1 2014-05-09 Richard Biener <rguenther@suse.de>
3 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
4 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
5 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
6 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
7 ssa_propagate): Adjust.
9 2014-05-08 Jeff Law <law@redhat.com>
11 PR tree-optimization/61009
12 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
13 tri-state rather than a boolean. When a block is too big to
14 thread through, inform caller via negative return value.
15 (thread_across_edge): If a block was too big for normal threading,
16 then it's too big for a joiner too, so remove temporary equivalences
17 and return immediately.
19 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
20 Matthias Klose <doko@ubuntu.com>
23 * optc-gen.awk: Fix option handling for -Wunused-parameter.
25 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
28 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
30 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
33 * config/alpha/alpha.c: Include gimple-iterator.h.
34 (alpha_gimple_fold_builtin): New function. Move
35 ALPHA_BUILTIN_UMULH folding from ...
36 (alpha_fold_builtin): ... here.
37 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
39 2014-05-08 Wei Mi <wmi@google.com>
42 * config/i386/i386.c (ix86_compute_frame_layout): Update
43 preferred_stack_boundary for call, expanded from tls descriptor.
44 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
45 to depend on SP register.
46 (*tls_local_dynamic_base_32_gnu): Ditto.
47 (*tls_local_dynamic_32_once): Ditto.
48 (tls_global_dynamic_64_<mode>): Set
49 ix86_tls_descriptor_calls_expanded_in_cfun.
50 (tls_local_dynamic_base_64_<mode>): Ditto.
51 (tls_global_dynamic_32): Set
52 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
53 to depend on SP register.
54 (tls_local_dynamic_base_32): Ditto.
56 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
58 * config/arm/arm_neon.h: Update comment.
59 * config/arm/neon-docgen.ml: Delete.
60 * config/arm/neon-gen.ml: Delete.
61 * doc/arm-neon-intrinsics.texi: Update comment.
63 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
65 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
67 (vand, vorr, veor, vorn, vbic): Remove.
68 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
70 (neon_vsub_unspec): Likewise.
71 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
73 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
75 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
85 (vaddq_s16): Likewise.
86 (vaddq_s32): Likewise.
87 (vaddq_s64): Likewise.
88 (vaddq_f32): Likewise.
90 (vaddq_u16): Likewise.
91 (vaddq_u32): Likewise.
92 (vaddq_u64): Likewise.
101 (vmulq_s8): Likewise.
102 (vmulq_s16): Likewise.
103 (vmulq_s32): Likewise.
104 (vmulq_f32): Likewise.
105 (vmulq_u8): Likewise.
106 (vmulq_u16): Likewise.
107 (vmulq_u32): Likewise.
109 (vsub_s16): Likewise.
110 (vsub_s32): Likewise.
111 (vsub_f32): Likewise.
113 (vsub_u16): Likewise.
114 (vsub_u32): Likewise.
115 (vsub_s64): Likewise.
116 (vsub_u64): Likewise.
117 (vsubq_s8): Likewise.
118 (vsubq_s16): Likewise.
119 (vsubq_s32): Likewise.
120 (vsubq_s64): Likewise.
121 (vsubq_f32): Likewise.
122 (vsubq_u8): Likewise.
123 (vsubq_u16): Likewise.
124 (vsubq_u32): Likewise.
125 (vsubq_u64): Likewise.
127 (vand_s16): Likewise.
128 (vand_s32): Likewise.
130 (vand_u16): Likewise.
131 (vand_u32): Likewise.
132 (vand_s64): Likewise.
133 (vand_u64): Likewise.
134 (vandq_s8): Likewise.
135 (vandq_s16): Likewise.
136 (vandq_s32): Likewise.
137 (vandq_s64): Likewise.
138 (vandq_u8): Likewise.
139 (vandq_u16): Likewise.
140 (vandq_u32): Likewise.
141 (vandq_u64): Likewise.
143 (vorr_s16): Likewise.
144 (vorr_s32): Likewise.
146 (vorr_u16): Likewise.
147 (vorr_u32): Likewise.
148 (vorr_s64): Likewise.
149 (vorr_u64): Likewise.
150 (vorrq_s8): Likewise.
151 (vorrq_s16): Likewise.
152 (vorrq_s32): Likewise.
153 (vorrq_s64): Likewise.
154 (vorrq_u8): Likewise.
155 (vorrq_u16): Likewise.
156 (vorrq_u32): Likewise.
157 (vorrq_u64): Likewise.
159 (veor_s16): Likewise.
160 (veor_s32): Likewise.
162 (veor_u16): Likewise.
163 (veor_u32): Likewise.
164 (veor_s64): Likewise.
165 (veor_u64): Likewise.
166 (veorq_s8): Likewise.
167 (veorq_s16): Likewise.
168 (veorq_s32): Likewise.
169 (veorq_s64): Likewise.
170 (veorq_u8): Likewise.
171 (veorq_u16): Likewise.
172 (veorq_u32): Likewise.
173 (veorq_u64): Likewise.
175 (vbic_s16): Likewise.
176 (vbic_s32): Likewise.
178 (vbic_u16): Likewise.
179 (vbic_u32): Likewise.
180 (vbic_s64): Likewise.
181 (vbic_u64): Likewise.
182 (vbicq_s8): Likewise.
183 (vbicq_s16): Likewise.
184 (vbicq_s32): Likewise.
185 (vbicq_s64): Likewise.
186 (vbicq_u8): Likewise.
187 (vbicq_u16): Likewise.
188 (vbicq_u32): Likewise.
189 (vbicq_u64): Likewise.
191 (vorn_s16): Likewise.
192 (vorn_s32): Likewise.
194 (vorn_u16): Likewise.
195 (vorn_u32): Likewise.
196 (vorn_s64): Likewise.
197 (vorn_u64): Likewise.
198 (vornq_s8): Likewise.
199 (vornq_s16): Likewise.
200 (vornq_s32): Likewise.
201 (vornq_s64): Likewise.
202 (vornq_u8): Likewise.
203 (vornq_u16): Likewise.
204 (vornq_u32): Likewise.
205 (vornq_u64): Likewise.
207 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
209 * wide-int.cc (UTItype): Define.
210 (UDWtype): Define for appropriate W_TYPE_SIZE.
212 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
214 PR tree-optimization/59100
215 * tree-ssa-phiopt.c: Include tree-inline.h.
216 (neutral_element_p, absorbing_element_p): New functions.
217 (value_replacement): Handle conditional binary operations with a
218 neutral or absorbing element.
220 2014-05-08 Richard Biener <rguenther@suse.de>
222 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
223 folding the expression.
224 (valueize_expr): Remove.
225 (visit_reference_op_load): Do not valueize the result of
227 (simplify_binary_expression): Likewise.
228 (simplify_unary_expression): Likewise.
230 2014-05-08 Richard Biener <rguenther@suse.de>
232 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
233 looking at TYPE_ARG_TYPES.
235 2014-05-08 Richard Biener <rguenther@suse.de>
237 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
238 pointer propagation special-case.
240 2014-05-08 Bin Cheng <bin.cheng@arm.com>
242 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
243 core part of address expressions.
245 2014-05-08 Alan Modra <amodra@gmail.com>
248 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
249 loads and stores when -mno-strict-align at any alignment.
250 (expand_block_clear): Similarly. Also correct calculation of
253 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
256 * tree-complex.c (expand_complex_move): Keep line info when expanding
258 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
259 of complex expression. Use new argument to display correct location
260 for values coming from phi statement.
261 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
262 (warn_uninitialized_phi): Pass location of phi argument to
264 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
265 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
267 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
269 * config/rs6000/predicates.md (indexed_address_mem): New.
270 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
271 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
272 fpstore_ux, fpstore_u.
273 (sign_extend, indexed, update): New.
274 (cell_micro): Adjust.
275 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
276 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
277 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
278 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
279 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
280 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
281 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
282 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
283 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
284 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
285 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
286 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
287 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
288 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
289 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
291 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
292 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
293 *vsx_extract_<mode>_store): Adjust.
294 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
295 is_cracked_insn, insn_must_be_first_in_group,
296 insn_must_be_last_in_group): Adjust.
298 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
300 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
301 ppc440-fpstore): Adjust.
302 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
303 ppc476-fpstore): Adjust.
304 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
305 ppc601-fpstore): Adjust.
306 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
308 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
310 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
311 ppc7450-fpstore): Adjust.
312 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
313 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
314 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
316 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
317 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
318 cell-fpstore, cell-fpstore-update): Adjust.
319 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
320 ppce300c3_store, ppce300c3_fpstore): Adjust.
321 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
322 e500mc_fpstore): Adjust.
323 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
324 e500mc64_store, e500mc64_fpstore): Adjust.
325 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
326 e5500_fpstore): Adjust.
327 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
328 e6500_fpstore): Adjust.
329 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
331 * config/rs6000/power4.md (power4-load, power4-load-ext,
332 power4-load-ext-update, power4-load-ext-update-indexed,
333 power4-load-update-indexed, power4-load-update, power4-fpload,
334 power4-fpload-update, power4-store, power4-store-update,
335 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
337 * config/rs6000/power5.md (power5-load, power5-load-ext,
338 power5-load-ext-update, power5-load-ext-update-indexed,
339 power5-load-update-indexed, power5-load-update, power5-fpload,
340 power5-fpload-update, power5-store, power5-store-update,
341 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
343 * config/rs6000/power6.md (power6-load, power6-load-ext,
344 power6-load-update, power6-load-update-indexed,
345 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
346 power6-fpload-update, power6-store, power6-store-update,
347 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
349 * config/rs6000/power7.md (power7-load, power7-load-ext,
350 power7-load-update, power7-load-update-indexed,
351 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
352 power7-fpload-update, power7-store, power7-store-update,
353 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
355 * config/rs6000/power8.md (power8-load, power8-load-update,
356 power8-load-ext, power8-load-ext-update, power8-fpload,
357 power8-fpload-update, power8-store, power8-store-update-indexed,
358 power8-fpstore, power8-fpstore-update): Adjust.
359 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
361 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
362 titan_lsu_store, titan_lsu_fpstore): Adjust.
363 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
365 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
368 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
369 unrolled byte insns. Emit address increments after move insns.
371 2014-05-07 David Malcolm <dmalcolm@redhat.com>
373 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
374 const_gimple, rather than a gimple.
375 (gimple_call_builtin_p): Likewise, for the three variants.
377 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
378 (gimple_call_builtin_p): Likewise, for the three variants.
380 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
382 PR tree-optimization/61095
383 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
385 2014-05-07 Richard Biener <rguenther@suse.de>
387 PR tree-optimization/61034
388 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
389 (maybe_skip_until): Use translate to take into account
390 lattices when trying to do disambiguations.
391 (get_continuation_for_phi_1): Likewise.
392 (get_continuation_for_phi): Adjust for added translate arguments.
393 (walk_non_aliased_vuses): Likewise.
394 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
395 (walk_non_aliased_vuses): Likewise.
396 (call_may_clobber_ref_p_1): Declare.
397 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
398 calls. Stop early if we are only supposed to disambiguate.
399 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
401 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
403 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
404 Emit an error when the function has arguments.
406 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
408 * cfgloop.h (unswitch_loops): Remove.
409 * doc/passes.texi: Remove references to loop-unswitch.c
410 * timevar.def (TV_LOOP_UNSWITCH): Remove.
412 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
414 * tree-vect-data-refs.c (vect_grouped_load_supported): New
415 check for loads group of length 3.
416 (vect_permute_load_chain): New permutations for loads group of
418 * tree-vect-stmts.c (vect_model_load_cost): Change cost
419 of vec_perm_shuffle for the new permutations.
421 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
423 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
424 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
425 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
426 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
427 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
428 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
429 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
430 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
432 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
434 * loop-unswitch.c: Delete.
436 2014-05-07 Richard Biener <rguenther@suse.de>
438 * config.gcc: Always set need_64bit_hwint to yes.
440 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
442 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
443 of using optimize_size.
445 2014-05-06 Mike Stump <mikestump@comcast.net>
447 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
449 2014-05-06 Joseph Myers <joseph@codesourcery.com>
451 * config/i386/sse.md (*mov<mode>_internal)
452 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
453 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
454 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
455 (*<code><mode>3, *andnot<mode>3<mask_name>)
456 (<mask_codefor><code><mode>3<mask_name>): Only consider
457 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
459 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
462 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
464 * lra-constraints.c (valid_address_p): Move earlier in file.
465 Add a constraint argument to the address_info version.
466 (satisfies_memory_constraint_p): New function.
467 (satisfies_address_constraint_p): Likewise.
468 (process_alt_operands, curr_insn_transform): Use them.
469 (process_address): Pass the constraint to valid_address_p when
470 checking address operands.
472 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
474 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
475 to their respective blocks. Fix inadvertent use of "node".
477 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
479 * emit-rtl.c (init_derived_machine_modes): New functionm, split
481 (init_emit_once): ...here.
482 * rtl.h (init_derived_machine_modes): Declare.
483 * toplev.c (do_compile): Call it even if no_backend.
485 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
486 Mike Stump <mikestump@comcast.net>
487 Richard Sandiford <rdsandiford@googlemail.com>
488 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
490 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
491 (rtx_equal_for_memref_p): Update comment.
492 (adjust_offset_for_component_ref): Use wide-int interfaces.
493 * builtins.c (get_object_alignment_2): Likewise.
494 (c_readstr): Likewise.
495 (target_char_cast): Add comment.
496 (determine_block_size): Use wide-int interfaces.
497 (expand_builtin_signbit): Likewise.
498 (fold_builtin_int_roundingfn): Likewise.
499 (fold_builtin_bitop): Likewise.
500 (fold_builtin_bswap): Likewise.
501 (fold_builtin_logarithm): Use signop.
502 (fold_builtin_pow): Likewise.
503 (fold_builtin_memory_op): Use wide-int interfaces.
504 (fold_builtin_object_size): Likewise.
505 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
506 nb_iterations_estimate.
507 (record_niter_bound): Use wide-int interfaces.
508 (get_estimated_loop_iterations_int): Likewise.
509 (get_estimated_loop_iterations): Likewise.
510 (get_max_loop_iterations): Likewise.
511 * cfgloop.h: Include wide-int.h.
512 (struct nb_iter_bound): Change bound to widest_int.
513 (struct loop): Change nb_iterations_upper_bound and
514 nb_iterations_estimate to widest_int.
515 (record_niter_bound): Switch to use widest_int.
516 (get_estimated_loop_iterations): Likewise.
517 (get_max_loop_iterations): Likewise.
518 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
520 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
521 * combine.c (try_combine): Likewise.
522 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
523 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
525 (aarch64_float_const_representable_p): Likewise.
526 * config/arc/arc.c: Include wide-int.h.
527 (arc_can_use_doloop_p): Use wide-int interfaces.
528 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
529 (vfp3_const_double_index): Likewise.
530 * config/avr/avr.c (avr_out_round): Likewise.
531 (avr_fold_builtin): Likewise.
532 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
533 (bfin_can_use_doloop_p): Likewise.
534 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
535 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
536 * config/i386/i386.c: Include wide-int.h.
537 (ix86_data_alignment): Use wide-int interfaces.
538 (ix86_local_alignment): Likewise.
539 (ix86_emit_swsqrtsf): Update real_from_integer.
540 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
541 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
542 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
543 (zero_constant): Likewise.
544 (input_operand): Likewise.
545 (splat_input_operand): Likewise.
546 (non_logical_cint_operand): Change const_double to const_wide_int.
547 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
548 (easy_altivec_constant): Remove comment.
549 (paired_expand_vector_init): Use CONSTANT_P.
550 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
551 (rs6000_emit_move): Update checks.
552 (rs6000_aggregate_candidate): Use wide-int interfaces.
553 (rs6000_expand_ternop_builtin): Likewise.
554 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
555 (rs6000_assemble_integer): Likewise.
556 (rs6000_hash_constant): Likewise.
557 (output_toc): Likewise.
558 (rs6000_rtx_costs): Likewise.
559 (rs6000_emit_swrsqrt); Update call to real_from_integer.
560 * config/rs6000/rs6000-c.c: Include wide-int.h.
561 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
562 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
563 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
564 Handle CONST_WIDE_INT.
565 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
566 Use tree_fits_uhwi_p.
567 * config/sparc/sparc.c: Include wide-int.h.
568 (sparc_fold_builtin): Use wide-int interfaces.
569 * config/vax/vax.c: Include wide-int.h.
570 (vax_float_literal): Use real_from_integer.
571 * coretypes.h (struct hwivec_def): New.
574 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
575 (equiv_constant): Handle CONST_WIDE_INT.
576 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
577 (cselib_hash_rtx): Handle CONST_WIDE_INT.
578 * dbxout.c (stabstr_U): Use wide-int interfaces.
579 (dbxout_type): Update to use cst_fits_shwi_p.
580 * defaults.h (LOG2_BITS_PER_UNIT): Define.
581 (TARGET_SUPPORTS_WIDE_INT): Add default.
582 * dfp.c: Include wide-int.h.
583 (decimal_real_to_integer2): Use wide-int interfaces and rename to
584 decimal_real_to_integer.
585 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
586 decimal_real_to_integer.
587 * doc/generic.texi (Constant expressions): Update for wide_int.
588 * doc/rtl.texi (const_double): Likewise.
589 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
590 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
591 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
592 (REAL_VALUE_FROM_INT): Remove.
593 (TARGET_SUPPORTS_WIDE_INT): New.
594 * doc/tm.texi: Regenerate.
595 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
596 * double-int.h: Include wide-int.h.
597 (struct wi::int_traits): New.
598 * dwarf2out.c (get_full_len): New.
599 (dw_val_equal_p): Add case dw_val_class_wide_int.
600 (size_of_loc_descr): Likewise.
601 (output_loc_operands): Likewise.
602 (insert_double): Remove.
603 (insert_wide_int): New.
605 (print_die): Add case dw_val_class_wide_int.
606 (attr_checksum): Likewise.
607 (attr_checksum_ordered): Likewise.
608 (same_dw_val_p): Likewise.
609 (size_of_die): Likewise.
610 (value_format): Likewise.
611 (output_die): Likewise.
612 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
614 (clz_loc_descriptor): Use wide-int interfaces.
615 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
616 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
617 (round_up_to_align): Use wide-int interfaces.
618 (field_byte_offset): Likewise.
619 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
620 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
621 CONST_DOUBLE handling. Use wide-int interfaces.
622 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
623 (gen_enumeration_type_die): Use add_AT_wide.
624 (hash_loc_operands): Add case dw_val_class_wide_int.
625 (compare_loc_operands): Likewise.
626 * dwarf2out.h: Include wide-int.h.
628 (enum dw_val_class): Add dw_val_class_wide_int.
629 (struct dw_val_struct): Add val_wide.
630 * emit-rtl.c (const_wide_int_htab): New.
631 (const_wide_int_htab_hash): New.
632 (const_wide_int_htab_eq): New.
633 (lookup_const_wide_int): New.
634 (const_double_htab_hash): Use wide-int interfaces.
635 (const_double_htab_eq): Likewise.
636 (rtx_to_double_int): Conditionally compile for wide-int.
637 (immed_double_int_const): Rename to immed_wide_int_const and
639 (immed_double_const): Conditionally compile for wide-int.
640 (init_emit_once): Use wide-int interfaces.
641 * explow.c (plus_constant): Likewise.
642 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
643 (lshift_value): Use wide-int interfaces.
644 (expand_mult): Likewise.
645 (choose_multiplier): Likewise.
646 (expand_smod_pow2): Likewise.
647 (make_tree): Likewise.
648 * expr.c (convert_modes): Consolidate handling of constants.
649 Use wide-int interfaces.
650 (emit_group_load_1): Add note.
651 (store_expr): Update comment.
652 (get_inner_reference): Use wide-int interfaces.
653 (expand_constructor): Update comment.
654 (expand_expr_real_2): Use wide-int interfaces.
655 (expand_expr_real_1): Likewise.
656 (reduce_to_bit_field_precision): Likewise.
657 (const_vector_from_tree): Likewise.
658 * final.c: Include wide-int-print.h.
659 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
660 * fixed-value.c: Include wide-int.h.
661 (fixed_from_string): Use wide-int interfaces.
662 (fixed_to_decimal): Likewise.
663 (fixed_convert_from_real): Likewise.
664 (real_convert_from_fixed): Likewise.
665 * fold-const.h (mem_ref_offset): Return an offset_int.
666 (div_if_zero_remainder): Remove code parameter.
667 * fold-const.c (div_if_zero_remainder): Remove code parameter.
668 Use wide-int interfaces.
669 (may_negate_without_overflow_p): Use wide-int interfaces.
670 (negate_expr_p): Likewise.
671 (fold_negate_expr): Likewise.
672 (int_const_binop_1): Likewise.
673 (const_binop): Likewise.
674 (fold_convert_const_int_from_int): Likewise.
675 (fold_convert_const_int_from_real): Likewise.
676 (fold_convert_const_int_from_fixed): Likewise.
677 (fold_convert_const_fixed_from_int): Likewise.
678 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
679 (sign_bit_p): Use wide-int interfaces.
680 (make_range_step): Likewise.
681 (build_range_check): Likewise. Pass an integer of the correct type
682 instead of using integer_one_node.
683 (range_predecessor): Pass an integer of the correct type instead
684 of using integer_one_node.
685 (range_successor): Likewise.
686 (merge_ranges): Likewise.
687 (unextend): Use wide-int interfaces.
688 (extract_muldiv_1): Likewise.
689 (fold_div_compare): Likewise.
690 (fold_single_bit_test): Likewise.
691 (fold_sign_changed_comparison): Likewise.
692 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
693 (fold_plusminus_mult_expr): Use wide-int interfaces.
694 (native_encode_int): Likewise.
695 (native_interpret_int): Likewise.
696 (fold_unary_loc): Likewise.
697 (pointer_may_wrap_p): Likewise.
698 (size_low_cst): Likewise.
699 (mask_with_tz): Likewise.
700 (fold_binary_loc): Likewise.
701 (fold_ternary_loc): Likewise.
702 (multiple_of_p): Likewise.
703 (tree_call_nonnegative_warnv_p): Update calls to
704 tree_int_cst_min_precision and real_from_integer.
705 (fold_negate_const): Use wide-int interfaces.
706 (fold_abs_const): Likewise.
707 (fold_relational_const): Use tree_int_cst_lt.
708 (round_up_loc): Use wide-int interfaces.
709 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
710 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
711 * gengtype.c: Remove include of double-int.h.
712 (do_typedef): Use wide-int interfaces.
713 (open_base_files): Add wide-int.h.
714 (main): Add offset_int and widest_int typedefs.
715 * gengtype-lex.l: Handle "^".
716 (CXX_KEYWORD): Add "static".
717 * gengtype-parse.c (require3): New.
718 (require_template_declaration): Handle constant template arguments
719 and nested templates.
720 * gengtype-state.c: Don't include "double-int.h".
721 * genpreds.c (write_one_predicate_function): Update comment.
722 (write_tm_constrs_h): Add check for hval and lval use in
724 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
725 (add_to_sequence): Likewise.
726 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
727 and const_double_operand.
728 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
730 * gimple-fold.c (get_base_constructor): Likewise.
731 (fold_array_ctor_reference): Likewise.
732 (fold_nonarray_ctor_reference): Likewise.
733 (fold_const_aggregate_ref_1): Likewise.
734 (gimple_val_nonnegative_real_p): Likewise.
735 (gimple_fold_indirect_ref): Likewise.
736 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
737 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
738 (struct slsr_cand_d): Change index to be widest_int.
739 (struct incr_info_d): Change incr to be widest_int.
740 (alloc_cand_and_find_basis): Use wide-int interfaces.
741 (slsr_process_phi): Likewise.
742 (backtrace_base_for_ref): Likewise. Return a widest_int.
743 (restructure_reference): Take a widest_int instead of a double_int.
744 (slsr_process_ref): Use wide-int interfaces.
745 (create_mul_ssa_cand): Likewise.
746 (create_mul_imm_cand): Likewise.
747 (create_add_ssa_cand): Likewise.
748 (create_add_imm_cand): Take a widest_int instead of a double_int.
749 (slsr_process_add): Use wide-int interfaces.
750 (slsr_process_cast): Likewise.
751 (slsr_process_copy): Likewise.
752 (dump_candidate): Likewise.
753 (dump_incr_vec): Likewise.
754 (replace_ref): Likewise.
755 (cand_increment): Likewise. Return a widest_int.
756 (cand_abs_increment): Likewise.
757 (replace_mult_candidate): Take a widest_int instead of a double_int.
758 (replace_unconditional_candidate): Use wide-int interfaces.
759 (incr_vec_index): Take a widest_int instead of a double_int.
760 (create_add_on_incoming_edge): Likewise.
761 (create_phi_basis): Use wide-int interfaces.
762 (replace_conditional_candidate): Likewise.
763 (record_increment): Take a widest_int instead of a double_int.
764 (record_phi_increments): Use wide-int interfaces.
765 (phi_incr_cost): Take a widest_int instead of a double_int.
766 (lowest_cost_path): Likewise.
767 (total_savings): Likewise.
768 (analyze_increments): Use wide-int interfaces.
769 (ncd_with_phi): Take a widest_int instead of a double_int.
770 (ncd_of_cand_and_phis): Likewise.
771 (nearest_common_dominator_for_cands): Likewise.
772 (insert_initializers): Use wide-int interfaces.
773 (all_phi_incrs_profitable): Likewise.
774 (replace_one_candidate): Likewise.
775 (replace_profitable_candidates): Likewise.
776 * godump.c: Include wide-int-print.h.
777 (go_output_typedef): Use wide-int interfaces.
778 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
779 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
780 (build_loop_iteration_domains): Likewise.
781 * hooks.h: Include wide-int.h rather than double-int.h.
782 (hook_bool_dint_dint_uint_bool_true): Delete.
783 (hook_bool_wint_wint_uint_bool_true): Declare.
784 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
785 (hook_bool_wint_wint_uint_bool_true): New.
786 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
788 (ubsan_expand_si_overflow_mul_check): Likewise.
789 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
790 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
791 (get_ancestor_addr_info): Likewise.
792 (ipa_modify_call_arguments): Likewise.
793 * loop-doloop.c (doloop_modify): Likewise.
794 (doloop_optimize): Likewise.
795 * loop-iv.c (iv_number_of_iterations): Likewise.
796 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
797 (unroll_loop_constant_iterations): Likewise.
798 (decide_unroll_runtime_iterations): Likewise.
799 (unroll_loop_runtime_iterations): Likewise.
800 (decide_peel_simple): Likewise.
801 (decide_unroll_stupid): Likewise.
802 * lto-streamer-in.c (streamer_read_wi): Add.
803 (input_cfg): Use wide-int interfaces.
804 (lto_input_tree_1): Likewise.
805 * lto-streamer-out.c (streamer_write_wi): Add.
806 (hash_tree): Use wide-int interfaces.
807 (output_cfg): Likewise.
808 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
809 (GTFILES): Add wide-int.h and signop.h.
810 (TAGS): Look for .cc files too.
811 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
812 * optabs.c (expand_subword_shift): Likewise.
813 (expand_doubleword_shift): Likewise.
814 (expand_absneg_bit): Likewise.
815 (expand_copysign_absneg): Likewise.
816 (expand_copysign_bit): Likewise.
817 * postreload.c (reload_cse_simplify_set): Likewise.
818 * predict.c (predict_iv_comparison): Likewise.
819 * pretty-print.h: Include wide-int-print.h.
821 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
822 * print-tree.c: Include wide-int-print.h.
823 (print_node_brief): Use wide-int interfaces.
824 (print_node): Likewise.
825 * read-rtl.c (validate_const_wide_int): New.
826 (read_rtx_code): Add CONST_WIDE_INT case.
827 * real.c: Include wide-int.h.
828 (real_to_integer2): Delete.
829 (real_to_integer): New function, returning a wide_int.
830 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
831 (ten_to_ptwo): Update call to real_from_integer.
832 (real_digit): Likewise.
833 * real.h: Include signop.h, wide-int.h and insn-modes.h.
834 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
835 (REAL_VALUE_TO_INT): Delete.
836 (real_to_integer): Declare a wide-int form.
837 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
838 * recog.c (const_int_operand): Improve comment.
839 (const_scalar_int_operand): New.
840 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
841 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
842 (split_double): Likewise.
843 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
844 (rtx_size): Likewise.
845 (rtx_alloc_stat_v): New.
846 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
847 (cwi_output_hex): New.
848 (iterative_hash_rtx): Handle CONST_WIDE_INT.
849 (cwi_check_failed_bounds): New.
850 * rtl.def (CONST_WIDE_INT): New.
851 * rtl.h: Include <utility> and wide-int.h.
852 (struct hwivec_def): New.
853 (CWI_GET_NUM_ELEM): New.
854 (CWI_PUT_NUM_ELEM): New.
855 (struct rtx_def): Add num_elem and hwiv.
856 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
857 (CASE_CONST_UNIQUE): Likewise.
858 (CASE_CONST_ANY): Likewise.
859 (CONST_SCALAR_INT_P): Likewise.
860 (CONST_WIDE_INT_P): New.
863 (cwi_check_failed_bounds): New.
866 (CONST_WIDE_INT_VEC) New.
867 (CONST_WIDE_INT_NUNITS) New.
868 (CONST_WIDE_INT_ELT) New.
869 (rtx_mode_t): New type.
870 (wi::int_traits <rtx_mode_t>): New.
872 (wi::min_value): New.
873 (wi::max_value): New.
875 (const_wide_int_alloc): New.
876 (immed_wide_int_const): New.
877 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
878 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
879 * signop.h: New file.
880 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
881 (simplify_const_unary_operation): Use wide-int interfaces.
882 (simplify_binary_operation_1): Likewise.
883 (simplify_const_binary_operation): Likewise.
884 (simplify_const_relational_operation): Likewise.
885 (simplify_immed_subreg): Likewise.
886 * stmt.c (expand_case): Likewise.
887 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
888 signop rather than a bool.
889 * stor-layout.c (layout_type): Use wide-int interfaces.
890 (initialize_sizetypes): Update calls to
891 set_min_and_max_values_for_integral_type.
892 (set_min_and_max_values_for_integral_type): Take a signop rather
893 than a bool. Use wide-int interfaces.
894 (fixup_signed_type): Update accordingly. Remove
895 HOST_BITS_PER_DOUBLE_INT limit.
896 (fixup_unsigned_type): Likewise.
897 * system.h (STATIC_CONSTANT_P): New.
898 (STATIC_ASSERT): New.
899 * target.def (can_use_doloop_p): Take widest_ints rather than
901 * target.h: Include wide-int.h rather than double-int.h.
902 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
904 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
905 rather than INT_CST_LT_UNSIGNED.
906 (can_use_doloop_if_innermost): Take widest_ints rather than
908 * tree-affine.c: Include wide-int-print.h.
909 (double_int_ext_for_comb): Delete.
910 (wide_int_ext_for_comb): New.
911 (aff_combination_zero): Use wide-int interfaces.
912 (aff_combination_const): Take a widest_int instead of a double_int.
913 (aff_combination_elt): Use wide-int interfaces.
914 (aff_combination_scale): Take a widest_int instead of a double_int.
915 (aff_combination_add_elt): Likewise.
916 (aff_combination_add_cst): Likewise.
917 (aff_combination_add): Use wide-int interfaces.
918 (aff_combination_convert): Likewise.
919 (tree_to_aff_combination): Likewise.
920 (add_elt_to_tree): Take a widest_int instead of a double_int.
921 (aff_combination_to_tree): Use wide-int interfaces.
922 (aff_combination_remove_elt): Likewise.
923 (aff_combination_add_product): Take a widest_int instead of
925 (aff_combination_mult): Use wide-int interfaces.
926 (aff_combination_expand): Likewise.
927 (double_int_constant_multiple_p): Delete.
928 (wide_int_constant_multiple_p): New.
929 (aff_combination_constant_multiple_p): Take a widest_int pointer
930 instead of a double_int pointer.
931 (print_aff): Use wide-int interfaces.
932 (get_inner_reference_aff): Take a widest_int pointer
933 instead of a double_int pointer.
934 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
935 * tree-affine.h: Include wide-int.h.
936 (struct aff_comb_elt): Change type of coef to widest_int.
937 (struct affine_tree_combination): Change type of offset to widest_int.
938 (double_int_ext_for_comb): Delete.
939 (wide_int_ext_for_comb): New.
940 (aff_combination_const): Use widest_int instead of double_int.
941 (aff_combination_scale): Likewise.
942 (aff_combination_add_elt): Likewise.
943 (aff_combination_constant_multiple_p): Likewise.
944 (get_inner_reference_aff): Likewise.
945 (aff_comb_cannot_overlap_p): Likewise.
946 (aff_combination_zero_p): Use wide-int interfaces.
947 * tree.c: Include tree.h.
948 (init_ttree): Use make_int_cst.
949 (tree_code_size): Removed code for INTEGER_CST case.
950 (tree_size): Add INTEGER_CST case.
951 (make_node_stat): Update comment.
952 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
953 (build_int_cst_type): Use wide-int interfaces.
954 (double_int_to_tree): Likewise.
955 (double_int_fits_to_tree_p): Delete.
956 (force_fit_type_double): Delete.
957 (force_fit_type): New.
958 (int_cst_hash_hash): Use wide-int interfaces.
959 (int_cst_hash_eq): Likewise.
960 (build_int_cst_wide): Delete.
961 (wide_int_to_tree): New.
962 (cache_integer_cst): Use wide-int interfaces.
963 (build_low_bits_mask): Likewise.
964 (cst_and_fits_in_hwi): Likewise.
965 (real_value_from_int_cst): Likewise.
966 (make_int_cst_stat): New.
967 (integer_zerop): Use wide_int interfaces.
968 (integer_onep): Likewise.
969 (integer_all_onesp): Likewise.
970 (integer_pow2p): Likewise.
971 (integer_nonzerop): Likewise.
972 (tree_log2): Likewise.
973 (tree_floor_log2): Likewise.
974 (tree_ctz): Likewise.
975 (int_size_in_bytes): Likewise.
976 (mem_ref_offset): Return an offset_int rather than a double_int.
977 (build_type_attribute_qual_variant): Use wide_int interfaces.
978 (type_hash_eq): Likewise
979 (tree_int_cst_equal): Likewise.
980 (tree_int_cst_lt): Delete.
981 (tree_int_cst_compare): Likewise.
982 (tree_fits_shwi_p): Use wide_int interfaces.
983 (tree_fits_uhwi_p): Likewise.
984 (tree_int_cst_sign_bit): Likewise.
985 (tree_int_cst_sgn): Likewise.
986 (tree_int_cst_min_precision): Take a signop rather than a bool.
987 (simple_cst_equal): Use wide_int interfaces.
988 (compare_tree_int): Likewise.
989 (iterative_hash_expr): Likewise.
990 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
992 (get_type_static_bounds): Use wide_int interfaces.
993 (tree_int_cst_elt_check_failed): New.
994 (build_common_tree_nodes): Reordered to set prec before filling in
996 (int_cst_value): Check cst_and_fits_in_hwi.
997 (widest_int_cst_value): Use wide_int interfaces.
998 (upper_bound_in_type): Likewise.
999 (lower_bound_in_type): Likewise.
1000 (num_ending_zeros): Likewise.
1001 (drop_tree_overflow): Likewise.
1002 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
1003 (gen_conditions_for_pow_cst_base): Likewise.
1004 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
1005 (group_case_labels_stmt): Use wide-int interfaces.
1006 (verify_gimple_assign_binary): Likewise.
1007 (print_loop): Likewise.
1008 * tree-chrec.c (tree_fold_binomial): Likewise.
1009 * tree-core.h (struct tree_base): Add int_length.
1010 (struct tree_int_cst): Change rep of value.
1011 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
1012 (dr_may_alias_p): Likewise.
1013 (max_stmt_executions_tree): Likewise.
1014 * tree.def (INTEGER_CST): Update comment.
1015 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
1016 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
1017 * tree-dump.c: Include wide-int.h and wide-int-print.h.
1018 (dequeue_and_dump): Use wide-int interfaces.
1019 * tree.h: Include wide-int.h.
1020 (NULL_TREE): Moved to earlier loc in file.
1021 (TREE_INT_CST_ELT_CHECK): New.
1022 (tree_int_cst_elt_check_failed): New.
1024 (TREE_INT_CST): Delete.
1025 (TREE_INT_CST_LOW): Use wide-int interfaces.
1026 (TREE_INT_CST_HIGH): Delete.
1027 (TREE_INT_CST_NUNITS): New.
1028 (TREE_INT_CST_EXT_NUNITS): Likewise.
1029 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
1030 (TREE_INT_CST_ELT): Likewise.
1031 (INT_CST_LT): Delete.
1032 (tree_int_cst_elt_check): New (two forms).
1033 (type_code_size): Update comment.
1034 (make_int_cst_stat, make_int_cst): New.
1035 (tree_to_double_int): Delete.
1036 (double_int_fits_to_tree_p): Delete.
1037 (force_fit_type_double): Delete.
1038 (build_int_cstu): Replace with out-of-line function.
1039 (build_int_cst_wide): Delete.
1040 (tree_int_cst_lt): Define inline.
1041 (tree_int_cst_le): New.
1042 (tree_int_cst_compare): Define inline.
1043 (tree_int_cst_min_precision): Take a signop rather than a bool.
1044 (wi::int_traits <const_tree>): New.
1045 (wi::int_traits <tree>): New.
1046 (wi::extended_tree): New.
1047 (wi::int_traits <wi::extended_tree>): New.
1048 (wi::to_widest): New.
1049 (wi::to_offset): New.
1050 (wi::fits_to_tree_p): New.
1051 (wi::min_value): New.
1052 (wi::max_value): New.
1053 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
1054 (copy_tree_body_r): Likewise.
1055 * tree-object-size.c (compute_object_offset): Likewise.
1056 (addr_object_size): Likewise.
1057 * tree-predcom.c: Include wide-int-print.h.
1058 (struct dref_d): Change type of offset to widest_int.
1059 (dump_dref): Call wide-int printer.
1060 (aff_combination_dr_offset): Use wide-int interfaces.
1061 (determine_offset): Take a widest_int pointer rather than a
1063 (split_data_refs_to_components): Use wide-int interfaces.
1064 (suitable_component_p): Likewise.
1065 (order_drefs): Likewise.
1066 (add_ref_to_chain): Likewise.
1067 (valid_initializer_p): Likewise.
1068 (determine_roots_comp): Likewise.
1069 * tree-pretty-print.c: Include wide-int-print.h.
1070 (dump_generic_node): Use wide-int interfaces.
1071 * tree-sra.c (sra_ipa_modify_expr): Likewise.
1072 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
1073 (move_fixed_address_to_symbol): Likewise.
1074 (move_hint_to_base): Likewise.
1075 (move_pointer_to_base): Likewise.
1076 (move_variant_to_index): Likewise.
1077 (most_expensive_mult_to_index): Likewise.
1078 (addr_to_parts): Likewise.
1079 (copy_ref_info): Likewise.
1080 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
1081 (indirect_refs_may_alias_p): Likewise.
1082 (stmt_kills_ref_p_1): Likewise.
1083 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
1084 * tree-ssa-ccp.c: Update comment at top of file. Include
1086 (struct prop_value_d): Change type of mask to widest_int.
1087 (extend_mask): New function.
1088 (dump_lattice_value): Use wide-int interfaces.
1089 (get_default_value): Likewise.
1090 (set_constant_value): Likewise.
1091 (set_value_varying): Likewise.
1092 (valid_lattice_transition): Likewise.
1093 (set_lattice_value): Likewise.
1094 (value_to_double_int): Delete.
1095 (value_to_wide_int): New.
1096 (get_value_from_alignment): Use wide-int interfaces.
1097 (get_value_for_expr): Likewise.
1098 (do_dbg_cnt): Likewise.
1099 (ccp_finalize): Likewise.
1100 (ccp_lattice_meet): Likewise.
1101 (bit_value_unop_1): Use widest_ints rather than double_ints.
1102 (bit_value_binop_1): Likewise.
1103 (bit_value_unop): Use wide-int interfaces.
1104 (bit_value_binop): Likewise.
1105 (bit_value_assume_aligned): Likewise.
1106 (evaluate_stmt): Likewise.
1107 (ccp_fold_stmt): Likewise.
1108 (visit_cond_stmt): Likewise.
1109 (ccp_visit_stmt): Likewise.
1110 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
1111 (constant_pointer_difference): Likewise.
1112 (associate_pointerplus): Likewise.
1113 (combine_conversions): Likewise.
1114 * tree-ssa-loop.h: Include wide-int.h.
1115 (struct tree_niter_desc): Change type of max to widest_int.
1116 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
1117 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
1118 (remove_redundant_iv_tests): Likewise.
1119 (canonicalize_loop_induction_variables): Likewise.
1120 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
1121 (constant_multiple_of): Take a widest_int pointer instead of
1122 a double_int pointer.
1123 (get_computation_aff): Use wide-int interfaces.
1124 (ptr_difference_cost): Likewise.
1125 (difference_cost): Likewise.
1126 (get_loop_invariant_expr_id): Likewise.
1127 (get_computation_cost_at): Likewise.
1128 (iv_elimination_compare_lt): Likewise.
1129 (may_eliminate_iv): Likewise.
1130 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
1131 instead of double_int.
1132 (max_loop_iterations): Likewise.
1133 (max_stmt_executions): Likewise.
1134 (estimated_stmt_executions): Likewise.
1135 * tree-ssa-loop-niter.c: Include wide-int-print.h.
1136 (split_to_var_and_offset): Use wide-int interfaces.
1137 (determine_value_range): Likewise.
1138 (bound_difference_of_offsetted_base): Likewise.
1139 (bounds_add): Take a widest_int instead of a double_int.
1140 (number_of_iterations_ne_max): Use wide-int interfaces.
1141 (number_of_iterations_ne): Likewise.
1142 (number_of_iterations_lt_to_ne): Likewise.
1143 (assert_loop_rolls_lt): Likewise.
1144 (number_of_iterations_lt): Likewise.
1145 (number_of_iterations_le): Likewise.
1146 (number_of_iterations_cond): Likewise.
1147 (number_of_iterations_exit): Likewise.
1148 (finite_loop_p): Likewise.
1149 (derive_constant_upper_bound_assign): Likewise.
1150 (derive_constant_upper_bound): Return a widest_int.
1151 (derive_constant_upper_bound_ops): Likewise.
1152 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
1153 (record_estimate): Take a widest_int rather than a double_int.
1154 (record_nonwrapping_iv): Use wide-int interfaces.
1155 (double_int_cmp): Delete.
1156 (wide_int_cmp): New.
1157 (bound_index): Take a widest_int rather than a double_int.
1158 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
1159 (maybe_lower_iteration_bound): Likewise.
1160 (estimate_numbers_of_iterations_loop): Likewise.
1161 (estimated_loop_iterations): Take a widest_int pointer than than
1162 a double_int pointer.
1163 (estimated_loop_iterations_int): Use wide-int interfaces.
1164 (max_loop_iterations): Take a widest_int pointer than than
1165 a double_int pointer.
1166 (max_loop_iterations_int): Use wide-int interfaces.
1167 (max_stmt_executions): Take a widest_int pointer than than
1168 a double_int pointer.
1169 (estimated_stmt_executions): Likewise.
1170 (n_of_executions_at_most): Use wide-int interfaces.
1171 (scev_probably_wraps_p): Likewise.
1172 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
1174 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
1176 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
1177 double_ints. Adjust for trailing_wide_ints <3> representation.
1178 (set_nonzero_bits): Likewise.
1179 (get_range_info): Return wide_ints rather than double_ints.
1180 Adjust for trailing_wide_ints <3> representation.
1181 (get_nonzero_bits): Likewise.
1182 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
1184 * tree-ssanames.h (struct range_info_def): Replace min, max and
1185 nonzero_bits with a trailing_wide_ints <3>.
1186 (set_range_info): Use wide_int_refs rather than double_ints.
1187 (set_nonzero_bits): Likewise.
1188 (get_range_info): Return wide_ints rather than double_ints.
1189 (get_nonzero_bits): Likewise.
1190 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
1191 * tree-ssa-pre.c (phi_translate_1): Likewise.
1192 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
1193 (acceptable_pow_call): Likewise.
1194 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
1196 (vn_reference_fold_indirect): Likewise.
1197 (vn_reference_maybe_forwprop_address): Likewise.
1198 (valueize_refs_1): Likewise.
1199 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
1200 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
1201 tree_int_cst_lt and tree_int_cst_le.
1202 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
1204 (streamer_alloc_tree): Likewise.
1205 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
1206 (streamer_write_tree_header): Likewise.
1207 (streamer_write_integer_cst): Likewise.
1208 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
1209 (build_constructors): Likewise.
1210 (array_value_type): Likewise.
1211 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
1212 (vect_check_gather): Likewise.
1213 * tree-vect-generic.c (build_replicated_const): Likewise.
1214 (expand_vector_divmod): Likewise.
1215 * tree-vect-loop.c (vect_transform_loop): Likewise.
1216 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
1217 (vect_do_peeling_for_alignment): Likewise.
1218 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
1219 * tree-vrp.c: Include wide-int.h.
1220 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
1221 (extract_range_from_assert): Use wide-int interfaces.
1222 (vrp_int_const_binop): Likewise.
1223 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
1224 double_int pointers.
1225 (ranges_from_anti_range): Use wide-int interfaces.
1226 (quad_int_cmp): Delete.
1227 (quad_int_pair_sort): Likewise.
1228 (extract_range_from_binary_expr_1): Use wide-int interfaces.
1229 (extract_range_from_unary_expr_1): Likewise.
1230 (adjust_range_with_scev): Likewise.
1231 (masked_increment): Take and return wide_ints rather than double_ints.
1232 (register_edge_assert_for_2): Use wide-int interfaces.
1233 (check_array_ref): Likewise.
1234 (search_for_addr_array): Likewise.
1235 (maybe_set_nonzero_bits): Likewise.
1236 (union_ranges): Pass an integer of the correct type instead of
1237 using integer_one_node.
1238 (intersect_ranges): Likewise.
1239 (simplify_truth_ops_using_ranges): Likewise.
1240 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
1241 (range_fits_type_p): Likewise.
1242 (simplify_cond_using_ranges): Likewise. Take a signop rather than
1244 (simplify_conversion_using_ranges): Use wide-int interfaces.
1245 (simplify_float_conversion_using_ranges): Likewise.
1246 (vrp_finalize): Likewise.
1247 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
1248 (gimple_stringops_transform): Likewise.
1249 * varasm.c (decode_addr_const): Likewise.
1250 (const_hash_1): Likewise.
1251 (const_rtx_hash_1): Likewise
1252 (output_constant): Likewise.
1253 (array_size_for_constructor): Likewise.
1254 (output_constructor_regular_field): Likewise.
1255 (output_constructor_bitfield): Likewise.
1256 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
1257 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
1259 * gencheck.c: Define BITS_PER_UNIT.
1262 * wide-int-print.cc: New.
1263 * wide-int-print.h: New.
1265 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1267 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
1269 2014-05-06 Richard Biener <rguenther@suse.de>
1271 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
1272 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
1273 (TODO_verify_all): Adjust.
1274 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
1275 TODO_verify_stmts and TODO_verify_rtl_sharing.
1276 * bb-reorder.c: Likewise.
1277 * cfgexpand.c: Likewise.
1278 * cprop.c: Likewise.
1280 * function.c: Likewise.
1281 * fwprop.c: Likewise.
1283 * gimple-ssa-isolate-paths.c: Likewise.
1284 * gimple-ssa-strength-reduction.c: Likewise.
1285 * ipa-split.c: Likewise.
1286 * loop-init.c: Likewise.
1287 * loop-unroll.c: Likewise.
1288 * lower-subreg.c: Likewise.
1289 * modulo-sched.c: Likewise.
1290 * postreload-gcse.c: Likewise.
1291 * predict.c: Likewise.
1292 * recog.c: Likewise.
1293 * sched-rgn.c: Likewise.
1294 * store-motion.c: Likewise.
1295 * tracer.c: Likewise.
1296 * trans-mem.c: Likewise.
1297 * tree-call-cdce.c: Likewise.
1298 * tree-cfg.c: Likewise.
1299 * tree-cfgcleanup.c: Likewise.
1300 * tree-complex.c: Likewise.
1301 * tree-eh.c: Likewise.
1302 * tree-emutls.c: Likewise.
1303 * tree-if-conv.c: Likewise.
1304 * tree-into-ssa.c: Likewise.
1305 * tree-loop-distribution.c: Likewise.
1306 * tree-object-size.c: Likewise.
1307 * tree-parloops.c: Likewise.
1308 * tree-pass.h: Likewise.
1309 * tree-sra.c: Likewise.
1310 * tree-ssa-ccp.c: Likewise.
1311 * tree-ssa-copy.c: Likewise.
1312 * tree-ssa-copyrename.c: Likewise.
1313 * tree-ssa-dce.c: Likewise.
1314 * tree-ssa-dom.c: Likewise.
1315 * tree-ssa-dse.c: Likewise.
1316 * tree-ssa-forwprop.c: Likewise.
1317 * tree-ssa-ifcombine.c: Likewise.
1318 * tree-ssa-loop-ch.c: Likewise.
1319 * tree-ssa-loop-ivcanon.c: Likewise.
1320 * tree-ssa-loop.c: Likewise.
1321 * tree-ssa-math-opts.c: Likewise.
1322 * tree-ssa-phiopt.c: Likewise.
1323 * tree-ssa-phiprop.c: Likewise.
1324 * tree-ssa-pre.c: Likewise.
1325 * tree-ssa-reassoc.c: Likewise.
1326 * tree-ssa-sink.c: Likewise.
1327 * tree-ssa-strlen.c: Likewise.
1328 * tree-ssa-tail-merge.c: Likewise.
1329 * tree-ssa-uncprop.c: Likewise.
1330 * tree-switch-conversion.c: Likewise.
1331 * tree-tailcall.c: Likewise.
1332 * tree-vect-generic.c: Likewise.
1333 * tree-vectorizer.c: Likewise.
1334 * tree-vrp.c: Likewise.
1336 * var-tracking.c: Likewise.
1337 * bt-load.c: Likewise.
1338 * cfgcleanup.c: Likewise.
1339 * combine-stack-adj.c: Likewise.
1340 * combine.c: Likewise.
1341 * compare-elim.c: Likewise.
1342 * config/epiphany/resolve-sw-modes.c: Likewise.
1343 * config/i386/i386.c: Likewise.
1344 * config/mips/mips.c: Likewise.
1345 * config/s390/s390.c: Likewise.
1346 * config/sh/sh_treg_combine.cc: Likewise.
1347 * config/sparc/sparc.c: Likewise.
1350 * final.c: Likewise.
1351 * ifcvt.c: Likewise.
1352 * mode-switching.c: Likewise.
1353 * passes.c: Likewise.
1354 * postreload.c: Likewise.
1356 * reg-stack.c: Likewise.
1357 * regcprop.c: Likewise.
1358 * regrename.c: Likewise.
1361 2014-05-06 Richard Biener <rguenther@suse.de>
1364 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
1365 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
1367 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
1370 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
1372 2014-05-05 Radovan Obradovic <robradovic@mips.com>
1373 Tom de Vries <tom@codesourcery.com>
1375 * target.def (call_fusage_contains_non_callee_clobbers): New
1377 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
1379 (@node Miscellaneous Register Hooks): New node.
1380 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
1381 * doc/tm.texi: Regenerate.
1383 2014-05-05 Marek Polacek <polacek@redhat.com>
1386 * opts.c (common_handle_option): Call error_at instead of warning_at.
1388 2014-05-05 Richard Biener <rguenther@suse.de>
1390 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
1391 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
1392 under the TODO_verify_il umbrella.
1394 2014-05-05 Richard Biener <rguenther@suse.de>
1396 * passes.c (execute_function_todo): Move TODO_verify_flow under
1397 the TODO_verify_ul umbrella.
1399 2014-05-05 Richard Biener <rguenther@suse.de>
1402 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
1403 X & CST away from a CST that is the mask of a mode.
1405 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1407 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
1408 int argument to enum machine_mode.
1409 (picochip_class_max_nregs): Ditto.
1410 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
1411 (picochip_class_max_nregs): Ditto.
1413 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1415 * target.def: Add new target hook.
1416 * doc/tm.texi: Regenerate.
1417 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
1418 * targhooks.c (default_keep_leaf_when_profiled): New function.
1420 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
1421 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
1423 2014-05-05 Bin Cheng <bin.cheng@arm.com>
1425 PR tree-optimization/60363
1426 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
1427 (copy_phi_args): New parameters. Call get_value_locus_in_path.
1428 (update_destination_phis): New parameter.
1429 (create_edge_and_update_destination_phis): Ditto.
1430 (ssa_fix_duplicate_block_edges): Pass new arguments.
1431 (thread_single_edge): Ditto.
1433 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
1435 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
1436 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
1437 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
1438 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
1439 Use RS6000_BTM_HARD_FLOAT.
1440 (BU_MISC_2): Likewise.
1441 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
1442 RS6000_BTM_HARD_FLOAT.
1443 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
1445 (rs6000_invalid_builtin): Add hard floating builtin support.
1446 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
1447 hard float builtins.
1448 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
1450 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1452 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
1453 Add missing function* argument.
1455 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1457 * lra-constraints.c (valid_address_p): Move earlier in file.
1458 Add a constraint argument to the address_info version.
1459 (satisfies_memory_constraint_p): New function.
1460 (satisfies_address_constraint_p): Likewise.
1461 (process_alt_operands, curr_insn_transform): Use them.
1462 (process_address): Pass the constraint to valid_address_p when
1463 checking address operands.
1465 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1467 * config/mips/mips.c (mips_isa_rev): New variable.
1468 (mips_set_architecture): Set it.
1469 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1471 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1472 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1473 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1474 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1475 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
1476 conditions in terms of mips_isa_rev.
1477 (mips_isa_rev): Declare.
1479 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1481 * config/sh/sh-mem.cc: Use tabs instead of spaces.
1482 (prob_unlikely, prob_likely): Make variables const.
1484 2014-05-03 Denis Chertykov <chertykov@gmail.com>
1486 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
1488 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1490 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
1492 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1494 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
1495 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
1496 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
1498 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
1499 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
1501 Replace usage of ROUND_REG with sh_round_reg.
1502 Use CEIL instead of ROUND_ADVANCE.
1504 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1507 * config/sh/sh.c: Include stdlib headers before everything else.
1509 2014-05-02 Jakub Jelinek <jakub@redhat.com>
1511 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
1512 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
1513 (gimplify_adjust_omp_clauses): Simd region is never
1514 directly nested in combined parallel. Instead, for linear
1515 with copyin/copyout, if in combined for simd loop, make decl
1516 firstprivate/lastprivate on OMP_FOR.
1517 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
1518 expand_omp_for_static_chunk): When setting endvar, also set
1519 fd->loop.v to the same value.
1521 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1523 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
1525 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
1527 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
1530 2014-05-02 Marek Polacek <polacek@redhat.com>
1532 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
1534 2014-05-02 Kito Cheng <kito@0xlab.org>
1536 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
1537 to a C expression marco.
1538 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
1539 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
1540 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
1541 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
1542 HONOR_REG_ALLOC_ORDER.
1543 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
1545 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1547 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
1549 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1551 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
1553 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
1555 * tree-if-conv.c (is_cond_scalar_reduction): New function.
1556 (convert_scalar_cond_reduction): Likewise.
1557 (predicate_scalar_phi): Add recognition and transformation
1558 of simple conditioanl reduction to be vectorizable.
1560 2014-05-01 Marek Polacek <polacek@redhat.com>
1563 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
1565 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
1567 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
1568 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
1569 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
1570 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
1571 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
1572 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
1573 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
1574 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
1576 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
1578 * config/arc/arc.opt (mlra): Move comment above option name
1579 to avoid mis-parsing as language options.
1581 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1583 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
1584 * config/sol2.h: ... here.
1585 * config/sol2-10.h: Remove.
1587 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
1588 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
1589 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
1590 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
1591 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
1592 * config/sol2.h: ... here.
1593 (SECTION_NAME_FORMAT): Don't redefine.
1594 (STARTFILE_ARCH32_SPEC): Rename to ...
1595 (STARTFILE_ARCH_SPEC): ... this.
1596 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
1597 * config/sparc/sol2.h: ... here.
1598 (SECTION_NAME_FORMAT): Don't undef.
1599 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
1600 (SUBTARGET_EXTRA_SPECS): Remove.
1601 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
1603 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
1604 (MD_STARTFILE_PREFIX): Remove.
1605 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
1606 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
1607 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
1608 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
1609 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
1610 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
1611 * config/i386/sol2.h: ... here.
1612 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
1613 * config/i386/sol2-bi.h: Remove.
1614 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
1615 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
1617 * config/i386/t-sol2-64: Rename to ...
1618 * config/i386/t-sol2: ... this.
1619 * config/sparc/t-sol2-64: Rename to ...
1620 * config/sparc/t-sol2: ... this.
1622 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
1623 sol2_tm_file_head, sol2_tm_file_tail.
1624 Include ${cpu_type}/sol2.h before sol2.h.
1626 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
1627 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
1628 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
1629 Reflect i386/t-sol2-64 renaming.
1630 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
1631 Reflect sparc/t-sol2-64 renaming.
1633 2014-04-30 Richard Biener <rguenther@suse.de>
1635 * passes.c (execute_function_todo): Move TODO_verify_stmts
1636 and TODO_verify_ssa under the TODO_verify_il umbrella.
1637 * tree-ssa.h (verify_ssa): Adjust prototype.
1638 * tree-ssa.c (verify_ssa): Add parameter to tell whether
1639 we should verify SSA operands.
1640 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
1641 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
1642 whether we should verify whether not throwing stmts have EH info.
1643 * graphite-scop-detection.c (create_sese_edges): Adjust.
1644 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
1645 * tree-eh.c (lower_try_finally_switch): Do not add the
1646 default case label twice.
1648 2014-04-30 Marek Polacek <polacek@redhat.com>
1650 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
1651 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
1652 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
1653 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
1655 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
1657 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
1658 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
1659 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
1660 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
1661 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
1662 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
1663 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
1664 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
1666 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1668 * tree-cfg.c (dump_function_to_file): Dump the return type of
1669 functions, in a line to itself before the function body, mimicking
1670 the layout of a C function.
1672 2014-04-29 Jakub Jelinek <jakub@redhat.com>
1674 PR tree-optimization/60971
1675 * tree-tailcall.c (process_assignment): Reject conversions which
1678 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
1680 * calls.c (initialize_argument_information): Always treat
1681 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
1682 (expand_call): Likewise.
1683 (emit_library_call_calue_1): Likewise.
1684 * expr.c (PUSH_ARGS_REVERSED): Do not define.
1685 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
1688 2014-04-29 Nick Clifton <nickc@redhat.com>
1690 * config/msp430/msp430.md (umulsidi): Fix typo.
1691 (mulhisi3): Enable even inside interrupt handlers.
1692 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
1693 bigger return address pushed in large mode.
1695 2014-04-29 Nick Clifton <nickc@redhat.com>
1697 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
1698 (arc_init_reg_tables): Use a machine_mode enum to iterate over
1700 * config/m32r/m32r.c (init_reg_tables): Likewise.
1701 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
1702 enum to hold the modes.
1704 2014-04-29 Richard Biener <rguenther@suse.de>
1706 * dominance.c (free_dominance_info): Add overload with
1708 (dom_info_state): Likewise.
1709 (dom_info_available_p): Likewise.
1710 * basic-block.h (free_dominance_info, dom_info_state,
1711 dom_info_available_p): Declare overloads.
1712 * passes.c (execute_function_todo): Verify that verifiers
1713 don't change dominator info state. Drop dominator info
1714 for IPA pass invocations.
1715 * cgraph.c (release_function_body): Restore asserts that
1716 dominator information is released.
1718 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
1720 * doc/invoke.texi: Fix typo.
1721 * tree-vrp.c: Fix typos.
1722 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
1724 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1726 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
1728 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1730 * config/aarch64/aarch64-builtins.c
1731 (aarch64_types_storestruct_lane_qualifiers): New.
1732 (TYPES_STORESTRUCT_LANE): Likewise.
1733 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
1734 (st3_lane): Likewise.
1735 (st4_lane): Likewise.
1736 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
1737 (vec_store_lanesci_lane<mode>): Likewise.
1738 (vec_store_lanesxi_lane<mode>): Likewise.
1739 (aarch64_st2_lane<VQ:mode>): Likewise.
1740 (aarch64_st3_lane<VQ:mode>): Likewise.
1741 (aarch64_st4_lane<VQ:mode>): Likewise.
1742 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
1743 * config/aarch64/arm_neon.h
1744 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
1745 use new macro arguments.
1746 (__ST3_LANE_FUNC): Likewise.
1747 (__ST4_LANE_FUNC): Likewise.
1748 * config/aarch64/iterators.md (V_TWO_ELEM): New.
1749 (V_THREE_ELEM): Likewise.
1750 (V_FOUR_ELEM): Likewise.
1752 2014-04-28 David Malcolm <dmalcolm@redhat.com>
1754 * doc/gimple.texi: Replace the description of the now-defunct
1755 union gimple_statement_d with a diagram showing the
1756 gimple_statement_base class hierarchy and its relationships to
1757 the GSS_ and GIMPLE_ enums.
1759 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1761 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
1762 * config/aarch64/aarch64.c
1763 (aarch64_cannot_change_mode_class): Weaken conditions.
1764 (aarch64_modes_tieable_p): New.
1765 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
1767 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
1769 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
1770 (loadsync_<mode>): Change mode.
1771 (load_quadpti, store_quadpti): New.
1772 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
1773 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
1775 2014-04-28 Martin Jambor <mjambor@suse.cz>
1777 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
1778 same alias type as the original statement.
1779 (subreplacement_assignment_data): New type.
1780 (handle_unscalarized_data_in_subtree): New type of parameter,
1781 generate new memory accesses with same alias type as the original
1783 (load_assign_lhs_subreplacements): Likewise.
1784 (sra_modify_constructor_assign): Generate new memory accesses with
1785 same alias type as the original statement.
1787 2014-04-28 Richard Biener <rguenther@suse.de>
1789 * tree-pass.h (TODO_verify_il): Define.
1790 (TODO_verify_all): Complete properly.
1791 * passes.c (execute_function_todo): Move existing loop-closed
1792 SSA verification under TODO_verify_il.
1793 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
1794 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
1795 Fix tree sharing issue.
1797 2014-04-28 Richard Biener <rguenther@suse.de>
1800 * builtins.def (DEF_C11_BUILTIN): Add.
1801 (BUILT_IN_ALIGNED_ALLOC): Likewise.
1802 * coretypes.h (enum function_class): Add function_c11_misc.
1803 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1804 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
1805 (call_may_clobber_ref_p_1): Likewise.
1806 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
1807 (mark_all_reaching_defs_necessary_1): Likewise.
1808 (propagate_necessity): Likewise.
1809 (eliminate_unnecessary_stmts): Likewise.
1810 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
1812 2014-04-28 Richard Biener <rguenther@suse.de>
1814 * tree-vrp.c (vrp_var_may_overflow): Remove.
1815 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
1816 with overflow immediately bump to one before that value and
1817 let iteration figure out overflow status.
1819 2014-04-28 Richard Biener <rguenther@suse.de>
1821 * configure.ac: Do valgrind header checks unconditionally.
1822 Add --enable-valgrind-annotations.
1823 * system.h: Guard valgrind header inclusion with
1824 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
1825 * alloc-pool.c (pool_alloc, pool_free): Use
1826 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
1827 to guard possibly dead code.
1828 * config.in: Regenerated.
1829 * configure: Likewise.
1831 2014-04-28 Jeff Law <law@redhat.com>
1833 PR tree-optimization/60902
1834 * tree-ssa-threadedge.c
1835 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
1836 over real defs when invalidating outputs from statements that do not
1837 produce useful outputs for threading.
1839 2014-04-28 Richard Biener <rguenther@suse.de>
1841 PR tree-optimization/60979
1842 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
1843 SCOPs that end in a block with a successor with abnormal
1846 2014-04-28 Richard Biener <rguenther@suse.de>
1848 * tree-pass.h (execute_pass_list): Adjust prototype.
1849 * passes.c (pass_manager::execute_early_local_passes): Adjust.
1850 (do_per_function): Change callback signature, push all actual
1851 work to the callbals.
1852 (do_per_function_toporder): Likewise.
1853 (execute_function_dump): Adjust.
1854 (execute_function_todo): Likewise.
1855 (clear_last_verified): Likewise.
1856 (verify_curr_properties): Likewise.
1857 (update_properties_after_pass): Likewise.
1858 (execute_pass_list_1): Split out from ...
1859 (execute_pass_list): ... here. Adjust.
1860 (execute_ipa_pass_list): Likewise.
1861 * cgraphunit.c (cgraph_add_new_function): Adjust.
1862 (analyze_function): Likewise.
1863 (expand_function): Likewise.
1864 * cgraph.c (release_function_body): Free dominance info
1865 here instead of asserting it was magically freed elsewhere.
1867 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
1869 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
1870 * configure: Regenerate.
1871 * config/sparc/sparc.opt (muser-mode): New option.
1872 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
1874 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
1875 * doc/invoke.texi (SPARC options): Document -muser-mode.
1877 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
1879 * cselib.c (find_slot_memmode): Delete.
1880 (cselib_hasher): Change compare_type to a struct.
1881 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
1883 (preserve_constants_and_equivs): Adjust for new compare_type.
1884 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
1885 (wrap_constant): Delete.
1886 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
1888 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
1890 * doc/install.texi (Building with profile feedback): Remove
1893 2014-04-26 Tom de Vries <tom@codesourcery.com>
1895 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
1898 2014-04-25 Cary Coutant <ccoutant@google.com>
1901 * dwarf2out.c (should_move_die_to_comdat): A type definition
1902 can contain a subprogram definition, but don't move it to a
1904 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
1905 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
1907 (clone_tree_hash): Rename to...
1908 (clone_tree_partial): ...this; change callers. Copy
1909 DW_TAG_subprogram DIEs as declarations.
1910 (copy_decls_walk): Don't copy children of a declaration into a
1913 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
1916 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
1919 2014-04-25 Jiong Wang <jiong.wang@arm.com>
1921 * config/arm/predicates.md (call_insn_operand): Add long_call check.
1922 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
1924 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
1927 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1929 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
1931 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1933 PR tree-optimization/60930
1934 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
1935 creating a multiply candidate by folding two constant
1936 multiplicands when the result overflows.
1938 2014-04-25 Jakub Jelinek <jakub@redhat.com>
1940 PR tree-optimization/60960
1941 * tree-vect-generic.c (expand_vector_operation): Only call
1942 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
1944 2014-04-25 Tom de Vries <tom@codesourcery.com>
1946 * expr.c (clobber_reg_mode): New function.
1947 * expr.h (clobber_reg): New function.
1949 2014-04-25 Tom de Vries <tom@codesourcery.com>
1951 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
1954 2014-04-25 Radovan Obradovic <robradovic@mips.com>
1955 Tom de Vries <tom@codesourcery.com>
1957 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
1959 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
1960 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
1961 new argument to find_all_hard_reg_sets call.
1963 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1965 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
1966 Use HOST_WIDE_INT_C for mask literal.
1967 (aarch_rev16_shleft_mask_imm_p): Likewise.
1969 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
1972 * config/sparc/sparc.md (ashlsi3_extend): Delete.
1974 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
1976 PR preprocessor/56540
1977 * config/i386/i386-c.c (ix86_target_macros): Define
1978 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
1980 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1982 * configure.ac (tga_func): Remove.
1983 (LIB_TLS_SPEC): Remove.
1984 * configure: Regenerate.
1985 * config.in: Regenerate.
1986 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
1988 2014-04-25 Richard Biener <rguenther@suse.de>
1991 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
1992 call stmt use/clobber sets during stmt walk instead of
1993 walking the possibly incomplete set of caller edges.
1995 2014-04-25 Richard Biener <rguenther@suse.de>
1998 * passes.c (apply_ipa_transforms): Inline into only caller ...
1999 (execute_one_pass): ... here. Properly bring in function
2000 bodies for nodes we want to apply IPA transforms to.
2002 2014-04-24 Cong Hou <congh@google.com>
2004 PR tree-optimization/60896
2005 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
2006 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
2007 (vect_mark_pattern_stmts): Set the def type of all statements in
2008 PATTERN_DEF_SEQ as vect_internal_def.
2010 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
2012 * doc/extend.texi (PowerPC Built-in Functions): Document new
2013 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
2014 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
2016 * config/rs6000/predicates.md (const_0_to_3_operand): New
2017 predicate to match 0..3 integer constants.
2019 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
2020 to support adding miscellaneous builtin functions.
2021 (BU_DFP_MISC_2): Likewise.
2022 (BU_P7_MISC_1): Likewise.
2023 (BU_P7_MISC_2): Likewise.
2024 (BU_P8V_MISC_3): Likewise.
2025 (BU_MISC_1): Likewise.
2026 (BU_MISC_2): Likewise.
2027 (DIVWE): Add extended divide builtin functions.
2030 (DIVWEUO): Likewise.
2034 (DIVDEUO): Likewise.
2035 (DXEX): Add decimal floating-point builtin functions.
2038 (DDEDPDQ): Likewise.
2040 (DENBCDQ): Likewise.
2047 (CDTBCD): Add new BCD builtin functions.
2051 (BCDADD_LT): Likewise.
2052 (BCDADD_EQ): Likewise.
2053 (BCDADD_GT): Likewise.
2054 (BCDADD_OV): Likewise.
2056 (BCDSUB_LT): Likewise.
2057 (BCDSUB_EQ): Likewise.
2058 (BCDSUB_GT): Likewise.
2059 (BCDSUB_OV): Likewise.
2060 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
2061 (UNPACK_TD): Likewise.
2062 (PACK_TF): Likewise.
2063 (UNPACK_TF): Likewise.
2064 (UNPACK_TF_0): Likewise.
2065 (UNPACK_TF_1): Likewise.
2066 (PACK_V1TI): Likewise.
2067 (UNPACK_V1TI): Likewise.
2069 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
2070 support for decimal floating point builtin functions.
2071 (rs6000_expand_ternop_builtin): Add checks for the new builtin
2072 functions that take constant arguments.
2073 (rs6000_invalid_builtin): Add decimal floating point builtin support.
2074 (rs6000_init_builtins): Setup long double, _Decimal64, and
2075 _Decimal128 types for new builtin functions.
2076 (builtin_function_type): Set the unsigned flags appropriately for
2077 the new builtin functions.
2078 (rs6000_opt_masks): Add support for decimal floating point builtin
2081 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
2082 floating point builtin functions.
2083 (RS6000_BTM_COMMON): Likewise.
2084 (RS6000_BTI_long_double): Likewise.
2085 (RS6000_BTI_dfloat64): Likewise.
2086 (RS6000_BTI_dfloat128): Likewise.
2087 (long_double_type_internal_node): Likewise.
2088 (dfloat64_type_internal_node): Likewise.
2089 (dfloat128_type_internal_node): Likewise.
2091 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
2092 2.07 bcd arithmetic instructions.
2093 (UNSPEC_BCDSUB): Likewise.
2094 (UNSPEC_BCD_OVERFLOW): Likewise.
2095 (UNSPEC_BCD_ADD_SUB): Likewise.
2096 (bcd_add_sub): Likewise.
2097 (BCD_TEST): Likewise.
2098 (bcd<bcd_add_sub>): Likewise.
2099 (bcd<bcd_add_sub>_test): Likewise.
2100 (bcd<bcd_add_sub>_test2): Likewise.
2101 (bcd<bcd_add_sub>_<code>): Likewise.
2102 (peephole2 for combined bcd ops): Likewise.
2104 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
2105 decimal floating point builtin functions.
2106 (UNSPEC_DENBCD): Likewise.
2107 (UNSPEC_DXEX): Likewise.
2108 (UNSPEC_DIEX): Likewise.
2109 (UNSPEC_DSCLI): Likewise.
2110 (UNSPEC_DSCRI): Likewise.
2111 (D64_D128): Likewise.
2112 (dfp_suffix): Likewise.
2113 (dfp_ddedpd_<mode>): Likewise.
2114 (dfp_denbcd_<mode>): Likewise.
2115 (dfp_dxex_<mode>): Likewise.
2116 (dfp_diex_<mode>): Likewise.
2117 (dfp_dscli_<mode>): Likewise.
2118 (dfp_dscri_<mode>): Likewise.
2120 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
2122 (UNSPEC_CDTBCD): Likewise.
2123 (UNSPEC_CBCDTD): Likewise.
2124 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
2125 (UNSPEC_DIVEO): Likewise.
2126 (UNSPEC_DIVEU): Likewise.
2127 (UNSPEC_DIVEUO): Likewise.
2128 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
2129 pack/unpack 128-bit types.
2130 (UNSPEC_PACK_128BIT): Likewise.
2131 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
2132 (udiv<mode>3): Use idiv_ldiv mode attribute.
2133 (div<mode>3): Likewise.
2134 (addg6s): Add new BCD builtin functions.
2137 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
2138 (div_extend): Likewise.
2139 (div<div_extend>_<mode>"): Likewise.
2140 (FP128_64): Add support for new builtin functions to pack/unpack
2142 (unpack<mode>): Likewise.
2143 (unpacktf_0): Likewise.
2144 (unpacktf_1): Likewise.
2145 (unpack<mode>_dm): Likewise.
2146 (unpack<mode>_nodm): Likewise.
2147 (pack<mode>): Likewise.
2148 (unpackv1ti): Likewise.
2149 (packv1ti): Likewise.
2151 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
2153 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
2156 2014-04-24 Jakub Jelinek <jakub@redhat.com>
2158 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
2159 * gimplify.c (omp_is_private): Change last argument's type to int.
2160 Only diagnose lastprivate if the simd argument is 1, only diagnose
2161 linear if the simd argument is 2.
2162 (gimplify_omp_for): Adjust omp_is_private callers. When adding
2163 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
2164 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
2165 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
2166 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
2167 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2168 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
2169 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2170 * tree-nested.c (convert_nonlocal_omp_clauses,
2171 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
2173 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
2176 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
2179 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
2181 * flag-types.h (enum ivar_visibility): Add.
2183 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
2185 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
2186 function * argument.
2188 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
2190 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
2192 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2193 Tom de Vries <tom@codesourcery.com>
2195 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
2196 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
2198 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
2199 * emit-rtl.c (try_split): Same.
2201 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2202 Tom de Vries <tom@codesourcery.com>
2204 * common.opt (fuse-caller-save): New option.
2206 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
2208 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
2209 elements for big-endian.
2211 2014-04-24 Richard Biener <rguenther@suse.de>
2213 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
2214 during TER and instead use the sepops interface for expanding
2215 non-GIMPLE_SINGLE_RHS.
2217 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2219 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
2220 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
2222 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2224 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
2225 assembler 64-bit option.
2226 * configure: Regenerate.
2228 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2230 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
2231 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
2232 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
2233 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
2234 (TARGET_CRYPTO): Take TARGET_SIMD into account.
2236 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2238 * config/aarch64/aarch64-builtins.c
2239 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
2240 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
2241 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
2242 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
2244 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
2245 (Vrevsuff): New mode attribute.
2247 2014-04-24 Terry Guo <terry.guo@arm.com>
2249 * config/arm/arm.h (machine_function): Define variable
2250 after_arm_reorg here.
2251 * config/arm/arm.c (after_arm_reorg): Remove the definition.
2252 (arm_split_constant): Update the way to access variable
2255 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
2257 2014-04-23 Tom de Vries <tom@codesourcery.com>
2259 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
2261 2014-04-23 David Malcolm <dmalcolm@redhat.com>
2263 * is-a.h: Update comments to reflect the following changes to the
2264 "pointerness" of the API, making the template parameter match the
2265 return type, allowing use of is-a.h with typedefs of pointers.
2266 (is_a_helper::cast): Return a T rather then a pointer to a T, so
2267 that the return type matches the parameter to the is_a_helper.
2269 (dyn_cast): Likewise.
2271 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
2272 pointer from the is-a.h API.
2274 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
2275 (is_a_helper <cgraph_node *>::test): ...this, matching change to
2277 (is_a_helper <varpool_node>::test): Likewise, convert to...
2278 (is_a_helper <varpool_node *>::test): ...this.
2280 (varpool_first_variable): Update for removal of implicit pointer
2281 from the is-a.h API.
2282 (varpool_next_variable): Likewise.
2283 (varpool_first_static_initializer): Likewise.
2284 (varpool_next_static_initializer): Likewise.
2285 (varpool_first_defined_variable): Likewise.
2286 (varpool_next_defined_variable): Likewise.
2287 (cgraph_first_defined_function): Likewise.
2288 (cgraph_next_defined_function): Likewise.
2289 (cgraph_first_function): Likewise.
2290 (cgraph_next_function): Likewise.
2291 (cgraph_first_function_with_gimple_body): Likewise.
2292 (cgraph_next_function_with_gimple_body): Likewise.
2293 (cgraph_alias_target): Likewise.
2294 (varpool_alias_target): Likewise.
2295 (cgraph_function_or_thunk_node): Likewise.
2296 (varpool_variable_node): Likewise.
2297 (symtab_real_symbol_p): Likewise.
2298 * cgraphunit.c (referred_to_p): Likewise.
2299 (analyze_functions): Likewise.
2300 (handle_alias_pairs): Likewise.
2301 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2302 * gimple-ssa.h (gimple_vuse_op): Likewise.
2303 (gimple_vdef_op): Likewise.
2304 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
2305 * gimple.c (gimple_build_asm_1): Likewise.
2306 (gimple_build_try): Likewise.
2307 (gimple_build_resx): Likewise.
2308 (gimple_build_eh_dispatch): Likewise.
2309 (gimple_build_omp_for): Likewise.
2310 (gimple_omp_for_set_clauses): Likewise.
2312 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
2313 (is_a_helper <gimple_statement_asm *>::test): ...this.
2314 (is_a_helper <gimple_statement_bind>::test): Convert to...
2315 (is_a_helper <gimple_statement_bind *>::test): ...this.
2316 (is_a_helper <gimple_statement_call>::test): Convert to...
2317 (is_a_helper <gimple_statement_call *>::test): ...this.
2318 (is_a_helper <gimple_statement_catch>::test): Convert to...
2319 (is_a_helper <gimple_statement_catch *>::test): ...this.
2320 (is_a_helper <gimple_statement_resx>::test): Convert to...
2321 (is_a_helper <gimple_statement_resx *>::test): ...this.
2322 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
2323 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
2324 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
2325 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
2326 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
2327 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
2328 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
2329 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
2330 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
2331 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
2332 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
2333 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
2334 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
2335 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
2336 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
2337 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
2338 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
2339 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
2340 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
2341 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
2342 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
2343 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
2344 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
2345 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
2346 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
2347 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
2348 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
2349 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
2350 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
2351 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
2352 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
2353 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
2354 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
2355 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
2356 (is_a_helper <gimple_statement_phi>::test): Convert to...
2357 (is_a_helper <gimple_statement_phi *>::test): ...this.
2358 (is_a_helper <gimple_statement_transaction>::test): Convert to...
2359 (is_a_helper <gimple_statement_transaction *>::test): ...this.
2360 (is_a_helper <gimple_statement_try>::test): Convert to...
2361 (is_a_helper <gimple_statement_try *>::test): ...this.
2362 (is_a_helper <gimple_statement_wce>::test): Convert to...
2363 (is_a_helper <gimple_statement_wce *>::test): ...this.
2364 (is_a_helper <const gimple_statement_asm>::test): Convert to...
2365 (is_a_helper <const gimple_statement_asm *>::test): ...this.
2366 (is_a_helper <const gimple_statement_bind>::test): Convert to...
2367 (is_a_helper <const gimple_statement_bind *>::test): ...this.
2368 (is_a_helper <const gimple_statement_call>::test): Convert to...
2369 (is_a_helper <const gimple_statement_call *>::test): ...this.
2370 (is_a_helper <const gimple_statement_catch>::test): Convert to...
2371 (is_a_helper <const gimple_statement_catch *>::test): ...this.
2372 (is_a_helper <const gimple_statement_resx>::test): Convert to...
2373 (is_a_helper <const gimple_statement_resx *>::test): ...this.
2374 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
2375 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
2376 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
2377 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
2378 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
2380 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
2382 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
2384 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
2386 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
2387 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
2388 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
2390 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
2391 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
2393 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
2394 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
2395 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
2396 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
2397 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
2398 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
2400 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
2401 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
2402 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
2403 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
2405 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
2406 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
2407 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
2408 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
2409 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
2410 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
2411 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
2412 (is_a_helper <const gimple_statement_phi>::test): Convert to...
2413 (is_a_helper <const gimple_statement_phi *>::test): ...this.
2414 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
2415 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
2416 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
2417 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
2418 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
2419 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
2420 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
2422 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
2424 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
2425 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
2427 (gimple_use_ops): Update for removal of implicit pointer from the
2429 (gimple_set_use_ops): Likewise.
2430 (gimple_vuse): Likewise.
2431 (gimple_vdef): Likewise.
2432 (gimple_vuse_ptr): Likewise.
2433 (gimple_vdef_ptr): Likewise.
2434 (gimple_set_vuse): Likewise.
2435 (gimple_set_vdef): Likewise.
2436 (gimple_omp_return_set_lhs): Likewise.
2437 (gimple_omp_return_lhs): Likewise.
2438 (gimple_omp_return_lhs_ptr): Likewise.
2439 (gimple_call_fntype): Likewise.
2440 (gimple_call_set_fntype): Likewise.
2441 (gimple_call_set_internal_fn): Likewise.
2442 (gimple_call_use_set): Likewise.
2443 (gimple_call_clobber_set): Likewise.
2444 (gimple_bind_vars): Likewise.
2445 (gimple_bind_set_vars): Likewise.
2446 (gimple_bind_body_ptr): Likewise.
2447 (gimple_bind_set_body): Likewise.
2448 (gimple_bind_add_stmt): Likewise.
2449 (gimple_bind_block): Likewise.
2450 (gimple_bind_set_block): Likewise.
2451 (gimple_asm_ninputs): Likewise.
2452 (gimple_asm_noutputs): Likewise.
2453 (gimple_asm_nclobbers): Likewise.
2454 (gimple_asm_nlabels): Likewise.
2455 (gimple_asm_input_op): Likewise.
2456 (gimple_asm_input_op_ptr): Likewise.
2457 (gimple_asm_output_op): Likewise.
2458 (gimple_asm_output_op_ptr): Likewise.
2459 (gimple_asm_set_output_op): Likewise.
2460 (gimple_asm_clobber_op): Likewise.
2461 (gimple_asm_set_clobber_op): Likewise.
2462 (gimple_asm_label_op): Likewise.
2463 (gimple_asm_set_label_op): Likewise.
2464 (gimple_asm_string): Likewise.
2465 (gimple_catch_types): Likewise.
2466 (gimple_catch_types_ptr): Likewise.
2467 (gimple_catch_handler_ptr): Likewise.
2468 (gimple_catch_set_types): Likewise.
2469 (gimple_catch_set_handler): Likewise.
2470 (gimple_eh_filter_types): Likewise.
2471 (gimple_eh_filter_types_ptr): Likewise.
2472 (gimple_eh_filter_failure_ptr): Likewise.
2473 (gimple_eh_filter_set_types): Likewise.
2474 (gimple_eh_filter_set_failure): Likewise.
2475 (gimple_eh_must_not_throw_fndecl): Likewise.
2476 (gimple_eh_must_not_throw_set_fndecl): Likewise.
2477 (gimple_eh_else_n_body_ptr): Likewise.
2478 (gimple_eh_else_e_body_ptr): Likewise.
2479 (gimple_eh_else_set_n_body): Likewise.
2480 (gimple_eh_else_set_e_body): Likewise.
2481 (gimple_try_eval_ptr): Likewise.
2482 (gimple_try_cleanup_ptr): Likewise.
2483 (gimple_try_set_eval): Likewise.
2484 (gimple_try_set_cleanup): Likewise.
2485 (gimple_wce_cleanup_ptr): Likewise.
2486 (gimple_wce_set_cleanup): Likewise.
2487 (gimple_phi_capacity): Likewise.
2488 (gimple_phi_num_args): Likewise.
2489 (gimple_phi_result): Likewise.
2490 (gimple_phi_result_ptr): Likewise.
2491 (gimple_phi_set_result): Likewise.
2492 (gimple_phi_arg): Likewise.
2493 (gimple_phi_set_arg): Likewise.
2494 (gimple_resx_region): Likewise.
2495 (gimple_resx_set_region): Likewise.
2496 (gimple_eh_dispatch_region): Likewise.
2497 (gimple_eh_dispatch_set_region): Likewise.
2498 (gimple_omp_critical_name): Likewise.
2499 (gimple_omp_critical_name_ptr): Likewise.
2500 (gimple_omp_critical_set_name): Likewise.
2501 (gimple_omp_for_clauses): Likewise.
2502 (gimple_omp_for_clauses_ptr): Likewise.
2503 (gimple_omp_for_set_clauses): Likewise.
2504 (gimple_omp_for_collapse): Likewise.
2505 (gimple_omp_for_index): Likewise.
2506 (gimple_omp_for_index_ptr): Likewise.
2507 (gimple_omp_for_set_index): Likewise.
2508 (gimple_omp_for_initial): Likewise.
2509 (gimple_omp_for_initial_ptr): Likewise.
2510 (gimple_omp_for_set_initial): Likewise.
2511 (gimple_omp_for_final): Likewise.
2512 (gimple_omp_for_final_ptr): Likewise.
2513 (gimple_omp_for_set_final): Likewise.
2514 (gimple_omp_for_incr): Likewise.
2515 (gimple_omp_for_incr_ptr): Likewise.
2516 (gimple_omp_for_set_incr): Likewise.
2517 (gimple_omp_for_pre_body_ptr): Likewise.
2518 (gimple_omp_for_set_pre_body): Likewise.
2519 (gimple_omp_parallel_clauses): Likewise.
2520 (gimple_omp_parallel_clauses_ptr): Likewise.
2521 (gimple_omp_parallel_set_clauses): Likewise.
2522 (gimple_omp_parallel_child_fn): Likewise.
2523 (gimple_omp_parallel_child_fn_ptr): Likewise.
2524 (gimple_omp_parallel_set_child_fn): Likewise.
2525 (gimple_omp_parallel_data_arg): Likewise.
2526 (gimple_omp_parallel_data_arg_ptr): Likewise.
2527 (gimple_omp_parallel_set_data_arg): Likewise.
2528 (gimple_omp_task_clauses): Likewise.
2529 (gimple_omp_task_clauses_ptr): Likewise.
2530 (gimple_omp_task_set_clauses): Likewise.
2531 (gimple_omp_task_child_fn): Likewise.
2532 (gimple_omp_task_child_fn_ptr): Likewise.
2533 (gimple_omp_task_set_child_fn): Likewise.
2534 (gimple_omp_task_data_arg): Likewise.
2535 (gimple_omp_task_data_arg_ptr): Likewise.
2536 (gimple_omp_task_set_data_arg): Likewise.
2537 (gimple_omp_taskreg_clauses): Likewise.
2538 (gimple_omp_taskreg_clauses_ptr): Likewise.
2539 (gimple_omp_taskreg_set_clauses): Likewise.
2540 (gimple_omp_taskreg_child_fn): Likewise.
2541 (gimple_omp_taskreg_child_fn_ptr): Likewise.
2542 (gimple_omp_taskreg_set_child_fn): Likewise.
2543 (gimple_omp_taskreg_data_arg): Likewise.
2544 (gimple_omp_taskreg_data_arg_ptr): Likewise.
2545 (gimple_omp_taskreg_set_data_arg): Likewise.
2546 (gimple_omp_task_copy_fn): Likewise.
2547 (gimple_omp_task_copy_fn_ptr): Likewise.
2548 (gimple_omp_task_set_copy_fn): Likewise.
2549 (gimple_omp_task_arg_size): Likewise.
2550 (gimple_omp_task_arg_size_ptr): Likewise.
2551 (gimple_omp_task_set_arg_size): Likewise.
2552 (gimple_omp_task_arg_align): Likewise.
2553 (gimple_omp_task_arg_align_ptr): Likewise.
2554 (gimple_omp_task_set_arg_align): Likewise.
2555 (gimple_omp_single_clauses): Likewise.
2556 (gimple_omp_single_clauses_ptr): Likewise.
2557 (gimple_omp_single_set_clauses): Likewise.
2558 (gimple_omp_target_clauses): Likewise.
2559 (gimple_omp_target_clauses_ptr): Likewise.
2560 (gimple_omp_target_set_clauses): Likewise.
2561 (gimple_omp_target_child_fn): Likewise.
2562 (gimple_omp_target_child_fn_ptr): Likewise.
2563 (gimple_omp_target_set_child_fn): Likewise.
2564 (gimple_omp_target_data_arg): Likewise.
2565 (gimple_omp_target_data_arg_ptr): Likewise.
2566 (gimple_omp_target_set_data_arg): Likewise.
2567 (gimple_omp_teams_clauses): Likewise.
2568 (gimple_omp_teams_clauses_ptr): Likewise.
2569 (gimple_omp_teams_set_clauses): Likewise.
2570 (gimple_omp_sections_clauses): Likewise.
2571 (gimple_omp_sections_clauses_ptr): Likewise.
2572 (gimple_omp_sections_set_clauses): Likewise.
2573 (gimple_omp_sections_control): Likewise.
2574 (gimple_omp_sections_control_ptr): Likewise.
2575 (gimple_omp_sections_set_control): Likewise.
2576 (gimple_omp_for_set_cond): Likewise.
2577 (gimple_omp_for_cond): Likewise.
2578 (gimple_omp_atomic_store_set_val): Likewise.
2579 (gimple_omp_atomic_store_val): Likewise.
2580 (gimple_omp_atomic_store_val_ptr): Likewise.
2581 (gimple_omp_atomic_load_set_lhs): Likewise.
2582 (gimple_omp_atomic_load_lhs): Likewise.
2583 (gimple_omp_atomic_load_lhs_ptr): Likewise.
2584 (gimple_omp_atomic_load_set_rhs): Likewise.
2585 (gimple_omp_atomic_load_rhs): Likewise.
2586 (gimple_omp_atomic_load_rhs_ptr): Likewise.
2587 (gimple_omp_continue_control_def): Likewise.
2588 (gimple_omp_continue_control_def_ptr): Likewise.
2589 (gimple_omp_continue_set_control_def): Likewise.
2590 (gimple_omp_continue_control_use): Likewise.
2591 (gimple_omp_continue_control_use_ptr): Likewise.
2592 (gimple_omp_continue_set_control_use): Likewise.
2593 (gimple_transaction_body_ptr): Likewise.
2594 (gimple_transaction_label): Likewise.
2595 (gimple_transaction_label_ptr): Likewise.
2596 (gimple_transaction_set_body): Likewise.
2597 (gimple_transaction_set_label): Likewise.
2599 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
2600 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2601 * ipa-ref.c (ipa_record_reference): Likewise.
2602 * ipa-reference.c (analyze_function): Likewise.
2603 (ipa_reference_write_optimization_summary): Likewise.
2604 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
2605 (address_taken_from_non_vtable_p): Likewise.
2606 (comdat_can_be_unshared_p_1): Likewise.
2607 * lto-cgraph.c (lto_output_ref): Likewise.
2608 (add_references): Likewise.
2609 (compute_ltrans_boundary): Likewise.
2610 (output_symtab): Likewise.
2611 (input_ref): Likewise.
2612 (input_cgraph_1): Likewise.
2613 (output_cgraph_opt_summary): Likewise.
2614 * lto-streamer-out.c (lto_output): Likewise.
2615 (output_symbol_p): Likewise.
2616 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
2617 (lsei_start_function_in_partition): Likewise.
2618 (lsei_next_variable_in_partition): Likewise.
2619 (lsei_start_variable_in_partition): Likewise.
2620 * symtab.c (insert_to_assembler_name_hash): Likewise.
2621 (unlink_from_assembler_name_hash): Likewise.
2622 (symtab_unregister_node): Likewise.
2623 (symtab_remove_node): Likewise.
2624 (dump_symtab_node): Likewise.
2625 (verify_symtab_base): Likewise.
2626 (verify_symtab_node): Likewise.
2627 (symtab_make_decl_local): Likewise.
2628 (symtab_alias_ultimate_target): Likewise.
2629 (symtab_resolve_alias): Likewise.
2630 (symtab_get_symbol_partitioning_class): Likewise.
2631 * tree-phinodes.c (allocate_phi_node): Likewise.
2632 (reserve_phi_args_for_new_edge): Likewise.
2633 (remove_phi_args): Likewise.
2634 * varpool.c (varpool_node_for_asm): Likewise.
2635 (varpool_remove_unreferenced_decls): Likewise.
2637 2014-04-23 Jeff Law <law@redhat.com>
2639 PR tree-optimization/60902
2640 * tree-ssa-threadedge.c
2641 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
2642 invalidate outputs from statements that do not produce useful
2643 outputs for threading.
2645 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2647 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
2648 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
2649 machine descriptions for Stack Smashing Protector.
2651 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
2653 * aarch64.md (<optab>_rol<mode>3): New pattern.
2654 (<optab>_rolsi3_uxtw): Likewise.
2655 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
2657 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
2659 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
2660 (arm_cortex_a12_tune): Likewise.
2662 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2664 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
2666 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2668 * config/arm/arm.md (arm_rev16si2): New pattern.
2669 (arm_rev16si2_alt): Likewise.
2670 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
2672 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2674 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
2675 (rev16<mode>2_alt): Likewise.
2676 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
2677 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
2678 (aarch_rev16_shleft_mask_imm_p): Likewise.
2679 (aarch_rev16_p_1): Likewise.
2680 (aarch_rev16_p): Likewise.
2681 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
2682 (aarch_rev16_shright_mask_imm_p): Likewise.
2683 (aarch_rev16_shleft_mask_imm_p): Likewise.
2685 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2687 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
2688 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
2690 (cortex_a53_extra_costs): Likewise.
2691 (cortex_a57_extra_costs): Likewise.
2692 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
2693 (cortexa7_extra_costs): Likewise.
2694 (cortexa8_extra_costs): Likewise.
2695 (cortexa12_extra_costs): Likewise.
2696 (cortexa15_extra_costs): Likewise.
2697 (v7m_extra_costs): Likewise.
2698 (arm_new_rtx_costs): Handle BSWAP.
2700 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2702 * config/arm/arm.c (cortexa8_extra_costs): New table.
2703 (arm_cortex_a8_tune): New tuning struct.
2704 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
2706 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2708 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
2710 2014-04-23 Richard Biener <rguenther@suse.de>
2712 * Makefile.in (OBJS): Remove loop-unswitch.o.
2713 * tree-pass.h (make_pass_rtl_unswitch): Remove.
2714 * passes.def (pass_rtl_unswitch): Likewise.
2715 * loop-init.c (gate_rtl_unswitch): Likewise.
2716 (rtl_unswitch): Likewise.
2717 (pass_data_rtl_unswitch): Likewise.
2718 (pass_rtl_unswitch): Likewise.
2719 (make_pass_rtl_unswitch): Likewise.
2720 * rtl.h (reversed_condition): Likewise.
2721 (compare_and_jump_seq): Likewise.
2722 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
2724 * loop-unroll.c (compare_and_jump_seq): Likewise.
2726 2014-04-23 Richard Biener <rguenther@suse.de>
2728 PR tree-optimization/60903
2729 * tree-ssa-loop-im.c (analyze_memory_references): Remove
2730 commented code block.
2731 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
2732 loop flags to newly created BBs and edges.
2734 2014-04-23 Nick Clifton <nickc@redhat.com>
2736 * config/msp430/msp430.c (msp430_handle_option): Move function
2738 (msp430_option_override): Simplify mcu and mcpu option handling.
2739 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
2740 support for -mhwmult command line option.
2741 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
2742 -mhwmult command line option.
2743 (msp430_hwmult_enabled): Delete.
2744 (msp43o_output_labelref): Add support for -mhwmult command line option.
2745 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
2746 (umulsidi3): Likewise.
2747 * config/msp430/msp430.opt (mmcu): Add Report attribute.
2748 (mcpu, mlarge, msmall): Likewise.
2749 (mhwmult): New option.
2750 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
2752 (msp430_is_f5_mcu): Remove prototype.
2753 (msp430_use_f5_series_hwmult): Add prototype.
2754 * config/msp430/msp430-opts.h: New file.
2755 * common/config/msp430: New directory.
2756 * common/config/msp430/msp430-common.c: New file.
2757 * config.gcc (msp430): Remove target_has_targetm_common.
2758 * doc/invoke.texi: Document -mhwmult command line option.
2760 2014-04-23 Nick Clifton <nickc@redhat.com>
2762 * config/i386/cygwin.h (ENDFILE_SPEC): Include
2763 default-manifest.o if it can be found in the search path.
2764 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
2766 2014-04-23 Terry Guo <terry.guo@arm.com>
2768 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
2770 2014-04-23 Richard Biener <rguenther@suse.de>
2773 * tree-inline.c (declare_return_variable): Use mark_addressable.
2775 2014-04-23 Richard Biener <rguenther@suse.de>
2778 * loop-init.c (loop_optimizer_init): Make sure to apply
2779 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
2781 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2784 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
2786 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
2787 if flag_sanitize_undefined_trap_on_error.
2788 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
2789 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
2790 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
2791 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
2792 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
2793 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
2794 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
2795 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
2796 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
2797 * ubsan.c (ubsan_instrument_unreachable): Return
2798 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
2799 (ubsan_expand_null_ifn): Emit __builtin_trap ()
2800 if flag_sanitize_undefined_trap_on_error and
2801 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
2802 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
2803 instrument_bool_enum_load): Emit __builtin_trap () if
2804 flag_sanitize_undefined_trap_on_error and
2805 __builtin_handle_*_abort () if !flag_sanitize_recover.
2806 * doc/invoke.texi (-fsanitize-recover,
2807 -fsanitize-undefined-trap-on-error): Document.
2809 2014-04-22 Christian Bruel <christian.bruel@st.com>
2811 * config/sh/sh.md (mov<mode>): Replace movQIHI.
2812 Force immediates to SImode.
2814 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
2816 * config/nios2/nios2.md (UNSPEC_ROUND): New.
2818 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
2819 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
2820 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
2821 (nios2_fpu_insn): Add entry for round.
2822 (N2FPU_NO_ERRNO_P): Define.
2823 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
2825 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
2827 2014-04-22 Richard Henderson <rth@redhat.com>
2829 * config/aarch64/aarch64 (addti3, subti3): New expanders.
2830 (add<GPI>3_compare0): Remove leading * from name.
2831 (add<GPI>3_carryin): Likewise.
2832 (sub<GPI>3_compare0): Likewise.
2833 (sub<GPI>3_carryin): Likewise.
2834 (<su_optab>mulditi3): New expander.
2835 (multi3): New expander.
2836 (madd<GPI>): Remove leading * from name.
2838 2014-04-22 Martin Jambor <mjambor@suse.cz>
2840 * cgraphclones.c (cgraph_function_versioning): Copy
2841 ipa_transforms_to_apply instead of asserting it is empty.
2843 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
2846 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
2847 on count_exp to get mode.
2849 2014-04-22 Andrew Pinski <apinski@cavium.com>
2851 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
2852 Handle TLS for ILP32.
2853 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
2854 (tlsie_small_<mode>): this and handle PTR.
2855 (tlsie_small_sidi): New pattern.
2856 (tlsle_small): Change to an expand to handle ILP32.
2857 (tlsle_small_<mode>): New pattern.
2858 (tlsdesc_small): Rename to ...
2859 (tlsdesc_small_<mode>): this and handle PTR.
2861 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2863 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
2865 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2867 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2868 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
2869 (aarch64_types_signed_poly_qualifiers): Likewise.
2870 (aarch64_types_unsigned_signed_qualifiers): Likewise.
2871 (aarch64_types_poly_signed_qualifiers): Likewise.
2872 (TYPES_REINTERP_SS): Type macro added.
2873 (TYPES_REINTERP_SU): Likewise.
2874 (TYPES_REINTERP_SP): Likewise.
2875 (TYPES_REINTERP_US): Likewise.
2876 (TYPES_REINTERP_PS): Likewise.
2877 (aarch64_fold_builtin): New expression folding added.
2878 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
2879 Declarations removed.
2880 (REINTERP_SS): Declarations added.
2881 (REINTERP_US): Likewise.
2882 (REINTERP_PS): Likewise.
2883 (REINTERP_SU): Likewise.
2884 (REINTERP_SP): Likewise.
2885 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
2886 (vreinterpretq_p8_f64): Likewise.
2887 (vreinterpret_p16_f64): Likewise.
2888 (vreinterpretq_p16_f64): Likewise.
2889 (vreinterpret_f32_f64): Likewise.
2890 (vreinterpretq_f32_f64): Likewise.
2891 (vreinterpret_f64_f32): Likewise.
2892 (vreinterpret_f64_p8): Likewise.
2893 (vreinterpret_f64_p16): Likewise.
2894 (vreinterpret_f64_s8): Likewise.
2895 (vreinterpret_f64_s16): Likewise.
2896 (vreinterpret_f64_s32): Likewise.
2897 (vreinterpret_f64_s64): Likewise.
2898 (vreinterpret_f64_u8): Likewise.
2899 (vreinterpret_f64_u16): Likewise.
2900 (vreinterpret_f64_u32): Likewise.
2901 (vreinterpret_f64_u64): Likewise.
2902 (vreinterpretq_f64_f32): Likewise.
2903 (vreinterpretq_f64_p8): Likewise.
2904 (vreinterpretq_f64_p16): Likewise.
2905 (vreinterpretq_f64_s8): Likewise.
2906 (vreinterpretq_f64_s16): Likewise.
2907 (vreinterpretq_f64_s32): Likewise.
2908 (vreinterpretq_f64_s64): Likewise.
2909 (vreinterpretq_f64_u8): Likewise.
2910 (vreinterpretq_f64_u16): Likewise.
2911 (vreinterpretq_f64_u32): Likewise.
2912 (vreinterpretq_f64_u64): Likewise.
2913 (vreinterpret_s64_f64): Likewise.
2914 (vreinterpretq_s64_f64): Likewise.
2915 (vreinterpret_u64_f64): Likewise.
2916 (vreinterpretq_u64_f64): Likewise.
2917 (vreinterpret_s8_f64): Likewise.
2918 (vreinterpretq_s8_f64): Likewise.
2919 (vreinterpret_s16_f64): Likewise.
2920 (vreinterpretq_s16_f64): Likewise.
2921 (vreinterpret_s32_f64): Likewise.
2922 (vreinterpretq_s32_f64): Likewise.
2923 (vreinterpret_u8_f64): Likewise.
2924 (vreinterpretq_u8_f64): Likewise.
2925 (vreinterpret_u16_f64): Likewise.
2926 (vreinterpretq_u16_f64): Likewise.
2927 (vreinterpret_u32_f64): Likewise.
2928 (vreinterpretq_u32_f64): Likewise.
2930 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2932 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2933 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
2934 (vreinterpret_p8_s8): Likewise.
2935 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
2936 (vreinterpret_p8_s16): Likewise.
2937 (vreinterpret_p8_s32): Likewise.
2938 (vreinterpret_p8_s64): Likewise.
2939 (vreinterpret_p8_f32): Likewise.
2940 (vreinterpret_p8_u8): Likewise.
2941 (vreinterpret_p8_u16): Likewise.
2942 (vreinterpret_p8_u32): Likewise.
2943 (vreinterpret_p8_u64): Likewise.
2944 (vreinterpret_p8_p16): Likewise.
2945 (vreinterpretq_p8_s8): Likewise.
2946 (vreinterpretq_p8_s16): Likewise.
2947 (vreinterpretq_p8_s32): Likewise.
2948 (vreinterpretq_p8_s64): Likewise.
2949 (vreinterpretq_p8_f32): Likewise.
2950 (vreinterpretq_p8_u8): Likewise.
2951 (vreinterpretq_p8_u16): Likewise.
2952 (vreinterpretq_p8_u32): Likewise.
2953 (vreinterpretq_p8_u64): Likewise.
2954 (vreinterpretq_p8_p16): Likewise.
2955 (vreinterpret_p16_s8): Likewise.
2956 (vreinterpret_p16_s16): Likewise.
2957 (vreinterpret_p16_s32): Likewise.
2958 (vreinterpret_p16_s64): Likewise.
2959 (vreinterpret_p16_f32): Likewise.
2960 (vreinterpret_p16_u8): Likewise.
2961 (vreinterpret_p16_u16): Likewise.
2962 (vreinterpret_p16_u32): Likewise.
2963 (vreinterpret_p16_u64): Likewise.
2964 (vreinterpret_p16_p8): Likewise.
2965 (vreinterpretq_p16_s8): Likewise.
2966 (vreinterpretq_p16_s16): Likewise.
2967 (vreinterpretq_p16_s32): Likewise.
2968 (vreinterpretq_p16_s64): Likewise.
2969 (vreinterpretq_p16_f32): Likewise.
2970 (vreinterpretq_p16_u8): Likewise.
2971 (vreinterpretq_p16_u16): Likewise.
2972 (vreinterpretq_p16_u32): Likewise.
2973 (vreinterpretq_p16_u64): Likewise.
2974 (vreinterpretq_p16_p8): Likewise.
2975 (vreinterpret_f32_s8): Likewise.
2976 (vreinterpret_f32_s16): Likewise.
2977 (vreinterpret_f32_s32): Likewise.
2978 (vreinterpret_f32_s64): Likewise.
2979 (vreinterpret_f32_u8): Likewise.
2980 (vreinterpret_f32_u16): Likewise.
2981 (vreinterpret_f32_u32): Likewise.
2982 (vreinterpret_f32_u64): Likewise.
2983 (vreinterpret_f32_p8): Likewise.
2984 (vreinterpret_f32_p16): Likewise.
2985 (vreinterpretq_f32_s8): Likewise.
2986 (vreinterpretq_f32_s16): Likewise.
2987 (vreinterpretq_f32_s32): Likewise.
2988 (vreinterpretq_f32_s64): Likewise.
2989 (vreinterpretq_f32_u8): Likewise.
2990 (vreinterpretq_f32_u16): Likewise.
2991 (vreinterpretq_f32_u32): Likewise.
2992 (vreinterpretq_f32_u64): Likewise.
2993 (vreinterpretq_f32_p8): Likewise.
2994 (vreinterpretq_f32_p16): Likewise.
2995 (vreinterpret_s64_s8): Likewise.
2996 (vreinterpret_s64_s16): Likewise.
2997 (vreinterpret_s64_s32): Likewise.
2998 (vreinterpret_s64_f32): Likewise.
2999 (vreinterpret_s64_u8): Likewise.
3000 (vreinterpret_s64_u16): Likewise.
3001 (vreinterpret_s64_u32): Likewise.
3002 (vreinterpret_s64_u64): Likewise.
3003 (vreinterpret_s64_p8): Likewise.
3004 (vreinterpret_s64_p16): Likewise.
3005 (vreinterpretq_s64_s8): Likewise.
3006 (vreinterpretq_s64_s16): Likewise.
3007 (vreinterpretq_s64_s32): Likewise.
3008 (vreinterpretq_s64_f32): Likewise.
3009 (vreinterpretq_s64_u8): Likewise.
3010 (vreinterpretq_s64_u16): Likewise.
3011 (vreinterpretq_s64_u32): Likewise.
3012 (vreinterpretq_s64_u64): Likewise.
3013 (vreinterpretq_s64_p8): Likewise.
3014 (vreinterpretq_s64_p16): Likewise.
3015 (vreinterpret_u64_s8): Likewise.
3016 (vreinterpret_u64_s16): Likewise.
3017 (vreinterpret_u64_s32): Likewise.
3018 (vreinterpret_u64_s64): Likewise.
3019 (vreinterpret_u64_f32): Likewise.
3020 (vreinterpret_u64_u8): Likewise.
3021 (vreinterpret_u64_u16): Likewise.
3022 (vreinterpret_u64_u32): Likewise.
3023 (vreinterpret_u64_p8): Likewise.
3024 (vreinterpret_u64_p16): Likewise.
3025 (vreinterpretq_u64_s8): Likewise.
3026 (vreinterpretq_u64_s16): Likewise.
3027 (vreinterpretq_u64_s32): Likewise.
3028 (vreinterpretq_u64_s64): Likewise.
3029 (vreinterpretq_u64_f32): Likewise.
3030 (vreinterpretq_u64_u8): Likewise.
3031 (vreinterpretq_u64_u16): Likewise.
3032 (vreinterpretq_u64_u32): Likewise.
3033 (vreinterpretq_u64_p8): Likewise.
3034 (vreinterpretq_u64_p16): Likewise.
3035 (vreinterpret_s8_s16): Likewise.
3036 (vreinterpret_s8_s32): Likewise.
3037 (vreinterpret_s8_s64): Likewise.
3038 (vreinterpret_s8_f32): Likewise.
3039 (vreinterpret_s8_u8): Likewise.
3040 (vreinterpret_s8_u16): Likewise.
3041 (vreinterpret_s8_u32): Likewise.
3042 (vreinterpret_s8_u64): Likewise.
3043 (vreinterpret_s8_p8): Likewise.
3044 (vreinterpret_s8_p16): Likewise.
3045 (vreinterpretq_s8_s16): Likewise.
3046 (vreinterpretq_s8_s32): Likewise.
3047 (vreinterpretq_s8_s64): Likewise.
3048 (vreinterpretq_s8_f32): Likewise.
3049 (vreinterpretq_s8_u8): Likewise.
3050 (vreinterpretq_s8_u16): Likewise.
3051 (vreinterpretq_s8_u32): Likewise.
3052 (vreinterpretq_s8_u64): Likewise.
3053 (vreinterpretq_s8_p8): Likewise.
3054 (vreinterpretq_s8_p16): Likewise.
3055 (vreinterpret_s16_s8): Likewise.
3056 (vreinterpret_s16_s32): Likewise.
3057 (vreinterpret_s16_s64): Likewise.
3058 (vreinterpret_s16_f32): Likewise.
3059 (vreinterpret_s16_u8): Likewise.
3060 (vreinterpret_s16_u16): Likewise.
3061 (vreinterpret_s16_u32): Likewise.
3062 (vreinterpret_s16_u64): Likewise.
3063 (vreinterpret_s16_p8): Likewise.
3064 (vreinterpret_s16_p16): Likewise.
3065 (vreinterpretq_s16_s8): Likewise.
3066 (vreinterpretq_s16_s32): Likewise.
3067 (vreinterpretq_s16_s64): Likewise.
3068 (vreinterpretq_s16_f32): Likewise.
3069 (vreinterpretq_s16_u8): Likewise.
3070 (vreinterpretq_s16_u16): Likewise.
3071 (vreinterpretq_s16_u32): Likewise.
3072 (vreinterpretq_s16_u64): Likewise.
3073 (vreinterpretq_s16_p8): Likewise.
3074 (vreinterpretq_s16_p16): Likewise.
3075 (vreinterpret_s32_s8): Likewise.
3076 (vreinterpret_s32_s16): Likewise.
3077 (vreinterpret_s32_s64): Likewise.
3078 (vreinterpret_s32_f32): Likewise.
3079 (vreinterpret_s32_u8): Likewise.
3080 (vreinterpret_s32_u16): Likewise.
3081 (vreinterpret_s32_u32): Likewise.
3082 (vreinterpret_s32_u64): Likewise.
3083 (vreinterpret_s32_p8): Likewise.
3084 (vreinterpret_s32_p16): Likewise.
3085 (vreinterpretq_s32_s8): Likewise.
3086 (vreinterpretq_s32_s16): Likewise.
3087 (vreinterpretq_s32_s64): Likewise.
3088 (vreinterpretq_s32_f32): Likewise.
3089 (vreinterpretq_s32_u8): Likewise.
3090 (vreinterpretq_s32_u16): Likewise.
3091 (vreinterpretq_s32_u32): Likewise.
3092 (vreinterpretq_s32_u64): Likewise.
3093 (vreinterpretq_s32_p8): Likewise.
3094 (vreinterpretq_s32_p16): Likewise.
3095 (vreinterpret_u8_s8): Likewise.
3096 (vreinterpret_u8_s16): Likewise.
3097 (vreinterpret_u8_s32): Likewise.
3098 (vreinterpret_u8_s64): Likewise.
3099 (vreinterpret_u8_f32): Likewise.
3100 (vreinterpret_u8_u16): Likewise.
3101 (vreinterpret_u8_u32): Likewise.
3102 (vreinterpret_u8_u64): Likewise.
3103 (vreinterpret_u8_p8): Likewise.
3104 (vreinterpret_u8_p16): Likewise.
3105 (vreinterpretq_u8_s8): Likewise.
3106 (vreinterpretq_u8_s16): Likewise.
3107 (vreinterpretq_u8_s32): Likewise.
3108 (vreinterpretq_u8_s64): Likewise.
3109 (vreinterpretq_u8_f32): Likewise.
3110 (vreinterpretq_u8_u16): Likewise.
3111 (vreinterpretq_u8_u32): Likewise.
3112 (vreinterpretq_u8_u64): Likewise.
3113 (vreinterpretq_u8_p8): Likewise.
3114 (vreinterpretq_u8_p16): Likewise.
3115 (vreinterpret_u16_s8): Likewise.
3116 (vreinterpret_u16_s16): Likewise.
3117 (vreinterpret_u16_s32): Likewise.
3118 (vreinterpret_u16_s64): Likewise.
3119 (vreinterpret_u16_f32): Likewise.
3120 (vreinterpret_u16_u8): Likewise.
3121 (vreinterpret_u16_u32): Likewise.
3122 (vreinterpret_u16_u64): Likewise.
3123 (vreinterpret_u16_p8): Likewise.
3124 (vreinterpret_u16_p16): Likewise.
3125 (vreinterpretq_u16_s8): Likewise.
3126 (vreinterpretq_u16_s16): Likewise.
3127 (vreinterpretq_u16_s32): Likewise.
3128 (vreinterpretq_u16_s64): Likewise.
3129 (vreinterpretq_u16_f32): Likewise.
3130 (vreinterpretq_u16_u8): Likewise.
3131 (vreinterpretq_u16_u32): Likewise.
3132 (vreinterpretq_u16_u64): Likewise.
3133 (vreinterpretq_u16_p8): Likewise.
3134 (vreinterpretq_u16_p16): Likewise.
3135 (vreinterpret_u32_s8): Likewise.
3136 (vreinterpret_u32_s16): Likewise.
3137 (vreinterpret_u32_s32): Likewise.
3138 (vreinterpret_u32_s64): Likewise.
3139 (vreinterpret_u32_f32): Likewise.
3140 (vreinterpret_u32_u8): Likewise.
3141 (vreinterpret_u32_u16): Likewise.
3142 (vreinterpret_u32_u64): Likewise.
3143 (vreinterpret_u32_p8): Likewise.
3144 (vreinterpret_u32_p16): Likewise.
3145 (vreinterpretq_u32_s8): Likewise.
3146 (vreinterpretq_u32_s16): Likewise.
3147 (vreinterpretq_u32_s32): Likewise.
3148 (vreinterpretq_u32_s64): Likewise.
3149 (vreinterpretq_u32_f32): Likewise.
3150 (vreinterpretq_u32_u8): Likewise.
3151 (vreinterpretq_u32_u16): Likewise.
3152 (vreinterpretq_u32_u64): Likewise.
3153 (vreinterpretq_u32_p8): Likewise.
3154 (vreinterpretq_u32_p16): Likewise.
3156 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3158 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
3160 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
3162 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
3163 (vqnegd_s64): Likewise.
3164 (vqabs_s64): Likewise.
3165 (vqabsd_s64): Likewise.
3167 2014-04-22 Richard Henderson <rth@redhat.com>
3169 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
3170 computation to the top of the loop.
3172 2014-04-22 Renlin <renlin.li@arm.com>
3173 Jiong Wang <jiong.wang@arm.com>
3175 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
3176 * config/aarch64/aarch64.c (aarch64_layout_frame)
3177 (aarch64_initial_elimination_offset): Likewise.
3179 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
3181 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
3184 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
3186 * machmode.h (bitwise_mode_for_mode): Declare.
3187 * stor-layout.h (bitwise_type_for_mode): Likewise.
3188 * stor-layout.c (bitwise_mode_for_mode): New function.
3189 (bitwise_type_for_mode): Likewise.
3190 * builtins.c (fold_builtin_memory_op): Use it instead of
3191 int_mode_for_mode and build_nonstandard_integer_type.
3193 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3195 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
3196 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
3197 (*-*-solaris2*): Simplify.
3198 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
3199 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
3200 *-*-solaris2.9* handling.
3202 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
3204 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
3205 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
3207 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
3208 * configure: Regenerate.
3210 * config/i386/sol2-9.h: Remove.
3212 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
3213 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
3214 Remove Solaris 9 references.
3216 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
3218 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
3219 (floatuns<GPI:mode><GPF:mode>2): Remove.
3220 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
3221 and floatuns conversions.
3222 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
3223 and floatuns conversions.
3224 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
3225 (w1,w2): New mode attributes for inequal width conversions.
3227 2014-04-22 Renlin Li <Renlin.Li@arm.com>
3229 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
3230 the output asm format.
3232 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
3234 * config/aarch64/aarch64-simd.md
3235 (aarch64_cm<optab>di): Always split.
3236 (*aarch64_cm<optab>di): New.
3237 (aarch64_cmtstdi): Always split.
3238 (*aarch64_cmtstdi): New.
3240 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3242 PR tree-optimization/60823
3243 * omp-low.c (ipa_simd_modify_function_body): Go through
3244 all SSA_NAMEs and for those refering to vector arguments
3245 which are going to be replaced adjust SSA_NAME_VAR and,
3246 if it is a default definition, change it into a non-default
3247 definition assigned at the beginning of function from new_decl.
3248 (ipa_simd_modify_stmt_ops): Rewritten.
3249 * tree-dfa.c (set_ssa_default_def): When removing default def,
3250 check for NULL loc instead of NULL *loc.
3252 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3254 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
3255 restrictions on core registers for DImode values in Thumb2.
3257 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3259 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
3260 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
3262 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3264 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
3265 (*iordi_notzesidi_di): Likewise.
3266 (*iordi_notsesidi_di): Likewise.
3268 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3270 * config/arm/arm-protos.h (tune_params): New struct members.
3271 * config/arm/arm.c: Initialise tune_params per processor.
3272 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
3273 for speed, based on new tune_params.
3275 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3277 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
3278 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
3279 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
3280 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
3281 * config/aarch64/arm_neon.h (vrnd_f64): Added.
3282 (vrnda_f64): Likewise.
3283 (vrndi_f64): Likewise.
3284 (vrndm_f64): Likewise.
3285 (vrndn_f64): Likewise.
3286 (vrndp_f64): Likewise.
3287 (vrndx_f64): Likewise.
3289 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3291 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
3292 GET_MODE_SIZE argument is enum machine_mode.
3294 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3297 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
3298 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
3300 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
3303 * asan.c (asan_emit_stack_protection): Force the base to align to
3304 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
3305 appropriate bits if STRICT_ALIGNMENT.
3306 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
3308 (expand_used_vars): Leave a space in the stack frame for alignment
3309 if STRICT_ALIGNMENT.
3311 2014-04-21 David Malcolm <dmalcolm@redhat.com>
3313 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
3315 (gimple_store_p): Likewise.
3316 (gimple_assign_load_p): Likewise.
3317 (gimple_assign_cast_p): Likewise.
3318 (gimple_clobber_p): Likewise.
3320 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
3321 rather than a gimple.
3322 (gimple_assign_cast_p): Likewise.
3324 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
3327 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
3328 If mode is DDmode and TARGET_E500_DOUBLE allow move.
3330 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
3331 more debug information for E500 if -mdebug=reg.
3333 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
3336 * config/i386/i386.c (ix86_expand_builtin)
3337 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
3338 register for target RTX.
3339 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
3341 2014-04-18 Cong Hou <congh@google.com>
3343 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
3344 the widen-mult pattern by handling two operands with different sizes,
3345 and operands whose size is smaller than half of the result type.
3347 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3349 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
3350 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
3351 (do_estimate_edge_time): Compute it.
3352 * ipa-inline.c (want_inline_small_function_p): Bypass
3353 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
3355 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3357 * ipa-inline.c (spec_rem): New static variable.
3358 (dump_overall_stats): New function.
3359 (dump_inline_stats): New function.
3361 2014-04-18 Richard Henderson <rth@redhat.com>
3363 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
3364 to GET_MODE_SIZE, not a reg_class_t.
3366 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3368 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
3369 (vsx_xxmrglw_<mode>): Likewise.
3371 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3374 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
3375 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
3376 (rs6000_init_hard_regno_mode_ok): Likewise.
3378 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
3380 * ipa-inline.c (inline_small_functions): Account only non-cold
3382 * doc/invoke.texi (inline-unit-growth): Update documentation.
3384 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
3386 * config/rs6000/rs6000.md (addti3, subti3): New.
3388 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
3391 * config/i386/i386.c (ix86_expand_clear): Remove outdated
3392 comment. Check optimize_insn_for_size_p instead of
3393 optimize_insn_for_speed_p.
3395 2014-04-17 Martin Jambor <mjambor@suse.cz>
3397 * gimple-iterator.c (gsi_start_edge): New function.
3398 * gimple-iterator.h (gsi_start_edge): Declare.
3399 * tree-sra.c (single_non_eh_succ): New function.
3400 (disqualify_ops_if_throwing_stmt): Renamed to
3401 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
3402 having one non-EH successor BB.
3403 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
3404 generate loads into replacements.
3405 (sra_modify_assign): Likewise and and also use the simple path for
3407 (sra_modify_function_body): Commit statements on edges.
3409 2014-04-17 Richard Biener <rguenther@suse.de>
3412 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
3413 comparison results and add clarifying comment.
3415 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3417 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
3418 (blank_mode): Initialize it.
3419 (emit_mode_size_inline, emit_mode_nunits_inline,
3420 emit_mode_inner_inline): New functions.
3421 (emit_insn_modes_h): Call them and surround their output with
3422 #if GCC_VERSION >= 4001 ... #endif.
3423 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
3424 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
3425 mode_* arrays if the argument is __builtin_constant_p.
3426 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
3427 is enum machine_mode.
3429 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3431 * passes.c (opt_pass::execute): Adjust.
3432 (pass_manager::execute_pass_mode_switching): Likewise.
3433 (early_local_passes::execute): Likewise.
3434 (execute_one_pass): Pass cfun to the pass's execute method.
3435 * tree-pass.h (opt_pass::execute): Add function * argument.
3436 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3437 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3438 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3439 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3440 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3441 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
3442 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
3443 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3444 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3445 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3446 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3447 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3448 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3449 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3450 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3451 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3452 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3453 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3454 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3455 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3456 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3457 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3458 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3459 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3460 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3461 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3462 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3463 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3464 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3467 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3469 * passes.c (opt_pass::gate): Take function * argument.
3470 (gate_all_early_local_passes): Merge into
3471 (early_local_passes::gate): this.
3472 (gate_all_early_optimizations): Merge into
3473 (all_early_optimizations::gate): this.
3474 (gate_all_optimizations): Mege into
3475 (all_optimizations::gate): this.
3476 (gate_all_optimizations_g): Merge into
3477 (all_optimizations_g::gate): this.
3478 (gate_rest_of_compilation): Mege into
3479 (rest_of_compilation::gate): this.
3480 (gate_postreload): Merge into
3481 (postreload::gate): this.
3482 (dump_one_pass): Pass cfun to the pass's gate method.
3483 (execute_ipa_summary_passes): Likewise.
3484 (execute_one_pass): Likewise.
3485 (ipa_write_summaries_2): Likewise.
3486 (ipa_write_optimization_summaries_1): Likewise.
3487 (ipa_read_summaries_1): Likewise.
3488 (ipa_read_optimization_summaries_1): Likewise.
3489 (execute_ipa_stmt_fixups): Likewise.
3490 * tree-pass.h (opt_pass::gate): Add function * argument.
3491 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
3492 combine-stack-adj.c, combine.c, compare-elim.c,
3493 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3494 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
3495 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
3496 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
3497 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3498 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3499 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3500 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
3501 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
3502 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
3503 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3504 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3505 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
3506 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3507 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3508 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3509 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3510 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3511 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3512 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3513 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3514 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3515 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3516 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
3517 var-tracking.c, vtable-verify.c, web.c: Adjust.
3519 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3521 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
3522 * configure: Regenerate.
3524 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3526 * passes.c (dump_one_pass): don't check pass->has_gate.
3527 (execute_ipa_summary_passes): Likewise.
3528 (execute_one_pass): Likewise.
3529 (ipa_write_summaries_2): Likewise.
3530 (ipa_write_optimization_summaries_1): Likewise.
3531 (ipa_read_optimization_summaries_1): Likewise.
3532 (execute_ipa_stmt_fixups): Likewise.
3533 * tree-pass.h (pass_data::has_gate): Remove.
3534 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3535 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3536 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3537 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3538 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3539 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
3540 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
3541 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
3542 gimple-low.c, gimple-ssa-isolate-paths.c,
3543 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
3544 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3545 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3546 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
3547 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
3548 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
3549 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
3550 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3551 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
3552 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3553 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3554 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3555 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3556 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3557 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3558 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3559 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3560 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3561 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3562 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3563 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3564 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3567 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3569 * pass_manager.h (pass_manager::register_dump_files_1): Remove
3571 * passes.c (pass_manager::register_dump_files_1): Merge into
3572 (pass_manager::register_dump_files): this, and remove its handling of
3573 properties since the pass always has the properties anyway.
3574 (pass_manager::pass_manager): Adjust.
3576 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3578 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
3579 * passes.c (pass_manager::register_dump_files_1): Remove dead code
3580 dealing with properties.
3581 (pass_manager::register_dump_files): Adjust.
3583 2014-03-20 Mark Wielaard <mjw@redhat.com>
3585 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
3586 then represent the bound as normal constant value.
3588 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3591 Forward port from 4.8 branch
3592 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
3594 * config/i386/bmiintrin.h (_blsi_u32): New.
3598 (_blsmsk_u32): Ditto.
3599 (_blsmsk_u64): Ditto.
3600 (_tzcnt_u32): Ditto.
3601 (_tzcnt_u64): Ditto.
3603 2014-04-17 Kito Cheng <kito@0xlab.org>
3605 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
3607 2014-04-17 Richard Biener <rguenther@suse.de>
3610 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
3611 boolean results for comparisons.
3613 2014-04-17 Richard Biener <rguenther@suse.de>
3615 PR tree-optimization/60836
3616 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
3617 initial PHI args to be gimple values.
3619 2014-04-17 Richard Biener <rguenther@suse.de>
3621 PR tree-optimization/60841
3622 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
3623 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
3624 of stmts to SLP build.
3625 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
3626 (vect_analyze_slp): Likewise.
3627 (vect_analyze_slp_instance): Likewise.
3628 (vect_build_slp_tree): Limit overall SLP tree growth.
3629 * tree-vectorizer.h (vect_analyze_data_refs,
3630 vect_analyze_slp): Adjust prototypes.
3632 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3634 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
3637 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3639 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
3640 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
3641 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
3642 for TARGET_SLOW_PSHUFB
3644 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3646 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
3647 * config/i386/i386.c (intel_cost): Ditto.
3649 2014-04-17 Joey Ye <joey.ye@arm.com>
3651 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
3653 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3655 * opts.c (common_handle_option): Disable -fipa-reference coorectly
3658 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3660 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
3661 (type_all_derivations_known_p): New predicate.
3662 (type_all_ctors_visible_p): New predicate.
3663 (type_possibly_instantiated_p): New predicate.
3664 (get_odr_type): Compute all_derivations_known.
3665 (dump_odr_type): Dump the flag.
3666 (maybe_record_type): Cleanup.
3667 (record_target_from_binfo): Add bases_to_consider array;
3668 record bases for types w/o instances and skip CXX destructor.
3669 (possible_polymorphic_call_targets_1): Add bases_to_consider
3670 and consider_construction parameters; check if type may have instance.
3671 (get_polymorphic_call_info): Set maybe_in_construction to true
3672 when we know nothing.
3673 (record_targets_from_bases): Skip CXX destructors; they are
3674 never called for types in construction.
3675 (possible_polymorphic_call_targets): Do not record target when
3676 type may not have instance.
3678 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3681 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
3682 external aliases alive, too.
3684 2014-04-16 Andrew Pinski <apinski@cavium.com>
3686 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
3689 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3691 * final.c (compute_alignments): Do not apply loop alignment to a block
3692 falling through to the exit.
3694 2014-04-16 Catherine Moore <clm@codesourcery.com>
3696 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
3697 Adjust constraints for microMIPS store patterns.
3699 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3701 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
3703 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3705 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
3706 (append_use): Run at -O0.
3707 (append_vdef): Likewise.
3708 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
3709 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
3711 2014-04-16 Jakub Jelinek <jakub@redhat.com>
3713 PR tree-optimization/60844
3714 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
3715 (propagate_op_to_single_use, remove_visited_stmt_chain,
3716 linearize_expr, repropagate_negates, reassociate_bb): Use it
3717 instead of gsi_remove.
3719 2014-04-16 Martin Jambor <mjambor@suse.cz>
3721 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
3722 ipa_transforms_to_apply.
3723 (cgraph_function_versioning): Assert that old_node has empty
3724 ipa_transforms_to_apply.
3725 * trans-mem.c (ipa_tm_create_version): Likewise.
3726 * tree-inline.c (tree_function_versioning): Do not duplicate
3727 ipa_transforms_to_apply.
3729 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3732 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
3734 Pass necessary as flags on 64-bit Solaris/x86.
3735 Use lowercase relocs for x86_64-*-*.
3736 * configure: Regenerate.
3738 2014-04-15 Jan Hubicka <jh@suse.cz>
3740 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
3741 (maybe_record_node, likely_target_p): Use it.
3743 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3746 Revert following patch
3748 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3751 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3752 software floating point or no floating point registers, do not
3753 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3754 in GPRs that occurs after we tested for GPRs that would never be
3757 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3758 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3759 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3760 specifically allow DDmode, since that does not use the SPE SIMD
3763 2014-03-21 Mark Wielaard <mjw@redhat.com>
3765 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
3766 as unsigned or int depending on type and value used.
3768 2014-04-15 Richard Biener <rguenther@suse.de>
3770 PR rtl-optimization/56965
3771 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
3772 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
3774 * alias.c (true_dependence_1): Do not call
3775 nonoverlapping_component_refs_p.
3776 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
3777 nonoverlapping_component_refs_p.
3778 (indirect_refs_may_alias_p): Likewise.
3780 2014-04-15 Teresa Johnson <tejohnson@google.com>
3782 * cfg.c (dump_bb_info): Fix flags check.
3783 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
3785 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3787 PR rtl-optimization/60663
3788 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
3791 2014-04-15 Richard Biener <rguenther@suse.de>
3793 * lto-streamer.h (LTO_major_version): Bump to 4.
3795 2014-04-15 Richard Biener <rguenther@suse.de>
3797 * common.opt (lto_partition_model): New enum.
3798 (flto-partition=): Merge separate options with a single with argument,
3799 add -flto-partition=one support.
3800 * flag-types.h (enum lto_partition_model): Declare.
3801 * opts.c (finish_options): Remove duplicate -flto-partition=
3803 * lto-wrapper.c (run_gcc): Adjust.
3805 2014-04-15 Richard Biener <rguenther@suse.de>
3807 * alias.c (ncr_compar): New function.
3808 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
3810 2014-04-15 Richard Biener <rguenther@suse.de>
3812 * alias.c (record_component_aliases): Do not walk BINFOs.
3814 2014-04-15 Richard Biener <rguenther@suse.de>
3816 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
3817 Add struct function argument and adjust.
3818 (find_func_aliases_for_call): Likewise.
3819 (find_func_aliases): Likewise.
3820 (find_func_clobbers): Likewise.
3821 (intra_create_variable_infos): Likewise.
3822 (compute_points_to_sets): Likewise.
3823 (ipa_pta_execute): Adjust. Do not push/pop cfun.
3825 2014-04-15 Richard Biener <rguenther@suse.de>
3827 * tree.c (iterative_hash_expr): Use enum tree_code_class
3828 to store TREE_CODE_CLASS.
3829 (tree_block): Likewise.
3830 (tree_set_block): Likewise.
3831 * tree.h (fold_build_pointer_plus_loc): Use
3832 convert_to_ptrofftype_loc.
3834 2014-04-15 Jakub Jelinek <jakub@redhat.com>
3837 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
3840 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
3842 * cfgloop.h (struct loop): Move force_vectorize down.
3843 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
3844 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
3845 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
3846 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
3847 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
3848 * tree-core.h (enum annot_expr_kind): Add new kind values.
3849 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
3850 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
3852 * tree.def (ANNOTATE_EXPR): Tweak comment.
3854 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3856 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
3859 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
3861 * tree.h (TYPE_IDENTIFIER): Declare.
3862 * tree.c (subrange_type_for_debug_p): Use it.
3863 * godump.c (go_format_type): Likewise.
3864 * dwarf2out.c (is_cxx_auto, modified_type_die,
3865 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
3866 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
3868 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3871 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
3873 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
3875 * config/i386/i386.c (examine_argument): Return bool. Return true if
3876 parameter should be passed in memory.
3877 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
3878 (construct_container): Update calls to examine_argument.
3879 (function_arg_advance_64): Ditto.
3880 (return_in_memory_32): Merge with ix86_return_in_memory.
3881 (return_in_memory_64): Ditto.
3882 (return_in_memory_ms_64): Ditto.
3884 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3886 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
3887 * coverage.c (coverage_compute_profile_id): Handle externally visible
3890 2014-04-14 Martin Jambor <mjambor@suse.cz>
3892 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
3893 DECL_DISREGARD_INLINE_LIMITS functions.
3895 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3898 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
3900 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3903 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
3904 optimize_insn_for_speed_p instead of
3905 optimize_function_for_speed_p.
3907 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
3909 * doc/invoke.texi (free): Document AArch64.
3911 2014-04-14 Richard Biener <rguenther@suse.de>
3913 PR tree-optimization/60042
3914 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
3915 (insert_into_preds_of_block): Do not prevent PHI insertion
3916 for REFERENCE exprs here ...
3917 (eliminate_dom_walker::before_dom_children): ... but prevent
3918 their use here under similar conditions when applied to the
3919 IL after PRE optimizations.
3921 2014-04-14 Richard Biener <rguenther@suse.de>
3923 * passes.def: Move early points-to after early SRA.
3925 2014-04-14 Richard Biener <rguenther@suse.de>
3927 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
3928 check for which sign-changes we allow when forwarding
3929 a converted value into a switch.
3931 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3933 * stor-layout.c (place_field): Finalize non-constant offset for the
3936 2014-04-14 Richard Biener <rguenther@suse.de>
3938 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
3940 (expand_switch_using_bit_tests_p): Likewise.
3941 (process_switch): Compute and pass on speed_p based on the
3943 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
3944 optimize_bb_for_speed_p.
3946 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
3948 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
3949 * function.h (struct function): Rename has_force_vect_loops into
3950 has_force_vectorize_loops.
3951 * lto-streamer-in.c (input_cfg): Adjust for renaming.
3952 (input_struct_function_base): Likewise.
3953 * lto-streamer-out.c (output_cfg): Likewise.
3954 (output_struct_function_base): Likewise.
3955 * omp-low.c (expand_omp_simd): Likewise.
3956 * tree-cfg.c (move_sese_region_to_fn): Likewise.
3957 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
3958 (version_loop_for_if_conversion): Likewise.
3959 (tree_if_conversion): Likewise.
3960 (main_tree_if_conversion): Likewise.
3961 (gate_tree_if_conversion): Likewise.
3962 * tree-inline.c (copy_loops): Likewise.
3963 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
3964 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
3965 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
3966 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
3967 * tree-vectorizer.c (vectorize_loops): Likewise.
3968 * tree-vectorizer.h (unlimited_cost_model): Likewise.
3970 2014-04-14 Richard Biener <rguenther@suse.de>
3973 * lto-streamer-out.c (wrap_refs): New function.
3974 (lto_output): Wrap symbol references in global initializes in
3975 type-preserving MEM_REFs.
3977 2014-04-14 Christian Bruel <christian.bruel@st.com>
3979 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
3981 2014-04-14 Christian Bruel <christian.bruel@st.com>
3983 * config/sh/sh.md (setmemqi): New expand pattern.
3984 * config/sh/sh.h (CLEAR_RATIO): Define.
3985 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
3986 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
3988 2014-04-14 Richard Biener <rguenther@suse.de>
3991 * fold-const.c (negate_expr_p): Don't negate directional rounding
3993 (fold_negate_expr): Likewise.
3995 2014-04-14 Richard Biener <rguenther@suse.de>
3997 PR tree-optimization/59817
3998 PR tree-optimization/60453
3999 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
4000 recursion to catch all CHRECs in the scalar evolution and restrict
4001 the predicate for the remains appropriately.
4003 2014-04-12 Catherine Moore <clm@codesourcery.com>
4005 * config/mips/constraints.md: Add new register constraint "kb".
4006 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
4007 (*movhi_internal): Likewise.
4008 (*movqi_internal): Likewise.
4009 * config/mips/mips.h (M16_STORE_REGS): New register class.
4010 (REG_CLASS_NAMES): Add M16_STORE_REGS.
4011 (REG_CLASS_CONTENTS): Likewise.
4012 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
4014 2014-04-11 Tobias Burnus <burnus@net-b.de>
4017 * doc/invoke.texi (-Wformat-signedness): Document it.
4018 (Wformat=2): Mention that this enables -Wformat-signedness.
4020 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4022 * common/config/epiphany/epiphany-common.c
4023 (epiphany_option_optimization_table): Enable section anchors by
4024 default at -O1 or higher.
4025 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
4026 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
4027 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
4028 carries no extra cost.
4029 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
4030 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
4031 * config/epiphany/predicates.md (memclob_operand): New predicate.
4032 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
4033 Use memclob_operand predicate and X constraint for operand 3.
4035 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4037 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
4038 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
4041 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4043 PR rtl-optimization/60651
4044 * mode-switching.c (optimize_mode_switching): Make sure to emit
4045 sets of a lower numbered entity before sets of a higher numbered
4046 entity to a mode of the same or lower priority.
4047 When creating a seginfo for a basic block that starts with a code
4048 label, move the insertion point past the code label.
4049 (new_seginfo): Document and enforce requirement that
4050 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
4051 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
4052 * doc/tm.texi: Regenerate.
4054 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
4057 * config/arc/arc.c (arc_save_restore): Fix assert typo.
4059 2013-04-11 Jakub Jelinek <jakub@redhat.com>
4061 * BASE-VER: Set to 4.10.0.
4063 2014-04-11 Tobias Burnus <burnus@net-b.de>
4066 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
4067 * doc/gcc.texi (Service): Update description in the @menu
4068 * doc/invoke.texi (Option Summary): Remove misplaced and
4071 2014-04-11 Steve Ellcey <sellcey@mips.com>
4072 Jakub Jelinek <jakub@redhat.com>
4075 * expr.c (convert_move): Use emit_store_flag_force instead of
4076 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
4079 2014-04-11 Richard Biener <rguenther@suse.de>
4082 * varasm.c (assemble_alias): Avoid endless error reporting
4083 recursion by setting TREE_ASM_WRITTEN.
4085 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4087 * config/s390/s390.md: Add a splitter for NOT rtx.
4089 2014-04-11 Jakub Jelinek <jakub@redhat.com>
4091 PR rtl-optimization/60663
4092 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
4094 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
4095 Jakub Jelinek <jakub@redhat.com>
4098 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
4099 flag from decl_node to node.
4101 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4104 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
4105 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
4106 ameliorating the cases where it can be.
4108 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
4111 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4113 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4114 (loadsync_<mode>): Change mode.
4115 (load_quadpti, store_quadpti): New.
4116 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4117 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4118 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
4120 2014-04-09 Cong Hou <congh@google.com>
4123 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
4126 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4128 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
4129 instead of vnor to exploit possible fusion opportunity in the
4131 (altivec_expand_vec_perm_const_le): Likewise.
4133 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4135 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4136 (loadsync_<mode>): Change mode.
4137 (load_quadpti, store_quadpti): New.
4138 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4139 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4141 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
4144 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
4145 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
4146 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
4148 2014-04-08 Richard Biener <rguenther@suse.de>
4151 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
4152 a 64bit widest int print double-int similar to on HWI64 hosts.
4154 2014-04-08 Richard Biener <rguenther@suse.de>
4156 PR tree-optimization/60785
4157 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
4158 default defs properly.
4160 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
4162 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
4163 (Weffc++): Likewise.
4165 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
4167 * ipa-devirt.c (maybe_record_node): When node is not recorded,
4168 set completep to false rather than true.
4170 2014-04-07 Douglas B Rupp <rupp@adacore.com>
4173 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
4174 ARM_TARGET2_DWARF_FORMAT.
4176 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
4179 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
4180 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
4183 2014-04-07 Richard Biener <rguenther@suse.de>
4185 PR tree-optimization/60766
4186 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
4187 (may_eliminate_iv): Convert cand_value_at result to desired type.
4189 2014-04-07 Jason Merrill <jason@redhat.com>
4192 * common.opt (-fno-gnu-unique): Add.
4193 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
4195 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4197 * haifa-sched.c: Fix outdated function reference and minor
4198 grammar errors in introductory comment.
4200 2014-04-07 Richard Biener <rguenther@suse.de>
4203 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
4205 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
4207 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
4210 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
4211 size accounting for thunks.
4212 (pa_asm_output_mi_thunk): Use final_start_function() and
4213 final_end_function() to output function start and end directives.
4215 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4217 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
4218 device specific ISA/ feature information. Remove short_sp and
4219 errata_skip ds. Add avr_device_specific_features enum to have device
4221 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
4222 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
4223 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
4224 updated device specific info.
4225 * config/avr/avr-mcus.def: Merge device specific details to
4226 dev_attribute field.
4227 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
4229 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
4230 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
4231 assembler if RMW isa supported by current device.
4232 * config/avr/genmultilib.awk: Update as device info structure changed.
4233 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
4235 2014-04-04 Cong Hou <congh@google.com>
4237 PR tree-optimization/60656
4238 * tree-vect-stmts.c (supportable_widening_operation):
4239 Fix a bug that elements in a vector with vect_used_by_reduction
4240 property are incorrectly reordered when the operation on it is not
4241 consistant with the one in reduction operation.
4243 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
4245 PR rtl-optimization/60155
4246 * gcse.c (record_set_data): New function.
4247 (single_set_gcse): New function.
4248 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
4249 (hoist_code): Likewise.
4250 (get_pressure_class_and_nregs): Likewise.
4252 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
4254 * explow.c (probe_stack_range): Emit a final optimization blockage.
4256 2014-04-04 Anthony Green <green@moxielogic.com>
4258 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
4261 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
4264 * lto-cgraph.c (input_overwrite_node): Check that partitioning
4265 flags are set only during streaming.
4266 * ipa.c (process_references, walk_polymorphic_call_targets,
4267 symtab_remove_unreachable_nodes): Drop bodies of always inline
4268 after early inlining.
4269 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
4271 2014-04-04 Jakub Jelinek <jakub@redhat.com>
4272 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4275 * dwarf2out.c (const_ok_for_output_1): Reject expressions
4278 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4281 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
4283 (cortex_a53_fdivd): Likewise.
4285 2014-04-04 Martin Jambor <mjambor@suse.cz>
4288 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
4290 * cgraph.c (clone_of_p): Also return true if thunks match.
4291 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
4292 cgraph_function_or_thunk_node and an obsolete comment.
4293 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
4295 (build_function_decl_skip_args): Likewise.
4296 (set_new_clone_decl_and_node_flags): New function.
4297 (duplicate_thunk_for_node): Likewise.
4298 (redirect_edge_duplicating_thunks): Likewise.
4299 (cgraph_clone_node): New parameter args_to_skip, pass it to
4300 redirect_edge_duplicating_thunks which is called instead of
4301 cgraph_redirect_edge_callee.
4302 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
4303 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
4305 2014-04-04 Jeff Law <law@redhat.com>
4308 * config/arm/predicates.md (const_int_I_operand): New predicate.
4309 (const_int_M_operand): Similarly.
4310 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
4312 (insv_t2, extv_reg, extzv_t2): Likewise.
4313 (load_multiple_with_writeback): Similarly for const_int_I_operand.
4314 (pop_multiple_with_writeback_and_return): Likewise.
4315 (vfp_pop_multiple_with_writeback): Likewise
4317 2014-04-04 Richard Biener <rguenther@suse.de>
4320 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
4321 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
4323 * gimplify.h (gimple_add_tmp_var_fn): Declare.
4324 * gimplify.c (gimple_add_tmp_var_fn): New function.
4325 * gimple-expr.h (create_tmp_reg_fn): Declare.
4326 * gimple-expr.c (create_tmp_reg_fn): New function.
4327 * gimple-low.c (record_vars_into): Don't change cfun.
4328 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
4329 code generation without cfun.
4331 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
4334 * Makefile.in (install-driver): Fix shell scripting.
4336 2014-04-03 Cong Hou <congh@google.com>
4338 PR tree-optimization/60505
4339 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
4340 threshold of number of iterations below which no vectorization
4342 * tree-vect-loop.c (new_loop_vec_info):
4343 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
4344 * tree-vect-loop.c (vect_analyze_loop_operations):
4345 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
4346 * tree-vect-loop.c (vect_transform_loop):
4347 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
4348 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
4349 of iterations of the loop and see if we should build the epilogue.
4351 2014-04-03 Richard Biener <rguenther@suse.de>
4353 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
4354 (streamer_tree_cache_create): Adjust.
4355 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
4356 to allow optional nodes array.
4357 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
4358 (streamer_tree_cache_append): Likewise.
4359 (streamer_tree_cache_create): Create nodes array optionally
4360 as specified by parameter.
4361 * lto-streamer-out.c (create_output_block): Avoid maintaining
4362 the node array in the writer cache.
4363 (DFS_write_tree): Remove assertion.
4364 (produce_asm_for_decls): Free the out decl state hash table early.
4365 * lto-streamer-in.c (lto_data_in_create): Adjust for
4366 streamer_tree_cache_create prototype change.
4368 2014-04-03 Richard Biener <rguenther@suse.de>
4370 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
4371 set TREE_CHAIN to NULL_TREE.
4373 2014-04-03 Richard Biener <rguenther@suse.de>
4375 PR tree-optimization/60740
4376 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
4377 over all GIMPLE_COND operands.
4379 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
4381 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
4382 (Weffc++): Remove Scott's numbering, merge lists and reference
4385 2014-04-03 Nick Clifton <nickc@redhat.com>
4387 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
4390 2014-04-03 Martin Jambor <mjambor@suse.cz>
4392 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
4393 mention gcc_unreachable before failing.
4394 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
4397 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
4400 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
4401 inconsistent code and instead mark the context inconsistent.
4402 (possible_polymorphic_call_targets): For inconsistent contexts
4403 return empty complete list.
4405 2014-04-02 Anthony Green <green@moxielogic.com>
4407 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
4408 (extendqisi2, extendhisi2): Define.
4409 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
4410 (WCHAR_TYPE): Change to unsigned int.
4412 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4414 PR tree-optimization/60733
4415 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
4416 insertion point for PHI candidates to be the end of the feeding
4417 block for the PHI argument.
4419 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
4421 PR rtl-optimization/60650
4422 * lra-constraints.c (process_alt_operands): Decrease reject for
4423 earlyclobber matching.
4425 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4427 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
4429 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4431 * config/spu/spu.c (pad_bb): Do not crash when the last
4432 insn is CODE_FOR_blockage.
4434 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4436 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
4437 lies outside the target mode.
4439 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4442 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4443 software floating point or no floating point registers, do not
4444 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4445 in GPRs that occurs after we tested for GPRs that would never be
4448 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4449 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4450 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4451 specifically allow DDmode, since that does not use the SPE SIMD
4454 2014-04-02 Richard Biener <rguenther@suse.de>
4457 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4458 MODE_INTs. Properly use negv_optab.
4459 (expand_abs): Likewise.
4461 2014-04-02 Richard Biener <rguenther@suse.de>
4464 * Makefile.in (install-driver): Guard extra installs with special
4467 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4469 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4472 2014-04-01 Richard Henderson <rth@redhat.com>
4475 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
4476 alternative enabled before register allocation.
4478 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
4480 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
4481 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
4483 (nios2_large_got_address): Remove unneeded 'sym' parameter.
4484 (nios2_got_address): Update nios2_large_got_address call site.
4485 (nios2_delegitimize_address): New function.
4486 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
4487 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
4488 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
4490 2014-04-01 Martin Husemann <martin@duskware.de>
4492 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
4495 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
4497 PR rtl-optimization/60604
4498 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
4499 check from register_operand.
4500 (register_operand): Redefine in terms of general_operand.
4501 (nonmemory_operand): Use register_operand for the non-constant cases.
4503 2014-04-01 Richard Biener <rguenther@suse.de>
4505 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
4507 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
4509 * doc/invoke.texi (mapp-regs): Clarify.
4511 2014-03-31 Ulrich Drepper <drepper@gmail.com>
4513 * config/i386/avx512fintrin.h (__v32hi): Define type.
4514 (__v64qi): Likewise.
4515 (_mm512_set1_epi8): Define.
4516 (_mm512_set1_epi16): Define.
4517 (_mm512_set4_epi32): Define.
4518 (_mm512_set4_epi64): Define.
4519 (_mm512_set4_pd): Define.
4520 (_mm512_set4_ps): Define.
4521 (_mm512_setr4_epi64): Define.
4522 (_mm512_setr4_epi32): Define.
4523 (_mm512_setr4_pd): Define.
4524 (_mm512_setr4_ps): Define.
4525 (_mm512_setzero_epi32): Define.
4527 2014-03-31 Martin Jambor <mjambor@suse.cz>
4530 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
4531 callsite_arguments_match_p. Updated all callers. Also check types of
4532 corresponding formal parameters and actual arguments.
4533 (not_all_callers_have_enough_arguments_p) Renamed to
4534 some_callers_have_mismatched_arguments_p.
4536 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
4538 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
4540 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
4543 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
4546 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
4548 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
4549 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
4551 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
4552 Use FMAMODE_NOVF512 mode iterator.
4553 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
4554 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
4555 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
4556 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
4558 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
4559 Use VF_128_256 mode iterator.
4560 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
4563 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4565 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
4566 static chain if needed.
4568 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4571 * lra-constraints.c (index_part_to_reg): New.
4572 (process_address): Use it.
4574 2014-03-27 Jeff Law <law@redhat.com>
4575 Jakub Jelinek <jakub@redhat.com>
4578 * expr.c (do_tablejump): Use simplify_gen_binary rather than
4579 gen_rtx_{PLUS,MULT} to build up the address expression.
4581 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
4582 creating non-canonical RTL.
4584 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4587 * ipa-inline.c (want_inline_small_function_p): Short circuit large
4588 functions; reorganize to make cheap checks first.
4589 (inline_small_functions): Do not estimate growth when dumping;
4591 * ipa-inline.h (inline_summary): Add min_size.
4592 (growth_likely_positive): New function.
4593 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
4594 (set_cond_stmt_execution_predicate): Cleanup.
4595 (estimate_edge_size_and_time): Compute min_size.
4596 (estimate_calls_size_and_time): Likewise.
4597 (estimate_node_size_and_time): Likewise.
4598 (inline_update_overall_summary): Update min_size.
4599 (do_estimate_edge_time): Likewise.
4600 (do_estimate_edge_size): Update.
4601 (do_estimate_edge_hints): Update.
4602 (growth_likely_positive): New function.
4604 2014-03-28 Jakub Jelinek <jakub@redhat.com>
4607 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
4608 also if addr has VOIDmode.
4610 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4612 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
4613 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
4615 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
4616 instructions as well as AdvancedSIMD loads.
4618 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4620 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
4621 Use crypto_aese type.
4622 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
4623 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
4624 crypto_aese, crypto_aesmc. Move to types.md.
4625 * config/arm/types.md (crypto_aes): Split into crypto_aese,
4627 * config/arm/iterators.md (crypto_type): Likewise.
4629 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4631 * cgraph.c: Include expr.h and tree-dfa.h.
4632 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
4635 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4638 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
4639 regs from checking multi-reg pseudos.
4641 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4643 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
4645 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4647 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
4648 if it would clobber the stack pointer, even temporarily.
4650 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
4652 * mode-switching.c: Make small adjustments to the top comment.
4654 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4656 * config/rs6000/constraints.md (wD constraint): New constraint to
4657 match the constant integer to get the top DImode/DFmode out of a
4658 vector in a VSX register.
4660 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
4661 match the constant integer to get the top DImode/DFmode out of a
4662 vector in a VSX register.
4664 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
4667 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
4670 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
4671 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
4673 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
4674 Optimize vec_extract of 64-bit values, where the value being
4675 extracted is in the top word, where we can use scalar
4676 instructions. Add direct move and store support. Combine the big
4677 endian/little endian vector select load support into a single insn.
4678 (vsx_extract_<mode>_internal1): Likewise.
4679 (vsx_extract_<mode>_internal2): Likewise.
4680 (vsx_extract_<mode>_load): Likewise.
4681 (vsx_extract_<mode>_store): Likewise.
4682 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
4683 combined into vsx_extract_<mode>_load.
4684 (vsx_extract_<mode>_one_le): Likewise.
4686 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
4687 define the top 64-bit vector element.
4689 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
4692 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4693 Document vec_vbpermq builtin.
4696 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
4697 enable use of xxsldwi and xxpermdi builtin functions.
4698 (vec_xxpermdi): Likewise.
4700 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4701 Document use of vec_xxsldwi and vec_xxpermdi builtins.
4703 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
4705 PR rtl-optimization/60650
4706 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
4708 (find_spills_for): New.
4709 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
4710 Spill all pseudos on the second iteration.
4712 2014-03-27 Marek Polacek <polacek@redhat.com>
4715 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
4718 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4720 * config/s390/s390.c (s390_can_use_return_insn): Check for
4721 call-saved FPRs on 31 bit.
4723 2014-03-27 Jakub Jelinek <jakub@redhat.com>
4726 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
4727 if they need regimplification, just drop them instead of
4728 calling gimple_regimplify_operands on them.
4730 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
4733 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
4734 (aarch64_frame_pointer_required): Adjust logic.
4735 (aarch64_can_eliminate): Adjust logic.
4736 (aarch64_override_options_after_change): Adjust logic.
4738 2014-03-27 Dehao Chen <dehao@google.com>
4740 * ipa-inline.c (early_inliner): Update node's inline info.
4742 2014-03-26 Dehao Chen <dehao@google.com>
4744 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
4745 compiler inserted conditional jumps for NAN float check.
4747 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4749 * ubsan.h (ubsan_create_data): Change second argument's type
4750 to const location_t *.
4751 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
4753 (ubsan_create_data): Change second argument to const location_t *PLOC.
4754 Create Loc field whenever PLOC is non-NULL.
4755 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
4756 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
4760 * real.c (real_to_integer2): Change type of low to UHWI.
4762 2014-03-26 Tobias Burnus <burnus@net-b.de>
4764 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
4765 (CILK_SELF_SPECS): New define.
4766 (driver_self_specs): Use it.
4768 2014-03-26 Richard Biener <rguenther@suse.de>
4770 * tree-pretty-print.c (percent_K_format): Implement special
4771 case for LTO and its stripped down BLOCK tree.
4773 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4776 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
4778 * tree-vrp.c (simplify_internal_call_using_ranges): If only
4779 one range is range_int_cst_p, but not both, at least optimize
4780 addition/subtraction of 0 and multiplication by 0 or 1.
4781 * gimple-fold.c (gimple_fold_call): Fold
4782 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
4783 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
4784 INTEGER_CSTs, try to fold at least x * 0 and y - y.
4786 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
4788 PR rtl-optimization/60452
4789 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
4790 <case REG>: Return 1 for invalid offsets from the frame pointer.
4792 2014-03-26 Marek Polacek <polacek@redhat.com>
4795 * doc/extend.texi (C Extensions): Mention variable-length arrays in
4798 2014-03-26 Marek Polacek <polacek@redhat.com>
4801 * doc/extend.texi (Designated Inits): Describe what happens to omitted
4804 2014-03-26 Marek Polacek <polacek@redhat.com>
4807 * ira-color.c (update_conflict_hard_regno_costs): Perform the
4808 multiplication in unsigned type.
4810 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4812 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
4814 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4816 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
4818 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4821 * cif-code.def (UNREACHABLE) New code.
4822 * ipa-inline.c (inline_small_functions): Skip edges to
4823 __builtlin_unreachable.
4824 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
4825 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
4826 predicate to __bulitin_unreachable.
4827 (set_cond_stmt_execution_predicate): Fix issue when
4828 invert_tree_comparison returns ERROR_MARK.
4829 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
4830 propagate to inline clones.
4831 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
4833 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
4834 * cgraphclones.c (cgraph_clone_node): If call destination is already
4835 ureachable, do not redirect it back.
4836 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
4839 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4841 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
4842 Do not modify inline clones.
4844 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4846 * config/i386/i386.md (general_sext_operand): New mode attr.
4847 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
4848 don't generate (sign_extend (const_int)).
4849 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
4850 operands[2]. Use We constraint instead of <i> and
4851 <general_sext_operand> predicate instead of <general_operand>.
4852 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
4853 * config/i386/constraints.md (We): New constraint.
4854 * config/i386/predicates.md (x86_64_sext_operand,
4855 sext_operand): New predicates.
4857 2014-03-25 Martin Jambor <mjambor@suse.cz>
4860 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
4861 inconsistent devirtualizations to __builtin_unreachable.
4863 2014-03-25 Marek Polacek <polacek@redhat.com>
4866 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
4868 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
4870 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
4871 order of elements for big-endian.
4873 2014-03-25 Richard Biener <rguenther@suse.de>
4876 * gimplify-me.c (gimple_regimplify_operands): Update the
4879 2014-03-25 Martin Jambor <mjambor@suse.cz>
4882 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
4883 (lto_output_varpool_node): Likewise.
4884 (input_overwrite_node): Likewise.
4885 (input_varpool_node): Likewise.
4887 2014-03-25 Richard Biener <rguenther@suse.de>
4889 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
4890 (run_gcc): Likewise.
4892 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4894 * combine.c (simplify_compare_const): Add MODE argument.
4895 Handle mode_width 0 as very large mode_width.
4896 (try_combine, simplify_comparison): Adjust callers.
4898 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
4899 type to avoid signed integer overflow.
4900 * explow.c (plus_constant): Likewise.
4902 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
4904 * doc/generic.texi: Correct typos.
4906 2014-03-24 Tobias Burnus <burnus@net-b.de>
4908 * doc/invoke.texi (-flto): Expand section about
4909 using static libraries with LTO.
4911 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4913 PR rtl-optimization/60501
4914 * optabs.def (addptr3_optab): New optab.
4915 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
4916 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
4917 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
4919 * lra.c (emit_add3_insn): Use the addptr pattern if available.
4921 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
4923 2014-03-24 Ulrich Drepper <drepper@gmail.com>
4925 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
4928 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
4929 (_mm256_undefined_ps): Define.
4930 (_mm256_undefined_pd): Define.
4931 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
4932 (_mm_undefined_pd): Define.
4933 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
4934 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
4935 (_mm512_undefined_ps): Define.
4936 (_mm512_undefined_pd): Define.
4937 Use _mm*_undefined_*.
4938 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
4940 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
4942 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
4943 (lshr_simd): DI mode added.
4944 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
4945 (aarch64_ushr_simddi): Likewise.
4946 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
4947 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
4948 (vshrd_n_u64): Likewise.
4950 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4952 * Makefile.in (s-macro_list): Depend on cc1.
4954 2014-03-23 Teresa Johnson <tejohnson@google.com>
4956 * ipa-utils.c (ipa_print_order): Use specified dump file.
4958 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
4960 PR rtl-optimization/60601
4961 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
4963 * gcc.c (eval_spec_function): Initialize save_growing_value.
4965 2014-03-22 Jakub Jelinek <jakub@redhat.com>
4968 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
4969 code == MINUS_EXPR, never swap op0 with op1.
4971 * toplev.c (init_local_tick): Avoid signed integer multiplication
4973 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
4974 shift by first operand's bitsize.
4976 2014-03-21 Jakub Jelinek <jakub@redhat.com>
4979 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
4981 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
4982 TARGET_ISA_64BIT_P(x).
4984 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4986 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
4987 pattern for vector nor instead of subtract from splat(-1).
4988 (altivec_expand_vec_perm_const_le): Likewise.
4990 2014-03-21 Richard Henderson <rth@twiddle.net>
4993 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
4994 related insns after epilogue_completed.
4996 2014-03-21 Martin Jambor <mjambor@suse.cz>
4999 * cgraph.h (symtab_node): New flag body_removed.
5000 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
5001 when removing bodies.
5002 * symtab.c (dump_symtab_base): Dump body_removed flag.
5003 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
5004 had their bodies removed.
5006 2014-03-21 Martin Jambor <mjambor@suse.cz>
5009 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
5012 2014-03-21 Richard Biener <rguenther@suse.de>
5014 PR tree-optimization/60577
5015 * tree-core.h (struct tree_base): Document nothrow_flag use
5016 in VAR_DECL_NONALIASED.
5017 * tree.h (VAR_DECL_NONALIASED): New.
5018 (may_be_aliased): Adjust.
5019 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
5021 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5023 * expr.c (expand_expr_real_1): Remove outdated comment.
5025 2014-03-20 Jakub Jelinek <jakub@redhat.com>
5028 * ira.c (adjust_cleared_regs): Call copy_rtx on
5029 *reg_equiv[REGNO (loc)].src_p before passing it to
5030 simplify_replace_fn_rtx.
5033 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
5034 into CONST, put pic register as first operand of PLUS. Use
5035 gen_const_mem for both 32-bit and 64-bit PIC got loads.
5037 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5039 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
5041 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5043 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
5044 around for store forwarding issue in the FPU on the UT699.
5045 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
5046 loads and operations if -mfix-ut699 is specified.
5047 (divtf3_hq): Tweak attribute.
5048 (sqrttf2_hq): Likewise.
5050 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5052 * calls.c (store_one_arg): Remove incorrect const qualification on the
5053 type of the temporary.
5054 * cfgexpand.c (expand_return): Likewise.
5055 * expr.c (expand_constructor): Likewise.
5056 (expand_expr_real_1): Likewise.
5058 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5060 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
5063 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
5066 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
5068 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
5070 * config/arm/aarch-common-protos.h
5071 (alu_cost_table): Fix spelling of "extend".
5072 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
5074 2014-03-19 Richard Biener <rguenther@suse.de>
5077 * tree-core.h (tree_type_common): Re-order pointer members
5078 to reduce recursion depth during GC walks.
5080 2014-03-19 Marek Polacek <polacek@redhat.com>
5083 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
5084 before accessing it.
5086 2014-03-19 Richard Biener <rguenther@suse.de>
5089 * lto-streamer-in.c (input_function): In WPA stage do not drop
5092 2014-03-19 Jakub Jelinek <jakub@redhat.com>
5094 PR tree-optimization/60559
5095 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
5096 with build_zero_cst assignment.
5098 2014-03-18 Kai Tietz <ktietz@redhat.com>
5100 PR rtl-optimization/56356
5101 * sdbout.c (sdbout_parms): Verify that parms'
5102 incoming argument is valid.
5103 (sdbout_reg_parms): Likewise.
5105 2014-03-18 Richard Henderson <rth@redhat.com>
5108 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
5109 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
5110 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
5112 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
5114 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
5115 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
5116 Italicize plugin event names in description. Explain that
5117 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
5118 Remind that no GCC functions should be called after PLUGIN_FINISH.
5119 Explain what pragmas with expansion are.
5121 2014-03-18 Martin Liska <mliska@suse.cz>
5123 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
5124 gimple call statement is update.
5125 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
5126 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
5128 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5131 * ubsan.c (ubsan_instrument_unreachable): Call
5132 initialize_sanitizer_builtins.
5133 (ubsan_pass): Likewise.
5136 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
5137 varpool_finalize_decl instead of rest_of_decl_compilation.
5139 2014-03-18 Richard Biener <rguenther@suse.de>
5141 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
5142 by using bitmap_and_compl instead of bitmap_and_compl_into.
5143 (df_rd_transfer_function): Likewise.
5145 2014-03-18 Richard Biener <rguenther@suse.de>
5147 * doc/lto.texi (fresolution): Fix typo.
5149 2014-03-18 Richard Biener <rguenther@suse.de>
5151 * doc/invoke.texi (flto): Update for changes in 4.9.
5153 2014-03-18 Richard Biener <rguenther@suse.de>
5155 * doc/loop.texi: Remove section on the removed lambda framework.
5156 Update loop docs with recent changes in preserving loop structure.
5158 2014-03-18 Richard Biener <rguenther@suse.de>
5160 * doc/lto.texi (-fresolution): Document.
5162 2014-03-18 Richard Biener <rguenther@suse.de>
5164 * doc/contrib.texi: Adjust my name.
5166 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5169 * internal-fn.c: Include diagnostic-core.h.
5170 (expand_BUILTIN_EXPECT): New function.
5171 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
5172 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
5173 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
5174 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
5176 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
5177 Revert 3 argument __builtin_expect code.
5178 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
5179 * gimple-fold.c (gimple_fold_call): Likewise.
5180 * tree.h (fold_builtin_expect): New prototype.
5181 * builtins.c (build_builtin_expect_predicate): Add predictor
5182 argument, if non-NULL, create 3 argument __builtin_expect.
5183 (fold_builtin_expect): No longer static. Add ARG2 argument,
5184 pass it through to build_builtin_expect_predicate.
5185 (fold_builtin_2): Adjust caller.
5186 (fold_builtin_3): Handle BUILT_IN_EXPECT.
5187 * internal-fn.def (BUILTIN_EXPECT): New.
5189 2014-03-18 Tobias Burnus <burnus@net-b.de>
5192 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
5193 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
5194 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
5196 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
5199 * predict.c (combine_predictions_for_bb): Fix up formatting.
5200 (expr_expected_value_1, expr_expected_value): Add predictor argument,
5201 fill what it points to if non-NULL.
5202 (tree_predict_by_opcode): Adjust caller, use the predictor.
5203 * predict.def (PRED_COMPARE_AND_SWAP): Add.
5205 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
5207 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
5208 proper constant for the store mode.
5210 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
5212 * symtab.c (change_decl_assembler_name): Fix transparent alias
5215 2014-03-16 Renlin Li <Renlin.Li@arm.com>
5217 * config/aarch64/aarch64.c: Correct the comments about the
5218 aarch64 stack layout.
5220 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
5222 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
5223 check for GF_OMP_FOR_KIND_FOR.
5225 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
5227 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
5228 ymm and zmm register names.
5230 2014-03-17 Jakub Jelinek <jakub@redhat.com>
5233 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
5234 note creation for the 2010-08-31 changes.
5236 2014-03-17 Marek Polacek <polacek@redhat.com>
5239 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
5240 as -fno-tree-loop-vectorize.
5241 (expand_omp_simd): Likewise.
5243 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
5245 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
5246 (eligible_for_call_delay): New prototype.
5247 * config/sparc/sparc.c (tls_call_delay): Rename into...
5248 (eligible_for_call_delay): ...this. Return false if the instruction
5249 cannot be put in the delay slot of a branch.
5250 (eligible_for_restore_insn): Simplify.
5251 (eligible_for_return_delay): Return false if the instruction cannot be
5252 put in the delay slot of a branch and simplify.
5253 (eligible_for_sibcall_delay): Return false if the instruction cannot be
5254 put in the delay slot of a branch.
5255 * config/sparc/sparc.md (fix_ut699): New attribute.
5256 (tls_call_delay): Delete.
5257 (in_call_delay): Reimplement.
5258 (eligible_for_sibcall_delay): Rename into...
5259 (in_sibcall_delay): ...this.
5260 (eligible_for_return_delay): Rename into...
5261 (in_return_delay): ...this.
5262 (in_branch_delay): Reimplement.
5263 (in_uncond_branch_delay): Delete.
5264 (in_annul_branch_delay): Delete.
5266 2014-03-14 Richard Henderson <rth@redhat.com>
5269 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
5270 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
5271 (*floathi<X87MODEF>2_i387_with_temp): Remove.
5272 (floathi splitters): Remove.
5273 (float<SWI48x>xf2): New pattern.
5274 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
5275 code that tried to handle DImode for 32-bit, but which was excluded
5276 by the pattern's condition. Drop allocation of stack temporary.
5277 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
5278 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
5279 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
5280 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
5281 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
5282 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
5283 (*float<SWI48><MODEF>2_sse_interunit): Remove.
5284 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
5285 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
5286 (*float<SWI48x><X87MODEF>2_i387): Remove.
5287 (all float _with_temp splitters): Remove.
5288 (*float<SWI48x><MODEF>2_i387): New pattern.
5289 (*float<SWI48><MODEF>2_sse): New pattern.
5290 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
5291 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
5293 2014-03-14 Jakub Jelinek <jakub@redhat.com>
5294 Marek Polacek <polacek@redhat.com>
5297 * common.opt (dump_base_name_prefixed): New Variable.
5298 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
5299 if x_dump_base_name_prefixed is already set, set it at the end.
5301 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
5303 PR rtl-optimization/60508
5304 * lra-constraints.c (get_reload_reg): Add new parameter
5306 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
5307 Pass the new parameter values.
5309 2014-03-14 Richard Biener <rguenther@suse.de>
5311 * common.opt: Revert unintented changes from r205065.
5314 2014-03-14 Richard Biener <rguenther@suse.de>
5317 * cfghooks.c (split_block): Properly adjust all loops the
5318 block was a latch of.
5320 2014-03-14 Martin Jambor <mjambor@suse.cz>
5323 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
5326 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
5329 * config/avr/avr.c (avr_set_current_function): Pass function name
5330 through default_strip_name_encoding before sanity checking instead
5331 of skipping the first char of the assembler name.
5333 2014-03-13 Richard Henderson <rth@redhat.com>
5336 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
5337 (ix86_force_to_memory, ix86_free_from_memory): Remove.
5338 * config/i386/i386-protos.h: Likewise.
5339 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
5340 in the expander instead of a splitter.
5341 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
5342 any possibility of requiring a memory.
5343 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
5344 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
5345 (fp branch splitters): Update for ix86_split_fp_branch.
5346 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
5347 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
5348 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
5349 (*fop_<MODEF>_2_i387): Remove f/r alternative.
5350 (*fop_<MODEF>_3_i387): Likewise.
5351 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
5352 (splitters for the fop_* register patterns): Remove.
5353 (fscalexf4_i387): Rename from *fscalexf4_i387.
5354 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
5356 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5358 PR tree-optimization/59779
5359 * tree-dfa.c (get_ref_base_and_extent): Use double_int
5360 type for bitsize and maxsize instead of HOST_WIDE_INT.
5362 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
5364 PR rtl-optimization/57320
5365 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
5366 the CFG after thread_prologue_and_epilogue_insns.
5368 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
5370 PR rtl-optimization/57189
5371 * lra-constraints.c (process_alt_operands): Disfavor spilling
5374 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
5376 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
5378 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5380 PR tree-optimization/59025
5382 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
5383 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
5385 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
5388 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
5389 calls of avr_out_plus_1.
5391 2014-03-13 Bin Cheng <bin.cheng@arm.com>
5393 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
5394 BB's single pred and update the father loop's latch info later.
5396 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5398 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
5402 (VEC_base): Likewise.
5403 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
5404 registers, we need to swap double words in little endian mode.
5406 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
5407 to be a container mode for 128-bit integer operations added in ISA
5408 2.07. Unlike TImode and PTImode, the preferred register set is
5409 the Altivec/VMX registers for the 128-bit operations.
5411 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
5413 (rs6000_split_128bit_ok_p): Likewise.
5415 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
5416 macros for creating ISA 2.07 normal and overloaded builtin
5417 functions with 3 arguments.
5418 (BU_P8V_OVERLOAD_3): Likewise.
5419 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
5420 for use as overloaded functions.
5421 (VPERM_1TI_UNS): Likewise.
5422 (VSEL_1TI): Likewise.
5423 (VSEL_1TI_UNS): Likewise.
5424 (ST_INTERNAL_1ti): Likewise.
5425 (LD_INTERNAL_1ti): Likewise.
5426 (XXSEL_1TI): Likewise.
5427 (XXSEL_1TI_UNS): Likewise.
5428 (VPERM_1TI): Likewise.
5429 (VPERM_1TI_UNS): Likewise.
5430 (XXPERMDI_1TI): Likewise.
5431 (SET_1TI): Likewise.
5432 (LXVD2X_V1TI): Likewise.
5433 (STXVD2X_V1TI): Likewise.
5434 (VEC_INIT_V1TI): Likewise.
5435 (VEC_SET_V1TI): Likewise.
5436 (VEC_EXT_V1TI): Likewise.
5437 (EQV_V1TI): Likewise.
5438 (NAND_V1TI): Likewise.
5439 (ORC_V1TI): Likewise.
5440 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
5441 added in ISA 2.07. Add both normal 'altivec' builtins, and the
5443 (VADDUQM): Likewise.
5444 (VSUBCUQ): Likewise.
5445 (VADDEUQM): Likewise.
5446 (VADDECUQ): Likewise.
5447 (VSUBEUQM): Likewise.
5448 (VSUBECUQ): Likewise.
5450 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5451 __int128_t and __uint128_t types.
5452 (__uint128_type): Likewise.
5453 (altivec_categorize_keyword): Add support for vector __int128_t,
5454 vector __uint128_t, vector __int128, and vector unsigned __int128
5455 as a container type for TImode operations that need to be done in
5456 VSX/Altivec registers.
5457 (rs6000_macro_to_expand): Likewise.
5458 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5459 to support 128-bit integer instructions vaddcuq, vadduqm,
5460 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5461 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5463 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5464 for V1TImode, and set up preferences to use VSX/Altivec registers.
5465 Setup VSX reload handlers.
5466 (rs6000_debug_reg_global): Likewise.
5467 (rs6000_init_hard_regno_mode_ok): Likewise.
5468 (rs6000_preferred_simd_mode): Likewise.
5469 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5470 (easy_altivec_constant): Likewise.
5471 (output_vec_const_move): Likewise.
5472 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5474 (rs6000_expand_vector_extract): Likewise.
5475 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
5477 (rs6000_const_vec): Add support for V1TImode.
5478 (rs6000_emit_le_vsx_load): Swap double words when loading or
5479 storing TImode/V1TImode.
5480 (rs6000_emit_le_vsx_store): Likewise.
5481 (rs6000_emit_le_vsx_move): Likewise.
5482 (rs6000_emit_move): Add support for V1TImode.
5483 (altivec_expand_ld_builtin): Likewise.
5484 (altivec_expand_st_builtin): Likewise.
5485 (altivec_expand_vec_init_builtin): Likewise.
5486 (altivec_expand_builtin): Likewise.
5487 (rs6000_init_builtins): Add support for V1TImode type. Add
5488 support for ISA 2.07 128-bit integer builtins. Define type names
5489 for the VSX/Altivec vector types.
5490 (altivec_init_builtins): Add support for overloaded vector
5491 functions with V1TImode type.
5492 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
5493 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
5495 (rs6000_split_128bit_ok_p): Likewise.
5496 (rs6000_handle_altivec_attribute): Create V1TImode from vector
5497 __int128_t and vector __uint128_t.
5499 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
5500 and mode attributes.
5507 (VS_scalar): Likewise.
5508 (VS_double): Likewise.
5509 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
5511 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
5512 we support the ISA 2.07 128-bit integer arithmetic instructions.
5513 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
5514 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
5515 and TImode types for use with the builtin functions.
5516 (V1TI_type_node): Likewise.
5517 (unsigned_V1TI_type_node): Likewise.
5518 (intTI_type_internal_node): Likewise.
5519 (uintTI_type_internal_node): Likewise.
5521 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
5522 128-bit builtin functions.
5523 (UNSPEC_VADDEUQM): Likewise.
5524 (UNSPEC_VADDECUQ): Likewise.
5525 (UNSPEC_VSUBCUQ): Likewise.
5526 (UNSPEC_VSUBEUQM): Likewise.
5527 (UNSPEC_VSUBECUQ): Likewise.
5528 (VM): Add V1TImode to vector mode iterators.
5530 (VI_unit): Likewise.
5531 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
5532 (altivec_vaddcuq): Likewise.
5533 (altivec_vsubuqm): Likewise.
5534 (altivec_vsubcuq): Likewise.
5535 (altivec_vaddeuqm): Likewise.
5536 (altivec_vaddecuq): Likewise.
5537 (altivec_vsubeuqm): Likewise.
5538 (altivec_vsubecuq): Likewise.
5540 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
5542 (BOOL_128): Likewise.
5543 (BOOL_REGS_OUTPUT): Likewise.
5544 (BOOL_REGS_OP1): Likewise.
5545 (BOOL_REGS_OP2): Likewise.
5546 (BOOL_REGS_UNARY): Likewise.
5547 (BOOL_REGS_AND_CR0): Likewise.
5549 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
5550 128-bit integer builtin support.
5551 (vec_vadduqm): Likewise.
5552 (vec_vaddecuq): Likewise.
5553 (vec_vaddeuqm): Likewise.
5554 (vec_vsubecuq): Likewise.
5555 (vec_vsubeuqm): Likewise.
5556 (vec_vsubcuq): Likewise.
5557 (vec_vsubuqm): Likewise.
5559 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5560 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
5561 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
5562 128-bit integer add/subtract to ISA 2.07.
5564 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
5566 * config/arc/arc.c (arc_predicate_delay_insns):
5567 Fix third argument passed to conditionalize_nonjump.
5569 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
5571 * config/aarch64/aarch64-builtins.c
5572 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
5573 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
5574 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
5575 instead of __builtin_lfloor.
5576 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
5578 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5580 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
5581 (tree_ssa_ifcombine_bb_1): New function.
5582 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
5583 is an empty forwarder block to then_bb or vice versa and then_bb
5584 and else_bb are effectively swapped.
5586 2014-03-12 Christian Bruel <christian.bruel@st.com>
5589 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
5590 REG_CFA_DEF_CFA note.
5591 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
5592 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
5594 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
5596 PR tree-optimization/60454
5597 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
5599 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5601 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
5602 Do not define target_cpu_default2 to generic.
5603 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
5604 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
5605 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
5607 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5608 Marc Glisse <marc.glisse@inria.fr>
5610 PR tree-optimization/60502
5611 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
5612 instead of build_low_bits_mask.
5614 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5617 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
5618 if there are multiple uses, but op doesn't live on E edge.
5619 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
5620 clobber stmts before __builtin_unreachable.
5622 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
5624 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
5625 hard_frame_pointer_rtx.
5626 * cse.c (cse_insn): Remove volatile check.
5627 * cselib.c (cselib_process_insn): Likewise.
5628 * dse.c (scan_insn): Likewise.
5630 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5632 * config/arc/arc.c (conditionalize_nonjump): New function,
5634 (arc_ifcvt): ... this.
5635 (arc_predicate_delay_insns): Use it.
5637 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5639 * config/arc/predicates.md (extend_operand): During/after reload,
5640 allow const_int_operand.
5641 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
5642 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
5643 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
5645 (umulsi3_highpart_i): Likewise.
5647 2014-03-11 Richard Biener <rguenther@suse.de>
5649 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
5650 Add asserts to guard possible wrong-code bugs.
5652 2014-03-11 Richard Biener <rguenther@suse.de>
5654 PR tree-optimization/60429
5655 PR tree-optimization/60485
5656 * tree-ssa-structalias.c (set_union_with_increment): Properly
5657 take into account all fields that overlap the shifted vars.
5658 (do_sd_constraint): Likewise.
5659 (do_ds_constraint): Likewise.
5660 (get_constraint_for_ptr_offset): Likewise.
5662 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
5664 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
5665 (nios2_compute_frame_layout):
5666 Add calculation of cfun->machine->fp_save_offset.
5667 (nios2_expand_prologue): Correct setting of frame pointer register
5669 (nios2_expand_epilogue): Update recovery of stack pointer from
5670 frame pointer accordingly.
5671 (nios2_initial_elimination_offset): Update calculation of offset
5672 for eliminating to HARD_FRAME_POINTER_REGNUM.
5674 2014-03-10 Jakub Jelinek <jakub@redhat.com>
5677 * ipa.c (symtab_remove_unreachable_nodes): Don't call
5678 cgraph_get_create_node on VAR_DECLs.
5680 2014-03-10 Richard Biener <rguenther@suse.de>
5683 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
5685 2014-03-08 Douglas B Rupp <rupp@gnat.com>
5687 * config/vms/vms.opt (vms_float_format): New variable.
5689 2014-03-08 Tobias Burnus <burnus@net-b.de>
5691 * doc/invoke.texi (-fcilkplus): Update implementation status.
5693 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
5694 Richard Biener <rguenther@suse.de>
5696 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
5697 consistently accross all TUs.
5698 (run_gcc): Enable -fshort-double automatically at link at link-time
5699 and disallow override.
5701 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
5704 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
5705 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
5706 if they can't be used.
5708 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5710 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
5711 for Solaris 11/x86 ld.
5712 * configure: Regenerate.
5714 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5716 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
5717 (LIB_TLS_SPEC): Save as ld_tls_libs.
5718 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
5719 (HAVE_AS_IX86_TLSLDM): New test.
5720 * configure, config.in: Regenerate.
5721 * config/i386/i386.c (legitimize_tls_address): Fall back to
5722 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
5723 cannot support TLS_MODEL_LOCAL_DYNAMIC.
5724 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
5725 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
5727 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
5729 * common.opt (fira-loop-pressure): Mark as optimization.
5731 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
5733 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
5734 an OpenMP mappable type.
5736 2014-03-06 Matthias Klose <doko@ubuntu.com>
5738 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
5739 MULTILIB_OSDIRNAMES is not defined.
5741 2014-03-06 Jakub Jelinek <jakub@redhat.com>
5742 Meador Inge <meadori@codesourcery.com>
5745 * config/arm/arm.c (arm_tls_symbol_p): Remove.
5746 (arm_legitimize_address): Call legitimize_tls_address for any
5747 arm_tls_referenced_p expression, handle constant addend. Call it
5748 before testing for !TARGET_ARM.
5749 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
5751 2014-03-06 Richard Biener <rguenther@suse.de>
5757 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5759 * tree-streamer.c (record_common_node): Assert we don't record
5760 nodes with type double.
5761 (preload_common_node): Skip type double, complex double and double
5762 pointer since it is now frontend dependent due to fshort-double option.
5764 2014-03-06 Richard Biener <rguenther@suse.de>
5766 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
5767 or -fno-lto is specified and the linker has full plugin support.
5768 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
5769 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
5770 * lto-wrapper.c (merge_and_complain): Merge compile-time
5771 optimization levels.
5772 (run_gcc): And pass it through to the link options.
5774 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
5778 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5780 * cselib.c (remove_useless_values): Skip to avoid quadratic
5781 behavior if the condition moved from...
5782 (cselib_process_insn): ... here holds.
5784 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5787 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
5788 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
5791 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
5792 (TM_H): Add x86-tune.def.
5794 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5796 * config/aarch64/aarch64.c (generic_tunings):
5797 Use cortexa57_extra_costs.
5799 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5802 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
5803 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
5804 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
5807 2014-03-04 Heiher <r@hev.cc>
5809 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
5810 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
5812 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
5814 * config/i386/predicates.md (const2356_operand): Change to ...
5815 (const2367_operand): ... this.
5816 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
5818 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5819 (*avx512pf_scatterpf<mode>sf): Ditto.
5820 (avx512pf_scatterpf<mode>df): Ditto.
5821 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5822 (*avx512pf_scatterpf<mode>df): Ditto.
5823 * config/i386/i386.c (ix86_expand_builtin): Update
5824 incorrect hint operand error message.
5826 2014-03-04 Richard Biener <rguenther@suse.de>
5828 * lto-section-in.c (lto_get_section_data): Fix const cast.
5830 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5832 * tree-streamer.c (record_common_node): Assert we don't record
5833 nodes with type double.
5834 (preload_common_node): Skip type double, complex double and double
5835 pointer since it is now frontend dependent due to fshort-double option.
5837 2014-03-04 Richard Biener <rguenther@suse.de>
5840 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
5841 (lto_input_toplevel_asms): Likewise.
5842 * lto-section-in.c (lto_get_section_data): Instead do it here
5845 2014-03-04 Richard Biener <rguenther@suse.de>
5847 PR tree-optimization/60382
5848 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
5849 dead PHIs a reduction.
5851 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
5853 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
5855 (_mm_prefetch): Move out of GCC target("sse") pragma.
5856 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
5857 GCC target("prfchw") pragma.
5858 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
5860 * config/i386/i386.c (ix86_option_override_internal): Enable
5861 -mprfchw with -mprefetchwt1.
5863 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5865 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
5868 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5870 * opts.h (CL_PCH_IGNORE): Define.
5871 * targhooks.c (option_affects_pch_p):
5872 Return false for options that have CL_PCH_IGNORE set.
5873 * opt-functions.awk: Process PchIgnore.
5874 * doc/options.texi: Document PchIgnore.
5876 * config/arc/arc.opt (misize): Add PchIgnore property.
5878 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5880 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
5881 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
5882 constraint on constants to permit them being loaded into
5883 GENERAL_REGS or BASE_REGS.
5885 2014-03-03 Nick Clifton <nickc@redhat.com>
5887 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
5888 anti-cacnonical alternatives.
5889 (negandhi3_real): New pattern.
5890 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
5892 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5894 * config/avr/avr-mcus.def: Remove atxmega16x1.
5895 * config/avr/avr-tables.opt: Regenerate.
5896 * config/avr/t-multilib: Regenerate.
5897 * doc/avr-mmcu.texi: Regenerate.
5899 2014-03-03 Tobias Grosser <tobias@grosser.es>
5900 Mircea Namolaru <mircea.namolaru@inria.fr>
5902 PR tree-optimization/58028
5903 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
5906 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5908 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
5909 not handled by recognizers.
5911 2014-03-03 Jakub Jelinek <jakub@redhat.com>
5914 * function.c (expand_function_end): Don't emit
5915 clobber_return_register sequence if clobber_after is a BARRIER.
5916 * cfgexpand.c (construct_exit_block): Append instructions before
5917 return_label to prev_bb.
5919 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5921 * config/rs6000/constraints.md: Document reserved use of "wc".
5923 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5926 * ipa.c (function_and_variable_visibility): When dissolving comdat
5927 group, also set all symbols to local.
5929 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
5934 2013-12-14 Jan Hubicka <jh@suse.cz>
5936 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
5938 2014-03-02 Jon Beniston <jon@beniston.com>
5944 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
5945 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
5946 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
5947 (simple_return, *simple_return): New patterns
5948 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
5949 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
5951 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
5953 * dwarf2out.c (gen_subprogram_die): Tidy.
5955 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
5958 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
5959 (*mov_t_msb_neg_negc): ... this new insn.
5961 2014-02-28 Jason Merrill <jason@redhat.com>
5964 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
5967 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
5970 * dwarf2out.c (decltype_auto_die): New static.
5971 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
5972 (gen_type_die_with_usage): Handle 'decltype(auto)'.
5973 (is_cxx_auto): Likewise.
5975 2014-02-28 Ian Bolton <ian.bolton@arm.com>
5977 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
5978 we are not using general regs only.
5980 2014-02-28 Richard Biener <rguenther@suse.de>
5983 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
5984 previous fix and only allow to remove trivial pre-headers
5985 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
5986 (remove_forwarder_block): Properly update the latch of a loop.
5988 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5991 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
5992 (cselib_preserved_hash_table): New.
5993 (preserve_constants_and_equivs): Move preserved vals to it.
5994 (cselib_find_slot): Look it up first.
5995 (cselib_init): Initialize it.
5996 (cselib_finish): Release it.
5997 (dump_cselib_table): Dump it.
5999 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6002 * cselib.c (remove_useless_values): Skip to avoid quadratic
6003 behavior if the condition moved from...
6004 (cselib_process_insn): ... here holds.
6006 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6009 * var-tracking.c (vt_initialize): Apply the same condition to
6010 preserve the CFA base value.
6012 2014-02-28 Joey Ye <joey.ye@arm.com>
6015 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
6016 if reload in progress or completed.
6018 2014-02-28 Tobias Burnus <burnus@net-b.de>
6021 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
6024 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
6026 * doc/tm.texi.in (Condition Code Status): Update documention for
6027 relative locations of cc0-setter and cc0-user.
6029 2014-02-27 Jeff Law <law@redhat.com>
6031 PR rtl-optimization/52714
6032 * combine.c (try_combine): When splitting an unrecognized PARALLEL
6033 into two independent simple sets, if I3 is a jump, ensure the
6034 pattern we place into I3 is a (set (pc) ...).
6036 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
6037 Jeff Law <law@redhat.com>
6039 PR rtl-optimization/49847
6040 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
6041 are in different blocks.
6042 * doc/tm.texi (Condition Code Status): Update documention for
6043 relative locations of cc0-setter and cc0-user.
6045 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
6048 * lra.c (lra_emit_add): Check SUBREG too.
6050 2014-02-27 Andreas Schwab <schwab@suse.de>
6052 * config/m68k/m68k.c (m68k_option_override): Disable
6053 -flive-range-shrinkage for classic m68k.
6054 (m68k_override_options_after_change): Likewise.
6056 2014-02-27 Marek Polacek <polacek@redhat.com>
6059 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
6060 -Wmaybe-uninitialized.
6062 2014-02-27 Alan Modra <amodra@gmail.com>
6065 * reload1.c (emit_input_reload_insns): When reload_override_in,
6066 set old to rl->in_reg when rl->in_reg is a subreg.
6068 2014-02-26 Richard Biener <rguenther@suse.de>
6071 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
6073 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6075 * common/config/i386/predicates.md (const1256_operand): Remove.
6076 (const2356_operand): New.
6077 (const_1_to_2_operand): Remove.
6078 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6079 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6080 (*avx512pf_gatherpf<mode>sf): Ditto.
6081 (avx512pf_gatherpf<mode>df): Ditto.
6082 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6083 (*avx512pf_gatherpf<mode>df): Ditto.
6084 (avx512pf_scatterpf<mode>sf): Ditto.
6085 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6086 (*avx512pf_scatterpf<mode>sf): Ditto.
6087 (avx512pf_scatterpf<mode>df): Ditto.
6088 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6089 (*avx512pf_scatterpf<mode>df): Ditto.
6090 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
6092 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
6094 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
6095 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
6096 (_mm512_mask_testn_epi64_mask): Move to ...
6097 * config/i386/avx512cdintrin.h: Here.
6098 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
6099 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
6100 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
6101 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
6102 TARGET_AVX512F from TARGET_AVX512CD.
6104 2014-02-26 Richard Biener <rguenther@suse.de>
6107 * ipa.c (walk_polymorphic_call_targets): Properly guard
6108 call to inline_update_overall_summary.
6110 2014-02-26 Bin Cheng <bin.cheng@arm.com>
6113 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
6114 and latches only if requested. Fix latch if it is removed.
6115 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
6116 LOOPS_HAVE_PREHEADERS.
6118 2014-02-25 Andrew Pinski <apinski@cavium.com>
6120 * builtins.c (expand_builtin_thread_pointer): Create a new target
6121 when the target is NULL.
6123 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
6125 PR rtl-optimization/60317
6126 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6127 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6128 * lra-assigns.c: Include params.h.
6129 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
6130 other reload pseudos considerations.
6132 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6134 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
6135 to use canonical form for nor<mode>3.
6137 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6140 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
6143 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6145 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
6146 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
6147 (ix86_handle_option): Handle OPT_mprefetchwt1.
6148 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
6149 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
6151 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
6152 OPTION_MASK_ISA_PREFETCHWT1.
6153 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
6154 (PTA_PREFETCHWT1): New.
6155 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
6156 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
6157 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
6158 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
6159 (*prefetch_avx512pf_<mode>_: Change into ...
6160 (*prefetch_prefetchwt1_<mode>: This.
6161 * config/i386/i386.opt (mprefetchwt1): New.
6162 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
6163 (_mm_prefetch): Handle intent to write.
6164 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
6166 2014-02-25 Richard Biener <rguenther@suse.de>
6169 * emit-rtl.c (mem_attrs_htab): Remove.
6170 (mem_attrs_htab_hash): Likewise.
6171 (mem_attrs_htab_eq): Likewise.
6172 (set_mem_attrs): Always allocate new mem-attrs when something changed.
6173 (init_emit_once): Do not allocate mem_attrs_htab.
6175 2014-02-25 Richard Biener <rguenther@suse.de>
6178 * lto-opts.c (lto_write_options): Output non-explicit conservative
6179 -fwrapv, -fno-trapv and -fno-strict-overflow.
6180 * lto-wrapper.c (merge_and_complain): Handle merging those options.
6181 (run_gcc): And pass them through.
6183 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6185 * sel-sched.c (calculate_new_fences): New parameter ptime.
6186 Calculate it as a maximum over all fence cycles.
6187 (sel_sched_region_2): Adjust the call to calculate_new_fences.
6188 Print the final schedule timing when sched_verbose.
6190 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6192 PR rtl-optimization/60292
6193 * sel-sched.c (fill_vec_av_set): Do not reset target availability
6194 bit fot the fence instruction.
6196 2014-02-24 Alangi Derick <alangiderick@gmail.com>
6198 * calls.h: Fix typo in comment.
6200 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
6202 * config/pa/pa.c (pa_output_move_double): Don't valididate when
6203 adjusting offsetable addresses.
6205 2014-02-24 Guozhi Wei <carrot@google.com>
6207 * sparseset.h (sparseset_pop): Fix the wrong index.
6209 2014-02-24 Walter Lee <walt@tilera.com>
6211 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
6212 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
6214 * common/config/tilegx/tilegx-common.c
6215 (TARGET_DEFAULT_TARGET_FLAGS): Define.
6216 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
6218 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
6219 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
6220 (tilegx_gimplify_va_arg_expr): Handle big endian.
6221 (tilegx_expand_unaligned_load): Ditto.
6222 (tilegx_expand_unaligned_store): Ditto.
6223 (TARGET_RETURN_IN_MSB): New.
6224 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
6225 (TARGET_ENDIAN_DEFAULT): New.
6226 (TARGET_BIG_ENDIAN): Handle big endian.
6227 (BYTES_BIG_ENDIAN): Ditto.
6228 (WORDS_BIG_ENDIAN): Ditto.
6229 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
6232 * config/tilegx/tilegx.md (extv): Handle big endian.
6234 (insn_st<n>): Ditto.
6235 (insn_st<n>_add<bitsuffix>): Ditto.
6236 (insn_stnt<n>): Ditto.
6237 (insn_stnt<n>_add<bitsuffix>):Ditto.
6238 (vec_interleave_highv8qi): Handle big endian.
6239 (vec_interleave_highv8qi_be): New.
6240 (vec_interleave_highv8qi_le): New.
6241 (insn_v1int_h): Handle big endian.
6242 (vec_interleave_lowv8qi): Handle big endian.
6243 (vec_interleave_lowv8qi_be): New.
6244 (vec_interleave_lowv8qi_le): New.
6245 (insn_v1int_l): Handle big endian.
6246 (vec_interleave_highv4hi): Handle big endian.
6247 (vec_interleave_highv4hi_be): New.
6248 (vec_interleave_highv4hi_le): New.
6249 (insn_v2int_h): Handle big endian.
6250 (vec_interleave_lowv4hi): Handle big endian.
6251 (vec_interleave_lowv4hi_be): New.
6252 (vec_interleave_lowv4hi_le): New.
6253 (insn_v2int_l): Handle big endian.
6254 (vec_interleave_highv2si): Handle big endian.
6255 (vec_interleave_highv2si_be): New.
6256 (vec_interleave_highv2si_le): New.
6257 (insn_v4int_h): Handle big endian.
6258 (vec_interleave_lowv2si): Handle big endian.
6259 (vec_interleave_lowv2si_be): New.
6260 (vec_interleave_lowv2si_le): New.
6261 (insn_v4int_l): Handle big endian.
6262 * config/tilegx/tilegx.opt (mbig-endian): New option.
6263 (mlittle-endian): New option.
6264 * doc/install.texi: Document tilegxbe-linux.
6265 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
6267 2014-02-24 Martin Jambor <mjambor@suse.cz>
6270 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
6271 there are no parameter descriptors.
6273 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
6275 PR rtl-optimization/60268
6276 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
6277 initialization to ...
6278 (sched_rgn_init): ... here.
6279 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
6281 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6283 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
6286 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
6288 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
6291 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6293 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
6294 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
6296 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6298 * config/microblaze/predicates.md: Add cmp_op predicate.
6299 * config/microblaze/microblaze.md: Add branch_compare instruction
6300 which uses cmp_op predicate and emits cmp insn before branch.
6301 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
6302 to microblaze_expand_conditional_branch and consolidate logic.
6303 (microblaze_expand_conditional_branch): emit branch_compare
6304 insn instead of handling cmp op separate from branch insn.
6306 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6308 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
6311 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6313 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
6314 define_insn with define_expand and new define_insn
6315 *altivec_lve<VI_char>x_internal.
6316 (altivec_stve<VI_char>x): Replace define_insn with define_expand
6317 and new define_insn *altivec_stve<VI_char>x_internal.
6318 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
6320 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
6322 (altivec_expand_stvex_be): New function.
6324 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
6326 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
6327 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
6328 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
6329 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
6331 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
6334 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
6335 instead of emit_move_insn.
6337 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6339 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
6341 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
6342 gen_altivec_vsumsws.
6344 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6346 * config/rs6000/altivec.md (altivec_lvxl): Rename as
6347 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
6348 (altivec_lvxl_<mode>): New define_expand incorporating
6349 -maltivec=be semantics where needed.
6350 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
6351 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
6352 semantics where needed.
6353 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
6354 (altivec_stvx_<mode>): New define_expand incorporating
6355 -maltivec=be semantics where needed.
6356 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
6357 VM2 iterator instead of V4SI.
6358 (altivec_stvxl_<mode>): New define_expand incorporating
6359 -maltivec=be semantics where needed.
6360 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
6361 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
6362 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
6363 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
6364 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
6365 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
6366 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
6367 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
6368 ALTIVEC_BUILTIN_STVXL.
6369 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
6370 (altivec_expand_stvx_be): Likewise.
6371 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
6372 (altivec_expand_lvx_be): Likewise.
6373 (altivec_expand_stvx_be): Likewise.
6374 (altivec_expand_builtin): Add cases for
6375 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
6376 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
6377 (altivec_init_builtins): Add definitions for
6378 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
6379 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
6381 2014-02-21 Catherine Moore <clm@codesourcery.com>
6383 * doc/invoke.texi (mvirt, mno-virt): Document.
6384 * config/mips/mips.opt (mvirt): New option.
6385 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
6387 2014-02-21 Richard Biener <rguenther@suse.de>
6389 PR tree-optimization/60276
6390 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
6391 (STMT_VINFO_MIN_NEG_DIST): New macro.
6392 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
6393 STMT_VINFO_MIN_NEG_DIST.
6394 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
6395 made for negative dependence distances still hold.
6397 2014-02-21 Richard Biener <rguenther@suse.de>
6400 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
6401 DECL_INITIAL for globals not in the current function context.
6403 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6405 PR tree-optimization/56490
6406 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
6407 * tree-ssa-uninit.c: Include params.h.
6408 (compute_control_dep_chain): Add num_calls argument, return false
6409 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
6410 num_calls to recursive call.
6411 (find_predicates): Change dep_chain into normal array,
6412 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
6413 variable and adjust compute_control_dep_chain caller.
6414 (find_def_preds): Likewise.
6416 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
6418 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
6419 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
6421 2014-02-21 Nick Clifton <nickc@redhat.com>
6423 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
6424 (pushhi1): Likewise.
6425 (popqi1): Add mode to pre_dec.
6428 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6430 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
6431 mode for mask of V8SFmode permutation.
6433 2014-02-20 Richard Henderson <rth@redhat.com>
6436 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
6437 a new pseudo for OLDVAL.
6439 2014-02-20 Jakub Jelinek <jakub@redhat.com>
6442 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
6443 gen_reg_rtx if d->testing_p.
6444 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
6445 if d->testing_p and we will certainly return true.
6446 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6449 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6451 * emit-rtl.c (gen_reg_rtx): Assert that
6452 crtl->emit.regno_pointer_align_length is non-zero.
6454 2014-02-20 Richard Henderson <rth@redhat.com>
6457 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6458 on failure the store back into EXPECT.
6460 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6461 Sandra Loosemore <sandra@codesourcery.com>
6463 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6464 * config/nios2/nios2.c (nios2_function_profiler): Add
6465 -fPIC (flag_pic == 2) support.
6466 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6467 (nios2_large_offset_p): New function.
6468 (nios2_unspec_reloc_p): Move up position, update to use
6469 nios2_large_offset_p.
6470 (nios2_unspec_address): Remove function.
6471 (nios2_unspec_offset): New function.
6472 (nios2_large_got_address): New function.
6473 (nios2_got_address): Add large offset support.
6474 (nios2_legitimize_tls_address): Update usage of removed and new
6476 (nios2_symbol_binds_local_p): New function.
6477 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
6478 (nios2_legitimize_address): Update to use nios2_large_offset_p.
6479 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
6480 (nios2_print_operand): Merge H/L processing, add hiadj/lo
6481 processing for (const (unspec ...)).
6482 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
6484 2014-02-20 Richard Biener <rguenther@suse.de>
6486 * tree-cfg.c (replace_uses_by): Mark altered BBs before
6487 doing the substitution.
6488 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
6490 2014-02-20 Martin Jambor <mjambor@suse.cz>
6493 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
6494 info when checking whether lattices are bottom.
6496 2014-02-20 Richard Biener <rguenther@suse.de>
6499 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
6502 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
6505 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
6506 parameter specifying the scaling.
6507 (inline_call): Update.
6508 (want_inline_recursively): Guard division by zero.
6509 (recursive_inlining): Update.
6510 * ipa-inline.h (clone_inlined_nodes): Update.
6512 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6515 * config/i386/i386.c (classify_argument): Pass structures of size
6516 64 bytes or less in register.
6518 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6519 Kirill Yukhin <kirill.yukhin@intel.com>
6521 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
6522 (_mm_rcp28_round_ss): Ditto.
6523 (_mm_rsqrt28_round_sd): Ditto.
6524 (_mm_rsqrt28_round_ss): Ditto.
6525 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
6526 (_mm_rcp14_round_ss): Ditto.
6527 (_mm_rsqrt14_round_sd): Ditto.
6528 (_mm_rsqrt14_round_ss): Ditto.
6529 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
6530 the first input operand, get rid of match_dup.
6531 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
6533 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
6535 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
6536 operand as the first input operand, set type attribute.
6537 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
6539 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
6540 operand as the first input operand, set type attribute.
6542 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6544 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
6547 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
6550 * config/i386/i386.c (construct_container): Remove TFmode check
6551 for X86_64_INTEGER_CLASS.
6553 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
6556 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
6557 only when -Wpsabi is enabled.
6559 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
6562 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
6563 passing arrays in registers are the same as for structs, so remove the
6564 special case for them.
6566 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
6568 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
6569 destination type, extract only the valid bits if the source type is not
6570 integral and has a different mode.
6572 2014-02-19 Richard Biener <rguenther@suse.de>
6575 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
6578 2014-02-19 Richard Biener <rguenther@suse.de>
6581 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
6582 (ipa_modify_call_arguments): Emit an argument load explicitely and
6583 preserve virtual SSA form there and for the replacement call.
6584 Do not update SSA form nor free dominance info.
6586 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6588 * ipa.c (function_and_variable_visibility): Also clear WEAK
6589 flag when disolving COMDAT_GROUP.
6591 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6593 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
6594 * ipa-prop.c (ipa_set_jf_known_type): Return early when
6596 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
6597 do more sanity checks.
6598 (detect_type_change): Return true when giving up early.
6599 (compute_complex_assign_jump_func): Fix type parameter of
6600 ipa_set_ancestor_jf.
6601 (compute_complex_ancestor_jump_func): Likewise.
6602 (update_jump_functions_after_inlining): Fix updating of
6604 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
6606 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6608 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
6609 inline clones when edge disappears.
6611 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6614 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
6615 Split 64-bit moves into 2 patterns. Do not allow the use of
6616 direct move for TDmode in little endian, since the decimal value
6617 has little endian bytes within a word, but the 64-bit pieces are
6618 ordered in a big endian fashion, and normal subreg's of TDmode are
6620 (mov<mode>_64bit_dm): Likewise.
6621 (movtd_64bit_nodm): Likewise.
6623 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6625 PR tree-optimization/60174
6626 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
6627 statement of an SSA_NAME that occurs in an abnormal PHI node.
6629 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6632 * final.c (SEEN_BB): Remove.
6633 (SEEN_NOTE, SEEN_EMITTED): Renumber.
6634 (final_scan_insn): Don't force_source_line on second
6635 NOTE_INSN_BASIC_BLOCK.
6637 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
6640 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
6641 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
6642 (type_natural_mode): Warn ABI change when %zmm register is not
6643 available for AVX512F vector value passing.
6645 2014-02-18 Kai Tietz <ktietz@redhat.com>
6648 * config/i386/i386.c (ix86_expand_prologue): Use value in
6649 rax register as displacement when restoring %r10 or %rax.
6650 Fix wrong offset when restoring both registers.
6652 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6654 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
6655 assertion with conditional return.
6657 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6658 Uros Bizjak <ubizjak@gmail.com>
6661 * config/i386/driver-i386.c (host_detect_local_cpu): If
6662 YMM state is not saved by the OS, also clear has_f16c. Move
6663 CPUID 0x80000001 handling before YMM state saving checking.
6665 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
6667 PR rtl-optimization/58960
6668 * haifa-sched.c (alloc_global_sched_pressure_data): New,
6669 factored out from ...
6670 (sched_init): ... here.
6671 (free_global_sched_pressure_data): New, factored out from ...
6672 (sched_finish): ... here.
6673 * sched-int.h (free_global_sched_pressure_data): Declare.
6674 * sched-rgn.c (nr_regions_initial): New static global.
6675 (haifa_find_rgns): Initialize it.
6676 (schedule_region): Disable sched-pressure for the newly
6679 2014-02-17 Richard Biener <rguenther@suse.de>
6681 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
6682 release SSA defs of pattern stmts.
6684 2014-02-17 Richard Biener <rguenther@suse.de>
6686 * tree-inline.c (expand_call_inline): Release the virtual
6687 operand defined by the call we are about to inline.
6689 2014-02-17 Richard Biener <rguenther@suse.de>
6691 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
6693 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
6694 Ilya Tocar <ilya.tocar@intel.com>
6696 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
6697 arguments order in builtin.
6698 (_mm512_permutexvar_epi64): Ditto.
6699 (_mm512_mask_permutexvar_epi64): Ditto
6700 (_mm512_maskz_permutexvar_epi32): Ditto
6701 (_mm512_permutexvar_epi32): Ditto
6702 (_mm512_mask_permutexvar_epi32): Ditto
6704 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6706 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
6707 (p8_vmrgow): Likewise.
6709 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6711 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
6714 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
6717 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
6718 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
6719 into 64-bit and 32-bit moves. On 64-bit moves, add support for
6720 using direct move instructions on ISA 2.07. Also adjust
6721 instruction length for 64-bit.
6722 (mov<mode>_64bit, TFmode/TDmode): Likewise.
6723 (mov<mode>_32bit, TFmode/TDmode): Likewise.
6725 2014-02-15 Alan Modra <amodra@gmail.com>
6729 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
6730 find_replacement on parts of insn rtl that might be reloaded.
6732 2014-02-15 Richard Biener <rguenther@suse.de>
6734 PR tree-optimization/60183
6735 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
6736 (tree_ssa_phiprop): Calculate and free post-dominators.
6738 2014-02-14 Jeff Law <law@redhat.com>
6740 PR rtl-optimization/60131
6741 * ree.c (get_extended_src_reg): New function.
6742 (combine_reaching_defs): Use it rather than assuming location of REG.
6743 (find_and_remove_re): Verify first operand of extension is
6744 a REG before adding the insns to the copy list.
6746 2014-02-14 Roland McGrath <mcgrathr@google.com>
6748 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
6749 * configure: Regenerated.
6750 * config.in: Regenerated.
6751 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
6752 instead of ASM_SHORT.
6754 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
6755 Richard Earnshaw <rearnsha@arm.com>
6757 PR rtl-optimization/59535
6758 * lra-constraints.c (process_alt_operands): Encourage alternative
6759 when unassigned pseudo class is superset of the alternative class.
6760 (inherit_reload_reg): Don't inherit when optimizing for code size.
6761 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
6762 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
6763 modes not less than 4 for Thumb1.
6765 2014-02-14 Kyle McMartin <kyle@redhat.com>
6768 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
6770 2014-02-14 Richard Biener <rguenther@suse.de>
6772 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
6773 (get_frame_arg): Drop the assert with langhook types_compatible_p.
6774 Do not strip INDIRECT_REFs.
6776 2014-02-14 Richard Biener <rguenther@suse.de>
6779 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
6780 DECL_FUNCTION_SPECIFIC_TARGET.
6781 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
6782 * tree-streamer-out.c (pack_ts_target_option): Remove.
6783 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
6784 (write_ts_function_decl_tree_pointers): Do not stream
6785 DECL_FUNCTION_SPECIFIC_TARGET.
6786 * tree-streamer-in.c (unpack_ts_target_option): Remove.
6787 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
6788 (lto_input_ts_function_decl_tree_pointers): Do not stream
6789 DECL_FUNCTION_SPECIFIC_TARGET.
6791 2014-02-14 Jakub Jelinek <jakub@redhat.com>
6793 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
6794 (get_initial_def_for_induction, vectorizable_induction): Ignore
6795 debug stmts when looking for exit_phi.
6796 (vectorizable_live_operation): Fix up condition.
6798 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6800 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
6801 nreverse() because it changes the content of original tree list.
6803 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6805 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
6806 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
6808 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6810 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
6811 GNU coding standards.
6813 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6816 * dwarf2out.c (gen_subprogram_die): Don't call
6817 add_calling_convention_attribute if subr_die is old_die.
6819 2014-02-13 Sharad Singhai <singhai@google.com>
6821 * doc/optinfo.texi: Fix order of nodes.
6823 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
6825 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
6826 operands[2], not operands[3].
6828 2014-02-13 Richard Biener <rguenther@suse.de>
6831 * doc/install.texi (ISL): Update recommended version to 0.12.2,
6832 mention the possibility of an in-tree build.
6833 (CLooG): Update recommended version to 0.18.1, mention the
6834 possibility of an in-tree build and clarify that the ISL
6835 bundled with CLooG does not work.
6837 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6840 * expr.c (compress_float_constant): If x is a hard register,
6841 extend into a pseudo and then move to x.
6843 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
6845 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
6846 caused by bad second argument to warning_at() with -mhotpatch and
6847 nested functions (e.g. with gfortran).
6849 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
6851 * opts.c (option_name): Remove "enabled by default" rider.
6853 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
6855 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
6857 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
6858 Uros Bizjak <ubizjak@gmail.com>
6861 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
6862 * configure: Regenerated.
6864 2014-02-12 Richard Biener <rguenther@suse.de>
6866 * vec.c (vec_prefix::calculate_allocation): Move as
6867 inline variant to vec.h.
6868 (vec_prefix::calculate_allocation_1): New out-of-line version.
6869 * vec.h (vec_prefix::calculate_allocation_1): Declare.
6870 (vec_prefix::m_has_auto_buf): Rename to ...
6871 (vec_prefix::m_using_auto_storage): ... this.
6872 (vec_prefix::calculate_allocation): Inline the easy cases
6873 and dispatch to calculate_allocation_1 which doesn't need the
6875 (va_heap::reserve): Use gcc_checking_assert.
6876 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
6877 m_using_auto_storage.
6878 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
6880 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
6881 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
6882 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
6884 2014-02-12 Richard Biener <rguenther@suse.de>
6886 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
6887 when we found a dependence.
6889 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
6891 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
6893 (maybe_fold_stmt): ... into this new function.
6894 * omp-low.c (lower_omp): Update comment.
6896 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
6899 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
6902 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
6904 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
6905 identifiers in comments.
6906 (cortexa53_extra_costs): Likewise.
6907 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
6908 (cortexa7_extra_costs): Likewise.
6909 (cortexa12_extra_costs): Likewise.
6910 (cortexa15_extra_costs): Likewise.
6911 (v7m_extra_costs): Likewise.
6913 2014-02-12 Richard Biener <rguenther@suse.de>
6916 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
6917 of posix_memalign being successful.
6918 (lower_stmt): Restrict lowering of posix_memalign to when
6919 -ftree-bit-ccp is enabled.
6921 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6923 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
6925 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
6927 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
6929 PR rtl-optimization/60116
6930 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
6931 other_insn once the combination has been validated.
6933 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
6936 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
6938 * ipa-devirt.c: Include demangle.h
6939 (odr_violation_reported): New static variable.
6940 (add_type_duplicate): Update odr_violations.
6941 (maybe_record_node): Add completep parameter; update it.
6942 (record_target_from_binfo): Add COMPLETEP parameter;
6943 update it as needed.
6944 (possible_polymorphic_call_targets_1): Likewise.
6945 (struct polymorphic_call_target_d): Add nonconstruction_targets;
6946 rename FINAL to COMPLETE.
6947 (record_targets_from_bases): Sanity check we found the binfo;
6948 fix COMPLETEP updating.
6949 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
6950 parameter, fix computing of COMPLETEP.
6951 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
6952 at LTO time do demangling.
6953 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
6954 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
6956 (gimple_get_virt_method_for_binfo): Likewise.
6957 * gimple-fold.h (gimple_get_virt_method_for_binfo,
6958 gimple_get_virt_method_for_vtable): Update prototypes.
6960 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
6963 * genautomata.c (add_presence_absence): Fix typo with
6964 {final_}presence_list.
6966 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
6969 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
6970 for VSX/Altivec vectors that land in GPR registers.
6972 2014-02-11 Richard Henderson <rth@redhat.com>
6973 Jakub Jelinek <jakub@redhat.com>
6976 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
6977 around drhs if type conversion to lacc->type is not useless.
6979 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6981 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
6983 (cortex-a57.cortex-a53): Likewise.
6984 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
6986 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6988 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
6991 2014-02-11 Renlin Li <Renlin.Li@arm.com>
6993 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
6994 add_options_for_arm_vfp3.
6996 2014-02-11 Jeff Law <law@redhat.com>
6999 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
7000 object with an undesirable mode.
7002 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7005 * config/i386/sol2-9.h: New file.
7006 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
7007 *-*-solaris2.9*): Use it.
7009 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7011 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
7012 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
7014 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7016 * config/microblaze/microblaze.c: Extend mcpu version format
7018 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
7020 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
7022 2014-02-10 Richard Henderson <rth@redhat.com>
7025 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
7026 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
7027 ms-abi vs -mno-accumulate-outgoing-args.
7028 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
7029 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
7032 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
7035 * cfgexpand.c (expand_asm_operands): Attach source location to
7036 ASM_INPUT rtx objects.
7037 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
7039 2014-02-10 Nick Clifton <nickc@redhat.com>
7041 * config/mn10300/mn10300.c (popcount): New function.
7042 (mn10300_expand_prologue): Include saved registers in stack usage
7045 2014-02-10 Jeff Law <law@redhat.com>
7048 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
7049 when changing the SET_DEST of a prior insn to avoid an input reload.
7051 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7053 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
7054 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
7055 -mcall-openbsd, or -mcall-linux.
7056 (CC1_ENDIAN_BIG_SPEC): Remove.
7057 (CC1_ENDIAN_LITTLE_SPEC): Remove.
7058 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7059 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
7060 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
7061 and %cc1_endian_default.
7062 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7064 2014-02-10 Richard Biener <rguenther@suse.de>
7066 PR tree-optimization/60115
7067 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
7068 MEM_REF handling. Properly verify that the accesses are not
7069 out of the objects bound.
7071 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7073 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
7076 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
7078 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
7079 proper constants and fix formatting.
7080 (possible_polymorphic_call_targets): Fix formatting.
7082 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
7083 Ilya Tocar <ilya.tocar@intel.com>
7085 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
7086 (_mm512_loadu_epi32): Renamed into...
7087 (_mm512_loadu_si512): This.
7088 (_mm512_storeu_epi32): Renamed into...
7089 (_mm512_storeu_si512): This.
7090 (_mm512_maskz_ceil_ps): Removed.
7091 (_mm512_maskz_ceil_pd): Ditto.
7092 (_mm512_maskz_floor_ps): Ditto.
7093 (_mm512_maskz_floor_pd): Ditto.
7094 (_mm512_floor_round_ps): Ditto.
7095 (_mm512_floor_round_pd): Ditto.
7096 (_mm512_ceil_round_ps): Ditto.
7097 (_mm512_ceil_round_pd): Ditto.
7098 (_mm512_mask_floor_round_ps): Ditto.
7099 (_mm512_mask_floor_round_pd): Ditto.
7100 (_mm512_mask_ceil_round_ps): Ditto.
7101 (_mm512_mask_ceil_round_pd): Ditto.
7102 (_mm512_maskz_floor_round_ps): Ditto.
7103 (_mm512_maskz_floor_round_pd): Ditto.
7104 (_mm512_maskz_ceil_round_ps): Ditto.
7105 (_mm512_maskz_ceil_round_pd): Ditto.
7106 (_mm512_expand_pd): Ditto.
7107 (_mm512_expand_ps): Ditto.
7108 * config/i386/i386.c (ix86_builtins): Remove
7109 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
7110 (bdesc_args): Ditto.
7111 * config/i386/predicates.md (const1256_operand): New.
7112 (const_1_to_2_operand): Ditto.
7113 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7114 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7115 (*avx512pf_gatherpf<mode>sf): Ditto.
7116 (avx512pf_gatherpf<mode>df): Ditto.
7117 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7118 (*avx512pf_gatherpf<mode>df): Ditto.
7119 (avx512pf_scatterpf<mode>sf): Ditto.
7120 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7121 (*avx512pf_scatterpf<mode>sf): Ditto.
7122 (avx512pf_scatterpf<mode>df): Ditto.
7123 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7124 (*avx512pf_scatterpf<mode>df): Ditto.
7125 (avx512f_expand<mode>): Removed.
7126 (<shift_insn><mode>3<mask_name>): Change predicate type.
7128 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7130 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
7131 not at the end of datarefs vector use ordered_remove to avoid
7132 reordering datarefs vector.
7135 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
7136 mark local addressable non-static vars as GOVD_PRIVATE
7137 instead of GOVD_LOCAL.
7138 * omp-low.c (lower_omp_for): Move gimple_bind_vars
7139 and BLOCK_VARS of gimple_bind_block to new_stmt rather
7143 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
7144 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
7145 assume_aligned or alloc_align attributes.
7146 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
7147 arguments. Handle also assume_aligned and alloc_align attributes.
7148 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
7149 calls to functions with assume_aligned or alloc_align attributes.
7150 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
7152 2014-02-08 Terry Guo <terry.guo@arm.com>
7154 * doc/invoke.texi: Document ARM -march=armv7e-m.
7156 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7158 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
7159 flag on __cilkrts_rethrow builtin.
7162 * ipa-cp.c (determine_versionability): Fail at -O0
7163 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
7164 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
7167 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7170 * tree-inline.c (copy_forbidden): Fail for
7171 __attribute__((optimize (0))) functions.
7173 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7175 * varpool.c: Include pointer-set.h.
7176 (varpool_remove_unreferenced_decls): Variables in other partitions
7177 will not be output; be however careful to not lose information
7180 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7182 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
7183 lookup in the vtable constructor.
7185 2014-02-07 Jeff Law <law@redhat.com>
7188 * config/m68k/m68k.md (ashldi_extsi): Turn into a
7189 define_insn_and_split.
7191 * ipa-inline.c (inline_small_functions): Fix typos.
7193 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7195 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
7196 (s390_can_use_return_insn): Declare.
7197 * config/s390/s390.h (EPILOGUE_USES): Define.
7198 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
7200 (s390_chunkify_start): Handle return JUMP_LABELs.
7201 (s390_early_mach): Emit a main_pool instruction on the entry edge.
7202 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
7203 (s390_can_use_return_insn): New functions.
7204 (s390_fix_long_loop_prediction): Handle conditional returns.
7205 (TARGET_SET_UP_BY_PROLOGUE): Define.
7206 * config/s390/s390.md (ANY_RETURN): New code iterator.
7207 (*creturn, *csimple_return, return, simple_return): New patterns.
7209 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7211 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
7212 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
7213 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
7214 REG_CFA_RESTORE list when deciding not to restore a register.
7216 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7218 * config/s390/s390.c: Include tree-pass.h and context.h.
7219 (s390_early_mach): New function, split out from...
7220 (s390_emit_prologue): ...here.
7221 (pass_data_s390_early_mach): New pass structure.
7222 (pass_s390_early_mach): New class.
7223 (s390_option_override): Create and register early_mach pass.
7224 Move to end of file.
7226 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7228 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
7229 to match for the exit block.
7231 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7233 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
7234 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
7235 Reject misaligned operands.
7237 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7239 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
7241 2014-02-07 Richard Biener <rguenther@suse.de>
7244 * gimple-low.c (lower_builtin_posix_memalign): New function.
7245 (lower_stmt): Call it to lower posix_memalign in a way
7246 to make alignment info accessible.
7248 2014-02-07 Jakub Jelinek <jakub@redhat.com>
7251 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
7252 __builtin_setjmp_receiver.
7254 2014-02-07 Richard Biener <rguenther@suse.de>
7257 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
7258 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
7259 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7260 Handle BUILT_IN_POSIX_MEMALIGN.
7261 (find_func_clobbers): Likewise.
7262 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
7263 (call_may_clobber_ref_p_1): Likewise.
7265 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7268 * ipa-devirt.c (record_target_from_binfo): Remove overactive
7271 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7274 * lto-cgraph.c (lto_output_node): Use
7275 symtab_get_symbol_partitioning_class.
7276 (lto_output_varpool_node): likewise.
7277 (symtab_get_symbol_partitioning_class): Move here from
7279 * cgraph.h (symbol_partitioning_class): Likewise.
7280 (symtab_get_symbol_partitioning_class): Declare.
7282 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7284 * ggc.h (ggc_internal_cleared_alloc): New macro.
7285 * vec.h (vec_safe_copy): Handle memory stats.
7286 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
7287 * target-globals.c (save_target_globals): Likewise.
7289 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7292 * expr.c (emit_move_resolve_push): Export; be bit more selective
7293 on when to clear alias set.
7294 * expr.h (emit_move_resolve_push): Declare.
7295 * function.h (struct function): Add tail_call_marked.
7296 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
7297 * config/i386/i386-protos.h (ix86_expand_push): Remove.
7298 * config/i386/i386.md (TImode move expander): De not call
7300 (FP push expanders): Preserve memory attributes.
7301 * config/i386/sse.md (push<mode>1): Remove.
7302 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
7303 (ix86_expand_push): Remove.
7304 * config/i386/mmx.md (push<mode>1): Remove.
7306 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7308 PR rtl-optimization/60030
7309 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
7310 lopart with paradoxical subreg before shifting it up by hprec.
7312 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7314 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
7315 Remove extra newline at end of file.
7316 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
7317 (arm_issue_rate): Handle cortexa57.
7318 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
7319 (cortex-a57.cortex-a53): Likewise.
7321 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7324 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
7325 don't record in REG_FRAME_RELATED_EXPR registers not set in that
7327 (arm_expand_prologue): Adjust all callers.
7328 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
7329 info, registers also at the lowest numbered registers side. Use
7330 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
7334 * var-tracking.c (adjust_mems): Before adding a SET to
7335 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
7337 2014-02-06 Alan Modra <amodra@gmail.com>
7340 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
7341 change SDmode to DDmode when lra_in_progress.
7343 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7346 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
7347 free_data_ref on the dr first, and before goto again also set dr
7348 to the next dr. For simd_lane_access, free old datarefs[i] before
7349 overwriting it. For get_vectype_for_scalar_type failure, don't
7350 free_data_ref if simd_lane_access.
7352 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
7355 * tree.h (opts_for_fn): New inline function.
7356 (opt_for_fn): Define.
7357 * config/i386/i386.c (ix86_function_regparm): Use
7358 opt_for_fn (decl, optimize) instead of optimize.
7360 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
7362 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
7363 for SYMBOL_REF in large memory model.
7365 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7367 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
7369 (cortex-a57): Likewise.
7370 (cortex-a57.cortex-a53): Likewise.
7372 2014-02-06 Yury Gribov <y.gribov@samsung.com>
7373 Kugan Vivekanandarajah <kuganv@linaro.org>
7375 * config/arm/arm.c (arm_vector_alignment_reachable): Check
7377 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
7379 2014-02-06 Richard Biener <rguenther@suse.de>
7381 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
7382 set_loop_copy and initialize_original_copy_tables.
7384 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
7386 * config/aarch64/aarch64-simd.md
7387 (aarch64_ashr_simddi): Change QI to SI.
7389 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7390 Jakub Jelinek <jakub@redhat.com>
7393 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
7396 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7398 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
7399 CODE_FOR_altivec_vpku[hw]um to
7400 CODE_FOR_altivec_vpku[hw]um_direct.
7401 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
7402 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
7403 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
7404 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
7406 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7408 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
7409 generation for -maltivec=be.
7410 (altivec_vsumsws): Simplify redundant test.
7412 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7414 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
7415 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
7416 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
7417 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
7418 gen_altivec_vpkuwum.
7419 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
7421 (altivec_vpks<VI_char>ss): Likewise.
7422 (altivec_vpks<VI_char>us): Likewise.
7423 (altivec_vpku<VI_char>us): Likewise.
7424 (altivec_vpku<VI_char>um): Likewise.
7425 (altivec_vpku<VI_char>um_direct): New (copy of
7426 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
7428 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
7429 target is little endian and -maltivec=be is not specified.
7430 (*altivec_vupkhs<VU_char>_direct): New (copy of
7431 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
7432 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
7433 target is little endian and -maltivec=be is not specified.
7434 (*altivec_vupkls<VU_char>_direct): New (copy of
7435 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
7436 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
7437 little endian and -maltivec=be is not specified.
7438 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
7439 little endian and -maltivec=be is not specified.
7441 2014-02-05 Richard Henderson <rth@redhat.com>
7444 * combine-stack-adj.c: Revert r206943.
7445 * sched-int.h (struct deps_desc): Add last_args_size.
7446 * sched-deps.c (init_deps): Initialize it.
7447 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7448 contain REG_ARGS_SIZE notes.
7450 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7452 * lto-cgraph.c (asm_nodes_output): Make global.
7453 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7454 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7455 (driver_handle_option): Handle OPT_fwpa.
7457 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7460 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7461 a comment typo and formatting issue. If odr_hash hasn't been
7462 created, return vNULL and set *completep to false.
7465 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7466 bb with no successors.
7468 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7471 * doc/invoke.texi (-march): Clarify documentation for ARM.
7475 2014-02-05 Richard Biener <rguenther@suse.de>
7477 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
7478 when not vectorizing because of too many alias checks.
7479 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
7480 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
7482 2014-02-05 Nick Clifton <nickc@redhat.com>
7484 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
7485 accept extended registers in any mode when compiling for the MN10300.
7487 2014-02-05 Yury Gribov <y.gribov@samsung.com>
7489 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
7490 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
7491 sanitization attributes.
7492 (can_inline_edge_p): Likewise.
7493 (sanitize_attrs_match_for_inline_p): New function.
7495 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7497 * ipa-prop.c (detect_type_change): Shor circuit testing of
7498 type changes on THIS pointer.
7500 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
7503 * config/pa/pa.c (legitimize_tls_address): Return original address
7504 if not passed a SYMBOL_REF rtx.
7505 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
7507 (pa_emit_move_sequence): Simplify TLS source operands.
7508 (pa_legitimate_constant_p): Reject all TLS constants.
7509 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
7510 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
7512 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7514 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
7515 groups when we know they are controlled by LTO.
7516 * varasm.c (default_binds_local_p_1): If object is in other partition,
7517 it will be resolved locally.
7519 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7521 * config/host-linux.c (linux_gt_pch_use_address): Don't
7522 use SSIZE_MAX because it is not always defined.
7524 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
7527 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
7528 threshold for pseudo splitting.
7529 (update_ebb_live_info): Process call argument hard registers and
7530 hard registers from insn definition too.
7531 (max_small_class_regs_num): New constant.
7532 (inherit_in_ebb): Update live hard regs through EBBs. Update
7533 reloads_num only for small register classes. Don't split for
7536 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
7539 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
7542 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7544 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
7547 2014-02-04 Marek Polacek <polacek@redhat.com>
7549 * gdbinit.in (pel): Define.
7551 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7553 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
7556 2014-02-04 Richard Biener <rguenther@suse.de>
7559 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
7560 in function context local.
7561 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
7562 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
7563 similar to LTO_imported_decl_ref.
7565 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7567 PR tree-optimization/60002
7568 * cgraphclones.c (build_function_decl_skip_args): Clear
7571 PR tree-optimization/60023
7572 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
7573 false to gsi_replace.
7574 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
7575 has been in some EH region and vec_stmt could throw, add
7576 vec_stmt into the same EH region.
7577 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
7578 has no lhs, ignore it.
7579 * internal-fn.c (expand_MASK_LOAD): Likewise.
7582 * tree-inline.c (copy_forbidden): Fail for
7583 __attribute__((optimize (0))) functions.
7586 * omp-low.c (simd_clone_struct_copy): If from->inbranch
7587 is set, copy one less argument.
7588 (expand_simd_clones): Don't subtract clone_info->inbranch
7589 from simd_clone_struct_alloc argument.
7591 PR rtl-optimization/57915
7592 * recog.c (simplify_while_replacing): If all unary/binary/relational
7593 operation arguments are constant, attempt to simplify those.
7596 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
7597 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
7599 2014-02-04 Richard Biener <rguenther@suse.de>
7601 PR tree-optimization/60012
7602 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
7603 TBAA disambiguation to all DDRs.
7605 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7608 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
7609 (LINK_SPEC): Use it for -shared, -shared-libgcc.
7611 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7614 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
7616 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7618 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
7619 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
7621 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7624 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
7625 to figure out targets of polymorphic calls with known decl.
7626 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7627 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
7628 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
7629 (get_polymorphic_call_info): ... here.
7630 (get_polymorphic_call_info_from_invariant): New function.
7632 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7634 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
7635 lookup via vtable pointer; check for type consistency
7636 and turn inconsitent facts into UNREACHABLE.
7637 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7638 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
7639 type inconsistent querries; return UNREACHABLE instead.
7641 2014-02-03 Richard Henderson <rth@twiddle.net>
7644 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
7645 already processed this node.
7646 (normalize_one_pred_1): Pass along mark_set.
7647 (normalize_one_pred): Create and destroy a pointer_set_t.
7648 (normalize_one_pred_chain): Likewise.
7650 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
7652 PR gcov-profile/58602
7653 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
7655 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7658 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
7659 -fno-devirtualize; try to devirtualize by the knowledge of
7660 virtual table pointer given by aggregate propagation.
7661 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7662 (ipa_print_node_jump_functions): Dump also offset that
7663 is relevant for polymorphic calls.
7664 (determine_known_aggregate_parts): Add arg_type parameter; use it
7665 instead of determining the type from pointer type.
7666 (ipa_compute_jump_functions_for_edge): Update call of
7667 determine_known_aggregate_parts.
7668 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
7669 (gimple_get_virt_method_for_binfo): ... here; simplify using
7670 vtable_pointer_value_to_vtable.
7671 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
7672 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
7673 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
7674 (vtable_pointer_value_to_vtable): Break out from ...; handle also
7676 (vtable_pointer_value_to_binfo): ... here.
7677 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
7679 2014-02-03 Teresa Johnson <tejohnson@google.com>
7681 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
7682 redef of outer loop index variable.
7684 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
7688 * doc/extend.texi (Function Attributes): Typo.
7690 2014-02-03 Cong Hou <congh@google.com>
7692 PR tree-optimization/60000
7693 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
7694 if the vectorized statement is a store. A store statement can only
7695 appear at the end of pattern statements.
7697 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7699 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
7700 (ix86_option_override_internal): Default long double to 64-bit for
7701 32-bit Bionic and to 128-bit for 64-bit Bionic.
7703 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
7704 TARGET_LONG_DOUBLE_128 is true.
7705 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
7707 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
7708 (mlong-double-64): Negate -mlong-double-128.
7709 (mlong-double-128): New option.
7711 * config/i386/i386-c.c (ix86_target_macros): Define
7712 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
7714 * doc/invoke.texi: Document -mlong-double-128.
7716 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7718 PR rtl-optimization/60024
7719 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
7721 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
7723 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
7725 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7727 PR rtl-optimization/57662
7728 * sel-sched.c (code_motion_path_driver): Do not mark already not
7729 existing blocks in the visiting bitmap.
7731 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7733 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
7734 on the insn being emitted.
7736 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
7737 Will Deacon <will.deacon@arm.com>
7739 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
7741 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7743 * config/arm/arm-tables.opt: Regenerate.
7745 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7747 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
7748 for vector types other than V16QImode.
7749 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
7750 define_expand, and call altivec_expand_vec_perm_le when producing
7751 code with little endian element order.
7752 (*altivec_vperm_<mode>_internal): New insn having previous
7753 behavior of altivec_vperm_<mode>.
7754 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
7755 altivec_expand_vec_perm_le when producing code with little endian
7757 (*altivec_vperm_<mode>_uns_internal): New insn having previous
7758 behavior of altivec_vperm_<mode>_uns.
7760 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7762 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
7763 (altivec_vsumsws): Add handling for -maltivec=be with a little
7765 (altivec_vsumsws_direct): New.
7766 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
7767 gen_altivec_vsumsws.
7769 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7771 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
7772 vtable_pointer_value_to_binfo): New functions.
7773 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
7774 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
7776 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
7778 * config/nios2/nios2.md (load_got_register): Initialize GOT
7779 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
7780 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
7782 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7784 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
7785 preserverd by passthrough, do not propagate the type.
7787 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7789 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
7790 (mips_atomic_assign_expand_fenv): New function.
7791 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
7793 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7795 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
7796 (__builtin_mips_set_fcsr): Likewise.
7797 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
7798 MIPS_USI_FTYPE_VOID.
7799 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
7800 (mips16_expand_set_fcsr): Likewise.
7801 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
7802 (mips16_set_fcsr_stub): Likewise.
7803 (mips16_get_fcsr_one_only_stub): New class.
7804 (mips16_set_fcsr_one_only_stub): Likewise.
7805 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
7806 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
7807 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
7808 (hard_float): New availability predicate.
7809 (mips_builtins): Add get_fcsr and set_fcsr.
7810 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
7811 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
7812 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
7813 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
7814 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
7817 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7819 * config/mips/mips.c (mips_one_only_stub): New class.
7820 (mips_need_mips16_rdhwr_p): Replace with...
7821 (mips16_rdhwr_stub): ...this new variable.
7822 (mips16_stub_call_address): New function.
7823 (mips16_rdhwr_one_only_stub): New class.
7824 (mips_expand_thread_pointer): Use mips16_stub_call_address.
7825 (mips_output_mips16_rdhwr): Delete.
7826 (mips_finish_stub): New function.
7827 (mips_code_end): Use it to handle rdhwr stubs.
7829 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
7832 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
7833 when calculating size of integer atomic types.
7835 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
7837 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
7839 2014-02-01 Jakub Jelinek <jakub@redhat.com>
7841 PR tree-optimization/60003
7842 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
7843 * profile.c (branch_prob): Use gimple_call_builtin_p
7844 to check for BUILT_IN_SETJMP_RECEIVER.
7845 * tree-inline.c (copy_bb): Call notice_special_calls.
7847 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
7850 * lra-constraints.c (process_alt_operands): Update reload_sum only
7853 2014-01-31 Richard Henderson <rth@redhat.com>
7856 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
7857 until after else_eh is processed.
7859 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7861 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
7862 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
7863 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
7864 in smmintrin.h, remove them.
7865 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
7866 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
7867 * config/i386/i386.md (ROUND_SAE): Fix value.
7868 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
7869 (const48_operand): New.
7870 * config/i386/subst.md (round), (round_expand): Use
7871 const_4_or_8_to_11_operand.
7872 (round_saeonly), (round_saeonly_expand): Use const48_operand.
7874 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7876 * config/i386/constraints.md (Yk): Swap meaning with k.
7877 * config/i386/i386.md (movhi_internal): Change Yk to k.
7878 (movqi_internal): Ditto.
7879 (*k<logic><mode>): Ditto.
7882 (kandn<mode>): Ditto.
7883 (*<code>hi_1): Ditto.
7884 (*<code>qi_1): Ditto.
7885 (kxnor<mode>): Ditto.
7886 (kortestzhi): Ditto.
7887 (kortestchi): Ditto.
7889 (*one_cmplhi2_1): Ditto.
7890 (*one_cmplqi2_1): Ditto.
7891 * config/i386/sse.md (): Change k to Yk.
7892 (avx512f_load<mode>_mask): Ditto.
7893 (avx512f_blendm<mode>): Ditto.
7894 (avx512f_store<mode>_mask): Ditto.
7895 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
7896 (avx512f_storedqu<mode>_mask): Ditto.
7897 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
7899 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
7900 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
7901 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
7902 (avx512f_maskcmp<mode>3): Ditto.
7903 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
7904 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
7905 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
7906 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
7907 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
7908 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
7909 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
7910 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
7911 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
7912 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
7913 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
7914 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
7915 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
7916 (vec_extract_lo_<mode>_maskm): Ditto.
7917 (vec_extract_hi_<mode>_maskm): Ditto.
7918 (avx512f_vternlog<mode>_mask): Ditto.
7919 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
7920 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
7921 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
7922 (avx512f_<code>v8div16qi2_mask): Ditto.
7923 (avx512f_<code>v8div16qi2_mask_store): Ditto.
7924 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
7925 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
7926 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
7927 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
7928 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7929 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7930 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7931 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7932 (avx512cd_maskb_vec_dupv8di): Ditto.
7933 (avx512cd_maskw_vec_dupv16si): Ditto.
7934 (avx512f_vpermi2var<mode>3_maskz): Ditto.
7935 (avx512f_vpermi2var<mode>3_mask): Ditto.
7936 (avx512f_vpermi2var<mode>3_mask): Ditto.
7937 (avx512f_vpermt2var<mode>3_maskz): Ditto.
7938 (*avx512f_gathersi<mode>): Ditto.
7939 (*avx512f_gathersi<mode>_2): Ditto.
7940 (*avx512f_gatherdi<mode>): Ditto.
7941 (*avx512f_gatherdi<mode>_2): Ditto.
7942 (*avx512f_scattersi<mode>): Ditto.
7943 (*avx512f_scatterdi<mode>): Ditto.
7944 (avx512f_compress<mode>_mask): Ditto.
7945 (avx512f_compressstore<mode>_mask): Ditto.
7946 (avx512f_expand<mode>_mask): Ditto.
7947 * config/i386/subst.md (mask): Change k to Yk.
7948 (mask_scalar_merge): Ditto.
7951 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
7953 * doc/extend.texi (Vector Extensions): Document ?: in C++.
7955 2014-01-31 Richard Biener <rguenther@suse.de>
7958 * builtins.c (fold_builtin_memory_op): Make sure to not
7959 use a floating-point mode or a boolean or enumeral type for
7962 2014-01-30 DJ Delorie <dj@redhat.com>
7964 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
7965 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
7966 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
7967 whenever main() has an epilogue.
7969 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7971 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
7972 unused variable "field".
7973 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
7974 (vsx_mergeh_<mode>): Likewise.
7975 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
7976 (altivec_vmrghh): Likewise.
7977 (altivec_vmrghw): Likewise.
7978 (altivec_vmrglb): Likewise.
7979 (altivec_vmrglh): Likewise.
7980 (altivec_vmrglw): Likewise.
7981 (altivec_vspltb): Add missing uses.
7982 (altivec_vsplth): Likewise.
7983 (altivec_vspltw): Likewise.
7984 (altivec_vspltsf): Likewise.
7986 2014-01-30 Jakub Jelinek <jakub@redhat.com>
7989 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
7990 frame related instructions.
7992 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
7994 PR rtl-optimization/59959
7995 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
7996 any reload of register whose subreg is invalid.
7998 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8000 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
8001 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
8002 Add missing return type - void.
8004 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8006 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
8007 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
8008 remove element index adjustment for endian (now handled in vsx.md
8010 (altivec_expand_vec_perm_const): Use
8011 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
8012 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
8013 (vsx_xxspltw_<mode>): Adjust element index for little endian.
8014 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
8015 define_expand and a new define_insn *altivec_vspltb_internal;
8016 adjust for -maltivec=be on a little endian target.
8017 (altivec_vspltb_direct): New.
8018 (altivec_vsplth): Divide into a define_expand and a new
8019 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
8020 little endian target.
8021 (altivec_vsplth_direct): New.
8022 (altivec_vspltw): Divide into a define_expand and a new
8023 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
8024 little endian target.
8025 (altivec_vspltw_direct): New.
8026 (altivec_vspltsf): Divide into a define_expand and a new
8027 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
8028 a little endian target.
8030 2014-01-30 Richard Biener <rguenther@suse.de>
8032 PR tree-optimization/59993
8033 * tree-ssa-forwprop.c (associate_pointerplus): Check we
8034 can propagate form the earlier stmt and avoid the transform
8035 when the intermediate result is needed.
8037 2014-01-30 Alangi Derick <alangiderick@gmail.com>
8039 * README.Portability: Fix typo.
8041 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
8043 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
8044 comparison_operator with ordered_comparison_operator.
8046 2014-01-30 Nick Clifton <nickc@redhat.com>
8048 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
8049 Rename to mn10300_store_multiple_regs.
8050 * config/mn10300/mn10300.c: Likewise.
8051 * config/mn10300/mn10300.md (store_movm): Fix typo: call
8052 store_multiple_regs.
8053 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
8054 Call mn10300_store_multiple_regs.
8056 2014-01-30 Nick Clifton <nickc@redhat.com>
8057 DJ Delorie <dj@redhat.com>
8059 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
8060 %fp 2 to keep registers after it properly word-aligned.
8061 (rl78_alloc_physical_registers_umul): Handle the case where both
8062 input operands are the same.
8064 2014-01-30 Richard Biener <rguenther@suse.de>
8066 PR tree-optimization/59903
8067 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
8070 2014-01-30 Jason Merrill <jason@redhat.com>
8073 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
8076 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
8078 2014-01-30 Richard Biener <rguenther@suse.de>
8080 PR tree-optimization/59951
8081 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
8083 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
8086 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
8087 SFmode to DFmode case.
8089 2014-01-29 DJ Delorie <dj@redhat.com>
8091 * config/msp430/msp430.opt (-minrt): New.
8092 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
8094 (ENDFILE_SPEC): Likewise.
8096 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
8098 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
8099 (estimate_function_body_sizes): Use it.
8101 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
8104 * dwarf2out.c (is_cxx_auto): New.
8105 (is_base_type): Use it.
8106 (gen_type_die_with_usage): Likewise.
8108 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8110 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
8111 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
8112 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
8113 -maltivec=be with LE targets.
8114 (vsx_mergeh_<mode>): Likewise.
8115 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
8116 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
8117 (altivec_vmrghb): Replace with define_expand and new
8118 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
8119 (altivec_vmrghb_direct): New define_insn.
8120 (altivec_vmrghh): Replace with define_expand and new
8121 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
8122 (altivec_vmrghh_direct): New define_insn.
8123 (altivec_vmrghw): Replace with define_expand and new
8124 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
8125 (altivec_vmrghw_direct): New define_insn.
8126 (*altivec_vmrghsf): Adjust for endianness.
8127 (altivec_vmrglb): Replace with define_expand and new
8128 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
8129 (altivec_vmrglb_direct): New define_insn.
8130 (altivec_vmrglh): Replace with define_expand and new
8131 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
8132 (altivec_vmrglh_direct): New define_insn.
8133 (altivec_vmrglw): Replace with define_expand and new
8134 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
8135 (altivec_vmrglw_direct): New define_insn.
8136 (*altivec_vmrglsf): Adjust for endianness.
8137 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8138 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8139 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8140 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8141 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8142 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8143 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8144 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8146 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
8148 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
8149 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
8152 2014-01-29 Richard Biener <rguenther@suse.de>
8154 PR tree-optimization/58742
8155 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
8156 associate_pointerplus_align.
8157 (associate_pointerplus_diff): New function.
8158 (associate_pointerplus): Likewise. Call associate_pointerplus_align
8159 and associate_pointerplus_diff.
8161 2014-01-29 Richard Biener <rguenther@suse.de>
8163 * lto-streamer.h (LTO_major_version): Bump to 3.
8164 (LTO_minor_version): Reset to 0.
8166 2014-01-29 Renlin Li <Renlin.Li@arm.com>
8168 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
8169 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
8170 (arm_file_start): Generate correct asm header for armv7ve.
8171 * config/arm/bpabi.h: Add multilib support for armv7ve.
8172 * config/arm/driver-arm.c: Change the architectures of cortex-a7
8173 and cortex-a15 to armv7ve.
8174 * config/arm/t-aprofile: Add multilib support for armv7ve.
8175 * doc/invoke.texi: Document -march=armv7ve.
8177 2014-01-29 Richard Biener <rguenther@suse.de>
8179 PR tree-optimization/58742
8180 * tree-ssa-forwprop.c (associate_plusminus): Return true
8181 if we changed sth, defer EH cleanup to ...
8182 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
8183 (simplify_mult): New function.
8185 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8188 PR tree-optimization/59920
8189 * tree.c (build_common_builtin_nodes): Remove
8190 __builtin_setjmp_dispatcher initialization.
8191 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
8192 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
8193 instead of gsi_after_labels + manually skipping debug stmts.
8194 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
8195 ignore bbs with IFN_ABNORMAL_DISPATCHER.
8196 * tree-inline.c (copy_edges_for_bb): Remove
8197 can_make_abnormal_goto argument, instead add abnormal_goto_dest
8198 argument. Ignore computed_goto_p stmts. Don't call
8199 make_abnormal_goto_edges. If a call might need abnormal edges
8200 for non-local gotos, see if it already has an edge to
8201 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
8202 with true argument, don't do anything then, otherwise add
8203 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
8204 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
8206 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
8207 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
8208 (lower_stmt): Don't set data->calls_builtin_setjmp.
8209 (lower_builtin_setjmp): Adjust comment.
8210 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
8211 * tree-cfg.c (found_computed_goto): Remove.
8212 (factor_computed_gotos): Remove.
8213 (make_goto_expr_edges): Return bool, true for computed gotos.
8214 Don't call make_abnormal_goto_edges.
8215 (build_gimple_cfg): Don't set found_computed_goto, don't call
8216 factor_computed_gotos.
8217 (computed_goto_p): No longer static.
8218 (make_blocks): Don't set found_computed_goto.
8219 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
8220 (make_edges): If make_goto_expr_edges returns true, push bb
8221 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
8222 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
8223 vector. Record mapping between bbs and OpenMP regions if there
8224 are any, adjust make_gimple_omp_edges caller. Call
8225 handle_abnormal_edges.
8226 (make_abnormal_goto_edges): Remove.
8227 * tree-cfg.h (make_abnormal_goto_edges): Remove.
8228 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
8229 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
8230 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
8231 * internal-fn.def (ABNORMAL_DISPATCHER): New.
8232 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
8233 filling *region also set *region_idx to (*region)->entry->index.
8236 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
8237 For REGs set ORIGINAL_REGNO.
8239 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
8241 * doc/md.texi: Mention that a target shouldn't implement
8242 vec_widen_(s|u)mul_even/odd pair if it is less efficient
8245 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8247 PR tree-optimization/59594
8248 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
8249 a copy of the datarefs vector rather than the vector itself.
8251 2014-01-28 Jason Merrill <jason@redhat.com>
8254 * dwarf2out.c (auto_die): New static.
8255 (gen_type_die_with_usage): Handle C++1y 'auto'.
8256 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
8259 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
8262 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
8263 (SPEC_X32): Likewise.
8264 (SPEC_64): Likewise.
8265 * config/i386/i386.c (ix86_option_override_internal): Turn off
8266 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
8268 (x86_file_start): Output .code16gcc for TARGET_16BIT.
8269 * config/i386/i386.h (TARGET_16BIT): New macro.
8270 (TARGET_16BIT_P): Likewise.
8271 * config/i386/i386.opt: Add m16.
8272 * doc/invoke.texi: Document -m16.
8274 2014-01-28 Jakub Jelinek <jakub@redhat.com>
8276 PR preprocessor/59935
8277 * input.c (location_get_source_line): Bail out on when line number
8278 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
8280 2014-01-28 Richard Biener <rguenther@suse.de>
8282 PR tree-optimization/58742
8283 * tree-ssa-forwprop.c (associate_plusminus): Handle
8284 pointer subtraction of the form (T)(P + A) - (T)P.
8286 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8288 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
8291 2014-01-28 Richard Biener <rguenther@suse.de>
8294 2014-01-28 Richard Biener <rguenther@suse.de>
8296 PR rtl-optimization/45364
8297 PR rtl-optimization/59890
8298 * var-tracking.c (local_get_addr_clear_given_value): Handle
8299 already cleared slot.
8300 (val_reset): Handle not allocated local_get_addr_cache.
8301 (vt_find_locations): Use post-order on the inverted CFG.
8303 2014-01-28 Richard Biener <rguenther@suse.de>
8305 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
8307 2014-01-28 Richard Biener <rguenther@suse.de>
8309 PR rtl-optimization/45364
8310 PR rtl-optimization/59890
8311 * var-tracking.c (local_get_addr_clear_given_value): Handle
8312 already cleared slot.
8313 (val_reset): Handle not allocated local_get_addr_cache.
8314 (vt_find_locations): Use post-order on the inverted CFG.
8316 2014-01-28 Alan Modra <amodra@gmail.com>
8318 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
8319 * configure.ac <recursive call for build != host>: Define
8320 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
8321 and LD_FOR_BUILD too.
8322 * configure: Regenerate.
8324 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
8326 * config/i386/i386.c (get_builtin_code_for_version): Separate
8327 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
8328 Broadwell from Haswell.
8330 2014-01-27 Steve Ellcey <sellcey@mips.com>
8332 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
8333 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
8334 * config/mips/mips.c (mips_option_override): Change setting
8336 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
8337 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
8338 Change from Mask to Var.
8340 2014-01-27 Jeff Law <law@redhat.com>
8342 * ipa-inline.c (inline_small_functions): Fix typo.
8344 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
8346 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
8347 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
8348 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
8349 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
8350 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
8351 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
8352 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
8353 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
8354 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
8355 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
8356 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
8357 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
8358 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
8359 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
8360 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
8361 (_mm512_storeu_epi64): Ditto.
8362 (_mm512_cmpge_epi32_mask): Ditto.
8363 (_mm512_cmpge_epu32_mask): Ditto.
8364 (_mm512_cmpge_epi64_mask): Ditto.
8365 (_mm512_cmpge_epu64_mask): Ditto.
8366 (_mm512_cmple_epi32_mask): Ditto.
8367 (_mm512_cmple_epu32_mask): Ditto.
8368 (_mm512_cmple_epi64_mask): Ditto.
8369 (_mm512_cmple_epu64_mask): Ditto.
8370 (_mm512_cmplt_epi32_mask): Ditto.
8371 (_mm512_cmplt_epu32_mask): Ditto.
8372 (_mm512_cmplt_epi64_mask): Ditto.
8373 (_mm512_cmplt_epu64_mask): Ditto.
8374 (_mm512_cmpneq_epi32_mask): Ditto.
8375 (_mm512_cmpneq_epu32_mask): Ditto.
8376 (_mm512_cmpneq_epi64_mask): Ditto.
8377 (_mm512_cmpneq_epu64_mask): Ditto.
8378 (_mm512_expand_pd): Ditto.
8379 (_mm512_expand_ps): Ditto.
8380 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
8381 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
8382 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
8383 * config/i386/i386.c (ix86_builtins): Add
8384 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
8385 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
8386 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
8387 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
8388 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
8389 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
8390 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
8391 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
8392 IX86_BUILTIN_PMOVUSQW512_MEM.
8393 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
8394 __builtin_ia32_pmovsqd512mem_mask,
8395 __builtin_ia32_pmovqd512mem_mask,
8396 __builtin_ia32_pmovusqw512mem_mask,
8397 __builtin_ia32_pmovsqw512mem_mask,
8398 __builtin_ia32_pmovqw512mem_mask,
8399 __builtin_ia32_pmovusdw512mem_mask,
8400 __builtin_ia32_pmovsdw512mem_mask,
8401 __builtin_ia32_pmovdw512mem_mask,
8402 __builtin_ia32_pmovqb512mem_mask,
8403 __builtin_ia32_pmovusqb512mem_mask,
8404 __builtin_ia32_pmovsqb512mem_mask,
8405 __builtin_ia32_pmovusdb512mem_mask,
8406 __builtin_ia32_pmovsdb512mem_mask,
8407 __builtin_ia32_pmovdb512mem_mask.
8408 (bdesc_args): Add __builtin_ia32_expanddf512,
8409 __builtin_ia32_expandsf512.
8410 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
8411 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
8412 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
8413 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
8414 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
8415 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
8416 (avx512f_<code>v8div16qi2_mask_store): This.
8417 (avx512f_expand<mode>): New.
8419 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
8421 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
8423 (_mm512_mask_prefetch_i64gather_pd): Ditto.
8424 (_mm512_prefetch_i32scatter_pd): Ditto.
8425 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
8426 (_mm512_prefetch_i64scatter_pd): Ditto.
8427 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
8428 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
8429 (_mm512_mask_prefetch_i64gather_ps): Ditto.
8430 (_mm512_prefetch_i32scatter_ps): Ditto.
8431 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
8432 (_mm512_prefetch_i64scatter_ps): Ditto.
8433 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
8434 * config/i386/i386-builtin-types.def: Define
8435 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
8436 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
8437 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
8438 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
8439 IX86_BUILTIN_SCATTERPFQPD.
8440 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
8441 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
8442 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
8443 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
8444 __builtin_ia32_scatterpfqps.
8445 (ix86_expand_builtin): Expand new built-ins.
8446 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8447 fix memory access data type.
8448 (*avx512pf_gatherpf<mode>_mask): Ditto.
8449 (*avx512pf_gatherpf<mode>): Ditto.
8450 (avx512pf_scatterpf<mode>): Ditto.
8451 (*avx512pf_scatterpf<mode>_mask): Ditto.
8452 (*avx512pf_scatterpf<mode>): Ditto.
8453 (GATHER_SCATTER_SF_MEM_MODE): New.
8454 (avx512pf_gatherpf<mode>df): Ditto.
8455 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8456 (*avx512pf_scatterpf<mode>df): Ditto.
8458 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8461 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8462 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8465 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8467 * common/config/arm/arm-common.c
8468 (arm_rewrite_mcpu): Handle multiple names.
8470 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8472 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8474 * gimple-builder.h (create_gimple_tmp): Delete.
8476 2014-01-27 Christian Bruel <christian.bruel@st.com>
8478 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
8481 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
8483 * config/pa/pa.md (call): Generate indirect long calls to non-local
8484 functions when outputing 32-bit code.
8485 (call_value): Likewise except for special call to buggy powf function.
8487 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
8488 portable runtime and PIC indirect calls.
8489 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
8490 and PIC call sequences. Use ldo instead of blr to set return register
8491 in PIC call sequence.
8493 2014-01-25 Walter Lee <walt@tilera.com>
8495 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
8496 avoid clobbering a live register.
8498 2014-01-25 Walter Lee <walt@tilera.com>
8500 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
8501 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
8502 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
8503 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
8505 2014-01-25 Walter Lee <walt@tilera.com>
8507 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
8508 arguments on even registers.
8509 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
8511 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
8512 (BIGGEST_ALIGNMENT): Ditto.
8513 (BIGGEST_FIELD_ALIGNMENT): Ditto.
8515 2014-01-25 Walter Lee <walt@tilera.com>
8517 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
8518 insns before bundling.
8519 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8521 2014-01-25 Walter Lee <walt@tilera.com>
8523 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
8524 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
8525 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
8527 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8529 * config/mips/constraints.md (kl): Delete.
8530 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
8531 define expands, using...
8532 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
8533 instructions for MIPS16.
8534 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
8535 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
8537 2014-01-25 Walter Lee <walt@tilera.com>
8539 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
8543 2014-01-25 Walter Lee <walt@tilera.com>
8545 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
8546 (TARGET_EXPAND_TO_RTL_HOOK): Define.
8548 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8550 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
8553 2014-01-25 Jakub Jelinek <jakub@redhat.com>
8555 * print-rtl.c (in_call_function_usage): New var.
8556 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
8557 EXPR_LIST mode as mode and not as reg note name.
8560 * cfgloopmanip.c (copy_loop_info): If
8561 loop->warned_aggressive_loop_optimizations, make sure
8562 the flag is set in target loop too.
8564 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
8566 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
8568 * builtins.def: Likewise.
8569 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
8570 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
8571 * ira.c (ira_setup_eliminable_regset): Likewise.
8572 * omp-low.c (gate_expand_omp): Likewise.
8573 (execute_lower_omp): Likewise.
8574 (diagnose_sb_0): Likewise.
8575 (gate_diagnose_omp_blocks): Likewise.
8576 (simd_clone_clauses_extract): Likewise.
8579 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8581 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
8582 correction for little endian...
8583 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
8586 2014-01-24 Jeff Law <law@redhat.com>
8588 PR tree-optimization/59919
8589 * tree-vrp.c (find_assert_locations_1): Do not register asserts
8590 for non-returning calls.
8592 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
8594 * common/config/aarch64/aarch64-common.c
8595 (aarch64_rewrite_mcpu): Handle multiple names.
8596 * config/aarch64/aarch64.h
8597 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8599 2014-01-24 Dodji Seketeli <dodji@redhat.com>
8601 * input.c (add_file_to_cache_tab): Handle the case where fopen
8604 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
8607 * config/i386/i386.md (pushsf splitter): Get stack adjustment
8608 from push operand if code of push isn't PRE_DEC.
8610 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8613 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
8614 -mquad-memory-atomic. Update -mquad-memory documentation to say
8615 it is only used for non-atomic loads/stores.
8617 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
8618 -mquad-memory or -mquad-memory-atomic switches.
8620 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
8621 -mquad-memory-atomic to ISA 2.07 support.
8623 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
8624 to separate support of normal quad word memory operations (ldq, stq)
8625 from the atomic quad word memory operations.
8627 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
8628 support to separate non-atomic quad word operations from atomic
8629 quad word operations. Disable non-atomic quad word operations in
8630 little endian mode so that we don't have to swap words after the
8631 load and before the store.
8632 (quad_load_store_p): Add comment about atomic quad word support.
8633 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
8634 options printed with -mdebug=reg.
8636 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
8637 -mquad-memory-atomic as the test for whether we have quad word
8638 atomic instructions.
8639 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
8640 or -mp8-vector are used, allow byte/half-word atomic operations.
8642 * config/rs6000/sync.md (load_lockedti): Insure that the address
8643 is a proper indexed or indirect address for the lqarx instruction.
8644 On little endian systems, swap the hi/lo registers after the lqarx
8646 (load_lockedpti): Use indexed_or_indirect_operand predicate to
8647 insure the address is valid for the lqarx instruction.
8648 (store_conditionalti): Insure that the address is a proper indexed
8649 or indirect address for the stqcrx. instruction. On little endian
8650 systems, swap the hi/lo registers before doing the stqcrx.
8652 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
8653 insure the address is valid for the stqcrx. instruction.
8655 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8656 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
8657 type of quad memory support is available.
8659 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
8662 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
8663 there is a danger of looping.
8665 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
8667 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8668 force flag_ira_loop_pressure if set via command line.
8670 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8672 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
8673 (ashr_simd): New builtin handling DI mode.
8674 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
8675 (aarch64_sshr_simddi): New match pattern.
8676 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
8677 (vshrd_n_s64): Likewise.
8678 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
8680 2014-01-23 Nick Clifton <nickc@redhat.com>
8682 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
8683 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
8684 favour of mcu specific scripts.
8685 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
8688 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8689 Alex Velenko <Alex.Velenko@arm.com>
8691 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
8692 (vaddv_s16): Likewise.
8693 (vaddv_s32): Likewise.
8694 (vaddv_u8): Likewise.
8695 (vaddv_u16): Likewise.
8696 (vaddv_u32): Likewise.
8697 (vaddvq_s8): Likewise.
8698 (vaddvq_s16): Likewise.
8699 (vaddvq_s32): Likewise.
8700 (vaddvq_s64): Likewise.
8701 (vaddvq_u8): Likewise.
8702 (vaddvq_u16): Likewise.
8703 (vaddvq_u32): Likewise.
8704 (vaddvq_u64): Likewise.
8705 (vaddv_f32): Likewise.
8706 (vaddvq_f32): Likewise.
8707 (vaddvq_f64): Likewise.
8708 (vmaxv_f32): Likewise.
8709 (vmaxv_s8): Likewise.
8710 (vmaxv_s16): Likewise.
8711 (vmaxv_s32): Likewise.
8712 (vmaxv_u8): Likewise.
8713 (vmaxv_u16): Likewise.
8714 (vmaxv_u32): Likewise.
8715 (vmaxvq_f32): Likewise.
8716 (vmaxvq_f64): Likewise.
8717 (vmaxvq_s8): Likewise.
8718 (vmaxvq_s16): Likewise.
8719 (vmaxvq_s32): Likewise.
8720 (vmaxvq_u8): Likewise.
8721 (vmaxvq_u16): Likewise.
8722 (vmaxvq_u32): Likewise.
8723 (vmaxnmv_f32): Likewise.
8724 (vmaxnmvq_f32): Likewise.
8725 (vmaxnmvq_f64): Likewise.
8726 (vminv_f32): Likewise.
8727 (vminv_s8): Likewise.
8728 (vminv_s16): Likewise.
8729 (vminv_s32): Likewise.
8730 (vminv_u8): Likewise.
8731 (vminv_u16): Likewise.
8732 (vminv_u32): Likewise.
8733 (vminvq_f32): Likewise.
8734 (vminvq_f64): Likewise.
8735 (vminvq_s8): Likewise.
8736 (vminvq_s16): Likewise.
8737 (vminvq_s32): Likewise.
8738 (vminvq_u8): Likewise.
8739 (vminvq_u16): Likewise.
8740 (vminvq_u32): Likewise.
8741 (vminnmv_f32): Likewise.
8742 (vminnmvq_f32): Likewise.
8743 (vminnmvq_f64): Likewise.
8745 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8747 * config/aarch64/aarch64-simd.md
8748 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
8749 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
8750 (*aarch64_mul3_elt<mode>): Likewise.
8751 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
8752 (*aarch64_mul3_elt_to_64v2df): Likewise.
8753 (*aarch64_mla_elt<mode>): Likewise.
8754 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8755 (*aarch64_mls_elt<mode>): Likewise.
8756 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8757 (*aarch64_fma4_elt<mode>): Likewise.
8758 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8759 (*aarch64_fma4_elt_to_64v2df): Likewise.
8760 (*aarch64_fnma4_elt<mode>): Likewise.
8761 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8762 (*aarch64_fnma4_elt_to_64v2df): Likewise.
8763 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
8764 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
8765 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8766 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8767 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
8768 (aarch64_sqdmull_lane<mode>_internal): Likewise.
8769 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
8771 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
8773 * config/aarch64/aarch64-simd.md
8774 (aarch64_be_checked_get_lane<mode>): New define_expand.
8775 * config/aarch64/aarch64-simd-builtins.def
8776 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
8777 New builtin definition.
8778 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
8779 Use new safe be builtin.
8781 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8783 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
8785 (aarch64_be_st1<mode>): Likewise.
8786 (aarch_ld1<VALL:mode>): Define_expand modified.
8787 (aarch_st1<VALL:mode>): Likewise.
8788 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
8789 (UNSPEC_ST1): Likewise.
8791 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
8793 * config/microblaze/microblaze.md: Add trap insn and attribute
8795 2014-01-23 Dodji Seketeli <dodji@redhat.com>
8797 PR preprocessor/58580
8798 * input.h (location_get_source_line): Take an additional line_size
8800 (void diagnostics_file_cache_fini): Declare new function.
8801 * input.c (struct fcache): New type.
8802 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
8803 New static constants.
8804 (diagnostic_file_cache_init, total_lines_num)
8805 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
8806 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
8807 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
8808 (get_next_line, read_next_line, goto_next_line, read_line_num):
8809 New static function definitions.
8810 (diagnostic_file_cache_fini): New function.
8811 (location_get_source_line): Take an additional output line_len
8812 parameter. Re-write using lookup_or_add_file_to_cache_tab and
8814 * diagnostic.c (diagnostic_finish): Call
8815 diagnostic_file_cache_fini.
8816 (adjust_line): Take an additional input parameter for the length
8817 of the line, rather than calculating it with strlen.
8818 (diagnostic_show_locus): Adjust the use of
8819 location_get_source_line and adjust_line with respect to their new
8820 signature. While displaying a line now, do not stop at the first
8821 null byte. Rather, display the zero byte as a space and keep
8822 going until we reach the size of the line.
8823 * Makefile.in: Add vec.o to OBJS-libcommon
8825 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8826 Ilya Tocar <ilya.tocar@intel.com>
8828 * config/i386/avx512fintrin.h (_mm512_kmov): New.
8829 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
8830 (__builtin_ia32_kmov16): Ditto.
8831 * config/i386/i386.md (UNSPEC_KMOV): New.
8834 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8836 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
8837 (_mm512_storeu_si512): Ditto.
8839 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
8842 * rtl.h (get_referenced_operands): Declare.
8843 * recog.c (get_referenced_operands): New function.
8844 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
8845 operands have been referenced when recording LO_SUM references.
8847 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
8849 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
8851 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8853 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
8854 Enable for generic and recent AMD targets.
8856 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8858 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
8859 ARG_SIZE note when adjustment was eliminated.
8861 2014-01-22 Jeff Law <law@redhat.com>
8863 PR tree-optimization/59597
8864 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
8865 in file. Accept new argument REGISTERING and use it to modify
8866 dump output appropriately.
8867 (register_jump_thread): Corresponding changes.
8868 (mark_threaded_blocks): Reinstate code to cancel unprofitable
8869 thread paths involving joiner blocks. Add code to dump cancelled
8870 jump threading paths.
8872 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
8874 PR rtl-optimization/59477
8875 * lra-constraints.c (inherit_in_ebb): Process call for living hard
8876 regs. Update reloads_num and potential_reload_hard_regs for all insns.
8878 2014-01-22 Tom Tromey <tromey@redhat.com>
8880 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
8882 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
8884 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8886 PR rtl-optimization/59896
8887 * lra-constraints.c (process_alt_operands): Check unused note for
8888 matched operands of insn with no output reloads.
8890 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
8892 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
8893 (mips_move_from_gpr_cost): Likewise.
8895 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8897 PR rtl-optimization/59858
8898 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
8899 ira_class_hard_regs_num.
8900 (process_alt_operands): Increase reject for dying matched operand.
8902 2014-01-21 Jakub Jelinek <jakub@redhat.com>
8905 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
8906 smaller than size, perform several stores or loads and stores
8907 at dst + count - size to store or copy all of size bytes, rather
8908 than just last modesize bytes.
8910 2014-01-20 DJ Delorie <dj@redhat.com>
8912 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
8913 that CLOBBERs are REGs before propogating their values.
8915 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
8918 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
8919 (cgraph_inline_failed_type): New function.
8920 * cgraph.h (DEFCIFCODE): Add type.
8921 (cgraph_inline_failed_type_t): New enum.
8922 (cgraph_inline_failed_type): New prototype.
8923 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
8924 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
8925 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
8926 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
8927 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
8928 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
8929 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
8930 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
8931 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
8932 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
8933 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
8934 OPTIMIZATION_MISMATCH.
8935 * tree-inline.c (expand_call_inline): Emit errors during
8936 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
8938 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8941 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
8942 mode attribute in insn output.
8944 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
8946 * output.h (output_constant): Delete.
8947 * varasm.c (output_constant): Make private.
8949 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
8951 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
8953 2014-01-20 Jakub Jelinek <jakub@redhat.com>
8956 * tree.h (fold_builtin_strcat): New prototype.
8957 * builtins.c (fold_builtin_strcat): No longer static. Add len
8958 argument, if non-NULL, don't call c_strlen. Optimize
8959 directly into __builtin_memcpy instead of __builtin_strcpy.
8960 (fold_builtin_2): Adjust fold_builtin_strcat caller.
8961 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
8963 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
8965 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
8966 for SImode_address_operand operands, having only a REG argument.
8968 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
8970 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
8971 loader name using mbig-endian.
8972 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
8974 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8976 * doc/invoke.texi (-march): Clarify documentation for AArch64.
8980 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
8982 * config/aarch64/aarch64-protos.h
8983 (aarch64_cannot_change_mode_class_ptr): Declare.
8984 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
8985 aarch64_cannot_change_mode_class_ptr): New.
8986 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
8987 backend hook aarch64_cannot_change_mode_class.
8989 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
8991 * common/config/aarch64/aarch64-common.c
8992 (aarch64_handle_option): Don't handle any option order logic here.
8993 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
8994 selected_cpu, warn on architecture version mismatch.
8995 (aarch64_override_options): Fix parsing order for option strings.
8997 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8998 Iain Sandoe <iain@codesourcery.com>
9001 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
9002 warning. Amend comment to reflect current functionality.
9004 2014-01-20 Richard Biener <rguenther@suse.de>
9007 * builtins.c (fold_builtin_strcat): Remove case better handled
9008 by tree-ssa-strlen.c.
9010 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
9012 * config/aarch64/aarch64.opt
9013 (mcpu, march, mtune): Make case-insensitive.
9015 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9018 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9019 if operands[1] is a REG or ZERO_EXTEND of a REG.
9021 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
9023 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
9025 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
9027 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
9028 long non-pic millicode calls.
9030 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9032 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
9034 2014-01-19 Kito Cheng <kito@0xlab.org>
9036 * builtins.c (expand_movstr): Check movstr expand done or fail.
9038 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9039 H.J. Lu <hongjiu.lu@intel.com>
9042 * config/i386/i386.md (*lea<mode>): Zero-extend return register
9043 to DImode for zero-extended addresses.
9045 2014-01-19 Jakub Jelinek <jakub@redhat.com>
9047 PR rtl-optimization/57763
9048 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
9049 on the new indirect jump_insn and increment LABEL_NUSES (label).
9051 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
9055 * config.gcc (x86_archs): New variable.
9056 (x86_64_archs): Likewise.
9057 (x86_cpus): Likewise.
9058 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
9059 --with-arch/--with-cpu= options.
9060 Support --with-arch=/--with-cpu={nehalem,westmere,
9061 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
9063 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9065 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
9066 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
9068 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9070 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
9072 2014-01-18 Jakub Jelinek <jakub@redhat.com>
9075 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
9076 clear cpp_get_options (parse_in)->warn_unused_macros for
9077 ix86_target_macros_internal with cpp_define.
9079 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
9081 * jump.c (delete_related_insns): Keep (use (insn))s.
9082 * reorg.c (redundant_insn): Check for barriers too.
9084 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9086 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
9088 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
9090 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
9091 call to $$dyncall when TARGET_LONG_CALLS is true.
9093 2014-01-17 Jeff Law <law@redhat.com>
9095 * ree.c (combine_set_extension): Temporarily disable test for
9096 changing number of hard registers.
9098 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9101 * ipa-inline-analysis.c (inline_free_summary):
9102 Do not free summary of aliases.
9104 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9107 * gimplify.c (gimplify_expr): Use create_tmp_var
9108 instead of create_tmp_var_raw. If cond doesn't have
9109 integral type, don't add the IFN_ANNOTATE builtin at all.
9111 2014-01-17 Martin Jambor <mjambor@suse.cz>
9114 * ipa-cp.c (prev_edge_clone): New variable.
9115 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
9116 Also resize prev_edge_clone vector.
9117 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
9118 (ipcp_edge_removal_hook): New function.
9119 (ipcp_driver): Register ipcp_edge_removal_hook.
9121 2014-01-17 Andrew Pinski <apinski@cavium.com>
9122 Steve Ellcey <sellcey@mips.com>
9125 * config/mips/mips.c (mips_print_operand): Check operand mode instead
9128 2014-01-17 Jeff Law <law@redhat.com>
9131 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
9132 so that pass_ccp runs first.
9134 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9136 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
9137 (ix86_adjust_cost): Use !TARGET_XXX.
9138 (do_reorder_for_imul): Likewise.
9139 (swap_top_of_ready_list): Likewise.
9140 (ix86_sched_reorder): Likewise.
9142 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9144 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9145 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
9146 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
9147 (intel_memset): New. Duplicate slm_memset.
9148 (intel_cost): New. Duplicate slm_cost.
9149 (m_INTEL): New macro.
9150 (processor_target_table): Add "intel".
9151 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
9152 with PROCESSOR_INTEL for "intel".
9153 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
9154 PROCESSOR_SILVERMONT.
9155 (ix86_issue_rate): Likewise.
9156 (ix86_adjust_cost): Likewise.
9157 (ia32_multipass_dfa_lookahead): Likewise.
9158 (swap_top_of_ready_list): Likewise.
9159 (ix86_sched_reorder): Likewise.
9160 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
9161 instead of TARGET_OPT_AGU.
9162 * config/i386/i386.h (TARGET_INTEL): New.
9163 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
9164 (processor_type): Add PROCESSOR_INTEL.
9165 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
9166 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
9168 2014-01-17 Marek Polacek <polacek@redhat.com>
9171 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
9174 2014-01-17 Richard Biener <rguenther@suse.de>
9176 PR tree-optimization/46590
9177 * opts.c (default_options_table): Add entries for
9178 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
9179 all enabled at -O1 but not for -Og.
9180 * common.opt (fbranch-count-reg): Remove Init(1).
9181 (fmove-loop-invariants): Likewise.
9182 (ftree-pta): Likewise.
9184 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9186 * config/i386/i386.c (ix86_data_alignment): For compatibility with
9187 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
9188 decls to at least the GCC 4.8 used alignments.
9191 * tree-nested.c (convert_nonlocal_reference_stmt,
9192 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
9193 of GIMPLE_BIND stmts, adjust associated decls.
9195 2014-01-17 Richard Biener <rguenther@suse.de>
9197 PR tree-optimization/46590
9198 * vec.h (vec<>::bseach): New member function implementing
9199 binary search according to C89 bsearch.
9200 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
9201 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
9202 bitmap pointer again. Make accesses_in_loop a flat array.
9203 (mem_ref_obstack): New global.
9204 (outermost_indep_loop): Adjust for mem_ref->stored changes.
9205 (mark_ref_stored): Likewise.
9206 (ref_indep_loop_p_2): Likewise.
9207 (set_ref_stored_in_loop): New helper function.
9208 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
9209 (memref_free): Adjust.
9210 (record_mem_ref_loc): Simplify.
9211 (gather_mem_refs_stmt): Adjust.
9212 (sort_locs_in_loop_postorder_cmp): New function.
9213 (analyze_memory_references): Sort accesses_in_loop after
9214 loop postorder number.
9215 (find_ref_loc_in_loop_cmp): New function.
9216 (for_all_locs_in_loop): Find relevant cluster of locs in
9217 accesses_in_loop and iterate without recursion.
9218 (execute_sm): Avoid uninit warning.
9219 (struct ref_always_accessed): Simplify.
9220 (ref_always_accessed::operator ()): Likewise.
9221 (ref_always_accessed_p): Likewise.
9222 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
9223 loop postorder numbers here.
9224 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
9227 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9230 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
9231 on decls for which assemble_alias has been called.
9233 2014-01-17 Nick Clifton <nickc@redhat.com>
9235 * config/msp430/msp430.opt: (mcpu): New option.
9236 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
9237 (msp430_option_override): Parse target_cpu. If the MCU name
9238 matches a generic string, clear target_mcu.
9239 (msp430_attr): Allow numeric interrupt values up to 63.
9240 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
9241 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
9243 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
9245 * config/msp430/msp430.md (popm): Use %J rather than %I.
9246 (addsi3): Use msp430_nonimmediate_operand for operand 2.
9247 (addhi_cy_i): Use immediate_operand for operand 2.
9248 * doc/invoke.texi: Document -mcpu option.
9250 2014-01-17 Richard Biener <rguenther@suse.de>
9252 PR rtl-optimization/38518
9253 * df.h (df_analyze_loop): Declare.
9254 * df-core.c: Include cfgloop.h.
9255 (df_analyze_1): Split out main part of df_analyze.
9256 (df_analyze): Adjust.
9257 (loop_inverted_post_order_compute): New function.
9258 (loop_post_order_compute): Likewise.
9259 (df_analyze_loop): New function avoiding whole-function
9261 * loop-invariant.c (find_defs): Use df_analyze_loop.
9262 (find_invariants): Adjust.
9263 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
9265 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
9267 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
9268 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
9270 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
9272 * ipa-ref.c (ipa_remove_stmt_references): Fix references
9273 traversal when removing references.
9275 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
9278 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
9280 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
9283 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
9284 pushing a reload for an autoinc when we had previously reloaded an
9285 inner part of the address.
9287 2014-01-16 Jakub Jelinek <jakub@redhat.com>
9289 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
9291 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
9292 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
9293 when not giving up or versioning for alias only because of
9295 (vect_analyze_data_ref_dependences): Set to true.
9296 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
9298 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
9299 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
9303 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
9306 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
9307 operand 0 predicate for gathers, use a new pseudo as subtarget.
9309 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9312 * lra-constraints.c (process_alt_operands): Add printing debug info.
9313 Check absence of input/output reloads for matched operands too.
9315 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9317 PR rtl-optimization/59835
9318 * ira.c (ira_init_register_move_cost): Increase cost for
9321 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
9323 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
9325 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
9328 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
9329 non-register objects. Use gen_(high/low)part more consistently.
9332 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9335 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
9336 endian support, remove tests for WORDS_BIG_ENDIAN.
9337 (p8_mfvsrd_3_<mode>): Likewise.
9338 (reload_gpr_from_vsx<mode>): Likewise.
9339 (reload_gpr_from_vsxsf): Likewise.
9340 (p8_mfvsrd_4_disf): Likewise.
9342 2014-01-16 Richard Biener <rguenther@suse.de>
9344 PR rtl-optimization/46590
9345 * lcm.c (compute_antinout_edge): Use postorder iteration.
9346 (compute_laterin): Use inverted postorder iteration.
9348 2014-01-16 Nick Clifton <nickc@redhat.com>
9351 * varasm.c (output_constant): Return the number of bytes actually
9353 (output_constructor_array_range): Update the field size with the
9354 number of bytes emitted by output_constant.
9355 (output_constructor_regular_field): Likewise. Also do not
9356 complain if the total number of bytes emitted is now greater
9357 than the expected fieldpos.
9358 * output.h (output_constant): Update prototype and descriptive comment.
9360 2014-01-16 Marek Polacek <polacek@redhat.com>
9363 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
9364 it is error_mark_node.
9366 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
9368 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
9369 VALID_AVX256_REG_OR_OI_MODE.
9371 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
9373 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
9374 current procedure should be profiled.
9376 2014-01-15 Andrew Pinski <apinski@cavium.com>
9378 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
9379 of moving from/to the STACK_REG register class.
9381 2014-01-15 Richard Henderson <rth@redhat.com>
9384 * reginfo.c (global_regs_decl): Globalize.
9385 * rtl.h (global_regs_decl): Declare.
9386 * ira.c (do_reload): Diagnose frame_pointer_needed and it
9387 reserved via global_regs.
9389 2014-01-15 Teresa Johnson <tejohnson@google.com>
9391 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
9393 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9395 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
9396 and vmulosh rather than call gen_vec_widen_smult_*.
9397 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
9398 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
9399 (vec_widen_smult_even_v16qi): Likewise.
9400 (vec_widen_umult_even_v8hi): Likewise.
9401 (vec_widen_smult_even_v8hi): Likewise.
9402 (vec_widen_umult_odd_v16qi): Likewise.
9403 (vec_widen_smult_odd_v16qi): Likewise.
9404 (vec_widen_umult_odd_v8hi): Likewise.
9405 (vec_widen_smult_odd_v8hi): Likewise.
9406 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
9407 vmuloub rather than call gen_vec_widen_umult_*.
9408 (vec_widen_umult_lo_v16qi): Likewise.
9409 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
9410 vmulosb rather than call gen_vec_widen_smult_*.
9411 (vec_widen_smult_lo_v16qi): Likewise.
9412 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
9413 rather than call gen_vec_widen_umult_*.
9414 (vec_widen_umult_lo_v8hi): Likewise.
9415 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
9416 rather than call gen_vec_widen_smult_*.
9417 (vec_widen_smult_lo_v8hi): Likewise.
9419 2014-01-15 Jeff Law <law@redhat.com>
9421 PR tree-optimization/59747
9422 * ree.c (find_and_remove_re): Properly handle case where a second
9423 eliminated extension requires widening a copy created for elimination
9424 of a prior extension.
9425 (combine_set_extension): Ensure that the number of hard regs needed
9426 for a destination register does not change when we widen it.
9428 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
9430 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
9431 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
9432 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
9433 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
9434 (avr-*-rtems*): Likewise.
9435 (bfin*-rtems*): Likewise.
9436 (moxie-*-rtems*): Likewise.
9437 (h8300-*-rtems*): Likewise.
9438 (i[34567]86-*-rtems*): Likewise.
9439 (lm32-*-rtems*): Likewise.
9440 (m32r-*-rtems*): Likewise.
9441 (m68k-*-rtems*): Likewise.
9442 (microblaze*-*-rtems*): Likewise.
9443 (mips*-*-rtems*): Likewise.
9444 (powerpc-*-rtems*): Likewise.
9445 (sh-*-rtems*): Likewise.
9446 (sparc-*-rtems*): Likewise.
9447 (sparc64-*-rtems*): Likewise.
9448 (v850-*-rtems*): Likewise.
9449 (m32c-*-rtems*): Likewise.
9451 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9453 PR rtl-optimization/59511
9454 * ira.c (ira_init_register_move_cost): Use memory costs for some
9455 cases of register move cost calculations.
9456 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9457 instead of BB frequency.
9458 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9459 * lra-assigns.c (find_hard_regno_for): Ditto.
9461 2014-01-15 Richard Biener <rguenther@suse.de>
9463 PR tree-optimization/59822
9464 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9465 (vectorizable_load): Use it to hoist defs of uses of invariant
9466 loads out of the loop.
9468 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9469 Kugan Vivekanandarajah <kuganv@linaro.org>
9472 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9475 2014-01-15 Richard Biener <rguenther@suse.de>
9477 PR rtl-optimization/59802
9478 * lcm.c (compute_available): Use inverted postorder to seed
9479 the initial worklist.
9481 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9484 * config/s390/s390.c (s390_preferred_reload_class): Don't return
9485 ADDR_REGS for invalid symrefs in non-PIC code.
9487 2014-01-15 Jakub Jelinek <jakub@redhat.com>
9490 * builtins.c (determine_block_size): Initialize *probable_max_size
9491 even if len_rtx is CONST_INT.
9493 2014-01-14 Andrew Pinski <apinski@cavium.com>
9495 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
9496 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
9497 (cortexa53_tunings): Likewise.
9498 (aarch64_sched_issue_rate): New function.
9499 (TARGET_SCHED_ISSUE_RATE): Define.
9501 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9503 * ira-costs.c (find_costs_and_classes): Add missed
9504 ira_init_register_move_cost_if_necessary.
9506 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9509 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
9511 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
9514 * config/i386/i386.c (type_natural_mode): Add a bool parameter
9515 to indicate if type is used for function return value. Warn ABI
9516 change if the vector mode isn't available for function return value.
9517 (ix86_function_arg_advance): Pass false to type_natural_mode.
9518 (ix86_function_arg): Likewise.
9519 (ix86_gimplify_va_arg): Likewise.
9520 (function_arg_32): Don't warn ABI change.
9521 (ix86_function_value): Pass true to type_natural_mode.
9522 (ix86_return_in_memory): Likewise.
9523 (ix86_struct_value_rtx): Removed.
9524 (TARGET_STRUCT_VALUE_RTX): Likewise.
9526 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9528 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
9529 converting a conditional jump into a conditional return.
9531 2014-01-14 Richard Biener <rguenther@suse.de>
9533 PR tree-optimization/58921
9534 PR tree-optimization/59006
9535 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
9536 hoisting invariant stmts.
9537 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
9538 invariant loads on the preheader edge if possible.
9540 2014-01-14 Joey Ye <joey.ye@arm.com>
9542 * doc/plugin.texi (Building GCC plugins): Update to C++.
9544 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
9546 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
9547 (_mm_rcp28_round_ss): Ditto.
9548 (_mm_rsqrt28_round_sd): Ditto.
9549 (_mm_rsqrt28_round_ss): Ditto.
9550 (_mm_rcp28_sd): Ditto.
9551 (_mm_rcp28_ss): Ditto.
9552 (_mm_rsqrt28_sd): Ditto.
9553 (_mm_rsqrt28_ss): Ditto.
9554 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
9555 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
9556 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
9557 (IX86_BUILTIN_RCP28SD): Ditto.
9558 (IX86_BUILTIN_RCP28SS): Ditto.
9559 (IX86_BUILTIN_RSQRT28SD): Ditto.
9560 (IX86_BUILTIN_RSQRT28SS): Ditto.
9561 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
9562 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
9563 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
9564 (ix86_expand_special_args_builtin): Expand new FTYPE.
9565 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
9566 (srcp14<mode>): Make insn unary.
9567 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
9568 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
9569 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
9570 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
9571 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
9572 Fix rounding: make it SAE only.
9573 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9575 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9577 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
9578 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
9579 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
9580 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
9581 (round_saeonly_mask_scalar_operand4): Ditto.
9582 (round_saeonly_mask_scalar_op3): Ditto.
9583 (round_saeonly_mask_scalar_op4): Ditto.
9585 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9587 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9588 Implement -maltivec=be for vec_insert and vec_extract.
9590 2014-01-10 DJ Delorie <dj@redhat.com>
9592 * config/msp430/msp430.md (call_internal): Don't allow memory
9593 references with SP as the base register.
9594 (call_value_internal): Likewise.
9595 * config/msp430/constraints.md (Yc): New. For memory references
9596 that don't use SP as a base register.
9598 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
9599 "an integer without a # prefix"
9600 * config/msp430/msp430.md (epilogue_helper): Use it.
9602 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9605 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
9606 AVX512F gather builtins.
9607 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
9608 on gather decls with INTEGER_TYPE masktype.
9609 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
9610 directly into the builtin rather than hoisting it before loop.
9612 PR tree-optimization/59387
9613 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
9614 (scev_const_prop): If folded_casts and type has undefined overflow,
9615 use force_gimple_operand instead of force_gimple_operand_gsi and
9616 for each added stmt if it is assign with
9617 arith_code_with_undefined_signed_overflow, call
9618 rewrite_to_defined_overflow.
9619 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
9620 gimple-fold.h instead.
9621 (arith_code_with_undefined_signed_overflow,
9622 rewrite_to_defined_overflow): Moved to ...
9623 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
9624 rewrite_to_defined_overflow): ... here. No longer static.
9625 Include gimplify-me.h.
9626 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
9627 rewrite_to_defined_overflow): New prototypes.
9629 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9631 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
9633 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
9635 * builtins.c (get_object_alignment_2): Minor tweak.
9636 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
9638 2014-01-13 Christian Bruel <christian.bruel@st.com>
9640 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
9641 optimized non constant lengths.
9643 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9646 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
9647 load as __atomic_load_N if possible.
9649 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9651 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
9653 (rs6000_expand_builtin): Adjust call.
9655 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9658 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
9659 * config/rs6000/rs6000.c: Include target-globals.h.
9660 (rs6000_set_current_function): Instead of doing target_reinit
9661 unconditionally, use save_target_globals_default_opts and
9662 restore_target_globals.
9664 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
9666 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
9667 (rs6000_expand_builtin): Handle mffs and mtfsf.
9668 (rs6000_init_builtins): Define mffs and mtfsf.
9669 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
9670 (rs6000_mffs): New pattern.
9671 (rs6000_mtfsf): New pattern.
9673 2014-01-11 Bin Cheng <bin.cheng@arm.com>
9675 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
9676 Start narrowing with START. Apply candidate-use pair
9677 and check overall cost in narrowing.
9678 (iv_ca_prune): Pass new argument.
9680 2014-01-10 Jeff Law <law@redhat.com>
9683 * ree.c (combine_reaching_defs): Ensure the defining statement
9684 occurs before the extension when optimizing extensions with
9685 different source and destination hard registers.
9687 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9690 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
9691 vtables into the type inheritance graph.
9693 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9695 PR rtl-optimization/59754
9696 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
9697 modes in the REGNO != REGNO case.
9699 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9701 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
9703 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9705 PR tree-optimization/59745
9706 * tree-predcom.c (tree_predictive_commoning_loop): Call
9707 free_affine_expand_cache if giving up because components is NULL.
9709 * target-globals.c (save_target_globals): Allocate < 4KB structs using
9710 GC in payload of target_globals struct instead of allocating them on
9711 the heap and the larger structs separately using GC.
9712 * target-globals.h (struct target_globals): Make regs, hard_regs,
9713 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
9714 of GTY((skip)) and change type to void *.
9715 (reset_target_globals): Cast loads from those fields to corresponding
9718 2014-01-10 Steve Ellcey <sellcey@mips.com>
9721 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
9722 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
9723 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
9725 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
9728 * aarch64-modes.def (CC_Zmode): New flags mode.
9729 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
9730 represents an equality.
9731 (aarch64_get_condition_code): Handle CC_Zmode.
9732 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
9734 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9736 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
9737 extraction in good case.
9739 2014-01-10 Richard Biener <rguenther@suse.de>
9741 PR tree-optimization/59374
9742 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
9743 checking after SLP discovery. Mark stmts not participating
9744 in any SLP instance properly.
9746 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9748 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
9749 when handling a SET rtx.
9751 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9753 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
9754 (cortex-a57): Likewise.
9755 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
9757 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9759 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
9760 non-iwmmxt builtins.
9762 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9766 * ipa-devirt.c record_target_from_binfo): Take as argument
9767 stack of binfos and lookup matching one for virtual inheritance.
9768 (possible_polymorphic_call_targets_1): Update.
9770 2014-01-10 Huacai Chen <chenhc@lemote.com>
9772 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
9773 kernel strings for Loongson-2E/2F/3A.
9775 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9778 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
9779 is_gimple_call before calling gimple_call_internal_p.
9781 2014-01-09 Steve Ellcey <sellcey@mips.com>
9783 * Makefile.in (TREE_FLOW_H): Remove.
9784 (TREE_SSA_H): Add file names from tree-flow.h.
9785 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
9786 * tree.h: Remove tree-flow.h reference.
9787 * hash-table.h: Remove tree-flow.h reference.
9788 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
9789 reference with tree-ssa-loop.h.
9791 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9793 * doc/invoke.texi: Add -maltivec={be,le} options, and document
9794 default element-order behavior for -maltivec.
9795 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
9796 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
9797 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
9798 when targeting big endian, at least for now.
9799 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
9801 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9804 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
9805 var satisfies use_register_for_decl, just take into account type
9806 alignment, rather than decl alignment.
9808 PR tree-optimization/59622
9809 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
9810 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
9811 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
9812 Don't devirtualize for inplace at all. For targets.length () == 1,
9813 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
9815 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9817 * config/i386/i386.md (cpu): Remove the unused btver1.
9819 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9821 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
9823 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9826 * tree-core.h (struct target_globals): New forward declaration.
9827 (struct tree_target_option): Add globals field.
9828 * tree.h (TREE_TARGET_GLOBALS): Define.
9829 (prepare_target_option_nodes_for_pch): New prototype.
9830 * target-globals.h (struct target_globals): Define even if
9832 * tree.c (prepare_target_option_node_for_pch,
9833 prepare_target_option_nodes_for_pch): New functions.
9834 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
9835 * config/i386/i386.c: Include target-globals.h.
9836 (ix86_set_current_function): Instead of doing target_reinit
9837 unconditionally, use save_target_globals_default_opts and
9838 restore_target_globals.
9840 2014-01-09 Richard Biener <rguenther@suse.de>
9842 PR tree-optimization/59715
9843 * tree-cfg.h (split_critical_edges): Declare.
9844 * tree-cfg.c (split_critical_edges): Export.
9845 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
9847 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
9849 * cfgexpand.c (expand_stack_vars): Optionally disable
9850 asan stack protection.
9851 (expand_used_vars): Likewise.
9852 (partition_stack_vars): Likewise.
9853 * asan.c (asan_emit_stack_protection): Optionally disable
9854 after return stack usage.
9855 (instrument_derefs): Optionally disable memory access instrumentation.
9856 (instrument_builtin_call): Likewise.
9857 (instrument_strlen_call): Likewise.
9858 (asan_protect_global): Optionally disable global variables protection.
9859 * doc/invoke.texi: Added doc for new options.
9860 * params.def: Added new options.
9861 * params.h: Likewise.
9863 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9865 PR rtl-optimization/59724
9866 * ifcvt.c (cond_exec_process_if_block): Don't call
9867 flow_find_head_matching_sequence with 0 longest_match.
9868 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
9869 non-active insns if !stop_after.
9870 (try_head_merge_bb): Revert 2014-01-07 changes.
9872 2014-01-08 Jeff Law <law@redhat.com>
9874 * ree.c (get_sub_rtx): New function, extracted from...
9875 (merge_def_and_ext): Here.
9876 (combine_reaching_defs): Use get_sub_rtx.
9878 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
9880 * cgraph.h (varpool_variable_node): Do not choke on null node.
9882 2014-01-08 Catherine Moore <clm@codesourcery.com>
9884 * config/mips/mips.md (simple_return): Attempt to use JRC
9886 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
9888 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9890 PR rtl-optimization/59137
9891 * reorg.c (steal_delay_list_from_target): Call update_block for
9893 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
9895 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9897 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
9898 two duplicate entries.
9900 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9903 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
9905 * config/mips/mips.c (mips_truncated_op_cost): New function.
9906 (mips_rtx_costs): Adjust test for BADDU.
9907 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
9909 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
9911 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
9912 (*baddu_si): ...this new pattern.
9914 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9917 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
9919 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9922 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
9924 (expand_expr, expand_normal): Adjust.
9925 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
9926 inner_reference_p. Use inner_reference_p to expand inner references.
9927 (store_expr): Adjust.
9928 * cfgexpand.c (expand_call_stmt): Adjust.
9930 2014-01-08 Rong Xu <xur@google.com>
9932 * gcov-io.c (gcov_var): Move from gcov-io.h.
9933 (gcov_position): Ditto.
9934 (gcov_is_error): Ditto.
9935 (gcov_rewrite): Ditto.
9936 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
9937 only part to libgcc/libgcov.h.
9939 2014-01-08 Marek Polacek <polacek@redhat.com>
9942 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
9944 2014-01-08 Marek Polacek <polacek@redhat.com>
9947 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
9949 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9951 PR rtl-optimization/59649
9952 * stor-layout.c (get_mode_bounds): For BImode return
9953 0 and STORE_FLAG_VALUE.
9955 2014-01-08 Richard Biener <rguenther@suse.de>
9958 * gimple.h (is_gimple_builtin_call): Remove.
9959 (gimple_builtin_call_types_compatible_p): New.
9960 (gimple_call_builtin_p): New overload.
9961 * gimple.c (is_gimple_builtin_call): Remove.
9962 (validate_call): Rename to ...
9963 (gimple_builtin_call_types_compatible_p): ... this and export. Also
9965 (validate_type): New static function.
9966 (gimple_call_builtin_p): New overload and adjust.
9967 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
9968 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
9969 (gimple_fold_stmt_to_constant_1): Likewise.
9970 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
9972 2014-01-08 Richard Biener <rguenther@suse.de>
9975 * gimplify.c (gimplify_expr): Gimplify register-register type
9976 VIEW_CONVERT_EXPRs to separate stmts.
9978 2014-01-07 Jeff Law <law@redhat.com>
9981 * ree.c (combine_set_extension): Handle case where source
9982 and destination registers in an extension insn are different.
9983 (combine_reaching_defs): Allow source and destination registers
9984 in extension to be different under limited circumstances.
9985 (add_removable_extension): Remove restriction that the
9986 source and destination registers in the extension are the same.
9987 (find_and_remove_re): Emit a copy from the extension's
9988 destination to its source after the defining insn if
9989 the source and destination registers are different.
9992 * ifcvt.c (merge_if_block): If we are merging a block with more than
9993 one successor with a block with no successors, remove any BARRIER
9994 after the second block.
9996 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
9998 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
10000 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
10003 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
10004 for 14-bit register offsets when INT14_OK_STRICT is false.
10006 2014-01-07 Roland Stigge <stigge@antcom.de>
10007 Michael Meissner <meissner@linux.vnet.ibm.com>
10010 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
10011 Only check TFmode for SPE constants. Don't check TImode or TDmode.
10013 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
10015 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
10018 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
10020 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
10021 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
10022 rtx is const0_rtx or not.
10024 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
10027 * target-globals.c (save_target_globals): Remove this_fn_optab
10029 * toplev.c: Include optabs.h.
10030 (target_reinit): Temporarily restore the global options if another
10031 set of options are in force.
10033 2014-01-07 Jakub Jelinek <jakub@redhat.com>
10035 PR rtl-optimization/58668
10036 * cfgcleanup.c (flow_find_cross_jump): Don't count
10037 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
10038 to determine what is counted.
10039 (flow_find_head_matching_sequence): Use active_insn_p to determine
10041 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
10043 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
10044 determine what is counted.
10046 PR tree-optimization/59643
10047 * tree-predcom.c (split_data_refs_to_components): If one dr is
10048 read and one write, determine_offset fails and the write isn't
10049 in the bad component, just put the read into the bad component.
10051 2014-01-07 Mike Stump <mikestump@comcast.net>
10052 Jakub Jelinek <jakub@redhat.com>
10055 * tree-core.h (struct tree_optimization_option): Change optabs
10056 type from unsigned char * to void *.
10057 * optabs.c (init_tree_optimization_optabs): Adjust
10058 TREE_OPTIMIZATION_OPTABS initialization.
10060 2014-01-06 Jakub Jelinek <jakub@redhat.com>
10063 * config/i386/i386.h (struct machine_function): Add
10064 no_drap_save_restore field.
10065 * config/i386/i386.c (ix86_save_reg): Use
10066 !cfun->machine->no_drap_save_restore instead of
10067 crtl->stack_realign_needed.
10068 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
10069 this function clears frame_pointer_needed. Set
10070 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
10071 and DRAP reg is needed.
10073 2014-01-06 Marek Polacek <polacek@redhat.com>
10076 * doc/implement-c.texi: Mention that other integer types are
10077 permitted as bit-field types in strictly conforming mode.
10079 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
10081 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
10082 is newly allocated.
10084 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
10086 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
10088 2014-01-06 Martin Jambor <mjambor@suse.cz>
10091 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
10093 * ipa-prop.c (ipa_print_node_params): Fix indentation.
10095 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
10099 * var-tracking.c (add_stores): Preserve the value of the source even if
10100 we don't record the store.
10102 2014-01-06 Terry Guo <terry.guo@arm.com>
10104 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
10106 2014-01-05 Iain Sandoe <iain@codesourcery.com>
10109 * config/darwin.c (darwin_function_section): Adjust return values to
10110 correspond to optimisation changes made in r206070.
10112 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
10114 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
10115 from prefetch_block tune setting.
10116 (nocona_cost): Correct size of prefetch block to 64.
10118 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
10120 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
10121 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
10122 used to save the static chain register in the computation of the offset
10123 from which the FP registers need to be restored.
10125 2014-01-04 Jakub Jelinek <jakub@redhat.com>
10127 PR tree-optimization/59519
10128 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
10129 ICE if get_current_def (current_new_name) is already non-NULL, as long
10130 as it is a phi result of some other phi in *new_exit_bb that has
10133 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
10134 or vmovdqu* for misaligned_operand.
10135 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
10136 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
10137 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
10138 aligned_mem for AVX512F masked aligned load and store builtins and for
10139 non-temporal moves.
10141 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
10143 PR tree-optimization/59651
10144 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
10145 Address range for negative step should be added by TYPE_SIZE_UNIT.
10147 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
10149 * config/m68k/m68k.c (handle_move_double): Handle pushes with
10150 overlapping registers also for registers other than the stack pointer.
10152 2014-01-03 Marek Polacek <polacek@redhat.com>
10155 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
10158 2014-01-03 Jakub Jelinek <jakub@redhat.com>
10161 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
10164 * config/i386/i386.md (MODE_SIZE): New mode attribute.
10165 (push splitter): Use <P:MODE_SIZE> instead of
10166 GET_MODE_SIZE (<P:MODE>mode).
10167 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
10168 (mov -1, reg peephole2): Likewise.
10169 * config/i386/sse.md (*mov<mode>_internal,
10170 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
10171 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
10172 *<code><mode>3, *andnot<mode>3<mask_name>,
10173 <mask_codefor><code><mode>3<mask_name>): Likewise.
10174 * config/i386/subst.md (mask_mode512bit_condition,
10175 sd_mask_mode512bit_condition): Likewise.
10177 2014-01-02 Xinliang David Li <davidxl@google.com>
10179 PR tree-optimization/59303
10180 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
10181 (dump_predicates): Better output format.
10182 (pred_equal_p): New function.
10183 (is_neq_relop_p): Ditto.
10184 (is_neq_zero_form_p): Ditto.
10185 (pred_expr_equal_p): Ditto.
10186 (pred_neg_p): Ditto.
10187 (simplify_pred): Ditto.
10188 (simplify_preds_2): Ditto.
10189 (simplify_preds_3): Ditto.
10190 (simplify_preds_4): Ditto.
10191 (simplify_preds): Ditto.
10192 (push_pred): Ditto.
10193 (push_to_worklist): Ditto.
10194 (get_pred_info_from_cmp): Ditto.
10195 (is_degenerated_phi): Ditto.
10196 (normalize_one_pred_1): Ditto.
10197 (normalize_one_pred): Ditto.
10198 (normalize_one_pred_chain): Ditto.
10199 (normalize_preds): Ditto.
10200 (normalize_cond_1): Remove function.
10201 (normalize_cond): Ditto.
10202 (is_gcond_subset_of): Ditto.
10203 (is_subset_of_any): Ditto.
10204 (is_or_set_subset_of): Ditto.
10205 (is_and_set_subset_of): Ditto.
10206 (is_norm_cond_subset_of): Ditto.
10207 (pred_chain_length_cmp): Ditto.
10208 (convert_control_dep_chain_into_preds): Type change.
10209 (find_predicates): Ditto.
10210 (find_def_preds): Ditto.
10211 (destroy_predicates_vecs): Ditto.
10212 (find_matching_predicates_in_rest_chains): Ditto.
10213 (use_pred_not_overlap_with_undef_path_pred): Ditto.
10214 (is_pred_expr_subset): Ditto.
10215 (is_pred_chain_subset_of): Ditto.
10216 (is_included_in): Ditto.
10217 (is_superset_of): Ditto.
10219 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10221 Update copyright years.
10223 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10225 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
10226 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
10227 config/arc/arc.md, config/arc/arc.opt,
10228 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
10229 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
10230 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
10231 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
10232 config/linux-protos.h, config/linux.c, config/winnt-c.c,
10233 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
10234 vtable-verify.c, vtable-verify.h: Use the standard form for the
10237 2014-01-02 Tobias Burnus <burnus@net-b.de>
10239 * gcc.c (process_command): Update copyright notice dates.
10240 * gcov-dump.c: Ditto.
10242 * doc/cpp.texi: Bump @copying's copyright year.
10243 * doc/cppinternals.texi: Ditto.
10244 * doc/gcc.texi: Ditto.
10245 * doc/gccint.texi: Ditto.
10246 * doc/gcov.texi: Ditto.
10247 * doc/install.texi: Ditto.
10248 * doc/invoke.texi: Ditto.
10250 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10252 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
10254 2014-01-01 Jakub Jelinek <jakub@redhat.com>
10256 * config/i386/sse.md (*mov<mode>_internal): Guard
10257 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
10259 PR rtl-optimization/59647
10260 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
10261 new_rtx into UNSIGNED_FLOAT rtxes.
10263 Copyright (C) 2014 Free Software Foundation, Inc.
10265 Copying and distribution of this file, with or without modification,
10266 are permitted in any medium without royalty provided the copyright
10267 notice and this notice are preserved.