1 2014-05-12 Mike Stump <mikestump@comcast.net>
4 * genattrtab.c (filename): Add.
5 (convert_set_attr_alternative): Improve error message.
6 (check_defs): Restore read_md_filename for error messages.
7 (gen_insn): Save filename.
9 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
11 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
12 -fno-local-ivars and -fivar-visibility.
13 * c-family/c.opt: Make -Wshadow also implicitly enable
16 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
18 * doc/tm.texi: Remove reference to deleted macro.
19 * doc/tm.texi.in: Likewise.
21 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
24 * config/avr/avr.c (avr_out_store_psi): Use correct constant
27 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
29 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
30 * config/arm/aout.h (License): Same.
31 * config/arm/bpabi.h (License): Same.
32 * config/arm/elf.h (License): Same.
33 * config/arm/linux-elf.h (License): Same.
34 * config/arm/linux-gas.h (License): Same.
35 * config/arm/netbsd-elf.h (License): Same.
36 * config/arm/uclinux-eabi.h (License): Same.
37 * config/arm/uclinux-elf.h (License): Same.
38 * config/arm/vxworks.h (License): Same.
40 2014-05-11 Jakub Jelinek <jakub@redhat.com>
42 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
43 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
44 number of operands to 3.
45 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
46 * tree-nested.c (convert_nonlocal_omp_clauses,
47 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
48 * gimplify.c (gimplify_scan_omp_clauses): Handle
49 OMP_CLAUSE_LINEAR_STMT.
50 * omp-low.c (lower_rec_input_clauses): Fix typo.
51 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
52 cast between Fortran boolean_type_node and C _Bool if
55 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
57 PR tree-optimization/61136
58 * wide-int.h (multiple_of_p): Define a version that doesn't return
60 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
61 integer_zerop/const_binop pair.
62 (multiple_of_p): Likewise, converting both operands to widest_int
65 2014-05-09 Teresa Johnson <tejohnson@google.com>
67 * cgraphunit.c (analyze_functions): Use correct dump file.
69 2014-05-09 Florian Weimer <fweimer@redhat.com>
71 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
73 (stack_protect_return_slot_p): New function.
74 (expand_used_vars): Call stack_protect_decl_p and
75 stack_protect_return_slot_p for -fstack-protector-strong.
77 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
80 * fold-const.c (fold_binary_loc): Changed width of mask.
82 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
84 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
85 unsigned int initializers for regno_in, regno_out.
87 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
90 * config/avr/avr.md (cc): Add new attribute set_vzn.
91 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
92 Set cc insn attribute to set_vzn instead of set_zn for alternatives
94 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
95 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
96 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
98 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
101 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
103 * wide-int.cc (UTItype): Define.
104 (UDWtype): Define for appropriate W_TYPE_SIZE.
106 2014-05-09 Richard Biener <rguenther@suse.de>
108 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
109 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
110 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
111 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
112 ssa_propagate): Adjust.
114 2014-05-08 Jeff Law <law@redhat.com>
116 PR tree-optimization/61009
117 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
118 tri-state rather than a boolean. When a block is too big to
119 thread through, inform caller via negative return value.
120 (thread_across_edge): If a block was too big for normal threading,
121 then it's too big for a joiner too, so remove temporary equivalences
122 and return immediately.
124 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
125 Matthias Klose <doko@ubuntu.com>
128 * optc-gen.awk: Fix option handling for -Wunused-parameter.
130 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
133 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
135 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
138 * config/alpha/alpha.c: Include gimple-iterator.h.
139 (alpha_gimple_fold_builtin): New function. Move
140 ALPHA_BUILTIN_UMULH folding from ...
141 (alpha_fold_builtin): ... here.
142 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
144 2014-05-08 Wei Mi <wmi@google.com>
147 * config/i386/i386.c (ix86_compute_frame_layout): Update
148 preferred_stack_boundary for call, expanded from tls descriptor.
149 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
150 to depend on SP register.
151 (*tls_local_dynamic_base_32_gnu): Ditto.
152 (*tls_local_dynamic_32_once): Ditto.
153 (tls_global_dynamic_64_<mode>): Set
154 ix86_tls_descriptor_calls_expanded_in_cfun.
155 (tls_local_dynamic_base_64_<mode>): Ditto.
156 (tls_global_dynamic_32): Set
157 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
158 to depend on SP register.
159 (tls_local_dynamic_base_32): Ditto.
161 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
163 * config/arm/arm_neon.h: Update comment.
164 * config/arm/neon-docgen.ml: Delete.
165 * config/arm/neon-gen.ml: Delete.
166 * doc/arm-neon-intrinsics.texi: Update comment.
168 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
170 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
172 (vand, vorr, veor, vorn, vbic): Remove.
173 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
175 (neon_vsub_unspec): Likewise.
176 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
178 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
180 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
181 (vadd_s16): Likewise.
182 (vadd_s32): Likewise.
183 (vadd_f32): Likewise.
185 (vadd_u16): Likewise.
186 (vadd_u32): Likewise.
187 (vadd_s64): Likewise.
188 (vadd_u64): Likewise.
189 (vaddq_s8): Likewise.
190 (vaddq_s16): Likewise.
191 (vaddq_s32): Likewise.
192 (vaddq_s64): Likewise.
193 (vaddq_f32): Likewise.
194 (vaddq_u8): Likewise.
195 (vaddq_u16): Likewise.
196 (vaddq_u32): Likewise.
197 (vaddq_u64): Likewise.
199 (vmul_s16): Likewise.
200 (vmul_s32): Likewise.
201 (vmul_f32): Likewise.
203 (vmul_u16): Likewise.
204 (vmul_u32): Likewise.
206 (vmulq_s8): Likewise.
207 (vmulq_s16): Likewise.
208 (vmulq_s32): Likewise.
209 (vmulq_f32): Likewise.
210 (vmulq_u8): Likewise.
211 (vmulq_u16): Likewise.
212 (vmulq_u32): Likewise.
214 (vsub_s16): Likewise.
215 (vsub_s32): Likewise.
216 (vsub_f32): Likewise.
218 (vsub_u16): Likewise.
219 (vsub_u32): Likewise.
220 (vsub_s64): Likewise.
221 (vsub_u64): Likewise.
222 (vsubq_s8): Likewise.
223 (vsubq_s16): Likewise.
224 (vsubq_s32): Likewise.
225 (vsubq_s64): Likewise.
226 (vsubq_f32): Likewise.
227 (vsubq_u8): Likewise.
228 (vsubq_u16): Likewise.
229 (vsubq_u32): Likewise.
230 (vsubq_u64): Likewise.
232 (vand_s16): Likewise.
233 (vand_s32): Likewise.
235 (vand_u16): Likewise.
236 (vand_u32): Likewise.
237 (vand_s64): Likewise.
238 (vand_u64): Likewise.
239 (vandq_s8): Likewise.
240 (vandq_s16): Likewise.
241 (vandq_s32): Likewise.
242 (vandq_s64): Likewise.
243 (vandq_u8): Likewise.
244 (vandq_u16): Likewise.
245 (vandq_u32): Likewise.
246 (vandq_u64): Likewise.
248 (vorr_s16): Likewise.
249 (vorr_s32): Likewise.
251 (vorr_u16): Likewise.
252 (vorr_u32): Likewise.
253 (vorr_s64): Likewise.
254 (vorr_u64): Likewise.
255 (vorrq_s8): Likewise.
256 (vorrq_s16): Likewise.
257 (vorrq_s32): Likewise.
258 (vorrq_s64): Likewise.
259 (vorrq_u8): Likewise.
260 (vorrq_u16): Likewise.
261 (vorrq_u32): Likewise.
262 (vorrq_u64): Likewise.
264 (veor_s16): Likewise.
265 (veor_s32): Likewise.
267 (veor_u16): Likewise.
268 (veor_u32): Likewise.
269 (veor_s64): Likewise.
270 (veor_u64): Likewise.
271 (veorq_s8): Likewise.
272 (veorq_s16): Likewise.
273 (veorq_s32): Likewise.
274 (veorq_s64): Likewise.
275 (veorq_u8): Likewise.
276 (veorq_u16): Likewise.
277 (veorq_u32): Likewise.
278 (veorq_u64): Likewise.
280 (vbic_s16): Likewise.
281 (vbic_s32): Likewise.
283 (vbic_u16): Likewise.
284 (vbic_u32): Likewise.
285 (vbic_s64): Likewise.
286 (vbic_u64): Likewise.
287 (vbicq_s8): Likewise.
288 (vbicq_s16): Likewise.
289 (vbicq_s32): Likewise.
290 (vbicq_s64): Likewise.
291 (vbicq_u8): Likewise.
292 (vbicq_u16): Likewise.
293 (vbicq_u32): Likewise.
294 (vbicq_u64): Likewise.
296 (vorn_s16): Likewise.
297 (vorn_s32): Likewise.
299 (vorn_u16): Likewise.
300 (vorn_u32): Likewise.
301 (vorn_s64): Likewise.
302 (vorn_u64): Likewise.
303 (vornq_s8): Likewise.
304 (vornq_s16): Likewise.
305 (vornq_s32): Likewise.
306 (vornq_s64): Likewise.
307 (vornq_u8): Likewise.
308 (vornq_u16): Likewise.
309 (vornq_u32): Likewise.
310 (vornq_u64): Likewise.
312 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
314 * wide-int.cc (UTItype): Define.
315 (UDWtype): Define for appropriate W_TYPE_SIZE.
317 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
319 PR tree-optimization/59100
320 * tree-ssa-phiopt.c: Include tree-inline.h.
321 (neutral_element_p, absorbing_element_p): New functions.
322 (value_replacement): Handle conditional binary operations with a
323 neutral or absorbing element.
325 2014-05-08 Richard Biener <rguenther@suse.de>
327 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
328 folding the expression.
329 (valueize_expr): Remove.
330 (visit_reference_op_load): Do not valueize the result of
332 (simplify_binary_expression): Likewise.
333 (simplify_unary_expression): Likewise.
335 2014-05-08 Richard Biener <rguenther@suse.de>
337 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
338 looking at TYPE_ARG_TYPES.
340 2014-05-08 Richard Biener <rguenther@suse.de>
342 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
343 pointer propagation special-case.
345 2014-05-08 Bin Cheng <bin.cheng@arm.com>
347 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
348 core part of address expressions.
350 2014-05-08 Alan Modra <amodra@gmail.com>
353 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
354 loads and stores when -mno-strict-align at any alignment.
355 (expand_block_clear): Similarly. Also correct calculation of
358 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
361 * tree-complex.c (expand_complex_move): Keep line info when expanding
363 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
364 of complex expression. Use new argument to display correct location
365 for values coming from phi statement.
366 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
367 (warn_uninitialized_phi): Pass location of phi argument to
369 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
370 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
372 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
374 * config/rs6000/predicates.md (indexed_address_mem): New.
375 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
376 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
377 fpstore_ux, fpstore_u.
378 (sign_extend, indexed, update): New.
379 (cell_micro): Adjust.
380 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
381 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
382 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
383 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
384 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
385 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
386 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
387 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
388 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
389 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
390 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
391 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
392 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
393 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
394 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
396 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
397 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
398 *vsx_extract_<mode>_store): Adjust.
399 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
400 is_cracked_insn, insn_must_be_first_in_group,
401 insn_must_be_last_in_group): Adjust.
403 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
405 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
406 ppc440-fpstore): Adjust.
407 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
408 ppc476-fpstore): Adjust.
409 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
410 ppc601-fpstore): Adjust.
411 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
413 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
415 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
416 ppc7450-fpstore): Adjust.
417 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
418 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
419 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
421 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
422 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
423 cell-fpstore, cell-fpstore-update): Adjust.
424 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
425 ppce300c3_store, ppce300c3_fpstore): Adjust.
426 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
427 e500mc_fpstore): Adjust.
428 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
429 e500mc64_store, e500mc64_fpstore): Adjust.
430 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
431 e5500_fpstore): Adjust.
432 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
433 e6500_fpstore): Adjust.
434 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
436 * config/rs6000/power4.md (power4-load, power4-load-ext,
437 power4-load-ext-update, power4-load-ext-update-indexed,
438 power4-load-update-indexed, power4-load-update, power4-fpload,
439 power4-fpload-update, power4-store, power4-store-update,
440 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
442 * config/rs6000/power5.md (power5-load, power5-load-ext,
443 power5-load-ext-update, power5-load-ext-update-indexed,
444 power5-load-update-indexed, power5-load-update, power5-fpload,
445 power5-fpload-update, power5-store, power5-store-update,
446 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
448 * config/rs6000/power6.md (power6-load, power6-load-ext,
449 power6-load-update, power6-load-update-indexed,
450 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
451 power6-fpload-update, power6-store, power6-store-update,
452 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
454 * config/rs6000/power7.md (power7-load, power7-load-ext,
455 power7-load-update, power7-load-update-indexed,
456 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
457 power7-fpload-update, power7-store, power7-store-update,
458 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
460 * config/rs6000/power8.md (power8-load, power8-load-update,
461 power8-load-ext, power8-load-ext-update, power8-fpload,
462 power8-fpload-update, power8-store, power8-store-update-indexed,
463 power8-fpstore, power8-fpstore-update): Adjust.
464 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
466 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
467 titan_lsu_store, titan_lsu_fpstore): Adjust.
468 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
470 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
473 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
474 unrolled byte insns. Emit address increments after move insns.
476 2014-05-07 David Malcolm <dmalcolm@redhat.com>
478 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
479 const_gimple, rather than a gimple.
480 (gimple_call_builtin_p): Likewise, for the three variants.
482 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
483 (gimple_call_builtin_p): Likewise, for the three variants.
485 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
487 PR tree-optimization/61095
488 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
490 2014-05-07 Richard Biener <rguenther@suse.de>
492 PR tree-optimization/61034
493 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
494 (maybe_skip_until): Use translate to take into account
495 lattices when trying to do disambiguations.
496 (get_continuation_for_phi_1): Likewise.
497 (get_continuation_for_phi): Adjust for added translate arguments.
498 (walk_non_aliased_vuses): Likewise.
499 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
500 (walk_non_aliased_vuses): Likewise.
501 (call_may_clobber_ref_p_1): Declare.
502 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
503 calls. Stop early if we are only supposed to disambiguate.
504 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
506 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
508 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
509 Emit an error when the function has arguments.
511 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
513 * cfgloop.h (unswitch_loops): Remove.
514 * doc/passes.texi: Remove references to loop-unswitch.c
515 * timevar.def (TV_LOOP_UNSWITCH): Remove.
517 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
519 * tree-vect-data-refs.c (vect_grouped_load_supported): New
520 check for loads group of length 3.
521 (vect_permute_load_chain): New permutations for loads group of
523 * tree-vect-stmts.c (vect_model_load_cost): Change cost
524 of vec_perm_shuffle for the new permutations.
526 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
528 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
529 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
530 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
531 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
532 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
533 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
534 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
535 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
537 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
539 * loop-unswitch.c: Delete.
541 2014-05-07 Richard Biener <rguenther@suse.de>
543 * config.gcc: Always set need_64bit_hwint to yes.
545 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
547 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
548 of using optimize_size.
550 2014-05-06 Mike Stump <mikestump@comcast.net>
552 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
554 2014-05-06 Joseph Myers <joseph@codesourcery.com>
556 * config/i386/sse.md (*mov<mode>_internal)
557 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
558 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
559 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
560 (*<code><mode>3, *andnot<mode>3<mask_name>)
561 (<mask_codefor><code><mode>3<mask_name>): Only consider
562 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
564 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
567 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
569 * lra-constraints.c (valid_address_p): Move earlier in file.
570 Add a constraint argument to the address_info version.
571 (satisfies_memory_constraint_p): New function.
572 (satisfies_address_constraint_p): Likewise.
573 (process_alt_operands, curr_insn_transform): Use them.
574 (process_address): Pass the constraint to valid_address_p when
575 checking address operands.
577 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
579 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
580 to their respective blocks. Fix inadvertent use of "node".
582 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
584 * emit-rtl.c (init_derived_machine_modes): New functionm, split
586 (init_emit_once): ...here.
587 * rtl.h (init_derived_machine_modes): Declare.
588 * toplev.c (do_compile): Call it even if no_backend.
590 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
591 Mike Stump <mikestump@comcast.net>
592 Richard Sandiford <rdsandiford@googlemail.com>
593 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
595 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
596 (rtx_equal_for_memref_p): Update comment.
597 (adjust_offset_for_component_ref): Use wide-int interfaces.
598 * builtins.c (get_object_alignment_2): Likewise.
599 (c_readstr): Likewise.
600 (target_char_cast): Add comment.
601 (determine_block_size): Use wide-int interfaces.
602 (expand_builtin_signbit): Likewise.
603 (fold_builtin_int_roundingfn): Likewise.
604 (fold_builtin_bitop): Likewise.
605 (fold_builtin_bswap): Likewise.
606 (fold_builtin_logarithm): Use signop.
607 (fold_builtin_pow): Likewise.
608 (fold_builtin_memory_op): Use wide-int interfaces.
609 (fold_builtin_object_size): Likewise.
610 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
611 nb_iterations_estimate.
612 (record_niter_bound): Use wide-int interfaces.
613 (get_estimated_loop_iterations_int): Likewise.
614 (get_estimated_loop_iterations): Likewise.
615 (get_max_loop_iterations): Likewise.
616 * cfgloop.h: Include wide-int.h.
617 (struct nb_iter_bound): Change bound to widest_int.
618 (struct loop): Change nb_iterations_upper_bound and
619 nb_iterations_estimate to widest_int.
620 (record_niter_bound): Switch to use widest_int.
621 (get_estimated_loop_iterations): Likewise.
622 (get_max_loop_iterations): Likewise.
623 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
625 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
626 * combine.c (try_combine): Likewise.
627 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
628 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
630 (aarch64_float_const_representable_p): Likewise.
631 * config/arc/arc.c: Include wide-int.h.
632 (arc_can_use_doloop_p): Use wide-int interfaces.
633 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
634 (vfp3_const_double_index): Likewise.
635 * config/avr/avr.c (avr_out_round): Likewise.
636 (avr_fold_builtin): Likewise.
637 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
638 (bfin_can_use_doloop_p): Likewise.
639 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
640 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
641 * config/i386/i386.c: Include wide-int.h.
642 (ix86_data_alignment): Use wide-int interfaces.
643 (ix86_local_alignment): Likewise.
644 (ix86_emit_swsqrtsf): Update real_from_integer.
645 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
646 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
647 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
648 (zero_constant): Likewise.
649 (input_operand): Likewise.
650 (splat_input_operand): Likewise.
651 (non_logical_cint_operand): Change const_double to const_wide_int.
652 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
653 (easy_altivec_constant): Remove comment.
654 (paired_expand_vector_init): Use CONSTANT_P.
655 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
656 (rs6000_emit_move): Update checks.
657 (rs6000_aggregate_candidate): Use wide-int interfaces.
658 (rs6000_expand_ternop_builtin): Likewise.
659 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
660 (rs6000_assemble_integer): Likewise.
661 (rs6000_hash_constant): Likewise.
662 (output_toc): Likewise.
663 (rs6000_rtx_costs): Likewise.
664 (rs6000_emit_swrsqrt); Update call to real_from_integer.
665 * config/rs6000/rs6000-c.c: Include wide-int.h.
666 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
667 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
668 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
669 Handle CONST_WIDE_INT.
670 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
671 Use tree_fits_uhwi_p.
672 * config/sparc/sparc.c: Include wide-int.h.
673 (sparc_fold_builtin): Use wide-int interfaces.
674 * config/vax/vax.c: Include wide-int.h.
675 (vax_float_literal): Use real_from_integer.
676 * coretypes.h (struct hwivec_def): New.
679 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
680 (equiv_constant): Handle CONST_WIDE_INT.
681 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
682 (cselib_hash_rtx): Handle CONST_WIDE_INT.
683 * dbxout.c (stabstr_U): Use wide-int interfaces.
684 (dbxout_type): Update to use cst_fits_shwi_p.
685 * defaults.h (LOG2_BITS_PER_UNIT): Define.
686 (TARGET_SUPPORTS_WIDE_INT): Add default.
687 * dfp.c: Include wide-int.h.
688 (decimal_real_to_integer2): Use wide-int interfaces and rename to
689 decimal_real_to_integer.
690 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
691 decimal_real_to_integer.
692 * doc/generic.texi (Constant expressions): Update for wide_int.
693 * doc/rtl.texi (const_double): Likewise.
694 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
695 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
696 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
697 (REAL_VALUE_FROM_INT): Remove.
698 (TARGET_SUPPORTS_WIDE_INT): New.
699 * doc/tm.texi: Regenerate.
700 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
701 * double-int.h: Include wide-int.h.
702 (struct wi::int_traits): New.
703 * dwarf2out.c (get_full_len): New.
704 (dw_val_equal_p): Add case dw_val_class_wide_int.
705 (size_of_loc_descr): Likewise.
706 (output_loc_operands): Likewise.
707 (insert_double): Remove.
708 (insert_wide_int): New.
710 (print_die): Add case dw_val_class_wide_int.
711 (attr_checksum): Likewise.
712 (attr_checksum_ordered): Likewise.
713 (same_dw_val_p): Likewise.
714 (size_of_die): Likewise.
715 (value_format): Likewise.
716 (output_die): Likewise.
717 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
719 (clz_loc_descriptor): Use wide-int interfaces.
720 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
721 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
722 (round_up_to_align): Use wide-int interfaces.
723 (field_byte_offset): Likewise.
724 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
725 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
726 CONST_DOUBLE handling. Use wide-int interfaces.
727 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
728 (gen_enumeration_type_die): Use add_AT_wide.
729 (hash_loc_operands): Add case dw_val_class_wide_int.
730 (compare_loc_operands): Likewise.
731 * dwarf2out.h: Include wide-int.h.
733 (enum dw_val_class): Add dw_val_class_wide_int.
734 (struct dw_val_struct): Add val_wide.
735 * emit-rtl.c (const_wide_int_htab): New.
736 (const_wide_int_htab_hash): New.
737 (const_wide_int_htab_eq): New.
738 (lookup_const_wide_int): New.
739 (const_double_htab_hash): Use wide-int interfaces.
740 (const_double_htab_eq): Likewise.
741 (rtx_to_double_int): Conditionally compile for wide-int.
742 (immed_double_int_const): Rename to immed_wide_int_const and
744 (immed_double_const): Conditionally compile for wide-int.
745 (init_emit_once): Use wide-int interfaces.
746 * explow.c (plus_constant): Likewise.
747 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
748 (lshift_value): Use wide-int interfaces.
749 (expand_mult): Likewise.
750 (choose_multiplier): Likewise.
751 (expand_smod_pow2): Likewise.
752 (make_tree): Likewise.
753 * expr.c (convert_modes): Consolidate handling of constants.
754 Use wide-int interfaces.
755 (emit_group_load_1): Add note.
756 (store_expr): Update comment.
757 (get_inner_reference): Use wide-int interfaces.
758 (expand_constructor): Update comment.
759 (expand_expr_real_2): Use wide-int interfaces.
760 (expand_expr_real_1): Likewise.
761 (reduce_to_bit_field_precision): Likewise.
762 (const_vector_from_tree): Likewise.
763 * final.c: Include wide-int-print.h.
764 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
765 * fixed-value.c: Include wide-int.h.
766 (fixed_from_string): Use wide-int interfaces.
767 (fixed_to_decimal): Likewise.
768 (fixed_convert_from_real): Likewise.
769 (real_convert_from_fixed): Likewise.
770 * fold-const.h (mem_ref_offset): Return an offset_int.
771 (div_if_zero_remainder): Remove code parameter.
772 * fold-const.c (div_if_zero_remainder): Remove code parameter.
773 Use wide-int interfaces.
774 (may_negate_without_overflow_p): Use wide-int interfaces.
775 (negate_expr_p): Likewise.
776 (fold_negate_expr): Likewise.
777 (int_const_binop_1): Likewise.
778 (const_binop): Likewise.
779 (fold_convert_const_int_from_int): Likewise.
780 (fold_convert_const_int_from_real): Likewise.
781 (fold_convert_const_int_from_fixed): Likewise.
782 (fold_convert_const_fixed_from_int): Likewise.
783 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
784 (sign_bit_p): Use wide-int interfaces.
785 (make_range_step): Likewise.
786 (build_range_check): Likewise. Pass an integer of the correct type
787 instead of using integer_one_node.
788 (range_predecessor): Pass an integer of the correct type instead
789 of using integer_one_node.
790 (range_successor): Likewise.
791 (merge_ranges): Likewise.
792 (unextend): Use wide-int interfaces.
793 (extract_muldiv_1): Likewise.
794 (fold_div_compare): Likewise.
795 (fold_single_bit_test): Likewise.
796 (fold_sign_changed_comparison): Likewise.
797 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
798 (fold_plusminus_mult_expr): Use wide-int interfaces.
799 (native_encode_int): Likewise.
800 (native_interpret_int): Likewise.
801 (fold_unary_loc): Likewise.
802 (pointer_may_wrap_p): Likewise.
803 (size_low_cst): Likewise.
804 (mask_with_tz): Likewise.
805 (fold_binary_loc): Likewise.
806 (fold_ternary_loc): Likewise.
807 (multiple_of_p): Likewise.
808 (tree_call_nonnegative_warnv_p): Update calls to
809 tree_int_cst_min_precision and real_from_integer.
810 (fold_negate_const): Use wide-int interfaces.
811 (fold_abs_const): Likewise.
812 (fold_relational_const): Use tree_int_cst_lt.
813 (round_up_loc): Use wide-int interfaces.
814 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
815 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
816 * gengtype.c: Remove include of double-int.h.
817 (do_typedef): Use wide-int interfaces.
818 (open_base_files): Add wide-int.h.
819 (main): Add offset_int and widest_int typedefs.
820 * gengtype-lex.l: Handle "^".
821 (CXX_KEYWORD): Add "static".
822 * gengtype-parse.c (require3): New.
823 (require_template_declaration): Handle constant template arguments
824 and nested templates.
825 * gengtype-state.c: Don't include "double-int.h".
826 * genpreds.c (write_one_predicate_function): Update comment.
827 (write_tm_constrs_h): Add check for hval and lval use in
829 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
830 (add_to_sequence): Likewise.
831 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
832 and const_double_operand.
833 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
835 * gimple-fold.c (get_base_constructor): Likewise.
836 (fold_array_ctor_reference): Likewise.
837 (fold_nonarray_ctor_reference): Likewise.
838 (fold_const_aggregate_ref_1): Likewise.
839 (gimple_val_nonnegative_real_p): Likewise.
840 (gimple_fold_indirect_ref): Likewise.
841 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
842 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
843 (struct slsr_cand_d): Change index to be widest_int.
844 (struct incr_info_d): Change incr to be widest_int.
845 (alloc_cand_and_find_basis): Use wide-int interfaces.
846 (slsr_process_phi): Likewise.
847 (backtrace_base_for_ref): Likewise. Return a widest_int.
848 (restructure_reference): Take a widest_int instead of a double_int.
849 (slsr_process_ref): Use wide-int interfaces.
850 (create_mul_ssa_cand): Likewise.
851 (create_mul_imm_cand): Likewise.
852 (create_add_ssa_cand): Likewise.
853 (create_add_imm_cand): Take a widest_int instead of a double_int.
854 (slsr_process_add): Use wide-int interfaces.
855 (slsr_process_cast): Likewise.
856 (slsr_process_copy): Likewise.
857 (dump_candidate): Likewise.
858 (dump_incr_vec): Likewise.
859 (replace_ref): Likewise.
860 (cand_increment): Likewise. Return a widest_int.
861 (cand_abs_increment): Likewise.
862 (replace_mult_candidate): Take a widest_int instead of a double_int.
863 (replace_unconditional_candidate): Use wide-int interfaces.
864 (incr_vec_index): Take a widest_int instead of a double_int.
865 (create_add_on_incoming_edge): Likewise.
866 (create_phi_basis): Use wide-int interfaces.
867 (replace_conditional_candidate): Likewise.
868 (record_increment): Take a widest_int instead of a double_int.
869 (record_phi_increments): Use wide-int interfaces.
870 (phi_incr_cost): Take a widest_int instead of a double_int.
871 (lowest_cost_path): Likewise.
872 (total_savings): Likewise.
873 (analyze_increments): Use wide-int interfaces.
874 (ncd_with_phi): Take a widest_int instead of a double_int.
875 (ncd_of_cand_and_phis): Likewise.
876 (nearest_common_dominator_for_cands): Likewise.
877 (insert_initializers): Use wide-int interfaces.
878 (all_phi_incrs_profitable): Likewise.
879 (replace_one_candidate): Likewise.
880 (replace_profitable_candidates): Likewise.
881 * godump.c: Include wide-int-print.h.
882 (go_output_typedef): Use wide-int interfaces.
883 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
884 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
885 (build_loop_iteration_domains): Likewise.
886 * hooks.h: Include wide-int.h rather than double-int.h.
887 (hook_bool_dint_dint_uint_bool_true): Delete.
888 (hook_bool_wint_wint_uint_bool_true): Declare.
889 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
890 (hook_bool_wint_wint_uint_bool_true): New.
891 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
893 (ubsan_expand_si_overflow_mul_check): Likewise.
894 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
895 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
896 (get_ancestor_addr_info): Likewise.
897 (ipa_modify_call_arguments): Likewise.
898 * loop-doloop.c (doloop_modify): Likewise.
899 (doloop_optimize): Likewise.
900 * loop-iv.c (iv_number_of_iterations): Likewise.
901 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
902 (unroll_loop_constant_iterations): Likewise.
903 (decide_unroll_runtime_iterations): Likewise.
904 (unroll_loop_runtime_iterations): Likewise.
905 (decide_peel_simple): Likewise.
906 (decide_unroll_stupid): Likewise.
907 * lto-streamer-in.c (streamer_read_wi): Add.
908 (input_cfg): Use wide-int interfaces.
909 (lto_input_tree_1): Likewise.
910 * lto-streamer-out.c (streamer_write_wi): Add.
911 (hash_tree): Use wide-int interfaces.
912 (output_cfg): Likewise.
913 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
914 (GTFILES): Add wide-int.h and signop.h.
915 (TAGS): Look for .cc files too.
916 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
917 * optabs.c (expand_subword_shift): Likewise.
918 (expand_doubleword_shift): Likewise.
919 (expand_absneg_bit): Likewise.
920 (expand_copysign_absneg): Likewise.
921 (expand_copysign_bit): Likewise.
922 * postreload.c (reload_cse_simplify_set): Likewise.
923 * predict.c (predict_iv_comparison): Likewise.
924 * pretty-print.h: Include wide-int-print.h.
926 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
927 * print-tree.c: Include wide-int-print.h.
928 (print_node_brief): Use wide-int interfaces.
929 (print_node): Likewise.
930 * read-rtl.c (validate_const_wide_int): New.
931 (read_rtx_code): Add CONST_WIDE_INT case.
932 * real.c: Include wide-int.h.
933 (real_to_integer2): Delete.
934 (real_to_integer): New function, returning a wide_int.
935 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
936 (ten_to_ptwo): Update call to real_from_integer.
937 (real_digit): Likewise.
938 * real.h: Include signop.h, wide-int.h and insn-modes.h.
939 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
940 (REAL_VALUE_TO_INT): Delete.
941 (real_to_integer): Declare a wide-int form.
942 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
943 * recog.c (const_int_operand): Improve comment.
944 (const_scalar_int_operand): New.
945 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
946 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
947 (split_double): Likewise.
948 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
949 (rtx_size): Likewise.
950 (rtx_alloc_stat_v): New.
951 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
952 (cwi_output_hex): New.
953 (iterative_hash_rtx): Handle CONST_WIDE_INT.
954 (cwi_check_failed_bounds): New.
955 * rtl.def (CONST_WIDE_INT): New.
956 * rtl.h: Include <utility> and wide-int.h.
957 (struct hwivec_def): New.
958 (CWI_GET_NUM_ELEM): New.
959 (CWI_PUT_NUM_ELEM): New.
960 (struct rtx_def): Add num_elem and hwiv.
961 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
962 (CASE_CONST_UNIQUE): Likewise.
963 (CASE_CONST_ANY): Likewise.
964 (CONST_SCALAR_INT_P): Likewise.
965 (CONST_WIDE_INT_P): New.
968 (cwi_check_failed_bounds): New.
971 (CONST_WIDE_INT_VEC) New.
972 (CONST_WIDE_INT_NUNITS) New.
973 (CONST_WIDE_INT_ELT) New.
974 (rtx_mode_t): New type.
975 (wi::int_traits <rtx_mode_t>): New.
977 (wi::min_value): New.
978 (wi::max_value): New.
980 (const_wide_int_alloc): New.
981 (immed_wide_int_const): New.
982 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
983 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
984 * signop.h: New file.
985 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
986 (simplify_const_unary_operation): Use wide-int interfaces.
987 (simplify_binary_operation_1): Likewise.
988 (simplify_const_binary_operation): Likewise.
989 (simplify_const_relational_operation): Likewise.
990 (simplify_immed_subreg): Likewise.
991 * stmt.c (expand_case): Likewise.
992 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
993 signop rather than a bool.
994 * stor-layout.c (layout_type): Use wide-int interfaces.
995 (initialize_sizetypes): Update calls to
996 set_min_and_max_values_for_integral_type.
997 (set_min_and_max_values_for_integral_type): Take a signop rather
998 than a bool. Use wide-int interfaces.
999 (fixup_signed_type): Update accordingly. Remove
1000 HOST_BITS_PER_DOUBLE_INT limit.
1001 (fixup_unsigned_type): Likewise.
1002 * system.h (STATIC_CONSTANT_P): New.
1003 (STATIC_ASSERT): New.
1004 * target.def (can_use_doloop_p): Take widest_ints rather than
1006 * target.h: Include wide-int.h rather than double-int.h.
1007 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
1009 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
1010 rather than INT_CST_LT_UNSIGNED.
1011 (can_use_doloop_if_innermost): Take widest_ints rather than
1013 * tree-affine.c: Include wide-int-print.h.
1014 (double_int_ext_for_comb): Delete.
1015 (wide_int_ext_for_comb): New.
1016 (aff_combination_zero): Use wide-int interfaces.
1017 (aff_combination_const): Take a widest_int instead of a double_int.
1018 (aff_combination_elt): Use wide-int interfaces.
1019 (aff_combination_scale): Take a widest_int instead of a double_int.
1020 (aff_combination_add_elt): Likewise.
1021 (aff_combination_add_cst): Likewise.
1022 (aff_combination_add): Use wide-int interfaces.
1023 (aff_combination_convert): Likewise.
1024 (tree_to_aff_combination): Likewise.
1025 (add_elt_to_tree): Take a widest_int instead of a double_int.
1026 (aff_combination_to_tree): Use wide-int interfaces.
1027 (aff_combination_remove_elt): Likewise.
1028 (aff_combination_add_product): Take a widest_int instead of
1030 (aff_combination_mult): Use wide-int interfaces.
1031 (aff_combination_expand): Likewise.
1032 (double_int_constant_multiple_p): Delete.
1033 (wide_int_constant_multiple_p): New.
1034 (aff_combination_constant_multiple_p): Take a widest_int pointer
1035 instead of a double_int pointer.
1036 (print_aff): Use wide-int interfaces.
1037 (get_inner_reference_aff): Take a widest_int pointer
1038 instead of a double_int pointer.
1039 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
1040 * tree-affine.h: Include wide-int.h.
1041 (struct aff_comb_elt): Change type of coef to widest_int.
1042 (struct affine_tree_combination): Change type of offset to widest_int.
1043 (double_int_ext_for_comb): Delete.
1044 (wide_int_ext_for_comb): New.
1045 (aff_combination_const): Use widest_int instead of double_int.
1046 (aff_combination_scale): Likewise.
1047 (aff_combination_add_elt): Likewise.
1048 (aff_combination_constant_multiple_p): Likewise.
1049 (get_inner_reference_aff): Likewise.
1050 (aff_comb_cannot_overlap_p): Likewise.
1051 (aff_combination_zero_p): Use wide-int interfaces.
1052 * tree.c: Include tree.h.
1053 (init_ttree): Use make_int_cst.
1054 (tree_code_size): Removed code for INTEGER_CST case.
1055 (tree_size): Add INTEGER_CST case.
1056 (make_node_stat): Update comment.
1057 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
1058 (build_int_cst_type): Use wide-int interfaces.
1059 (double_int_to_tree): Likewise.
1060 (double_int_fits_to_tree_p): Delete.
1061 (force_fit_type_double): Delete.
1062 (force_fit_type): New.
1063 (int_cst_hash_hash): Use wide-int interfaces.
1064 (int_cst_hash_eq): Likewise.
1065 (build_int_cst_wide): Delete.
1066 (wide_int_to_tree): New.
1067 (cache_integer_cst): Use wide-int interfaces.
1068 (build_low_bits_mask): Likewise.
1069 (cst_and_fits_in_hwi): Likewise.
1070 (real_value_from_int_cst): Likewise.
1071 (make_int_cst_stat): New.
1072 (integer_zerop): Use wide_int interfaces.
1073 (integer_onep): Likewise.
1074 (integer_all_onesp): Likewise.
1075 (integer_pow2p): Likewise.
1076 (integer_nonzerop): Likewise.
1077 (tree_log2): Likewise.
1078 (tree_floor_log2): Likewise.
1079 (tree_ctz): Likewise.
1080 (int_size_in_bytes): Likewise.
1081 (mem_ref_offset): Return an offset_int rather than a double_int.
1082 (build_type_attribute_qual_variant): Use wide_int interfaces.
1083 (type_hash_eq): Likewise
1084 (tree_int_cst_equal): Likewise.
1085 (tree_int_cst_lt): Delete.
1086 (tree_int_cst_compare): Likewise.
1087 (tree_fits_shwi_p): Use wide_int interfaces.
1088 (tree_fits_uhwi_p): Likewise.
1089 (tree_int_cst_sign_bit): Likewise.
1090 (tree_int_cst_sgn): Likewise.
1091 (tree_int_cst_min_precision): Take a signop rather than a bool.
1092 (simple_cst_equal): Use wide_int interfaces.
1093 (compare_tree_int): Likewise.
1094 (iterative_hash_expr): Likewise.
1095 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
1097 (get_type_static_bounds): Use wide_int interfaces.
1098 (tree_int_cst_elt_check_failed): New.
1099 (build_common_tree_nodes): Reordered to set prec before filling in
1101 (int_cst_value): Check cst_and_fits_in_hwi.
1102 (widest_int_cst_value): Use wide_int interfaces.
1103 (upper_bound_in_type): Likewise.
1104 (lower_bound_in_type): Likewise.
1105 (num_ending_zeros): Likewise.
1106 (drop_tree_overflow): Likewise.
1107 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
1108 (gen_conditions_for_pow_cst_base): Likewise.
1109 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
1110 (group_case_labels_stmt): Use wide-int interfaces.
1111 (verify_gimple_assign_binary): Likewise.
1112 (print_loop): Likewise.
1113 * tree-chrec.c (tree_fold_binomial): Likewise.
1114 * tree-core.h (struct tree_base): Add int_length.
1115 (struct tree_int_cst): Change rep of value.
1116 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
1117 (dr_may_alias_p): Likewise.
1118 (max_stmt_executions_tree): Likewise.
1119 * tree.def (INTEGER_CST): Update comment.
1120 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
1121 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
1122 * tree-dump.c: Include wide-int.h and wide-int-print.h.
1123 (dequeue_and_dump): Use wide-int interfaces.
1124 * tree.h: Include wide-int.h.
1125 (NULL_TREE): Moved to earlier loc in file.
1126 (TREE_INT_CST_ELT_CHECK): New.
1127 (tree_int_cst_elt_check_failed): New.
1129 (TREE_INT_CST): Delete.
1130 (TREE_INT_CST_LOW): Use wide-int interfaces.
1131 (TREE_INT_CST_HIGH): Delete.
1132 (TREE_INT_CST_NUNITS): New.
1133 (TREE_INT_CST_EXT_NUNITS): Likewise.
1134 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
1135 (TREE_INT_CST_ELT): Likewise.
1136 (INT_CST_LT): Delete.
1137 (tree_int_cst_elt_check): New (two forms).
1138 (type_code_size): Update comment.
1139 (make_int_cst_stat, make_int_cst): New.
1140 (tree_to_double_int): Delete.
1141 (double_int_fits_to_tree_p): Delete.
1142 (force_fit_type_double): Delete.
1143 (build_int_cstu): Replace with out-of-line function.
1144 (build_int_cst_wide): Delete.
1145 (tree_int_cst_lt): Define inline.
1146 (tree_int_cst_le): New.
1147 (tree_int_cst_compare): Define inline.
1148 (tree_int_cst_min_precision): Take a signop rather than a bool.
1149 (wi::int_traits <const_tree>): New.
1150 (wi::int_traits <tree>): New.
1151 (wi::extended_tree): New.
1152 (wi::int_traits <wi::extended_tree>): New.
1153 (wi::to_widest): New.
1154 (wi::to_offset): New.
1155 (wi::fits_to_tree_p): New.
1156 (wi::min_value): New.
1157 (wi::max_value): New.
1158 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
1159 (copy_tree_body_r): Likewise.
1160 * tree-object-size.c (compute_object_offset): Likewise.
1161 (addr_object_size): Likewise.
1162 * tree-predcom.c: Include wide-int-print.h.
1163 (struct dref_d): Change type of offset to widest_int.
1164 (dump_dref): Call wide-int printer.
1165 (aff_combination_dr_offset): Use wide-int interfaces.
1166 (determine_offset): Take a widest_int pointer rather than a
1168 (split_data_refs_to_components): Use wide-int interfaces.
1169 (suitable_component_p): Likewise.
1170 (order_drefs): Likewise.
1171 (add_ref_to_chain): Likewise.
1172 (valid_initializer_p): Likewise.
1173 (determine_roots_comp): Likewise.
1174 * tree-pretty-print.c: Include wide-int-print.h.
1175 (dump_generic_node): Use wide-int interfaces.
1176 * tree-sra.c (sra_ipa_modify_expr): Likewise.
1177 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
1178 (move_fixed_address_to_symbol): Likewise.
1179 (move_hint_to_base): Likewise.
1180 (move_pointer_to_base): Likewise.
1181 (move_variant_to_index): Likewise.
1182 (most_expensive_mult_to_index): Likewise.
1183 (addr_to_parts): Likewise.
1184 (copy_ref_info): Likewise.
1185 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
1186 (indirect_refs_may_alias_p): Likewise.
1187 (stmt_kills_ref_p_1): Likewise.
1188 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
1189 * tree-ssa-ccp.c: Update comment at top of file. Include
1191 (struct prop_value_d): Change type of mask to widest_int.
1192 (extend_mask): New function.
1193 (dump_lattice_value): Use wide-int interfaces.
1194 (get_default_value): Likewise.
1195 (set_constant_value): Likewise.
1196 (set_value_varying): Likewise.
1197 (valid_lattice_transition): Likewise.
1198 (set_lattice_value): Likewise.
1199 (value_to_double_int): Delete.
1200 (value_to_wide_int): New.
1201 (get_value_from_alignment): Use wide-int interfaces.
1202 (get_value_for_expr): Likewise.
1203 (do_dbg_cnt): Likewise.
1204 (ccp_finalize): Likewise.
1205 (ccp_lattice_meet): Likewise.
1206 (bit_value_unop_1): Use widest_ints rather than double_ints.
1207 (bit_value_binop_1): Likewise.
1208 (bit_value_unop): Use wide-int interfaces.
1209 (bit_value_binop): Likewise.
1210 (bit_value_assume_aligned): Likewise.
1211 (evaluate_stmt): Likewise.
1212 (ccp_fold_stmt): Likewise.
1213 (visit_cond_stmt): Likewise.
1214 (ccp_visit_stmt): Likewise.
1215 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
1216 (constant_pointer_difference): Likewise.
1217 (associate_pointerplus): Likewise.
1218 (combine_conversions): Likewise.
1219 * tree-ssa-loop.h: Include wide-int.h.
1220 (struct tree_niter_desc): Change type of max to widest_int.
1221 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
1222 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
1223 (remove_redundant_iv_tests): Likewise.
1224 (canonicalize_loop_induction_variables): Likewise.
1225 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
1226 (constant_multiple_of): Take a widest_int pointer instead of
1227 a double_int pointer.
1228 (get_computation_aff): Use wide-int interfaces.
1229 (ptr_difference_cost): Likewise.
1230 (difference_cost): Likewise.
1231 (get_loop_invariant_expr_id): Likewise.
1232 (get_computation_cost_at): Likewise.
1233 (iv_elimination_compare_lt): Likewise.
1234 (may_eliminate_iv): Likewise.
1235 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
1236 instead of double_int.
1237 (max_loop_iterations): Likewise.
1238 (max_stmt_executions): Likewise.
1239 (estimated_stmt_executions): Likewise.
1240 * tree-ssa-loop-niter.c: Include wide-int-print.h.
1241 (split_to_var_and_offset): Use wide-int interfaces.
1242 (determine_value_range): Likewise.
1243 (bound_difference_of_offsetted_base): Likewise.
1244 (bounds_add): Take a widest_int instead of a double_int.
1245 (number_of_iterations_ne_max): Use wide-int interfaces.
1246 (number_of_iterations_ne): Likewise.
1247 (number_of_iterations_lt_to_ne): Likewise.
1248 (assert_loop_rolls_lt): Likewise.
1249 (number_of_iterations_lt): Likewise.
1250 (number_of_iterations_le): Likewise.
1251 (number_of_iterations_cond): Likewise.
1252 (number_of_iterations_exit): Likewise.
1253 (finite_loop_p): Likewise.
1254 (derive_constant_upper_bound_assign): Likewise.
1255 (derive_constant_upper_bound): Return a widest_int.
1256 (derive_constant_upper_bound_ops): Likewise.
1257 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
1258 (record_estimate): Take a widest_int rather than a double_int.
1259 (record_nonwrapping_iv): Use wide-int interfaces.
1260 (double_int_cmp): Delete.
1261 (wide_int_cmp): New.
1262 (bound_index): Take a widest_int rather than a double_int.
1263 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
1264 (maybe_lower_iteration_bound): Likewise.
1265 (estimate_numbers_of_iterations_loop): Likewise.
1266 (estimated_loop_iterations): Take a widest_int pointer than than
1267 a double_int pointer.
1268 (estimated_loop_iterations_int): Use wide-int interfaces.
1269 (max_loop_iterations): Take a widest_int pointer than than
1270 a double_int pointer.
1271 (max_loop_iterations_int): Use wide-int interfaces.
1272 (max_stmt_executions): Take a widest_int pointer than than
1273 a double_int pointer.
1274 (estimated_stmt_executions): Likewise.
1275 (n_of_executions_at_most): Use wide-int interfaces.
1276 (scev_probably_wraps_p): Likewise.
1277 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
1279 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
1281 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
1282 double_ints. Adjust for trailing_wide_ints <3> representation.
1283 (set_nonzero_bits): Likewise.
1284 (get_range_info): Return wide_ints rather than double_ints.
1285 Adjust for trailing_wide_ints <3> representation.
1286 (get_nonzero_bits): Likewise.
1287 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
1289 * tree-ssanames.h (struct range_info_def): Replace min, max and
1290 nonzero_bits with a trailing_wide_ints <3>.
1291 (set_range_info): Use wide_int_refs rather than double_ints.
1292 (set_nonzero_bits): Likewise.
1293 (get_range_info): Return wide_ints rather than double_ints.
1294 (get_nonzero_bits): Likewise.
1295 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
1296 * tree-ssa-pre.c (phi_translate_1): Likewise.
1297 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
1298 (acceptable_pow_call): Likewise.
1299 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
1301 (vn_reference_fold_indirect): Likewise.
1302 (vn_reference_maybe_forwprop_address): Likewise.
1303 (valueize_refs_1): Likewise.
1304 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
1305 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
1306 tree_int_cst_lt and tree_int_cst_le.
1307 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
1309 (streamer_alloc_tree): Likewise.
1310 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
1311 (streamer_write_tree_header): Likewise.
1312 (streamer_write_integer_cst): Likewise.
1313 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
1314 (build_constructors): Likewise.
1315 (array_value_type): Likewise.
1316 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
1317 (vect_check_gather): Likewise.
1318 * tree-vect-generic.c (build_replicated_const): Likewise.
1319 (expand_vector_divmod): Likewise.
1320 * tree-vect-loop.c (vect_transform_loop): Likewise.
1321 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
1322 (vect_do_peeling_for_alignment): Likewise.
1323 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
1324 * tree-vrp.c: Include wide-int.h.
1325 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
1326 (extract_range_from_assert): Use wide-int interfaces.
1327 (vrp_int_const_binop): Likewise.
1328 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
1329 double_int pointers.
1330 (ranges_from_anti_range): Use wide-int interfaces.
1331 (quad_int_cmp): Delete.
1332 (quad_int_pair_sort): Likewise.
1333 (extract_range_from_binary_expr_1): Use wide-int interfaces.
1334 (extract_range_from_unary_expr_1): Likewise.
1335 (adjust_range_with_scev): Likewise.
1336 (masked_increment): Take and return wide_ints rather than double_ints.
1337 (register_edge_assert_for_2): Use wide-int interfaces.
1338 (check_array_ref): Likewise.
1339 (search_for_addr_array): Likewise.
1340 (maybe_set_nonzero_bits): Likewise.
1341 (union_ranges): Pass an integer of the correct type instead of
1342 using integer_one_node.
1343 (intersect_ranges): Likewise.
1344 (simplify_truth_ops_using_ranges): Likewise.
1345 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
1346 (range_fits_type_p): Likewise.
1347 (simplify_cond_using_ranges): Likewise. Take a signop rather than
1349 (simplify_conversion_using_ranges): Use wide-int interfaces.
1350 (simplify_float_conversion_using_ranges): Likewise.
1351 (vrp_finalize): Likewise.
1352 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
1353 (gimple_stringops_transform): Likewise.
1354 * varasm.c (decode_addr_const): Likewise.
1355 (const_hash_1): Likewise.
1356 (const_rtx_hash_1): Likewise
1357 (output_constant): Likewise.
1358 (array_size_for_constructor): Likewise.
1359 (output_constructor_regular_field): Likewise.
1360 (output_constructor_bitfield): Likewise.
1361 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
1362 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
1364 * gencheck.c: Define BITS_PER_UNIT.
1367 * wide-int-print.cc: New.
1368 * wide-int-print.h: New.
1370 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1372 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
1374 2014-05-06 Richard Biener <rguenther@suse.de>
1376 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
1377 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
1378 (TODO_verify_all): Adjust.
1379 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
1380 TODO_verify_stmts and TODO_verify_rtl_sharing.
1381 * bb-reorder.c: Likewise.
1382 * cfgexpand.c: Likewise.
1383 * cprop.c: Likewise.
1385 * function.c: Likewise.
1386 * fwprop.c: Likewise.
1388 * gimple-ssa-isolate-paths.c: Likewise.
1389 * gimple-ssa-strength-reduction.c: Likewise.
1390 * ipa-split.c: Likewise.
1391 * loop-init.c: Likewise.
1392 * loop-unroll.c: Likewise.
1393 * lower-subreg.c: Likewise.
1394 * modulo-sched.c: Likewise.
1395 * postreload-gcse.c: Likewise.
1396 * predict.c: Likewise.
1397 * recog.c: Likewise.
1398 * sched-rgn.c: Likewise.
1399 * store-motion.c: Likewise.
1400 * tracer.c: Likewise.
1401 * trans-mem.c: Likewise.
1402 * tree-call-cdce.c: Likewise.
1403 * tree-cfg.c: Likewise.
1404 * tree-cfgcleanup.c: Likewise.
1405 * tree-complex.c: Likewise.
1406 * tree-eh.c: Likewise.
1407 * tree-emutls.c: Likewise.
1408 * tree-if-conv.c: Likewise.
1409 * tree-into-ssa.c: Likewise.
1410 * tree-loop-distribution.c: Likewise.
1411 * tree-object-size.c: Likewise.
1412 * tree-parloops.c: Likewise.
1413 * tree-pass.h: Likewise.
1414 * tree-sra.c: Likewise.
1415 * tree-ssa-ccp.c: Likewise.
1416 * tree-ssa-copy.c: Likewise.
1417 * tree-ssa-copyrename.c: Likewise.
1418 * tree-ssa-dce.c: Likewise.
1419 * tree-ssa-dom.c: Likewise.
1420 * tree-ssa-dse.c: Likewise.
1421 * tree-ssa-forwprop.c: Likewise.
1422 * tree-ssa-ifcombine.c: Likewise.
1423 * tree-ssa-loop-ch.c: Likewise.
1424 * tree-ssa-loop-ivcanon.c: Likewise.
1425 * tree-ssa-loop.c: Likewise.
1426 * tree-ssa-math-opts.c: Likewise.
1427 * tree-ssa-phiopt.c: Likewise.
1428 * tree-ssa-phiprop.c: Likewise.
1429 * tree-ssa-pre.c: Likewise.
1430 * tree-ssa-reassoc.c: Likewise.
1431 * tree-ssa-sink.c: Likewise.
1432 * tree-ssa-strlen.c: Likewise.
1433 * tree-ssa-tail-merge.c: Likewise.
1434 * tree-ssa-uncprop.c: Likewise.
1435 * tree-switch-conversion.c: Likewise.
1436 * tree-tailcall.c: Likewise.
1437 * tree-vect-generic.c: Likewise.
1438 * tree-vectorizer.c: Likewise.
1439 * tree-vrp.c: Likewise.
1441 * var-tracking.c: Likewise.
1442 * bt-load.c: Likewise.
1443 * cfgcleanup.c: Likewise.
1444 * combine-stack-adj.c: Likewise.
1445 * combine.c: Likewise.
1446 * compare-elim.c: Likewise.
1447 * config/epiphany/resolve-sw-modes.c: Likewise.
1448 * config/i386/i386.c: Likewise.
1449 * config/mips/mips.c: Likewise.
1450 * config/s390/s390.c: Likewise.
1451 * config/sh/sh_treg_combine.cc: Likewise.
1452 * config/sparc/sparc.c: Likewise.
1455 * final.c: Likewise.
1456 * ifcvt.c: Likewise.
1457 * mode-switching.c: Likewise.
1458 * passes.c: Likewise.
1459 * postreload.c: Likewise.
1461 * reg-stack.c: Likewise.
1462 * regcprop.c: Likewise.
1463 * regrename.c: Likewise.
1466 2014-05-06 Richard Biener <rguenther@suse.de>
1469 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
1470 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
1472 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
1475 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
1477 2014-05-05 Radovan Obradovic <robradovic@mips.com>
1478 Tom de Vries <tom@codesourcery.com>
1480 * target.def (call_fusage_contains_non_callee_clobbers): New
1482 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
1484 (@node Miscellaneous Register Hooks): New node.
1485 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
1486 * doc/tm.texi: Regenerate.
1488 2014-05-05 Marek Polacek <polacek@redhat.com>
1491 * opts.c (common_handle_option): Call error_at instead of warning_at.
1493 2014-05-05 Richard Biener <rguenther@suse.de>
1495 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
1496 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
1497 under the TODO_verify_il umbrella.
1499 2014-05-05 Richard Biener <rguenther@suse.de>
1501 * passes.c (execute_function_todo): Move TODO_verify_flow under
1502 the TODO_verify_ul umbrella.
1504 2014-05-05 Richard Biener <rguenther@suse.de>
1507 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
1508 X & CST away from a CST that is the mask of a mode.
1510 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1512 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
1513 int argument to enum machine_mode.
1514 (picochip_class_max_nregs): Ditto.
1515 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
1516 (picochip_class_max_nregs): Ditto.
1518 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1520 * target.def: Add new target hook.
1521 * doc/tm.texi: Regenerate.
1522 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
1523 * targhooks.c (default_keep_leaf_when_profiled): New function.
1525 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
1526 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
1528 2014-05-05 Bin Cheng <bin.cheng@arm.com>
1530 PR tree-optimization/60363
1531 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
1532 (copy_phi_args): New parameters. Call get_value_locus_in_path.
1533 (update_destination_phis): New parameter.
1534 (create_edge_and_update_destination_phis): Ditto.
1535 (ssa_fix_duplicate_block_edges): Pass new arguments.
1536 (thread_single_edge): Ditto.
1538 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
1540 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
1541 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
1542 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
1543 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
1544 Use RS6000_BTM_HARD_FLOAT.
1545 (BU_MISC_2): Likewise.
1546 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
1547 RS6000_BTM_HARD_FLOAT.
1548 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
1550 (rs6000_invalid_builtin): Add hard floating builtin support.
1551 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
1552 hard float builtins.
1553 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
1555 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1557 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
1558 Add missing function* argument.
1560 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1562 * lra-constraints.c (valid_address_p): Move earlier in file.
1563 Add a constraint argument to the address_info version.
1564 (satisfies_memory_constraint_p): New function.
1565 (satisfies_address_constraint_p): Likewise.
1566 (process_alt_operands, curr_insn_transform): Use them.
1567 (process_address): Pass the constraint to valid_address_p when
1568 checking address operands.
1570 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1572 * config/mips/mips.c (mips_isa_rev): New variable.
1573 (mips_set_architecture): Set it.
1574 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1576 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1577 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1578 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1579 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1580 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
1581 conditions in terms of mips_isa_rev.
1582 (mips_isa_rev): Declare.
1584 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1586 * config/sh/sh-mem.cc: Use tabs instead of spaces.
1587 (prob_unlikely, prob_likely): Make variables const.
1589 2014-05-03 Denis Chertykov <chertykov@gmail.com>
1591 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
1593 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1595 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
1597 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1599 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
1600 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
1601 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
1603 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
1604 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
1606 Replace usage of ROUND_REG with sh_round_reg.
1607 Use CEIL instead of ROUND_ADVANCE.
1609 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1612 * config/sh/sh.c: Include stdlib headers before everything else.
1614 2014-05-02 Jakub Jelinek <jakub@redhat.com>
1616 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
1617 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
1618 (gimplify_adjust_omp_clauses): Simd region is never
1619 directly nested in combined parallel. Instead, for linear
1620 with copyin/copyout, if in combined for simd loop, make decl
1621 firstprivate/lastprivate on OMP_FOR.
1622 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
1623 expand_omp_for_static_chunk): When setting endvar, also set
1624 fd->loop.v to the same value.
1626 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1628 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
1630 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
1632 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
1635 2014-05-02 Marek Polacek <polacek@redhat.com>
1637 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
1639 2014-05-02 Kito Cheng <kito@0xlab.org>
1641 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
1642 to a C expression marco.
1643 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
1644 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
1645 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
1646 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
1647 HONOR_REG_ALLOC_ORDER.
1648 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
1650 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1652 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
1654 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1656 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
1658 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
1660 * tree-if-conv.c (is_cond_scalar_reduction): New function.
1661 (convert_scalar_cond_reduction): Likewise.
1662 (predicate_scalar_phi): Add recognition and transformation
1663 of simple conditioanl reduction to be vectorizable.
1665 2014-05-01 Marek Polacek <polacek@redhat.com>
1668 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
1670 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
1672 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
1673 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
1674 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
1675 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
1676 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
1677 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
1678 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
1679 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
1681 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
1683 * config/arc/arc.opt (mlra): Move comment above option name
1684 to avoid mis-parsing as language options.
1686 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1688 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
1689 * config/sol2.h: ... here.
1690 * config/sol2-10.h: Remove.
1692 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
1693 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
1694 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
1695 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
1696 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
1697 * config/sol2.h: ... here.
1698 (SECTION_NAME_FORMAT): Don't redefine.
1699 (STARTFILE_ARCH32_SPEC): Rename to ...
1700 (STARTFILE_ARCH_SPEC): ... this.
1701 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
1702 * config/sparc/sol2.h: ... here.
1703 (SECTION_NAME_FORMAT): Don't undef.
1704 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
1705 (SUBTARGET_EXTRA_SPECS): Remove.
1706 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
1708 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
1709 (MD_STARTFILE_PREFIX): Remove.
1710 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
1711 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
1712 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
1713 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
1714 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
1715 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
1716 * config/i386/sol2.h: ... here.
1717 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
1718 * config/i386/sol2-bi.h: Remove.
1719 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
1720 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
1722 * config/i386/t-sol2-64: Rename to ...
1723 * config/i386/t-sol2: ... this.
1724 * config/sparc/t-sol2-64: Rename to ...
1725 * config/sparc/t-sol2: ... this.
1727 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
1728 sol2_tm_file_head, sol2_tm_file_tail.
1729 Include ${cpu_type}/sol2.h before sol2.h.
1731 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
1732 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
1733 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
1734 Reflect i386/t-sol2-64 renaming.
1735 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
1736 Reflect sparc/t-sol2-64 renaming.
1738 2014-04-30 Richard Biener <rguenther@suse.de>
1740 * passes.c (execute_function_todo): Move TODO_verify_stmts
1741 and TODO_verify_ssa under the TODO_verify_il umbrella.
1742 * tree-ssa.h (verify_ssa): Adjust prototype.
1743 * tree-ssa.c (verify_ssa): Add parameter to tell whether
1744 we should verify SSA operands.
1745 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
1746 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
1747 whether we should verify whether not throwing stmts have EH info.
1748 * graphite-scop-detection.c (create_sese_edges): Adjust.
1749 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
1750 * tree-eh.c (lower_try_finally_switch): Do not add the
1751 default case label twice.
1753 2014-04-30 Marek Polacek <polacek@redhat.com>
1755 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
1756 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
1757 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
1758 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
1760 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
1762 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
1763 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
1764 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
1765 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
1766 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
1767 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
1768 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
1769 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
1771 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1773 * tree-cfg.c (dump_function_to_file): Dump the return type of
1774 functions, in a line to itself before the function body, mimicking
1775 the layout of a C function.
1777 2014-04-29 Jakub Jelinek <jakub@redhat.com>
1779 PR tree-optimization/60971
1780 * tree-tailcall.c (process_assignment): Reject conversions which
1783 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
1785 * calls.c (initialize_argument_information): Always treat
1786 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
1787 (expand_call): Likewise.
1788 (emit_library_call_calue_1): Likewise.
1789 * expr.c (PUSH_ARGS_REVERSED): Do not define.
1790 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
1793 2014-04-29 Nick Clifton <nickc@redhat.com>
1795 * config/msp430/msp430.md (umulsidi): Fix typo.
1796 (mulhisi3): Enable even inside interrupt handlers.
1797 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
1798 bigger return address pushed in large mode.
1800 2014-04-29 Nick Clifton <nickc@redhat.com>
1802 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
1803 (arc_init_reg_tables): Use a machine_mode enum to iterate over
1805 * config/m32r/m32r.c (init_reg_tables): Likewise.
1806 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
1807 enum to hold the modes.
1809 2014-04-29 Richard Biener <rguenther@suse.de>
1811 * dominance.c (free_dominance_info): Add overload with
1813 (dom_info_state): Likewise.
1814 (dom_info_available_p): Likewise.
1815 * basic-block.h (free_dominance_info, dom_info_state,
1816 dom_info_available_p): Declare overloads.
1817 * passes.c (execute_function_todo): Verify that verifiers
1818 don't change dominator info state. Drop dominator info
1819 for IPA pass invocations.
1820 * cgraph.c (release_function_body): Restore asserts that
1821 dominator information is released.
1823 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
1825 * doc/invoke.texi: Fix typo.
1826 * tree-vrp.c: Fix typos.
1827 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
1829 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1831 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
1833 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1835 * config/aarch64/aarch64-builtins.c
1836 (aarch64_types_storestruct_lane_qualifiers): New.
1837 (TYPES_STORESTRUCT_LANE): Likewise.
1838 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
1839 (st3_lane): Likewise.
1840 (st4_lane): Likewise.
1841 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
1842 (vec_store_lanesci_lane<mode>): Likewise.
1843 (vec_store_lanesxi_lane<mode>): Likewise.
1844 (aarch64_st2_lane<VQ:mode>): Likewise.
1845 (aarch64_st3_lane<VQ:mode>): Likewise.
1846 (aarch64_st4_lane<VQ:mode>): Likewise.
1847 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
1848 * config/aarch64/arm_neon.h
1849 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
1850 use new macro arguments.
1851 (__ST3_LANE_FUNC): Likewise.
1852 (__ST4_LANE_FUNC): Likewise.
1853 * config/aarch64/iterators.md (V_TWO_ELEM): New.
1854 (V_THREE_ELEM): Likewise.
1855 (V_FOUR_ELEM): Likewise.
1857 2014-04-28 David Malcolm <dmalcolm@redhat.com>
1859 * doc/gimple.texi: Replace the description of the now-defunct
1860 union gimple_statement_d with a diagram showing the
1861 gimple_statement_base class hierarchy and its relationships to
1862 the GSS_ and GIMPLE_ enums.
1864 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1866 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
1867 * config/aarch64/aarch64.c
1868 (aarch64_cannot_change_mode_class): Weaken conditions.
1869 (aarch64_modes_tieable_p): New.
1870 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
1872 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
1874 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
1875 (loadsync_<mode>): Change mode.
1876 (load_quadpti, store_quadpti): New.
1877 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
1878 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
1880 2014-04-28 Martin Jambor <mjambor@suse.cz>
1882 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
1883 same alias type as the original statement.
1884 (subreplacement_assignment_data): New type.
1885 (handle_unscalarized_data_in_subtree): New type of parameter,
1886 generate new memory accesses with same alias type as the original
1888 (load_assign_lhs_subreplacements): Likewise.
1889 (sra_modify_constructor_assign): Generate new memory accesses with
1890 same alias type as the original statement.
1892 2014-04-28 Richard Biener <rguenther@suse.de>
1894 * tree-pass.h (TODO_verify_il): Define.
1895 (TODO_verify_all): Complete properly.
1896 * passes.c (execute_function_todo): Move existing loop-closed
1897 SSA verification under TODO_verify_il.
1898 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
1899 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
1900 Fix tree sharing issue.
1902 2014-04-28 Richard Biener <rguenther@suse.de>
1905 * builtins.def (DEF_C11_BUILTIN): Add.
1906 (BUILT_IN_ALIGNED_ALLOC): Likewise.
1907 * coretypes.h (enum function_class): Add function_c11_misc.
1908 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1909 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
1910 (call_may_clobber_ref_p_1): Likewise.
1911 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
1912 (mark_all_reaching_defs_necessary_1): Likewise.
1913 (propagate_necessity): Likewise.
1914 (eliminate_unnecessary_stmts): Likewise.
1915 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
1917 2014-04-28 Richard Biener <rguenther@suse.de>
1919 * tree-vrp.c (vrp_var_may_overflow): Remove.
1920 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
1921 with overflow immediately bump to one before that value and
1922 let iteration figure out overflow status.
1924 2014-04-28 Richard Biener <rguenther@suse.de>
1926 * configure.ac: Do valgrind header checks unconditionally.
1927 Add --enable-valgrind-annotations.
1928 * system.h: Guard valgrind header inclusion with
1929 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
1930 * alloc-pool.c (pool_alloc, pool_free): Use
1931 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
1932 to guard possibly dead code.
1933 * config.in: Regenerated.
1934 * configure: Likewise.
1936 2014-04-28 Jeff Law <law@redhat.com>
1938 PR tree-optimization/60902
1939 * tree-ssa-threadedge.c
1940 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
1941 over real defs when invalidating outputs from statements that do not
1942 produce useful outputs for threading.
1944 2014-04-28 Richard Biener <rguenther@suse.de>
1946 PR tree-optimization/60979
1947 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
1948 SCOPs that end in a block with a successor with abnormal
1951 2014-04-28 Richard Biener <rguenther@suse.de>
1953 * tree-pass.h (execute_pass_list): Adjust prototype.
1954 * passes.c (pass_manager::execute_early_local_passes): Adjust.
1955 (do_per_function): Change callback signature, push all actual
1956 work to the callbals.
1957 (do_per_function_toporder): Likewise.
1958 (execute_function_dump): Adjust.
1959 (execute_function_todo): Likewise.
1960 (clear_last_verified): Likewise.
1961 (verify_curr_properties): Likewise.
1962 (update_properties_after_pass): Likewise.
1963 (execute_pass_list_1): Split out from ...
1964 (execute_pass_list): ... here. Adjust.
1965 (execute_ipa_pass_list): Likewise.
1966 * cgraphunit.c (cgraph_add_new_function): Adjust.
1967 (analyze_function): Likewise.
1968 (expand_function): Likewise.
1969 * cgraph.c (release_function_body): Free dominance info
1970 here instead of asserting it was magically freed elsewhere.
1972 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
1974 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
1975 * configure: Regenerate.
1976 * config/sparc/sparc.opt (muser-mode): New option.
1977 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
1979 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
1980 * doc/invoke.texi (SPARC options): Document -muser-mode.
1982 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
1984 * cselib.c (find_slot_memmode): Delete.
1985 (cselib_hasher): Change compare_type to a struct.
1986 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
1988 (preserve_constants_and_equivs): Adjust for new compare_type.
1989 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
1990 (wrap_constant): Delete.
1991 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
1993 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
1995 * doc/install.texi (Building with profile feedback): Remove
1998 2014-04-26 Tom de Vries <tom@codesourcery.com>
2000 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
2003 2014-04-25 Cary Coutant <ccoutant@google.com>
2006 * dwarf2out.c (should_move_die_to_comdat): A type definition
2007 can contain a subprogram definition, but don't move it to a
2009 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
2010 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
2012 (clone_tree_hash): Rename to...
2013 (clone_tree_partial): ...this; change callers. Copy
2014 DW_TAG_subprogram DIEs as declarations.
2015 (copy_decls_walk): Don't copy children of a declaration into a
2018 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
2021 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
2024 2014-04-25 Jiong Wang <jiong.wang@arm.com>
2026 * config/arm/predicates.md (call_insn_operand): Add long_call check.
2027 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
2029 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
2032 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2034 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
2036 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2038 PR tree-optimization/60930
2039 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
2040 creating a multiply candidate by folding two constant
2041 multiplicands when the result overflows.
2043 2014-04-25 Jakub Jelinek <jakub@redhat.com>
2045 PR tree-optimization/60960
2046 * tree-vect-generic.c (expand_vector_operation): Only call
2047 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
2049 2014-04-25 Tom de Vries <tom@codesourcery.com>
2051 * expr.c (clobber_reg_mode): New function.
2052 * expr.h (clobber_reg): New function.
2054 2014-04-25 Tom de Vries <tom@codesourcery.com>
2056 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
2059 2014-04-25 Radovan Obradovic <robradovic@mips.com>
2060 Tom de Vries <tom@codesourcery.com>
2062 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
2064 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
2065 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
2066 new argument to find_all_hard_reg_sets call.
2068 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2070 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
2071 Use HOST_WIDE_INT_C for mask literal.
2072 (aarch_rev16_shleft_mask_imm_p): Likewise.
2074 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
2077 * config/sparc/sparc.md (ashlsi3_extend): Delete.
2079 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
2081 PR preprocessor/56540
2082 * config/i386/i386-c.c (ix86_target_macros): Define
2083 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
2085 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2087 * configure.ac (tga_func): Remove.
2088 (LIB_TLS_SPEC): Remove.
2089 * configure: Regenerate.
2090 * config.in: Regenerate.
2091 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
2093 2014-04-25 Richard Biener <rguenther@suse.de>
2096 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
2097 call stmt use/clobber sets during stmt walk instead of
2098 walking the possibly incomplete set of caller edges.
2100 2014-04-25 Richard Biener <rguenther@suse.de>
2103 * passes.c (apply_ipa_transforms): Inline into only caller ...
2104 (execute_one_pass): ... here. Properly bring in function
2105 bodies for nodes we want to apply IPA transforms to.
2107 2014-04-24 Cong Hou <congh@google.com>
2109 PR tree-optimization/60896
2110 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
2111 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
2112 (vect_mark_pattern_stmts): Set the def type of all statements in
2113 PATTERN_DEF_SEQ as vect_internal_def.
2115 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
2117 * doc/extend.texi (PowerPC Built-in Functions): Document new
2118 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
2119 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
2121 * config/rs6000/predicates.md (const_0_to_3_operand): New
2122 predicate to match 0..3 integer constants.
2124 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
2125 to support adding miscellaneous builtin functions.
2126 (BU_DFP_MISC_2): Likewise.
2127 (BU_P7_MISC_1): Likewise.
2128 (BU_P7_MISC_2): Likewise.
2129 (BU_P8V_MISC_3): Likewise.
2130 (BU_MISC_1): Likewise.
2131 (BU_MISC_2): Likewise.
2132 (DIVWE): Add extended divide builtin functions.
2135 (DIVWEUO): Likewise.
2139 (DIVDEUO): Likewise.
2140 (DXEX): Add decimal floating-point builtin functions.
2143 (DDEDPDQ): Likewise.
2145 (DENBCDQ): Likewise.
2152 (CDTBCD): Add new BCD builtin functions.
2156 (BCDADD_LT): Likewise.
2157 (BCDADD_EQ): Likewise.
2158 (BCDADD_GT): Likewise.
2159 (BCDADD_OV): Likewise.
2161 (BCDSUB_LT): Likewise.
2162 (BCDSUB_EQ): Likewise.
2163 (BCDSUB_GT): Likewise.
2164 (BCDSUB_OV): Likewise.
2165 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
2166 (UNPACK_TD): Likewise.
2167 (PACK_TF): Likewise.
2168 (UNPACK_TF): Likewise.
2169 (UNPACK_TF_0): Likewise.
2170 (UNPACK_TF_1): Likewise.
2171 (PACK_V1TI): Likewise.
2172 (UNPACK_V1TI): Likewise.
2174 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
2175 support for decimal floating point builtin functions.
2176 (rs6000_expand_ternop_builtin): Add checks for the new builtin
2177 functions that take constant arguments.
2178 (rs6000_invalid_builtin): Add decimal floating point builtin support.
2179 (rs6000_init_builtins): Setup long double, _Decimal64, and
2180 _Decimal128 types for new builtin functions.
2181 (builtin_function_type): Set the unsigned flags appropriately for
2182 the new builtin functions.
2183 (rs6000_opt_masks): Add support for decimal floating point builtin
2186 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
2187 floating point builtin functions.
2188 (RS6000_BTM_COMMON): Likewise.
2189 (RS6000_BTI_long_double): Likewise.
2190 (RS6000_BTI_dfloat64): Likewise.
2191 (RS6000_BTI_dfloat128): Likewise.
2192 (long_double_type_internal_node): Likewise.
2193 (dfloat64_type_internal_node): Likewise.
2194 (dfloat128_type_internal_node): Likewise.
2196 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
2197 2.07 bcd arithmetic instructions.
2198 (UNSPEC_BCDSUB): Likewise.
2199 (UNSPEC_BCD_OVERFLOW): Likewise.
2200 (UNSPEC_BCD_ADD_SUB): Likewise.
2201 (bcd_add_sub): Likewise.
2202 (BCD_TEST): Likewise.
2203 (bcd<bcd_add_sub>): Likewise.
2204 (bcd<bcd_add_sub>_test): Likewise.
2205 (bcd<bcd_add_sub>_test2): Likewise.
2206 (bcd<bcd_add_sub>_<code>): Likewise.
2207 (peephole2 for combined bcd ops): Likewise.
2209 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
2210 decimal floating point builtin functions.
2211 (UNSPEC_DENBCD): Likewise.
2212 (UNSPEC_DXEX): Likewise.
2213 (UNSPEC_DIEX): Likewise.
2214 (UNSPEC_DSCLI): Likewise.
2215 (UNSPEC_DSCRI): Likewise.
2216 (D64_D128): Likewise.
2217 (dfp_suffix): Likewise.
2218 (dfp_ddedpd_<mode>): Likewise.
2219 (dfp_denbcd_<mode>): Likewise.
2220 (dfp_dxex_<mode>): Likewise.
2221 (dfp_diex_<mode>): Likewise.
2222 (dfp_dscli_<mode>): Likewise.
2223 (dfp_dscri_<mode>): Likewise.
2225 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
2227 (UNSPEC_CDTBCD): Likewise.
2228 (UNSPEC_CBCDTD): Likewise.
2229 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
2230 (UNSPEC_DIVEO): Likewise.
2231 (UNSPEC_DIVEU): Likewise.
2232 (UNSPEC_DIVEUO): Likewise.
2233 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
2234 pack/unpack 128-bit types.
2235 (UNSPEC_PACK_128BIT): Likewise.
2236 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
2237 (udiv<mode>3): Use idiv_ldiv mode attribute.
2238 (div<mode>3): Likewise.
2239 (addg6s): Add new BCD builtin functions.
2242 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
2243 (div_extend): Likewise.
2244 (div<div_extend>_<mode>"): Likewise.
2245 (FP128_64): Add support for new builtin functions to pack/unpack
2247 (unpack<mode>): Likewise.
2248 (unpacktf_0): Likewise.
2249 (unpacktf_1): Likewise.
2250 (unpack<mode>_dm): Likewise.
2251 (unpack<mode>_nodm): Likewise.
2252 (pack<mode>): Likewise.
2253 (unpackv1ti): Likewise.
2254 (packv1ti): Likewise.
2256 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
2258 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
2261 2014-04-24 Jakub Jelinek <jakub@redhat.com>
2263 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
2264 * gimplify.c (omp_is_private): Change last argument's type to int.
2265 Only diagnose lastprivate if the simd argument is 1, only diagnose
2266 linear if the simd argument is 2.
2267 (gimplify_omp_for): Adjust omp_is_private callers. When adding
2268 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
2269 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
2270 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
2271 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
2272 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2273 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
2274 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2275 * tree-nested.c (convert_nonlocal_omp_clauses,
2276 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
2278 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
2281 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
2284 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
2286 * flag-types.h (enum ivar_visibility): Add.
2288 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
2290 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
2291 function * argument.
2293 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
2295 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
2297 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2298 Tom de Vries <tom@codesourcery.com>
2300 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
2301 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
2303 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
2304 * emit-rtl.c (try_split): Same.
2306 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2307 Tom de Vries <tom@codesourcery.com>
2309 * common.opt (fuse-caller-save): New option.
2311 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
2313 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
2314 elements for big-endian.
2316 2014-04-24 Richard Biener <rguenther@suse.de>
2318 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
2319 during TER and instead use the sepops interface for expanding
2320 non-GIMPLE_SINGLE_RHS.
2322 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2324 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
2325 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
2327 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2329 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
2330 assembler 64-bit option.
2331 * configure: Regenerate.
2333 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2335 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
2336 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
2337 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
2338 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
2339 (TARGET_CRYPTO): Take TARGET_SIMD into account.
2341 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2343 * config/aarch64/aarch64-builtins.c
2344 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
2345 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
2346 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
2347 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
2349 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
2350 (Vrevsuff): New mode attribute.
2352 2014-04-24 Terry Guo <terry.guo@arm.com>
2354 * config/arm/arm.h (machine_function): Define variable
2355 after_arm_reorg here.
2356 * config/arm/arm.c (after_arm_reorg): Remove the definition.
2357 (arm_split_constant): Update the way to access variable
2360 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
2362 2014-04-23 Tom de Vries <tom@codesourcery.com>
2364 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
2366 2014-04-23 David Malcolm <dmalcolm@redhat.com>
2368 * is-a.h: Update comments to reflect the following changes to the
2369 "pointerness" of the API, making the template parameter match the
2370 return type, allowing use of is-a.h with typedefs of pointers.
2371 (is_a_helper::cast): Return a T rather then a pointer to a T, so
2372 that the return type matches the parameter to the is_a_helper.
2374 (dyn_cast): Likewise.
2376 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
2377 pointer from the is-a.h API.
2379 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
2380 (is_a_helper <cgraph_node *>::test): ...this, matching change to
2382 (is_a_helper <varpool_node>::test): Likewise, convert to...
2383 (is_a_helper <varpool_node *>::test): ...this.
2385 (varpool_first_variable): Update for removal of implicit pointer
2386 from the is-a.h API.
2387 (varpool_next_variable): Likewise.
2388 (varpool_first_static_initializer): Likewise.
2389 (varpool_next_static_initializer): Likewise.
2390 (varpool_first_defined_variable): Likewise.
2391 (varpool_next_defined_variable): Likewise.
2392 (cgraph_first_defined_function): Likewise.
2393 (cgraph_next_defined_function): Likewise.
2394 (cgraph_first_function): Likewise.
2395 (cgraph_next_function): Likewise.
2396 (cgraph_first_function_with_gimple_body): Likewise.
2397 (cgraph_next_function_with_gimple_body): Likewise.
2398 (cgraph_alias_target): Likewise.
2399 (varpool_alias_target): Likewise.
2400 (cgraph_function_or_thunk_node): Likewise.
2401 (varpool_variable_node): Likewise.
2402 (symtab_real_symbol_p): Likewise.
2403 * cgraphunit.c (referred_to_p): Likewise.
2404 (analyze_functions): Likewise.
2405 (handle_alias_pairs): Likewise.
2406 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2407 * gimple-ssa.h (gimple_vuse_op): Likewise.
2408 (gimple_vdef_op): Likewise.
2409 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
2410 * gimple.c (gimple_build_asm_1): Likewise.
2411 (gimple_build_try): Likewise.
2412 (gimple_build_resx): Likewise.
2413 (gimple_build_eh_dispatch): Likewise.
2414 (gimple_build_omp_for): Likewise.
2415 (gimple_omp_for_set_clauses): Likewise.
2417 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
2418 (is_a_helper <gimple_statement_asm *>::test): ...this.
2419 (is_a_helper <gimple_statement_bind>::test): Convert to...
2420 (is_a_helper <gimple_statement_bind *>::test): ...this.
2421 (is_a_helper <gimple_statement_call>::test): Convert to...
2422 (is_a_helper <gimple_statement_call *>::test): ...this.
2423 (is_a_helper <gimple_statement_catch>::test): Convert to...
2424 (is_a_helper <gimple_statement_catch *>::test): ...this.
2425 (is_a_helper <gimple_statement_resx>::test): Convert to...
2426 (is_a_helper <gimple_statement_resx *>::test): ...this.
2427 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
2428 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
2429 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
2430 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
2431 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
2432 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
2433 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
2434 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
2435 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
2436 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
2437 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
2438 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
2439 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
2440 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
2441 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
2442 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
2443 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
2444 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
2445 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
2446 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
2447 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
2448 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
2449 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
2450 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
2451 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
2452 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
2453 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
2454 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
2455 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
2456 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
2457 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
2458 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
2459 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
2460 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
2461 (is_a_helper <gimple_statement_phi>::test): Convert to...
2462 (is_a_helper <gimple_statement_phi *>::test): ...this.
2463 (is_a_helper <gimple_statement_transaction>::test): Convert to...
2464 (is_a_helper <gimple_statement_transaction *>::test): ...this.
2465 (is_a_helper <gimple_statement_try>::test): Convert to...
2466 (is_a_helper <gimple_statement_try *>::test): ...this.
2467 (is_a_helper <gimple_statement_wce>::test): Convert to...
2468 (is_a_helper <gimple_statement_wce *>::test): ...this.
2469 (is_a_helper <const gimple_statement_asm>::test): Convert to...
2470 (is_a_helper <const gimple_statement_asm *>::test): ...this.
2471 (is_a_helper <const gimple_statement_bind>::test): Convert to...
2472 (is_a_helper <const gimple_statement_bind *>::test): ...this.
2473 (is_a_helper <const gimple_statement_call>::test): Convert to...
2474 (is_a_helper <const gimple_statement_call *>::test): ...this.
2475 (is_a_helper <const gimple_statement_catch>::test): Convert to...
2476 (is_a_helper <const gimple_statement_catch *>::test): ...this.
2477 (is_a_helper <const gimple_statement_resx>::test): Convert to...
2478 (is_a_helper <const gimple_statement_resx *>::test): ...this.
2479 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
2480 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
2481 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
2482 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
2483 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
2485 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
2487 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
2489 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
2491 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
2492 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
2493 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
2495 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
2496 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
2498 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
2499 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
2500 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
2501 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
2502 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
2503 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
2505 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
2506 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
2507 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
2508 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
2510 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
2511 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
2512 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
2513 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
2514 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
2515 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
2516 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
2517 (is_a_helper <const gimple_statement_phi>::test): Convert to...
2518 (is_a_helper <const gimple_statement_phi *>::test): ...this.
2519 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
2520 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
2521 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
2522 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
2523 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
2524 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
2525 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
2527 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
2529 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
2530 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
2532 (gimple_use_ops): Update for removal of implicit pointer from the
2534 (gimple_set_use_ops): Likewise.
2535 (gimple_vuse): Likewise.
2536 (gimple_vdef): Likewise.
2537 (gimple_vuse_ptr): Likewise.
2538 (gimple_vdef_ptr): Likewise.
2539 (gimple_set_vuse): Likewise.
2540 (gimple_set_vdef): Likewise.
2541 (gimple_omp_return_set_lhs): Likewise.
2542 (gimple_omp_return_lhs): Likewise.
2543 (gimple_omp_return_lhs_ptr): Likewise.
2544 (gimple_call_fntype): Likewise.
2545 (gimple_call_set_fntype): Likewise.
2546 (gimple_call_set_internal_fn): Likewise.
2547 (gimple_call_use_set): Likewise.
2548 (gimple_call_clobber_set): Likewise.
2549 (gimple_bind_vars): Likewise.
2550 (gimple_bind_set_vars): Likewise.
2551 (gimple_bind_body_ptr): Likewise.
2552 (gimple_bind_set_body): Likewise.
2553 (gimple_bind_add_stmt): Likewise.
2554 (gimple_bind_block): Likewise.
2555 (gimple_bind_set_block): Likewise.
2556 (gimple_asm_ninputs): Likewise.
2557 (gimple_asm_noutputs): Likewise.
2558 (gimple_asm_nclobbers): Likewise.
2559 (gimple_asm_nlabels): Likewise.
2560 (gimple_asm_input_op): Likewise.
2561 (gimple_asm_input_op_ptr): Likewise.
2562 (gimple_asm_output_op): Likewise.
2563 (gimple_asm_output_op_ptr): Likewise.
2564 (gimple_asm_set_output_op): Likewise.
2565 (gimple_asm_clobber_op): Likewise.
2566 (gimple_asm_set_clobber_op): Likewise.
2567 (gimple_asm_label_op): Likewise.
2568 (gimple_asm_set_label_op): Likewise.
2569 (gimple_asm_string): Likewise.
2570 (gimple_catch_types): Likewise.
2571 (gimple_catch_types_ptr): Likewise.
2572 (gimple_catch_handler_ptr): Likewise.
2573 (gimple_catch_set_types): Likewise.
2574 (gimple_catch_set_handler): Likewise.
2575 (gimple_eh_filter_types): Likewise.
2576 (gimple_eh_filter_types_ptr): Likewise.
2577 (gimple_eh_filter_failure_ptr): Likewise.
2578 (gimple_eh_filter_set_types): Likewise.
2579 (gimple_eh_filter_set_failure): Likewise.
2580 (gimple_eh_must_not_throw_fndecl): Likewise.
2581 (gimple_eh_must_not_throw_set_fndecl): Likewise.
2582 (gimple_eh_else_n_body_ptr): Likewise.
2583 (gimple_eh_else_e_body_ptr): Likewise.
2584 (gimple_eh_else_set_n_body): Likewise.
2585 (gimple_eh_else_set_e_body): Likewise.
2586 (gimple_try_eval_ptr): Likewise.
2587 (gimple_try_cleanup_ptr): Likewise.
2588 (gimple_try_set_eval): Likewise.
2589 (gimple_try_set_cleanup): Likewise.
2590 (gimple_wce_cleanup_ptr): Likewise.
2591 (gimple_wce_set_cleanup): Likewise.
2592 (gimple_phi_capacity): Likewise.
2593 (gimple_phi_num_args): Likewise.
2594 (gimple_phi_result): Likewise.
2595 (gimple_phi_result_ptr): Likewise.
2596 (gimple_phi_set_result): Likewise.
2597 (gimple_phi_arg): Likewise.
2598 (gimple_phi_set_arg): Likewise.
2599 (gimple_resx_region): Likewise.
2600 (gimple_resx_set_region): Likewise.
2601 (gimple_eh_dispatch_region): Likewise.
2602 (gimple_eh_dispatch_set_region): Likewise.
2603 (gimple_omp_critical_name): Likewise.
2604 (gimple_omp_critical_name_ptr): Likewise.
2605 (gimple_omp_critical_set_name): Likewise.
2606 (gimple_omp_for_clauses): Likewise.
2607 (gimple_omp_for_clauses_ptr): Likewise.
2608 (gimple_omp_for_set_clauses): Likewise.
2609 (gimple_omp_for_collapse): Likewise.
2610 (gimple_omp_for_index): Likewise.
2611 (gimple_omp_for_index_ptr): Likewise.
2612 (gimple_omp_for_set_index): Likewise.
2613 (gimple_omp_for_initial): Likewise.
2614 (gimple_omp_for_initial_ptr): Likewise.
2615 (gimple_omp_for_set_initial): Likewise.
2616 (gimple_omp_for_final): Likewise.
2617 (gimple_omp_for_final_ptr): Likewise.
2618 (gimple_omp_for_set_final): Likewise.
2619 (gimple_omp_for_incr): Likewise.
2620 (gimple_omp_for_incr_ptr): Likewise.
2621 (gimple_omp_for_set_incr): Likewise.
2622 (gimple_omp_for_pre_body_ptr): Likewise.
2623 (gimple_omp_for_set_pre_body): Likewise.
2624 (gimple_omp_parallel_clauses): Likewise.
2625 (gimple_omp_parallel_clauses_ptr): Likewise.
2626 (gimple_omp_parallel_set_clauses): Likewise.
2627 (gimple_omp_parallel_child_fn): Likewise.
2628 (gimple_omp_parallel_child_fn_ptr): Likewise.
2629 (gimple_omp_parallel_set_child_fn): Likewise.
2630 (gimple_omp_parallel_data_arg): Likewise.
2631 (gimple_omp_parallel_data_arg_ptr): Likewise.
2632 (gimple_omp_parallel_set_data_arg): Likewise.
2633 (gimple_omp_task_clauses): Likewise.
2634 (gimple_omp_task_clauses_ptr): Likewise.
2635 (gimple_omp_task_set_clauses): Likewise.
2636 (gimple_omp_task_child_fn): Likewise.
2637 (gimple_omp_task_child_fn_ptr): Likewise.
2638 (gimple_omp_task_set_child_fn): Likewise.
2639 (gimple_omp_task_data_arg): Likewise.
2640 (gimple_omp_task_data_arg_ptr): Likewise.
2641 (gimple_omp_task_set_data_arg): Likewise.
2642 (gimple_omp_taskreg_clauses): Likewise.
2643 (gimple_omp_taskreg_clauses_ptr): Likewise.
2644 (gimple_omp_taskreg_set_clauses): Likewise.
2645 (gimple_omp_taskreg_child_fn): Likewise.
2646 (gimple_omp_taskreg_child_fn_ptr): Likewise.
2647 (gimple_omp_taskreg_set_child_fn): Likewise.
2648 (gimple_omp_taskreg_data_arg): Likewise.
2649 (gimple_omp_taskreg_data_arg_ptr): Likewise.
2650 (gimple_omp_taskreg_set_data_arg): Likewise.
2651 (gimple_omp_task_copy_fn): Likewise.
2652 (gimple_omp_task_copy_fn_ptr): Likewise.
2653 (gimple_omp_task_set_copy_fn): Likewise.
2654 (gimple_omp_task_arg_size): Likewise.
2655 (gimple_omp_task_arg_size_ptr): Likewise.
2656 (gimple_omp_task_set_arg_size): Likewise.
2657 (gimple_omp_task_arg_align): Likewise.
2658 (gimple_omp_task_arg_align_ptr): Likewise.
2659 (gimple_omp_task_set_arg_align): Likewise.
2660 (gimple_omp_single_clauses): Likewise.
2661 (gimple_omp_single_clauses_ptr): Likewise.
2662 (gimple_omp_single_set_clauses): Likewise.
2663 (gimple_omp_target_clauses): Likewise.
2664 (gimple_omp_target_clauses_ptr): Likewise.
2665 (gimple_omp_target_set_clauses): Likewise.
2666 (gimple_omp_target_child_fn): Likewise.
2667 (gimple_omp_target_child_fn_ptr): Likewise.
2668 (gimple_omp_target_set_child_fn): Likewise.
2669 (gimple_omp_target_data_arg): Likewise.
2670 (gimple_omp_target_data_arg_ptr): Likewise.
2671 (gimple_omp_target_set_data_arg): Likewise.
2672 (gimple_omp_teams_clauses): Likewise.
2673 (gimple_omp_teams_clauses_ptr): Likewise.
2674 (gimple_omp_teams_set_clauses): Likewise.
2675 (gimple_omp_sections_clauses): Likewise.
2676 (gimple_omp_sections_clauses_ptr): Likewise.
2677 (gimple_omp_sections_set_clauses): Likewise.
2678 (gimple_omp_sections_control): Likewise.
2679 (gimple_omp_sections_control_ptr): Likewise.
2680 (gimple_omp_sections_set_control): Likewise.
2681 (gimple_omp_for_set_cond): Likewise.
2682 (gimple_omp_for_cond): Likewise.
2683 (gimple_omp_atomic_store_set_val): Likewise.
2684 (gimple_omp_atomic_store_val): Likewise.
2685 (gimple_omp_atomic_store_val_ptr): Likewise.
2686 (gimple_omp_atomic_load_set_lhs): Likewise.
2687 (gimple_omp_atomic_load_lhs): Likewise.
2688 (gimple_omp_atomic_load_lhs_ptr): Likewise.
2689 (gimple_omp_atomic_load_set_rhs): Likewise.
2690 (gimple_omp_atomic_load_rhs): Likewise.
2691 (gimple_omp_atomic_load_rhs_ptr): Likewise.
2692 (gimple_omp_continue_control_def): Likewise.
2693 (gimple_omp_continue_control_def_ptr): Likewise.
2694 (gimple_omp_continue_set_control_def): Likewise.
2695 (gimple_omp_continue_control_use): Likewise.
2696 (gimple_omp_continue_control_use_ptr): Likewise.
2697 (gimple_omp_continue_set_control_use): Likewise.
2698 (gimple_transaction_body_ptr): Likewise.
2699 (gimple_transaction_label): Likewise.
2700 (gimple_transaction_label_ptr): Likewise.
2701 (gimple_transaction_set_body): Likewise.
2702 (gimple_transaction_set_label): Likewise.
2704 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
2705 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2706 * ipa-ref.c (ipa_record_reference): Likewise.
2707 * ipa-reference.c (analyze_function): Likewise.
2708 (ipa_reference_write_optimization_summary): Likewise.
2709 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
2710 (address_taken_from_non_vtable_p): Likewise.
2711 (comdat_can_be_unshared_p_1): Likewise.
2712 * lto-cgraph.c (lto_output_ref): Likewise.
2713 (add_references): Likewise.
2714 (compute_ltrans_boundary): Likewise.
2715 (output_symtab): Likewise.
2716 (input_ref): Likewise.
2717 (input_cgraph_1): Likewise.
2718 (output_cgraph_opt_summary): Likewise.
2719 * lto-streamer-out.c (lto_output): Likewise.
2720 (output_symbol_p): Likewise.
2721 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
2722 (lsei_start_function_in_partition): Likewise.
2723 (lsei_next_variable_in_partition): Likewise.
2724 (lsei_start_variable_in_partition): Likewise.
2725 * symtab.c (insert_to_assembler_name_hash): Likewise.
2726 (unlink_from_assembler_name_hash): Likewise.
2727 (symtab_unregister_node): Likewise.
2728 (symtab_remove_node): Likewise.
2729 (dump_symtab_node): Likewise.
2730 (verify_symtab_base): Likewise.
2731 (verify_symtab_node): Likewise.
2732 (symtab_make_decl_local): Likewise.
2733 (symtab_alias_ultimate_target): Likewise.
2734 (symtab_resolve_alias): Likewise.
2735 (symtab_get_symbol_partitioning_class): Likewise.
2736 * tree-phinodes.c (allocate_phi_node): Likewise.
2737 (reserve_phi_args_for_new_edge): Likewise.
2738 (remove_phi_args): Likewise.
2739 * varpool.c (varpool_node_for_asm): Likewise.
2740 (varpool_remove_unreferenced_decls): Likewise.
2742 2014-04-23 Jeff Law <law@redhat.com>
2744 PR tree-optimization/60902
2745 * tree-ssa-threadedge.c
2746 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
2747 invalidate outputs from statements that do not produce useful
2748 outputs for threading.
2750 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2752 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
2753 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
2754 machine descriptions for Stack Smashing Protector.
2756 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
2758 * aarch64.md (<optab>_rol<mode>3): New pattern.
2759 (<optab>_rolsi3_uxtw): Likewise.
2760 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
2762 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
2764 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
2765 (arm_cortex_a12_tune): Likewise.
2767 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2769 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
2771 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2773 * config/arm/arm.md (arm_rev16si2): New pattern.
2774 (arm_rev16si2_alt): Likewise.
2775 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
2777 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2779 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
2780 (rev16<mode>2_alt): Likewise.
2781 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
2782 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
2783 (aarch_rev16_shleft_mask_imm_p): Likewise.
2784 (aarch_rev16_p_1): Likewise.
2785 (aarch_rev16_p): Likewise.
2786 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
2787 (aarch_rev16_shright_mask_imm_p): Likewise.
2788 (aarch_rev16_shleft_mask_imm_p): Likewise.
2790 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2792 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
2793 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
2795 (cortex_a53_extra_costs): Likewise.
2796 (cortex_a57_extra_costs): Likewise.
2797 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
2798 (cortexa7_extra_costs): Likewise.
2799 (cortexa8_extra_costs): Likewise.
2800 (cortexa12_extra_costs): Likewise.
2801 (cortexa15_extra_costs): Likewise.
2802 (v7m_extra_costs): Likewise.
2803 (arm_new_rtx_costs): Handle BSWAP.
2805 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2807 * config/arm/arm.c (cortexa8_extra_costs): New table.
2808 (arm_cortex_a8_tune): New tuning struct.
2809 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
2811 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2813 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
2815 2014-04-23 Richard Biener <rguenther@suse.de>
2817 * Makefile.in (OBJS): Remove loop-unswitch.o.
2818 * tree-pass.h (make_pass_rtl_unswitch): Remove.
2819 * passes.def (pass_rtl_unswitch): Likewise.
2820 * loop-init.c (gate_rtl_unswitch): Likewise.
2821 (rtl_unswitch): Likewise.
2822 (pass_data_rtl_unswitch): Likewise.
2823 (pass_rtl_unswitch): Likewise.
2824 (make_pass_rtl_unswitch): Likewise.
2825 * rtl.h (reversed_condition): Likewise.
2826 (compare_and_jump_seq): Likewise.
2827 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
2829 * loop-unroll.c (compare_and_jump_seq): Likewise.
2831 2014-04-23 Richard Biener <rguenther@suse.de>
2833 PR tree-optimization/60903
2834 * tree-ssa-loop-im.c (analyze_memory_references): Remove
2835 commented code block.
2836 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
2837 loop flags to newly created BBs and edges.
2839 2014-04-23 Nick Clifton <nickc@redhat.com>
2841 * config/msp430/msp430.c (msp430_handle_option): Move function
2843 (msp430_option_override): Simplify mcu and mcpu option handling.
2844 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
2845 support for -mhwmult command line option.
2846 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
2847 -mhwmult command line option.
2848 (msp430_hwmult_enabled): Delete.
2849 (msp43o_output_labelref): Add support for -mhwmult command line option.
2850 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
2851 (umulsidi3): Likewise.
2852 * config/msp430/msp430.opt (mmcu): Add Report attribute.
2853 (mcpu, mlarge, msmall): Likewise.
2854 (mhwmult): New option.
2855 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
2857 (msp430_is_f5_mcu): Remove prototype.
2858 (msp430_use_f5_series_hwmult): Add prototype.
2859 * config/msp430/msp430-opts.h: New file.
2860 * common/config/msp430: New directory.
2861 * common/config/msp430/msp430-common.c: New file.
2862 * config.gcc (msp430): Remove target_has_targetm_common.
2863 * doc/invoke.texi: Document -mhwmult command line option.
2865 2014-04-23 Nick Clifton <nickc@redhat.com>
2867 * config/i386/cygwin.h (ENDFILE_SPEC): Include
2868 default-manifest.o if it can be found in the search path.
2869 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
2871 2014-04-23 Terry Guo <terry.guo@arm.com>
2873 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
2875 2014-04-23 Richard Biener <rguenther@suse.de>
2878 * tree-inline.c (declare_return_variable): Use mark_addressable.
2880 2014-04-23 Richard Biener <rguenther@suse.de>
2883 * loop-init.c (loop_optimizer_init): Make sure to apply
2884 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
2886 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2889 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
2891 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
2892 if flag_sanitize_undefined_trap_on_error.
2893 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
2894 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
2895 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
2896 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
2897 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
2898 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
2899 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
2900 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
2901 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
2902 * ubsan.c (ubsan_instrument_unreachable): Return
2903 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
2904 (ubsan_expand_null_ifn): Emit __builtin_trap ()
2905 if flag_sanitize_undefined_trap_on_error and
2906 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
2907 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
2908 instrument_bool_enum_load): Emit __builtin_trap () if
2909 flag_sanitize_undefined_trap_on_error and
2910 __builtin_handle_*_abort () if !flag_sanitize_recover.
2911 * doc/invoke.texi (-fsanitize-recover,
2912 -fsanitize-undefined-trap-on-error): Document.
2914 2014-04-22 Christian Bruel <christian.bruel@st.com>
2916 * config/sh/sh.md (mov<mode>): Replace movQIHI.
2917 Force immediates to SImode.
2919 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
2921 * config/nios2/nios2.md (UNSPEC_ROUND): New.
2923 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
2924 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
2925 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
2926 (nios2_fpu_insn): Add entry for round.
2927 (N2FPU_NO_ERRNO_P): Define.
2928 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
2930 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
2932 2014-04-22 Richard Henderson <rth@redhat.com>
2934 * config/aarch64/aarch64 (addti3, subti3): New expanders.
2935 (add<GPI>3_compare0): Remove leading * from name.
2936 (add<GPI>3_carryin): Likewise.
2937 (sub<GPI>3_compare0): Likewise.
2938 (sub<GPI>3_carryin): Likewise.
2939 (<su_optab>mulditi3): New expander.
2940 (multi3): New expander.
2941 (madd<GPI>): Remove leading * from name.
2943 2014-04-22 Martin Jambor <mjambor@suse.cz>
2945 * cgraphclones.c (cgraph_function_versioning): Copy
2946 ipa_transforms_to_apply instead of asserting it is empty.
2948 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
2951 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
2952 on count_exp to get mode.
2954 2014-04-22 Andrew Pinski <apinski@cavium.com>
2956 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
2957 Handle TLS for ILP32.
2958 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
2959 (tlsie_small_<mode>): this and handle PTR.
2960 (tlsie_small_sidi): New pattern.
2961 (tlsle_small): Change to an expand to handle ILP32.
2962 (tlsle_small_<mode>): New pattern.
2963 (tlsdesc_small): Rename to ...
2964 (tlsdesc_small_<mode>): this and handle PTR.
2966 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2968 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
2970 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2972 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2973 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
2974 (aarch64_types_signed_poly_qualifiers): Likewise.
2975 (aarch64_types_unsigned_signed_qualifiers): Likewise.
2976 (aarch64_types_poly_signed_qualifiers): Likewise.
2977 (TYPES_REINTERP_SS): Type macro added.
2978 (TYPES_REINTERP_SU): Likewise.
2979 (TYPES_REINTERP_SP): Likewise.
2980 (TYPES_REINTERP_US): Likewise.
2981 (TYPES_REINTERP_PS): Likewise.
2982 (aarch64_fold_builtin): New expression folding added.
2983 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
2984 Declarations removed.
2985 (REINTERP_SS): Declarations added.
2986 (REINTERP_US): Likewise.
2987 (REINTERP_PS): Likewise.
2988 (REINTERP_SU): Likewise.
2989 (REINTERP_SP): Likewise.
2990 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
2991 (vreinterpretq_p8_f64): Likewise.
2992 (vreinterpret_p16_f64): Likewise.
2993 (vreinterpretq_p16_f64): Likewise.
2994 (vreinterpret_f32_f64): Likewise.
2995 (vreinterpretq_f32_f64): Likewise.
2996 (vreinterpret_f64_f32): Likewise.
2997 (vreinterpret_f64_p8): Likewise.
2998 (vreinterpret_f64_p16): Likewise.
2999 (vreinterpret_f64_s8): Likewise.
3000 (vreinterpret_f64_s16): Likewise.
3001 (vreinterpret_f64_s32): Likewise.
3002 (vreinterpret_f64_s64): Likewise.
3003 (vreinterpret_f64_u8): Likewise.
3004 (vreinterpret_f64_u16): Likewise.
3005 (vreinterpret_f64_u32): Likewise.
3006 (vreinterpret_f64_u64): Likewise.
3007 (vreinterpretq_f64_f32): Likewise.
3008 (vreinterpretq_f64_p8): Likewise.
3009 (vreinterpretq_f64_p16): Likewise.
3010 (vreinterpretq_f64_s8): Likewise.
3011 (vreinterpretq_f64_s16): Likewise.
3012 (vreinterpretq_f64_s32): Likewise.
3013 (vreinterpretq_f64_s64): Likewise.
3014 (vreinterpretq_f64_u8): Likewise.
3015 (vreinterpretq_f64_u16): Likewise.
3016 (vreinterpretq_f64_u32): Likewise.
3017 (vreinterpretq_f64_u64): Likewise.
3018 (vreinterpret_s64_f64): Likewise.
3019 (vreinterpretq_s64_f64): Likewise.
3020 (vreinterpret_u64_f64): Likewise.
3021 (vreinterpretq_u64_f64): Likewise.
3022 (vreinterpret_s8_f64): Likewise.
3023 (vreinterpretq_s8_f64): Likewise.
3024 (vreinterpret_s16_f64): Likewise.
3025 (vreinterpretq_s16_f64): Likewise.
3026 (vreinterpret_s32_f64): Likewise.
3027 (vreinterpretq_s32_f64): Likewise.
3028 (vreinterpret_u8_f64): Likewise.
3029 (vreinterpretq_u8_f64): Likewise.
3030 (vreinterpret_u16_f64): Likewise.
3031 (vreinterpretq_u16_f64): Likewise.
3032 (vreinterpret_u32_f64): Likewise.
3033 (vreinterpretq_u32_f64): Likewise.
3035 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3037 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
3038 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
3039 (vreinterpret_p8_s8): Likewise.
3040 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
3041 (vreinterpret_p8_s16): Likewise.
3042 (vreinterpret_p8_s32): Likewise.
3043 (vreinterpret_p8_s64): Likewise.
3044 (vreinterpret_p8_f32): Likewise.
3045 (vreinterpret_p8_u8): Likewise.
3046 (vreinterpret_p8_u16): Likewise.
3047 (vreinterpret_p8_u32): Likewise.
3048 (vreinterpret_p8_u64): Likewise.
3049 (vreinterpret_p8_p16): Likewise.
3050 (vreinterpretq_p8_s8): Likewise.
3051 (vreinterpretq_p8_s16): Likewise.
3052 (vreinterpretq_p8_s32): Likewise.
3053 (vreinterpretq_p8_s64): Likewise.
3054 (vreinterpretq_p8_f32): Likewise.
3055 (vreinterpretq_p8_u8): Likewise.
3056 (vreinterpretq_p8_u16): Likewise.
3057 (vreinterpretq_p8_u32): Likewise.
3058 (vreinterpretq_p8_u64): Likewise.
3059 (vreinterpretq_p8_p16): Likewise.
3060 (vreinterpret_p16_s8): Likewise.
3061 (vreinterpret_p16_s16): Likewise.
3062 (vreinterpret_p16_s32): Likewise.
3063 (vreinterpret_p16_s64): Likewise.
3064 (vreinterpret_p16_f32): Likewise.
3065 (vreinterpret_p16_u8): Likewise.
3066 (vreinterpret_p16_u16): Likewise.
3067 (vreinterpret_p16_u32): Likewise.
3068 (vreinterpret_p16_u64): Likewise.
3069 (vreinterpret_p16_p8): Likewise.
3070 (vreinterpretq_p16_s8): Likewise.
3071 (vreinterpretq_p16_s16): Likewise.
3072 (vreinterpretq_p16_s32): Likewise.
3073 (vreinterpretq_p16_s64): Likewise.
3074 (vreinterpretq_p16_f32): Likewise.
3075 (vreinterpretq_p16_u8): Likewise.
3076 (vreinterpretq_p16_u16): Likewise.
3077 (vreinterpretq_p16_u32): Likewise.
3078 (vreinterpretq_p16_u64): Likewise.
3079 (vreinterpretq_p16_p8): Likewise.
3080 (vreinterpret_f32_s8): Likewise.
3081 (vreinterpret_f32_s16): Likewise.
3082 (vreinterpret_f32_s32): Likewise.
3083 (vreinterpret_f32_s64): Likewise.
3084 (vreinterpret_f32_u8): Likewise.
3085 (vreinterpret_f32_u16): Likewise.
3086 (vreinterpret_f32_u32): Likewise.
3087 (vreinterpret_f32_u64): Likewise.
3088 (vreinterpret_f32_p8): Likewise.
3089 (vreinterpret_f32_p16): Likewise.
3090 (vreinterpretq_f32_s8): Likewise.
3091 (vreinterpretq_f32_s16): Likewise.
3092 (vreinterpretq_f32_s32): Likewise.
3093 (vreinterpretq_f32_s64): Likewise.
3094 (vreinterpretq_f32_u8): Likewise.
3095 (vreinterpretq_f32_u16): Likewise.
3096 (vreinterpretq_f32_u32): Likewise.
3097 (vreinterpretq_f32_u64): Likewise.
3098 (vreinterpretq_f32_p8): Likewise.
3099 (vreinterpretq_f32_p16): Likewise.
3100 (vreinterpret_s64_s8): Likewise.
3101 (vreinterpret_s64_s16): Likewise.
3102 (vreinterpret_s64_s32): Likewise.
3103 (vreinterpret_s64_f32): Likewise.
3104 (vreinterpret_s64_u8): Likewise.
3105 (vreinterpret_s64_u16): Likewise.
3106 (vreinterpret_s64_u32): Likewise.
3107 (vreinterpret_s64_u64): Likewise.
3108 (vreinterpret_s64_p8): Likewise.
3109 (vreinterpret_s64_p16): Likewise.
3110 (vreinterpretq_s64_s8): Likewise.
3111 (vreinterpretq_s64_s16): Likewise.
3112 (vreinterpretq_s64_s32): Likewise.
3113 (vreinterpretq_s64_f32): Likewise.
3114 (vreinterpretq_s64_u8): Likewise.
3115 (vreinterpretq_s64_u16): Likewise.
3116 (vreinterpretq_s64_u32): Likewise.
3117 (vreinterpretq_s64_u64): Likewise.
3118 (vreinterpretq_s64_p8): Likewise.
3119 (vreinterpretq_s64_p16): Likewise.
3120 (vreinterpret_u64_s8): Likewise.
3121 (vreinterpret_u64_s16): Likewise.
3122 (vreinterpret_u64_s32): Likewise.
3123 (vreinterpret_u64_s64): Likewise.
3124 (vreinterpret_u64_f32): Likewise.
3125 (vreinterpret_u64_u8): Likewise.
3126 (vreinterpret_u64_u16): Likewise.
3127 (vreinterpret_u64_u32): Likewise.
3128 (vreinterpret_u64_p8): Likewise.
3129 (vreinterpret_u64_p16): Likewise.
3130 (vreinterpretq_u64_s8): Likewise.
3131 (vreinterpretq_u64_s16): Likewise.
3132 (vreinterpretq_u64_s32): Likewise.
3133 (vreinterpretq_u64_s64): Likewise.
3134 (vreinterpretq_u64_f32): Likewise.
3135 (vreinterpretq_u64_u8): Likewise.
3136 (vreinterpretq_u64_u16): Likewise.
3137 (vreinterpretq_u64_u32): Likewise.
3138 (vreinterpretq_u64_p8): Likewise.
3139 (vreinterpretq_u64_p16): Likewise.
3140 (vreinterpret_s8_s16): Likewise.
3141 (vreinterpret_s8_s32): Likewise.
3142 (vreinterpret_s8_s64): Likewise.
3143 (vreinterpret_s8_f32): Likewise.
3144 (vreinterpret_s8_u8): Likewise.
3145 (vreinterpret_s8_u16): Likewise.
3146 (vreinterpret_s8_u32): Likewise.
3147 (vreinterpret_s8_u64): Likewise.
3148 (vreinterpret_s8_p8): Likewise.
3149 (vreinterpret_s8_p16): Likewise.
3150 (vreinterpretq_s8_s16): Likewise.
3151 (vreinterpretq_s8_s32): Likewise.
3152 (vreinterpretq_s8_s64): Likewise.
3153 (vreinterpretq_s8_f32): Likewise.
3154 (vreinterpretq_s8_u8): Likewise.
3155 (vreinterpretq_s8_u16): Likewise.
3156 (vreinterpretq_s8_u32): Likewise.
3157 (vreinterpretq_s8_u64): Likewise.
3158 (vreinterpretq_s8_p8): Likewise.
3159 (vreinterpretq_s8_p16): Likewise.
3160 (vreinterpret_s16_s8): Likewise.
3161 (vreinterpret_s16_s32): Likewise.
3162 (vreinterpret_s16_s64): Likewise.
3163 (vreinterpret_s16_f32): Likewise.
3164 (vreinterpret_s16_u8): Likewise.
3165 (vreinterpret_s16_u16): Likewise.
3166 (vreinterpret_s16_u32): Likewise.
3167 (vreinterpret_s16_u64): Likewise.
3168 (vreinterpret_s16_p8): Likewise.
3169 (vreinterpret_s16_p16): Likewise.
3170 (vreinterpretq_s16_s8): Likewise.
3171 (vreinterpretq_s16_s32): Likewise.
3172 (vreinterpretq_s16_s64): Likewise.
3173 (vreinterpretq_s16_f32): Likewise.
3174 (vreinterpretq_s16_u8): Likewise.
3175 (vreinterpretq_s16_u16): Likewise.
3176 (vreinterpretq_s16_u32): Likewise.
3177 (vreinterpretq_s16_u64): Likewise.
3178 (vreinterpretq_s16_p8): Likewise.
3179 (vreinterpretq_s16_p16): Likewise.
3180 (vreinterpret_s32_s8): Likewise.
3181 (vreinterpret_s32_s16): Likewise.
3182 (vreinterpret_s32_s64): Likewise.
3183 (vreinterpret_s32_f32): Likewise.
3184 (vreinterpret_s32_u8): Likewise.
3185 (vreinterpret_s32_u16): Likewise.
3186 (vreinterpret_s32_u32): Likewise.
3187 (vreinterpret_s32_u64): Likewise.
3188 (vreinterpret_s32_p8): Likewise.
3189 (vreinterpret_s32_p16): Likewise.
3190 (vreinterpretq_s32_s8): Likewise.
3191 (vreinterpretq_s32_s16): Likewise.
3192 (vreinterpretq_s32_s64): Likewise.
3193 (vreinterpretq_s32_f32): Likewise.
3194 (vreinterpretq_s32_u8): Likewise.
3195 (vreinterpretq_s32_u16): Likewise.
3196 (vreinterpretq_s32_u32): Likewise.
3197 (vreinterpretq_s32_u64): Likewise.
3198 (vreinterpretq_s32_p8): Likewise.
3199 (vreinterpretq_s32_p16): Likewise.
3200 (vreinterpret_u8_s8): Likewise.
3201 (vreinterpret_u8_s16): Likewise.
3202 (vreinterpret_u8_s32): Likewise.
3203 (vreinterpret_u8_s64): Likewise.
3204 (vreinterpret_u8_f32): Likewise.
3205 (vreinterpret_u8_u16): Likewise.
3206 (vreinterpret_u8_u32): Likewise.
3207 (vreinterpret_u8_u64): Likewise.
3208 (vreinterpret_u8_p8): Likewise.
3209 (vreinterpret_u8_p16): Likewise.
3210 (vreinterpretq_u8_s8): Likewise.
3211 (vreinterpretq_u8_s16): Likewise.
3212 (vreinterpretq_u8_s32): Likewise.
3213 (vreinterpretq_u8_s64): Likewise.
3214 (vreinterpretq_u8_f32): Likewise.
3215 (vreinterpretq_u8_u16): Likewise.
3216 (vreinterpretq_u8_u32): Likewise.
3217 (vreinterpretq_u8_u64): Likewise.
3218 (vreinterpretq_u8_p8): Likewise.
3219 (vreinterpretq_u8_p16): Likewise.
3220 (vreinterpret_u16_s8): Likewise.
3221 (vreinterpret_u16_s16): Likewise.
3222 (vreinterpret_u16_s32): Likewise.
3223 (vreinterpret_u16_s64): Likewise.
3224 (vreinterpret_u16_f32): Likewise.
3225 (vreinterpret_u16_u8): Likewise.
3226 (vreinterpret_u16_u32): Likewise.
3227 (vreinterpret_u16_u64): Likewise.
3228 (vreinterpret_u16_p8): Likewise.
3229 (vreinterpret_u16_p16): Likewise.
3230 (vreinterpretq_u16_s8): Likewise.
3231 (vreinterpretq_u16_s16): Likewise.
3232 (vreinterpretq_u16_s32): Likewise.
3233 (vreinterpretq_u16_s64): Likewise.
3234 (vreinterpretq_u16_f32): Likewise.
3235 (vreinterpretq_u16_u8): Likewise.
3236 (vreinterpretq_u16_u32): Likewise.
3237 (vreinterpretq_u16_u64): Likewise.
3238 (vreinterpretq_u16_p8): Likewise.
3239 (vreinterpretq_u16_p16): Likewise.
3240 (vreinterpret_u32_s8): Likewise.
3241 (vreinterpret_u32_s16): Likewise.
3242 (vreinterpret_u32_s32): Likewise.
3243 (vreinterpret_u32_s64): Likewise.
3244 (vreinterpret_u32_f32): Likewise.
3245 (vreinterpret_u32_u8): Likewise.
3246 (vreinterpret_u32_u16): Likewise.
3247 (vreinterpret_u32_u64): Likewise.
3248 (vreinterpret_u32_p8): Likewise.
3249 (vreinterpret_u32_p16): Likewise.
3250 (vreinterpretq_u32_s8): Likewise.
3251 (vreinterpretq_u32_s16): Likewise.
3252 (vreinterpretq_u32_s32): Likewise.
3253 (vreinterpretq_u32_s64): Likewise.
3254 (vreinterpretq_u32_f32): Likewise.
3255 (vreinterpretq_u32_u8): Likewise.
3256 (vreinterpretq_u32_u16): Likewise.
3257 (vreinterpretq_u32_u64): Likewise.
3258 (vreinterpretq_u32_p8): Likewise.
3259 (vreinterpretq_u32_p16): Likewise.
3261 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3263 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
3265 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
3267 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
3268 (vqnegd_s64): Likewise.
3269 (vqabs_s64): Likewise.
3270 (vqabsd_s64): Likewise.
3272 2014-04-22 Richard Henderson <rth@redhat.com>
3274 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
3275 computation to the top of the loop.
3277 2014-04-22 Renlin <renlin.li@arm.com>
3278 Jiong Wang <jiong.wang@arm.com>
3280 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
3281 * config/aarch64/aarch64.c (aarch64_layout_frame)
3282 (aarch64_initial_elimination_offset): Likewise.
3284 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
3286 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
3289 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
3291 * machmode.h (bitwise_mode_for_mode): Declare.
3292 * stor-layout.h (bitwise_type_for_mode): Likewise.
3293 * stor-layout.c (bitwise_mode_for_mode): New function.
3294 (bitwise_type_for_mode): Likewise.
3295 * builtins.c (fold_builtin_memory_op): Use it instead of
3296 int_mode_for_mode and build_nonstandard_integer_type.
3298 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3300 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
3301 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
3302 (*-*-solaris2*): Simplify.
3303 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
3304 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
3305 *-*-solaris2.9* handling.
3307 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
3309 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
3310 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
3312 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
3313 * configure: Regenerate.
3315 * config/i386/sol2-9.h: Remove.
3317 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
3318 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
3319 Remove Solaris 9 references.
3321 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
3323 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
3324 (floatuns<GPI:mode><GPF:mode>2): Remove.
3325 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
3326 and floatuns conversions.
3327 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
3328 and floatuns conversions.
3329 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
3330 (w1,w2): New mode attributes for inequal width conversions.
3332 2014-04-22 Renlin Li <Renlin.Li@arm.com>
3334 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
3335 the output asm format.
3337 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
3339 * config/aarch64/aarch64-simd.md
3340 (aarch64_cm<optab>di): Always split.
3341 (*aarch64_cm<optab>di): New.
3342 (aarch64_cmtstdi): Always split.
3343 (*aarch64_cmtstdi): New.
3345 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3347 PR tree-optimization/60823
3348 * omp-low.c (ipa_simd_modify_function_body): Go through
3349 all SSA_NAMEs and for those refering to vector arguments
3350 which are going to be replaced adjust SSA_NAME_VAR and,
3351 if it is a default definition, change it into a non-default
3352 definition assigned at the beginning of function from new_decl.
3353 (ipa_simd_modify_stmt_ops): Rewritten.
3354 * tree-dfa.c (set_ssa_default_def): When removing default def,
3355 check for NULL loc instead of NULL *loc.
3357 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3359 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
3360 restrictions on core registers for DImode values in Thumb2.
3362 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3364 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
3365 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
3367 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3369 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
3370 (*iordi_notzesidi_di): Likewise.
3371 (*iordi_notsesidi_di): Likewise.
3373 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3375 * config/arm/arm-protos.h (tune_params): New struct members.
3376 * config/arm/arm.c: Initialise tune_params per processor.
3377 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
3378 for speed, based on new tune_params.
3380 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3382 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
3383 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
3384 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
3385 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
3386 * config/aarch64/arm_neon.h (vrnd_f64): Added.
3387 (vrnda_f64): Likewise.
3388 (vrndi_f64): Likewise.
3389 (vrndm_f64): Likewise.
3390 (vrndn_f64): Likewise.
3391 (vrndp_f64): Likewise.
3392 (vrndx_f64): Likewise.
3394 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3396 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
3397 GET_MODE_SIZE argument is enum machine_mode.
3399 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3402 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
3403 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
3405 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
3408 * asan.c (asan_emit_stack_protection): Force the base to align to
3409 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
3410 appropriate bits if STRICT_ALIGNMENT.
3411 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
3413 (expand_used_vars): Leave a space in the stack frame for alignment
3414 if STRICT_ALIGNMENT.
3416 2014-04-21 David Malcolm <dmalcolm@redhat.com>
3418 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
3420 (gimple_store_p): Likewise.
3421 (gimple_assign_load_p): Likewise.
3422 (gimple_assign_cast_p): Likewise.
3423 (gimple_clobber_p): Likewise.
3425 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
3426 rather than a gimple.
3427 (gimple_assign_cast_p): Likewise.
3429 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
3432 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
3433 If mode is DDmode and TARGET_E500_DOUBLE allow move.
3435 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
3436 more debug information for E500 if -mdebug=reg.
3438 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
3441 * config/i386/i386.c (ix86_expand_builtin)
3442 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
3443 register for target RTX.
3444 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
3446 2014-04-18 Cong Hou <congh@google.com>
3448 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
3449 the widen-mult pattern by handling two operands with different sizes,
3450 and operands whose size is smaller than half of the result type.
3452 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3454 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
3455 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
3456 (do_estimate_edge_time): Compute it.
3457 * ipa-inline.c (want_inline_small_function_p): Bypass
3458 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
3460 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3462 * ipa-inline.c (spec_rem): New static variable.
3463 (dump_overall_stats): New function.
3464 (dump_inline_stats): New function.
3466 2014-04-18 Richard Henderson <rth@redhat.com>
3468 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
3469 to GET_MODE_SIZE, not a reg_class_t.
3471 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3473 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
3474 (vsx_xxmrglw_<mode>): Likewise.
3476 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3479 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
3480 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
3481 (rs6000_init_hard_regno_mode_ok): Likewise.
3483 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
3485 * ipa-inline.c (inline_small_functions): Account only non-cold
3487 * doc/invoke.texi (inline-unit-growth): Update documentation.
3489 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
3491 * config/rs6000/rs6000.md (addti3, subti3): New.
3493 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
3496 * config/i386/i386.c (ix86_expand_clear): Remove outdated
3497 comment. Check optimize_insn_for_size_p instead of
3498 optimize_insn_for_speed_p.
3500 2014-04-17 Martin Jambor <mjambor@suse.cz>
3502 * gimple-iterator.c (gsi_start_edge): New function.
3503 * gimple-iterator.h (gsi_start_edge): Declare.
3504 * tree-sra.c (single_non_eh_succ): New function.
3505 (disqualify_ops_if_throwing_stmt): Renamed to
3506 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
3507 having one non-EH successor BB.
3508 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
3509 generate loads into replacements.
3510 (sra_modify_assign): Likewise and and also use the simple path for
3512 (sra_modify_function_body): Commit statements on edges.
3514 2014-04-17 Richard Biener <rguenther@suse.de>
3517 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
3518 comparison results and add clarifying comment.
3520 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3522 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
3523 (blank_mode): Initialize it.
3524 (emit_mode_size_inline, emit_mode_nunits_inline,
3525 emit_mode_inner_inline): New functions.
3526 (emit_insn_modes_h): Call them and surround their output with
3527 #if GCC_VERSION >= 4001 ... #endif.
3528 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
3529 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
3530 mode_* arrays if the argument is __builtin_constant_p.
3531 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
3532 is enum machine_mode.
3534 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3536 * passes.c (opt_pass::execute): Adjust.
3537 (pass_manager::execute_pass_mode_switching): Likewise.
3538 (early_local_passes::execute): Likewise.
3539 (execute_one_pass): Pass cfun to the pass's execute method.
3540 * tree-pass.h (opt_pass::execute): Add function * argument.
3541 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3542 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3543 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3544 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3545 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3546 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
3547 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
3548 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3549 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3550 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3551 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3552 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3553 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3554 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3555 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3556 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3557 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3558 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3559 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3560 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3561 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3562 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3563 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3564 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3565 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3566 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3567 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3568 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3569 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3572 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3574 * passes.c (opt_pass::gate): Take function * argument.
3575 (gate_all_early_local_passes): Merge into
3576 (early_local_passes::gate): this.
3577 (gate_all_early_optimizations): Merge into
3578 (all_early_optimizations::gate): this.
3579 (gate_all_optimizations): Mege into
3580 (all_optimizations::gate): this.
3581 (gate_all_optimizations_g): Merge into
3582 (all_optimizations_g::gate): this.
3583 (gate_rest_of_compilation): Mege into
3584 (rest_of_compilation::gate): this.
3585 (gate_postreload): Merge into
3586 (postreload::gate): this.
3587 (dump_one_pass): Pass cfun to the pass's gate method.
3588 (execute_ipa_summary_passes): Likewise.
3589 (execute_one_pass): Likewise.
3590 (ipa_write_summaries_2): Likewise.
3591 (ipa_write_optimization_summaries_1): Likewise.
3592 (ipa_read_summaries_1): Likewise.
3593 (ipa_read_optimization_summaries_1): Likewise.
3594 (execute_ipa_stmt_fixups): Likewise.
3595 * tree-pass.h (opt_pass::gate): Add function * argument.
3596 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
3597 combine-stack-adj.c, combine.c, compare-elim.c,
3598 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3599 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
3600 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
3601 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
3602 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3603 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3604 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3605 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
3606 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
3607 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
3608 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3609 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3610 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
3611 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3612 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3613 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3614 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3615 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3616 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3617 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3618 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3619 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3620 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3621 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
3622 var-tracking.c, vtable-verify.c, web.c: Adjust.
3624 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3626 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
3627 * configure: Regenerate.
3629 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3631 * passes.c (dump_one_pass): don't check pass->has_gate.
3632 (execute_ipa_summary_passes): Likewise.
3633 (execute_one_pass): Likewise.
3634 (ipa_write_summaries_2): Likewise.
3635 (ipa_write_optimization_summaries_1): Likewise.
3636 (ipa_read_optimization_summaries_1): Likewise.
3637 (execute_ipa_stmt_fixups): Likewise.
3638 * tree-pass.h (pass_data::has_gate): Remove.
3639 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3640 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3641 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3642 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3643 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3644 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
3645 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
3646 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
3647 gimple-low.c, gimple-ssa-isolate-paths.c,
3648 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
3649 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3650 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3651 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
3652 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
3653 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
3654 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
3655 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3656 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
3657 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3658 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3659 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3660 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3661 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3662 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3663 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3664 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3665 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3666 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3667 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3668 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3669 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3672 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3674 * pass_manager.h (pass_manager::register_dump_files_1): Remove
3676 * passes.c (pass_manager::register_dump_files_1): Merge into
3677 (pass_manager::register_dump_files): this, and remove its handling of
3678 properties since the pass always has the properties anyway.
3679 (pass_manager::pass_manager): Adjust.
3681 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3683 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
3684 * passes.c (pass_manager::register_dump_files_1): Remove dead code
3685 dealing with properties.
3686 (pass_manager::register_dump_files): Adjust.
3688 2014-03-20 Mark Wielaard <mjw@redhat.com>
3690 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
3691 then represent the bound as normal constant value.
3693 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3696 Forward port from 4.8 branch
3697 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
3699 * config/i386/bmiintrin.h (_blsi_u32): New.
3703 (_blsmsk_u32): Ditto.
3704 (_blsmsk_u64): Ditto.
3705 (_tzcnt_u32): Ditto.
3706 (_tzcnt_u64): Ditto.
3708 2014-04-17 Kito Cheng <kito@0xlab.org>
3710 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
3712 2014-04-17 Richard Biener <rguenther@suse.de>
3715 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
3716 boolean results for comparisons.
3718 2014-04-17 Richard Biener <rguenther@suse.de>
3720 PR tree-optimization/60836
3721 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
3722 initial PHI args to be gimple values.
3724 2014-04-17 Richard Biener <rguenther@suse.de>
3726 PR tree-optimization/60841
3727 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
3728 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
3729 of stmts to SLP build.
3730 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
3731 (vect_analyze_slp): Likewise.
3732 (vect_analyze_slp_instance): Likewise.
3733 (vect_build_slp_tree): Limit overall SLP tree growth.
3734 * tree-vectorizer.h (vect_analyze_data_refs,
3735 vect_analyze_slp): Adjust prototypes.
3737 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3739 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
3742 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3744 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
3745 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
3746 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
3747 for TARGET_SLOW_PSHUFB
3749 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3751 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
3752 * config/i386/i386.c (intel_cost): Ditto.
3754 2014-04-17 Joey Ye <joey.ye@arm.com>
3756 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
3758 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3760 * opts.c (common_handle_option): Disable -fipa-reference coorectly
3763 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3765 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
3766 (type_all_derivations_known_p): New predicate.
3767 (type_all_ctors_visible_p): New predicate.
3768 (type_possibly_instantiated_p): New predicate.
3769 (get_odr_type): Compute all_derivations_known.
3770 (dump_odr_type): Dump the flag.
3771 (maybe_record_type): Cleanup.
3772 (record_target_from_binfo): Add bases_to_consider array;
3773 record bases for types w/o instances and skip CXX destructor.
3774 (possible_polymorphic_call_targets_1): Add bases_to_consider
3775 and consider_construction parameters; check if type may have instance.
3776 (get_polymorphic_call_info): Set maybe_in_construction to true
3777 when we know nothing.
3778 (record_targets_from_bases): Skip CXX destructors; they are
3779 never called for types in construction.
3780 (possible_polymorphic_call_targets): Do not record target when
3781 type may not have instance.
3783 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3786 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
3787 external aliases alive, too.
3789 2014-04-16 Andrew Pinski <apinski@cavium.com>
3791 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
3794 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3796 * final.c (compute_alignments): Do not apply loop alignment to a block
3797 falling through to the exit.
3799 2014-04-16 Catherine Moore <clm@codesourcery.com>
3801 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
3802 Adjust constraints for microMIPS store patterns.
3804 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3806 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
3808 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3810 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
3811 (append_use): Run at -O0.
3812 (append_vdef): Likewise.
3813 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
3814 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
3816 2014-04-16 Jakub Jelinek <jakub@redhat.com>
3818 PR tree-optimization/60844
3819 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
3820 (propagate_op_to_single_use, remove_visited_stmt_chain,
3821 linearize_expr, repropagate_negates, reassociate_bb): Use it
3822 instead of gsi_remove.
3824 2014-04-16 Martin Jambor <mjambor@suse.cz>
3826 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
3827 ipa_transforms_to_apply.
3828 (cgraph_function_versioning): Assert that old_node has empty
3829 ipa_transforms_to_apply.
3830 * trans-mem.c (ipa_tm_create_version): Likewise.
3831 * tree-inline.c (tree_function_versioning): Do not duplicate
3832 ipa_transforms_to_apply.
3834 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3837 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
3839 Pass necessary as flags on 64-bit Solaris/x86.
3840 Use lowercase relocs for x86_64-*-*.
3841 * configure: Regenerate.
3843 2014-04-15 Jan Hubicka <jh@suse.cz>
3845 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
3846 (maybe_record_node, likely_target_p): Use it.
3848 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3851 Revert following patch
3853 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3856 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3857 software floating point or no floating point registers, do not
3858 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3859 in GPRs that occurs after we tested for GPRs that would never be
3862 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3863 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3864 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3865 specifically allow DDmode, since that does not use the SPE SIMD
3868 2014-03-21 Mark Wielaard <mjw@redhat.com>
3870 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
3871 as unsigned or int depending on type and value used.
3873 2014-04-15 Richard Biener <rguenther@suse.de>
3875 PR rtl-optimization/56965
3876 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
3877 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
3879 * alias.c (true_dependence_1): Do not call
3880 nonoverlapping_component_refs_p.
3881 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
3882 nonoverlapping_component_refs_p.
3883 (indirect_refs_may_alias_p): Likewise.
3885 2014-04-15 Teresa Johnson <tejohnson@google.com>
3887 * cfg.c (dump_bb_info): Fix flags check.
3888 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
3890 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3892 PR rtl-optimization/60663
3893 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
3896 2014-04-15 Richard Biener <rguenther@suse.de>
3898 * lto-streamer.h (LTO_major_version): Bump to 4.
3900 2014-04-15 Richard Biener <rguenther@suse.de>
3902 * common.opt (lto_partition_model): New enum.
3903 (flto-partition=): Merge separate options with a single with argument,
3904 add -flto-partition=one support.
3905 * flag-types.h (enum lto_partition_model): Declare.
3906 * opts.c (finish_options): Remove duplicate -flto-partition=
3908 * lto-wrapper.c (run_gcc): Adjust.
3910 2014-04-15 Richard Biener <rguenther@suse.de>
3912 * alias.c (ncr_compar): New function.
3913 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
3915 2014-04-15 Richard Biener <rguenther@suse.de>
3917 * alias.c (record_component_aliases): Do not walk BINFOs.
3919 2014-04-15 Richard Biener <rguenther@suse.de>
3921 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
3922 Add struct function argument and adjust.
3923 (find_func_aliases_for_call): Likewise.
3924 (find_func_aliases): Likewise.
3925 (find_func_clobbers): Likewise.
3926 (intra_create_variable_infos): Likewise.
3927 (compute_points_to_sets): Likewise.
3928 (ipa_pta_execute): Adjust. Do not push/pop cfun.
3930 2014-04-15 Richard Biener <rguenther@suse.de>
3932 * tree.c (iterative_hash_expr): Use enum tree_code_class
3933 to store TREE_CODE_CLASS.
3934 (tree_block): Likewise.
3935 (tree_set_block): Likewise.
3936 * tree.h (fold_build_pointer_plus_loc): Use
3937 convert_to_ptrofftype_loc.
3939 2014-04-15 Jakub Jelinek <jakub@redhat.com>
3942 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
3945 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
3947 * cfgloop.h (struct loop): Move force_vectorize down.
3948 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
3949 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
3950 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
3951 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
3952 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
3953 * tree-core.h (enum annot_expr_kind): Add new kind values.
3954 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
3955 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
3957 * tree.def (ANNOTATE_EXPR): Tweak comment.
3959 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3961 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
3964 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
3966 * tree.h (TYPE_IDENTIFIER): Declare.
3967 * tree.c (subrange_type_for_debug_p): Use it.
3968 * godump.c (go_format_type): Likewise.
3969 * dwarf2out.c (is_cxx_auto, modified_type_die,
3970 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
3971 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
3973 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3976 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
3978 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
3980 * config/i386/i386.c (examine_argument): Return bool. Return true if
3981 parameter should be passed in memory.
3982 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
3983 (construct_container): Update calls to examine_argument.
3984 (function_arg_advance_64): Ditto.
3985 (return_in_memory_32): Merge with ix86_return_in_memory.
3986 (return_in_memory_64): Ditto.
3987 (return_in_memory_ms_64): Ditto.
3989 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3991 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
3992 * coverage.c (coverage_compute_profile_id): Handle externally visible
3995 2014-04-14 Martin Jambor <mjambor@suse.cz>
3997 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
3998 DECL_DISREGARD_INLINE_LIMITS functions.
4000 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
4003 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
4005 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
4008 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
4009 optimize_insn_for_speed_p instead of
4010 optimize_function_for_speed_p.
4012 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
4014 * doc/invoke.texi (free): Document AArch64.
4016 2014-04-14 Richard Biener <rguenther@suse.de>
4018 PR tree-optimization/60042
4019 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
4020 (insert_into_preds_of_block): Do not prevent PHI insertion
4021 for REFERENCE exprs here ...
4022 (eliminate_dom_walker::before_dom_children): ... but prevent
4023 their use here under similar conditions when applied to the
4024 IL after PRE optimizations.
4026 2014-04-14 Richard Biener <rguenther@suse.de>
4028 * passes.def: Move early points-to after early SRA.
4030 2014-04-14 Richard Biener <rguenther@suse.de>
4032 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
4033 check for which sign-changes we allow when forwarding
4034 a converted value into a switch.
4036 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
4038 * stor-layout.c (place_field): Finalize non-constant offset for the
4041 2014-04-14 Richard Biener <rguenther@suse.de>
4043 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
4045 (expand_switch_using_bit_tests_p): Likewise.
4046 (process_switch): Compute and pass on speed_p based on the
4048 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
4049 optimize_bb_for_speed_p.
4051 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
4053 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
4054 * function.h (struct function): Rename has_force_vect_loops into
4055 has_force_vectorize_loops.
4056 * lto-streamer-in.c (input_cfg): Adjust for renaming.
4057 (input_struct_function_base): Likewise.
4058 * lto-streamer-out.c (output_cfg): Likewise.
4059 (output_struct_function_base): Likewise.
4060 * omp-low.c (expand_omp_simd): Likewise.
4061 * tree-cfg.c (move_sese_region_to_fn): Likewise.
4062 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
4063 (version_loop_for_if_conversion): Likewise.
4064 (tree_if_conversion): Likewise.
4065 (main_tree_if_conversion): Likewise.
4066 (gate_tree_if_conversion): Likewise.
4067 * tree-inline.c (copy_loops): Likewise.
4068 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
4069 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
4070 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
4071 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
4072 * tree-vectorizer.c (vectorize_loops): Likewise.
4073 * tree-vectorizer.h (unlimited_cost_model): Likewise.
4075 2014-04-14 Richard Biener <rguenther@suse.de>
4078 * lto-streamer-out.c (wrap_refs): New function.
4079 (lto_output): Wrap symbol references in global initializes in
4080 type-preserving MEM_REFs.
4082 2014-04-14 Christian Bruel <christian.bruel@st.com>
4084 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
4086 2014-04-14 Christian Bruel <christian.bruel@st.com>
4088 * config/sh/sh.md (setmemqi): New expand pattern.
4089 * config/sh/sh.h (CLEAR_RATIO): Define.
4090 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
4091 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
4093 2014-04-14 Richard Biener <rguenther@suse.de>
4096 * fold-const.c (negate_expr_p): Don't negate directional rounding
4098 (fold_negate_expr): Likewise.
4100 2014-04-14 Richard Biener <rguenther@suse.de>
4102 PR tree-optimization/59817
4103 PR tree-optimization/60453
4104 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
4105 recursion to catch all CHRECs in the scalar evolution and restrict
4106 the predicate for the remains appropriately.
4108 2014-04-12 Catherine Moore <clm@codesourcery.com>
4110 * config/mips/constraints.md: Add new register constraint "kb".
4111 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
4112 (*movhi_internal): Likewise.
4113 (*movqi_internal): Likewise.
4114 * config/mips/mips.h (M16_STORE_REGS): New register class.
4115 (REG_CLASS_NAMES): Add M16_STORE_REGS.
4116 (REG_CLASS_CONTENTS): Likewise.
4117 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
4119 2014-04-11 Tobias Burnus <burnus@net-b.de>
4122 * doc/invoke.texi (-Wformat-signedness): Document it.
4123 (Wformat=2): Mention that this enables -Wformat-signedness.
4125 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4127 * common/config/epiphany/epiphany-common.c
4128 (epiphany_option_optimization_table): Enable section anchors by
4129 default at -O1 or higher.
4130 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
4131 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
4132 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
4133 carries no extra cost.
4134 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
4135 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
4136 * config/epiphany/predicates.md (memclob_operand): New predicate.
4137 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
4138 Use memclob_operand predicate and X constraint for operand 3.
4140 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4142 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
4143 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
4146 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4148 PR rtl-optimization/60651
4149 * mode-switching.c (optimize_mode_switching): Make sure to emit
4150 sets of a lower numbered entity before sets of a higher numbered
4151 entity to a mode of the same or lower priority.
4152 When creating a seginfo for a basic block that starts with a code
4153 label, move the insertion point past the code label.
4154 (new_seginfo): Document and enforce requirement that
4155 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
4156 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
4157 * doc/tm.texi: Regenerate.
4159 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
4162 * config/arc/arc.c (arc_save_restore): Fix assert typo.
4164 2013-04-11 Jakub Jelinek <jakub@redhat.com>
4166 * BASE-VER: Set to 4.10.0.
4168 2014-04-11 Tobias Burnus <burnus@net-b.de>
4171 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
4172 * doc/gcc.texi (Service): Update description in the @menu
4173 * doc/invoke.texi (Option Summary): Remove misplaced and
4176 2014-04-11 Steve Ellcey <sellcey@mips.com>
4177 Jakub Jelinek <jakub@redhat.com>
4180 * expr.c (convert_move): Use emit_store_flag_force instead of
4181 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
4184 2014-04-11 Richard Biener <rguenther@suse.de>
4187 * varasm.c (assemble_alias): Avoid endless error reporting
4188 recursion by setting TREE_ASM_WRITTEN.
4190 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4192 * config/s390/s390.md: Add a splitter for NOT rtx.
4194 2014-04-11 Jakub Jelinek <jakub@redhat.com>
4196 PR rtl-optimization/60663
4197 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
4199 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
4200 Jakub Jelinek <jakub@redhat.com>
4203 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
4204 flag from decl_node to node.
4206 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4209 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
4210 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
4211 ameliorating the cases where it can be.
4213 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
4216 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4218 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4219 (loadsync_<mode>): Change mode.
4220 (load_quadpti, store_quadpti): New.
4221 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4222 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4223 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
4225 2014-04-09 Cong Hou <congh@google.com>
4228 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
4231 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4233 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
4234 instead of vnor to exploit possible fusion opportunity in the
4236 (altivec_expand_vec_perm_const_le): Likewise.
4238 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4240 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4241 (loadsync_<mode>): Change mode.
4242 (load_quadpti, store_quadpti): New.
4243 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4244 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4246 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
4249 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
4250 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
4251 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
4253 2014-04-08 Richard Biener <rguenther@suse.de>
4256 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
4257 a 64bit widest int print double-int similar to on HWI64 hosts.
4259 2014-04-08 Richard Biener <rguenther@suse.de>
4261 PR tree-optimization/60785
4262 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
4263 default defs properly.
4265 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
4267 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
4268 (Weffc++): Likewise.
4270 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
4272 * ipa-devirt.c (maybe_record_node): When node is not recorded,
4273 set completep to false rather than true.
4275 2014-04-07 Douglas B Rupp <rupp@adacore.com>
4278 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
4279 ARM_TARGET2_DWARF_FORMAT.
4281 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
4284 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
4285 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
4288 2014-04-07 Richard Biener <rguenther@suse.de>
4290 PR tree-optimization/60766
4291 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
4292 (may_eliminate_iv): Convert cand_value_at result to desired type.
4294 2014-04-07 Jason Merrill <jason@redhat.com>
4297 * common.opt (-fno-gnu-unique): Add.
4298 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
4300 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4302 * haifa-sched.c: Fix outdated function reference and minor
4303 grammar errors in introductory comment.
4305 2014-04-07 Richard Biener <rguenther@suse.de>
4308 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
4310 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
4312 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
4315 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
4316 size accounting for thunks.
4317 (pa_asm_output_mi_thunk): Use final_start_function() and
4318 final_end_function() to output function start and end directives.
4320 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4322 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
4323 device specific ISA/ feature information. Remove short_sp and
4324 errata_skip ds. Add avr_device_specific_features enum to have device
4326 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
4327 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
4328 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
4329 updated device specific info.
4330 * config/avr/avr-mcus.def: Merge device specific details to
4331 dev_attribute field.
4332 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
4334 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
4335 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
4336 assembler if RMW isa supported by current device.
4337 * config/avr/genmultilib.awk: Update as device info structure changed.
4338 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
4340 2014-04-04 Cong Hou <congh@google.com>
4342 PR tree-optimization/60656
4343 * tree-vect-stmts.c (supportable_widening_operation):
4344 Fix a bug that elements in a vector with vect_used_by_reduction
4345 property are incorrectly reordered when the operation on it is not
4346 consistant with the one in reduction operation.
4348 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
4350 PR rtl-optimization/60155
4351 * gcse.c (record_set_data): New function.
4352 (single_set_gcse): New function.
4353 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
4354 (hoist_code): Likewise.
4355 (get_pressure_class_and_nregs): Likewise.
4357 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
4359 * explow.c (probe_stack_range): Emit a final optimization blockage.
4361 2014-04-04 Anthony Green <green@moxielogic.com>
4363 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
4366 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
4369 * lto-cgraph.c (input_overwrite_node): Check that partitioning
4370 flags are set only during streaming.
4371 * ipa.c (process_references, walk_polymorphic_call_targets,
4372 symtab_remove_unreachable_nodes): Drop bodies of always inline
4373 after early inlining.
4374 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
4376 2014-04-04 Jakub Jelinek <jakub@redhat.com>
4377 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4380 * dwarf2out.c (const_ok_for_output_1): Reject expressions
4383 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4386 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
4388 (cortex_a53_fdivd): Likewise.
4390 2014-04-04 Martin Jambor <mjambor@suse.cz>
4393 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
4395 * cgraph.c (clone_of_p): Also return true if thunks match.
4396 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
4397 cgraph_function_or_thunk_node and an obsolete comment.
4398 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
4400 (build_function_decl_skip_args): Likewise.
4401 (set_new_clone_decl_and_node_flags): New function.
4402 (duplicate_thunk_for_node): Likewise.
4403 (redirect_edge_duplicating_thunks): Likewise.
4404 (cgraph_clone_node): New parameter args_to_skip, pass it to
4405 redirect_edge_duplicating_thunks which is called instead of
4406 cgraph_redirect_edge_callee.
4407 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
4408 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
4410 2014-04-04 Jeff Law <law@redhat.com>
4413 * config/arm/predicates.md (const_int_I_operand): New predicate.
4414 (const_int_M_operand): Similarly.
4415 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
4417 (insv_t2, extv_reg, extzv_t2): Likewise.
4418 (load_multiple_with_writeback): Similarly for const_int_I_operand.
4419 (pop_multiple_with_writeback_and_return): Likewise.
4420 (vfp_pop_multiple_with_writeback): Likewise
4422 2014-04-04 Richard Biener <rguenther@suse.de>
4425 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
4426 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
4428 * gimplify.h (gimple_add_tmp_var_fn): Declare.
4429 * gimplify.c (gimple_add_tmp_var_fn): New function.
4430 * gimple-expr.h (create_tmp_reg_fn): Declare.
4431 * gimple-expr.c (create_tmp_reg_fn): New function.
4432 * gimple-low.c (record_vars_into): Don't change cfun.
4433 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
4434 code generation without cfun.
4436 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
4439 * Makefile.in (install-driver): Fix shell scripting.
4441 2014-04-03 Cong Hou <congh@google.com>
4443 PR tree-optimization/60505
4444 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
4445 threshold of number of iterations below which no vectorization
4447 * tree-vect-loop.c (new_loop_vec_info):
4448 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
4449 * tree-vect-loop.c (vect_analyze_loop_operations):
4450 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
4451 * tree-vect-loop.c (vect_transform_loop):
4452 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
4453 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
4454 of iterations of the loop and see if we should build the epilogue.
4456 2014-04-03 Richard Biener <rguenther@suse.de>
4458 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
4459 (streamer_tree_cache_create): Adjust.
4460 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
4461 to allow optional nodes array.
4462 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
4463 (streamer_tree_cache_append): Likewise.
4464 (streamer_tree_cache_create): Create nodes array optionally
4465 as specified by parameter.
4466 * lto-streamer-out.c (create_output_block): Avoid maintaining
4467 the node array in the writer cache.
4468 (DFS_write_tree): Remove assertion.
4469 (produce_asm_for_decls): Free the out decl state hash table early.
4470 * lto-streamer-in.c (lto_data_in_create): Adjust for
4471 streamer_tree_cache_create prototype change.
4473 2014-04-03 Richard Biener <rguenther@suse.de>
4475 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
4476 set TREE_CHAIN to NULL_TREE.
4478 2014-04-03 Richard Biener <rguenther@suse.de>
4480 PR tree-optimization/60740
4481 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
4482 over all GIMPLE_COND operands.
4484 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
4486 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
4487 (Weffc++): Remove Scott's numbering, merge lists and reference
4490 2014-04-03 Nick Clifton <nickc@redhat.com>
4492 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
4495 2014-04-03 Martin Jambor <mjambor@suse.cz>
4497 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
4498 mention gcc_unreachable before failing.
4499 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
4502 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
4505 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
4506 inconsistent code and instead mark the context inconsistent.
4507 (possible_polymorphic_call_targets): For inconsistent contexts
4508 return empty complete list.
4510 2014-04-02 Anthony Green <green@moxielogic.com>
4512 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
4513 (extendqisi2, extendhisi2): Define.
4514 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
4515 (WCHAR_TYPE): Change to unsigned int.
4517 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4519 PR tree-optimization/60733
4520 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
4521 insertion point for PHI candidates to be the end of the feeding
4522 block for the PHI argument.
4524 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
4526 PR rtl-optimization/60650
4527 * lra-constraints.c (process_alt_operands): Decrease reject for
4528 earlyclobber matching.
4530 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4532 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
4534 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4536 * config/spu/spu.c (pad_bb): Do not crash when the last
4537 insn is CODE_FOR_blockage.
4539 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4541 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
4542 lies outside the target mode.
4544 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4547 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4548 software floating point or no floating point registers, do not
4549 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4550 in GPRs that occurs after we tested for GPRs that would never be
4553 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4554 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4555 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4556 specifically allow DDmode, since that does not use the SPE SIMD
4559 2014-04-02 Richard Biener <rguenther@suse.de>
4562 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4563 MODE_INTs. Properly use negv_optab.
4564 (expand_abs): Likewise.
4566 2014-04-02 Richard Biener <rguenther@suse.de>
4569 * Makefile.in (install-driver): Guard extra installs with special
4572 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4574 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4577 2014-04-01 Richard Henderson <rth@redhat.com>
4580 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
4581 alternative enabled before register allocation.
4583 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
4585 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
4586 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
4588 (nios2_large_got_address): Remove unneeded 'sym' parameter.
4589 (nios2_got_address): Update nios2_large_got_address call site.
4590 (nios2_delegitimize_address): New function.
4591 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
4592 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
4593 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
4595 2014-04-01 Martin Husemann <martin@duskware.de>
4597 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
4600 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
4602 PR rtl-optimization/60604
4603 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
4604 check from register_operand.
4605 (register_operand): Redefine in terms of general_operand.
4606 (nonmemory_operand): Use register_operand for the non-constant cases.
4608 2014-04-01 Richard Biener <rguenther@suse.de>
4610 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
4612 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
4614 * doc/invoke.texi (mapp-regs): Clarify.
4616 2014-03-31 Ulrich Drepper <drepper@gmail.com>
4618 * config/i386/avx512fintrin.h (__v32hi): Define type.
4619 (__v64qi): Likewise.
4620 (_mm512_set1_epi8): Define.
4621 (_mm512_set1_epi16): Define.
4622 (_mm512_set4_epi32): Define.
4623 (_mm512_set4_epi64): Define.
4624 (_mm512_set4_pd): Define.
4625 (_mm512_set4_ps): Define.
4626 (_mm512_setr4_epi64): Define.
4627 (_mm512_setr4_epi32): Define.
4628 (_mm512_setr4_pd): Define.
4629 (_mm512_setr4_ps): Define.
4630 (_mm512_setzero_epi32): Define.
4632 2014-03-31 Martin Jambor <mjambor@suse.cz>
4635 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
4636 callsite_arguments_match_p. Updated all callers. Also check types of
4637 corresponding formal parameters and actual arguments.
4638 (not_all_callers_have_enough_arguments_p) Renamed to
4639 some_callers_have_mismatched_arguments_p.
4641 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
4643 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
4645 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
4648 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
4651 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
4653 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
4654 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
4656 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
4657 Use FMAMODE_NOVF512 mode iterator.
4658 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
4659 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
4660 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
4661 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
4663 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
4664 Use VF_128_256 mode iterator.
4665 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
4668 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4670 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
4671 static chain if needed.
4673 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4676 * lra-constraints.c (index_part_to_reg): New.
4677 (process_address): Use it.
4679 2014-03-27 Jeff Law <law@redhat.com>
4680 Jakub Jelinek <jakub@redhat.com>
4683 * expr.c (do_tablejump): Use simplify_gen_binary rather than
4684 gen_rtx_{PLUS,MULT} to build up the address expression.
4686 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
4687 creating non-canonical RTL.
4689 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4692 * ipa-inline.c (want_inline_small_function_p): Short circuit large
4693 functions; reorganize to make cheap checks first.
4694 (inline_small_functions): Do not estimate growth when dumping;
4696 * ipa-inline.h (inline_summary): Add min_size.
4697 (growth_likely_positive): New function.
4698 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
4699 (set_cond_stmt_execution_predicate): Cleanup.
4700 (estimate_edge_size_and_time): Compute min_size.
4701 (estimate_calls_size_and_time): Likewise.
4702 (estimate_node_size_and_time): Likewise.
4703 (inline_update_overall_summary): Update min_size.
4704 (do_estimate_edge_time): Likewise.
4705 (do_estimate_edge_size): Update.
4706 (do_estimate_edge_hints): Update.
4707 (growth_likely_positive): New function.
4709 2014-03-28 Jakub Jelinek <jakub@redhat.com>
4712 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
4713 also if addr has VOIDmode.
4715 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4717 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
4718 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
4720 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
4721 instructions as well as AdvancedSIMD loads.
4723 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4725 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
4726 Use crypto_aese type.
4727 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
4728 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
4729 crypto_aese, crypto_aesmc. Move to types.md.
4730 * config/arm/types.md (crypto_aes): Split into crypto_aese,
4732 * config/arm/iterators.md (crypto_type): Likewise.
4734 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4736 * cgraph.c: Include expr.h and tree-dfa.h.
4737 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
4740 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4743 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
4744 regs from checking multi-reg pseudos.
4746 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4748 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
4750 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4752 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
4753 if it would clobber the stack pointer, even temporarily.
4755 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
4757 * mode-switching.c: Make small adjustments to the top comment.
4759 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4761 * config/rs6000/constraints.md (wD constraint): New constraint to
4762 match the constant integer to get the top DImode/DFmode out of a
4763 vector in a VSX register.
4765 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
4766 match the constant integer to get the top DImode/DFmode out of a
4767 vector in a VSX register.
4769 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
4772 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
4775 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
4776 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
4778 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
4779 Optimize vec_extract of 64-bit values, where the value being
4780 extracted is in the top word, where we can use scalar
4781 instructions. Add direct move and store support. Combine the big
4782 endian/little endian vector select load support into a single insn.
4783 (vsx_extract_<mode>_internal1): Likewise.
4784 (vsx_extract_<mode>_internal2): Likewise.
4785 (vsx_extract_<mode>_load): Likewise.
4786 (vsx_extract_<mode>_store): Likewise.
4787 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
4788 combined into vsx_extract_<mode>_load.
4789 (vsx_extract_<mode>_one_le): Likewise.
4791 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
4792 define the top 64-bit vector element.
4794 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
4797 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4798 Document vec_vbpermq builtin.
4801 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
4802 enable use of xxsldwi and xxpermdi builtin functions.
4803 (vec_xxpermdi): Likewise.
4805 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4806 Document use of vec_xxsldwi and vec_xxpermdi builtins.
4808 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
4810 PR rtl-optimization/60650
4811 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
4813 (find_spills_for): New.
4814 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
4815 Spill all pseudos on the second iteration.
4817 2014-03-27 Marek Polacek <polacek@redhat.com>
4820 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
4823 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4825 * config/s390/s390.c (s390_can_use_return_insn): Check for
4826 call-saved FPRs on 31 bit.
4828 2014-03-27 Jakub Jelinek <jakub@redhat.com>
4831 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
4832 if they need regimplification, just drop them instead of
4833 calling gimple_regimplify_operands on them.
4835 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
4838 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
4839 (aarch64_frame_pointer_required): Adjust logic.
4840 (aarch64_can_eliminate): Adjust logic.
4841 (aarch64_override_options_after_change): Adjust logic.
4843 2014-03-27 Dehao Chen <dehao@google.com>
4845 * ipa-inline.c (early_inliner): Update node's inline info.
4847 2014-03-26 Dehao Chen <dehao@google.com>
4849 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
4850 compiler inserted conditional jumps for NAN float check.
4852 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4854 * ubsan.h (ubsan_create_data): Change second argument's type
4855 to const location_t *.
4856 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
4858 (ubsan_create_data): Change second argument to const location_t *PLOC.
4859 Create Loc field whenever PLOC is non-NULL.
4860 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
4861 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
4865 * real.c (real_to_integer2): Change type of low to UHWI.
4867 2014-03-26 Tobias Burnus <burnus@net-b.de>
4869 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
4870 (CILK_SELF_SPECS): New define.
4871 (driver_self_specs): Use it.
4873 2014-03-26 Richard Biener <rguenther@suse.de>
4875 * tree-pretty-print.c (percent_K_format): Implement special
4876 case for LTO and its stripped down BLOCK tree.
4878 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4881 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
4883 * tree-vrp.c (simplify_internal_call_using_ranges): If only
4884 one range is range_int_cst_p, but not both, at least optimize
4885 addition/subtraction of 0 and multiplication by 0 or 1.
4886 * gimple-fold.c (gimple_fold_call): Fold
4887 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
4888 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
4889 INTEGER_CSTs, try to fold at least x * 0 and y - y.
4891 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
4893 PR rtl-optimization/60452
4894 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
4895 <case REG>: Return 1 for invalid offsets from the frame pointer.
4897 2014-03-26 Marek Polacek <polacek@redhat.com>
4900 * doc/extend.texi (C Extensions): Mention variable-length arrays in
4903 2014-03-26 Marek Polacek <polacek@redhat.com>
4906 * doc/extend.texi (Designated Inits): Describe what happens to omitted
4909 2014-03-26 Marek Polacek <polacek@redhat.com>
4912 * ira-color.c (update_conflict_hard_regno_costs): Perform the
4913 multiplication in unsigned type.
4915 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4917 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
4919 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4921 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
4923 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4926 * cif-code.def (UNREACHABLE) New code.
4927 * ipa-inline.c (inline_small_functions): Skip edges to
4928 __builtlin_unreachable.
4929 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
4930 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
4931 predicate to __bulitin_unreachable.
4932 (set_cond_stmt_execution_predicate): Fix issue when
4933 invert_tree_comparison returns ERROR_MARK.
4934 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
4935 propagate to inline clones.
4936 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
4938 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
4939 * cgraphclones.c (cgraph_clone_node): If call destination is already
4940 ureachable, do not redirect it back.
4941 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
4944 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4946 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
4947 Do not modify inline clones.
4949 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4951 * config/i386/i386.md (general_sext_operand): New mode attr.
4952 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
4953 don't generate (sign_extend (const_int)).
4954 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
4955 operands[2]. Use We constraint instead of <i> and
4956 <general_sext_operand> predicate instead of <general_operand>.
4957 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
4958 * config/i386/constraints.md (We): New constraint.
4959 * config/i386/predicates.md (x86_64_sext_operand,
4960 sext_operand): New predicates.
4962 2014-03-25 Martin Jambor <mjambor@suse.cz>
4965 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
4966 inconsistent devirtualizations to __builtin_unreachable.
4968 2014-03-25 Marek Polacek <polacek@redhat.com>
4971 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
4973 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
4975 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
4976 order of elements for big-endian.
4978 2014-03-25 Richard Biener <rguenther@suse.de>
4981 * gimplify-me.c (gimple_regimplify_operands): Update the
4984 2014-03-25 Martin Jambor <mjambor@suse.cz>
4987 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
4988 (lto_output_varpool_node): Likewise.
4989 (input_overwrite_node): Likewise.
4990 (input_varpool_node): Likewise.
4992 2014-03-25 Richard Biener <rguenther@suse.de>
4994 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
4995 (run_gcc): Likewise.
4997 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4999 * combine.c (simplify_compare_const): Add MODE argument.
5000 Handle mode_width 0 as very large mode_width.
5001 (try_combine, simplify_comparison): Adjust callers.
5003 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
5004 type to avoid signed integer overflow.
5005 * explow.c (plus_constant): Likewise.
5007 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
5009 * doc/generic.texi: Correct typos.
5011 2014-03-24 Tobias Burnus <burnus@net-b.de>
5013 * doc/invoke.texi (-flto): Expand section about
5014 using static libraries with LTO.
5016 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5018 PR rtl-optimization/60501
5019 * optabs.def (addptr3_optab): New optab.
5020 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
5021 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
5022 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
5024 * lra.c (emit_add3_insn): Use the addptr pattern if available.
5026 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
5028 2014-03-24 Ulrich Drepper <drepper@gmail.com>
5030 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
5033 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
5034 (_mm256_undefined_ps): Define.
5035 (_mm256_undefined_pd): Define.
5036 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
5037 (_mm_undefined_pd): Define.
5038 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
5039 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
5040 (_mm512_undefined_ps): Define.
5041 (_mm512_undefined_pd): Define.
5042 Use _mm*_undefined_*.
5043 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
5045 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
5047 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
5048 (lshr_simd): DI mode added.
5049 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
5050 (aarch64_ushr_simddi): Likewise.
5051 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
5052 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
5053 (vshrd_n_u64): Likewise.
5055 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5057 * Makefile.in (s-macro_list): Depend on cc1.
5059 2014-03-23 Teresa Johnson <tejohnson@google.com>
5061 * ipa-utils.c (ipa_print_order): Use specified dump file.
5063 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
5065 PR rtl-optimization/60601
5066 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
5068 * gcc.c (eval_spec_function): Initialize save_growing_value.
5070 2014-03-22 Jakub Jelinek <jakub@redhat.com>
5073 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
5074 code == MINUS_EXPR, never swap op0 with op1.
5076 * toplev.c (init_local_tick): Avoid signed integer multiplication
5078 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
5079 shift by first operand's bitsize.
5081 2014-03-21 Jakub Jelinek <jakub@redhat.com>
5084 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
5086 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
5087 TARGET_ISA_64BIT_P(x).
5089 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5091 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
5092 pattern for vector nor instead of subtract from splat(-1).
5093 (altivec_expand_vec_perm_const_le): Likewise.
5095 2014-03-21 Richard Henderson <rth@twiddle.net>
5098 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
5099 related insns after epilogue_completed.
5101 2014-03-21 Martin Jambor <mjambor@suse.cz>
5104 * cgraph.h (symtab_node): New flag body_removed.
5105 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
5106 when removing bodies.
5107 * symtab.c (dump_symtab_base): Dump body_removed flag.
5108 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
5109 had their bodies removed.
5111 2014-03-21 Martin Jambor <mjambor@suse.cz>
5114 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
5117 2014-03-21 Richard Biener <rguenther@suse.de>
5119 PR tree-optimization/60577
5120 * tree-core.h (struct tree_base): Document nothrow_flag use
5121 in VAR_DECL_NONALIASED.
5122 * tree.h (VAR_DECL_NONALIASED): New.
5123 (may_be_aliased): Adjust.
5124 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
5126 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5128 * expr.c (expand_expr_real_1): Remove outdated comment.
5130 2014-03-20 Jakub Jelinek <jakub@redhat.com>
5133 * ira.c (adjust_cleared_regs): Call copy_rtx on
5134 *reg_equiv[REGNO (loc)].src_p before passing it to
5135 simplify_replace_fn_rtx.
5138 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
5139 into CONST, put pic register as first operand of PLUS. Use
5140 gen_const_mem for both 32-bit and 64-bit PIC got loads.
5142 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5144 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
5146 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5148 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
5149 around for store forwarding issue in the FPU on the UT699.
5150 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
5151 loads and operations if -mfix-ut699 is specified.
5152 (divtf3_hq): Tweak attribute.
5153 (sqrttf2_hq): Likewise.
5155 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5157 * calls.c (store_one_arg): Remove incorrect const qualification on the
5158 type of the temporary.
5159 * cfgexpand.c (expand_return): Likewise.
5160 * expr.c (expand_constructor): Likewise.
5161 (expand_expr_real_1): Likewise.
5163 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5165 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
5168 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
5171 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
5173 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
5175 * config/arm/aarch-common-protos.h
5176 (alu_cost_table): Fix spelling of "extend".
5177 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
5179 2014-03-19 Richard Biener <rguenther@suse.de>
5182 * tree-core.h (tree_type_common): Re-order pointer members
5183 to reduce recursion depth during GC walks.
5185 2014-03-19 Marek Polacek <polacek@redhat.com>
5188 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
5189 before accessing it.
5191 2014-03-19 Richard Biener <rguenther@suse.de>
5194 * lto-streamer-in.c (input_function): In WPA stage do not drop
5197 2014-03-19 Jakub Jelinek <jakub@redhat.com>
5199 PR tree-optimization/60559
5200 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
5201 with build_zero_cst assignment.
5203 2014-03-18 Kai Tietz <ktietz@redhat.com>
5205 PR rtl-optimization/56356
5206 * sdbout.c (sdbout_parms): Verify that parms'
5207 incoming argument is valid.
5208 (sdbout_reg_parms): Likewise.
5210 2014-03-18 Richard Henderson <rth@redhat.com>
5213 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
5214 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
5215 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
5217 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
5219 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
5220 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
5221 Italicize plugin event names in description. Explain that
5222 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
5223 Remind that no GCC functions should be called after PLUGIN_FINISH.
5224 Explain what pragmas with expansion are.
5226 2014-03-18 Martin Liska <mliska@suse.cz>
5228 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
5229 gimple call statement is update.
5230 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
5231 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
5233 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5236 * ubsan.c (ubsan_instrument_unreachable): Call
5237 initialize_sanitizer_builtins.
5238 (ubsan_pass): Likewise.
5241 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
5242 varpool_finalize_decl instead of rest_of_decl_compilation.
5244 2014-03-18 Richard Biener <rguenther@suse.de>
5246 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
5247 by using bitmap_and_compl instead of bitmap_and_compl_into.
5248 (df_rd_transfer_function): Likewise.
5250 2014-03-18 Richard Biener <rguenther@suse.de>
5252 * doc/lto.texi (fresolution): Fix typo.
5254 2014-03-18 Richard Biener <rguenther@suse.de>
5256 * doc/invoke.texi (flto): Update for changes in 4.9.
5258 2014-03-18 Richard Biener <rguenther@suse.de>
5260 * doc/loop.texi: Remove section on the removed lambda framework.
5261 Update loop docs with recent changes in preserving loop structure.
5263 2014-03-18 Richard Biener <rguenther@suse.de>
5265 * doc/lto.texi (-fresolution): Document.
5267 2014-03-18 Richard Biener <rguenther@suse.de>
5269 * doc/contrib.texi: Adjust my name.
5271 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5274 * internal-fn.c: Include diagnostic-core.h.
5275 (expand_BUILTIN_EXPECT): New function.
5276 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
5277 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
5278 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
5279 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
5281 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
5282 Revert 3 argument __builtin_expect code.
5283 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
5284 * gimple-fold.c (gimple_fold_call): Likewise.
5285 * tree.h (fold_builtin_expect): New prototype.
5286 * builtins.c (build_builtin_expect_predicate): Add predictor
5287 argument, if non-NULL, create 3 argument __builtin_expect.
5288 (fold_builtin_expect): No longer static. Add ARG2 argument,
5289 pass it through to build_builtin_expect_predicate.
5290 (fold_builtin_2): Adjust caller.
5291 (fold_builtin_3): Handle BUILT_IN_EXPECT.
5292 * internal-fn.def (BUILTIN_EXPECT): New.
5294 2014-03-18 Tobias Burnus <burnus@net-b.de>
5297 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
5298 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
5299 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
5301 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
5304 * predict.c (combine_predictions_for_bb): Fix up formatting.
5305 (expr_expected_value_1, expr_expected_value): Add predictor argument,
5306 fill what it points to if non-NULL.
5307 (tree_predict_by_opcode): Adjust caller, use the predictor.
5308 * predict.def (PRED_COMPARE_AND_SWAP): Add.
5310 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
5312 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
5313 proper constant for the store mode.
5315 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
5317 * symtab.c (change_decl_assembler_name): Fix transparent alias
5320 2014-03-16 Renlin Li <Renlin.Li@arm.com>
5322 * config/aarch64/aarch64.c: Correct the comments about the
5323 aarch64 stack layout.
5325 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
5327 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
5328 check for GF_OMP_FOR_KIND_FOR.
5330 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
5332 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
5333 ymm and zmm register names.
5335 2014-03-17 Jakub Jelinek <jakub@redhat.com>
5338 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
5339 note creation for the 2010-08-31 changes.
5341 2014-03-17 Marek Polacek <polacek@redhat.com>
5344 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
5345 as -fno-tree-loop-vectorize.
5346 (expand_omp_simd): Likewise.
5348 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
5350 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
5351 (eligible_for_call_delay): New prototype.
5352 * config/sparc/sparc.c (tls_call_delay): Rename into...
5353 (eligible_for_call_delay): ...this. Return false if the instruction
5354 cannot be put in the delay slot of a branch.
5355 (eligible_for_restore_insn): Simplify.
5356 (eligible_for_return_delay): Return false if the instruction cannot be
5357 put in the delay slot of a branch and simplify.
5358 (eligible_for_sibcall_delay): Return false if the instruction cannot be
5359 put in the delay slot of a branch.
5360 * config/sparc/sparc.md (fix_ut699): New attribute.
5361 (tls_call_delay): Delete.
5362 (in_call_delay): Reimplement.
5363 (eligible_for_sibcall_delay): Rename into...
5364 (in_sibcall_delay): ...this.
5365 (eligible_for_return_delay): Rename into...
5366 (in_return_delay): ...this.
5367 (in_branch_delay): Reimplement.
5368 (in_uncond_branch_delay): Delete.
5369 (in_annul_branch_delay): Delete.
5371 2014-03-14 Richard Henderson <rth@redhat.com>
5374 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
5375 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
5376 (*floathi<X87MODEF>2_i387_with_temp): Remove.
5377 (floathi splitters): Remove.
5378 (float<SWI48x>xf2): New pattern.
5379 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
5380 code that tried to handle DImode for 32-bit, but which was excluded
5381 by the pattern's condition. Drop allocation of stack temporary.
5382 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
5383 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
5384 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
5385 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
5386 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
5387 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
5388 (*float<SWI48><MODEF>2_sse_interunit): Remove.
5389 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
5390 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
5391 (*float<SWI48x><X87MODEF>2_i387): Remove.
5392 (all float _with_temp splitters): Remove.
5393 (*float<SWI48x><MODEF>2_i387): New pattern.
5394 (*float<SWI48><MODEF>2_sse): New pattern.
5395 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
5396 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
5398 2014-03-14 Jakub Jelinek <jakub@redhat.com>
5399 Marek Polacek <polacek@redhat.com>
5402 * common.opt (dump_base_name_prefixed): New Variable.
5403 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
5404 if x_dump_base_name_prefixed is already set, set it at the end.
5406 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
5408 PR rtl-optimization/60508
5409 * lra-constraints.c (get_reload_reg): Add new parameter
5411 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
5412 Pass the new parameter values.
5414 2014-03-14 Richard Biener <rguenther@suse.de>
5416 * common.opt: Revert unintented changes from r205065.
5419 2014-03-14 Richard Biener <rguenther@suse.de>
5422 * cfghooks.c (split_block): Properly adjust all loops the
5423 block was a latch of.
5425 2014-03-14 Martin Jambor <mjambor@suse.cz>
5428 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
5431 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
5434 * config/avr/avr.c (avr_set_current_function): Pass function name
5435 through default_strip_name_encoding before sanity checking instead
5436 of skipping the first char of the assembler name.
5438 2014-03-13 Richard Henderson <rth@redhat.com>
5441 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
5442 (ix86_force_to_memory, ix86_free_from_memory): Remove.
5443 * config/i386/i386-protos.h: Likewise.
5444 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
5445 in the expander instead of a splitter.
5446 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
5447 any possibility of requiring a memory.
5448 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
5449 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
5450 (fp branch splitters): Update for ix86_split_fp_branch.
5451 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
5452 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
5453 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
5454 (*fop_<MODEF>_2_i387): Remove f/r alternative.
5455 (*fop_<MODEF>_3_i387): Likewise.
5456 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
5457 (splitters for the fop_* register patterns): Remove.
5458 (fscalexf4_i387): Rename from *fscalexf4_i387.
5459 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
5461 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5463 PR tree-optimization/59779
5464 * tree-dfa.c (get_ref_base_and_extent): Use double_int
5465 type for bitsize and maxsize instead of HOST_WIDE_INT.
5467 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
5469 PR rtl-optimization/57320
5470 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
5471 the CFG after thread_prologue_and_epilogue_insns.
5473 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
5475 PR rtl-optimization/57189
5476 * lra-constraints.c (process_alt_operands): Disfavor spilling
5479 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
5481 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
5483 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5485 PR tree-optimization/59025
5487 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
5488 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
5490 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
5493 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
5494 calls of avr_out_plus_1.
5496 2014-03-13 Bin Cheng <bin.cheng@arm.com>
5498 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
5499 BB's single pred and update the father loop's latch info later.
5501 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5503 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
5507 (VEC_base): Likewise.
5508 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
5509 registers, we need to swap double words in little endian mode.
5511 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
5512 to be a container mode for 128-bit integer operations added in ISA
5513 2.07. Unlike TImode and PTImode, the preferred register set is
5514 the Altivec/VMX registers for the 128-bit operations.
5516 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
5518 (rs6000_split_128bit_ok_p): Likewise.
5520 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
5521 macros for creating ISA 2.07 normal and overloaded builtin
5522 functions with 3 arguments.
5523 (BU_P8V_OVERLOAD_3): Likewise.
5524 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
5525 for use as overloaded functions.
5526 (VPERM_1TI_UNS): Likewise.
5527 (VSEL_1TI): Likewise.
5528 (VSEL_1TI_UNS): Likewise.
5529 (ST_INTERNAL_1ti): Likewise.
5530 (LD_INTERNAL_1ti): Likewise.
5531 (XXSEL_1TI): Likewise.
5532 (XXSEL_1TI_UNS): Likewise.
5533 (VPERM_1TI): Likewise.
5534 (VPERM_1TI_UNS): Likewise.
5535 (XXPERMDI_1TI): Likewise.
5536 (SET_1TI): Likewise.
5537 (LXVD2X_V1TI): Likewise.
5538 (STXVD2X_V1TI): Likewise.
5539 (VEC_INIT_V1TI): Likewise.
5540 (VEC_SET_V1TI): Likewise.
5541 (VEC_EXT_V1TI): Likewise.
5542 (EQV_V1TI): Likewise.
5543 (NAND_V1TI): Likewise.
5544 (ORC_V1TI): Likewise.
5545 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
5546 added in ISA 2.07. Add both normal 'altivec' builtins, and the
5548 (VADDUQM): Likewise.
5549 (VSUBCUQ): Likewise.
5550 (VADDEUQM): Likewise.
5551 (VADDECUQ): Likewise.
5552 (VSUBEUQM): Likewise.
5553 (VSUBECUQ): Likewise.
5555 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5556 __int128_t and __uint128_t types.
5557 (__uint128_type): Likewise.
5558 (altivec_categorize_keyword): Add support for vector __int128_t,
5559 vector __uint128_t, vector __int128, and vector unsigned __int128
5560 as a container type for TImode operations that need to be done in
5561 VSX/Altivec registers.
5562 (rs6000_macro_to_expand): Likewise.
5563 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5564 to support 128-bit integer instructions vaddcuq, vadduqm,
5565 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5566 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5568 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5569 for V1TImode, and set up preferences to use VSX/Altivec registers.
5570 Setup VSX reload handlers.
5571 (rs6000_debug_reg_global): Likewise.
5572 (rs6000_init_hard_regno_mode_ok): Likewise.
5573 (rs6000_preferred_simd_mode): Likewise.
5574 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5575 (easy_altivec_constant): Likewise.
5576 (output_vec_const_move): Likewise.
5577 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5579 (rs6000_expand_vector_extract): Likewise.
5580 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
5582 (rs6000_const_vec): Add support for V1TImode.
5583 (rs6000_emit_le_vsx_load): Swap double words when loading or
5584 storing TImode/V1TImode.
5585 (rs6000_emit_le_vsx_store): Likewise.
5586 (rs6000_emit_le_vsx_move): Likewise.
5587 (rs6000_emit_move): Add support for V1TImode.
5588 (altivec_expand_ld_builtin): Likewise.
5589 (altivec_expand_st_builtin): Likewise.
5590 (altivec_expand_vec_init_builtin): Likewise.
5591 (altivec_expand_builtin): Likewise.
5592 (rs6000_init_builtins): Add support for V1TImode type. Add
5593 support for ISA 2.07 128-bit integer builtins. Define type names
5594 for the VSX/Altivec vector types.
5595 (altivec_init_builtins): Add support for overloaded vector
5596 functions with V1TImode type.
5597 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
5598 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
5600 (rs6000_split_128bit_ok_p): Likewise.
5601 (rs6000_handle_altivec_attribute): Create V1TImode from vector
5602 __int128_t and vector __uint128_t.
5604 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
5605 and mode attributes.
5612 (VS_scalar): Likewise.
5613 (VS_double): Likewise.
5614 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
5616 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
5617 we support the ISA 2.07 128-bit integer arithmetic instructions.
5618 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
5619 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
5620 and TImode types for use with the builtin functions.
5621 (V1TI_type_node): Likewise.
5622 (unsigned_V1TI_type_node): Likewise.
5623 (intTI_type_internal_node): Likewise.
5624 (uintTI_type_internal_node): Likewise.
5626 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
5627 128-bit builtin functions.
5628 (UNSPEC_VADDEUQM): Likewise.
5629 (UNSPEC_VADDECUQ): Likewise.
5630 (UNSPEC_VSUBCUQ): Likewise.
5631 (UNSPEC_VSUBEUQM): Likewise.
5632 (UNSPEC_VSUBECUQ): Likewise.
5633 (VM): Add V1TImode to vector mode iterators.
5635 (VI_unit): Likewise.
5636 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
5637 (altivec_vaddcuq): Likewise.
5638 (altivec_vsubuqm): Likewise.
5639 (altivec_vsubcuq): Likewise.
5640 (altivec_vaddeuqm): Likewise.
5641 (altivec_vaddecuq): Likewise.
5642 (altivec_vsubeuqm): Likewise.
5643 (altivec_vsubecuq): Likewise.
5645 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
5647 (BOOL_128): Likewise.
5648 (BOOL_REGS_OUTPUT): Likewise.
5649 (BOOL_REGS_OP1): Likewise.
5650 (BOOL_REGS_OP2): Likewise.
5651 (BOOL_REGS_UNARY): Likewise.
5652 (BOOL_REGS_AND_CR0): Likewise.
5654 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
5655 128-bit integer builtin support.
5656 (vec_vadduqm): Likewise.
5657 (vec_vaddecuq): Likewise.
5658 (vec_vaddeuqm): Likewise.
5659 (vec_vsubecuq): Likewise.
5660 (vec_vsubeuqm): Likewise.
5661 (vec_vsubcuq): Likewise.
5662 (vec_vsubuqm): Likewise.
5664 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5665 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
5666 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
5667 128-bit integer add/subtract to ISA 2.07.
5669 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
5671 * config/arc/arc.c (arc_predicate_delay_insns):
5672 Fix third argument passed to conditionalize_nonjump.
5674 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
5676 * config/aarch64/aarch64-builtins.c
5677 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
5678 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
5679 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
5680 instead of __builtin_lfloor.
5681 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
5683 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5685 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
5686 (tree_ssa_ifcombine_bb_1): New function.
5687 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
5688 is an empty forwarder block to then_bb or vice versa and then_bb
5689 and else_bb are effectively swapped.
5691 2014-03-12 Christian Bruel <christian.bruel@st.com>
5694 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
5695 REG_CFA_DEF_CFA note.
5696 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
5697 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
5699 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
5701 PR tree-optimization/60454
5702 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
5704 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5706 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
5707 Do not define target_cpu_default2 to generic.
5708 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
5709 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
5710 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
5712 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5713 Marc Glisse <marc.glisse@inria.fr>
5715 PR tree-optimization/60502
5716 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
5717 instead of build_low_bits_mask.
5719 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5722 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
5723 if there are multiple uses, but op doesn't live on E edge.
5724 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
5725 clobber stmts before __builtin_unreachable.
5727 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
5729 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
5730 hard_frame_pointer_rtx.
5731 * cse.c (cse_insn): Remove volatile check.
5732 * cselib.c (cselib_process_insn): Likewise.
5733 * dse.c (scan_insn): Likewise.
5735 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5737 * config/arc/arc.c (conditionalize_nonjump): New function,
5739 (arc_ifcvt): ... this.
5740 (arc_predicate_delay_insns): Use it.
5742 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5744 * config/arc/predicates.md (extend_operand): During/after reload,
5745 allow const_int_operand.
5746 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
5747 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
5748 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
5750 (umulsi3_highpart_i): Likewise.
5752 2014-03-11 Richard Biener <rguenther@suse.de>
5754 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
5755 Add asserts to guard possible wrong-code bugs.
5757 2014-03-11 Richard Biener <rguenther@suse.de>
5759 PR tree-optimization/60429
5760 PR tree-optimization/60485
5761 * tree-ssa-structalias.c (set_union_with_increment): Properly
5762 take into account all fields that overlap the shifted vars.
5763 (do_sd_constraint): Likewise.
5764 (do_ds_constraint): Likewise.
5765 (get_constraint_for_ptr_offset): Likewise.
5767 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
5769 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
5770 (nios2_compute_frame_layout):
5771 Add calculation of cfun->machine->fp_save_offset.
5772 (nios2_expand_prologue): Correct setting of frame pointer register
5774 (nios2_expand_epilogue): Update recovery of stack pointer from
5775 frame pointer accordingly.
5776 (nios2_initial_elimination_offset): Update calculation of offset
5777 for eliminating to HARD_FRAME_POINTER_REGNUM.
5779 2014-03-10 Jakub Jelinek <jakub@redhat.com>
5782 * ipa.c (symtab_remove_unreachable_nodes): Don't call
5783 cgraph_get_create_node on VAR_DECLs.
5785 2014-03-10 Richard Biener <rguenther@suse.de>
5788 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
5790 2014-03-08 Douglas B Rupp <rupp@gnat.com>
5792 * config/vms/vms.opt (vms_float_format): New variable.
5794 2014-03-08 Tobias Burnus <burnus@net-b.de>
5796 * doc/invoke.texi (-fcilkplus): Update implementation status.
5798 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
5799 Richard Biener <rguenther@suse.de>
5801 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
5802 consistently accross all TUs.
5803 (run_gcc): Enable -fshort-double automatically at link at link-time
5804 and disallow override.
5806 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
5809 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
5810 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
5811 if they can't be used.
5813 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5815 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
5816 for Solaris 11/x86 ld.
5817 * configure: Regenerate.
5819 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5821 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
5822 (LIB_TLS_SPEC): Save as ld_tls_libs.
5823 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
5824 (HAVE_AS_IX86_TLSLDM): New test.
5825 * configure, config.in: Regenerate.
5826 * config/i386/i386.c (legitimize_tls_address): Fall back to
5827 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
5828 cannot support TLS_MODEL_LOCAL_DYNAMIC.
5829 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
5830 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
5832 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
5834 * common.opt (fira-loop-pressure): Mark as optimization.
5836 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
5838 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
5839 an OpenMP mappable type.
5841 2014-03-06 Matthias Klose <doko@ubuntu.com>
5843 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
5844 MULTILIB_OSDIRNAMES is not defined.
5846 2014-03-06 Jakub Jelinek <jakub@redhat.com>
5847 Meador Inge <meadori@codesourcery.com>
5850 * config/arm/arm.c (arm_tls_symbol_p): Remove.
5851 (arm_legitimize_address): Call legitimize_tls_address for any
5852 arm_tls_referenced_p expression, handle constant addend. Call it
5853 before testing for !TARGET_ARM.
5854 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
5856 2014-03-06 Richard Biener <rguenther@suse.de>
5862 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5864 * tree-streamer.c (record_common_node): Assert we don't record
5865 nodes with type double.
5866 (preload_common_node): Skip type double, complex double and double
5867 pointer since it is now frontend dependent due to fshort-double option.
5869 2014-03-06 Richard Biener <rguenther@suse.de>
5871 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
5872 or -fno-lto is specified and the linker has full plugin support.
5873 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
5874 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
5875 * lto-wrapper.c (merge_and_complain): Merge compile-time
5876 optimization levels.
5877 (run_gcc): And pass it through to the link options.
5879 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
5883 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5885 * cselib.c (remove_useless_values): Skip to avoid quadratic
5886 behavior if the condition moved from...
5887 (cselib_process_insn): ... here holds.
5889 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5892 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
5893 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
5896 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
5897 (TM_H): Add x86-tune.def.
5899 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5901 * config/aarch64/aarch64.c (generic_tunings):
5902 Use cortexa57_extra_costs.
5904 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5907 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
5908 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
5909 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
5912 2014-03-04 Heiher <r@hev.cc>
5914 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
5915 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
5917 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
5919 * config/i386/predicates.md (const2356_operand): Change to ...
5920 (const2367_operand): ... this.
5921 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
5923 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5924 (*avx512pf_scatterpf<mode>sf): Ditto.
5925 (avx512pf_scatterpf<mode>df): Ditto.
5926 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5927 (*avx512pf_scatterpf<mode>df): Ditto.
5928 * config/i386/i386.c (ix86_expand_builtin): Update
5929 incorrect hint operand error message.
5931 2014-03-04 Richard Biener <rguenther@suse.de>
5933 * lto-section-in.c (lto_get_section_data): Fix const cast.
5935 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5937 * tree-streamer.c (record_common_node): Assert we don't record
5938 nodes with type double.
5939 (preload_common_node): Skip type double, complex double and double
5940 pointer since it is now frontend dependent due to fshort-double option.
5942 2014-03-04 Richard Biener <rguenther@suse.de>
5945 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
5946 (lto_input_toplevel_asms): Likewise.
5947 * lto-section-in.c (lto_get_section_data): Instead do it here
5950 2014-03-04 Richard Biener <rguenther@suse.de>
5952 PR tree-optimization/60382
5953 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
5954 dead PHIs a reduction.
5956 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
5958 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
5960 (_mm_prefetch): Move out of GCC target("sse") pragma.
5961 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
5962 GCC target("prfchw") pragma.
5963 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
5965 * config/i386/i386.c (ix86_option_override_internal): Enable
5966 -mprfchw with -mprefetchwt1.
5968 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5970 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
5973 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5975 * opts.h (CL_PCH_IGNORE): Define.
5976 * targhooks.c (option_affects_pch_p):
5977 Return false for options that have CL_PCH_IGNORE set.
5978 * opt-functions.awk: Process PchIgnore.
5979 * doc/options.texi: Document PchIgnore.
5981 * config/arc/arc.opt (misize): Add PchIgnore property.
5983 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5985 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
5986 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
5987 constraint on constants to permit them being loaded into
5988 GENERAL_REGS or BASE_REGS.
5990 2014-03-03 Nick Clifton <nickc@redhat.com>
5992 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
5993 anti-cacnonical alternatives.
5994 (negandhi3_real): New pattern.
5995 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
5997 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5999 * config/avr/avr-mcus.def: Remove atxmega16x1.
6000 * config/avr/avr-tables.opt: Regenerate.
6001 * config/avr/t-multilib: Regenerate.
6002 * doc/avr-mmcu.texi: Regenerate.
6004 2014-03-03 Tobias Grosser <tobias@grosser.es>
6005 Mircea Namolaru <mircea.namolaru@inria.fr>
6007 PR tree-optimization/58028
6008 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
6011 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6013 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
6014 not handled by recognizers.
6016 2014-03-03 Jakub Jelinek <jakub@redhat.com>
6019 * function.c (expand_function_end): Don't emit
6020 clobber_return_register sequence if clobber_after is a BARRIER.
6021 * cfgexpand.c (construct_exit_block): Append instructions before
6022 return_label to prev_bb.
6024 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6026 * config/rs6000/constraints.md: Document reserved use of "wc".
6028 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
6031 * ipa.c (function_and_variable_visibility): When dissolving comdat
6032 group, also set all symbols to local.
6034 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
6039 2013-12-14 Jan Hubicka <jh@suse.cz>
6041 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
6043 2014-03-02 Jon Beniston <jon@beniston.com>
6049 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
6050 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
6051 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
6052 (simple_return, *simple_return): New patterns
6053 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
6054 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
6056 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
6058 * dwarf2out.c (gen_subprogram_die): Tidy.
6060 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
6063 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
6064 (*mov_t_msb_neg_negc): ... this new insn.
6066 2014-02-28 Jason Merrill <jason@redhat.com>
6069 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
6072 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
6075 * dwarf2out.c (decltype_auto_die): New static.
6076 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
6077 (gen_type_die_with_usage): Handle 'decltype(auto)'.
6078 (is_cxx_auto): Likewise.
6080 2014-02-28 Ian Bolton <ian.bolton@arm.com>
6082 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
6083 we are not using general regs only.
6085 2014-02-28 Richard Biener <rguenther@suse.de>
6088 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
6089 previous fix and only allow to remove trivial pre-headers
6090 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
6091 (remove_forwarder_block): Properly update the latch of a loop.
6093 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6096 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
6097 (cselib_preserved_hash_table): New.
6098 (preserve_constants_and_equivs): Move preserved vals to it.
6099 (cselib_find_slot): Look it up first.
6100 (cselib_init): Initialize it.
6101 (cselib_finish): Release it.
6102 (dump_cselib_table): Dump it.
6104 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6107 * cselib.c (remove_useless_values): Skip to avoid quadratic
6108 behavior if the condition moved from...
6109 (cselib_process_insn): ... here holds.
6111 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6114 * var-tracking.c (vt_initialize): Apply the same condition to
6115 preserve the CFA base value.
6117 2014-02-28 Joey Ye <joey.ye@arm.com>
6120 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
6121 if reload in progress or completed.
6123 2014-02-28 Tobias Burnus <burnus@net-b.de>
6126 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
6129 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
6131 * doc/tm.texi.in (Condition Code Status): Update documention for
6132 relative locations of cc0-setter and cc0-user.
6134 2014-02-27 Jeff Law <law@redhat.com>
6136 PR rtl-optimization/52714
6137 * combine.c (try_combine): When splitting an unrecognized PARALLEL
6138 into two independent simple sets, if I3 is a jump, ensure the
6139 pattern we place into I3 is a (set (pc) ...).
6141 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
6142 Jeff Law <law@redhat.com>
6144 PR rtl-optimization/49847
6145 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
6146 are in different blocks.
6147 * doc/tm.texi (Condition Code Status): Update documention for
6148 relative locations of cc0-setter and cc0-user.
6150 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
6153 * lra.c (lra_emit_add): Check SUBREG too.
6155 2014-02-27 Andreas Schwab <schwab@suse.de>
6157 * config/m68k/m68k.c (m68k_option_override): Disable
6158 -flive-range-shrinkage for classic m68k.
6159 (m68k_override_options_after_change): Likewise.
6161 2014-02-27 Marek Polacek <polacek@redhat.com>
6164 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
6165 -Wmaybe-uninitialized.
6167 2014-02-27 Alan Modra <amodra@gmail.com>
6170 * reload1.c (emit_input_reload_insns): When reload_override_in,
6171 set old to rl->in_reg when rl->in_reg is a subreg.
6173 2014-02-26 Richard Biener <rguenther@suse.de>
6176 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
6178 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6180 * common/config/i386/predicates.md (const1256_operand): Remove.
6181 (const2356_operand): New.
6182 (const_1_to_2_operand): Remove.
6183 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6184 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6185 (*avx512pf_gatherpf<mode>sf): Ditto.
6186 (avx512pf_gatherpf<mode>df): Ditto.
6187 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6188 (*avx512pf_gatherpf<mode>df): Ditto.
6189 (avx512pf_scatterpf<mode>sf): Ditto.
6190 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6191 (*avx512pf_scatterpf<mode>sf): Ditto.
6192 (avx512pf_scatterpf<mode>df): Ditto.
6193 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6194 (*avx512pf_scatterpf<mode>df): Ditto.
6195 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
6197 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
6199 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
6200 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
6201 (_mm512_mask_testn_epi64_mask): Move to ...
6202 * config/i386/avx512cdintrin.h: Here.
6203 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
6204 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
6205 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
6206 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
6207 TARGET_AVX512F from TARGET_AVX512CD.
6209 2014-02-26 Richard Biener <rguenther@suse.de>
6212 * ipa.c (walk_polymorphic_call_targets): Properly guard
6213 call to inline_update_overall_summary.
6215 2014-02-26 Bin Cheng <bin.cheng@arm.com>
6218 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
6219 and latches only if requested. Fix latch if it is removed.
6220 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
6221 LOOPS_HAVE_PREHEADERS.
6223 2014-02-25 Andrew Pinski <apinski@cavium.com>
6225 * builtins.c (expand_builtin_thread_pointer): Create a new target
6226 when the target is NULL.
6228 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
6230 PR rtl-optimization/60317
6231 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6232 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6233 * lra-assigns.c: Include params.h.
6234 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
6235 other reload pseudos considerations.
6237 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6239 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
6240 to use canonical form for nor<mode>3.
6242 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6245 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
6248 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6250 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
6251 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
6252 (ix86_handle_option): Handle OPT_mprefetchwt1.
6253 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
6254 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
6256 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
6257 OPTION_MASK_ISA_PREFETCHWT1.
6258 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
6259 (PTA_PREFETCHWT1): New.
6260 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
6261 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
6262 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
6263 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
6264 (*prefetch_avx512pf_<mode>_: Change into ...
6265 (*prefetch_prefetchwt1_<mode>: This.
6266 * config/i386/i386.opt (mprefetchwt1): New.
6267 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
6268 (_mm_prefetch): Handle intent to write.
6269 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
6271 2014-02-25 Richard Biener <rguenther@suse.de>
6274 * emit-rtl.c (mem_attrs_htab): Remove.
6275 (mem_attrs_htab_hash): Likewise.
6276 (mem_attrs_htab_eq): Likewise.
6277 (set_mem_attrs): Always allocate new mem-attrs when something changed.
6278 (init_emit_once): Do not allocate mem_attrs_htab.
6280 2014-02-25 Richard Biener <rguenther@suse.de>
6283 * lto-opts.c (lto_write_options): Output non-explicit conservative
6284 -fwrapv, -fno-trapv and -fno-strict-overflow.
6285 * lto-wrapper.c (merge_and_complain): Handle merging those options.
6286 (run_gcc): And pass them through.
6288 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6290 * sel-sched.c (calculate_new_fences): New parameter ptime.
6291 Calculate it as a maximum over all fence cycles.
6292 (sel_sched_region_2): Adjust the call to calculate_new_fences.
6293 Print the final schedule timing when sched_verbose.
6295 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6297 PR rtl-optimization/60292
6298 * sel-sched.c (fill_vec_av_set): Do not reset target availability
6299 bit fot the fence instruction.
6301 2014-02-24 Alangi Derick <alangiderick@gmail.com>
6303 * calls.h: Fix typo in comment.
6305 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
6307 * config/pa/pa.c (pa_output_move_double): Don't valididate when
6308 adjusting offsetable addresses.
6310 2014-02-24 Guozhi Wei <carrot@google.com>
6312 * sparseset.h (sparseset_pop): Fix the wrong index.
6314 2014-02-24 Walter Lee <walt@tilera.com>
6316 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
6317 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
6319 * common/config/tilegx/tilegx-common.c
6320 (TARGET_DEFAULT_TARGET_FLAGS): Define.
6321 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
6323 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
6324 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
6325 (tilegx_gimplify_va_arg_expr): Handle big endian.
6326 (tilegx_expand_unaligned_load): Ditto.
6327 (tilegx_expand_unaligned_store): Ditto.
6328 (TARGET_RETURN_IN_MSB): New.
6329 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
6330 (TARGET_ENDIAN_DEFAULT): New.
6331 (TARGET_BIG_ENDIAN): Handle big endian.
6332 (BYTES_BIG_ENDIAN): Ditto.
6333 (WORDS_BIG_ENDIAN): Ditto.
6334 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
6337 * config/tilegx/tilegx.md (extv): Handle big endian.
6339 (insn_st<n>): Ditto.
6340 (insn_st<n>_add<bitsuffix>): Ditto.
6341 (insn_stnt<n>): Ditto.
6342 (insn_stnt<n>_add<bitsuffix>):Ditto.
6343 (vec_interleave_highv8qi): Handle big endian.
6344 (vec_interleave_highv8qi_be): New.
6345 (vec_interleave_highv8qi_le): New.
6346 (insn_v1int_h): Handle big endian.
6347 (vec_interleave_lowv8qi): Handle big endian.
6348 (vec_interleave_lowv8qi_be): New.
6349 (vec_interleave_lowv8qi_le): New.
6350 (insn_v1int_l): Handle big endian.
6351 (vec_interleave_highv4hi): Handle big endian.
6352 (vec_interleave_highv4hi_be): New.
6353 (vec_interleave_highv4hi_le): New.
6354 (insn_v2int_h): Handle big endian.
6355 (vec_interleave_lowv4hi): Handle big endian.
6356 (vec_interleave_lowv4hi_be): New.
6357 (vec_interleave_lowv4hi_le): New.
6358 (insn_v2int_l): Handle big endian.
6359 (vec_interleave_highv2si): Handle big endian.
6360 (vec_interleave_highv2si_be): New.
6361 (vec_interleave_highv2si_le): New.
6362 (insn_v4int_h): Handle big endian.
6363 (vec_interleave_lowv2si): Handle big endian.
6364 (vec_interleave_lowv2si_be): New.
6365 (vec_interleave_lowv2si_le): New.
6366 (insn_v4int_l): Handle big endian.
6367 * config/tilegx/tilegx.opt (mbig-endian): New option.
6368 (mlittle-endian): New option.
6369 * doc/install.texi: Document tilegxbe-linux.
6370 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
6372 2014-02-24 Martin Jambor <mjambor@suse.cz>
6375 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
6376 there are no parameter descriptors.
6378 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
6380 PR rtl-optimization/60268
6381 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
6382 initialization to ...
6383 (sched_rgn_init): ... here.
6384 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
6386 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6388 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
6391 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
6393 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
6396 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6398 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
6399 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
6401 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6403 * config/microblaze/predicates.md: Add cmp_op predicate.
6404 * config/microblaze/microblaze.md: Add branch_compare instruction
6405 which uses cmp_op predicate and emits cmp insn before branch.
6406 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
6407 to microblaze_expand_conditional_branch and consolidate logic.
6408 (microblaze_expand_conditional_branch): emit branch_compare
6409 insn instead of handling cmp op separate from branch insn.
6411 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6413 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
6416 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6418 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
6419 define_insn with define_expand and new define_insn
6420 *altivec_lve<VI_char>x_internal.
6421 (altivec_stve<VI_char>x): Replace define_insn with define_expand
6422 and new define_insn *altivec_stve<VI_char>x_internal.
6423 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
6425 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
6427 (altivec_expand_stvex_be): New function.
6429 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
6431 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
6432 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
6433 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
6434 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
6436 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
6439 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
6440 instead of emit_move_insn.
6442 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6444 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
6446 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
6447 gen_altivec_vsumsws.
6449 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6451 * config/rs6000/altivec.md (altivec_lvxl): Rename as
6452 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
6453 (altivec_lvxl_<mode>): New define_expand incorporating
6454 -maltivec=be semantics where needed.
6455 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
6456 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
6457 semantics where needed.
6458 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
6459 (altivec_stvx_<mode>): New define_expand incorporating
6460 -maltivec=be semantics where needed.
6461 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
6462 VM2 iterator instead of V4SI.
6463 (altivec_stvxl_<mode>): New define_expand incorporating
6464 -maltivec=be semantics where needed.
6465 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
6466 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
6467 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
6468 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
6469 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
6470 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
6471 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
6472 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
6473 ALTIVEC_BUILTIN_STVXL.
6474 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
6475 (altivec_expand_stvx_be): Likewise.
6476 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
6477 (altivec_expand_lvx_be): Likewise.
6478 (altivec_expand_stvx_be): Likewise.
6479 (altivec_expand_builtin): Add cases for
6480 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
6481 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
6482 (altivec_init_builtins): Add definitions for
6483 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
6484 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
6486 2014-02-21 Catherine Moore <clm@codesourcery.com>
6488 * doc/invoke.texi (mvirt, mno-virt): Document.
6489 * config/mips/mips.opt (mvirt): New option.
6490 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
6492 2014-02-21 Richard Biener <rguenther@suse.de>
6494 PR tree-optimization/60276
6495 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
6496 (STMT_VINFO_MIN_NEG_DIST): New macro.
6497 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
6498 STMT_VINFO_MIN_NEG_DIST.
6499 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
6500 made for negative dependence distances still hold.
6502 2014-02-21 Richard Biener <rguenther@suse.de>
6505 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
6506 DECL_INITIAL for globals not in the current function context.
6508 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6510 PR tree-optimization/56490
6511 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
6512 * tree-ssa-uninit.c: Include params.h.
6513 (compute_control_dep_chain): Add num_calls argument, return false
6514 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
6515 num_calls to recursive call.
6516 (find_predicates): Change dep_chain into normal array,
6517 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
6518 variable and adjust compute_control_dep_chain caller.
6519 (find_def_preds): Likewise.
6521 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
6523 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
6524 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
6526 2014-02-21 Nick Clifton <nickc@redhat.com>
6528 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
6529 (pushhi1): Likewise.
6530 (popqi1): Add mode to pre_dec.
6533 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6535 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
6536 mode for mask of V8SFmode permutation.
6538 2014-02-20 Richard Henderson <rth@redhat.com>
6541 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
6542 a new pseudo for OLDVAL.
6544 2014-02-20 Jakub Jelinek <jakub@redhat.com>
6547 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
6548 gen_reg_rtx if d->testing_p.
6549 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
6550 if d->testing_p and we will certainly return true.
6551 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6554 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6556 * emit-rtl.c (gen_reg_rtx): Assert that
6557 crtl->emit.regno_pointer_align_length is non-zero.
6559 2014-02-20 Richard Henderson <rth@redhat.com>
6562 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6563 on failure the store back into EXPECT.
6565 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6566 Sandra Loosemore <sandra@codesourcery.com>
6568 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6569 * config/nios2/nios2.c (nios2_function_profiler): Add
6570 -fPIC (flag_pic == 2) support.
6571 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6572 (nios2_large_offset_p): New function.
6573 (nios2_unspec_reloc_p): Move up position, update to use
6574 nios2_large_offset_p.
6575 (nios2_unspec_address): Remove function.
6576 (nios2_unspec_offset): New function.
6577 (nios2_large_got_address): New function.
6578 (nios2_got_address): Add large offset support.
6579 (nios2_legitimize_tls_address): Update usage of removed and new
6581 (nios2_symbol_binds_local_p): New function.
6582 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
6583 (nios2_legitimize_address): Update to use nios2_large_offset_p.
6584 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
6585 (nios2_print_operand): Merge H/L processing, add hiadj/lo
6586 processing for (const (unspec ...)).
6587 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
6589 2014-02-20 Richard Biener <rguenther@suse.de>
6591 * tree-cfg.c (replace_uses_by): Mark altered BBs before
6592 doing the substitution.
6593 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
6595 2014-02-20 Martin Jambor <mjambor@suse.cz>
6598 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
6599 info when checking whether lattices are bottom.
6601 2014-02-20 Richard Biener <rguenther@suse.de>
6604 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
6607 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
6610 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
6611 parameter specifying the scaling.
6612 (inline_call): Update.
6613 (want_inline_recursively): Guard division by zero.
6614 (recursive_inlining): Update.
6615 * ipa-inline.h (clone_inlined_nodes): Update.
6617 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6620 * config/i386/i386.c (classify_argument): Pass structures of size
6621 64 bytes or less in register.
6623 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6624 Kirill Yukhin <kirill.yukhin@intel.com>
6626 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
6627 (_mm_rcp28_round_ss): Ditto.
6628 (_mm_rsqrt28_round_sd): Ditto.
6629 (_mm_rsqrt28_round_ss): Ditto.
6630 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
6631 (_mm_rcp14_round_ss): Ditto.
6632 (_mm_rsqrt14_round_sd): Ditto.
6633 (_mm_rsqrt14_round_ss): Ditto.
6634 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
6635 the first input operand, get rid of match_dup.
6636 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
6638 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
6640 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
6641 operand as the first input operand, set type attribute.
6642 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
6644 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
6645 operand as the first input operand, set type attribute.
6647 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6649 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
6652 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
6655 * config/i386/i386.c (construct_container): Remove TFmode check
6656 for X86_64_INTEGER_CLASS.
6658 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
6661 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
6662 only when -Wpsabi is enabled.
6664 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
6667 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
6668 passing arrays in registers are the same as for structs, so remove the
6669 special case for them.
6671 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
6673 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
6674 destination type, extract only the valid bits if the source type is not
6675 integral and has a different mode.
6677 2014-02-19 Richard Biener <rguenther@suse.de>
6680 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
6683 2014-02-19 Richard Biener <rguenther@suse.de>
6686 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
6687 (ipa_modify_call_arguments): Emit an argument load explicitely and
6688 preserve virtual SSA form there and for the replacement call.
6689 Do not update SSA form nor free dominance info.
6691 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6693 * ipa.c (function_and_variable_visibility): Also clear WEAK
6694 flag when disolving COMDAT_GROUP.
6696 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6698 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
6699 * ipa-prop.c (ipa_set_jf_known_type): Return early when
6701 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
6702 do more sanity checks.
6703 (detect_type_change): Return true when giving up early.
6704 (compute_complex_assign_jump_func): Fix type parameter of
6705 ipa_set_ancestor_jf.
6706 (compute_complex_ancestor_jump_func): Likewise.
6707 (update_jump_functions_after_inlining): Fix updating of
6709 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
6711 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6713 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
6714 inline clones when edge disappears.
6716 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6719 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
6720 Split 64-bit moves into 2 patterns. Do not allow the use of
6721 direct move for TDmode in little endian, since the decimal value
6722 has little endian bytes within a word, but the 64-bit pieces are
6723 ordered in a big endian fashion, and normal subreg's of TDmode are
6725 (mov<mode>_64bit_dm): Likewise.
6726 (movtd_64bit_nodm): Likewise.
6728 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6730 PR tree-optimization/60174
6731 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
6732 statement of an SSA_NAME that occurs in an abnormal PHI node.
6734 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6737 * final.c (SEEN_BB): Remove.
6738 (SEEN_NOTE, SEEN_EMITTED): Renumber.
6739 (final_scan_insn): Don't force_source_line on second
6740 NOTE_INSN_BASIC_BLOCK.
6742 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
6745 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
6746 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
6747 (type_natural_mode): Warn ABI change when %zmm register is not
6748 available for AVX512F vector value passing.
6750 2014-02-18 Kai Tietz <ktietz@redhat.com>
6753 * config/i386/i386.c (ix86_expand_prologue): Use value in
6754 rax register as displacement when restoring %r10 or %rax.
6755 Fix wrong offset when restoring both registers.
6757 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6759 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
6760 assertion with conditional return.
6762 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6763 Uros Bizjak <ubizjak@gmail.com>
6766 * config/i386/driver-i386.c (host_detect_local_cpu): If
6767 YMM state is not saved by the OS, also clear has_f16c. Move
6768 CPUID 0x80000001 handling before YMM state saving checking.
6770 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
6772 PR rtl-optimization/58960
6773 * haifa-sched.c (alloc_global_sched_pressure_data): New,
6774 factored out from ...
6775 (sched_init): ... here.
6776 (free_global_sched_pressure_data): New, factored out from ...
6777 (sched_finish): ... here.
6778 * sched-int.h (free_global_sched_pressure_data): Declare.
6779 * sched-rgn.c (nr_regions_initial): New static global.
6780 (haifa_find_rgns): Initialize it.
6781 (schedule_region): Disable sched-pressure for the newly
6784 2014-02-17 Richard Biener <rguenther@suse.de>
6786 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
6787 release SSA defs of pattern stmts.
6789 2014-02-17 Richard Biener <rguenther@suse.de>
6791 * tree-inline.c (expand_call_inline): Release the virtual
6792 operand defined by the call we are about to inline.
6794 2014-02-17 Richard Biener <rguenther@suse.de>
6796 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
6798 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
6799 Ilya Tocar <ilya.tocar@intel.com>
6801 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
6802 arguments order in builtin.
6803 (_mm512_permutexvar_epi64): Ditto.
6804 (_mm512_mask_permutexvar_epi64): Ditto
6805 (_mm512_maskz_permutexvar_epi32): Ditto
6806 (_mm512_permutexvar_epi32): Ditto
6807 (_mm512_mask_permutexvar_epi32): Ditto
6809 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6811 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
6812 (p8_vmrgow): Likewise.
6814 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6816 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
6819 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
6822 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
6823 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
6824 into 64-bit and 32-bit moves. On 64-bit moves, add support for
6825 using direct move instructions on ISA 2.07. Also adjust
6826 instruction length for 64-bit.
6827 (mov<mode>_64bit, TFmode/TDmode): Likewise.
6828 (mov<mode>_32bit, TFmode/TDmode): Likewise.
6830 2014-02-15 Alan Modra <amodra@gmail.com>
6834 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
6835 find_replacement on parts of insn rtl that might be reloaded.
6837 2014-02-15 Richard Biener <rguenther@suse.de>
6839 PR tree-optimization/60183
6840 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
6841 (tree_ssa_phiprop): Calculate and free post-dominators.
6843 2014-02-14 Jeff Law <law@redhat.com>
6845 PR rtl-optimization/60131
6846 * ree.c (get_extended_src_reg): New function.
6847 (combine_reaching_defs): Use it rather than assuming location of REG.
6848 (find_and_remove_re): Verify first operand of extension is
6849 a REG before adding the insns to the copy list.
6851 2014-02-14 Roland McGrath <mcgrathr@google.com>
6853 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
6854 * configure: Regenerated.
6855 * config.in: Regenerated.
6856 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
6857 instead of ASM_SHORT.
6859 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
6860 Richard Earnshaw <rearnsha@arm.com>
6862 PR rtl-optimization/59535
6863 * lra-constraints.c (process_alt_operands): Encourage alternative
6864 when unassigned pseudo class is superset of the alternative class.
6865 (inherit_reload_reg): Don't inherit when optimizing for code size.
6866 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
6867 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
6868 modes not less than 4 for Thumb1.
6870 2014-02-14 Kyle McMartin <kyle@redhat.com>
6873 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
6875 2014-02-14 Richard Biener <rguenther@suse.de>
6877 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
6878 (get_frame_arg): Drop the assert with langhook types_compatible_p.
6879 Do not strip INDIRECT_REFs.
6881 2014-02-14 Richard Biener <rguenther@suse.de>
6884 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
6885 DECL_FUNCTION_SPECIFIC_TARGET.
6886 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
6887 * tree-streamer-out.c (pack_ts_target_option): Remove.
6888 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
6889 (write_ts_function_decl_tree_pointers): Do not stream
6890 DECL_FUNCTION_SPECIFIC_TARGET.
6891 * tree-streamer-in.c (unpack_ts_target_option): Remove.
6892 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
6893 (lto_input_ts_function_decl_tree_pointers): Do not stream
6894 DECL_FUNCTION_SPECIFIC_TARGET.
6896 2014-02-14 Jakub Jelinek <jakub@redhat.com>
6898 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
6899 (get_initial_def_for_induction, vectorizable_induction): Ignore
6900 debug stmts when looking for exit_phi.
6901 (vectorizable_live_operation): Fix up condition.
6903 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6905 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
6906 nreverse() because it changes the content of original tree list.
6908 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6910 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
6911 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
6913 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6915 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
6916 GNU coding standards.
6918 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6921 * dwarf2out.c (gen_subprogram_die): Don't call
6922 add_calling_convention_attribute if subr_die is old_die.
6924 2014-02-13 Sharad Singhai <singhai@google.com>
6926 * doc/optinfo.texi: Fix order of nodes.
6928 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
6930 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
6931 operands[2], not operands[3].
6933 2014-02-13 Richard Biener <rguenther@suse.de>
6936 * doc/install.texi (ISL): Update recommended version to 0.12.2,
6937 mention the possibility of an in-tree build.
6938 (CLooG): Update recommended version to 0.18.1, mention the
6939 possibility of an in-tree build and clarify that the ISL
6940 bundled with CLooG does not work.
6942 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6945 * expr.c (compress_float_constant): If x is a hard register,
6946 extend into a pseudo and then move to x.
6948 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
6950 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
6951 caused by bad second argument to warning_at() with -mhotpatch and
6952 nested functions (e.g. with gfortran).
6954 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
6956 * opts.c (option_name): Remove "enabled by default" rider.
6958 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
6960 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
6962 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
6963 Uros Bizjak <ubizjak@gmail.com>
6966 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
6967 * configure: Regenerated.
6969 2014-02-12 Richard Biener <rguenther@suse.de>
6971 * vec.c (vec_prefix::calculate_allocation): Move as
6972 inline variant to vec.h.
6973 (vec_prefix::calculate_allocation_1): New out-of-line version.
6974 * vec.h (vec_prefix::calculate_allocation_1): Declare.
6975 (vec_prefix::m_has_auto_buf): Rename to ...
6976 (vec_prefix::m_using_auto_storage): ... this.
6977 (vec_prefix::calculate_allocation): Inline the easy cases
6978 and dispatch to calculate_allocation_1 which doesn't need the
6980 (va_heap::reserve): Use gcc_checking_assert.
6981 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
6982 m_using_auto_storage.
6983 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
6985 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
6986 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
6987 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
6989 2014-02-12 Richard Biener <rguenther@suse.de>
6991 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
6992 when we found a dependence.
6994 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
6996 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
6998 (maybe_fold_stmt): ... into this new function.
6999 * omp-low.c (lower_omp): Update comment.
7001 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
7004 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
7007 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
7009 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
7010 identifiers in comments.
7011 (cortexa53_extra_costs): Likewise.
7012 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
7013 (cortexa7_extra_costs): Likewise.
7014 (cortexa12_extra_costs): Likewise.
7015 (cortexa15_extra_costs): Likewise.
7016 (v7m_extra_costs): Likewise.
7018 2014-02-12 Richard Biener <rguenther@suse.de>
7021 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
7022 of posix_memalign being successful.
7023 (lower_stmt): Restrict lowering of posix_memalign to when
7024 -ftree-bit-ccp is enabled.
7026 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
7028 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
7030 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
7032 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
7034 PR rtl-optimization/60116
7035 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
7036 other_insn once the combination has been validated.
7038 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
7041 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
7043 * ipa-devirt.c: Include demangle.h
7044 (odr_violation_reported): New static variable.
7045 (add_type_duplicate): Update odr_violations.
7046 (maybe_record_node): Add completep parameter; update it.
7047 (record_target_from_binfo): Add COMPLETEP parameter;
7048 update it as needed.
7049 (possible_polymorphic_call_targets_1): Likewise.
7050 (struct polymorphic_call_target_d): Add nonconstruction_targets;
7051 rename FINAL to COMPLETE.
7052 (record_targets_from_bases): Sanity check we found the binfo;
7053 fix COMPLETEP updating.
7054 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
7055 parameter, fix computing of COMPLETEP.
7056 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
7057 at LTO time do demangling.
7058 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
7059 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
7061 (gimple_get_virt_method_for_binfo): Likewise.
7062 * gimple-fold.h (gimple_get_virt_method_for_binfo,
7063 gimple_get_virt_method_for_vtable): Update prototypes.
7065 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
7068 * genautomata.c (add_presence_absence): Fix typo with
7069 {final_}presence_list.
7071 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
7074 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
7075 for VSX/Altivec vectors that land in GPR registers.
7077 2014-02-11 Richard Henderson <rth@redhat.com>
7078 Jakub Jelinek <jakub@redhat.com>
7081 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
7082 around drhs if type conversion to lacc->type is not useless.
7084 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7086 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
7088 (cortex-a57.cortex-a53): Likewise.
7089 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
7091 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7093 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
7096 2014-02-11 Renlin Li <Renlin.Li@arm.com>
7098 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
7099 add_options_for_arm_vfp3.
7101 2014-02-11 Jeff Law <law@redhat.com>
7104 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
7105 object with an undesirable mode.
7107 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7110 * config/i386/sol2-9.h: New file.
7111 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
7112 *-*-solaris2.9*): Use it.
7114 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7116 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
7117 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
7119 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7121 * config/microblaze/microblaze.c: Extend mcpu version format
7123 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
7125 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
7127 2014-02-10 Richard Henderson <rth@redhat.com>
7130 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
7131 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
7132 ms-abi vs -mno-accumulate-outgoing-args.
7133 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
7134 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
7137 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
7140 * cfgexpand.c (expand_asm_operands): Attach source location to
7141 ASM_INPUT rtx objects.
7142 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
7144 2014-02-10 Nick Clifton <nickc@redhat.com>
7146 * config/mn10300/mn10300.c (popcount): New function.
7147 (mn10300_expand_prologue): Include saved registers in stack usage
7150 2014-02-10 Jeff Law <law@redhat.com>
7153 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
7154 when changing the SET_DEST of a prior insn to avoid an input reload.
7156 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7158 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
7159 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
7160 -mcall-openbsd, or -mcall-linux.
7161 (CC1_ENDIAN_BIG_SPEC): Remove.
7162 (CC1_ENDIAN_LITTLE_SPEC): Remove.
7163 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7164 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
7165 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
7166 and %cc1_endian_default.
7167 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7169 2014-02-10 Richard Biener <rguenther@suse.de>
7171 PR tree-optimization/60115
7172 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
7173 MEM_REF handling. Properly verify that the accesses are not
7174 out of the objects bound.
7176 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7178 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
7181 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
7183 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
7184 proper constants and fix formatting.
7185 (possible_polymorphic_call_targets): Fix formatting.
7187 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
7188 Ilya Tocar <ilya.tocar@intel.com>
7190 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
7191 (_mm512_loadu_epi32): Renamed into...
7192 (_mm512_loadu_si512): This.
7193 (_mm512_storeu_epi32): Renamed into...
7194 (_mm512_storeu_si512): This.
7195 (_mm512_maskz_ceil_ps): Removed.
7196 (_mm512_maskz_ceil_pd): Ditto.
7197 (_mm512_maskz_floor_ps): Ditto.
7198 (_mm512_maskz_floor_pd): Ditto.
7199 (_mm512_floor_round_ps): Ditto.
7200 (_mm512_floor_round_pd): Ditto.
7201 (_mm512_ceil_round_ps): Ditto.
7202 (_mm512_ceil_round_pd): Ditto.
7203 (_mm512_mask_floor_round_ps): Ditto.
7204 (_mm512_mask_floor_round_pd): Ditto.
7205 (_mm512_mask_ceil_round_ps): Ditto.
7206 (_mm512_mask_ceil_round_pd): Ditto.
7207 (_mm512_maskz_floor_round_ps): Ditto.
7208 (_mm512_maskz_floor_round_pd): Ditto.
7209 (_mm512_maskz_ceil_round_ps): Ditto.
7210 (_mm512_maskz_ceil_round_pd): Ditto.
7211 (_mm512_expand_pd): Ditto.
7212 (_mm512_expand_ps): Ditto.
7213 * config/i386/i386.c (ix86_builtins): Remove
7214 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
7215 (bdesc_args): Ditto.
7216 * config/i386/predicates.md (const1256_operand): New.
7217 (const_1_to_2_operand): Ditto.
7218 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7219 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7220 (*avx512pf_gatherpf<mode>sf): Ditto.
7221 (avx512pf_gatherpf<mode>df): Ditto.
7222 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7223 (*avx512pf_gatherpf<mode>df): Ditto.
7224 (avx512pf_scatterpf<mode>sf): Ditto.
7225 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7226 (*avx512pf_scatterpf<mode>sf): Ditto.
7227 (avx512pf_scatterpf<mode>df): Ditto.
7228 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7229 (*avx512pf_scatterpf<mode>df): Ditto.
7230 (avx512f_expand<mode>): Removed.
7231 (<shift_insn><mode>3<mask_name>): Change predicate type.
7233 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7235 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
7236 not at the end of datarefs vector use ordered_remove to avoid
7237 reordering datarefs vector.
7240 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
7241 mark local addressable non-static vars as GOVD_PRIVATE
7242 instead of GOVD_LOCAL.
7243 * omp-low.c (lower_omp_for): Move gimple_bind_vars
7244 and BLOCK_VARS of gimple_bind_block to new_stmt rather
7248 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
7249 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
7250 assume_aligned or alloc_align attributes.
7251 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
7252 arguments. Handle also assume_aligned and alloc_align attributes.
7253 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
7254 calls to functions with assume_aligned or alloc_align attributes.
7255 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
7257 2014-02-08 Terry Guo <terry.guo@arm.com>
7259 * doc/invoke.texi: Document ARM -march=armv7e-m.
7261 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7263 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
7264 flag on __cilkrts_rethrow builtin.
7267 * ipa-cp.c (determine_versionability): Fail at -O0
7268 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
7269 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
7272 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7275 * tree-inline.c (copy_forbidden): Fail for
7276 __attribute__((optimize (0))) functions.
7278 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7280 * varpool.c: Include pointer-set.h.
7281 (varpool_remove_unreferenced_decls): Variables in other partitions
7282 will not be output; be however careful to not lose information
7285 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7287 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
7288 lookup in the vtable constructor.
7290 2014-02-07 Jeff Law <law@redhat.com>
7293 * config/m68k/m68k.md (ashldi_extsi): Turn into a
7294 define_insn_and_split.
7296 * ipa-inline.c (inline_small_functions): Fix typos.
7298 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7300 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
7301 (s390_can_use_return_insn): Declare.
7302 * config/s390/s390.h (EPILOGUE_USES): Define.
7303 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
7305 (s390_chunkify_start): Handle return JUMP_LABELs.
7306 (s390_early_mach): Emit a main_pool instruction on the entry edge.
7307 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
7308 (s390_can_use_return_insn): New functions.
7309 (s390_fix_long_loop_prediction): Handle conditional returns.
7310 (TARGET_SET_UP_BY_PROLOGUE): Define.
7311 * config/s390/s390.md (ANY_RETURN): New code iterator.
7312 (*creturn, *csimple_return, return, simple_return): New patterns.
7314 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7316 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
7317 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
7318 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
7319 REG_CFA_RESTORE list when deciding not to restore a register.
7321 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7323 * config/s390/s390.c: Include tree-pass.h and context.h.
7324 (s390_early_mach): New function, split out from...
7325 (s390_emit_prologue): ...here.
7326 (pass_data_s390_early_mach): New pass structure.
7327 (pass_s390_early_mach): New class.
7328 (s390_option_override): Create and register early_mach pass.
7329 Move to end of file.
7331 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7333 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
7334 to match for the exit block.
7336 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7338 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
7339 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
7340 Reject misaligned operands.
7342 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7344 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
7346 2014-02-07 Richard Biener <rguenther@suse.de>
7349 * gimple-low.c (lower_builtin_posix_memalign): New function.
7350 (lower_stmt): Call it to lower posix_memalign in a way
7351 to make alignment info accessible.
7353 2014-02-07 Jakub Jelinek <jakub@redhat.com>
7356 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
7357 __builtin_setjmp_receiver.
7359 2014-02-07 Richard Biener <rguenther@suse.de>
7362 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
7363 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
7364 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7365 Handle BUILT_IN_POSIX_MEMALIGN.
7366 (find_func_clobbers): Likewise.
7367 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
7368 (call_may_clobber_ref_p_1): Likewise.
7370 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7373 * ipa-devirt.c (record_target_from_binfo): Remove overactive
7376 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7379 * lto-cgraph.c (lto_output_node): Use
7380 symtab_get_symbol_partitioning_class.
7381 (lto_output_varpool_node): likewise.
7382 (symtab_get_symbol_partitioning_class): Move here from
7384 * cgraph.h (symbol_partitioning_class): Likewise.
7385 (symtab_get_symbol_partitioning_class): Declare.
7387 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7389 * ggc.h (ggc_internal_cleared_alloc): New macro.
7390 * vec.h (vec_safe_copy): Handle memory stats.
7391 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
7392 * target-globals.c (save_target_globals): Likewise.
7394 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7397 * expr.c (emit_move_resolve_push): Export; be bit more selective
7398 on when to clear alias set.
7399 * expr.h (emit_move_resolve_push): Declare.
7400 * function.h (struct function): Add tail_call_marked.
7401 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
7402 * config/i386/i386-protos.h (ix86_expand_push): Remove.
7403 * config/i386/i386.md (TImode move expander): De not call
7405 (FP push expanders): Preserve memory attributes.
7406 * config/i386/sse.md (push<mode>1): Remove.
7407 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
7408 (ix86_expand_push): Remove.
7409 * config/i386/mmx.md (push<mode>1): Remove.
7411 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7413 PR rtl-optimization/60030
7414 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
7415 lopart with paradoxical subreg before shifting it up by hprec.
7417 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7419 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
7420 Remove extra newline at end of file.
7421 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
7422 (arm_issue_rate): Handle cortexa57.
7423 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
7424 (cortex-a57.cortex-a53): Likewise.
7426 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7429 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
7430 don't record in REG_FRAME_RELATED_EXPR registers not set in that
7432 (arm_expand_prologue): Adjust all callers.
7433 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
7434 info, registers also at the lowest numbered registers side. Use
7435 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
7439 * var-tracking.c (adjust_mems): Before adding a SET to
7440 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
7442 2014-02-06 Alan Modra <amodra@gmail.com>
7445 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
7446 change SDmode to DDmode when lra_in_progress.
7448 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7451 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
7452 free_data_ref on the dr first, and before goto again also set dr
7453 to the next dr. For simd_lane_access, free old datarefs[i] before
7454 overwriting it. For get_vectype_for_scalar_type failure, don't
7455 free_data_ref if simd_lane_access.
7457 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
7460 * tree.h (opts_for_fn): New inline function.
7461 (opt_for_fn): Define.
7462 * config/i386/i386.c (ix86_function_regparm): Use
7463 opt_for_fn (decl, optimize) instead of optimize.
7465 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
7467 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
7468 for SYMBOL_REF in large memory model.
7470 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7472 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
7474 (cortex-a57): Likewise.
7475 (cortex-a57.cortex-a53): Likewise.
7477 2014-02-06 Yury Gribov <y.gribov@samsung.com>
7478 Kugan Vivekanandarajah <kuganv@linaro.org>
7480 * config/arm/arm.c (arm_vector_alignment_reachable): Check
7482 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
7484 2014-02-06 Richard Biener <rguenther@suse.de>
7486 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
7487 set_loop_copy and initialize_original_copy_tables.
7489 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
7491 * config/aarch64/aarch64-simd.md
7492 (aarch64_ashr_simddi): Change QI to SI.
7494 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7495 Jakub Jelinek <jakub@redhat.com>
7498 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
7501 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7503 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
7504 CODE_FOR_altivec_vpku[hw]um to
7505 CODE_FOR_altivec_vpku[hw]um_direct.
7506 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
7507 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
7508 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
7509 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
7511 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7513 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
7514 generation for -maltivec=be.
7515 (altivec_vsumsws): Simplify redundant test.
7517 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7519 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
7520 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
7521 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
7522 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
7523 gen_altivec_vpkuwum.
7524 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
7526 (altivec_vpks<VI_char>ss): Likewise.
7527 (altivec_vpks<VI_char>us): Likewise.
7528 (altivec_vpku<VI_char>us): Likewise.
7529 (altivec_vpku<VI_char>um): Likewise.
7530 (altivec_vpku<VI_char>um_direct): New (copy of
7531 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
7533 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
7534 target is little endian and -maltivec=be is not specified.
7535 (*altivec_vupkhs<VU_char>_direct): New (copy of
7536 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
7537 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
7538 target is little endian and -maltivec=be is not specified.
7539 (*altivec_vupkls<VU_char>_direct): New (copy of
7540 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
7541 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
7542 little endian and -maltivec=be is not specified.
7543 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
7544 little endian and -maltivec=be is not specified.
7546 2014-02-05 Richard Henderson <rth@redhat.com>
7549 * combine-stack-adj.c: Revert r206943.
7550 * sched-int.h (struct deps_desc): Add last_args_size.
7551 * sched-deps.c (init_deps): Initialize it.
7552 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7553 contain REG_ARGS_SIZE notes.
7555 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7557 * lto-cgraph.c (asm_nodes_output): Make global.
7558 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7559 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7560 (driver_handle_option): Handle OPT_fwpa.
7562 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7565 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7566 a comment typo and formatting issue. If odr_hash hasn't been
7567 created, return vNULL and set *completep to false.
7570 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7571 bb with no successors.
7573 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7576 * doc/invoke.texi (-march): Clarify documentation for ARM.
7580 2014-02-05 Richard Biener <rguenther@suse.de>
7582 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
7583 when not vectorizing because of too many alias checks.
7584 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
7585 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
7587 2014-02-05 Nick Clifton <nickc@redhat.com>
7589 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
7590 accept extended registers in any mode when compiling for the MN10300.
7592 2014-02-05 Yury Gribov <y.gribov@samsung.com>
7594 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
7595 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
7596 sanitization attributes.
7597 (can_inline_edge_p): Likewise.
7598 (sanitize_attrs_match_for_inline_p): New function.
7600 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7602 * ipa-prop.c (detect_type_change): Shor circuit testing of
7603 type changes on THIS pointer.
7605 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
7608 * config/pa/pa.c (legitimize_tls_address): Return original address
7609 if not passed a SYMBOL_REF rtx.
7610 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
7612 (pa_emit_move_sequence): Simplify TLS source operands.
7613 (pa_legitimate_constant_p): Reject all TLS constants.
7614 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
7615 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
7617 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7619 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
7620 groups when we know they are controlled by LTO.
7621 * varasm.c (default_binds_local_p_1): If object is in other partition,
7622 it will be resolved locally.
7624 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7626 * config/host-linux.c (linux_gt_pch_use_address): Don't
7627 use SSIZE_MAX because it is not always defined.
7629 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
7632 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
7633 threshold for pseudo splitting.
7634 (update_ebb_live_info): Process call argument hard registers and
7635 hard registers from insn definition too.
7636 (max_small_class_regs_num): New constant.
7637 (inherit_in_ebb): Update live hard regs through EBBs. Update
7638 reloads_num only for small register classes. Don't split for
7641 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
7644 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
7647 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7649 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
7652 2014-02-04 Marek Polacek <polacek@redhat.com>
7654 * gdbinit.in (pel): Define.
7656 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7658 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
7661 2014-02-04 Richard Biener <rguenther@suse.de>
7664 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
7665 in function context local.
7666 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
7667 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
7668 similar to LTO_imported_decl_ref.
7670 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7672 PR tree-optimization/60002
7673 * cgraphclones.c (build_function_decl_skip_args): Clear
7676 PR tree-optimization/60023
7677 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
7678 false to gsi_replace.
7679 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
7680 has been in some EH region and vec_stmt could throw, add
7681 vec_stmt into the same EH region.
7682 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
7683 has no lhs, ignore it.
7684 * internal-fn.c (expand_MASK_LOAD): Likewise.
7687 * tree-inline.c (copy_forbidden): Fail for
7688 __attribute__((optimize (0))) functions.
7691 * omp-low.c (simd_clone_struct_copy): If from->inbranch
7692 is set, copy one less argument.
7693 (expand_simd_clones): Don't subtract clone_info->inbranch
7694 from simd_clone_struct_alloc argument.
7696 PR rtl-optimization/57915
7697 * recog.c (simplify_while_replacing): If all unary/binary/relational
7698 operation arguments are constant, attempt to simplify those.
7701 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
7702 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
7704 2014-02-04 Richard Biener <rguenther@suse.de>
7706 PR tree-optimization/60012
7707 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
7708 TBAA disambiguation to all DDRs.
7710 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7713 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
7714 (LINK_SPEC): Use it for -shared, -shared-libgcc.
7716 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7719 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
7721 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7723 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
7724 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
7726 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7729 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
7730 to figure out targets of polymorphic calls with known decl.
7731 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7732 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
7733 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
7734 (get_polymorphic_call_info): ... here.
7735 (get_polymorphic_call_info_from_invariant): New function.
7737 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7739 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
7740 lookup via vtable pointer; check for type consistency
7741 and turn inconsitent facts into UNREACHABLE.
7742 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7743 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
7744 type inconsistent querries; return UNREACHABLE instead.
7746 2014-02-03 Richard Henderson <rth@twiddle.net>
7749 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
7750 already processed this node.
7751 (normalize_one_pred_1): Pass along mark_set.
7752 (normalize_one_pred): Create and destroy a pointer_set_t.
7753 (normalize_one_pred_chain): Likewise.
7755 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
7757 PR gcov-profile/58602
7758 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
7760 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7763 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
7764 -fno-devirtualize; try to devirtualize by the knowledge of
7765 virtual table pointer given by aggregate propagation.
7766 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7767 (ipa_print_node_jump_functions): Dump also offset that
7768 is relevant for polymorphic calls.
7769 (determine_known_aggregate_parts): Add arg_type parameter; use it
7770 instead of determining the type from pointer type.
7771 (ipa_compute_jump_functions_for_edge): Update call of
7772 determine_known_aggregate_parts.
7773 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
7774 (gimple_get_virt_method_for_binfo): ... here; simplify using
7775 vtable_pointer_value_to_vtable.
7776 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
7777 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
7778 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
7779 (vtable_pointer_value_to_vtable): Break out from ...; handle also
7781 (vtable_pointer_value_to_binfo): ... here.
7782 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
7784 2014-02-03 Teresa Johnson <tejohnson@google.com>
7786 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
7787 redef of outer loop index variable.
7789 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
7793 * doc/extend.texi (Function Attributes): Typo.
7795 2014-02-03 Cong Hou <congh@google.com>
7797 PR tree-optimization/60000
7798 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
7799 if the vectorized statement is a store. A store statement can only
7800 appear at the end of pattern statements.
7802 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7804 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
7805 (ix86_option_override_internal): Default long double to 64-bit for
7806 32-bit Bionic and to 128-bit for 64-bit Bionic.
7808 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
7809 TARGET_LONG_DOUBLE_128 is true.
7810 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
7812 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
7813 (mlong-double-64): Negate -mlong-double-128.
7814 (mlong-double-128): New option.
7816 * config/i386/i386-c.c (ix86_target_macros): Define
7817 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
7819 * doc/invoke.texi: Document -mlong-double-128.
7821 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7823 PR rtl-optimization/60024
7824 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
7826 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
7828 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
7830 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7832 PR rtl-optimization/57662
7833 * sel-sched.c (code_motion_path_driver): Do not mark already not
7834 existing blocks in the visiting bitmap.
7836 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7838 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
7839 on the insn being emitted.
7841 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
7842 Will Deacon <will.deacon@arm.com>
7844 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
7846 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7848 * config/arm/arm-tables.opt: Regenerate.
7850 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7852 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
7853 for vector types other than V16QImode.
7854 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
7855 define_expand, and call altivec_expand_vec_perm_le when producing
7856 code with little endian element order.
7857 (*altivec_vperm_<mode>_internal): New insn having previous
7858 behavior of altivec_vperm_<mode>.
7859 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
7860 altivec_expand_vec_perm_le when producing code with little endian
7862 (*altivec_vperm_<mode>_uns_internal): New insn having previous
7863 behavior of altivec_vperm_<mode>_uns.
7865 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7867 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
7868 (altivec_vsumsws): Add handling for -maltivec=be with a little
7870 (altivec_vsumsws_direct): New.
7871 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
7872 gen_altivec_vsumsws.
7874 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7876 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
7877 vtable_pointer_value_to_binfo): New functions.
7878 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
7879 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
7881 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
7883 * config/nios2/nios2.md (load_got_register): Initialize GOT
7884 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
7885 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
7887 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7889 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
7890 preserverd by passthrough, do not propagate the type.
7892 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7894 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
7895 (mips_atomic_assign_expand_fenv): New function.
7896 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
7898 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7900 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
7901 (__builtin_mips_set_fcsr): Likewise.
7902 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
7903 MIPS_USI_FTYPE_VOID.
7904 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
7905 (mips16_expand_set_fcsr): Likewise.
7906 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
7907 (mips16_set_fcsr_stub): Likewise.
7908 (mips16_get_fcsr_one_only_stub): New class.
7909 (mips16_set_fcsr_one_only_stub): Likewise.
7910 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
7911 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
7912 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
7913 (hard_float): New availability predicate.
7914 (mips_builtins): Add get_fcsr and set_fcsr.
7915 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
7916 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
7917 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
7918 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
7919 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
7922 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7924 * config/mips/mips.c (mips_one_only_stub): New class.
7925 (mips_need_mips16_rdhwr_p): Replace with...
7926 (mips16_rdhwr_stub): ...this new variable.
7927 (mips16_stub_call_address): New function.
7928 (mips16_rdhwr_one_only_stub): New class.
7929 (mips_expand_thread_pointer): Use mips16_stub_call_address.
7930 (mips_output_mips16_rdhwr): Delete.
7931 (mips_finish_stub): New function.
7932 (mips_code_end): Use it to handle rdhwr stubs.
7934 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
7937 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
7938 when calculating size of integer atomic types.
7940 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
7942 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
7944 2014-02-01 Jakub Jelinek <jakub@redhat.com>
7946 PR tree-optimization/60003
7947 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
7948 * profile.c (branch_prob): Use gimple_call_builtin_p
7949 to check for BUILT_IN_SETJMP_RECEIVER.
7950 * tree-inline.c (copy_bb): Call notice_special_calls.
7952 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
7955 * lra-constraints.c (process_alt_operands): Update reload_sum only
7958 2014-01-31 Richard Henderson <rth@redhat.com>
7961 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
7962 until after else_eh is processed.
7964 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7966 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
7967 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
7968 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
7969 in smmintrin.h, remove them.
7970 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
7971 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
7972 * config/i386/i386.md (ROUND_SAE): Fix value.
7973 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
7974 (const48_operand): New.
7975 * config/i386/subst.md (round), (round_expand): Use
7976 const_4_or_8_to_11_operand.
7977 (round_saeonly), (round_saeonly_expand): Use const48_operand.
7979 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7981 * config/i386/constraints.md (Yk): Swap meaning with k.
7982 * config/i386/i386.md (movhi_internal): Change Yk to k.
7983 (movqi_internal): Ditto.
7984 (*k<logic><mode>): Ditto.
7987 (kandn<mode>): Ditto.
7988 (*<code>hi_1): Ditto.
7989 (*<code>qi_1): Ditto.
7990 (kxnor<mode>): Ditto.
7991 (kortestzhi): Ditto.
7992 (kortestchi): Ditto.
7994 (*one_cmplhi2_1): Ditto.
7995 (*one_cmplqi2_1): Ditto.
7996 * config/i386/sse.md (): Change k to Yk.
7997 (avx512f_load<mode>_mask): Ditto.
7998 (avx512f_blendm<mode>): Ditto.
7999 (avx512f_store<mode>_mask): Ditto.
8000 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
8001 (avx512f_storedqu<mode>_mask): Ditto.
8002 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
8004 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
8005 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
8006 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
8007 (avx512f_maskcmp<mode>3): Ditto.
8008 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
8009 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
8010 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
8011 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
8012 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
8013 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
8014 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
8015 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
8016 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
8017 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
8018 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
8019 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
8020 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
8021 (vec_extract_lo_<mode>_maskm): Ditto.
8022 (vec_extract_hi_<mode>_maskm): Ditto.
8023 (avx512f_vternlog<mode>_mask): Ditto.
8024 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
8025 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
8026 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
8027 (avx512f_<code>v8div16qi2_mask): Ditto.
8028 (avx512f_<code>v8div16qi2_mask_store): Ditto.
8029 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
8030 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
8031 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
8032 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
8033 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8034 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8035 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8036 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8037 (avx512cd_maskb_vec_dupv8di): Ditto.
8038 (avx512cd_maskw_vec_dupv16si): Ditto.
8039 (avx512f_vpermi2var<mode>3_maskz): Ditto.
8040 (avx512f_vpermi2var<mode>3_mask): Ditto.
8041 (avx512f_vpermi2var<mode>3_mask): Ditto.
8042 (avx512f_vpermt2var<mode>3_maskz): Ditto.
8043 (*avx512f_gathersi<mode>): Ditto.
8044 (*avx512f_gathersi<mode>_2): Ditto.
8045 (*avx512f_gatherdi<mode>): Ditto.
8046 (*avx512f_gatherdi<mode>_2): Ditto.
8047 (*avx512f_scattersi<mode>): Ditto.
8048 (*avx512f_scatterdi<mode>): Ditto.
8049 (avx512f_compress<mode>_mask): Ditto.
8050 (avx512f_compressstore<mode>_mask): Ditto.
8051 (avx512f_expand<mode>_mask): Ditto.
8052 * config/i386/subst.md (mask): Change k to Yk.
8053 (mask_scalar_merge): Ditto.
8056 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
8058 * doc/extend.texi (Vector Extensions): Document ?: in C++.
8060 2014-01-31 Richard Biener <rguenther@suse.de>
8063 * builtins.c (fold_builtin_memory_op): Make sure to not
8064 use a floating-point mode or a boolean or enumeral type for
8067 2014-01-30 DJ Delorie <dj@redhat.com>
8069 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
8070 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
8071 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
8072 whenever main() has an epilogue.
8074 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8076 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
8077 unused variable "field".
8078 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
8079 (vsx_mergeh_<mode>): Likewise.
8080 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
8081 (altivec_vmrghh): Likewise.
8082 (altivec_vmrghw): Likewise.
8083 (altivec_vmrglb): Likewise.
8084 (altivec_vmrglh): Likewise.
8085 (altivec_vmrglw): Likewise.
8086 (altivec_vspltb): Add missing uses.
8087 (altivec_vsplth): Likewise.
8088 (altivec_vspltw): Likewise.
8089 (altivec_vspltsf): Likewise.
8091 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8094 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
8095 frame related instructions.
8097 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
8099 PR rtl-optimization/59959
8100 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
8101 any reload of register whose subreg is invalid.
8103 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8105 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
8106 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
8107 Add missing return type - void.
8109 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8111 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
8112 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
8113 remove element index adjustment for endian (now handled in vsx.md
8115 (altivec_expand_vec_perm_const): Use
8116 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
8117 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
8118 (vsx_xxspltw_<mode>): Adjust element index for little endian.
8119 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
8120 define_expand and a new define_insn *altivec_vspltb_internal;
8121 adjust for -maltivec=be on a little endian target.
8122 (altivec_vspltb_direct): New.
8123 (altivec_vsplth): Divide into a define_expand and a new
8124 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
8125 little endian target.
8126 (altivec_vsplth_direct): New.
8127 (altivec_vspltw): Divide into a define_expand and a new
8128 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
8129 little endian target.
8130 (altivec_vspltw_direct): New.
8131 (altivec_vspltsf): Divide into a define_expand and a new
8132 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
8133 a little endian target.
8135 2014-01-30 Richard Biener <rguenther@suse.de>
8137 PR tree-optimization/59993
8138 * tree-ssa-forwprop.c (associate_pointerplus): Check we
8139 can propagate form the earlier stmt and avoid the transform
8140 when the intermediate result is needed.
8142 2014-01-30 Alangi Derick <alangiderick@gmail.com>
8144 * README.Portability: Fix typo.
8146 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
8148 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
8149 comparison_operator with ordered_comparison_operator.
8151 2014-01-30 Nick Clifton <nickc@redhat.com>
8153 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
8154 Rename to mn10300_store_multiple_regs.
8155 * config/mn10300/mn10300.c: Likewise.
8156 * config/mn10300/mn10300.md (store_movm): Fix typo: call
8157 store_multiple_regs.
8158 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
8159 Call mn10300_store_multiple_regs.
8161 2014-01-30 Nick Clifton <nickc@redhat.com>
8162 DJ Delorie <dj@redhat.com>
8164 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
8165 %fp 2 to keep registers after it properly word-aligned.
8166 (rl78_alloc_physical_registers_umul): Handle the case where both
8167 input operands are the same.
8169 2014-01-30 Richard Biener <rguenther@suse.de>
8171 PR tree-optimization/59903
8172 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
8175 2014-01-30 Jason Merrill <jason@redhat.com>
8178 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
8181 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
8183 2014-01-30 Richard Biener <rguenther@suse.de>
8185 PR tree-optimization/59951
8186 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
8188 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
8191 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
8192 SFmode to DFmode case.
8194 2014-01-29 DJ Delorie <dj@redhat.com>
8196 * config/msp430/msp430.opt (-minrt): New.
8197 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
8199 (ENDFILE_SPEC): Likewise.
8201 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
8203 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
8204 (estimate_function_body_sizes): Use it.
8206 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
8209 * dwarf2out.c (is_cxx_auto): New.
8210 (is_base_type): Use it.
8211 (gen_type_die_with_usage): Likewise.
8213 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8215 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
8216 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
8217 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
8218 -maltivec=be with LE targets.
8219 (vsx_mergeh_<mode>): Likewise.
8220 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
8221 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
8222 (altivec_vmrghb): Replace with define_expand and new
8223 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
8224 (altivec_vmrghb_direct): New define_insn.
8225 (altivec_vmrghh): Replace with define_expand and new
8226 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
8227 (altivec_vmrghh_direct): New define_insn.
8228 (altivec_vmrghw): Replace with define_expand and new
8229 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
8230 (altivec_vmrghw_direct): New define_insn.
8231 (*altivec_vmrghsf): Adjust for endianness.
8232 (altivec_vmrglb): Replace with define_expand and new
8233 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
8234 (altivec_vmrglb_direct): New define_insn.
8235 (altivec_vmrglh): Replace with define_expand and new
8236 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
8237 (altivec_vmrglh_direct): New define_insn.
8238 (altivec_vmrglw): Replace with define_expand and new
8239 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
8240 (altivec_vmrglw_direct): New define_insn.
8241 (*altivec_vmrglsf): Adjust for endianness.
8242 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8243 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8244 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8245 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8246 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8247 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8248 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8249 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8251 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
8253 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
8254 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
8257 2014-01-29 Richard Biener <rguenther@suse.de>
8259 PR tree-optimization/58742
8260 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
8261 associate_pointerplus_align.
8262 (associate_pointerplus_diff): New function.
8263 (associate_pointerplus): Likewise. Call associate_pointerplus_align
8264 and associate_pointerplus_diff.
8266 2014-01-29 Richard Biener <rguenther@suse.de>
8268 * lto-streamer.h (LTO_major_version): Bump to 3.
8269 (LTO_minor_version): Reset to 0.
8271 2014-01-29 Renlin Li <Renlin.Li@arm.com>
8273 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
8274 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
8275 (arm_file_start): Generate correct asm header for armv7ve.
8276 * config/arm/bpabi.h: Add multilib support for armv7ve.
8277 * config/arm/driver-arm.c: Change the architectures of cortex-a7
8278 and cortex-a15 to armv7ve.
8279 * config/arm/t-aprofile: Add multilib support for armv7ve.
8280 * doc/invoke.texi: Document -march=armv7ve.
8282 2014-01-29 Richard Biener <rguenther@suse.de>
8284 PR tree-optimization/58742
8285 * tree-ssa-forwprop.c (associate_plusminus): Return true
8286 if we changed sth, defer EH cleanup to ...
8287 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
8288 (simplify_mult): New function.
8290 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8293 PR tree-optimization/59920
8294 * tree.c (build_common_builtin_nodes): Remove
8295 __builtin_setjmp_dispatcher initialization.
8296 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
8297 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
8298 instead of gsi_after_labels + manually skipping debug stmts.
8299 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
8300 ignore bbs with IFN_ABNORMAL_DISPATCHER.
8301 * tree-inline.c (copy_edges_for_bb): Remove
8302 can_make_abnormal_goto argument, instead add abnormal_goto_dest
8303 argument. Ignore computed_goto_p stmts. Don't call
8304 make_abnormal_goto_edges. If a call might need abnormal edges
8305 for non-local gotos, see if it already has an edge to
8306 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
8307 with true argument, don't do anything then, otherwise add
8308 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
8309 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
8311 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
8312 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
8313 (lower_stmt): Don't set data->calls_builtin_setjmp.
8314 (lower_builtin_setjmp): Adjust comment.
8315 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
8316 * tree-cfg.c (found_computed_goto): Remove.
8317 (factor_computed_gotos): Remove.
8318 (make_goto_expr_edges): Return bool, true for computed gotos.
8319 Don't call make_abnormal_goto_edges.
8320 (build_gimple_cfg): Don't set found_computed_goto, don't call
8321 factor_computed_gotos.
8322 (computed_goto_p): No longer static.
8323 (make_blocks): Don't set found_computed_goto.
8324 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
8325 (make_edges): If make_goto_expr_edges returns true, push bb
8326 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
8327 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
8328 vector. Record mapping between bbs and OpenMP regions if there
8329 are any, adjust make_gimple_omp_edges caller. Call
8330 handle_abnormal_edges.
8331 (make_abnormal_goto_edges): Remove.
8332 * tree-cfg.h (make_abnormal_goto_edges): Remove.
8333 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
8334 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
8335 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
8336 * internal-fn.def (ABNORMAL_DISPATCHER): New.
8337 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
8338 filling *region also set *region_idx to (*region)->entry->index.
8341 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
8342 For REGs set ORIGINAL_REGNO.
8344 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
8346 * doc/md.texi: Mention that a target shouldn't implement
8347 vec_widen_(s|u)mul_even/odd pair if it is less efficient
8350 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8352 PR tree-optimization/59594
8353 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
8354 a copy of the datarefs vector rather than the vector itself.
8356 2014-01-28 Jason Merrill <jason@redhat.com>
8359 * dwarf2out.c (auto_die): New static.
8360 (gen_type_die_with_usage): Handle C++1y 'auto'.
8361 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
8364 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
8367 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
8368 (SPEC_X32): Likewise.
8369 (SPEC_64): Likewise.
8370 * config/i386/i386.c (ix86_option_override_internal): Turn off
8371 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
8373 (x86_file_start): Output .code16gcc for TARGET_16BIT.
8374 * config/i386/i386.h (TARGET_16BIT): New macro.
8375 (TARGET_16BIT_P): Likewise.
8376 * config/i386/i386.opt: Add m16.
8377 * doc/invoke.texi: Document -m16.
8379 2014-01-28 Jakub Jelinek <jakub@redhat.com>
8381 PR preprocessor/59935
8382 * input.c (location_get_source_line): Bail out on when line number
8383 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
8385 2014-01-28 Richard Biener <rguenther@suse.de>
8387 PR tree-optimization/58742
8388 * tree-ssa-forwprop.c (associate_plusminus): Handle
8389 pointer subtraction of the form (T)(P + A) - (T)P.
8391 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8393 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
8396 2014-01-28 Richard Biener <rguenther@suse.de>
8399 2014-01-28 Richard Biener <rguenther@suse.de>
8401 PR rtl-optimization/45364
8402 PR rtl-optimization/59890
8403 * var-tracking.c (local_get_addr_clear_given_value): Handle
8404 already cleared slot.
8405 (val_reset): Handle not allocated local_get_addr_cache.
8406 (vt_find_locations): Use post-order on the inverted CFG.
8408 2014-01-28 Richard Biener <rguenther@suse.de>
8410 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
8412 2014-01-28 Richard Biener <rguenther@suse.de>
8414 PR rtl-optimization/45364
8415 PR rtl-optimization/59890
8416 * var-tracking.c (local_get_addr_clear_given_value): Handle
8417 already cleared slot.
8418 (val_reset): Handle not allocated local_get_addr_cache.
8419 (vt_find_locations): Use post-order on the inverted CFG.
8421 2014-01-28 Alan Modra <amodra@gmail.com>
8423 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
8424 * configure.ac <recursive call for build != host>: Define
8425 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
8426 and LD_FOR_BUILD too.
8427 * configure: Regenerate.
8429 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
8431 * config/i386/i386.c (get_builtin_code_for_version): Separate
8432 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
8433 Broadwell from Haswell.
8435 2014-01-27 Steve Ellcey <sellcey@mips.com>
8437 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
8438 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
8439 * config/mips/mips.c (mips_option_override): Change setting
8441 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
8442 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
8443 Change from Mask to Var.
8445 2014-01-27 Jeff Law <law@redhat.com>
8447 * ipa-inline.c (inline_small_functions): Fix typo.
8449 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
8451 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
8452 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
8453 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
8454 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
8455 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
8456 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
8457 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
8458 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
8459 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
8460 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
8461 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
8462 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
8463 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
8464 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
8465 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
8466 (_mm512_storeu_epi64): Ditto.
8467 (_mm512_cmpge_epi32_mask): Ditto.
8468 (_mm512_cmpge_epu32_mask): Ditto.
8469 (_mm512_cmpge_epi64_mask): Ditto.
8470 (_mm512_cmpge_epu64_mask): Ditto.
8471 (_mm512_cmple_epi32_mask): Ditto.
8472 (_mm512_cmple_epu32_mask): Ditto.
8473 (_mm512_cmple_epi64_mask): Ditto.
8474 (_mm512_cmple_epu64_mask): Ditto.
8475 (_mm512_cmplt_epi32_mask): Ditto.
8476 (_mm512_cmplt_epu32_mask): Ditto.
8477 (_mm512_cmplt_epi64_mask): Ditto.
8478 (_mm512_cmplt_epu64_mask): Ditto.
8479 (_mm512_cmpneq_epi32_mask): Ditto.
8480 (_mm512_cmpneq_epu32_mask): Ditto.
8481 (_mm512_cmpneq_epi64_mask): Ditto.
8482 (_mm512_cmpneq_epu64_mask): Ditto.
8483 (_mm512_expand_pd): Ditto.
8484 (_mm512_expand_ps): Ditto.
8485 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
8486 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
8487 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
8488 * config/i386/i386.c (ix86_builtins): Add
8489 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
8490 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
8491 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
8492 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
8493 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
8494 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
8495 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
8496 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
8497 IX86_BUILTIN_PMOVUSQW512_MEM.
8498 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
8499 __builtin_ia32_pmovsqd512mem_mask,
8500 __builtin_ia32_pmovqd512mem_mask,
8501 __builtin_ia32_pmovusqw512mem_mask,
8502 __builtin_ia32_pmovsqw512mem_mask,
8503 __builtin_ia32_pmovqw512mem_mask,
8504 __builtin_ia32_pmovusdw512mem_mask,
8505 __builtin_ia32_pmovsdw512mem_mask,
8506 __builtin_ia32_pmovdw512mem_mask,
8507 __builtin_ia32_pmovqb512mem_mask,
8508 __builtin_ia32_pmovusqb512mem_mask,
8509 __builtin_ia32_pmovsqb512mem_mask,
8510 __builtin_ia32_pmovusdb512mem_mask,
8511 __builtin_ia32_pmovsdb512mem_mask,
8512 __builtin_ia32_pmovdb512mem_mask.
8513 (bdesc_args): Add __builtin_ia32_expanddf512,
8514 __builtin_ia32_expandsf512.
8515 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
8516 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
8517 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
8518 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
8519 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
8520 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
8521 (avx512f_<code>v8div16qi2_mask_store): This.
8522 (avx512f_expand<mode>): New.
8524 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
8526 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
8528 (_mm512_mask_prefetch_i64gather_pd): Ditto.
8529 (_mm512_prefetch_i32scatter_pd): Ditto.
8530 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
8531 (_mm512_prefetch_i64scatter_pd): Ditto.
8532 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
8533 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
8534 (_mm512_mask_prefetch_i64gather_ps): Ditto.
8535 (_mm512_prefetch_i32scatter_ps): Ditto.
8536 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
8537 (_mm512_prefetch_i64scatter_ps): Ditto.
8538 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
8539 * config/i386/i386-builtin-types.def: Define
8540 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
8541 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
8542 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
8543 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
8544 IX86_BUILTIN_SCATTERPFQPD.
8545 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
8546 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
8547 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
8548 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
8549 __builtin_ia32_scatterpfqps.
8550 (ix86_expand_builtin): Expand new built-ins.
8551 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8552 fix memory access data type.
8553 (*avx512pf_gatherpf<mode>_mask): Ditto.
8554 (*avx512pf_gatherpf<mode>): Ditto.
8555 (avx512pf_scatterpf<mode>): Ditto.
8556 (*avx512pf_scatterpf<mode>_mask): Ditto.
8557 (*avx512pf_scatterpf<mode>): Ditto.
8558 (GATHER_SCATTER_SF_MEM_MODE): New.
8559 (avx512pf_gatherpf<mode>df): Ditto.
8560 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8561 (*avx512pf_scatterpf<mode>df): Ditto.
8563 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8566 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8567 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8570 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8572 * common/config/arm/arm-common.c
8573 (arm_rewrite_mcpu): Handle multiple names.
8575 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8577 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8579 * gimple-builder.h (create_gimple_tmp): Delete.
8581 2014-01-27 Christian Bruel <christian.bruel@st.com>
8583 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
8586 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
8588 * config/pa/pa.md (call): Generate indirect long calls to non-local
8589 functions when outputing 32-bit code.
8590 (call_value): Likewise except for special call to buggy powf function.
8592 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
8593 portable runtime and PIC indirect calls.
8594 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
8595 and PIC call sequences. Use ldo instead of blr to set return register
8596 in PIC call sequence.
8598 2014-01-25 Walter Lee <walt@tilera.com>
8600 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
8601 avoid clobbering a live register.
8603 2014-01-25 Walter Lee <walt@tilera.com>
8605 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
8606 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
8607 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
8608 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
8610 2014-01-25 Walter Lee <walt@tilera.com>
8612 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
8613 arguments on even registers.
8614 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
8616 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
8617 (BIGGEST_ALIGNMENT): Ditto.
8618 (BIGGEST_FIELD_ALIGNMENT): Ditto.
8620 2014-01-25 Walter Lee <walt@tilera.com>
8622 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
8623 insns before bundling.
8624 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8626 2014-01-25 Walter Lee <walt@tilera.com>
8628 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
8629 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
8630 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
8632 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8634 * config/mips/constraints.md (kl): Delete.
8635 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
8636 define expands, using...
8637 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
8638 instructions for MIPS16.
8639 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
8640 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
8642 2014-01-25 Walter Lee <walt@tilera.com>
8644 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
8648 2014-01-25 Walter Lee <walt@tilera.com>
8650 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
8651 (TARGET_EXPAND_TO_RTL_HOOK): Define.
8653 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8655 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
8658 2014-01-25 Jakub Jelinek <jakub@redhat.com>
8660 * print-rtl.c (in_call_function_usage): New var.
8661 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
8662 EXPR_LIST mode as mode and not as reg note name.
8665 * cfgloopmanip.c (copy_loop_info): If
8666 loop->warned_aggressive_loop_optimizations, make sure
8667 the flag is set in target loop too.
8669 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
8671 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
8673 * builtins.def: Likewise.
8674 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
8675 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
8676 * ira.c (ira_setup_eliminable_regset): Likewise.
8677 * omp-low.c (gate_expand_omp): Likewise.
8678 (execute_lower_omp): Likewise.
8679 (diagnose_sb_0): Likewise.
8680 (gate_diagnose_omp_blocks): Likewise.
8681 (simd_clone_clauses_extract): Likewise.
8684 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8686 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
8687 correction for little endian...
8688 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
8691 2014-01-24 Jeff Law <law@redhat.com>
8693 PR tree-optimization/59919
8694 * tree-vrp.c (find_assert_locations_1): Do not register asserts
8695 for non-returning calls.
8697 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
8699 * common/config/aarch64/aarch64-common.c
8700 (aarch64_rewrite_mcpu): Handle multiple names.
8701 * config/aarch64/aarch64.h
8702 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8704 2014-01-24 Dodji Seketeli <dodji@redhat.com>
8706 * input.c (add_file_to_cache_tab): Handle the case where fopen
8709 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
8712 * config/i386/i386.md (pushsf splitter): Get stack adjustment
8713 from push operand if code of push isn't PRE_DEC.
8715 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8718 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
8719 -mquad-memory-atomic. Update -mquad-memory documentation to say
8720 it is only used for non-atomic loads/stores.
8722 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
8723 -mquad-memory or -mquad-memory-atomic switches.
8725 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
8726 -mquad-memory-atomic to ISA 2.07 support.
8728 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
8729 to separate support of normal quad word memory operations (ldq, stq)
8730 from the atomic quad word memory operations.
8732 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
8733 support to separate non-atomic quad word operations from atomic
8734 quad word operations. Disable non-atomic quad word operations in
8735 little endian mode so that we don't have to swap words after the
8736 load and before the store.
8737 (quad_load_store_p): Add comment about atomic quad word support.
8738 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
8739 options printed with -mdebug=reg.
8741 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
8742 -mquad-memory-atomic as the test for whether we have quad word
8743 atomic instructions.
8744 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
8745 or -mp8-vector are used, allow byte/half-word atomic operations.
8747 * config/rs6000/sync.md (load_lockedti): Insure that the address
8748 is a proper indexed or indirect address for the lqarx instruction.
8749 On little endian systems, swap the hi/lo registers after the lqarx
8751 (load_lockedpti): Use indexed_or_indirect_operand predicate to
8752 insure the address is valid for the lqarx instruction.
8753 (store_conditionalti): Insure that the address is a proper indexed
8754 or indirect address for the stqcrx. instruction. On little endian
8755 systems, swap the hi/lo registers before doing the stqcrx.
8757 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
8758 insure the address is valid for the stqcrx. instruction.
8760 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8761 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
8762 type of quad memory support is available.
8764 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
8767 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
8768 there is a danger of looping.
8770 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
8772 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8773 force flag_ira_loop_pressure if set via command line.
8775 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8777 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
8778 (ashr_simd): New builtin handling DI mode.
8779 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
8780 (aarch64_sshr_simddi): New match pattern.
8781 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
8782 (vshrd_n_s64): Likewise.
8783 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
8785 2014-01-23 Nick Clifton <nickc@redhat.com>
8787 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
8788 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
8789 favour of mcu specific scripts.
8790 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
8793 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8794 Alex Velenko <Alex.Velenko@arm.com>
8796 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
8797 (vaddv_s16): Likewise.
8798 (vaddv_s32): Likewise.
8799 (vaddv_u8): Likewise.
8800 (vaddv_u16): Likewise.
8801 (vaddv_u32): Likewise.
8802 (vaddvq_s8): Likewise.
8803 (vaddvq_s16): Likewise.
8804 (vaddvq_s32): Likewise.
8805 (vaddvq_s64): Likewise.
8806 (vaddvq_u8): Likewise.
8807 (vaddvq_u16): Likewise.
8808 (vaddvq_u32): Likewise.
8809 (vaddvq_u64): Likewise.
8810 (vaddv_f32): Likewise.
8811 (vaddvq_f32): Likewise.
8812 (vaddvq_f64): Likewise.
8813 (vmaxv_f32): Likewise.
8814 (vmaxv_s8): Likewise.
8815 (vmaxv_s16): Likewise.
8816 (vmaxv_s32): Likewise.
8817 (vmaxv_u8): Likewise.
8818 (vmaxv_u16): Likewise.
8819 (vmaxv_u32): Likewise.
8820 (vmaxvq_f32): Likewise.
8821 (vmaxvq_f64): Likewise.
8822 (vmaxvq_s8): Likewise.
8823 (vmaxvq_s16): Likewise.
8824 (vmaxvq_s32): Likewise.
8825 (vmaxvq_u8): Likewise.
8826 (vmaxvq_u16): Likewise.
8827 (vmaxvq_u32): Likewise.
8828 (vmaxnmv_f32): Likewise.
8829 (vmaxnmvq_f32): Likewise.
8830 (vmaxnmvq_f64): Likewise.
8831 (vminv_f32): Likewise.
8832 (vminv_s8): Likewise.
8833 (vminv_s16): Likewise.
8834 (vminv_s32): Likewise.
8835 (vminv_u8): Likewise.
8836 (vminv_u16): Likewise.
8837 (vminv_u32): Likewise.
8838 (vminvq_f32): Likewise.
8839 (vminvq_f64): Likewise.
8840 (vminvq_s8): Likewise.
8841 (vminvq_s16): Likewise.
8842 (vminvq_s32): Likewise.
8843 (vminvq_u8): Likewise.
8844 (vminvq_u16): Likewise.
8845 (vminvq_u32): Likewise.
8846 (vminnmv_f32): Likewise.
8847 (vminnmvq_f32): Likewise.
8848 (vminnmvq_f64): Likewise.
8850 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8852 * config/aarch64/aarch64-simd.md
8853 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
8854 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
8855 (*aarch64_mul3_elt<mode>): Likewise.
8856 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
8857 (*aarch64_mul3_elt_to_64v2df): Likewise.
8858 (*aarch64_mla_elt<mode>): Likewise.
8859 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8860 (*aarch64_mls_elt<mode>): Likewise.
8861 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8862 (*aarch64_fma4_elt<mode>): Likewise.
8863 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8864 (*aarch64_fma4_elt_to_64v2df): Likewise.
8865 (*aarch64_fnma4_elt<mode>): Likewise.
8866 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8867 (*aarch64_fnma4_elt_to_64v2df): Likewise.
8868 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
8869 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
8870 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8871 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8872 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
8873 (aarch64_sqdmull_lane<mode>_internal): Likewise.
8874 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
8876 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
8878 * config/aarch64/aarch64-simd.md
8879 (aarch64_be_checked_get_lane<mode>): New define_expand.
8880 * config/aarch64/aarch64-simd-builtins.def
8881 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
8882 New builtin definition.
8883 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
8884 Use new safe be builtin.
8886 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8888 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
8890 (aarch64_be_st1<mode>): Likewise.
8891 (aarch_ld1<VALL:mode>): Define_expand modified.
8892 (aarch_st1<VALL:mode>): Likewise.
8893 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
8894 (UNSPEC_ST1): Likewise.
8896 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
8898 * config/microblaze/microblaze.md: Add trap insn and attribute
8900 2014-01-23 Dodji Seketeli <dodji@redhat.com>
8902 PR preprocessor/58580
8903 * input.h (location_get_source_line): Take an additional line_size
8905 (void diagnostics_file_cache_fini): Declare new function.
8906 * input.c (struct fcache): New type.
8907 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
8908 New static constants.
8909 (diagnostic_file_cache_init, total_lines_num)
8910 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
8911 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
8912 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
8913 (get_next_line, read_next_line, goto_next_line, read_line_num):
8914 New static function definitions.
8915 (diagnostic_file_cache_fini): New function.
8916 (location_get_source_line): Take an additional output line_len
8917 parameter. Re-write using lookup_or_add_file_to_cache_tab and
8919 * diagnostic.c (diagnostic_finish): Call
8920 diagnostic_file_cache_fini.
8921 (adjust_line): Take an additional input parameter for the length
8922 of the line, rather than calculating it with strlen.
8923 (diagnostic_show_locus): Adjust the use of
8924 location_get_source_line and adjust_line with respect to their new
8925 signature. While displaying a line now, do not stop at the first
8926 null byte. Rather, display the zero byte as a space and keep
8927 going until we reach the size of the line.
8928 * Makefile.in: Add vec.o to OBJS-libcommon
8930 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8931 Ilya Tocar <ilya.tocar@intel.com>
8933 * config/i386/avx512fintrin.h (_mm512_kmov): New.
8934 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
8935 (__builtin_ia32_kmov16): Ditto.
8936 * config/i386/i386.md (UNSPEC_KMOV): New.
8939 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8941 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
8942 (_mm512_storeu_si512): Ditto.
8944 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
8947 * rtl.h (get_referenced_operands): Declare.
8948 * recog.c (get_referenced_operands): New function.
8949 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
8950 operands have been referenced when recording LO_SUM references.
8952 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
8954 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
8956 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8958 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
8959 Enable for generic and recent AMD targets.
8961 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8963 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
8964 ARG_SIZE note when adjustment was eliminated.
8966 2014-01-22 Jeff Law <law@redhat.com>
8968 PR tree-optimization/59597
8969 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
8970 in file. Accept new argument REGISTERING and use it to modify
8971 dump output appropriately.
8972 (register_jump_thread): Corresponding changes.
8973 (mark_threaded_blocks): Reinstate code to cancel unprofitable
8974 thread paths involving joiner blocks. Add code to dump cancelled
8975 jump threading paths.
8977 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
8979 PR rtl-optimization/59477
8980 * lra-constraints.c (inherit_in_ebb): Process call for living hard
8981 regs. Update reloads_num and potential_reload_hard_regs for all insns.
8983 2014-01-22 Tom Tromey <tromey@redhat.com>
8985 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
8987 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
8989 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8991 PR rtl-optimization/59896
8992 * lra-constraints.c (process_alt_operands): Check unused note for
8993 matched operands of insn with no output reloads.
8995 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
8997 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
8998 (mips_move_from_gpr_cost): Likewise.
9000 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
9002 PR rtl-optimization/59858
9003 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
9004 ira_class_hard_regs_num.
9005 (process_alt_operands): Increase reject for dying matched operand.
9007 2014-01-21 Jakub Jelinek <jakub@redhat.com>
9010 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
9011 smaller than size, perform several stores or loads and stores
9012 at dst + count - size to store or copy all of size bytes, rather
9013 than just last modesize bytes.
9015 2014-01-20 DJ Delorie <dj@redhat.com>
9017 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
9018 that CLOBBERs are REGs before propogating their values.
9020 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
9023 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
9024 (cgraph_inline_failed_type): New function.
9025 * cgraph.h (DEFCIFCODE): Add type.
9026 (cgraph_inline_failed_type_t): New enum.
9027 (cgraph_inline_failed_type): New prototype.
9028 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
9029 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
9030 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
9031 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
9032 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
9033 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
9034 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
9035 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
9036 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
9037 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
9038 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
9039 OPTIMIZATION_MISMATCH.
9040 * tree-inline.c (expand_call_inline): Emit errors during
9041 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
9043 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9046 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
9047 mode attribute in insn output.
9049 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
9051 * output.h (output_constant): Delete.
9052 * varasm.c (output_constant): Make private.
9054 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
9056 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
9058 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9061 * tree.h (fold_builtin_strcat): New prototype.
9062 * builtins.c (fold_builtin_strcat): No longer static. Add len
9063 argument, if non-NULL, don't call c_strlen. Optimize
9064 directly into __builtin_memcpy instead of __builtin_strcpy.
9065 (fold_builtin_2): Adjust fold_builtin_strcat caller.
9066 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
9068 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9070 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9071 for SImode_address_operand operands, having only a REG argument.
9073 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
9075 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
9076 loader name using mbig-endian.
9077 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
9079 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9081 * doc/invoke.texi (-march): Clarify documentation for AArch64.
9085 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
9087 * config/aarch64/aarch64-protos.h
9088 (aarch64_cannot_change_mode_class_ptr): Declare.
9089 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
9090 aarch64_cannot_change_mode_class_ptr): New.
9091 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
9092 backend hook aarch64_cannot_change_mode_class.
9094 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9096 * common/config/aarch64/aarch64-common.c
9097 (aarch64_handle_option): Don't handle any option order logic here.
9098 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
9099 selected_cpu, warn on architecture version mismatch.
9100 (aarch64_override_options): Fix parsing order for option strings.
9102 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9103 Iain Sandoe <iain@codesourcery.com>
9106 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
9107 warning. Amend comment to reflect current functionality.
9109 2014-01-20 Richard Biener <rguenther@suse.de>
9112 * builtins.c (fold_builtin_strcat): Remove case better handled
9113 by tree-ssa-strlen.c.
9115 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
9117 * config/aarch64/aarch64.opt
9118 (mcpu, march, mtune): Make case-insensitive.
9120 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9123 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9124 if operands[1] is a REG or ZERO_EXTEND of a REG.
9126 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
9128 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
9130 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
9132 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
9133 long non-pic millicode calls.
9135 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9137 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
9139 2014-01-19 Kito Cheng <kito@0xlab.org>
9141 * builtins.c (expand_movstr): Check movstr expand done or fail.
9143 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9144 H.J. Lu <hongjiu.lu@intel.com>
9147 * config/i386/i386.md (*lea<mode>): Zero-extend return register
9148 to DImode for zero-extended addresses.
9150 2014-01-19 Jakub Jelinek <jakub@redhat.com>
9152 PR rtl-optimization/57763
9153 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
9154 on the new indirect jump_insn and increment LABEL_NUSES (label).
9156 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
9160 * config.gcc (x86_archs): New variable.
9161 (x86_64_archs): Likewise.
9162 (x86_cpus): Likewise.
9163 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
9164 --with-arch/--with-cpu= options.
9165 Support --with-arch=/--with-cpu={nehalem,westmere,
9166 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
9168 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9170 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
9171 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
9173 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9175 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
9177 2014-01-18 Jakub Jelinek <jakub@redhat.com>
9180 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
9181 clear cpp_get_options (parse_in)->warn_unused_macros for
9182 ix86_target_macros_internal with cpp_define.
9184 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
9186 * jump.c (delete_related_insns): Keep (use (insn))s.
9187 * reorg.c (redundant_insn): Check for barriers too.
9189 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9191 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
9193 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
9195 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
9196 call to $$dyncall when TARGET_LONG_CALLS is true.
9198 2014-01-17 Jeff Law <law@redhat.com>
9200 * ree.c (combine_set_extension): Temporarily disable test for
9201 changing number of hard registers.
9203 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9206 * ipa-inline-analysis.c (inline_free_summary):
9207 Do not free summary of aliases.
9209 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9212 * gimplify.c (gimplify_expr): Use create_tmp_var
9213 instead of create_tmp_var_raw. If cond doesn't have
9214 integral type, don't add the IFN_ANNOTATE builtin at all.
9216 2014-01-17 Martin Jambor <mjambor@suse.cz>
9219 * ipa-cp.c (prev_edge_clone): New variable.
9220 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
9221 Also resize prev_edge_clone vector.
9222 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
9223 (ipcp_edge_removal_hook): New function.
9224 (ipcp_driver): Register ipcp_edge_removal_hook.
9226 2014-01-17 Andrew Pinski <apinski@cavium.com>
9227 Steve Ellcey <sellcey@mips.com>
9230 * config/mips/mips.c (mips_print_operand): Check operand mode instead
9233 2014-01-17 Jeff Law <law@redhat.com>
9236 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
9237 so that pass_ccp runs first.
9239 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9241 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
9242 (ix86_adjust_cost): Use !TARGET_XXX.
9243 (do_reorder_for_imul): Likewise.
9244 (swap_top_of_ready_list): Likewise.
9245 (ix86_sched_reorder): Likewise.
9247 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9249 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9250 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
9251 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
9252 (intel_memset): New. Duplicate slm_memset.
9253 (intel_cost): New. Duplicate slm_cost.
9254 (m_INTEL): New macro.
9255 (processor_target_table): Add "intel".
9256 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
9257 with PROCESSOR_INTEL for "intel".
9258 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
9259 PROCESSOR_SILVERMONT.
9260 (ix86_issue_rate): Likewise.
9261 (ix86_adjust_cost): Likewise.
9262 (ia32_multipass_dfa_lookahead): Likewise.
9263 (swap_top_of_ready_list): Likewise.
9264 (ix86_sched_reorder): Likewise.
9265 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
9266 instead of TARGET_OPT_AGU.
9267 * config/i386/i386.h (TARGET_INTEL): New.
9268 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
9269 (processor_type): Add PROCESSOR_INTEL.
9270 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
9271 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
9273 2014-01-17 Marek Polacek <polacek@redhat.com>
9276 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
9279 2014-01-17 Richard Biener <rguenther@suse.de>
9281 PR tree-optimization/46590
9282 * opts.c (default_options_table): Add entries for
9283 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
9284 all enabled at -O1 but not for -Og.
9285 * common.opt (fbranch-count-reg): Remove Init(1).
9286 (fmove-loop-invariants): Likewise.
9287 (ftree-pta): Likewise.
9289 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9291 * config/i386/i386.c (ix86_data_alignment): For compatibility with
9292 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
9293 decls to at least the GCC 4.8 used alignments.
9296 * tree-nested.c (convert_nonlocal_reference_stmt,
9297 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
9298 of GIMPLE_BIND stmts, adjust associated decls.
9300 2014-01-17 Richard Biener <rguenther@suse.de>
9302 PR tree-optimization/46590
9303 * vec.h (vec<>::bseach): New member function implementing
9304 binary search according to C89 bsearch.
9305 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
9306 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
9307 bitmap pointer again. Make accesses_in_loop a flat array.
9308 (mem_ref_obstack): New global.
9309 (outermost_indep_loop): Adjust for mem_ref->stored changes.
9310 (mark_ref_stored): Likewise.
9311 (ref_indep_loop_p_2): Likewise.
9312 (set_ref_stored_in_loop): New helper function.
9313 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
9314 (memref_free): Adjust.
9315 (record_mem_ref_loc): Simplify.
9316 (gather_mem_refs_stmt): Adjust.
9317 (sort_locs_in_loop_postorder_cmp): New function.
9318 (analyze_memory_references): Sort accesses_in_loop after
9319 loop postorder number.
9320 (find_ref_loc_in_loop_cmp): New function.
9321 (for_all_locs_in_loop): Find relevant cluster of locs in
9322 accesses_in_loop and iterate without recursion.
9323 (execute_sm): Avoid uninit warning.
9324 (struct ref_always_accessed): Simplify.
9325 (ref_always_accessed::operator ()): Likewise.
9326 (ref_always_accessed_p): Likewise.
9327 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
9328 loop postorder numbers here.
9329 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
9332 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9335 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
9336 on decls for which assemble_alias has been called.
9338 2014-01-17 Nick Clifton <nickc@redhat.com>
9340 * config/msp430/msp430.opt: (mcpu): New option.
9341 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
9342 (msp430_option_override): Parse target_cpu. If the MCU name
9343 matches a generic string, clear target_mcu.
9344 (msp430_attr): Allow numeric interrupt values up to 63.
9345 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
9346 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
9348 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
9350 * config/msp430/msp430.md (popm): Use %J rather than %I.
9351 (addsi3): Use msp430_nonimmediate_operand for operand 2.
9352 (addhi_cy_i): Use immediate_operand for operand 2.
9353 * doc/invoke.texi: Document -mcpu option.
9355 2014-01-17 Richard Biener <rguenther@suse.de>
9357 PR rtl-optimization/38518
9358 * df.h (df_analyze_loop): Declare.
9359 * df-core.c: Include cfgloop.h.
9360 (df_analyze_1): Split out main part of df_analyze.
9361 (df_analyze): Adjust.
9362 (loop_inverted_post_order_compute): New function.
9363 (loop_post_order_compute): Likewise.
9364 (df_analyze_loop): New function avoiding whole-function
9366 * loop-invariant.c (find_defs): Use df_analyze_loop.
9367 (find_invariants): Adjust.
9368 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
9370 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
9372 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
9373 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
9375 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
9377 * ipa-ref.c (ipa_remove_stmt_references): Fix references
9378 traversal when removing references.
9380 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
9383 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
9385 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
9388 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
9389 pushing a reload for an autoinc when we had previously reloaded an
9390 inner part of the address.
9392 2014-01-16 Jakub Jelinek <jakub@redhat.com>
9394 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
9396 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
9397 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
9398 when not giving up or versioning for alias only because of
9400 (vect_analyze_data_ref_dependences): Set to true.
9401 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
9403 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
9404 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
9408 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
9411 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
9412 operand 0 predicate for gathers, use a new pseudo as subtarget.
9414 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9417 * lra-constraints.c (process_alt_operands): Add printing debug info.
9418 Check absence of input/output reloads for matched operands too.
9420 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9422 PR rtl-optimization/59835
9423 * ira.c (ira_init_register_move_cost): Increase cost for
9426 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
9428 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
9430 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
9433 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
9434 non-register objects. Use gen_(high/low)part more consistently.
9437 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9440 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
9441 endian support, remove tests for WORDS_BIG_ENDIAN.
9442 (p8_mfvsrd_3_<mode>): Likewise.
9443 (reload_gpr_from_vsx<mode>): Likewise.
9444 (reload_gpr_from_vsxsf): Likewise.
9445 (p8_mfvsrd_4_disf): Likewise.
9447 2014-01-16 Richard Biener <rguenther@suse.de>
9449 PR rtl-optimization/46590
9450 * lcm.c (compute_antinout_edge): Use postorder iteration.
9451 (compute_laterin): Use inverted postorder iteration.
9453 2014-01-16 Nick Clifton <nickc@redhat.com>
9456 * varasm.c (output_constant): Return the number of bytes actually
9458 (output_constructor_array_range): Update the field size with the
9459 number of bytes emitted by output_constant.
9460 (output_constructor_regular_field): Likewise. Also do not
9461 complain if the total number of bytes emitted is now greater
9462 than the expected fieldpos.
9463 * output.h (output_constant): Update prototype and descriptive comment.
9465 2014-01-16 Marek Polacek <polacek@redhat.com>
9468 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
9469 it is error_mark_node.
9471 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
9473 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
9474 VALID_AVX256_REG_OR_OI_MODE.
9476 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
9478 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
9479 current procedure should be profiled.
9481 2014-01-15 Andrew Pinski <apinski@cavium.com>
9483 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
9484 of moving from/to the STACK_REG register class.
9486 2014-01-15 Richard Henderson <rth@redhat.com>
9489 * reginfo.c (global_regs_decl): Globalize.
9490 * rtl.h (global_regs_decl): Declare.
9491 * ira.c (do_reload): Diagnose frame_pointer_needed and it
9492 reserved via global_regs.
9494 2014-01-15 Teresa Johnson <tejohnson@google.com>
9496 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
9498 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9500 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
9501 and vmulosh rather than call gen_vec_widen_smult_*.
9502 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
9503 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
9504 (vec_widen_smult_even_v16qi): Likewise.
9505 (vec_widen_umult_even_v8hi): Likewise.
9506 (vec_widen_smult_even_v8hi): Likewise.
9507 (vec_widen_umult_odd_v16qi): Likewise.
9508 (vec_widen_smult_odd_v16qi): Likewise.
9509 (vec_widen_umult_odd_v8hi): Likewise.
9510 (vec_widen_smult_odd_v8hi): Likewise.
9511 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
9512 vmuloub rather than call gen_vec_widen_umult_*.
9513 (vec_widen_umult_lo_v16qi): Likewise.
9514 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
9515 vmulosb rather than call gen_vec_widen_smult_*.
9516 (vec_widen_smult_lo_v16qi): Likewise.
9517 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
9518 rather than call gen_vec_widen_umult_*.
9519 (vec_widen_umult_lo_v8hi): Likewise.
9520 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
9521 rather than call gen_vec_widen_smult_*.
9522 (vec_widen_smult_lo_v8hi): Likewise.
9524 2014-01-15 Jeff Law <law@redhat.com>
9526 PR tree-optimization/59747
9527 * ree.c (find_and_remove_re): Properly handle case where a second
9528 eliminated extension requires widening a copy created for elimination
9529 of a prior extension.
9530 (combine_set_extension): Ensure that the number of hard regs needed
9531 for a destination register does not change when we widen it.
9533 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
9535 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
9536 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
9537 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
9538 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
9539 (avr-*-rtems*): Likewise.
9540 (bfin*-rtems*): Likewise.
9541 (moxie-*-rtems*): Likewise.
9542 (h8300-*-rtems*): Likewise.
9543 (i[34567]86-*-rtems*): Likewise.
9544 (lm32-*-rtems*): Likewise.
9545 (m32r-*-rtems*): Likewise.
9546 (m68k-*-rtems*): Likewise.
9547 (microblaze*-*-rtems*): Likewise.
9548 (mips*-*-rtems*): Likewise.
9549 (powerpc-*-rtems*): Likewise.
9550 (sh-*-rtems*): Likewise.
9551 (sparc-*-rtems*): Likewise.
9552 (sparc64-*-rtems*): Likewise.
9553 (v850-*-rtems*): Likewise.
9554 (m32c-*-rtems*): Likewise.
9556 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9558 PR rtl-optimization/59511
9559 * ira.c (ira_init_register_move_cost): Use memory costs for some
9560 cases of register move cost calculations.
9561 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9562 instead of BB frequency.
9563 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9564 * lra-assigns.c (find_hard_regno_for): Ditto.
9566 2014-01-15 Richard Biener <rguenther@suse.de>
9568 PR tree-optimization/59822
9569 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9570 (vectorizable_load): Use it to hoist defs of uses of invariant
9571 loads out of the loop.
9573 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9574 Kugan Vivekanandarajah <kuganv@linaro.org>
9577 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9580 2014-01-15 Richard Biener <rguenther@suse.de>
9582 PR rtl-optimization/59802
9583 * lcm.c (compute_available): Use inverted postorder to seed
9584 the initial worklist.
9586 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9589 * config/s390/s390.c (s390_preferred_reload_class): Don't return
9590 ADDR_REGS for invalid symrefs in non-PIC code.
9592 2014-01-15 Jakub Jelinek <jakub@redhat.com>
9595 * builtins.c (determine_block_size): Initialize *probable_max_size
9596 even if len_rtx is CONST_INT.
9598 2014-01-14 Andrew Pinski <apinski@cavium.com>
9600 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
9601 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
9602 (cortexa53_tunings): Likewise.
9603 (aarch64_sched_issue_rate): New function.
9604 (TARGET_SCHED_ISSUE_RATE): Define.
9606 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9608 * ira-costs.c (find_costs_and_classes): Add missed
9609 ira_init_register_move_cost_if_necessary.
9611 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9614 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
9616 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
9619 * config/i386/i386.c (type_natural_mode): Add a bool parameter
9620 to indicate if type is used for function return value. Warn ABI
9621 change if the vector mode isn't available for function return value.
9622 (ix86_function_arg_advance): Pass false to type_natural_mode.
9623 (ix86_function_arg): Likewise.
9624 (ix86_gimplify_va_arg): Likewise.
9625 (function_arg_32): Don't warn ABI change.
9626 (ix86_function_value): Pass true to type_natural_mode.
9627 (ix86_return_in_memory): Likewise.
9628 (ix86_struct_value_rtx): Removed.
9629 (TARGET_STRUCT_VALUE_RTX): Likewise.
9631 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9633 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
9634 converting a conditional jump into a conditional return.
9636 2014-01-14 Richard Biener <rguenther@suse.de>
9638 PR tree-optimization/58921
9639 PR tree-optimization/59006
9640 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
9641 hoisting invariant stmts.
9642 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
9643 invariant loads on the preheader edge if possible.
9645 2014-01-14 Joey Ye <joey.ye@arm.com>
9647 * doc/plugin.texi (Building GCC plugins): Update to C++.
9649 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
9651 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
9652 (_mm_rcp28_round_ss): Ditto.
9653 (_mm_rsqrt28_round_sd): Ditto.
9654 (_mm_rsqrt28_round_ss): Ditto.
9655 (_mm_rcp28_sd): Ditto.
9656 (_mm_rcp28_ss): Ditto.
9657 (_mm_rsqrt28_sd): Ditto.
9658 (_mm_rsqrt28_ss): Ditto.
9659 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
9660 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
9661 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
9662 (IX86_BUILTIN_RCP28SD): Ditto.
9663 (IX86_BUILTIN_RCP28SS): Ditto.
9664 (IX86_BUILTIN_RSQRT28SD): Ditto.
9665 (IX86_BUILTIN_RSQRT28SS): Ditto.
9666 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
9667 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
9668 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
9669 (ix86_expand_special_args_builtin): Expand new FTYPE.
9670 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
9671 (srcp14<mode>): Make insn unary.
9672 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
9673 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
9674 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
9675 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
9676 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
9677 Fix rounding: make it SAE only.
9678 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9680 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9682 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
9683 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
9684 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
9685 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
9686 (round_saeonly_mask_scalar_operand4): Ditto.
9687 (round_saeonly_mask_scalar_op3): Ditto.
9688 (round_saeonly_mask_scalar_op4): Ditto.
9690 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9692 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9693 Implement -maltivec=be for vec_insert and vec_extract.
9695 2014-01-10 DJ Delorie <dj@redhat.com>
9697 * config/msp430/msp430.md (call_internal): Don't allow memory
9698 references with SP as the base register.
9699 (call_value_internal): Likewise.
9700 * config/msp430/constraints.md (Yc): New. For memory references
9701 that don't use SP as a base register.
9703 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
9704 "an integer without a # prefix"
9705 * config/msp430/msp430.md (epilogue_helper): Use it.
9707 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9710 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
9711 AVX512F gather builtins.
9712 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
9713 on gather decls with INTEGER_TYPE masktype.
9714 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
9715 directly into the builtin rather than hoisting it before loop.
9717 PR tree-optimization/59387
9718 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
9719 (scev_const_prop): If folded_casts and type has undefined overflow,
9720 use force_gimple_operand instead of force_gimple_operand_gsi and
9721 for each added stmt if it is assign with
9722 arith_code_with_undefined_signed_overflow, call
9723 rewrite_to_defined_overflow.
9724 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
9725 gimple-fold.h instead.
9726 (arith_code_with_undefined_signed_overflow,
9727 rewrite_to_defined_overflow): Moved to ...
9728 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
9729 rewrite_to_defined_overflow): ... here. No longer static.
9730 Include gimplify-me.h.
9731 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
9732 rewrite_to_defined_overflow): New prototypes.
9734 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9736 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
9738 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
9740 * builtins.c (get_object_alignment_2): Minor tweak.
9741 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
9743 2014-01-13 Christian Bruel <christian.bruel@st.com>
9745 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
9746 optimized non constant lengths.
9748 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9751 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
9752 load as __atomic_load_N if possible.
9754 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9756 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
9758 (rs6000_expand_builtin): Adjust call.
9760 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9763 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
9764 * config/rs6000/rs6000.c: Include target-globals.h.
9765 (rs6000_set_current_function): Instead of doing target_reinit
9766 unconditionally, use save_target_globals_default_opts and
9767 restore_target_globals.
9769 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
9771 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
9772 (rs6000_expand_builtin): Handle mffs and mtfsf.
9773 (rs6000_init_builtins): Define mffs and mtfsf.
9774 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
9775 (rs6000_mffs): New pattern.
9776 (rs6000_mtfsf): New pattern.
9778 2014-01-11 Bin Cheng <bin.cheng@arm.com>
9780 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
9781 Start narrowing with START. Apply candidate-use pair
9782 and check overall cost in narrowing.
9783 (iv_ca_prune): Pass new argument.
9785 2014-01-10 Jeff Law <law@redhat.com>
9788 * ree.c (combine_reaching_defs): Ensure the defining statement
9789 occurs before the extension when optimizing extensions with
9790 different source and destination hard registers.
9792 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9795 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
9796 vtables into the type inheritance graph.
9798 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9800 PR rtl-optimization/59754
9801 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
9802 modes in the REGNO != REGNO case.
9804 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9806 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
9808 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9810 PR tree-optimization/59745
9811 * tree-predcom.c (tree_predictive_commoning_loop): Call
9812 free_affine_expand_cache if giving up because components is NULL.
9814 * target-globals.c (save_target_globals): Allocate < 4KB structs using
9815 GC in payload of target_globals struct instead of allocating them on
9816 the heap and the larger structs separately using GC.
9817 * target-globals.h (struct target_globals): Make regs, hard_regs,
9818 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
9819 of GTY((skip)) and change type to void *.
9820 (reset_target_globals): Cast loads from those fields to corresponding
9823 2014-01-10 Steve Ellcey <sellcey@mips.com>
9826 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
9827 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
9828 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
9830 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
9833 * aarch64-modes.def (CC_Zmode): New flags mode.
9834 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
9835 represents an equality.
9836 (aarch64_get_condition_code): Handle CC_Zmode.
9837 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
9839 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9841 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
9842 extraction in good case.
9844 2014-01-10 Richard Biener <rguenther@suse.de>
9846 PR tree-optimization/59374
9847 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
9848 checking after SLP discovery. Mark stmts not participating
9849 in any SLP instance properly.
9851 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9853 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
9854 when handling a SET rtx.
9856 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9858 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
9859 (cortex-a57): Likewise.
9860 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
9862 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9864 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
9865 non-iwmmxt builtins.
9867 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9871 * ipa-devirt.c record_target_from_binfo): Take as argument
9872 stack of binfos and lookup matching one for virtual inheritance.
9873 (possible_polymorphic_call_targets_1): Update.
9875 2014-01-10 Huacai Chen <chenhc@lemote.com>
9877 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
9878 kernel strings for Loongson-2E/2F/3A.
9880 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9883 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
9884 is_gimple_call before calling gimple_call_internal_p.
9886 2014-01-09 Steve Ellcey <sellcey@mips.com>
9888 * Makefile.in (TREE_FLOW_H): Remove.
9889 (TREE_SSA_H): Add file names from tree-flow.h.
9890 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
9891 * tree.h: Remove tree-flow.h reference.
9892 * hash-table.h: Remove tree-flow.h reference.
9893 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
9894 reference with tree-ssa-loop.h.
9896 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9898 * doc/invoke.texi: Add -maltivec={be,le} options, and document
9899 default element-order behavior for -maltivec.
9900 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
9901 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
9902 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
9903 when targeting big endian, at least for now.
9904 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
9906 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9909 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
9910 var satisfies use_register_for_decl, just take into account type
9911 alignment, rather than decl alignment.
9913 PR tree-optimization/59622
9914 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
9915 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
9916 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
9917 Don't devirtualize for inplace at all. For targets.length () == 1,
9918 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
9920 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9922 * config/i386/i386.md (cpu): Remove the unused btver1.
9924 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9926 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
9928 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9931 * tree-core.h (struct target_globals): New forward declaration.
9932 (struct tree_target_option): Add globals field.
9933 * tree.h (TREE_TARGET_GLOBALS): Define.
9934 (prepare_target_option_nodes_for_pch): New prototype.
9935 * target-globals.h (struct target_globals): Define even if
9937 * tree.c (prepare_target_option_node_for_pch,
9938 prepare_target_option_nodes_for_pch): New functions.
9939 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
9940 * config/i386/i386.c: Include target-globals.h.
9941 (ix86_set_current_function): Instead of doing target_reinit
9942 unconditionally, use save_target_globals_default_opts and
9943 restore_target_globals.
9945 2014-01-09 Richard Biener <rguenther@suse.de>
9947 PR tree-optimization/59715
9948 * tree-cfg.h (split_critical_edges): Declare.
9949 * tree-cfg.c (split_critical_edges): Export.
9950 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
9952 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
9954 * cfgexpand.c (expand_stack_vars): Optionally disable
9955 asan stack protection.
9956 (expand_used_vars): Likewise.
9957 (partition_stack_vars): Likewise.
9958 * asan.c (asan_emit_stack_protection): Optionally disable
9959 after return stack usage.
9960 (instrument_derefs): Optionally disable memory access instrumentation.
9961 (instrument_builtin_call): Likewise.
9962 (instrument_strlen_call): Likewise.
9963 (asan_protect_global): Optionally disable global variables protection.
9964 * doc/invoke.texi: Added doc for new options.
9965 * params.def: Added new options.
9966 * params.h: Likewise.
9968 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9970 PR rtl-optimization/59724
9971 * ifcvt.c (cond_exec_process_if_block): Don't call
9972 flow_find_head_matching_sequence with 0 longest_match.
9973 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
9974 non-active insns if !stop_after.
9975 (try_head_merge_bb): Revert 2014-01-07 changes.
9977 2014-01-08 Jeff Law <law@redhat.com>
9979 * ree.c (get_sub_rtx): New function, extracted from...
9980 (merge_def_and_ext): Here.
9981 (combine_reaching_defs): Use get_sub_rtx.
9983 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
9985 * cgraph.h (varpool_variable_node): Do not choke on null node.
9987 2014-01-08 Catherine Moore <clm@codesourcery.com>
9989 * config/mips/mips.md (simple_return): Attempt to use JRC
9991 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
9993 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9995 PR rtl-optimization/59137
9996 * reorg.c (steal_delay_list_from_target): Call update_block for
9998 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
10000 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10002 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
10003 two duplicate entries.
10005 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
10008 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
10010 * config/mips/mips.c (mips_truncated_op_cost): New function.
10011 (mips_rtx_costs): Adjust test for BADDU.
10012 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
10014 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
10016 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
10017 (*baddu_si): ...this new pattern.
10019 2014-01-08 Jakub Jelinek <jakub@redhat.com>
10022 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
10024 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
10026 PR middle-end/57748
10027 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
10029 (expand_expr, expand_normal): Adjust.
10030 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
10031 inner_reference_p. Use inner_reference_p to expand inner references.
10032 (store_expr): Adjust.
10033 * cfgexpand.c (expand_call_stmt): Adjust.
10035 2014-01-08 Rong Xu <xur@google.com>
10037 * gcov-io.c (gcov_var): Move from gcov-io.h.
10038 (gcov_position): Ditto.
10039 (gcov_is_error): Ditto.
10040 (gcov_rewrite): Ditto.
10041 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
10042 only part to libgcc/libgcov.h.
10044 2014-01-08 Marek Polacek <polacek@redhat.com>
10046 PR middle-end/59669
10047 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
10049 2014-01-08 Marek Polacek <polacek@redhat.com>
10052 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
10054 2014-01-08 Jakub Jelinek <jakub@redhat.com>
10056 PR rtl-optimization/59649
10057 * stor-layout.c (get_mode_bounds): For BImode return
10058 0 and STORE_FLAG_VALUE.
10060 2014-01-08 Richard Biener <rguenther@suse.de>
10062 PR middle-end/59630
10063 * gimple.h (is_gimple_builtin_call): Remove.
10064 (gimple_builtin_call_types_compatible_p): New.
10065 (gimple_call_builtin_p): New overload.
10066 * gimple.c (is_gimple_builtin_call): Remove.
10067 (validate_call): Rename to ...
10068 (gimple_builtin_call_types_compatible_p): ... this and export. Also
10069 check return types.
10070 (validate_type): New static function.
10071 (gimple_call_builtin_p): New overload and adjust.
10072 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
10073 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
10074 (gimple_fold_stmt_to_constant_1): Likewise.
10075 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
10077 2014-01-08 Richard Biener <rguenther@suse.de>
10079 PR middle-end/59471
10080 * gimplify.c (gimplify_expr): Gimplify register-register type
10081 VIEW_CONVERT_EXPRs to separate stmts.
10083 2014-01-07 Jeff Law <law@redhat.com>
10085 PR middle-end/53623
10086 * ree.c (combine_set_extension): Handle case where source
10087 and destination registers in an extension insn are different.
10088 (combine_reaching_defs): Allow source and destination registers
10089 in extension to be different under limited circumstances.
10090 (add_removable_extension): Remove restriction that the
10091 source and destination registers in the extension are the same.
10092 (find_and_remove_re): Emit a copy from the extension's
10093 destination to its source after the defining insn if
10094 the source and destination registers are different.
10096 PR middle-end/59285
10097 * ifcvt.c (merge_if_block): If we are merging a block with more than
10098 one successor with a block with no successors, remove any BARRIER
10099 after the second block.
10101 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
10103 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
10105 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
10108 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
10109 for 14-bit register offsets when INT14_OK_STRICT is false.
10111 2014-01-07 Roland Stigge <stigge@antcom.de>
10112 Michael Meissner <meissner@linux.vnet.ibm.com>
10115 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
10116 Only check TFmode for SPE constants. Don't check TImode or TDmode.
10118 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
10120 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
10123 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
10125 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
10126 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
10127 rtx is const0_rtx or not.
10129 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
10132 * target-globals.c (save_target_globals): Remove this_fn_optab
10134 * toplev.c: Include optabs.h.
10135 (target_reinit): Temporarily restore the global options if another
10136 set of options are in force.
10138 2014-01-07 Jakub Jelinek <jakub@redhat.com>
10140 PR rtl-optimization/58668
10141 * cfgcleanup.c (flow_find_cross_jump): Don't count
10142 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
10143 to determine what is counted.
10144 (flow_find_head_matching_sequence): Use active_insn_p to determine
10146 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
10148 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
10149 determine what is counted.
10151 PR tree-optimization/59643
10152 * tree-predcom.c (split_data_refs_to_components): If one dr is
10153 read and one write, determine_offset fails and the write isn't
10154 in the bad component, just put the read into the bad component.
10156 2014-01-07 Mike Stump <mikestump@comcast.net>
10157 Jakub Jelinek <jakub@redhat.com>
10160 * tree-core.h (struct tree_optimization_option): Change optabs
10161 type from unsigned char * to void *.
10162 * optabs.c (init_tree_optimization_optabs): Adjust
10163 TREE_OPTIMIZATION_OPTABS initialization.
10165 2014-01-06 Jakub Jelinek <jakub@redhat.com>
10168 * config/i386/i386.h (struct machine_function): Add
10169 no_drap_save_restore field.
10170 * config/i386/i386.c (ix86_save_reg): Use
10171 !cfun->machine->no_drap_save_restore instead of
10172 crtl->stack_realign_needed.
10173 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
10174 this function clears frame_pointer_needed. Set
10175 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
10176 and DRAP reg is needed.
10178 2014-01-06 Marek Polacek <polacek@redhat.com>
10181 * doc/implement-c.texi: Mention that other integer types are
10182 permitted as bit-field types in strictly conforming mode.
10184 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
10186 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
10187 is newly allocated.
10189 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
10191 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
10193 2014-01-06 Martin Jambor <mjambor@suse.cz>
10196 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
10198 * ipa-prop.c (ipa_print_node_params): Fix indentation.
10200 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
10204 * var-tracking.c (add_stores): Preserve the value of the source even if
10205 we don't record the store.
10207 2014-01-06 Terry Guo <terry.guo@arm.com>
10209 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
10211 2014-01-05 Iain Sandoe <iain@codesourcery.com>
10214 * config/darwin.c (darwin_function_section): Adjust return values to
10215 correspond to optimisation changes made in r206070.
10217 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
10219 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
10220 from prefetch_block tune setting.
10221 (nocona_cost): Correct size of prefetch block to 64.
10223 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
10225 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
10226 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
10227 used to save the static chain register in the computation of the offset
10228 from which the FP registers need to be restored.
10230 2014-01-04 Jakub Jelinek <jakub@redhat.com>
10232 PR tree-optimization/59519
10233 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
10234 ICE if get_current_def (current_new_name) is already non-NULL, as long
10235 as it is a phi result of some other phi in *new_exit_bb that has
10238 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
10239 or vmovdqu* for misaligned_operand.
10240 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
10241 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
10242 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
10243 aligned_mem for AVX512F masked aligned load and store builtins and for
10244 non-temporal moves.
10246 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
10248 PR tree-optimization/59651
10249 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
10250 Address range for negative step should be added by TYPE_SIZE_UNIT.
10252 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
10254 * config/m68k/m68k.c (handle_move_double): Handle pushes with
10255 overlapping registers also for registers other than the stack pointer.
10257 2014-01-03 Marek Polacek <polacek@redhat.com>
10260 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
10263 2014-01-03 Jakub Jelinek <jakub@redhat.com>
10266 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
10269 * config/i386/i386.md (MODE_SIZE): New mode attribute.
10270 (push splitter): Use <P:MODE_SIZE> instead of
10271 GET_MODE_SIZE (<P:MODE>mode).
10272 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
10273 (mov -1, reg peephole2): Likewise.
10274 * config/i386/sse.md (*mov<mode>_internal,
10275 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
10276 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
10277 *<code><mode>3, *andnot<mode>3<mask_name>,
10278 <mask_codefor><code><mode>3<mask_name>): Likewise.
10279 * config/i386/subst.md (mask_mode512bit_condition,
10280 sd_mask_mode512bit_condition): Likewise.
10282 2014-01-02 Xinliang David Li <davidxl@google.com>
10284 PR tree-optimization/59303
10285 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
10286 (dump_predicates): Better output format.
10287 (pred_equal_p): New function.
10288 (is_neq_relop_p): Ditto.
10289 (is_neq_zero_form_p): Ditto.
10290 (pred_expr_equal_p): Ditto.
10291 (pred_neg_p): Ditto.
10292 (simplify_pred): Ditto.
10293 (simplify_preds_2): Ditto.
10294 (simplify_preds_3): Ditto.
10295 (simplify_preds_4): Ditto.
10296 (simplify_preds): Ditto.
10297 (push_pred): Ditto.
10298 (push_to_worklist): Ditto.
10299 (get_pred_info_from_cmp): Ditto.
10300 (is_degenerated_phi): Ditto.
10301 (normalize_one_pred_1): Ditto.
10302 (normalize_one_pred): Ditto.
10303 (normalize_one_pred_chain): Ditto.
10304 (normalize_preds): Ditto.
10305 (normalize_cond_1): Remove function.
10306 (normalize_cond): Ditto.
10307 (is_gcond_subset_of): Ditto.
10308 (is_subset_of_any): Ditto.
10309 (is_or_set_subset_of): Ditto.
10310 (is_and_set_subset_of): Ditto.
10311 (is_norm_cond_subset_of): Ditto.
10312 (pred_chain_length_cmp): Ditto.
10313 (convert_control_dep_chain_into_preds): Type change.
10314 (find_predicates): Ditto.
10315 (find_def_preds): Ditto.
10316 (destroy_predicates_vecs): Ditto.
10317 (find_matching_predicates_in_rest_chains): Ditto.
10318 (use_pred_not_overlap_with_undef_path_pred): Ditto.
10319 (is_pred_expr_subset): Ditto.
10320 (is_pred_chain_subset_of): Ditto.
10321 (is_included_in): Ditto.
10322 (is_superset_of): Ditto.
10324 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10326 Update copyright years.
10328 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10330 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
10331 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
10332 config/arc/arc.md, config/arc/arc.opt,
10333 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
10334 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
10335 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
10336 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
10337 config/linux-protos.h, config/linux.c, config/winnt-c.c,
10338 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
10339 vtable-verify.c, vtable-verify.h: Use the standard form for the
10342 2014-01-02 Tobias Burnus <burnus@net-b.de>
10344 * gcc.c (process_command): Update copyright notice dates.
10345 * gcov-dump.c: Ditto.
10347 * doc/cpp.texi: Bump @copying's copyright year.
10348 * doc/cppinternals.texi: Ditto.
10349 * doc/gcc.texi: Ditto.
10350 * doc/gccint.texi: Ditto.
10351 * doc/gcov.texi: Ditto.
10352 * doc/install.texi: Ditto.
10353 * doc/invoke.texi: Ditto.
10355 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10357 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
10359 2014-01-01 Jakub Jelinek <jakub@redhat.com>
10361 * config/i386/sse.md (*mov<mode>_internal): Guard
10362 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
10364 PR rtl-optimization/59647
10365 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
10366 new_rtx into UNSIGNED_FLOAT rtxes.
10368 Copyright (C) 2014 Free Software Foundation, Inc.
10370 Copying and distribution of this file, with or without modification,
10371 are permitted in any medium without royalty provided the copyright
10372 notice and this notice are preserved.