
4545 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4547 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
4548 static chain if needed.
4550 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4553 * lra-constraints.c (index_part_to_reg): New.
4554 (process_address): Use it.
4556 2014-03-27 Jeff Law <law@redhat.com>
4557 Jakub Jelinek <jakub@redhat.com>
4560 * expr.c (do_tablejump): Use simplify_gen_binary rather than
4561 gen_rtx_{PLUS,MULT} to build up the address expression.
4563 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
4564 creating non-canonical RTL.
4566 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4569 * ipa-inline.c (want_inline_small_function_p): Short circuit large
4570 functions; reorganize to make cheap checks first.
4571 (inline_small_functions): Do not estimate growth when dumping;
4573 * ipa-inline.h (inline_summary): Add min_size.
4574 (growth_likely_positive): New function.
4575 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
4576 (set_cond_stmt_execution_predicate): Cleanup.
4577 (estimate_edge_size_and_time): Compute min_size.
4578 (estimate_calls_size_and_time): Likewise.
4579 (estimate_node_size_and_time): Likewise.
4580 (inline_update_overall_summary): Update min_size.
4581 (do_estimate_edge_time): Likewise.
4582 (do_estimate_edge_size): Update.
4583 (do_estimate_edge_hints): Update.
4584 (growth_likely_positive): New function.
4586 2014-03-28 Jakub Jelinek <jakub@redhat.com>
4589 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
4590 also if addr has VOIDmode.
4592 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4594 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
4595 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
4597 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
4598 instructions as well as AdvancedSIMD loads.
4600 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4602 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
4603 Use crypto_aese type.
4604 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
4605 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
4606 crypto_aese, crypto_aesmc. Move to types.md.
4607 * config/arm/types.md (crypto_aes): Split into crypto_aese,
4609 * config/arm/iterators.md (crypto_type): Likewise.
4611 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4613 * cgraph.c: Include expr.h and tree-dfa.h.
4614 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
4617 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4620 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
4621 regs from checking multi-reg pseudos.
4623 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4625 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
4627 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4629 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
4630 if it would clobber the stack pointer, even temporarily.
4632 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
4634 * mode-switching.c: Make small adjustments to the top comment.
4636 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4638 * config/rs6000/constraints.md (wD constraint): New constraint to
4639 match the constant integer to get the top DImode/DFmode out of a
4640 vector in a VSX register.
4642 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
4643 match the constant integer to get the top DImode/DFmode out of a
4644 vector in a VSX register.
4646 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
4649 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
4652 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
4653 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
4655 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
4656 Optimize vec_extract of 64-bit values, where the value being
4657 extracted is in the top word, where we can use scalar
4658 instructions. Add direct move and store support. Combine the big
4659 endian/little endian vector select load support into a single insn.
4660 (vsx_extract_<mode>_internal1): Likewise.
4661 (vsx_extract_<mode>_internal2): Likewise.
4662 (vsx_extract_<mode>_load): Likewise.
4663 (vsx_extract_<mode>_store): Likewise.
4664 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
4665 combined into vsx_extract_<mode>_load.
4666 (vsx_extract_<mode>_one_le): Likewise.
4668 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
4669 define the top 64-bit vector element.
4671 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
4674 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4675 Document vec_vbpermq builtin.
4678 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
4679 enable use of xxsldwi and xxpermdi builtin functions.
4680 (vec_xxpermdi): Likewise.
4682 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4683 Document use of vec_xxsldwi and vec_xxpermdi builtins.
4685 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
4687 PR rtl-optimization/60650
4688 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
4690 (find_spills_for): New.
4691 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
4692 Spill all pseudos on the second iteration.
4694 2014-03-27 Marek Polacek <polacek@redhat.com>
4697 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
4700 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4702 * config/s390/s390.c (s390_can_use_return_insn): Check for
4703 call-saved FPRs on 31 bit.
4705 2014-03-27 Jakub Jelinek <jakub@redhat.com>
4708 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
4709 if they need regimplification, just drop them instead of
4710 calling gimple_regimplify_operands on them.
4712 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
4715 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
4716 (aarch64_frame_pointer_required): Adjust logic.
4717 (aarch64_can_eliminate): Adjust logic.
4718 (aarch64_override_options_after_change): Adjust logic.
4720 2014-03-27 Dehao Chen <dehao@google.com>
4722 * ipa-inline.c (early_inliner): Update node's inline info.
4724 2014-03-26 Dehao Chen <dehao@google.com>
4726 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
4727 compiler inserted conditional jumps for NAN float check.
4729 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4731 * ubsan.h (ubsan_create_data): Change second argument's type
4732 to const location_t *.
4733 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
4735 (ubsan_create_data): Change second argument to const location_t *PLOC.
4736 Create Loc field whenever PLOC is non-NULL.
4737 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
4738 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
4742 * real.c (real_to_integer2): Change type of low to UHWI.
4744 2014-03-26 Tobias Burnus <burnus@net-b.de>
4746 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
4747 (CILK_SELF_SPECS): New define.
4748 (driver_self_specs): Use it.
4750 2014-03-26 Richard Biener <rguenther@suse.de>
4752 * tree-pretty-print.c (percent_K_format): Implement special
4753 case for LTO and its stripped down BLOCK tree.
4755 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4758 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
4760 * tree-vrp.c (simplify_internal_call_using_ranges): If only
4761 one range is range_int_cst_p, but not both, at least optimize
4762 addition/subtraction of 0 and multiplication by 0 or 1.
4763 * gimple-fold.c (gimple_fold_call): Fold
4764 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
4765 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
4766 INTEGER_CSTs, try to fold at least x * 0 and y - y.
4768 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
4770 PR rtl-optimization/60452
4771 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
4772 <case REG>: Return 1 for invalid offsets from the frame pointer.
4774 2014-03-26 Marek Polacek <polacek@redhat.com>
4777 * doc/extend.texi (C Extensions): Mention variable-length arrays in
4780 2014-03-26 Marek Polacek <polacek@redhat.com>
4783 * doc/extend.texi (Designated Inits): Describe what happens to omitted
4786 2014-03-26 Marek Polacek <polacek@redhat.com>
4789 * ira-color.c (update_conflict_hard_regno_costs): Perform the
4790 multiplication in unsigned type.
4792 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4794 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
4796 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4798 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
4800 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4803 * cif-code.def (UNREACHABLE) New code.
4804 * ipa-inline.c (inline_small_functions): Skip edges to
4805 __builtlin_unreachable.
4806 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
4807 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
4808 predicate to __bulitin_unreachable.
4809 (set_cond_stmt_execution_predicate): Fix issue when
4810 invert_tree_comparison returns ERROR_MARK.
4811 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
4812 propagate to inline clones.
4813 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
4815 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
4816 * cgraphclones.c (cgraph_clone_node): If call destination is already
4817 ureachable, do not redirect it back.
4818 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
4821 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4823 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
4824 Do not modify inline clones.
4826 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4828 * config/i386/i386.md (general_sext_operand): New mode attr.
4829 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
4830 don't generate (sign_extend (const_int)).
4831 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
4832 operands[2]. Use We constraint instead of <i> and
4833 <general_sext_operand> predicate instead of <general_operand>.
4834 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
4835 * config/i386/constraints.md (We): New constraint.
4836 * config/i386/predicates.md (x86_64_sext_operand,
4837 sext_operand): New predicates.
4839 2014-03-25 Martin Jambor <mjambor@suse.cz>
4842 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
4843 inconsistent devirtualizations to __builtin_unreachable.
4845 2014-03-25 Marek Polacek <polacek@redhat.com>
4848 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
4850 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
4852 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
4853 order of elements for big-endian.
4855 2014-03-25 Richard Biener <rguenther@suse.de>
4858 * gimplify-me.c (gimple_regimplify_operands): Update the
4861 2014-03-25 Martin Jambor <mjambor@suse.cz>
4864 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
4865 (lto_output_varpool_node): Likewise.
4866 (input_overwrite_node): Likewise.
4867 (input_varpool_node): Likewise.
4869 2014-03-25 Richard Biener <rguenther@suse.de>
4871 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
4872 (run_gcc): Likewise.
4874 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4876 * combine.c (simplify_compare_const): Add MODE argument.
4877 Handle mode_width 0 as very large mode_width.
4878 (try_combine, simplify_comparison): Adjust callers.
4880 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
4881 type to avoid signed integer overflow.
4882 * explow.c (plus_constant): Likewise.
4884 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
4886 * doc/generic.texi: Correct typos.
4888 2014-03-24 Tobias Burnus <burnus@net-b.de>
4890 * doc/invoke.texi (-flto): Expand section about
4891 using static libraries with LTO.
4893 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4895 PR rtl-optimization/60501
4896 * optabs.def (addptr3_optab): New optab.
4897 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
4898 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
4899 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
4901 * lra.c (emit_add3_insn): Use the addptr pattern if available.
4903 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
4905 2014-03-24 Ulrich Drepper <drepper@gmail.com>
4907 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
4910 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
4911 (_mm256_undefined_ps): Define.
4912 (_mm256_undefined_pd): Define.
4913 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
4914 (_mm_undefined_pd): Define.
4915 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
4916 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
4917 (_mm512_undefined_ps): Define.
4918 (_mm512_undefined_pd): Define.
4919 Use _mm*_undefined_*.
4920 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
4922 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
4924 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
4925 (lshr_simd): DI mode added.
4926 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
4927 (aarch64_ushr_simddi): Likewise.
4928 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
4929 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
4930 (vshrd_n_u64): Likewise.
4932 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4934 * Makefile.in (s-macro_list): Depend on cc1.
4936 2014-03-23 Teresa Johnson <tejohnson@google.com>
4938 * ipa-utils.c (ipa_print_order): Use specified dump file.
4940 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
4942 PR rtl-optimization/60601
4943 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
4945 * gcc.c (eval_spec_function): Initialize save_growing_value.
4947 2014-03-22 Jakub Jelinek <jakub@redhat.com>
4950 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
4951 code == MINUS_EXPR, never swap op0 with op1.
4953 * toplev.c (init_local_tick): Avoid signed integer multiplication
4955 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
4956 shift by first operand's bitsize.
4958 2014-03-21 Jakub Jelinek <jakub@redhat.com>
4961 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
4963 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
4964 TARGET_ISA_64BIT_P(x).
4966 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4968 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
4969 pattern for vector nor instead of subtract from splat(-1).
4970 (altivec_expand_vec_perm_const_le): Likewise.
4972 2014-03-21 Richard Henderson <rth@twiddle.net>
4975 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
4976 related insns after epilogue_completed.
4978 2014-03-21 Martin Jambor <mjambor@suse.cz>
4981 * cgraph.h (symtab_node): New flag body_removed.
4982 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
4983 when removing bodies.
4984 * symtab.c (dump_symtab_base): Dump body_removed flag.
4985 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
4986 had their bodies removed.
4988 2014-03-21 Martin Jambor <mjambor@suse.cz>
4991 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
4994 2014-03-21 Richard Biener <rguenther@suse.de>
4996 PR tree-optimization/60577
4997 * tree-core.h (struct tree_base): Document nothrow_flag use
4998 in VAR_DECL_NONALIASED.
4999 * tree.h (VAR_DECL_NONALIASED): New.
5000 (may_be_aliased): Adjust.
5001 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
5003 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5005 * expr.c (expand_expr_real_1): Remove outdated comment.
5007 2014-03-20 Jakub Jelinek <jakub@redhat.com>
5010 * ira.c (adjust_cleared_regs): Call copy_rtx on
5011 *reg_equiv[REGNO (loc)].src_p before passing it to
5012 simplify_replace_fn_rtx.
5015 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
5016 into CONST, put pic register as first operand of PLUS. Use
5017 gen_const_mem for both 32-bit and 64-bit PIC got loads.
5019 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5021 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
5023 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5025 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
5026 around for store forwarding issue in the FPU on the UT699.
5027 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
5028 loads and operations if -mfix-ut699 is specified.
5029 (divtf3_hq): Tweak attribute.
5030 (sqrttf2_hq): Likewise.
5032 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5034 * calls.c (store_one_arg): Remove incorrect const qualification on the
5035 type of the temporary.
5036 * cfgexpand.c (expand_return): Likewise.
5037 * expr.c (expand_constructor): Likewise.
5038 (expand_expr_real_1): Likewise.
5040 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5042 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
5045 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
5048 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
5050 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
5052 * config/arm/aarch-common-protos.h
5053 (alu_cost_table): Fix spelling of "extend".
5054 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
5056 2014-03-19 Richard Biener <rguenther@suse.de>
5059 * tree-core.h (tree_type_common): Re-order pointer members
5060 to reduce recursion depth during GC walks.
5062 2014-03-19 Marek Polacek <polacek@redhat.com>
5065 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
5066 before accessing it.
5068 2014-03-19 Richard Biener <rguenther@suse.de>
5071 * lto-streamer-in.c (input_function): In WPA stage do not drop
5074 2014-03-19 Jakub Jelinek <jakub@redhat.com>
5076 PR tree-optimization/60559
5077 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
5078 with build_zero_cst assignment.
5080 2014-03-18 Kai Tietz <ktietz@redhat.com>
5082 PR rtl-optimization/56356
5083 * sdbout.c (sdbout_parms): Verify that parms'
5084 incoming argument is valid.
5085 (sdbout_reg_parms): Likewise.
5087 2014-03-18 Richard Henderson <rth@redhat.com>
5090 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
5091 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
5092 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
5094 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
5096 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
5097 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
5098 Italicize plugin event names in description. Explain that
5099 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
5100 Remind that no GCC functions should be called after PLUGIN_FINISH.
5101 Explain what pragmas with expansion are.
5103 2014-03-18 Martin Liska <mliska@suse.cz>
5105 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
5106 gimple call statement is update.
5107 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
5108 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
5110 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5113 * ubsan.c (ubsan_instrument_unreachable): Call
5114 initialize_sanitizer_builtins.
5115 (ubsan_pass): Likewise.
5118 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
5119 varpool_finalize_decl instead of rest_of_decl_compilation.
5121 2014-03-18 Richard Biener <rguenther@suse.de>
5123 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
5124 by using bitmap_and_compl instead of bitmap_and_compl_into.
5125 (df_rd_transfer_function): Likewise.
5127 2014-03-18 Richard Biener <rguenther@suse.de>
5129 * doc/lto.texi (fresolution): Fix typo.
5131 2014-03-18 Richard Biener <rguenther@suse.de>
5133 * doc/invoke.texi (flto): Update for changes in 4.9.
5135 2014-03-18 Richard Biener <rguenther@suse.de>
5137 * doc/loop.texi: Remove section on the removed lambda framework.
5138 Update loop docs with recent changes in preserving loop structure.
5140 2014-03-18 Richard Biener <rguenther@suse.de>
5142 * doc/lto.texi (-fresolution): Document.
5144 2014-03-18 Richard Biener <rguenther@suse.de>
5146 * doc/contrib.texi: Adjust my name.
5148 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5151 * internal-fn.c: Include diagnostic-core.h.
5152 (expand_BUILTIN_EXPECT): New function.
5153 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
5154 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
5155 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
5156 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
5158 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
5159 Revert 3 argument __builtin_expect code.
5160 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
5161 * gimple-fold.c (gimple_fold_call): Likewise.
5162 * tree.h (fold_builtin_expect): New prototype.
5163 * builtins.c (build_builtin_expect_predicate): Add predictor
5164 argument, if non-NULL, create 3 argument __builtin_expect.
5165 (fold_builtin_expect): No longer static. Add ARG2 argument,
5166 pass it through to build_builtin_expect_predicate.
5167 (fold_builtin_2): Adjust caller.
5168 (fold_builtin_3): Handle BUILT_IN_EXPECT.
5169 * internal-fn.def (BUILTIN_EXPECT): New.
5171 2014-03-18 Tobias Burnus <burnus@net-b.de>
5174 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
5175 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
5176 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
5178 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
5181 * predict.c (combine_predictions_for_bb): Fix up formatting.
5182 (expr_expected_value_1, expr_expected_value): Add predictor argument,
5183 fill what it points to if non-NULL.
5184 (tree_predict_by_opcode): Adjust caller, use the predictor.
5185 * predict.def (PRED_COMPARE_AND_SWAP): Add.
5187 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
5189 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
5190 proper constant for the store mode.
5192 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
5194 * symtab.c (change_decl_assembler_name): Fix transparent alias
5197 2014-03-16 Renlin Li <Renlin.Li@arm.com>
5199 * config/aarch64/aarch64.c: Correct the comments about the
5200 aarch64 stack layout.
5202 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
5204 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
5205 check for GF_OMP_FOR_KIND_FOR.
5207 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
5209 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
5210 ymm and zmm register names.
5212 2014-03-17 Jakub Jelinek <jakub@redhat.com>
5215 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
5216 note creation for the 2010-08-31 changes.
5218 2014-03-17 Marek Polacek <polacek@redhat.com>
5221 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
5222 as -fno-tree-loop-vectorize.
5223 (expand_omp_simd): Likewise.
5225 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
5227 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
5228 (eligible_for_call_delay): New prototype.
5229 * config/sparc/sparc.c (tls_call_delay): Rename into...
5230 (eligible_for_call_delay): ...this. Return false if the instruction
5231 cannot be put in the delay slot of a branch.
5232 (eligible_for_restore_insn): Simplify.
5233 (eligible_for_return_delay): Return false if the instruction cannot be
5234 put in the delay slot of a branch and simplify.
5235 (eligible_for_sibcall_delay): Return false if the instruction cannot be
5236 put in the delay slot of a branch.
5237 * config/sparc/sparc.md (fix_ut699): New attribute.
5238 (tls_call_delay): Delete.
5239 (in_call_delay): Reimplement.
5240 (eligible_for_sibcall_delay): Rename into...
5241 (in_sibcall_delay): ...this.
5242 (eligible_for_return_delay): Rename into...
5243 (in_return_delay): ...this.
5244 (in_branch_delay): Reimplement.
5245 (in_uncond_branch_delay): Delete.
5246 (in_annul_branch_delay): Delete.
5248 2014-03-14 Richard Henderson <rth@redhat.com>
5251 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
5252 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
5253 (*floathi<X87MODEF>2_i387_with_temp): Remove.
5254 (floathi splitters): Remove.
5255 (float<SWI48x>xf2): New pattern.
5256 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
5257 code that tried to handle DImode for 32-bit, but which was excluded
5258 by the pattern's condition. Drop allocation of stack temporary.
5259 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
5260 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
5261 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
5262 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
5263 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
5264 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
5265 (*float<SWI48><MODEF>2_sse_interunit): Remove.
5266 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
5267 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
5268 (*float<SWI48x><X87MODEF>2_i387): Remove.
5269 (all float _with_temp splitters): Remove.
5270 (*float<SWI48x><MODEF>2_i387): New pattern.
5271 (*float<SWI48><MODEF>2_sse): New pattern.
5272 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
5273 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
5275 2014-03-14 Jakub Jelinek <jakub@redhat.com>
5276 Marek Polacek <polacek@redhat.com>
5279 * common.opt (dump_base_name_prefixed): New Variable.
5280 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
5281 if x_dump_base_name_prefixed is already set, set it at the end.
5283 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
5285 PR rtl-optimization/60508
5286 * lra-constraints.c (get_reload_reg): Add new parameter
5288 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
5289 Pass the new parameter values.
5291 2014-03-14 Richard Biener <rguenther@suse.de>
5293 * common.opt: Revert unintented changes from r205065.
5296 2014-03-14 Richard Biener <rguenther@suse.de>
5299 * cfghooks.c (split_block): Properly adjust all loops the
5300 block was a latch of.
5302 2014-03-14 Martin Jambor <mjambor@suse.cz>
5305 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
5308 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
5311 * config/avr/avr.c (avr_set_current_function): Pass function name
5312 through default_strip_name_encoding before sanity checking instead
5313 of skipping the first char of the assembler name.
5315 2014-03-13 Richard Henderson <rth@redhat.com>
5318 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
5319 (ix86_force_to_memory, ix86_free_from_memory): Remove.
5320 * config/i386/i386-protos.h: Likewise.
5321 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
5322 in the expander instead of a splitter.
5323 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
5324 any possibility of requiring a memory.
5325 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
5326 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
5327 (fp branch splitters): Update for ix86_split_fp_branch.
5328 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
5329 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
5330 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
5331 (*fop_<MODEF>_2_i387): Remove f/r alternative.
5332 (*fop_<MODEF>_3_i387): Likewise.
5333 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
5334 (splitters for the fop_* register patterns): Remove.
5335 (fscalexf4_i387): Rename from *fscalexf4_i387.
5336 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
5338 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5340 PR tree-optimization/59779
5341 * tree-dfa.c (get_ref_base_and_extent): Use double_int
5342 type for bitsize and maxsize instead of HOST_WIDE_INT.
5344 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
5346 PR rtl-optimization/57320
5347 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
5348 the CFG after thread_prologue_and_epilogue_insns.
5350 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
5352 PR rtl-optimization/57189
5353 * lra-constraints.c (process_alt_operands): Disfavor spilling
5356 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
5358 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
5360 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5362 PR tree-optimization/59025
5364 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
5365 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
5367 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
5370 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
5371 calls of avr_out_plus_1.
5373 2014-03-13 Bin Cheng <bin.cheng@arm.com>
5375 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
5376 BB's single pred and update the father loop's latch info later.
5378 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5380 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
5384 (VEC_base): Likewise.
5385 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
5386 registers, we need to swap double words in little endian mode.
5388 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
5389 to be a container mode for 128-bit integer operations added in ISA
5390 2.07. Unlike TImode and PTImode, the preferred register set is
5391 the Altivec/VMX registers for the 128-bit operations.
5393 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
5395 (rs6000_split_128bit_ok_p): Likewise.
5397 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
5398 macros for creating ISA 2.07 normal and overloaded builtin
5399 functions with 3 arguments.
5400 (BU_P8V_OVERLOAD_3): Likewise.
5401 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
5402 for use as overloaded functions.
5403 (VPERM_1TI_UNS): Likewise.
5404 (VSEL_1TI): Likewise.
5405 (VSEL_1TI_UNS): Likewise.
5406 (ST_INTERNAL_1ti): Likewise.
5407 (LD_INTERNAL_1ti): Likewise.
5408 (XXSEL_1TI): Likewise.
5409 (XXSEL_1TI_UNS): Likewise.
5410 (VPERM_1TI): Likewise.
5411 (VPERM_1TI_UNS): Likewise.
5412 (XXPERMDI_1TI): Likewise.
5413 (SET_1TI): Likewise.
5414 (LXVD2X_V1TI): Likewise.
5415 (STXVD2X_V1TI): Likewise.
5416 (VEC_INIT_V1TI): Likewise.
5417 (VEC_SET_V1TI): Likewise.
5418 (VEC_EXT_V1TI): Likewise.
5419 (EQV_V1TI): Likewise.
5420 (NAND_V1TI): Likewise.
5421 (ORC_V1TI): Likewise.
5422 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
5423 added in ISA 2.07. Add both normal 'altivec' builtins, and the
5425 (VADDUQM): Likewise.
5426 (VSUBCUQ): Likewise.
5427 (VADDEUQM): Likewise.
5428 (VADDECUQ): Likewise.
5429 (VSUBEUQM): Likewise.
5430 (VSUBECUQ): Likewise.
5432 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5433 __int128_t and __uint128_t types.
5434 (__uint128_type): Likewise.
5435 (altivec_categorize_keyword): Add support for vector __int128_t,
5436 vector __uint128_t, vector __int128, and vector unsigned __int128
5437 as a container type for TImode operations that need to be done in
5438 VSX/Altivec registers.
5439 (rs6000_macro_to_expand): Likewise.
5440 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5441 to support 128-bit integer instructions vaddcuq, vadduqm,
5442 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5443 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5445 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5446 for V1TImode, and set up preferences to use VSX/Altivec registers.
5447 Setup VSX reload handlers.
5448 (rs6000_debug_reg_global): Likewise.
5449 (rs6000_init_hard_regno_mode_ok): Likewise.
5450 (rs6000_preferred_simd_mode): Likewise.
5451 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5452 (easy_altivec_constant): Likewise.
5453 (output_vec_const_move): Likewise.
5454 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5456 (rs6000_expand_vector_extract): Likewise.
5457 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
5459 (rs6000_const_vec): Add support for V1TImode.
5460 (rs6000_emit_le_vsx_load): Swap double words when loading or
5461 storing TImode/V1TImode.
5462 (rs6000_emit_le_vsx_store): Likewise.
5463 (rs6000_emit_le_vsx_move): Likewise.
5464 (rs6000_emit_move): Add support for V1TImode.
5465 (altivec_expand_ld_builtin): Likewise.
5466 (altivec_expand_st_builtin): Likewise.
5467 (altivec_expand_vec_init_builtin): Likewise.
5468 (altivec_expand_builtin): Likewise.
5469 (rs6000_init_builtins): Add support for V1TImode type. Add
5470 support for ISA 2.07 128-bit integer builtins. Define type names
5471 for the VSX/Altivec vector types.
5472 (altivec_init_builtins): Add support for overloaded vector
5473 functions with V1TImode type.
5474 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
5475 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
5477 (rs6000_split_128bit_ok_p): Likewise.
5478 (rs6000_handle_altivec_attribute): Create V1TImode from vector
5479 __int128_t and vector __uint128_t.
5481 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
5482 and mode attributes.
5489 (VS_scalar): Likewise.
5490 (VS_double): Likewise.
5491 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
5493 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
5494 we support the ISA 2.07 128-bit integer arithmetic instructions.
5495 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
5496 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
5497 and TImode types for use with the builtin functions.
5498 (V1TI_type_node): Likewise.
5499 (unsigned_V1TI_type_node): Likewise.
5500 (intTI_type_internal_node): Likewise.
5501 (uintTI_type_internal_node): Likewise.
5503 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
5504 128-bit builtin functions.
5505 (UNSPEC_VADDEUQM): Likewise.
5506 (UNSPEC_VADDECUQ): Likewise.
5507 (UNSPEC_VSUBCUQ): Likewise.
5508 (UNSPEC_VSUBEUQM): Likewise.
5509 (UNSPEC_VSUBECUQ): Likewise.
5510 (VM): Add V1TImode to vector mode iterators.
5512 (VI_unit): Likewise.
5513 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
5514 (altivec_vaddcuq): Likewise.
5515 (altivec_vsubuqm): Likewise.
5516 (altivec_vsubcuq): Likewise.
5517 (altivec_vaddeuqm): Likewise.
5518 (altivec_vaddecuq): Likewise.
5519 (altivec_vsubeuqm): Likewise.
5520 (altivec_vsubecuq): Likewise.
5522 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
5524 (BOOL_128): Likewise.
5525 (BOOL_REGS_OUTPUT): Likewise.
5526 (BOOL_REGS_OP1): Likewise.
5527 (BOOL_REGS_OP2): Likewise.
5528 (BOOL_REGS_UNARY): Likewise.
5529 (BOOL_REGS_AND_CR0): Likewise.
5531 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
5532 128-bit integer builtin support.
5533 (vec_vadduqm): Likewise.
5534 (vec_vaddecuq): Likewise.
5535 (vec_vaddeuqm): Likewise.
5536 (vec_vsubecuq): Likewise.
5537 (vec_vsubeuqm): Likewise.
5538 (vec_vsubcuq): Likewise.
5539 (vec_vsubuqm): Likewise.
5541 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5542 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
5543 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
5544 128-bit integer add/subtract to ISA 2.07.
5546 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
5548 * config/arc/arc.c (arc_predicate_delay_insns):
5549 Fix third argument passed to conditionalize_nonjump.
5551 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
5553 * config/aarch64/aarch64-builtins.c
5554 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
5555 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
5556 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
5557 instead of __builtin_lfloor.
5558 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
5560 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5562 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
5563 (tree_ssa_ifcombine_bb_1): New function.
5564 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
5565 is an empty forwarder block to then_bb or vice versa and then_bb
5566 and else_bb are effectively swapped.
5568 2014-03-12 Christian Bruel <christian.bruel@st.com>
5571 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
5572 REG_CFA_DEF_CFA note.
5573 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
5574 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
5576 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
5578 PR tree-optimization/60454
5579 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
5581 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5583 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
5584 Do not define target_cpu_default2 to generic.
5585 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
5586 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
5587 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
5589 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5590 Marc Glisse <marc.glisse@inria.fr>
5592 PR tree-optimization/60502
5593 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
5594 instead of build_low_bits_mask.
5596 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5599 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
5600 if there are multiple uses, but op doesn't live on E edge.
5601 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
5602 clobber stmts before __builtin_unreachable.
5604 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
5606 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
5607 hard_frame_pointer_rtx.
5608 * cse.c (cse_insn): Remove volatile check.
5609 * cselib.c (cselib_process_insn): Likewise.
5610 * dse.c (scan_insn): Likewise.
5612 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5614 * config/arc/arc.c (conditionalize_nonjump): New function,
5616 (arc_ifcvt): ... this.
5617 (arc_predicate_delay_insns): Use it.
5619 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5621 * config/arc/predicates.md (extend_operand): During/after reload,
5622 allow const_int_operand.
5623 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
5624 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
5625 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
5627 (umulsi3_highpart_i): Likewise.
5629 2014-03-11 Richard Biener <rguenther@suse.de>
5631 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
5632 Add asserts to guard possible wrong-code bugs.
5634 2014-03-11 Richard Biener <rguenther@suse.de>
5636 PR tree-optimization/60429
5637 PR tree-optimization/60485
5638 * tree-ssa-structalias.c (set_union_with_increment): Properly
5639 take into account all fields that overlap the shifted vars.
5640 (do_sd_constraint): Likewise.
5641 (do_ds_constraint): Likewise.
5642 (get_constraint_for_ptr_offset): Likewise.
5644 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
5646 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
5647 (nios2_compute_frame_layout):
5648 Add calculation of cfun->machine->fp_save_offset.
5649 (nios2_expand_prologue): Correct setting of frame pointer register
5651 (nios2_expand_epilogue): Update recovery of stack pointer from
5652 frame pointer accordingly.
5653 (nios2_initial_elimination_offset): Update calculation of offset
5654 for eliminating to HARD_FRAME_POINTER_REGNUM.
5656 2014-03-10 Jakub Jelinek <jakub@redhat.com>
5659 * ipa.c (symtab_remove_unreachable_nodes): Don't call
5660 cgraph_get_create_node on VAR_DECLs.
5662 2014-03-10 Richard Biener <rguenther@suse.de>
5665 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
5667 2014-03-08 Douglas B Rupp <rupp@gnat.com>
5669 * config/vms/vms.opt (vms_float_format): New variable.
5671 2014-03-08 Tobias Burnus <burnus@net-b.de>
5673 * doc/invoke.texi (-fcilkplus): Update implementation status.
5675 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
5676 Richard Biener <rguenther@suse.de>
5678 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
5679 consistently accross all TUs.
5680 (run_gcc): Enable -fshort-double automatically at link at link-time
5681 and disallow override.
5683 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
5686 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
5687 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
5688 if they can't be used.
5690 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5692 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
5693 for Solaris 11/x86 ld.
5694 * configure: Regenerate.
5696 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5698 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
5699 (LIB_TLS_SPEC): Save as ld_tls_libs.
5700 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
5701 (HAVE_AS_IX86_TLSLDM): New test.
5702 * configure, config.in: Regenerate.
5703 * config/i386/i386.c (legitimize_tls_address): Fall back to
5704 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
5705 cannot support TLS_MODEL_LOCAL_DYNAMIC.
5706 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
5707 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
5709 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
5711 * common.opt (fira-loop-pressure): Mark as optimization.
5713 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
5715 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
5716 an OpenMP mappable type.
5718 2014-03-06 Matthias Klose <doko@ubuntu.com>
5720 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
5721 MULTILIB_OSDIRNAMES is not defined.
5723 2014-03-06 Jakub Jelinek <jakub@redhat.com>
5724 Meador Inge <meadori@codesourcery.com>
5727 * config/arm/arm.c (arm_tls_symbol_p): Remove.
5728 (arm_legitimize_address): Call legitimize_tls_address for any
5729 arm_tls_referenced_p expression, handle constant addend. Call it
5730 before testing for !TARGET_ARM.
5731 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
5733 2014-03-06 Richard Biener <rguenther@suse.de>
5739 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5741 * tree-streamer.c (record_common_node): Assert we don't record
5742 nodes with type double.
5743 (preload_common_node): Skip type double, complex double and double
5744 pointer since it is now frontend dependent due to fshort-double option.
5746 2014-03-06 Richard Biener <rguenther@suse.de>
5748 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
5749 or -fno-lto is specified and the linker has full plugin support.
5750 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
5751 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
5752 * lto-wrapper.c (merge_and_complain): Merge compile-time
5753 optimization levels.
5754 (run_gcc): And pass it through to the link options.
5756 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
5760 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5762 * cselib.c (remove_useless_values): Skip to avoid quadratic
5763 behavior if the condition moved from...
5764 (cselib_process_insn): ... here holds.
5766 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5769 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
5770 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
5773 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
5774 (TM_H): Add x86-tune.def.
5776 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5778 * config/aarch64/aarch64.c (generic_tunings):
5779 Use cortexa57_extra_costs.
5781 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5784 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
5785 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
5786 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
5789 2014-03-04 Heiher <r@hev.cc>
5791 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
5792 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
5794 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
5796 * config/i386/predicates.md (const2356_operand): Change to ...
5797 (const2367_operand): ... this.
5798 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
5800 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5801 (*avx512pf_scatterpf<mode>sf): Ditto.
5802 (avx512pf_scatterpf<mode>df): Ditto.
5803 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5804 (*avx512pf_scatterpf<mode>df): Ditto.
5805 * config/i386/i386.c (ix86_expand_builtin): Update
5806 incorrect hint operand error message.
5808 2014-03-04 Richard Biener <rguenther@suse.de>
5810 * lto-section-in.c (lto_get_section_data): Fix const cast.
5812 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5814 * tree-streamer.c (record_common_node): Assert we don't record
5815 nodes with type double.
5816 (preload_common_node): Skip type double, complex double and double
5817 pointer since it is now frontend dependent due to fshort-double option.
5819 2014-03-04 Richard Biener <rguenther@suse.de>
5822 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
5823 (lto_input_toplevel_asms): Likewise.
5824 * lto-section-in.c (lto_get_section_data): Instead do it here
5827 2014-03-04 Richard Biener <rguenther@suse.de>
5829 PR tree-optimization/60382
5830 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
5831 dead PHIs a reduction.
5833 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
5835 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
5837 (_mm_prefetch): Move out of GCC target("sse") pragma.
5838 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
5839 GCC target("prfchw") pragma.
5840 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
5842 * config/i386/i386.c (ix86_option_override_internal): Enable
5843 -mprfchw with -mprefetchwt1.
5845 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5847 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
5850 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5852 * opts.h (CL_PCH_IGNORE): Define.
5853 * targhooks.c (option_affects_pch_p):
5854 Return false for options that have CL_PCH_IGNORE set.
5855 * opt-functions.awk: Process PchIgnore.
5856 * doc/options.texi: Document PchIgnore.
5858 * config/arc/arc.opt (misize): Add PchIgnore property.
5860 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5862 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
5863 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
5864 constraint on constants to permit them being loaded into
5865 GENERAL_REGS or BASE_REGS.
5867 2014-03-03 Nick Clifton <nickc@redhat.com>
5869 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
5870 anti-cacnonical alternatives.
5871 (negandhi3_real): New pattern.
5872 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
5874 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5876 * config/avr/avr-mcus.def: Remove atxmega16x1.
5877 * config/avr/avr-tables.opt: Regenerate.
5878 * config/avr/t-multilib: Regenerate.
5879 * doc/avr-mmcu.texi: Regenerate.
5881 2014-03-03 Tobias Grosser <tobias@grosser.es>
5882 Mircea Namolaru <mircea.namolaru@inria.fr>
5884 PR tree-optimization/58028
5885 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
5888 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5890 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
5891 not handled by recognizers.
5893 2014-03-03 Jakub Jelinek <jakub@redhat.com>
5896 * function.c (expand_function_end): Don't emit
5897 clobber_return_register sequence if clobber_after is a BARRIER.
5898 * cfgexpand.c (construct_exit_block): Append instructions before
5899 return_label to prev_bb.
5901 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5903 * config/rs6000/constraints.md: Document reserved use of "wc".
5905 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5908 * ipa.c (function_and_variable_visibility): When dissolving comdat
5909 group, also set all symbols to local.
5911 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5916 2013-12-14 Jan Hubicka <jh@suse.cz>
5918 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
5920 2014-03-02 Jon Beniston <jon@beniston.com>
5926 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
5927 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
5928 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
5929 (simple_return, *simple_return): New patterns
5930 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
5931 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
5933 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
5935 * dwarf2out.c (gen_subprogram_die): Tidy.
5937 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
5940 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
5941 (*mov_t_msb_neg_negc): ... this new insn.
5943 2014-02-28 Jason Merrill <jason@redhat.com>
5946 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
5949 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
5952 * dwarf2out.c (decltype_auto_die): New static.
5953 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
5954 (gen_type_die_with_usage): Handle 'decltype(auto)'.
5955 (is_cxx_auto): Likewise.
5957 2014-02-28 Ian Bolton <ian.bolton@arm.com>
5959 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
5960 we are not using general regs only.
5962 2014-02-28 Richard Biener <rguenther@suse.de>
5965 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
5966 previous fix and only allow to remove trivial pre-headers
5967 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
5968 (remove_forwarder_block): Properly update the latch of a loop.
5970 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5973 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
5974 (cselib_preserved_hash_table): New.
5975 (preserve_constants_and_equivs): Move preserved vals to it.
5976 (cselib_find_slot): Look it up first.
5977 (cselib_init): Initialize it.
5978 (cselib_finish): Release it.
5979 (dump_cselib_table): Dump it.
5981 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5984 * cselib.c (remove_useless_values): Skip to avoid quadratic
5985 behavior if the condition moved from...
5986 (cselib_process_insn): ... here holds.
5988 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5991 * var-tracking.c (vt_initialize): Apply the same condition to
5992 preserve the CFA base value.
5994 2014-02-28 Joey Ye <joey.ye@arm.com>
5997 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
5998 if reload in progress or completed.
6000 2014-02-28 Tobias Burnus <burnus@net-b.de>
6003 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
6006 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
6008 * doc/tm.texi.in (Condition Code Status): Update documention for
6009 relative locations of cc0-setter and cc0-user.
6011 2014-02-27 Jeff Law <law@redhat.com>
6013 PR rtl-optimization/52714
6014 * combine.c (try_combine): When splitting an unrecognized PARALLEL
6015 into two independent simple sets, if I3 is a jump, ensure the
6016 pattern we place into I3 is a (set (pc) ...).
6018 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
6019 Jeff Law <law@redhat.com>
6021 PR rtl-optimization/49847
6022 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
6023 are in different blocks.
6024 * doc/tm.texi (Condition Code Status): Update documention for
6025 relative locations of cc0-setter and cc0-user.
6027 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
6030 * lra.c (lra_emit_add): Check SUBREG too.
6032 2014-02-27 Andreas Schwab <schwab@suse.de>
6034 * config/m68k/m68k.c (m68k_option_override): Disable
6035 -flive-range-shrinkage for classic m68k.
6036 (m68k_override_options_after_change): Likewise.
6038 2014-02-27 Marek Polacek <polacek@redhat.com>
6041 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
6042 -Wmaybe-uninitialized.
6044 2014-02-27 Alan Modra <amodra@gmail.com>
6047 * reload1.c (emit_input_reload_insns): When reload_override_in,
6048 set old to rl->in_reg when rl->in_reg is a subreg.
6050 2014-02-26 Richard Biener <rguenther@suse.de>
6053 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
6055 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6057 * common/config/i386/predicates.md (const1256_operand): Remove.
6058 (const2356_operand): New.
6059 (const_1_to_2_operand): Remove.
6060 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6061 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6062 (*avx512pf_gatherpf<mode>sf): Ditto.
6063 (avx512pf_gatherpf<mode>df): Ditto.
6064 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6065 (*avx512pf_gatherpf<mode>df): Ditto.
6066 (avx512pf_scatterpf<mode>sf): Ditto.
6067 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6068 (*avx512pf_scatterpf<mode>sf): Ditto.
6069 (avx512pf_scatterpf<mode>df): Ditto.
6070 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6071 (*avx512pf_scatterpf<mode>df): Ditto.
6072 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
6074 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
6076 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
6077 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
6078 (_mm512_mask_testn_epi64_mask): Move to ...
6079 * config/i386/avx512cdintrin.h: Here.
6080 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
6081 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
6082 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
6083 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
6084 TARGET_AVX512F from TARGET_AVX512CD.
6086 2014-02-26 Richard Biener <rguenther@suse.de>
6089 * ipa.c (walk_polymorphic_call_targets): Properly guard
6090 call to inline_update_overall_summary.
6092 2014-02-26 Bin Cheng <bin.cheng@arm.com>
6095 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
6096 and latches only if requested. Fix latch if it is removed.
6097 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
6098 LOOPS_HAVE_PREHEADERS.
6100 2014-02-25 Andrew Pinski <apinski@cavium.com>
6102 * builtins.c (expand_builtin_thread_pointer): Create a new target
6103 when the target is NULL.
6105 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
6107 PR rtl-optimization/60317
6108 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6109 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6110 * lra-assigns.c: Include params.h.
6111 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
6112 other reload pseudos considerations.
6114 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6116 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
6117 to use canonical form for nor<mode>3.
6119 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6122 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
6125 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6127 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
6128 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
6129 (ix86_handle_option): Handle OPT_mprefetchwt1.
6130 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
6131 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
6133 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
6134 OPTION_MASK_ISA_PREFETCHWT1.
6135 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
6136 (PTA_PREFETCHWT1): New.
6137 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
6138 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
6139 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
6140 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
6141 (*prefetch_avx512pf_<mode>_: Change into ...
6142 (*prefetch_prefetchwt1_<mode>: This.
6143 * config/i386/i386.opt (mprefetchwt1): New.
6144 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
6145 (_mm_prefetch): Handle intent to write.
6146 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
6148 2014-02-25 Richard Biener <rguenther@suse.de>
6151 * emit-rtl.c (mem_attrs_htab): Remove.
6152 (mem_attrs_htab_hash): Likewise.
6153 (mem_attrs_htab_eq): Likewise.
6154 (set_mem_attrs): Always allocate new mem-attrs when something changed.
6155 (init_emit_once): Do not allocate mem_attrs_htab.
6157 2014-02-25 Richard Biener <rguenther@suse.de>
6160 * lto-opts.c (lto_write_options): Output non-explicit conservative
6161 -fwrapv, -fno-trapv and -fno-strict-overflow.
6162 * lto-wrapper.c (merge_and_complain): Handle merging those options.
6163 (run_gcc): And pass them through.
6165 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6167 * sel-sched.c (calculate_new_fences): New parameter ptime.
6168 Calculate it as a maximum over all fence cycles.
6169 (sel_sched_region_2): Adjust the call to calculate_new_fences.
6170 Print the final schedule timing when sched_verbose.
6172 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6174 PR rtl-optimization/60292
6175 * sel-sched.c (fill_vec_av_set): Do not reset target availability
6176 bit fot the fence instruction.
6178 2014-02-24 Alangi Derick <alangiderick@gmail.com>
6180 * calls.h: Fix typo in comment.
6182 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
6184 * config/pa/pa.c (pa_output_move_double): Don't valididate when
6185 adjusting offsetable addresses.
6187 2014-02-24 Guozhi Wei <carrot@google.com>
6189 * sparseset.h (sparseset_pop): Fix the wrong index.
6191 2014-02-24 Walter Lee <walt@tilera.com>
6193 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
6194 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
6196 * common/config/tilegx/tilegx-common.c
6197 (TARGET_DEFAULT_TARGET_FLAGS): Define.
6198 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
6200 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
6201 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
6202 (tilegx_gimplify_va_arg_expr): Handle big endian.
6203 (tilegx_expand_unaligned_load): Ditto.
6204 (tilegx_expand_unaligned_store): Ditto.
6205 (TARGET_RETURN_IN_MSB): New.
6206 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
6207 (TARGET_ENDIAN_DEFAULT): New.
6208 (TARGET_BIG_ENDIAN): Handle big endian.
6209 (BYTES_BIG_ENDIAN): Ditto.
6210 (WORDS_BIG_ENDIAN): Ditto.
6211 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
6214 * config/tilegx/tilegx.md (extv): Handle big endian.
6216 (insn_st<n>): Ditto.
6217 (insn_st<n>_add<bitsuffix>): Ditto.
6218 (insn_stnt<n>): Ditto.
6219 (insn_stnt<n>_add<bitsuffix>):Ditto.
6220 (vec_interleave_highv8qi): Handle big endian.
6221 (vec_interleave_highv8qi_be): New.
6222 (vec_interleave_highv8qi_le): New.
6223 (insn_v1int_h): Handle big endian.
6224 (vec_interleave_lowv8qi): Handle big endian.
6225 (vec_interleave_lowv8qi_be): New.
6226 (vec_interleave_lowv8qi_le): New.
6227 (insn_v1int_l): Handle big endian.
6228 (vec_interleave_highv4hi): Handle big endian.
6229 (vec_interleave_highv4hi_be): New.
6230 (vec_interleave_highv4hi_le): New.
6231 (insn_v2int_h): Handle big endian.
6232 (vec_interleave_lowv4hi): Handle big endian.
6233 (vec_interleave_lowv4hi_be): New.
6234 (vec_interleave_lowv4hi_le): New.
6235 (insn_v2int_l): Handle big endian.
6236 (vec_interleave_highv2si): Handle big endian.
6237 (vec_interleave_highv2si_be): New.
6238 (vec_interleave_highv2si_le): New.
6239 (insn_v4int_h): Handle big endian.
6240 (vec_interleave_lowv2si): Handle big endian.
6241 (vec_interleave_lowv2si_be): New.
6242 (vec_interleave_lowv2si_le): New.
6243 (insn_v4int_l): Handle big endian.
6244 * config/tilegx/tilegx.opt (mbig-endian): New option.
6245 (mlittle-endian): New option.
6246 * doc/install.texi: Document tilegxbe-linux.
6247 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
6249 2014-02-24 Martin Jambor <mjambor@suse.cz>
6252 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
6253 there are no parameter descriptors.
6255 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
6257 PR rtl-optimization/60268
6258 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
6259 initialization to ...
6260 (sched_rgn_init): ... here.
6261 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
6263 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6265 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
6268 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
6270 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
6273 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6275 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
6276 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
6278 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6280 * config/microblaze/predicates.md: Add cmp_op predicate.
6281 * config/microblaze/microblaze.md: Add branch_compare instruction
6282 which uses cmp_op predicate and emits cmp insn before branch.
6283 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
6284 to microblaze_expand_conditional_branch and consolidate logic.
6285 (microblaze_expand_conditional_branch): emit branch_compare
6286 insn instead of handling cmp op separate from branch insn.
6288 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6290 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
6293 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6295 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
6296 define_insn with define_expand and new define_insn
6297 *altivec_lve<VI_char>x_internal.
6298 (altivec_stve<VI_char>x): Replace define_insn with define_expand
6299 and new define_insn *altivec_stve<VI_char>x_internal.
6300 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
6302 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
6304 (altivec_expand_stvex_be): New function.
6306 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
6308 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
6309 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
6310 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
6311 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
6313 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
6316 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
6317 instead of emit_move_insn.
6319 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6321 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
6323 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
6324 gen_altivec_vsumsws.
6326 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6328 * config/rs6000/altivec.md (altivec_lvxl): Rename as
6329 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
6330 (altivec_lvxl_<mode>): New define_expand incorporating
6331 -maltivec=be semantics where needed.
6332 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
6333 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
6334 semantics where needed.
6335 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
6336 (altivec_stvx_<mode>): New define_expand incorporating
6337 -maltivec=be semantics where needed.
6338 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
6339 VM2 iterator instead of V4SI.
6340 (altivec_stvxl_<mode>): New define_expand incorporating
6341 -maltivec=be semantics where needed.
6342 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
6343 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
6344 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
6345 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
6346 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
6347 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
6348 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
6349 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
6350 ALTIVEC_BUILTIN_STVXL.
6351 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
6352 (altivec_expand_stvx_be): Likewise.
6353 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
6354 (altivec_expand_lvx_be): Likewise.
6355 (altivec_expand_stvx_be): Likewise.
6356 (altivec_expand_builtin): Add cases for
6357 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
6358 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
6359 (altivec_init_builtins): Add definitions for
6360 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
6361 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
6363 2014-02-21 Catherine Moore <clm@codesourcery.com>
6365 * doc/invoke.texi (mvirt, mno-virt): Document.
6366 * config/mips/mips.opt (mvirt): New option.
6367 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
6369 2014-02-21 Richard Biener <rguenther@suse.de>
6371 PR tree-optimization/60276
6372 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
6373 (STMT_VINFO_MIN_NEG_DIST): New macro.
6374 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
6375 STMT_VINFO_MIN_NEG_DIST.
6376 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
6377 made for negative dependence distances still hold.
6379 2014-02-21 Richard Biener <rguenther@suse.de>
6382 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
6383 DECL_INITIAL for globals not in the current function context.
6385 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6387 PR tree-optimization/56490
6388 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
6389 * tree-ssa-uninit.c: Include params.h.
6390 (compute_control_dep_chain): Add num_calls argument, return false
6391 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
6392 num_calls to recursive call.
6393 (find_predicates): Change dep_chain into normal array,
6394 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
6395 variable and adjust compute_control_dep_chain caller.
6396 (find_def_preds): Likewise.
6398 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
6400 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
6401 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
6403 2014-02-21 Nick Clifton <nickc@redhat.com>
6405 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
6406 (pushhi1): Likewise.
6407 (popqi1): Add mode to pre_dec.
6410 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6412 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
6413 mode for mask of V8SFmode permutation.
6415 2014-02-20 Richard Henderson <rth@redhat.com>
6418 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
6419 a new pseudo for OLDVAL.
6421 2014-02-20 Jakub Jelinek <jakub@redhat.com>
6424 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
6425 gen_reg_rtx if d->testing_p.
6426 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
6427 if d->testing_p and we will certainly return true.
6428 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6431 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6433 * emit-rtl.c (gen_reg_rtx): Assert that
6434 crtl->emit.regno_pointer_align_length is non-zero.
6436 2014-02-20 Richard Henderson <rth@redhat.com>
6439 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6440 on failure the store back into EXPECT.
6442 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6443 Sandra Loosemore <sandra@codesourcery.com>
6445 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6446 * config/nios2/nios2.c (nios2_function_profiler): Add
6447 -fPIC (flag_pic == 2) support.
6448 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6449 (nios2_large_offset_p): New function.
6450 (nios2_unspec_reloc_p): Move up position, update to use
6451 nios2_large_offset_p.
6452 (nios2_unspec_address): Remove function.
6453 (nios2_unspec_offset): New function.
6454 (nios2_large_got_address): New function.
6455 (nios2_got_address): Add large offset support.
6456 (nios2_legitimize_tls_address): Update usage of removed and new
6458 (nios2_symbol_binds_local_p): New function.
6459 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
6460 (nios2_legitimize_address): Update to use nios2_large_offset_p.
6461 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
6462 (nios2_print_operand): Merge H/L processing, add hiadj/lo
6463 processing for (const (unspec ...)).
6464 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
6466 2014-02-20 Richard Biener <rguenther@suse.de>
6468 * tree-cfg.c (replace_uses_by): Mark altered BBs before
6469 doing the substitution.
6470 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
6472 2014-02-20 Martin Jambor <mjambor@suse.cz>
6475 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
6476 info when checking whether lattices are bottom.
6478 2014-02-20 Richard Biener <rguenther@suse.de>
6481 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
6484 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
6487 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
6488 parameter specifying the scaling.
6489 (inline_call): Update.
6490 (want_inline_recursively): Guard division by zero.
6491 (recursive_inlining): Update.
6492 * ipa-inline.h (clone_inlined_nodes): Update.
6494 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6497 * config/i386/i386.c (classify_argument): Pass structures of size
6498 64 bytes or less in register.
6500 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6501 Kirill Yukhin <kirill.yukhin@intel.com>
6503 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
6504 (_mm_rcp28_round_ss): Ditto.
6505 (_mm_rsqrt28_round_sd): Ditto.
6506 (_mm_rsqrt28_round_ss): Ditto.
6507 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
6508 (_mm_rcp14_round_ss): Ditto.
6509 (_mm_rsqrt14_round_sd): Ditto.
6510 (_mm_rsqrt14_round_ss): Ditto.
6511 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
6512 the first input operand, get rid of match_dup.
6513 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
6515 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
6517 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
6518 operand as the first input operand, set type attribute.
6519 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
6521 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
6522 operand as the first input operand, set type attribute.
6524 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6526 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
6529 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
6532 * config/i386/i386.c (construct_container): Remove TFmode check
6533 for X86_64_INTEGER_CLASS.
6535 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
6538 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
6539 only when -Wpsabi is enabled.
6541 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
6544 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
6545 passing arrays in registers are the same as for structs, so remove the
6546 special case for them.
6548 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
6550 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
6551 destination type, extract only the valid bits if the source type is not
6552 integral and has a different mode.
6554 2014-02-19 Richard Biener <rguenther@suse.de>
6557 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
6560 2014-02-19 Richard Biener <rguenther@suse.de>
6563 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
6564 (ipa_modify_call_arguments): Emit an argument load explicitely and
6565 preserve virtual SSA form there and for the replacement call.
6566 Do not update SSA form nor free dominance info.
6568 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6570 * ipa.c (function_and_variable_visibility): Also clear WEAK
6571 flag when disolving COMDAT_GROUP.
6573 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6575 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
6576 * ipa-prop.c (ipa_set_jf_known_type): Return early when
6578 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
6579 do more sanity checks.
6580 (detect_type_change): Return true when giving up early.
6581 (compute_complex_assign_jump_func): Fix type parameter of
6582 ipa_set_ancestor_jf.
6583 (compute_complex_ancestor_jump_func): Likewise.
6584 (update_jump_functions_after_inlining): Fix updating of
6586 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
6588 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6590 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
6591 inline clones when edge disappears.
6593 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6596 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
6597 Split 64-bit moves into 2 patterns. Do not allow the use of
6598 direct move for TDmode in little endian, since the decimal value
6599 has little endian bytes within a word, but the 64-bit pieces are
6600 ordered in a big endian fashion, and normal subreg's of TDmode are
6602 (mov<mode>_64bit_dm): Likewise.
6603 (movtd_64bit_nodm): Likewise.
6605 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6607 PR tree-optimization/60174
6608 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
6609 statement of an SSA_NAME that occurs in an abnormal PHI node.
6611 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6614 * final.c (SEEN_BB): Remove.
6615 (SEEN_NOTE, SEEN_EMITTED): Renumber.
6616 (final_scan_insn): Don't force_source_line on second
6617 NOTE_INSN_BASIC_BLOCK.
6619 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
6622 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
6623 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
6624 (type_natural_mode): Warn ABI change when %zmm register is not
6625 available for AVX512F vector value passing.
6627 2014-02-18 Kai Tietz <ktietz@redhat.com>
6630 * config/i386/i386.c (ix86_expand_prologue): Use value in
6631 rax register as displacement when restoring %r10 or %rax.
6632 Fix wrong offset when restoring both registers.
6634 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6636 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
6637 assertion with conditional return.
6639 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6640 Uros Bizjak <ubizjak@gmail.com>
6643 * config/i386/driver-i386.c (host_detect_local_cpu): If
6644 YMM state is not saved by the OS, also clear has_f16c. Move
6645 CPUID 0x80000001 handling before YMM state saving checking.
6647 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
6649 PR rtl-optimization/58960
6650 * haifa-sched.c (alloc_global_sched_pressure_data): New,
6651 factored out from ...
6652 (sched_init): ... here.
6653 (free_global_sched_pressure_data): New, factored out from ...
6654 (sched_finish): ... here.
6655 * sched-int.h (free_global_sched_pressure_data): Declare.
6656 * sched-rgn.c (nr_regions_initial): New static global.
6657 (haifa_find_rgns): Initialize it.
6658 (schedule_region): Disable sched-pressure for the newly
6661 2014-02-17 Richard Biener <rguenther@suse.de>
6663 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
6664 release SSA defs of pattern stmts.
6666 2014-02-17 Richard Biener <rguenther@suse.de>
6668 * tree-inline.c (expand_call_inline): Release the virtual
6669 operand defined by the call we are about to inline.
6671 2014-02-17 Richard Biener <rguenther@suse.de>
6673 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
6675 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
6676 Ilya Tocar <ilya.tocar@intel.com>
6678 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
6679 arguments order in builtin.
6680 (_mm512_permutexvar_epi64): Ditto.
6681 (_mm512_mask_permutexvar_epi64): Ditto
6682 (_mm512_maskz_permutexvar_epi32): Ditto
6683 (_mm512_permutexvar_epi32): Ditto
6684 (_mm512_mask_permutexvar_epi32): Ditto
6686 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6688 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
6689 (p8_vmrgow): Likewise.
6691 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6693 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
6696 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
6699 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
6700 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
6701 into 64-bit and 32-bit moves. On 64-bit moves, add support for
6702 using direct move instructions on ISA 2.07. Also adjust
6703 instruction length for 64-bit.
6704 (mov<mode>_64bit, TFmode/TDmode): Likewise.
6705 (mov<mode>_32bit, TFmode/TDmode): Likewise.
6707 2014-02-15 Alan Modra <amodra@gmail.com>
6711 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
6712 find_replacement on parts of insn rtl that might be reloaded.
6714 2014-02-15 Richard Biener <rguenther@suse.de>
6716 PR tree-optimization/60183
6717 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
6718 (tree_ssa_phiprop): Calculate and free post-dominators.
6720 2014-02-14 Jeff Law <law@redhat.com>
6722 PR rtl-optimization/60131
6723 * ree.c (get_extended_src_reg): New function.
6724 (combine_reaching_defs): Use it rather than assuming location of REG.
6725 (find_and_remove_re): Verify first operand of extension is
6726 a REG before adding the insns to the copy list.
6728 2014-02-14 Roland McGrath <mcgrathr@google.com>
6730 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
6731 * configure: Regenerated.
6732 * config.in: Regenerated.
6733 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
6734 instead of ASM_SHORT.
6736 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
6737 Richard Earnshaw <rearnsha@arm.com>
6739 PR rtl-optimization/59535
6740 * lra-constraints.c (process_alt_operands): Encourage alternative
6741 when unassigned pseudo class is superset of the alternative class.
6742 (inherit_reload_reg): Don't inherit when optimizing for code size.
6743 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
6744 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
6745 modes not less than 4 for Thumb1.
6747 2014-02-14 Kyle McMartin <kyle@redhat.com>
6750 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
6752 2014-02-14 Richard Biener <rguenther@suse.de>
6754 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
6755 (get_frame_arg): Drop the assert with langhook types_compatible_p.
6756 Do not strip INDIRECT_REFs.
6758 2014-02-14 Richard Biener <rguenther@suse.de>
6761 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
6762 DECL_FUNCTION_SPECIFIC_TARGET.
6763 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
6764 * tree-streamer-out.c (pack_ts_target_option): Remove.
6765 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
6766 (write_ts_function_decl_tree_pointers): Do not stream
6767 DECL_FUNCTION_SPECIFIC_TARGET.
6768 * tree-streamer-in.c (unpack_ts_target_option): Remove.
6769 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
6770 (lto_input_ts_function_decl_tree_pointers): Do not stream
6771 DECL_FUNCTION_SPECIFIC_TARGET.
6773 2014-02-14 Jakub Jelinek <jakub@redhat.com>
6775 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
6776 (get_initial_def_for_induction, vectorizable_induction): Ignore
6777 debug stmts when looking for exit_phi.
6778 (vectorizable_live_operation): Fix up condition.
6780 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6782 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
6783 nreverse() because it changes the content of original tree list.
6785 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6787 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
6788 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
6790 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6792 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
6793 GNU coding standards.
6795 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6798 * dwarf2out.c (gen_subprogram_die): Don't call
6799 add_calling_convention_attribute if subr_die is old_die.
6801 2014-02-13 Sharad Singhai <singhai@google.com>
6803 * doc/optinfo.texi: Fix order of nodes.
6805 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
6807 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
6808 operands[2], not operands[3].
6810 2014-02-13 Richard Biener <rguenther@suse.de>
6813 * doc/install.texi (ISL): Update recommended version to 0.12.2,
6814 mention the possibility of an in-tree build.
6815 (CLooG): Update recommended version to 0.18.1, mention the
6816 possibility of an in-tree build and clarify that the ISL
6817 bundled with CLooG does not work.
6819 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6822 * expr.c (compress_float_constant): If x is a hard register,
6823 extend into a pseudo and then move to x.
6825 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
6827 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
6828 caused by bad second argument to warning_at() with -mhotpatch and
6829 nested functions (e.g. with gfortran).
6831 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
6833 * opts.c (option_name): Remove "enabled by default" rider.
6835 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
6837 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
6839 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
6840 Uros Bizjak <ubizjak@gmail.com>
6843 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
6844 * configure: Regenerated.
6846 2014-02-12 Richard Biener <rguenther@suse.de>
6848 * vec.c (vec_prefix::calculate_allocation): Move as
6849 inline variant to vec.h.
6850 (vec_prefix::calculate_allocation_1): New out-of-line version.
6851 * vec.h (vec_prefix::calculate_allocation_1): Declare.
6852 (vec_prefix::m_has_auto_buf): Rename to ...
6853 (vec_prefix::m_using_auto_storage): ... this.
6854 (vec_prefix::calculate_allocation): Inline the easy cases
6855 and dispatch to calculate_allocation_1 which doesn't need the
6857 (va_heap::reserve): Use gcc_checking_assert.
6858 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
6859 m_using_auto_storage.
6860 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
6862 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
6863 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
6864 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
6866 2014-02-12 Richard Biener <rguenther@suse.de>
6868 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
6869 when we found a dependence.
6871 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
6873 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
6875 (maybe_fold_stmt): ... into this new function.
6876 * omp-low.c (lower_omp): Update comment.
6878 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
6881 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
6884 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
6886 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
6887 identifiers in comments.
6888 (cortexa53_extra_costs): Likewise.
6889 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
6890 (cortexa7_extra_costs): Likewise.
6891 (cortexa12_extra_costs): Likewise.
6892 (cortexa15_extra_costs): Likewise.
6893 (v7m_extra_costs): Likewise.
6895 2014-02-12 Richard Biener <rguenther@suse.de>
6898 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
6899 of posix_memalign being successful.
6900 (lower_stmt): Restrict lowering of posix_memalign to when
6901 -ftree-bit-ccp is enabled.
6903 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6905 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
6907 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
6909 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
6911 PR rtl-optimization/60116
6912 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
6913 other_insn once the combination has been validated.
6915 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
6918 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
6920 * ipa-devirt.c: Include demangle.h
6921 (odr_violation_reported): New static variable.
6922 (add_type_duplicate): Update odr_violations.
6923 (maybe_record_node): Add completep parameter; update it.
6924 (record_target_from_binfo): Add COMPLETEP parameter;
6925 update it as needed.
6926 (possible_polymorphic_call_targets_1): Likewise.
6927 (struct polymorphic_call_target_d): Add nonconstruction_targets;
6928 rename FINAL to COMPLETE.
6929 (record_targets_from_bases): Sanity check we found the binfo;
6930 fix COMPLETEP updating.
6931 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
6932 parameter, fix computing of COMPLETEP.
6933 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
6934 at LTO time do demangling.
6935 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
6936 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
6938 (gimple_get_virt_method_for_binfo): Likewise.
6939 * gimple-fold.h (gimple_get_virt_method_for_binfo,
6940 gimple_get_virt_method_for_vtable): Update prototypes.
6942 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
6945 * genautomata.c (add_presence_absence): Fix typo with
6946 {final_}presence_list.
6948 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
6951 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
6952 for VSX/Altivec vectors that land in GPR registers.
6954 2014-02-11 Richard Henderson <rth@redhat.com>
6955 Jakub Jelinek <jakub@redhat.com>
6958 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
6959 around drhs if type conversion to lacc->type is not useless.
6961 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6963 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
6965 (cortex-a57.cortex-a53): Likewise.
6966 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
6968 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6970 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
6973 2014-02-11 Renlin Li <Renlin.Li@arm.com>
6975 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
6976 add_options_for_arm_vfp3.
6978 2014-02-11 Jeff Law <law@redhat.com>
6981 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
6982 object with an undesirable mode.
6984 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6987 * config/i386/sol2-9.h: New file.
6988 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
6989 *-*-solaris2.9*): Use it.
6991 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6993 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
6994 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
6996 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
6998 * config/microblaze/microblaze.c: Extend mcpu version format
7000 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
7002 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
7004 2014-02-10 Richard Henderson <rth@redhat.com>
7007 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
7008 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
7009 ms-abi vs -mno-accumulate-outgoing-args.
7010 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
7011 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
7014 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
7017 * cfgexpand.c (expand_asm_operands): Attach source location to
7018 ASM_INPUT rtx objects.
7019 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
7021 2014-02-10 Nick Clifton <nickc@redhat.com>
7023 * config/mn10300/mn10300.c (popcount): New function.
7024 (mn10300_expand_prologue): Include saved registers in stack usage
7027 2014-02-10 Jeff Law <law@redhat.com>
7030 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
7031 when changing the SET_DEST of a prior insn to avoid an input reload.
7033 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7035 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
7036 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
7037 -mcall-openbsd, or -mcall-linux.
7038 (CC1_ENDIAN_BIG_SPEC): Remove.
7039 (CC1_ENDIAN_LITTLE_SPEC): Remove.
7040 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7041 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
7042 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
7043 and %cc1_endian_default.
7044 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7046 2014-02-10 Richard Biener <rguenther@suse.de>
7048 PR tree-optimization/60115
7049 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
7050 MEM_REF handling. Properly verify that the accesses are not
7051 out of the objects bound.
7053 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7055 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
7058 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
7060 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
7061 proper constants and fix formatting.
7062 (possible_polymorphic_call_targets): Fix formatting.
7064 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
7065 Ilya Tocar <ilya.tocar@intel.com>
7067 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
7068 (_mm512_loadu_epi32): Renamed into...
7069 (_mm512_loadu_si512): This.
7070 (_mm512_storeu_epi32): Renamed into...
7071 (_mm512_storeu_si512): This.
7072 (_mm512_maskz_ceil_ps): Removed.
7073 (_mm512_maskz_ceil_pd): Ditto.
7074 (_mm512_maskz_floor_ps): Ditto.
7075 (_mm512_maskz_floor_pd): Ditto.
7076 (_mm512_floor_round_ps): Ditto.
7077 (_mm512_floor_round_pd): Ditto.
7078 (_mm512_ceil_round_ps): Ditto.
7079 (_mm512_ceil_round_pd): Ditto.
7080 (_mm512_mask_floor_round_ps): Ditto.
7081 (_mm512_mask_floor_round_pd): Ditto.
7082 (_mm512_mask_ceil_round_ps): Ditto.
7083 (_mm512_mask_ceil_round_pd): Ditto.
7084 (_mm512_maskz_floor_round_ps): Ditto.
7085 (_mm512_maskz_floor_round_pd): Ditto.
7086 (_mm512_maskz_ceil_round_ps): Ditto.
7087 (_mm512_maskz_ceil_round_pd): Ditto.
7088 (_mm512_expand_pd): Ditto.
7089 (_mm512_expand_ps): Ditto.
7090 * config/i386/i386.c (ix86_builtins): Remove
7091 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
7092 (bdesc_args): Ditto.
7093 * config/i386/predicates.md (const1256_operand): New.
7094 (const_1_to_2_operand): Ditto.
7095 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7096 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7097 (*avx512pf_gatherpf<mode>sf): Ditto.
7098 (avx512pf_gatherpf<mode>df): Ditto.
7099 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7100 (*avx512pf_gatherpf<mode>df): Ditto.
7101 (avx512pf_scatterpf<mode>sf): Ditto.
7102 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7103 (*avx512pf_scatterpf<mode>sf): Ditto.
7104 (avx512pf_scatterpf<mode>df): Ditto.
7105 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7106 (*avx512pf_scatterpf<mode>df): Ditto.
7107 (avx512f_expand<mode>): Removed.
7108 (<shift_insn><mode>3<mask_name>): Change predicate type.
7110 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7112 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
7113 not at the end of datarefs vector use ordered_remove to avoid
7114 reordering datarefs vector.
7117 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
7118 mark local addressable non-static vars as GOVD_PRIVATE
7119 instead of GOVD_LOCAL.
7120 * omp-low.c (lower_omp_for): Move gimple_bind_vars
7121 and BLOCK_VARS of gimple_bind_block to new_stmt rather
7125 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
7126 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
7127 assume_aligned or alloc_align attributes.
7128 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
7129 arguments. Handle also assume_aligned and alloc_align attributes.
7130 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
7131 calls to functions with assume_aligned or alloc_align attributes.
7132 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
7134 2014-02-08 Terry Guo <terry.guo@arm.com>
7136 * doc/invoke.texi: Document ARM -march=armv7e-m.
7138 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7140 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
7141 flag on __cilkrts_rethrow builtin.
7144 * ipa-cp.c (determine_versionability): Fail at -O0
7145 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
7146 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
7149 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7152 * tree-inline.c (copy_forbidden): Fail for
7153 __attribute__((optimize (0))) functions.
7155 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7157 * varpool.c: Include pointer-set.h.
7158 (varpool_remove_unreferenced_decls): Variables in other partitions
7159 will not be output; be however careful to not lose information
7162 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7164 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
7165 lookup in the vtable constructor.
7167 2014-02-07 Jeff Law <law@redhat.com>
7170 * config/m68k/m68k.md (ashldi_extsi): Turn into a
7171 define_insn_and_split.
7173 * ipa-inline.c (inline_small_functions): Fix typos.
7175 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7177 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
7178 (s390_can_use_return_insn): Declare.
7179 * config/s390/s390.h (EPILOGUE_USES): Define.
7180 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
7182 (s390_chunkify_start): Handle return JUMP_LABELs.
7183 (s390_early_mach): Emit a main_pool instruction on the entry edge.
7184 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
7185 (s390_can_use_return_insn): New functions.
7186 (s390_fix_long_loop_prediction): Handle conditional returns.
7187 (TARGET_SET_UP_BY_PROLOGUE): Define.
7188 * config/s390/s390.md (ANY_RETURN): New code iterator.
7189 (*creturn, *csimple_return, return, simple_return): New patterns.
7191 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7193 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
7194 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
7195 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
7196 REG_CFA_RESTORE list when deciding not to restore a register.
7198 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7200 * config/s390/s390.c: Include tree-pass.h and context.h.
7201 (s390_early_mach): New function, split out from...
7202 (s390_emit_prologue): ...here.
7203 (pass_data_s390_early_mach): New pass structure.
7204 (pass_s390_early_mach): New class.
7205 (s390_option_override): Create and register early_mach pass.
7206 Move to end of file.
7208 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7210 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
7211 to match for the exit block.
7213 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7215 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
7216 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
7217 Reject misaligned operands.
7219 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7221 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
7223 2014-02-07 Richard Biener <rguenther@suse.de>
7226 * gimple-low.c (lower_builtin_posix_memalign): New function.
7227 (lower_stmt): Call it to lower posix_memalign in a way
7228 to make alignment info accessible.
7230 2014-02-07 Jakub Jelinek <jakub@redhat.com>
7233 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
7234 __builtin_setjmp_receiver.
7236 2014-02-07 Richard Biener <rguenther@suse.de>
7239 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
7240 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
7241 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7242 Handle BUILT_IN_POSIX_MEMALIGN.
7243 (find_func_clobbers): Likewise.
7244 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
7245 (call_may_clobber_ref_p_1): Likewise.
7247 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7250 * ipa-devirt.c (record_target_from_binfo): Remove overactive
7253 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7256 * lto-cgraph.c (lto_output_node): Use
7257 symtab_get_symbol_partitioning_class.
7258 (lto_output_varpool_node): likewise.
7259 (symtab_get_symbol_partitioning_class): Move here from
7261 * cgraph.h (symbol_partitioning_class): Likewise.
7262 (symtab_get_symbol_partitioning_class): Declare.
7264 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7266 * ggc.h (ggc_internal_cleared_alloc): New macro.
7267 * vec.h (vec_safe_copy): Handle memory stats.
7268 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
7269 * target-globals.c (save_target_globals): Likewise.
7271 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7274 * expr.c (emit_move_resolve_push): Export; be bit more selective
7275 on when to clear alias set.
7276 * expr.h (emit_move_resolve_push): Declare.
7277 * function.h (struct function): Add tail_call_marked.
7278 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
7279 * config/i386/i386-protos.h (ix86_expand_push): Remove.
7280 * config/i386/i386.md (TImode move expander): De not call
7282 (FP push expanders): Preserve memory attributes.
7283 * config/i386/sse.md (push<mode>1): Remove.
7284 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
7285 (ix86_expand_push): Remove.
7286 * config/i386/mmx.md (push<mode>1): Remove.
7288 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7290 PR rtl-optimization/60030
7291 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
7292 lopart with paradoxical subreg before shifting it up by hprec.
7294 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7296 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
7297 Remove extra newline at end of file.
7298 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
7299 (arm_issue_rate): Handle cortexa57.
7300 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
7301 (cortex-a57.cortex-a53): Likewise.
7303 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7306 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
7307 don't record in REG_FRAME_RELATED_EXPR registers not set in that
7309 (arm_expand_prologue): Adjust all callers.
7310 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
7311 info, registers also at the lowest numbered registers side. Use
7312 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
7316 * var-tracking.c (adjust_mems): Before adding a SET to
7317 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
7319 2014-02-06 Alan Modra <amodra@gmail.com>
7322 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
7323 change SDmode to DDmode when lra_in_progress.
7325 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7328 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
7329 free_data_ref on the dr first, and before goto again also set dr
7330 to the next dr. For simd_lane_access, free old datarefs[i] before
7331 overwriting it. For get_vectype_for_scalar_type failure, don't
7332 free_data_ref if simd_lane_access.
7334 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
7337 * tree.h (opts_for_fn): New inline function.
7338 (opt_for_fn): Define.
7339 * config/i386/i386.c (ix86_function_regparm): Use
7340 opt_for_fn (decl, optimize) instead of optimize.
7342 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
7344 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
7345 for SYMBOL_REF in large memory model.
7347 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7349 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
7351 (cortex-a57): Likewise.
7352 (cortex-a57.cortex-a53): Likewise.
7354 2014-02-06 Yury Gribov <y.gribov@samsung.com>
7355 Kugan Vivekanandarajah <kuganv@linaro.org>
7357 * config/arm/arm.c (arm_vector_alignment_reachable): Check
7359 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
7361 2014-02-06 Richard Biener <rguenther@suse.de>
7363 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
7364 set_loop_copy and initialize_original_copy_tables.
7366 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
7368 * config/aarch64/aarch64-simd.md
7369 (aarch64_ashr_simddi): Change QI to SI.
7371 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7372 Jakub Jelinek <jakub@redhat.com>
7375 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
7378 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7380 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
7381 CODE_FOR_altivec_vpku[hw]um to
7382 CODE_FOR_altivec_vpku[hw]um_direct.
7383 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
7384 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
7385 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
7386 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
7388 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7390 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
7391 generation for -maltivec=be.
7392 (altivec_vsumsws): Simplify redundant test.
7394 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7396 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
7397 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
7398 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
7399 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
7400 gen_altivec_vpkuwum.
7401 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
7403 (altivec_vpks<VI_char>ss): Likewise.
7404 (altivec_vpks<VI_char>us): Likewise.
7405 (altivec_vpku<VI_char>us): Likewise.
7406 (altivec_vpku<VI_char>um): Likewise.
7407 (altivec_vpku<VI_char>um_direct): New (copy of
7408 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
7410 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
7411 target is little endian and -maltivec=be is not specified.
7412 (*altivec_vupkhs<VU_char>_direct): New (copy of
7413 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
7414 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
7415 target is little endian and -maltivec=be is not specified.
7416 (*altivec_vupkls<VU_char>_direct): New (copy of
7417 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
7418 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
7419 little endian and -maltivec=be is not specified.
7420 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
7421 little endian and -maltivec=be is not specified.
7423 2014-02-05 Richard Henderson <rth@redhat.com>
7426 * combine-stack-adj.c: Revert r206943.
7427 * sched-int.h (struct deps_desc): Add last_args_size.
7428 * sched-deps.c (init_deps): Initialize it.
7429 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7430 contain REG_ARGS_SIZE notes.
7432 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7434 * lto-cgraph.c (asm_nodes_output): Make global.
7435 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7436 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7437 (driver_handle_option): Handle OPT_fwpa.
7439 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7442 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7443 a comment typo and formatting issue. If odr_hash hasn't been
7444 created, return vNULL and set *completep to false.
7447 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7448 bb with no successors.
7450 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7453 * doc/invoke.texi (-march): Clarify documentation for ARM.
7457 2014-02-05 Richard Biener <rguenther@suse.de>
7459 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
7460 when not vectorizing because of too many alias checks.
7461 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
7462 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
7464 2014-02-05 Nick Clifton <nickc@redhat.com>
7466 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
7467 accept extended registers in any mode when compiling for the MN10300.
7469 2014-02-05 Yury Gribov <y.gribov@samsung.com>
7471 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
7472 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
7473 sanitization attributes.
7474 (can_inline_edge_p): Likewise.
7475 (sanitize_attrs_match_for_inline_p): New function.
7477 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7479 * ipa-prop.c (detect_type_change): Shor circuit testing of
7480 type changes on THIS pointer.
7482 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
7485 * config/pa/pa.c (legitimize_tls_address): Return original address
7486 if not passed a SYMBOL_REF rtx.
7487 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
7489 (pa_emit_move_sequence): Simplify TLS source operands.
7490 (pa_legitimate_constant_p): Reject all TLS constants.
7491 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
7492 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
7494 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7496 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
7497 groups when we know they are controlled by LTO.
7498 * varasm.c (default_binds_local_p_1): If object is in other partition,
7499 it will be resolved locally.
7501 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7503 * config/host-linux.c (linux_gt_pch_use_address): Don't
7504 use SSIZE_MAX because it is not always defined.
7506 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
7509 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
7510 threshold for pseudo splitting.
7511 (update_ebb_live_info): Process call argument hard registers and
7512 hard registers from insn definition too.
7513 (max_small_class_regs_num): New constant.
7514 (inherit_in_ebb): Update live hard regs through EBBs. Update
7515 reloads_num only for small register classes. Don't split for
7518 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
7521 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
7524 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7526 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
7529 2014-02-04 Marek Polacek <polacek@redhat.com>
7531 * gdbinit.in (pel): Define.
7533 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7535 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
7538 2014-02-04 Richard Biener <rguenther@suse.de>
7541 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
7542 in function context local.
7543 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
7544 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
7545 similar to LTO_imported_decl_ref.
7547 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7549 PR tree-optimization/60002
7550 * cgraphclones.c (build_function_decl_skip_args): Clear
7553 PR tree-optimization/60023
7554 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
7555 false to gsi_replace.
7556 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
7557 has been in some EH region and vec_stmt could throw, add
7558 vec_stmt into the same EH region.
7559 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
7560 has no lhs, ignore it.
7561 * internal-fn.c (expand_MASK_LOAD): Likewise.
7564 * tree-inline.c (copy_forbidden): Fail for
7565 __attribute__((optimize (0))) functions.
7568 * omp-low.c (simd_clone_struct_copy): If from->inbranch
7569 is set, copy one less argument.
7570 (expand_simd_clones): Don't subtract clone_info->inbranch
7571 from simd_clone_struct_alloc argument.
7573 PR rtl-optimization/57915
7574 * recog.c (simplify_while_replacing): If all unary/binary/relational
7575 operation arguments are constant, attempt to simplify those.
7578 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
7579 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
7581 2014-02-04 Richard Biener <rguenther@suse.de>
7583 PR tree-optimization/60012
7584 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
7585 TBAA disambiguation to all DDRs.
7587 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7590 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
7591 (LINK_SPEC): Use it for -shared, -shared-libgcc.
7593 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7596 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
7598 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7600 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
7601 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
7603 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7606 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
7607 to figure out targets of polymorphic calls with known decl.
7608 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7609 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
7610 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
7611 (get_polymorphic_call_info): ... here.
7612 (get_polymorphic_call_info_from_invariant): New function.
7614 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7616 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
7617 lookup via vtable pointer; check for type consistency
7618 and turn inconsitent facts into UNREACHABLE.
7619 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7620 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
7621 type inconsistent querries; return UNREACHABLE instead.
7623 2014-02-03 Richard Henderson <rth@twiddle.net>
7626 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
7627 already processed this node.
7628 (normalize_one_pred_1): Pass along mark_set.
7629 (normalize_one_pred): Create and destroy a pointer_set_t.
7630 (normalize_one_pred_chain): Likewise.
7632 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
7634 PR gcov-profile/58602
7635 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
7637 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7640 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
7641 -fno-devirtualize; try to devirtualize by the knowledge of
7642 virtual table pointer given by aggregate propagation.
7643 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7644 (ipa_print_node_jump_functions): Dump also offset that
7645 is relevant for polymorphic calls.
7646 (determine_known_aggregate_parts): Add arg_type parameter; use it
7647 instead of determining the type from pointer type.
7648 (ipa_compute_jump_functions_for_edge): Update call of
7649 determine_known_aggregate_parts.
7650 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
7651 (gimple_get_virt_method_for_binfo): ... here; simplify using
7652 vtable_pointer_value_to_vtable.
7653 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
7654 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
7655 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
7656 (vtable_pointer_value_to_vtable): Break out from ...; handle also
7658 (vtable_pointer_value_to_binfo): ... here.
7659 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
7661 2014-02-03 Teresa Johnson <tejohnson@google.com>
7663 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
7664 redef of outer loop index variable.
7666 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
7670 * doc/extend.texi (Function Attributes): Typo.
7672 2014-02-03 Cong Hou <congh@google.com>
7674 PR tree-optimization/60000
7675 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
7676 if the vectorized statement is a store. A store statement can only
7677 appear at the end of pattern statements.
7679 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7681 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
7682 (ix86_option_override_internal): Default long double to 64-bit for
7683 32-bit Bionic and to 128-bit for 64-bit Bionic.
7685 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
7686 TARGET_LONG_DOUBLE_128 is true.
7687 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
7689 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
7690 (mlong-double-64): Negate -mlong-double-128.
7691 (mlong-double-128): New option.
7693 * config/i386/i386-c.c (ix86_target_macros): Define
7694 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
7696 * doc/invoke.texi: Document -mlong-double-128.
7698 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7700 PR rtl-optimization/60024
7701 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
7703 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
7705 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
7707 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7709 PR rtl-optimization/57662
7710 * sel-sched.c (code_motion_path_driver): Do not mark already not
7711 existing blocks in the visiting bitmap.
7713 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7715 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
7716 on the insn being emitted.
7718 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
7719 Will Deacon <will.deacon@arm.com>
7721 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
7723 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7725 * config/arm/arm-tables.opt: Regenerate.
7727 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7729 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
7730 for vector types other than V16QImode.
7731 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
7732 define_expand, and call altivec_expand_vec_perm_le when producing
7733 code with little endian element order.
7734 (*altivec_vperm_<mode>_internal): New insn having previous
7735 behavior of altivec_vperm_<mode>.
7736 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
7737 altivec_expand_vec_perm_le when producing code with little endian
7739 (*altivec_vperm_<mode>_uns_internal): New insn having previous
7740 behavior of altivec_vperm_<mode>_uns.
7742 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7744 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
7745 (altivec_vsumsws): Add handling for -maltivec=be with a little
7747 (altivec_vsumsws_direct): New.
7748 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
7749 gen_altivec_vsumsws.
7751 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7753 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
7754 vtable_pointer_value_to_binfo): New functions.
7755 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
7756 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
7758 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
7760 * config/nios2/nios2.md (load_got_register): Initialize GOT
7761 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
7762 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
7764 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7766 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
7767 preserverd by passthrough, do not propagate the type.
7769 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7771 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
7772 (mips_atomic_assign_expand_fenv): New function.
7773 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
7775 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7777 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
7778 (__builtin_mips_set_fcsr): Likewise.
7779 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
7780 MIPS_USI_FTYPE_VOID.
7781 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
7782 (mips16_expand_set_fcsr): Likewise.
7783 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
7784 (mips16_set_fcsr_stub): Likewise.
7785 (mips16_get_fcsr_one_only_stub): New class.
7786 (mips16_set_fcsr_one_only_stub): Likewise.
7787 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
7788 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
7789 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
7790 (hard_float): New availability predicate.
7791 (mips_builtins): Add get_fcsr and set_fcsr.
7792 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
7793 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
7794 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
7795 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
7796 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
7799 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7801 * config/mips/mips.c (mips_one_only_stub): New class.
7802 (mips_need_mips16_rdhwr_p): Replace with...
7803 (mips16_rdhwr_stub): ...this new variable.
7804 (mips16_stub_call_address): New function.
7805 (mips16_rdhwr_one_only_stub): New class.
7806 (mips_expand_thread_pointer): Use mips16_stub_call_address.
7807 (mips_output_mips16_rdhwr): Delete.
7808 (mips_finish_stub): New function.
7809 (mips_code_end): Use it to handle rdhwr stubs.
7811 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
7814 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
7815 when calculating size of integer atomic types.
7817 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
7819 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
7821 2014-02-01 Jakub Jelinek <jakub@redhat.com>
7823 PR tree-optimization/60003
7824 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
7825 * profile.c (branch_prob): Use gimple_call_builtin_p
7826 to check for BUILT_IN_SETJMP_RECEIVER.
7827 * tree-inline.c (copy_bb): Call notice_special_calls.
7829 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
7832 * lra-constraints.c (process_alt_operands): Update reload_sum only
7835 2014-01-31 Richard Henderson <rth@redhat.com>
7838 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
7839 until after else_eh is processed.
7841 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7843 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
7844 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
7845 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
7846 in smmintrin.h, remove them.
7847 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
7848 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
7849 * config/i386/i386.md (ROUND_SAE): Fix value.
7850 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
7851 (const48_operand): New.
7852 * config/i386/subst.md (round), (round_expand): Use
7853 const_4_or_8_to_11_operand.
7854 (round_saeonly), (round_saeonly_expand): Use const48_operand.
7856 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7858 * config/i386/constraints.md (Yk): Swap meaning with k.
7859 * config/i386/i386.md (movhi_internal): Change Yk to k.
7860 (movqi_internal): Ditto.
7861 (*k<logic><mode>): Ditto.
7864 (kandn<mode>): Ditto.
7865 (*<code>hi_1): Ditto.
7866 (*<code>qi_1): Ditto.
7867 (kxnor<mode>): Ditto.
7868 (kortestzhi): Ditto.
7869 (kortestchi): Ditto.
7871 (*one_cmplhi2_1): Ditto.
7872 (*one_cmplqi2_1): Ditto.
7873 * config/i386/sse.md (): Change k to Yk.
7874 (avx512f_load<mode>_mask): Ditto.
7875 (avx512f_blendm<mode>): Ditto.
7876 (avx512f_store<mode>_mask): Ditto.
7877 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
7878 (avx512f_storedqu<mode>_mask): Ditto.
7879 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
7881 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
7882 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
7883 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
7884 (avx512f_maskcmp<mode>3): Ditto.
7885 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
7886 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
7887 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
7888 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
7889 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
7890 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
7891 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
7892 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
7893 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
7894 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
7895 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
7896 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
7897 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
7898 (vec_extract_lo_<mode>_maskm): Ditto.
7899 (vec_extract_hi_<mode>_maskm): Ditto.
7900 (avx512f_vternlog<mode>_mask): Ditto.
7901 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
7902 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
7903 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
7904 (avx512f_<code>v8div16qi2_mask): Ditto.
7905 (avx512f_<code>v8div16qi2_mask_store): Ditto.
7906 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
7907 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
7908 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
7909 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
7910 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7911 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7912 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7913 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7914 (avx512cd_maskb_vec_dupv8di): Ditto.
7915 (avx512cd_maskw_vec_dupv16si): Ditto.
7916 (avx512f_vpermi2var<mode>3_maskz): Ditto.
7917 (avx512f_vpermi2var<mode>3_mask): Ditto.
7918 (avx512f_vpermi2var<mode>3_mask): Ditto.
7919 (avx512f_vpermt2var<mode>3_maskz): Ditto.
7920 (*avx512f_gathersi<mode>): Ditto.
7921 (*avx512f_gathersi<mode>_2): Ditto.
7922 (*avx512f_gatherdi<mode>): Ditto.
7923 (*avx512f_gatherdi<mode>_2): Ditto.
7924 (*avx512f_scattersi<mode>): Ditto.
7925 (*avx512f_scatterdi<mode>): Ditto.
7926 (avx512f_compress<mode>_mask): Ditto.
7927 (avx512f_compressstore<mode>_mask): Ditto.
7928 (avx512f_expand<mode>_mask): Ditto.
7929 * config/i386/subst.md (mask): Change k to Yk.
7930 (mask_scalar_merge): Ditto.
7933 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
7935 * doc/extend.texi (Vector Extensions): Document ?: in C++.
7937 2014-01-31 Richard Biener <rguenther@suse.de>
7940 * builtins.c (fold_builtin_memory_op): Make sure to not
7941 use a floating-point mode or a boolean or enumeral type for
7944 2014-01-30 DJ Delorie <dj@redhat.com>
7946 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
7947 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
7948 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
7949 whenever main() has an epilogue.
7951 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7953 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
7954 unused variable "field".
7955 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
7956 (vsx_mergeh_<mode>): Likewise.
7957 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
7958 (altivec_vmrghh): Likewise.
7959 (altivec_vmrghw): Likewise.
7960 (altivec_vmrglb): Likewise.
7961 (altivec_vmrglh): Likewise.
7962 (altivec_vmrglw): Likewise.
7963 (altivec_vspltb): Add missing uses.
7964 (altivec_vsplth): Likewise.
7965 (altivec_vspltw): Likewise.
7966 (altivec_vspltsf): Likewise.
7968 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7971 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
7972 frame related instructions.
7974 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
7976 PR rtl-optimization/59959
7977 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
7978 any reload of register whose subreg is invalid.
7980 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7982 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
7983 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
7984 Add missing return type - void.
7986 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7988 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
7989 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
7990 remove element index adjustment for endian (now handled in vsx.md
7992 (altivec_expand_vec_perm_const): Use
7993 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
7994 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
7995 (vsx_xxspltw_<mode>): Adjust element index for little endian.
7996 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
7997 define_expand and a new define_insn *altivec_vspltb_internal;
7998 adjust for -maltivec=be on a little endian target.
7999 (altivec_vspltb_direct): New.
8000 (altivec_vsplth): Divide into a define_expand and a new
8001 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
8002 little endian target.
8003 (altivec_vsplth_direct): New.
8004 (altivec_vspltw): Divide into a define_expand and a new
8005 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
8006 little endian target.
8007 (altivec_vspltw_direct): New.
8008 (altivec_vspltsf): Divide into a define_expand and a new
8009 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
8010 a little endian target.
8012 2014-01-30 Richard Biener <rguenther@suse.de>
8014 PR tree-optimization/59993
8015 * tree-ssa-forwprop.c (associate_pointerplus): Check we
8016 can propagate form the earlier stmt and avoid the transform
8017 when the intermediate result is needed.
8019 2014-01-30 Alangi Derick <alangiderick@gmail.com>
8021 * README.Portability: Fix typo.
8023 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
8025 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
8026 comparison_operator with ordered_comparison_operator.
8028 2014-01-30 Nick Clifton <nickc@redhat.com>
8030 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
8031 Rename to mn10300_store_multiple_regs.
8032 * config/mn10300/mn10300.c: Likewise.
8033 * config/mn10300/mn10300.md (store_movm): Fix typo: call
8034 store_multiple_regs.
8035 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
8036 Call mn10300_store_multiple_regs.
8038 2014-01-30 Nick Clifton <nickc@redhat.com>
8039 DJ Delorie <dj@redhat.com>
8041 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
8042 %fp 2 to keep registers after it properly word-aligned.
8043 (rl78_alloc_physical_registers_umul): Handle the case where both
8044 input operands are the same.
8046 2014-01-30 Richard Biener <rguenther@suse.de>
8048 PR tree-optimization/59903
8049 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
8052 2014-01-30 Jason Merrill <jason@redhat.com>
8055 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
8058 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
8060 2014-01-30 Richard Biener <rguenther@suse.de>
8062 PR tree-optimization/59951
8063 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
8065 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
8068 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
8069 SFmode to DFmode case.
8071 2014-01-29 DJ Delorie <dj@redhat.com>
8073 * config/msp430/msp430.opt (-minrt): New.
8074 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
8076 (ENDFILE_SPEC): Likewise.
8078 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
8080 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
8081 (estimate_function_body_sizes): Use it.
8083 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
8086 * dwarf2out.c (is_cxx_auto): New.
8087 (is_base_type): Use it.
8088 (gen_type_die_with_usage): Likewise.
8090 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8092 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
8093 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
8094 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
8095 -maltivec=be with LE targets.
8096 (vsx_mergeh_<mode>): Likewise.
8097 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
8098 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
8099 (altivec_vmrghb): Replace with define_expand and new
8100 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
8101 (altivec_vmrghb_direct): New define_insn.
8102 (altivec_vmrghh): Replace with define_expand and new
8103 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
8104 (altivec_vmrghh_direct): New define_insn.
8105 (altivec_vmrghw): Replace with define_expand and new
8106 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
8107 (altivec_vmrghw_direct): New define_insn.
8108 (*altivec_vmrghsf): Adjust for endianness.
8109 (altivec_vmrglb): Replace with define_expand and new
8110 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
8111 (altivec_vmrglb_direct): New define_insn.
8112 (altivec_vmrglh): Replace with define_expand and new
8113 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
8114 (altivec_vmrglh_direct): New define_insn.
8115 (altivec_vmrglw): Replace with define_expand and new
8116 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
8117 (altivec_vmrglw_direct): New define_insn.
8118 (*altivec_vmrglsf): Adjust for endianness.
8119 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8120 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8121 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8122 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8123 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8124 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8125 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8126 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8128 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
8130 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
8131 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
8134 2014-01-29 Richard Biener <rguenther@suse.de>
8136 PR tree-optimization/58742
8137 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
8138 associate_pointerplus_align.
8139 (associate_pointerplus_diff): New function.
8140 (associate_pointerplus): Likewise. Call associate_pointerplus_align
8141 and associate_pointerplus_diff.
8143 2014-01-29 Richard Biener <rguenther@suse.de>
8145 * lto-streamer.h (LTO_major_version): Bump to 3.
8146 (LTO_minor_version): Reset to 0.
8148 2014-01-29 Renlin Li <Renlin.Li@arm.com>
8150 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
8151 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
8152 (arm_file_start): Generate correct asm header for armv7ve.
8153 * config/arm/bpabi.h: Add multilib support for armv7ve.
8154 * config/arm/driver-arm.c: Change the architectures of cortex-a7
8155 and cortex-a15 to armv7ve.
8156 * config/arm/t-aprofile: Add multilib support for armv7ve.
8157 * doc/invoke.texi: Document -march=armv7ve.
8159 2014-01-29 Richard Biener <rguenther@suse.de>
8161 PR tree-optimization/58742
8162 * tree-ssa-forwprop.c (associate_plusminus): Return true
8163 if we changed sth, defer EH cleanup to ...
8164 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
8165 (simplify_mult): New function.
8167 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8170 PR tree-optimization/59920
8171 * tree.c (build_common_builtin_nodes): Remove
8172 __builtin_setjmp_dispatcher initialization.
8173 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
8174 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
8175 instead of gsi_after_labels + manually skipping debug stmts.
8176 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
8177 ignore bbs with IFN_ABNORMAL_DISPATCHER.
8178 * tree-inline.c (copy_edges_for_bb): Remove
8179 can_make_abnormal_goto argument, instead add abnormal_goto_dest
8180 argument. Ignore computed_goto_p stmts. Don't call
8181 make_abnormal_goto_edges. If a call might need abnormal edges
8182 for non-local gotos, see if it already has an edge to
8183 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
8184 with true argument, don't do anything then, otherwise add
8185 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
8186 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
8188 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
8189 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
8190 (lower_stmt): Don't set data->calls_builtin_setjmp.
8191 (lower_builtin_setjmp): Adjust comment.
8192 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
8193 * tree-cfg.c (found_computed_goto): Remove.
8194 (factor_computed_gotos): Remove.
8195 (make_goto_expr_edges): Return bool, true for computed gotos.
8196 Don't call make_abnormal_goto_edges.
8197 (build_gimple_cfg): Don't set found_computed_goto, don't call
8198 factor_computed_gotos.
8199 (computed_goto_p): No longer static.
8200 (make_blocks): Don't set found_computed_goto.
8201 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
8202 (make_edges): If make_goto_expr_edges returns true, push bb
8203 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
8204 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
8205 vector. Record mapping between bbs and OpenMP regions if there
8206 are any, adjust make_gimple_omp_edges caller. Call
8207 handle_abnormal_edges.
8208 (make_abnormal_goto_edges): Remove.
8209 * tree-cfg.h (make_abnormal_goto_edges): Remove.
8210 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
8211 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
8212 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
8213 * internal-fn.def (ABNORMAL_DISPATCHER): New.
8214 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
8215 filling *region also set *region_idx to (*region)->entry->index.
8218 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
8219 For REGs set ORIGINAL_REGNO.
8221 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
8223 * doc/md.texi: Mention that a target shouldn't implement
8224 vec_widen_(s|u)mul_even/odd pair if it is less efficient
8227 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8229 PR tree-optimization/59594
8230 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
8231 a copy of the datarefs vector rather than the vector itself.
8233 2014-01-28 Jason Merrill <jason@redhat.com>
8236 * dwarf2out.c (auto_die): New static.
8237 (gen_type_die_with_usage): Handle C++1y 'auto'.
8238 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
8241 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
8244 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
8245 (SPEC_X32): Likewise.
8246 (SPEC_64): Likewise.
8247 * config/i386/i386.c (ix86_option_override_internal): Turn off
8248 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
8250 (x86_file_start): Output .code16gcc for TARGET_16BIT.
8251 * config/i386/i386.h (TARGET_16BIT): New macro.
8252 (TARGET_16BIT_P): Likewise.
8253 * config/i386/i386.opt: Add m16.
8254 * doc/invoke.texi: Document -m16.
8256 2014-01-28 Jakub Jelinek <jakub@redhat.com>
8258 PR preprocessor/59935
8259 * input.c (location_get_source_line): Bail out on when line number
8260 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
8262 2014-01-28 Richard Biener <rguenther@suse.de>
8264 PR tree-optimization/58742
8265 * tree-ssa-forwprop.c (associate_plusminus): Handle
8266 pointer subtraction of the form (T)(P + A) - (T)P.
8268 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8270 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
8273 2014-01-28 Richard Biener <rguenther@suse.de>
8276 2014-01-28 Richard Biener <rguenther@suse.de>
8278 PR rtl-optimization/45364
8279 PR rtl-optimization/59890
8280 * var-tracking.c (local_get_addr_clear_given_value): Handle
8281 already cleared slot.
8282 (val_reset): Handle not allocated local_get_addr_cache.
8283 (vt_find_locations): Use post-order on the inverted CFG.
8285 2014-01-28 Richard Biener <rguenther@suse.de>
8287 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
8289 2014-01-28 Richard Biener <rguenther@suse.de>
8291 PR rtl-optimization/45364
8292 PR rtl-optimization/59890
8293 * var-tracking.c (local_get_addr_clear_given_value): Handle
8294 already cleared slot.
8295 (val_reset): Handle not allocated local_get_addr_cache.
8296 (vt_find_locations): Use post-order on the inverted CFG.
8298 2014-01-28 Alan Modra <amodra@gmail.com>
8300 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
8301 * configure.ac <recursive call for build != host>: Define
8302 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
8303 and LD_FOR_BUILD too.
8304 * configure: Regenerate.
8306 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
8308 * config/i386/i386.c (get_builtin_code_for_version): Separate
8309 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
8310 Broadwell from Haswell.
8312 2014-01-27 Steve Ellcey <sellcey@mips.com>
8314 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
8315 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
8316 * config/mips/mips.c (mips_option_override): Change setting
8318 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
8319 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
8320 Change from Mask to Var.
8322 2014-01-27 Jeff Law <law@redhat.com>
8324 * ipa-inline.c (inline_small_functions): Fix typo.
8326 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
8328 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
8329 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
8330 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
8331 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
8332 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
8333 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
8334 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
8335 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
8336 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
8337 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
8338 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
8339 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
8340 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
8341 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
8342 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
8343 (_mm512_storeu_epi64): Ditto.
8344 (_mm512_cmpge_epi32_mask): Ditto.
8345 (_mm512_cmpge_epu32_mask): Ditto.
8346 (_mm512_cmpge_epi64_mask): Ditto.
8347 (_mm512_cmpge_epu64_mask): Ditto.
8348 (_mm512_cmple_epi32_mask): Ditto.
8349 (_mm512_cmple_epu32_mask): Ditto.
8350 (_mm512_cmple_epi64_mask): Ditto.
8351 (_mm512_cmple_epu64_mask): Ditto.
8352 (_mm512_cmplt_epi32_mask): Ditto.
8353 (_mm512_cmplt_epu32_mask): Ditto.
8354 (_mm512_cmplt_epi64_mask): Ditto.
8355 (_mm512_cmplt_epu64_mask): Ditto.
8356 (_mm512_cmpneq_epi32_mask): Ditto.
8357 (_mm512_cmpneq_epu32_mask): Ditto.
8358 (_mm512_cmpneq_epi64_mask): Ditto.
8359 (_mm512_cmpneq_epu64_mask): Ditto.
8360 (_mm512_expand_pd): Ditto.
8361 (_mm512_expand_ps): Ditto.
8362 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
8363 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
8364 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
8365 * config/i386/i386.c (ix86_builtins): Add
8366 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
8367 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
8368 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
8369 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
8370 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
8371 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
8372 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
8373 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
8374 IX86_BUILTIN_PMOVUSQW512_MEM.
8375 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
8376 __builtin_ia32_pmovsqd512mem_mask,
8377 __builtin_ia32_pmovqd512mem_mask,
8378 __builtin_ia32_pmovusqw512mem_mask,
8379 __builtin_ia32_pmovsqw512mem_mask,
8380 __builtin_ia32_pmovqw512mem_mask,
8381 __builtin_ia32_pmovusdw512mem_mask,
8382 __builtin_ia32_pmovsdw512mem_mask,
8383 __builtin_ia32_pmovdw512mem_mask,
8384 __builtin_ia32_pmovqb512mem_mask,
8385 __builtin_ia32_pmovusqb512mem_mask,
8386 __builtin_ia32_pmovsqb512mem_mask,
8387 __builtin_ia32_pmovusdb512mem_mask,
8388 __builtin_ia32_pmovsdb512mem_mask,
8389 __builtin_ia32_pmovdb512mem_mask.
8390 (bdesc_args): Add __builtin_ia32_expanddf512,
8391 __builtin_ia32_expandsf512.
8392 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
8393 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
8394 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
8395 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
8396 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
8397 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
8398 (avx512f_<code>v8div16qi2_mask_store): This.
8399 (avx512f_expand<mode>): New.
8401 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
8403 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
8405 (_mm512_mask_prefetch_i64gather_pd): Ditto.
8406 (_mm512_prefetch_i32scatter_pd): Ditto.
8407 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
8408 (_mm512_prefetch_i64scatter_pd): Ditto.
8409 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
8410 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
8411 (_mm512_mask_prefetch_i64gather_ps): Ditto.
8412 (_mm512_prefetch_i32scatter_ps): Ditto.
8413 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
8414 (_mm512_prefetch_i64scatter_ps): Ditto.
8415 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
8416 * config/i386/i386-builtin-types.def: Define
8417 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
8418 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
8419 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
8420 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
8421 IX86_BUILTIN_SCATTERPFQPD.
8422 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
8423 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
8424 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
8425 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
8426 __builtin_ia32_scatterpfqps.
8427 (ix86_expand_builtin): Expand new built-ins.
8428 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8429 fix memory access data type.
8430 (*avx512pf_gatherpf<mode>_mask): Ditto.
8431 (*avx512pf_gatherpf<mode>): Ditto.
8432 (avx512pf_scatterpf<mode>): Ditto.
8433 (*avx512pf_scatterpf<mode>_mask): Ditto.
8434 (*avx512pf_scatterpf<mode>): Ditto.
8435 (GATHER_SCATTER_SF_MEM_MODE): New.
8436 (avx512pf_gatherpf<mode>df): Ditto.
8437 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8438 (*avx512pf_scatterpf<mode>df): Ditto.
8440 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8443 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8444 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8447 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8449 * common/config/arm/arm-common.c
8450 (arm_rewrite_mcpu): Handle multiple names.
8452 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8454 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8456 * gimple-builder.h (create_gimple_tmp): Delete.
8458 2014-01-27 Christian Bruel <christian.bruel@st.com>
8460 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
8463 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
8465 * config/pa/pa.md (call): Generate indirect long calls to non-local
8466 functions when outputing 32-bit code.
8467 (call_value): Likewise except for special call to buggy powf function.
8469 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
8470 portable runtime and PIC indirect calls.
8471 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
8472 and PIC call sequences. Use ldo instead of blr to set return register
8473 in PIC call sequence.
8475 2014-01-25 Walter Lee <walt@tilera.com>
8477 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
8478 avoid clobbering a live register.
8480 2014-01-25 Walter Lee <walt@tilera.com>
8482 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
8483 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
8484 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
8485 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
8487 2014-01-25 Walter Lee <walt@tilera.com>
8489 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
8490 arguments on even registers.
8491 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
8493 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
8494 (BIGGEST_ALIGNMENT): Ditto.
8495 (BIGGEST_FIELD_ALIGNMENT): Ditto.
8497 2014-01-25 Walter Lee <walt@tilera.com>
8499 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
8500 insns before bundling.
8501 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8503 2014-01-25 Walter Lee <walt@tilera.com>
8505 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
8506 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
8507 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
8509 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8511 * config/mips/constraints.md (kl): Delete.
8512 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
8513 define expands, using...
8514 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
8515 instructions for MIPS16.
8516 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
8517 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
8519 2014-01-25 Walter Lee <walt@tilera.com>
8521 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
8525 2014-01-25 Walter Lee <walt@tilera.com>
8527 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
8528 (TARGET_EXPAND_TO_RTL_HOOK): Define.
8530 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8532 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
8535 2014-01-25 Jakub Jelinek <jakub@redhat.com>
8537 * print-rtl.c (in_call_function_usage): New var.
8538 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
8539 EXPR_LIST mode as mode and not as reg note name.
8542 * cfgloopmanip.c (copy_loop_info): If
8543 loop->warned_aggressive_loop_optimizations, make sure
8544 the flag is set in target loop too.
8546 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
8548 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
8550 * builtins.def: Likewise.
8551 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
8552 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
8553 * ira.c (ira_setup_eliminable_regset): Likewise.
8554 * omp-low.c (gate_expand_omp): Likewise.
8555 (execute_lower_omp): Likewise.
8556 (diagnose_sb_0): Likewise.
8557 (gate_diagnose_omp_blocks): Likewise.
8558 (simd_clone_clauses_extract): Likewise.
8561 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8563 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
8564 correction for little endian...
8565 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
8568 2014-01-24 Jeff Law <law@redhat.com>
8570 PR tree-optimization/59919
8571 * tree-vrp.c (find_assert_locations_1): Do not register asserts
8572 for non-returning calls.
8574 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
8576 * common/config/aarch64/aarch64-common.c
8577 (aarch64_rewrite_mcpu): Handle multiple names.
8578 * config/aarch64/aarch64.h
8579 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8581 2014-01-24 Dodji Seketeli <dodji@redhat.com>
8583 * input.c (add_file_to_cache_tab): Handle the case where fopen
8586 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
8589 * config/i386/i386.md (pushsf splitter): Get stack adjustment
8590 from push operand if code of push isn't PRE_DEC.
8592 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8595 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
8596 -mquad-memory-atomic. Update -mquad-memory documentation to say
8597 it is only used for non-atomic loads/stores.
8599 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
8600 -mquad-memory or -mquad-memory-atomic switches.
8602 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
8603 -mquad-memory-atomic to ISA 2.07 support.
8605 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
8606 to separate support of normal quad word memory operations (ldq, stq)
8607 from the atomic quad word memory operations.
8609 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
8610 support to separate non-atomic quad word operations from atomic
8611 quad word operations. Disable non-atomic quad word operations in
8612 little endian mode so that we don't have to swap words after the
8613 load and before the store.
8614 (quad_load_store_p): Add comment about atomic quad word support.
8615 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
8616 options printed with -mdebug=reg.
8618 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
8619 -mquad-memory-atomic as the test for whether we have quad word
8620 atomic instructions.
8621 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
8622 or -mp8-vector are used, allow byte/half-word atomic operations.
8624 * config/rs6000/sync.md (load_lockedti): Insure that the address
8625 is a proper indexed or indirect address for the lqarx instruction.
8626 On little endian systems, swap the hi/lo registers after the lqarx
8628 (load_lockedpti): Use indexed_or_indirect_operand predicate to
8629 insure the address is valid for the lqarx instruction.
8630 (store_conditionalti): Insure that the address is a proper indexed
8631 or indirect address for the stqcrx. instruction. On little endian
8632 systems, swap the hi/lo registers before doing the stqcrx.
8634 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
8635 insure the address is valid for the stqcrx. instruction.
8637 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8638 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
8639 type of quad memory support is available.
8641 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
8644 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
8645 there is a danger of looping.
8647 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
8649 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8650 force flag_ira_loop_pressure if set via command line.
8652 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8654 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
8655 (ashr_simd): New builtin handling DI mode.
8656 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
8657 (aarch64_sshr_simddi): New match pattern.
8658 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
8659 (vshrd_n_s64): Likewise.
8660 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
8662 2014-01-23 Nick Clifton <nickc@redhat.com>
8664 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
8665 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
8666 favour of mcu specific scripts.
8667 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
8670 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8671 Alex Velenko <Alex.Velenko@arm.com>
8673 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
8674 (vaddv_s16): Likewise.
8675 (vaddv_s32): Likewise.
8676 (vaddv_u8): Likewise.
8677 (vaddv_u16): Likewise.
8678 (vaddv_u32): Likewise.
8679 (vaddvq_s8): Likewise.
8680 (vaddvq_s16): Likewise.
8681 (vaddvq_s32): Likewise.
8682 (vaddvq_s64): Likewise.
8683 (vaddvq_u8): Likewise.
8684 (vaddvq_u16): Likewise.
8685 (vaddvq_u32): Likewise.
8686 (vaddvq_u64): Likewise.
8687 (vaddv_f32): Likewise.
8688 (vaddvq_f32): Likewise.
8689 (vaddvq_f64): Likewise.
8690 (vmaxv_f32): Likewise.
8691 (vmaxv_s8): Likewise.
8692 (vmaxv_s16): Likewise.
8693 (vmaxv_s32): Likewise.
8694 (vmaxv_u8): Likewise.
8695 (vmaxv_u16): Likewise.
8696 (vmaxv_u32): Likewise.
8697 (vmaxvq_f32): Likewise.
8698 (vmaxvq_f64): Likewise.
8699 (vmaxvq_s8): Likewise.
8700 (vmaxvq_s16): Likewise.
8701 (vmaxvq_s32): Likewise.
8702 (vmaxvq_u8): Likewise.
8703 (vmaxvq_u16): Likewise.
8704 (vmaxvq_u32): Likewise.
8705 (vmaxnmv_f32): Likewise.
8706 (vmaxnmvq_f32): Likewise.
8707 (vmaxnmvq_f64): Likewise.
8708 (vminv_f32): Likewise.
8709 (vminv_s8): Likewise.
8710 (vminv_s16): Likewise.
8711 (vminv_s32): Likewise.
8712 (vminv_u8): Likewise.
8713 (vminv_u16): Likewise.
8714 (vminv_u32): Likewise.
8715 (vminvq_f32): Likewise.
8716 (vminvq_f64): Likewise.
8717 (vminvq_s8): Likewise.
8718 (vminvq_s16): Likewise.
8719 (vminvq_s32): Likewise.
8720 (vminvq_u8): Likewise.
8721 (vminvq_u16): Likewise.
8722 (vminvq_u32): Likewise.
8723 (vminnmv_f32): Likewise.
8724 (vminnmvq_f32): Likewise.
8725 (vminnmvq_f64): Likewise.
8727 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8729 * config/aarch64/aarch64-simd.md
8730 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
8731 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
8732 (*aarch64_mul3_elt<mode>): Likewise.
8733 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
8734 (*aarch64_mul3_elt_to_64v2df): Likewise.
8735 (*aarch64_mla_elt<mode>): Likewise.
8736 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8737 (*aarch64_mls_elt<mode>): Likewise.
8738 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8739 (*aarch64_fma4_elt<mode>): Likewise.
8740 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8741 (*aarch64_fma4_elt_to_64v2df): Likewise.
8742 (*aarch64_fnma4_elt<mode>): Likewise.
8743 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8744 (*aarch64_fnma4_elt_to_64v2df): Likewise.
8745 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
8746 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
8747 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8748 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8749 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
8750 (aarch64_sqdmull_lane<mode>_internal): Likewise.
8751 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
8753 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
8755 * config/aarch64/aarch64-simd.md
8756 (aarch64_be_checked_get_lane<mode>): New define_expand.
8757 * config/aarch64/aarch64-simd-builtins.def
8758 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
8759 New builtin definition.
8760 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
8761 Use new safe be builtin.
8763 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8765 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
8767 (aarch64_be_st1<mode>): Likewise.
8768 (aarch_ld1<VALL:mode>): Define_expand modified.
8769 (aarch_st1<VALL:mode>): Likewise.
8770 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
8771 (UNSPEC_ST1): Likewise.
8773 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
8775 * config/microblaze/microblaze.md: Add trap insn and attribute
8777 2014-01-23 Dodji Seketeli <dodji@redhat.com>
8779 PR preprocessor/58580
8780 * input.h (location_get_source_line): Take an additional line_size
8782 (void diagnostics_file_cache_fini): Declare new function.
8783 * input.c (struct fcache): New type.
8784 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
8785 New static constants.
8786 (diagnostic_file_cache_init, total_lines_num)
8787 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
8788 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
8789 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
8790 (get_next_line, read_next_line, goto_next_line, read_line_num):
8791 New static function definitions.
8792 (diagnostic_file_cache_fini): New function.
8793 (location_get_source_line): Take an additional output line_len
8794 parameter. Re-write using lookup_or_add_file_to_cache_tab and
8796 * diagnostic.c (diagnostic_finish): Call
8797 diagnostic_file_cache_fini.
8798 (adjust_line): Take an additional input parameter for the length
8799 of the line, rather than calculating it with strlen.
8800 (diagnostic_show_locus): Adjust the use of
8801 location_get_source_line and adjust_line with respect to their new
8802 signature. While displaying a line now, do not stop at the first
8803 null byte. Rather, display the zero byte as a space and keep
8804 going until we reach the size of the line.
8805 * Makefile.in: Add vec.o to OBJS-libcommon
8807 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8808 Ilya Tocar <ilya.tocar@intel.com>
8810 * config/i386/avx512fintrin.h (_mm512_kmov): New.
8811 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
8812 (__builtin_ia32_kmov16): Ditto.
8813 * config/i386/i386.md (UNSPEC_KMOV): New.
8816 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8818 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
8819 (_mm512_storeu_si512): Ditto.
8821 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
8824 * rtl.h (get_referenced_operands): Declare.
8825 * recog.c (get_referenced_operands): New function.
8826 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
8827 operands have been referenced when recording LO_SUM references.
8829 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
8831 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
8833 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8835 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
8836 Enable for generic and recent AMD targets.
8838 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8840 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
8841 ARG_SIZE note when adjustment was eliminated.
8843 2014-01-22 Jeff Law <law@redhat.com>
8845 PR tree-optimization/59597
8846 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
8847 in file. Accept new argument REGISTERING and use it to modify
8848 dump output appropriately.
8849 (register_jump_thread): Corresponding changes.
8850 (mark_threaded_blocks): Reinstate code to cancel unprofitable
8851 thread paths involving joiner blocks. Add code to dump cancelled
8852 jump threading paths.
8854 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
8856 PR rtl-optimization/59477
8857 * lra-constraints.c (inherit_in_ebb): Process call for living hard
8858 regs. Update reloads_num and potential_reload_hard_regs for all insns.
8860 2014-01-22 Tom Tromey <tromey@redhat.com>
8862 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
8864 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
8866 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8868 PR rtl-optimization/59896
8869 * lra-constraints.c (process_alt_operands): Check unused note for
8870 matched operands of insn with no output reloads.
8872 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
8874 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
8875 (mips_move_from_gpr_cost): Likewise.
8877 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8879 PR rtl-optimization/59858
8880 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
8881 ira_class_hard_regs_num.
8882 (process_alt_operands): Increase reject for dying matched operand.
8884 2014-01-21 Jakub Jelinek <jakub@redhat.com>
8887 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
8888 smaller than size, perform several stores or loads and stores
8889 at dst + count - size to store or copy all of size bytes, rather
8890 than just last modesize bytes.
8892 2014-01-20 DJ Delorie <dj@redhat.com>
8894 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
8895 that CLOBBERs are REGs before propogating their values.
8897 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
8900 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
8901 (cgraph_inline_failed_type): New function.
8902 * cgraph.h (DEFCIFCODE): Add type.
8903 (cgraph_inline_failed_type_t): New enum.
8904 (cgraph_inline_failed_type): New prototype.
8905 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
8906 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
8907 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
8908 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
8909 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
8910 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
8911 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
8912 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
8913 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
8914 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
8915 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
8916 OPTIMIZATION_MISMATCH.
8917 * tree-inline.c (expand_call_inline): Emit errors during
8918 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
8920 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8923 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
8924 mode attribute in insn output.
8926 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
8928 * output.h (output_constant): Delete.
8929 * varasm.c (output_constant): Make private.
8931 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
8933 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
8935 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8938 * tree.h (fold_builtin_strcat): New prototype.
8939 * builtins.c (fold_builtin_strcat): No longer static. Add len
8940 argument, if non-NULL, don't call c_strlen. Optimize
8941 directly into __builtin_memcpy instead of __builtin_strcpy.
8942 (fold_builtin_2): Adjust fold_builtin_strcat caller.
8943 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
8945 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8947 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8948 for SImode_address_operand operands, having only a REG argument.
8950 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
8952 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
8953 loader name using mbig-endian.
8954 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
8956 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8958 * doc/invoke.texi (-march): Clarify documentation for AArch64.
8962 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
8964 * config/aarch64/aarch64-protos.h
8965 (aarch64_cannot_change_mode_class_ptr): Declare.
8966 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
8967 aarch64_cannot_change_mode_class_ptr): New.
8968 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
8969 backend hook aarch64_cannot_change_mode_class.
8971 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8973 * common/config/aarch64/aarch64-common.c
8974 (aarch64_handle_option): Don't handle any option order logic here.
8975 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
8976 selected_cpu, warn on architecture version mismatch.
8977 (aarch64_override_options): Fix parsing order for option strings.
8979 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8980 Iain Sandoe <iain@codesourcery.com>
8983 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
8984 warning. Amend comment to reflect current functionality.
8986 2014-01-20 Richard Biener <rguenther@suse.de>
8989 * builtins.c (fold_builtin_strcat): Remove case better handled
8990 by tree-ssa-strlen.c.
8992 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
8994 * config/aarch64/aarch64.opt
8995 (mcpu, march, mtune): Make case-insensitive.
8997 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9000 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9001 if operands[1] is a REG or ZERO_EXTEND of a REG.
9003 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
9005 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
9007 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
9009 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
9010 long non-pic millicode calls.
9012 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9014 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
9016 2014-01-19 Kito Cheng <kito@0xlab.org>
9018 * builtins.c (expand_movstr): Check movstr expand done or fail.
9020 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9021 H.J. Lu <hongjiu.lu@intel.com>
9024 * config/i386/i386.md (*lea<mode>): Zero-extend return register
9025 to DImode for zero-extended addresses.
9027 2014-01-19 Jakub Jelinek <jakub@redhat.com>
9029 PR rtl-optimization/57763
9030 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
9031 on the new indirect jump_insn and increment LABEL_NUSES (label).
9033 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
9037 * config.gcc (x86_archs): New variable.
9038 (x86_64_archs): Likewise.
9039 (x86_cpus): Likewise.
9040 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
9041 --with-arch/--with-cpu= options.
9042 Support --with-arch=/--with-cpu={nehalem,westmere,
9043 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
9045 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9047 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
9048 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
9050 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9052 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
9054 2014-01-18 Jakub Jelinek <jakub@redhat.com>
9057 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
9058 clear cpp_get_options (parse_in)->warn_unused_macros for
9059 ix86_target_macros_internal with cpp_define.
9061 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
9063 * jump.c (delete_related_insns): Keep (use (insn))s.
9064 * reorg.c (redundant_insn): Check for barriers too.
9066 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9068 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
9070 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
9072 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
9073 call to $$dyncall when TARGET_LONG_CALLS is true.
9075 2014-01-17 Jeff Law <law@redhat.com>
9077 * ree.c (combine_set_extension): Temporarily disable test for
9078 changing number of hard registers.
9080 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9083 * ipa-inline-analysis.c (inline_free_summary):
9084 Do not free summary of aliases.
9086 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9089 * gimplify.c (gimplify_expr): Use create_tmp_var
9090 instead of create_tmp_var_raw. If cond doesn't have
9091 integral type, don't add the IFN_ANNOTATE builtin at all.
9093 2014-01-17 Martin Jambor <mjambor@suse.cz>
9096 * ipa-cp.c (prev_edge_clone): New variable.
9097 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
9098 Also resize prev_edge_clone vector.
9099 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
9100 (ipcp_edge_removal_hook): New function.
9101 (ipcp_driver): Register ipcp_edge_removal_hook.
9103 2014-01-17 Andrew Pinski <apinski@cavium.com>
9104 Steve Ellcey <sellcey@mips.com>
9107 * config/mips/mips.c (mips_print_operand): Check operand mode instead
9110 2014-01-17 Jeff Law <law@redhat.com>
9113 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
9114 so that pass_ccp runs first.
9116 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9118 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
9119 (ix86_adjust_cost): Use !TARGET_XXX.
9120 (do_reorder_for_imul): Likewise.
9121 (swap_top_of_ready_list): Likewise.
9122 (ix86_sched_reorder): Likewise.
9124 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9126 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9127 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
9128 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
9129 (intel_memset): New. Duplicate slm_memset.
9130 (intel_cost): New. Duplicate slm_cost.
9131 (m_INTEL): New macro.
9132 (processor_target_table): Add "intel".
9133 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
9134 with PROCESSOR_INTEL for "intel".
9135 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
9136 PROCESSOR_SILVERMONT.
9137 (ix86_issue_rate): Likewise.
9138 (ix86_adjust_cost): Likewise.
9139 (ia32_multipass_dfa_lookahead): Likewise.
9140 (swap_top_of_ready_list): Likewise.
9141 (ix86_sched_reorder): Likewise.
9142 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
9143 instead of TARGET_OPT_AGU.
9144 * config/i386/i386.h (TARGET_INTEL): New.
9145 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
9146 (processor_type): Add PROCESSOR_INTEL.
9147 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
9148 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
9150 2014-01-17 Marek Polacek <polacek@redhat.com>
9153 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
9156 2014-01-17 Richard Biener <rguenther@suse.de>
9158 PR tree-optimization/46590
9159 * opts.c (default_options_table): Add entries for
9160 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
9161 all enabled at -O1 but not for -Og.
9162 * common.opt (fbranch-count-reg): Remove Init(1).
9163 (fmove-loop-invariants): Likewise.
9164 (ftree-pta): Likewise.
9166 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9168 * config/i386/i386.c (ix86_data_alignment): For compatibility with
9169 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
9170 decls to at least the GCC 4.8 used alignments.
9173 * tree-nested.c (convert_nonlocal_reference_stmt,
9174 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
9175 of GIMPLE_BIND stmts, adjust associated decls.
9177 2014-01-17 Richard Biener <rguenther@suse.de>
9179 PR tree-optimization/46590
9180 * vec.h (vec<>::bseach): New member function implementing
9181 binary search according to C89 bsearch.
9182 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
9183 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
9184 bitmap pointer again. Make accesses_in_loop a flat array.
9185 (mem_ref_obstack): New global.
9186 (outermost_indep_loop): Adjust for mem_ref->stored changes.
9187 (mark_ref_stored): Likewise.
9188 (ref_indep_loop_p_2): Likewise.
9189 (set_ref_stored_in_loop): New helper function.
9190 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
9191 (memref_free): Adjust.
9192 (record_mem_ref_loc): Simplify.
9193 (gather_mem_refs_stmt): Adjust.
9194 (sort_locs_in_loop_postorder_cmp): New function.
9195 (analyze_memory_references): Sort accesses_in_loop after
9196 loop postorder number.
9197 (find_ref_loc_in_loop_cmp): New function.
9198 (for_all_locs_in_loop): Find relevant cluster of locs in
9199 accesses_in_loop and iterate without recursion.
9200 (execute_sm): Avoid uninit warning.
9201 (struct ref_always_accessed): Simplify.
9202 (ref_always_accessed::operator ()): Likewise.
9203 (ref_always_accessed_p): Likewise.
9204 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
9205 loop postorder numbers here.
9206 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
9209 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9212 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
9213 on decls for which assemble_alias has been called.
9215 2014-01-17 Nick Clifton <nickc@redhat.com>
9217 * config/msp430/msp430.opt: (mcpu): New option.
9218 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
9219 (msp430_option_override): Parse target_cpu. If the MCU name
9220 matches a generic string, clear target_mcu.
9221 (msp430_attr): Allow numeric interrupt values up to 63.
9222 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
9223 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
9225 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
9227 * config/msp430/msp430.md (popm): Use %J rather than %I.
9228 (addsi3): Use msp430_nonimmediate_operand for operand 2.
9229 (addhi_cy_i): Use immediate_operand for operand 2.
9230 * doc/invoke.texi: Document -mcpu option.
9232 2014-01-17 Richard Biener <rguenther@suse.de>
9234 PR rtl-optimization/38518
9235 * df.h (df_analyze_loop): Declare.
9236 * df-core.c: Include cfgloop.h.
9237 (df_analyze_1): Split out main part of df_analyze.
9238 (df_analyze): Adjust.
9239 (loop_inverted_post_order_compute): New function.
9240 (loop_post_order_compute): Likewise.
9241 (df_analyze_loop): New function avoiding whole-function
9243 * loop-invariant.c (find_defs): Use df_analyze_loop.
9244 (find_invariants): Adjust.
9245 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
9247 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
9249 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
9250 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
9252 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
9254 * ipa-ref.c (ipa_remove_stmt_references): Fix references
9255 traversal when removing references.
9257 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
9260 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
9262 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
9265 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
9266 pushing a reload for an autoinc when we had previously reloaded an
9267 inner part of the address.
9269 2014-01-16 Jakub Jelinek <jakub@redhat.com>
9271 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
9273 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
9274 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
9275 when not giving up or versioning for alias only because of
9277 (vect_analyze_data_ref_dependences): Set to true.
9278 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
9280 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
9281 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
9285 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
9288 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
9289 operand 0 predicate for gathers, use a new pseudo as subtarget.
9291 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9294 * lra-constraints.c (process_alt_operands): Add printing debug info.
9295 Check absence of input/output reloads for matched operands too.
9297 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9299 PR rtl-optimization/59835
9300 * ira.c (ira_init_register_move_cost): Increase cost for
9303 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
9305 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
9307 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
9310 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
9311 non-register objects. Use gen_(high/low)part more consistently.
9314 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9317 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
9318 endian support, remove tests for WORDS_BIG_ENDIAN.
9319 (p8_mfvsrd_3_<mode>): Likewise.
9320 (reload_gpr_from_vsx<mode>): Likewise.
9321 (reload_gpr_from_vsxsf): Likewise.
9322 (p8_mfvsrd_4_disf): Likewise.
9324 2014-01-16 Richard Biener <rguenther@suse.de>
9326 PR rtl-optimization/46590
9327 * lcm.c (compute_antinout_edge): Use postorder iteration.
9328 (compute_laterin): Use inverted postorder iteration.
9330 2014-01-16 Nick Clifton <nickc@redhat.com>
9333 * varasm.c (output_constant): Return the number of bytes actually
9335 (output_constructor_array_range): Update the field size with the
9336 number of bytes emitted by output_constant.
9337 (output_constructor_regular_field): Likewise. Also do not
9338 complain if the total number of bytes emitted is now greater
9339 than the expected fieldpos.
9340 * output.h (output_constant): Update prototype and descriptive comment.
9342 2014-01-16 Marek Polacek <polacek@redhat.com>
9345 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
9346 it is error_mark_node.
9348 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
9350 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
9351 VALID_AVX256_REG_OR_OI_MODE.
9353 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
9355 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
9356 current procedure should be profiled.
9358 2014-01-15 Andrew Pinski <apinski@cavium.com>
9360 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
9361 of moving from/to the STACK_REG register class.
9363 2014-01-15 Richard Henderson <rth@redhat.com>
9366 * reginfo.c (global_regs_decl): Globalize.
9367 * rtl.h (global_regs_decl): Declare.
9368 * ira.c (do_reload): Diagnose frame_pointer_needed and it
9369 reserved via global_regs.
9371 2014-01-15 Teresa Johnson <tejohnson@google.com>
9373 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
9375 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9377 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
9378 and vmulosh rather than call gen_vec_widen_smult_*.
9379 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
9380 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
9381 (vec_widen_smult_even_v16qi): Likewise.
9382 (vec_widen_umult_even_v8hi): Likewise.
9383 (vec_widen_smult_even_v8hi): Likewise.
9384 (vec_widen_umult_odd_v16qi): Likewise.
9385 (vec_widen_smult_odd_v16qi): Likewise.
9386 (vec_widen_umult_odd_v8hi): Likewise.
9387 (vec_widen_smult_odd_v8hi): Likewise.
9388 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
9389 vmuloub rather than call gen_vec_widen_umult_*.
9390 (vec_widen_umult_lo_v16qi): Likewise.
9391 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
9392 vmulosb rather than call gen_vec_widen_smult_*.
9393 (vec_widen_smult_lo_v16qi): Likewise.
9394 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
9395 rather than call gen_vec_widen_umult_*.
9396 (vec_widen_umult_lo_v8hi): Likewise.
9397 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
9398 rather than call gen_vec_widen_smult_*.
9399 (vec_widen_smult_lo_v8hi): Likewise.
9401 2014-01-15 Jeff Law <law@redhat.com>
9403 PR tree-optimization/59747
9404 * ree.c (find_and_remove_re): Properly handle case where a second
9405 eliminated extension requires widening a copy created for elimination
9406 of a prior extension.
9407 (combine_set_extension): Ensure that the number of hard regs needed
9408 for a destination register does not change when we widen it.
9410 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
9412 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
9413 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
9414 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
9415 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
9416 (avr-*-rtems*): Likewise.
9417 (bfin*-rtems*): Likewise.
9418 (moxie-*-rtems*): Likewise.
9419 (h8300-*-rtems*): Likewise.
9420 (i[34567]86-*-rtems*): Likewise.
9421 (lm32-*-rtems*): Likewise.
9422 (m32r-*-rtems*): Likewise.
9423 (m68k-*-rtems*): Likewise.
9424 (microblaze*-*-rtems*): Likewise.
9425 (mips*-*-rtems*): Likewise.
9426 (powerpc-*-rtems*): Likewise.
9427 (sh-*-rtems*): Likewise.
9428 (sparc-*-rtems*): Likewise.
9429 (sparc64-*-rtems*): Likewise.
9430 (v850-*-rtems*): Likewise.
9431 (m32c-*-rtems*): Likewise.
9433 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9435 PR rtl-optimization/59511
9436 * ira.c (ira_init_register_move_cost): Use memory costs for some
9437 cases of register move cost calculations.
9438 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9439 instead of BB frequency.
9440 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9441 * lra-assigns.c (find_hard_regno_for): Ditto.
9443 2014-01-15 Richard Biener <rguenther@suse.de>
9445 PR tree-optimization/59822
9446 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9447 (vectorizable_load): Use it to hoist defs of uses of invariant
9448 loads out of the loop.
9450 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9451 Kugan Vivekanandarajah <kuganv@linaro.org>
9454 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9457 2014-01-15 Richard Biener <rguenther@suse.de>
9459 PR rtl-optimization/59802
9460 * lcm.c (compute_available): Use inverted postorder to seed
9461 the initial worklist.
9463 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9466 * config/s390/s390.c (s390_preferred_reload_class): Don't return
9467 ADDR_REGS for invalid symrefs in non-PIC code.
9469 2014-01-15 Jakub Jelinek <jakub@redhat.com>
9472 * builtins.c (determine_block_size): Initialize *probable_max_size
9473 even if len_rtx is CONST_INT.
9475 2014-01-14 Andrew Pinski <apinski@cavium.com>
9477 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
9478 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
9479 (cortexa53_tunings): Likewise.
9480 (aarch64_sched_issue_rate): New function.
9481 (TARGET_SCHED_ISSUE_RATE): Define.
9483 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9485 * ira-costs.c (find_costs_and_classes): Add missed
9486 ira_init_register_move_cost_if_necessary.
9488 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9491 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
9493 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
9496 * config/i386/i386.c (type_natural_mode): Add a bool parameter
9497 to indicate if type is used for function return value. Warn ABI
9498 change if the vector mode isn't available for function return value.
9499 (ix86_function_arg_advance): Pass false to type_natural_mode.
9500 (ix86_function_arg): Likewise.
9501 (ix86_gimplify_va_arg): Likewise.
9502 (function_arg_32): Don't warn ABI change.
9503 (ix86_function_value): Pass true to type_natural_mode.
9504 (ix86_return_in_memory): Likewise.
9505 (ix86_struct_value_rtx): Removed.
9506 (TARGET_STRUCT_VALUE_RTX): Likewise.
9508 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9510 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
9511 converting a conditional jump into a conditional return.
9513 2014-01-14 Richard Biener <rguenther@suse.de>
9515 PR tree-optimization/58921
9516 PR tree-optimization/59006
9517 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
9518 hoisting invariant stmts.
9519 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
9520 invariant loads on the preheader edge if possible.
9522 2014-01-14 Joey Ye <joey.ye@arm.com>
9524 * doc/plugin.texi (Building GCC plugins): Update to C++.
9526 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
9528 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
9529 (_mm_rcp28_round_ss): Ditto.
9530 (_mm_rsqrt28_round_sd): Ditto.
9531 (_mm_rsqrt28_round_ss): Ditto.
9532 (_mm_rcp28_sd): Ditto.
9533 (_mm_rcp28_ss): Ditto.
9534 (_mm_rsqrt28_sd): Ditto.
9535 (_mm_rsqrt28_ss): Ditto.
9536 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
9537 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
9538 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
9539 (IX86_BUILTIN_RCP28SD): Ditto.
9540 (IX86_BUILTIN_RCP28SS): Ditto.
9541 (IX86_BUILTIN_RSQRT28SD): Ditto.
9542 (IX86_BUILTIN_RSQRT28SS): Ditto.
9543 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
9544 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
9545 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
9546 (ix86_expand_special_args_builtin): Expand new FTYPE.
9547 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
9548 (srcp14<mode>): Make insn unary.
9549 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
9550 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
9551 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
9552 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
9553 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
9554 Fix rounding: make it SAE only.
9555 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9557 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9559 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
9560 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
9561 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
9562 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
9563 (round_saeonly_mask_scalar_operand4): Ditto.
9564 (round_saeonly_mask_scalar_op3): Ditto.
9565 (round_saeonly_mask_scalar_op4): Ditto.
9567 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9569 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9570 Implement -maltivec=be for vec_insert and vec_extract.
9572 2014-01-10 DJ Delorie <dj@redhat.com>
9574 * config/msp430/msp430.md (call_internal): Don't allow memory
9575 references with SP as the base register.
9576 (call_value_internal): Likewise.
9577 * config/msp430/constraints.md (Yc): New. For memory references
9578 that don't use SP as a base register.
9580 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
9581 "an integer without a # prefix"
9582 * config/msp430/msp430.md (epilogue_helper): Use it.
9584 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9587 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
9588 AVX512F gather builtins.
9589 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
9590 on gather decls with INTEGER_TYPE masktype.
9591 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
9592 directly into the builtin rather than hoisting it before loop.
9594 PR tree-optimization/59387
9595 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
9596 (scev_const_prop): If folded_casts and type has undefined overflow,
9597 use force_gimple_operand instead of force_gimple_operand_gsi and
9598 for each added stmt if it is assign with
9599 arith_code_with_undefined_signed_overflow, call
9600 rewrite_to_defined_overflow.
9601 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
9602 gimple-fold.h instead.
9603 (arith_code_with_undefined_signed_overflow,
9604 rewrite_to_defined_overflow): Moved to ...
9605 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
9606 rewrite_to_defined_overflow): ... here. No longer static.
9607 Include gimplify-me.h.
9608 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
9609 rewrite_to_defined_overflow): New prototypes.
9611 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9613 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
9615 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
9617 * builtins.c (get_object_alignment_2): Minor tweak.
9618 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
9620 2014-01-13 Christian Bruel <christian.bruel@st.com>
9622 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
9623 optimized non constant lengths.
9625 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9628 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
9629 load as __atomic_load_N if possible.
9631 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9633 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
9635 (rs6000_expand_builtin): Adjust call.
9637 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9640 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
9641 * config/rs6000/rs6000.c: Include target-globals.h.
9642 (rs6000_set_current_function): Instead of doing target_reinit
9643 unconditionally, use save_target_globals_default_opts and
9644 restore_target_globals.
9646 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
9648 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
9649 (rs6000_expand_builtin): Handle mffs and mtfsf.
9650 (rs6000_init_builtins): Define mffs and mtfsf.
9651 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
9652 (rs6000_mffs): New pattern.
9653 (rs6000_mtfsf): New pattern.
9655 2014-01-11 Bin Cheng <bin.cheng@arm.com>
9657 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
9658 Start narrowing with START. Apply candidate-use pair
9659 and check overall cost in narrowing.
9660 (iv_ca_prune): Pass new argument.
9662 2014-01-10 Jeff Law <law@redhat.com>
9665 * ree.c (combine_reaching_defs): Ensure the defining statement
9666 occurs before the extension when optimizing extensions with
9667 different source and destination hard registers.
9669 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9672 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
9673 vtables into the type inheritance graph.
9675 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9677 PR rtl-optimization/59754
9678 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
9679 modes in the REGNO != REGNO case.
9681 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9683 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
9685 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9687 PR tree-optimization/59745
9688 * tree-predcom.c (tree_predictive_commoning_loop): Call
9689 free_affine_expand_cache if giving up because components is NULL.
9691 * target-globals.c (save_target_globals): Allocate < 4KB structs using
9692 GC in payload of target_globals struct instead of allocating them on
9693 the heap and the larger structs separately using GC.
9694 * target-globals.h (struct target_globals): Make regs, hard_regs,
9695 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
9696 of GTY((skip)) and change type to void *.
9697 (reset_target_globals): Cast loads from those fields to corresponding
9700 2014-01-10 Steve Ellcey <sellcey@mips.com>
9703 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
9704 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
9705 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
9707 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
9710 * aarch64-modes.def (CC_Zmode): New flags mode.
9711 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
9712 represents an equality.
9713 (aarch64_get_condition_code): Handle CC_Zmode.
9714 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
9716 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9718 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
9719 extraction in good case.
9721 2014-01-10 Richard Biener <rguenther@suse.de>
9723 PR tree-optimization/59374
9724 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
9725 checking after SLP discovery. Mark stmts not participating
9726 in any SLP instance properly.
9728 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9730 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
9731 when handling a SET rtx.
9733 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9735 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
9736 (cortex-a57): Likewise.
9737 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
9739 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9741 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
9742 non-iwmmxt builtins.
9744 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9748 * ipa-devirt.c record_target_from_binfo): Take as argument
9749 stack of binfos and lookup matching one for virtual inheritance.
9750 (possible_polymorphic_call_targets_1): Update.
9752 2014-01-10 Huacai Chen <chenhc@lemote.com>
9754 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
9755 kernel strings for Loongson-2E/2F/3A.
9757 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9760 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
9761 is_gimple_call before calling gimple_call_internal_p.
9763 2014-01-09 Steve Ellcey <sellcey@mips.com>
9765 * Makefile.in (TREE_FLOW_H): Remove.
9766 (TREE_SSA_H): Add file names from tree-flow.h.
9767 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
9768 * tree.h: Remove tree-flow.h reference.
9769 * hash-table.h: Remove tree-flow.h reference.
9770 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
9771 reference with tree-ssa-loop.h.
9773 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9775 * doc/invoke.texi: Add -maltivec={be,le} options, and document
9776 default element-order behavior for -maltivec.
9777 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
9778 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
9779 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
9780 when targeting big endian, at least for now.
9781 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
9783 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9786 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
9787 var satisfies use_register_for_decl, just take into account type
9788 alignment, rather than decl alignment.
9790 PR tree-optimization/59622
9791 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
9792 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
9793 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
9794 Don't devirtualize for inplace at all. For targets.length () == 1,
9795 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
9797 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9799 * config/i386/i386.md (cpu): Remove the unused btver1.
9801 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9803 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
9805 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9808 * tree-core.h (struct target_globals): New forward declaration.
9809 (struct tree_target_option): Add globals field.
9810 * tree.h (TREE_TARGET_GLOBALS): Define.
9811 (prepare_target_option_nodes_for_pch): New prototype.
9812 * target-globals.h (struct target_globals): Define even if
9814 * tree.c (prepare_target_option_node_for_pch,
9815 prepare_target_option_nodes_for_pch): New functions.
9816 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
9817 * config/i386/i386.c: Include target-globals.h.
9818 (ix86_set_current_function): Instead of doing target_reinit
9819 unconditionally, use save_target_globals_default_opts and
9820 restore_target_globals.
9822 2014-01-09 Richard Biener <rguenther@suse.de>
9824 PR tree-optimization/59715
9825 * tree-cfg.h (split_critical_edges): Declare.
9826 * tree-cfg.c (split_critical_edges): Export.
9827 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
9829 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
9831 * cfgexpand.c (expand_stack_vars): Optionally disable
9832 asan stack protection.
9833 (expand_used_vars): Likewise.
9834 (partition_stack_vars): Likewise.
9835 * asan.c (asan_emit_stack_protection): Optionally disable
9836 after return stack usage.
9837 (instrument_derefs): Optionally disable memory access instrumentation.
9838 (instrument_builtin_call): Likewise.
9839 (instrument_strlen_call): Likewise.
9840 (asan_protect_global): Optionally disable global variables protection.
9841 * doc/invoke.texi: Added doc for new options.
9842 * params.def: Added new options.
9843 * params.h: Likewise.
9845 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9847 PR rtl-optimization/59724
9848 * ifcvt.c (cond_exec_process_if_block): Don't call
9849 flow_find_head_matching_sequence with 0 longest_match.
9850 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
9851 non-active insns if !stop_after.
9852 (try_head_merge_bb): Revert 2014-01-07 changes.
9854 2014-01-08 Jeff Law <law@redhat.com>
9856 * ree.c (get_sub_rtx): New function, extracted from...
9857 (merge_def_and_ext): Here.
9858 (combine_reaching_defs): Use get_sub_rtx.
9860 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
9862 * cgraph.h (varpool_variable_node): Do not choke on null node.
9864 2014-01-08 Catherine Moore <clm@codesourcery.com>
9866 * config/mips/mips.md (simple_return): Attempt to use JRC
9868 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
9870 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9872 PR rtl-optimization/59137
9873 * reorg.c (steal_delay_list_from_target): Call update_block for
9875 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
9877 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9879 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
9880 two duplicate entries.
9882 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9885 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
9887 * config/mips/mips.c (mips_truncated_op_cost): New function.
9888 (mips_rtx_costs): Adjust test for BADDU.
9889 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
9891 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
9893 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
9894 (*baddu_si): ...this new pattern.
9896 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9899 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
9901 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9904 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
9906 (expand_expr, expand_normal): Adjust.
9907 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
9908 inner_reference_p. Use inner_reference_p to expand inner references.
9909 (store_expr): Adjust.
9910 * cfgexpand.c (expand_call_stmt): Adjust.
9912 2014-01-08 Rong Xu <xur@google.com>
9914 * gcov-io.c (gcov_var): Move from gcov-io.h.
9915 (gcov_position): Ditto.
9916 (gcov_is_error): Ditto.
9917 (gcov_rewrite): Ditto.
9918 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
9919 only part to libgcc/libgcov.h.
9921 2014-01-08 Marek Polacek <polacek@redhat.com>
9924 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
9926 2014-01-08 Marek Polacek <polacek@redhat.com>
9929 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
9931 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9933 PR rtl-optimization/59649
9934 * stor-layout.c (get_mode_bounds): For BImode return
9935 0 and STORE_FLAG_VALUE.
9937 2014-01-08 Richard Biener <rguenther@suse.de>
9940 * gimple.h (is_gimple_builtin_call): Remove.
9941 (gimple_builtin_call_types_compatible_p): New.
9942 (gimple_call_builtin_p): New overload.
9943 * gimple.c (is_gimple_builtin_call): Remove.
9944 (validate_call): Rename to ...
9945 (gimple_builtin_call_types_compatible_p): ... this and export. Also
9947 (validate_type): New static function.
9948 (gimple_call_builtin_p): New overload and adjust.
9949 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
9950 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
9951 (gimple_fold_stmt_to_constant_1): Likewise.
9952 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
9954 2014-01-08 Richard Biener <rguenther@suse.de>
9957 * gimplify.c (gimplify_expr): Gimplify register-register type
9958 VIEW_CONVERT_EXPRs to separate stmts.
9960 2014-01-07 Jeff Law <law@redhat.com>
9963 * ree.c (combine_set_extension): Handle case where source
9964 and destination registers in an extension insn are different.
9965 (combine_reaching_defs): Allow source and destination registers
9966 in extension to be different under limited circumstances.
9967 (add_removable_extension): Remove restriction that the
9968 source and destination registers in the extension are the same.
9969 (find_and_remove_re): Emit a copy from the extension's
9970 destination to its source after the defining insn if
9971 the source and destination registers are different.
9974 * ifcvt.c (merge_if_block): If we are merging a block with more than
9975 one successor with a block with no successors, remove any BARRIER
9976 after the second block.
9978 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
9980 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
9982 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
9985 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
9986 for 14-bit register offsets when INT14_OK_STRICT is false.
9988 2014-01-07 Roland Stigge <stigge@antcom.de>
9989 Michael Meissner <meissner@linux.vnet.ibm.com>
9992 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
9993 Only check TFmode for SPE constants. Don't check TImode or TDmode.
9995 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
9997 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
10000 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
10002 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
10003 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
10004 rtx is const0_rtx or not.
10006 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
10009 * target-globals.c (save_target_globals): Remove this_fn_optab
10011 * toplev.c: Include optabs.h.
10012 (target_reinit): Temporarily restore the global options if another
10013 set of options are in force.
10015 2014-01-07 Jakub Jelinek <jakub@redhat.com>
10017 PR rtl-optimization/58668
10018 * cfgcleanup.c (flow_find_cross_jump): Don't count
10019 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
10020 to determine what is counted.
10021 (flow_find_head_matching_sequence): Use active_insn_p to determine
10023 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
10025 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
10026 determine what is counted.
10028 PR tree-optimization/59643
10029 * tree-predcom.c (split_data_refs_to_components): If one dr is
10030 read and one write, determine_offset fails and the write isn't
10031 in the bad component, just put the read into the bad component.
10033 2014-01-07 Mike Stump <mikestump@comcast.net>
10034 Jakub Jelinek <jakub@redhat.com>
10037 * tree-core.h (struct tree_optimization_option): Change optabs
10038 type from unsigned char * to void *.
10039 * optabs.c (init_tree_optimization_optabs): Adjust
10040 TREE_OPTIMIZATION_OPTABS initialization.
10042 2014-01-06 Jakub Jelinek <jakub@redhat.com>
10045 * config/i386/i386.h (struct machine_function): Add
10046 no_drap_save_restore field.
10047 * config/i386/i386.c (ix86_save_reg): Use
10048 !cfun->machine->no_drap_save_restore instead of
10049 crtl->stack_realign_needed.
10050 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
10051 this function clears frame_pointer_needed. Set
10052 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
10053 and DRAP reg is needed.
10055 2014-01-06 Marek Polacek <polacek@redhat.com>
10058 * doc/implement-c.texi: Mention that other integer types are
10059 permitted as bit-field types in strictly conforming mode.
10061 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
10063 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
10064 is newly allocated.
10066 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
10068 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
10070 2014-01-06 Martin Jambor <mjambor@suse.cz>
10073 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
10075 * ipa-prop.c (ipa_print_node_params): Fix indentation.
10077 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
10081 * var-tracking.c (add_stores): Preserve the value of the source even if
10082 we don't record the store.
10084 2014-01-06 Terry Guo <terry.guo@arm.com>
10086 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
10088 2014-01-05 Iain Sandoe <iain@codesourcery.com>
10091 * config/darwin.c (darwin_function_section): Adjust return values to
10092 correspond to optimisation changes made in r206070.
10094 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
10096 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
10097 from prefetch_block tune setting.
10098 (nocona_cost): Correct size of prefetch block to 64.
10100 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
10102 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
10103 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
10104 used to save the static chain register in the computation of the offset
10105 from which the FP registers need to be restored.
10107 2014-01-04 Jakub Jelinek <jakub@redhat.com>
10109 PR tree-optimization/59519
10110 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
10111 ICE if get_current_def (current_new_name) is already non-NULL, as long
10112 as it is a phi result of some other phi in *new_exit_bb that has
10115 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
10116 or vmovdqu* for misaligned_operand.
10117 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
10118 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
10119 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
10120 aligned_mem for AVX512F masked aligned load and store builtins and for
10121 non-temporal moves.
10123 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
10125 PR tree-optimization/59651
10126 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
10127 Address range for negative step should be added by TYPE_SIZE_UNIT.
10129 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
10131 * config/m68k/m68k.c (handle_move_double): Handle pushes with
10132 overlapping registers also for registers other than the stack pointer.
10134 2014-01-03 Marek Polacek <polacek@redhat.com>
10137 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
10140 2014-01-03 Jakub Jelinek <jakub@redhat.com>
10143 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
10146 * config/i386/i386.md (MODE_SIZE): New mode attribute.
10147 (push splitter): Use <P:MODE_SIZE> instead of
10148 GET_MODE_SIZE (<P:MODE>mode).
10149 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
10150 (mov -1, reg peephole2): Likewise.
10151 * config/i386/sse.md (*mov<mode>_internal,
10152 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
10153 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
10154 *<code><mode>3, *andnot<mode>3<mask_name>,
10155 <mask_codefor><code><mode>3<mask_name>): Likewise.
10156 * config/i386/subst.md (mask_mode512bit_condition,
10157 sd_mask_mode512bit_condition): Likewise.
10159 2014-01-02 Xinliang David Li <davidxl@google.com>
10161 PR tree-optimization/59303
10162 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
10163 (dump_predicates): Better output format.
10164 (pred_equal_p): New function.
10165 (is_neq_relop_p): Ditto.
10166 (is_neq_zero_form_p): Ditto.
10167 (pred_expr_equal_p): Ditto.
10168 (pred_neg_p): Ditto.
10169 (simplify_pred): Ditto.
10170 (simplify_preds_2): Ditto.
10171 (simplify_preds_3): Ditto.
10172 (simplify_preds_4): Ditto.
10173 (simplify_preds): Ditto.
10174 (push_pred): Ditto.
10175 (push_to_worklist): Ditto.
10176 (get_pred_info_from_cmp): Ditto.
10177 (is_degenerated_phi): Ditto.
10178 (normalize_one_pred_1): Ditto.
10179 (normalize_one_pred): Ditto.
10180 (normalize_one_pred_chain): Ditto.
10181 (normalize_preds): Ditto.
10182 (normalize_cond_1): Remove function.
10183 (normalize_cond): Ditto.
10184 (is_gcond_subset_of): Ditto.
10185 (is_subset_of_any): Ditto.
10186 (is_or_set_subset_of): Ditto.
10187 (is_and_set_subset_of): Ditto.
10188 (is_norm_cond_subset_of): Ditto.
10189 (pred_chain_length_cmp): Ditto.
10190 (convert_control_dep_chain_into_preds): Type change.
10191 (find_predicates): Ditto.
10192 (find_def_preds): Ditto.
10193 (destroy_predicates_vecs): Ditto.
10194 (find_matching_predicates_in_rest_chains): Ditto.
10195 (use_pred_not_overlap_with_undef_path_pred): Ditto.
10196 (is_pred_expr_subset): Ditto.
10197 (is_pred_chain_subset_of): Ditto.
10198 (is_included_in): Ditto.
10199 (is_superset_of): Ditto.
10201 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10203 Update copyright years.
10205 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10207 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
10208 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
10209 config/arc/arc.md, config/arc/arc.opt,
10210 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
10211 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
10212 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
10213 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
10214 config/linux-protos.h, config/linux.c, config/winnt-c.c,
10215 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
10216 vtable-verify.c, vtable-verify.h: Use the standard form for the
10219 2014-01-02 Tobias Burnus <burnus@net-b.de>
10221 * gcc.c (process_command): Update copyright notice dates.
10222 * gcov-dump.c: Ditto.
10224 * doc/cpp.texi: Bump @copying's copyright year.
10225 * doc/cppinternals.texi: Ditto.
10226 * doc/gcc.texi: Ditto.
10227 * doc/gccint.texi: Ditto.
10228 * doc/gcov.texi: Ditto.
10229 * doc/install.texi: Ditto.
10230 * doc/invoke.texi: Ditto.
10232 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10234 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
10236 2014-01-01 Jakub Jelinek <jakub@redhat.com>
10238 * config/i386/sse.md (*mov<mode>_internal): Guard
10239 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
10241 PR rtl-optimization/59647
10242 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
10243 new_rtx into UNSIGNED_FLOAT rtxes.
10245 Copyright (C) 2014 Free Software Foundation, Inc.
10247 Copying and distribution of this file, with or without modification,
10248 are permitted in any medium without royalty provided the copyright
10249 notice and this notice are preserved.