1 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
3 * doc/extend.texi: Reflect current numbers of pragmas. Remove
6 2014-05-12 Mike Stump <mikestump@comcast.net>
9 * genattrtab.c (filename): Add.
10 (convert_set_attr_alternative): Improve error message.
11 (check_defs): Restore read_md_filename for error messages.
12 (gen_insn): Save filename.
14 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
16 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
17 -fno-local-ivars and -fivar-visibility.
18 * c-family/c.opt: Make -Wshadow also implicitly enable
21 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
23 * doc/tm.texi: Remove reference to deleted macro.
24 * doc/tm.texi.in: Likewise.
26 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
29 * config/avr/avr.c (avr_out_store_psi): Use correct constant
32 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
34 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
35 * config/arm/aout.h (License): Same.
36 * config/arm/bpabi.h (License): Same.
37 * config/arm/elf.h (License): Same.
38 * config/arm/linux-elf.h (License): Same.
39 * config/arm/linux-gas.h (License): Same.
40 * config/arm/netbsd-elf.h (License): Same.
41 * config/arm/uclinux-eabi.h (License): Same.
42 * config/arm/uclinux-elf.h (License): Same.
43 * config/arm/vxworks.h (License): Same.
45 2014-05-11 Jakub Jelinek <jakub@redhat.com>
47 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
48 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
49 number of operands to 3.
50 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
51 * tree-nested.c (convert_nonlocal_omp_clauses,
52 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
53 * gimplify.c (gimplify_scan_omp_clauses): Handle
54 OMP_CLAUSE_LINEAR_STMT.
55 * omp-low.c (lower_rec_input_clauses): Fix typo.
56 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
57 cast between Fortran boolean_type_node and C _Bool if
60 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
62 PR tree-optimization/61136
63 * wide-int.h (multiple_of_p): Define a version that doesn't return
65 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
66 integer_zerop/const_binop pair.
67 (multiple_of_p): Likewise, converting both operands to widest_int
70 2014-05-09 Teresa Johnson <tejohnson@google.com>
72 * cgraphunit.c (analyze_functions): Use correct dump file.
74 2014-05-09 Florian Weimer <fweimer@redhat.com>
76 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
78 (stack_protect_return_slot_p): New function.
79 (expand_used_vars): Call stack_protect_decl_p and
80 stack_protect_return_slot_p for -fstack-protector-strong.
82 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
85 * fold-const.c (fold_binary_loc): Changed width of mask.
87 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
89 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
90 unsigned int initializers for regno_in, regno_out.
92 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
95 * config/avr/avr.md (cc): Add new attribute set_vzn.
96 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
97 Set cc insn attribute to set_vzn instead of set_zn for alternatives
99 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
100 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
101 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
103 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
106 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
108 * wide-int.cc (UTItype): Define.
109 (UDWtype): Define for appropriate W_TYPE_SIZE.
111 2014-05-09 Richard Biener <rguenther@suse.de>
113 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
114 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
115 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
116 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
117 ssa_propagate): Adjust.
119 2014-05-08 Jeff Law <law@redhat.com>
121 PR tree-optimization/61009
122 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
123 tri-state rather than a boolean. When a block is too big to
124 thread through, inform caller via negative return value.
125 (thread_across_edge): If a block was too big for normal threading,
126 then it's too big for a joiner too, so remove temporary equivalences
127 and return immediately.
129 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
130 Matthias Klose <doko@ubuntu.com>
133 * optc-gen.awk: Fix option handling for -Wunused-parameter.
135 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
138 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
140 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
143 * config/alpha/alpha.c: Include gimple-iterator.h.
144 (alpha_gimple_fold_builtin): New function. Move
145 ALPHA_BUILTIN_UMULH folding from ...
146 (alpha_fold_builtin): ... here.
147 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
149 2014-05-08 Wei Mi <wmi@google.com>
152 * config/i386/i386.c (ix86_compute_frame_layout): Update
153 preferred_stack_boundary for call, expanded from tls descriptor.
154 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
155 to depend on SP register.
156 (*tls_local_dynamic_base_32_gnu): Ditto.
157 (*tls_local_dynamic_32_once): Ditto.
158 (tls_global_dynamic_64_<mode>): Set
159 ix86_tls_descriptor_calls_expanded_in_cfun.
160 (tls_local_dynamic_base_64_<mode>): Ditto.
161 (tls_global_dynamic_32): Set
162 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
163 to depend on SP register.
164 (tls_local_dynamic_base_32): Ditto.
166 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
168 * config/arm/arm_neon.h: Update comment.
169 * config/arm/neon-docgen.ml: Delete.
170 * config/arm/neon-gen.ml: Delete.
171 * doc/arm-neon-intrinsics.texi: Update comment.
173 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
175 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
177 (vand, vorr, veor, vorn, vbic): Remove.
178 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
180 (neon_vsub_unspec): Likewise.
181 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
183 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
185 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
186 (vadd_s16): Likewise.
187 (vadd_s32): Likewise.
188 (vadd_f32): Likewise.
190 (vadd_u16): Likewise.
191 (vadd_u32): Likewise.
192 (vadd_s64): Likewise.
193 (vadd_u64): Likewise.
194 (vaddq_s8): Likewise.
195 (vaddq_s16): Likewise.
196 (vaddq_s32): Likewise.
197 (vaddq_s64): Likewise.
198 (vaddq_f32): Likewise.
199 (vaddq_u8): Likewise.
200 (vaddq_u16): Likewise.
201 (vaddq_u32): Likewise.
202 (vaddq_u64): Likewise.
204 (vmul_s16): Likewise.
205 (vmul_s32): Likewise.
206 (vmul_f32): Likewise.
208 (vmul_u16): Likewise.
209 (vmul_u32): Likewise.
211 (vmulq_s8): Likewise.
212 (vmulq_s16): Likewise.
213 (vmulq_s32): Likewise.
214 (vmulq_f32): Likewise.
215 (vmulq_u8): Likewise.
216 (vmulq_u16): Likewise.
217 (vmulq_u32): Likewise.
219 (vsub_s16): Likewise.
220 (vsub_s32): Likewise.
221 (vsub_f32): Likewise.
223 (vsub_u16): Likewise.
224 (vsub_u32): Likewise.
225 (vsub_s64): Likewise.
226 (vsub_u64): Likewise.
227 (vsubq_s8): Likewise.
228 (vsubq_s16): Likewise.
229 (vsubq_s32): Likewise.
230 (vsubq_s64): Likewise.
231 (vsubq_f32): Likewise.
232 (vsubq_u8): Likewise.
233 (vsubq_u16): Likewise.
234 (vsubq_u32): Likewise.
235 (vsubq_u64): Likewise.
237 (vand_s16): Likewise.
238 (vand_s32): Likewise.
240 (vand_u16): Likewise.
241 (vand_u32): Likewise.
242 (vand_s64): Likewise.
243 (vand_u64): Likewise.
244 (vandq_s8): Likewise.
245 (vandq_s16): Likewise.
246 (vandq_s32): Likewise.
247 (vandq_s64): Likewise.
248 (vandq_u8): Likewise.
249 (vandq_u16): Likewise.
250 (vandq_u32): Likewise.
251 (vandq_u64): Likewise.
253 (vorr_s16): Likewise.
254 (vorr_s32): Likewise.
256 (vorr_u16): Likewise.
257 (vorr_u32): Likewise.
258 (vorr_s64): Likewise.
259 (vorr_u64): Likewise.
260 (vorrq_s8): Likewise.
261 (vorrq_s16): Likewise.
262 (vorrq_s32): Likewise.
263 (vorrq_s64): Likewise.
264 (vorrq_u8): Likewise.
265 (vorrq_u16): Likewise.
266 (vorrq_u32): Likewise.
267 (vorrq_u64): Likewise.
269 (veor_s16): Likewise.
270 (veor_s32): Likewise.
272 (veor_u16): Likewise.
273 (veor_u32): Likewise.
274 (veor_s64): Likewise.
275 (veor_u64): Likewise.
276 (veorq_s8): Likewise.
277 (veorq_s16): Likewise.
278 (veorq_s32): Likewise.
279 (veorq_s64): Likewise.
280 (veorq_u8): Likewise.
281 (veorq_u16): Likewise.
282 (veorq_u32): Likewise.
283 (veorq_u64): Likewise.
285 (vbic_s16): Likewise.
286 (vbic_s32): Likewise.
288 (vbic_u16): Likewise.
289 (vbic_u32): Likewise.
290 (vbic_s64): Likewise.
291 (vbic_u64): Likewise.
292 (vbicq_s8): Likewise.
293 (vbicq_s16): Likewise.
294 (vbicq_s32): Likewise.
295 (vbicq_s64): Likewise.
296 (vbicq_u8): Likewise.
297 (vbicq_u16): Likewise.
298 (vbicq_u32): Likewise.
299 (vbicq_u64): Likewise.
301 (vorn_s16): Likewise.
302 (vorn_s32): Likewise.
304 (vorn_u16): Likewise.
305 (vorn_u32): Likewise.
306 (vorn_s64): Likewise.
307 (vorn_u64): Likewise.
308 (vornq_s8): Likewise.
309 (vornq_s16): Likewise.
310 (vornq_s32): Likewise.
311 (vornq_s64): Likewise.
312 (vornq_u8): Likewise.
313 (vornq_u16): Likewise.
314 (vornq_u32): Likewise.
315 (vornq_u64): Likewise.
317 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
319 * wide-int.cc (UTItype): Define.
320 (UDWtype): Define for appropriate W_TYPE_SIZE.
322 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
324 PR tree-optimization/59100
325 * tree-ssa-phiopt.c: Include tree-inline.h.
326 (neutral_element_p, absorbing_element_p): New functions.
327 (value_replacement): Handle conditional binary operations with a
328 neutral or absorbing element.
330 2014-05-08 Richard Biener <rguenther@suse.de>
332 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
333 folding the expression.
334 (valueize_expr): Remove.
335 (visit_reference_op_load): Do not valueize the result of
337 (simplify_binary_expression): Likewise.
338 (simplify_unary_expression): Likewise.
340 2014-05-08 Richard Biener <rguenther@suse.de>
342 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
343 looking at TYPE_ARG_TYPES.
345 2014-05-08 Richard Biener <rguenther@suse.de>
347 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
348 pointer propagation special-case.
350 2014-05-08 Bin Cheng <bin.cheng@arm.com>
352 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
353 core part of address expressions.
355 2014-05-08 Alan Modra <amodra@gmail.com>
358 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
359 loads and stores when -mno-strict-align at any alignment.
360 (expand_block_clear): Similarly. Also correct calculation of
363 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
366 * tree-complex.c (expand_complex_move): Keep line info when expanding
368 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
369 of complex expression. Use new argument to display correct location
370 for values coming from phi statement.
371 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
372 (warn_uninitialized_phi): Pass location of phi argument to
374 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
375 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
377 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
379 * config/rs6000/predicates.md (indexed_address_mem): New.
380 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
381 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
382 fpstore_ux, fpstore_u.
383 (sign_extend, indexed, update): New.
384 (cell_micro): Adjust.
385 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
386 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
387 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
388 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
389 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
390 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
391 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
392 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
393 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
394 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
395 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
396 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
397 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
398 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
399 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
401 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
402 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
403 *vsx_extract_<mode>_store): Adjust.
404 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
405 is_cracked_insn, insn_must_be_first_in_group,
406 insn_must_be_last_in_group): Adjust.
408 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
410 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
411 ppc440-fpstore): Adjust.
412 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
413 ppc476-fpstore): Adjust.
414 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
415 ppc601-fpstore): Adjust.
416 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
418 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
420 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
421 ppc7450-fpstore): Adjust.
422 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
423 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
424 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
426 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
427 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
428 cell-fpstore, cell-fpstore-update): Adjust.
429 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
430 ppce300c3_store, ppce300c3_fpstore): Adjust.
431 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
432 e500mc_fpstore): Adjust.
433 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
434 e500mc64_store, e500mc64_fpstore): Adjust.
435 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
436 e5500_fpstore): Adjust.
437 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
438 e6500_fpstore): Adjust.
439 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
441 * config/rs6000/power4.md (power4-load, power4-load-ext,
442 power4-load-ext-update, power4-load-ext-update-indexed,
443 power4-load-update-indexed, power4-load-update, power4-fpload,
444 power4-fpload-update, power4-store, power4-store-update,
445 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
447 * config/rs6000/power5.md (power5-load, power5-load-ext,
448 power5-load-ext-update, power5-load-ext-update-indexed,
449 power5-load-update-indexed, power5-load-update, power5-fpload,
450 power5-fpload-update, power5-store, power5-store-update,
451 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
453 * config/rs6000/power6.md (power6-load, power6-load-ext,
454 power6-load-update, power6-load-update-indexed,
455 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
456 power6-fpload-update, power6-store, power6-store-update,
457 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
459 * config/rs6000/power7.md (power7-load, power7-load-ext,
460 power7-load-update, power7-load-update-indexed,
461 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
462 power7-fpload-update, power7-store, power7-store-update,
463 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
465 * config/rs6000/power8.md (power8-load, power8-load-update,
466 power8-load-ext, power8-load-ext-update, power8-fpload,
467 power8-fpload-update, power8-store, power8-store-update-indexed,
468 power8-fpstore, power8-fpstore-update): Adjust.
469 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
471 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
472 titan_lsu_store, titan_lsu_fpstore): Adjust.
473 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
475 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
478 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
479 unrolled byte insns. Emit address increments after move insns.
481 2014-05-07 David Malcolm <dmalcolm@redhat.com>
483 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
484 const_gimple, rather than a gimple.
485 (gimple_call_builtin_p): Likewise, for the three variants.
487 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
488 (gimple_call_builtin_p): Likewise, for the three variants.
490 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
492 PR tree-optimization/61095
493 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
495 2014-05-07 Richard Biener <rguenther@suse.de>
497 PR tree-optimization/61034
498 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
499 (maybe_skip_until): Use translate to take into account
500 lattices when trying to do disambiguations.
501 (get_continuation_for_phi_1): Likewise.
502 (get_continuation_for_phi): Adjust for added translate arguments.
503 (walk_non_aliased_vuses): Likewise.
504 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
505 (walk_non_aliased_vuses): Likewise.
506 (call_may_clobber_ref_p_1): Declare.
507 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
508 calls. Stop early if we are only supposed to disambiguate.
509 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
511 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
513 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
514 Emit an error when the function has arguments.
516 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
518 * cfgloop.h (unswitch_loops): Remove.
519 * doc/passes.texi: Remove references to loop-unswitch.c
520 * timevar.def (TV_LOOP_UNSWITCH): Remove.
522 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
524 * tree-vect-data-refs.c (vect_grouped_load_supported): New
525 check for loads group of length 3.
526 (vect_permute_load_chain): New permutations for loads group of
528 * tree-vect-stmts.c (vect_model_load_cost): Change cost
529 of vec_perm_shuffle for the new permutations.
531 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
533 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
534 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
535 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
536 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
537 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
538 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
539 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
540 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
542 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
544 * loop-unswitch.c: Delete.
546 2014-05-07 Richard Biener <rguenther@suse.de>
548 * config.gcc: Always set need_64bit_hwint to yes.
550 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
552 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
553 of using optimize_size.
555 2014-05-06 Mike Stump <mikestump@comcast.net>
557 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
559 2014-05-06 Joseph Myers <joseph@codesourcery.com>
561 * config/i386/sse.md (*mov<mode>_internal)
562 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
563 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
564 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
565 (*<code><mode>3, *andnot<mode>3<mask_name>)
566 (<mask_codefor><code><mode>3<mask_name>): Only consider
567 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
569 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
572 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
574 * lra-constraints.c (valid_address_p): Move earlier in file.
575 Add a constraint argument to the address_info version.
576 (satisfies_memory_constraint_p): New function.
577 (satisfies_address_constraint_p): Likewise.
578 (process_alt_operands, curr_insn_transform): Use them.
579 (process_address): Pass the constraint to valid_address_p when
580 checking address operands.
582 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
584 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
585 to their respective blocks. Fix inadvertent use of "node".
587 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
589 * emit-rtl.c (init_derived_machine_modes): New functionm, split
591 (init_emit_once): ...here.
592 * rtl.h (init_derived_machine_modes): Declare.
593 * toplev.c (do_compile): Call it even if no_backend.
595 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
596 Mike Stump <mikestump@comcast.net>
597 Richard Sandiford <rdsandiford@googlemail.com>
598 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
600 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
601 (rtx_equal_for_memref_p): Update comment.
602 (adjust_offset_for_component_ref): Use wide-int interfaces.
603 * builtins.c (get_object_alignment_2): Likewise.
604 (c_readstr): Likewise.
605 (target_char_cast): Add comment.
606 (determine_block_size): Use wide-int interfaces.
607 (expand_builtin_signbit): Likewise.
608 (fold_builtin_int_roundingfn): Likewise.
609 (fold_builtin_bitop): Likewise.
610 (fold_builtin_bswap): Likewise.
611 (fold_builtin_logarithm): Use signop.
612 (fold_builtin_pow): Likewise.
613 (fold_builtin_memory_op): Use wide-int interfaces.
614 (fold_builtin_object_size): Likewise.
615 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
616 nb_iterations_estimate.
617 (record_niter_bound): Use wide-int interfaces.
618 (get_estimated_loop_iterations_int): Likewise.
619 (get_estimated_loop_iterations): Likewise.
620 (get_max_loop_iterations): Likewise.
621 * cfgloop.h: Include wide-int.h.
622 (struct nb_iter_bound): Change bound to widest_int.
623 (struct loop): Change nb_iterations_upper_bound and
624 nb_iterations_estimate to widest_int.
625 (record_niter_bound): Switch to use widest_int.
626 (get_estimated_loop_iterations): Likewise.
627 (get_max_loop_iterations): Likewise.
628 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
630 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
631 * combine.c (try_combine): Likewise.
632 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
633 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
635 (aarch64_float_const_representable_p): Likewise.
636 * config/arc/arc.c: Include wide-int.h.
637 (arc_can_use_doloop_p): Use wide-int interfaces.
638 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
639 (vfp3_const_double_index): Likewise.
640 * config/avr/avr.c (avr_out_round): Likewise.
641 (avr_fold_builtin): Likewise.
642 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
643 (bfin_can_use_doloop_p): Likewise.
644 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
645 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
646 * config/i386/i386.c: Include wide-int.h.
647 (ix86_data_alignment): Use wide-int interfaces.
648 (ix86_local_alignment): Likewise.
649 (ix86_emit_swsqrtsf): Update real_from_integer.
650 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
651 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
652 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
653 (zero_constant): Likewise.
654 (input_operand): Likewise.
655 (splat_input_operand): Likewise.
656 (non_logical_cint_operand): Change const_double to const_wide_int.
657 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
658 (easy_altivec_constant): Remove comment.
659 (paired_expand_vector_init): Use CONSTANT_P.
660 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
661 (rs6000_emit_move): Update checks.
662 (rs6000_aggregate_candidate): Use wide-int interfaces.
663 (rs6000_expand_ternop_builtin): Likewise.
664 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
665 (rs6000_assemble_integer): Likewise.
666 (rs6000_hash_constant): Likewise.
667 (output_toc): Likewise.
668 (rs6000_rtx_costs): Likewise.
669 (rs6000_emit_swrsqrt); Update call to real_from_integer.
670 * config/rs6000/rs6000-c.c: Include wide-int.h.
671 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
672 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
673 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
674 Handle CONST_WIDE_INT.
675 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
676 Use tree_fits_uhwi_p.
677 * config/sparc/sparc.c: Include wide-int.h.
678 (sparc_fold_builtin): Use wide-int interfaces.
679 * config/vax/vax.c: Include wide-int.h.
680 (vax_float_literal): Use real_from_integer.
681 * coretypes.h (struct hwivec_def): New.
684 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
685 (equiv_constant): Handle CONST_WIDE_INT.
686 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
687 (cselib_hash_rtx): Handle CONST_WIDE_INT.
688 * dbxout.c (stabstr_U): Use wide-int interfaces.
689 (dbxout_type): Update to use cst_fits_shwi_p.
690 * defaults.h (LOG2_BITS_PER_UNIT): Define.
691 (TARGET_SUPPORTS_WIDE_INT): Add default.
692 * dfp.c: Include wide-int.h.
693 (decimal_real_to_integer2): Use wide-int interfaces and rename to
694 decimal_real_to_integer.
695 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
696 decimal_real_to_integer.
697 * doc/generic.texi (Constant expressions): Update for wide_int.
698 * doc/rtl.texi (const_double): Likewise.
699 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
700 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
701 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
702 (REAL_VALUE_FROM_INT): Remove.
703 (TARGET_SUPPORTS_WIDE_INT): New.
704 * doc/tm.texi: Regenerate.
705 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
706 * double-int.h: Include wide-int.h.
707 (struct wi::int_traits): New.
708 * dwarf2out.c (get_full_len): New.
709 (dw_val_equal_p): Add case dw_val_class_wide_int.
710 (size_of_loc_descr): Likewise.
711 (output_loc_operands): Likewise.
712 (insert_double): Remove.
713 (insert_wide_int): New.
715 (print_die): Add case dw_val_class_wide_int.
716 (attr_checksum): Likewise.
717 (attr_checksum_ordered): Likewise.
718 (same_dw_val_p): Likewise.
719 (size_of_die): Likewise.
720 (value_format): Likewise.
721 (output_die): Likewise.
722 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
724 (clz_loc_descriptor): Use wide-int interfaces.
725 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
726 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
727 (round_up_to_align): Use wide-int interfaces.
728 (field_byte_offset): Likewise.
729 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
730 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
731 CONST_DOUBLE handling. Use wide-int interfaces.
732 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
733 (gen_enumeration_type_die): Use add_AT_wide.
734 (hash_loc_operands): Add case dw_val_class_wide_int.
735 (compare_loc_operands): Likewise.
736 * dwarf2out.h: Include wide-int.h.
738 (enum dw_val_class): Add dw_val_class_wide_int.
739 (struct dw_val_struct): Add val_wide.
740 * emit-rtl.c (const_wide_int_htab): New.
741 (const_wide_int_htab_hash): New.
742 (const_wide_int_htab_eq): New.
743 (lookup_const_wide_int): New.
744 (const_double_htab_hash): Use wide-int interfaces.
745 (const_double_htab_eq): Likewise.
746 (rtx_to_double_int): Conditionally compile for wide-int.
747 (immed_double_int_const): Rename to immed_wide_int_const and
749 (immed_double_const): Conditionally compile for wide-int.
750 (init_emit_once): Use wide-int interfaces.
751 * explow.c (plus_constant): Likewise.
752 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
753 (lshift_value): Use wide-int interfaces.
754 (expand_mult): Likewise.
755 (choose_multiplier): Likewise.
756 (expand_smod_pow2): Likewise.
757 (make_tree): Likewise.
758 * expr.c (convert_modes): Consolidate handling of constants.
759 Use wide-int interfaces.
760 (emit_group_load_1): Add note.
761 (store_expr): Update comment.
762 (get_inner_reference): Use wide-int interfaces.
763 (expand_constructor): Update comment.
764 (expand_expr_real_2): Use wide-int interfaces.
765 (expand_expr_real_1): Likewise.
766 (reduce_to_bit_field_precision): Likewise.
767 (const_vector_from_tree): Likewise.
768 * final.c: Include wide-int-print.h.
769 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
770 * fixed-value.c: Include wide-int.h.
771 (fixed_from_string): Use wide-int interfaces.
772 (fixed_to_decimal): Likewise.
773 (fixed_convert_from_real): Likewise.
774 (real_convert_from_fixed): Likewise.
775 * fold-const.h (mem_ref_offset): Return an offset_int.
776 (div_if_zero_remainder): Remove code parameter.
777 * fold-const.c (div_if_zero_remainder): Remove code parameter.
778 Use wide-int interfaces.
779 (may_negate_without_overflow_p): Use wide-int interfaces.
780 (negate_expr_p): Likewise.
781 (fold_negate_expr): Likewise.
782 (int_const_binop_1): Likewise.
783 (const_binop): Likewise.
784 (fold_convert_const_int_from_int): Likewise.
785 (fold_convert_const_int_from_real): Likewise.
786 (fold_convert_const_int_from_fixed): Likewise.
787 (fold_convert_const_fixed_from_int): Likewise.
788 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
789 (sign_bit_p): Use wide-int interfaces.
790 (make_range_step): Likewise.
791 (build_range_check): Likewise. Pass an integer of the correct type
792 instead of using integer_one_node.
793 (range_predecessor): Pass an integer of the correct type instead
794 of using integer_one_node.
795 (range_successor): Likewise.
796 (merge_ranges): Likewise.
797 (unextend): Use wide-int interfaces.
798 (extract_muldiv_1): Likewise.
799 (fold_div_compare): Likewise.
800 (fold_single_bit_test): Likewise.
801 (fold_sign_changed_comparison): Likewise.
802 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
803 (fold_plusminus_mult_expr): Use wide-int interfaces.
804 (native_encode_int): Likewise.
805 (native_interpret_int): Likewise.
806 (fold_unary_loc): Likewise.
807 (pointer_may_wrap_p): Likewise.
808 (size_low_cst): Likewise.
809 (mask_with_tz): Likewise.
810 (fold_binary_loc): Likewise.
811 (fold_ternary_loc): Likewise.
812 (multiple_of_p): Likewise.
813 (tree_call_nonnegative_warnv_p): Update calls to
814 tree_int_cst_min_precision and real_from_integer.
815 (fold_negate_const): Use wide-int interfaces.
816 (fold_abs_const): Likewise.
817 (fold_relational_const): Use tree_int_cst_lt.
818 (round_up_loc): Use wide-int interfaces.
819 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
820 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
821 * gengtype.c: Remove include of double-int.h.
822 (do_typedef): Use wide-int interfaces.
823 (open_base_files): Add wide-int.h.
824 (main): Add offset_int and widest_int typedefs.
825 * gengtype-lex.l: Handle "^".
826 (CXX_KEYWORD): Add "static".
827 * gengtype-parse.c (require3): New.
828 (require_template_declaration): Handle constant template arguments
829 and nested templates.
830 * gengtype-state.c: Don't include "double-int.h".
831 * genpreds.c (write_one_predicate_function): Update comment.
832 (write_tm_constrs_h): Add check for hval and lval use in
834 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
835 (add_to_sequence): Likewise.
836 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
837 and const_double_operand.
838 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
840 * gimple-fold.c (get_base_constructor): Likewise.
841 (fold_array_ctor_reference): Likewise.
842 (fold_nonarray_ctor_reference): Likewise.
843 (fold_const_aggregate_ref_1): Likewise.
844 (gimple_val_nonnegative_real_p): Likewise.
845 (gimple_fold_indirect_ref): Likewise.
846 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
847 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
848 (struct slsr_cand_d): Change index to be widest_int.
849 (struct incr_info_d): Change incr to be widest_int.
850 (alloc_cand_and_find_basis): Use wide-int interfaces.
851 (slsr_process_phi): Likewise.
852 (backtrace_base_for_ref): Likewise. Return a widest_int.
853 (restructure_reference): Take a widest_int instead of a double_int.
854 (slsr_process_ref): Use wide-int interfaces.
855 (create_mul_ssa_cand): Likewise.
856 (create_mul_imm_cand): Likewise.
857 (create_add_ssa_cand): Likewise.
858 (create_add_imm_cand): Take a widest_int instead of a double_int.
859 (slsr_process_add): Use wide-int interfaces.
860 (slsr_process_cast): Likewise.
861 (slsr_process_copy): Likewise.
862 (dump_candidate): Likewise.
863 (dump_incr_vec): Likewise.
864 (replace_ref): Likewise.
865 (cand_increment): Likewise. Return a widest_int.
866 (cand_abs_increment): Likewise.
867 (replace_mult_candidate): Take a widest_int instead of a double_int.
868 (replace_unconditional_candidate): Use wide-int interfaces.
869 (incr_vec_index): Take a widest_int instead of a double_int.
870 (create_add_on_incoming_edge): Likewise.
871 (create_phi_basis): Use wide-int interfaces.
872 (replace_conditional_candidate): Likewise.
873 (record_increment): Take a widest_int instead of a double_int.
874 (record_phi_increments): Use wide-int interfaces.
875 (phi_incr_cost): Take a widest_int instead of a double_int.
876 (lowest_cost_path): Likewise.
877 (total_savings): Likewise.
878 (analyze_increments): Use wide-int interfaces.
879 (ncd_with_phi): Take a widest_int instead of a double_int.
880 (ncd_of_cand_and_phis): Likewise.
881 (nearest_common_dominator_for_cands): Likewise.
882 (insert_initializers): Use wide-int interfaces.
883 (all_phi_incrs_profitable): Likewise.
884 (replace_one_candidate): Likewise.
885 (replace_profitable_candidates): Likewise.
886 * godump.c: Include wide-int-print.h.
887 (go_output_typedef): Use wide-int interfaces.
888 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
889 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
890 (build_loop_iteration_domains): Likewise.
891 * hooks.h: Include wide-int.h rather than double-int.h.
892 (hook_bool_dint_dint_uint_bool_true): Delete.
893 (hook_bool_wint_wint_uint_bool_true): Declare.
894 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
895 (hook_bool_wint_wint_uint_bool_true): New.
896 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
898 (ubsan_expand_si_overflow_mul_check): Likewise.
899 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
900 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
901 (get_ancestor_addr_info): Likewise.
902 (ipa_modify_call_arguments): Likewise.
903 * loop-doloop.c (doloop_modify): Likewise.
904 (doloop_optimize): Likewise.
905 * loop-iv.c (iv_number_of_iterations): Likewise.
906 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
907 (unroll_loop_constant_iterations): Likewise.
908 (decide_unroll_runtime_iterations): Likewise.
909 (unroll_loop_runtime_iterations): Likewise.
910 (decide_peel_simple): Likewise.
911 (decide_unroll_stupid): Likewise.
912 * lto-streamer-in.c (streamer_read_wi): Add.
913 (input_cfg): Use wide-int interfaces.
914 (lto_input_tree_1): Likewise.
915 * lto-streamer-out.c (streamer_write_wi): Add.
916 (hash_tree): Use wide-int interfaces.
917 (output_cfg): Likewise.
918 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
919 (GTFILES): Add wide-int.h and signop.h.
920 (TAGS): Look for .cc files too.
921 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
922 * optabs.c (expand_subword_shift): Likewise.
923 (expand_doubleword_shift): Likewise.
924 (expand_absneg_bit): Likewise.
925 (expand_copysign_absneg): Likewise.
926 (expand_copysign_bit): Likewise.
927 * postreload.c (reload_cse_simplify_set): Likewise.
928 * predict.c (predict_iv_comparison): Likewise.
929 * pretty-print.h: Include wide-int-print.h.
931 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
932 * print-tree.c: Include wide-int-print.h.
933 (print_node_brief): Use wide-int interfaces.
934 (print_node): Likewise.
935 * read-rtl.c (validate_const_wide_int): New.
936 (read_rtx_code): Add CONST_WIDE_INT case.
937 * real.c: Include wide-int.h.
938 (real_to_integer2): Delete.
939 (real_to_integer): New function, returning a wide_int.
940 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
941 (ten_to_ptwo): Update call to real_from_integer.
942 (real_digit): Likewise.
943 * real.h: Include signop.h, wide-int.h and insn-modes.h.
944 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
945 (REAL_VALUE_TO_INT): Delete.
946 (real_to_integer): Declare a wide-int form.
947 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
948 * recog.c (const_int_operand): Improve comment.
949 (const_scalar_int_operand): New.
950 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
951 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
952 (split_double): Likewise.
953 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
954 (rtx_size): Likewise.
955 (rtx_alloc_stat_v): New.
956 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
957 (cwi_output_hex): New.
958 (iterative_hash_rtx): Handle CONST_WIDE_INT.
959 (cwi_check_failed_bounds): New.
960 * rtl.def (CONST_WIDE_INT): New.
961 * rtl.h: Include <utility> and wide-int.h.
962 (struct hwivec_def): New.
963 (CWI_GET_NUM_ELEM): New.
964 (CWI_PUT_NUM_ELEM): New.
965 (struct rtx_def): Add num_elem and hwiv.
966 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
967 (CASE_CONST_UNIQUE): Likewise.
968 (CASE_CONST_ANY): Likewise.
969 (CONST_SCALAR_INT_P): Likewise.
970 (CONST_WIDE_INT_P): New.
973 (cwi_check_failed_bounds): New.
976 (CONST_WIDE_INT_VEC) New.
977 (CONST_WIDE_INT_NUNITS) New.
978 (CONST_WIDE_INT_ELT) New.
979 (rtx_mode_t): New type.
980 (wi::int_traits <rtx_mode_t>): New.
982 (wi::min_value): New.
983 (wi::max_value): New.
985 (const_wide_int_alloc): New.
986 (immed_wide_int_const): New.
987 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
988 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
989 * signop.h: New file.
990 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
991 (simplify_const_unary_operation): Use wide-int interfaces.
992 (simplify_binary_operation_1): Likewise.
993 (simplify_const_binary_operation): Likewise.
994 (simplify_const_relational_operation): Likewise.
995 (simplify_immed_subreg): Likewise.
996 * stmt.c (expand_case): Likewise.
997 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
998 signop rather than a bool.
999 * stor-layout.c (layout_type): Use wide-int interfaces.
1000 (initialize_sizetypes): Update calls to
1001 set_min_and_max_values_for_integral_type.
1002 (set_min_and_max_values_for_integral_type): Take a signop rather
1003 than a bool. Use wide-int interfaces.
1004 (fixup_signed_type): Update accordingly. Remove
1005 HOST_BITS_PER_DOUBLE_INT limit.
1006 (fixup_unsigned_type): Likewise.
1007 * system.h (STATIC_CONSTANT_P): New.
1008 (STATIC_ASSERT): New.
1009 * target.def (can_use_doloop_p): Take widest_ints rather than
1011 * target.h: Include wide-int.h rather than double-int.h.
1012 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
1014 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
1015 rather than INT_CST_LT_UNSIGNED.
1016 (can_use_doloop_if_innermost): Take widest_ints rather than
1018 * tree-affine.c: Include wide-int-print.h.
1019 (double_int_ext_for_comb): Delete.
1020 (wide_int_ext_for_comb): New.
1021 (aff_combination_zero): Use wide-int interfaces.
1022 (aff_combination_const): Take a widest_int instead of a double_int.
1023 (aff_combination_elt): Use wide-int interfaces.
1024 (aff_combination_scale): Take a widest_int instead of a double_int.
1025 (aff_combination_add_elt): Likewise.
1026 (aff_combination_add_cst): Likewise.
1027 (aff_combination_add): Use wide-int interfaces.
1028 (aff_combination_convert): Likewise.
1029 (tree_to_aff_combination): Likewise.
1030 (add_elt_to_tree): Take a widest_int instead of a double_int.
1031 (aff_combination_to_tree): Use wide-int interfaces.
1032 (aff_combination_remove_elt): Likewise.
1033 (aff_combination_add_product): Take a widest_int instead of
1035 (aff_combination_mult): Use wide-int interfaces.
1036 (aff_combination_expand): Likewise.
1037 (double_int_constant_multiple_p): Delete.
1038 (wide_int_constant_multiple_p): New.
1039 (aff_combination_constant_multiple_p): Take a widest_int pointer
1040 instead of a double_int pointer.
1041 (print_aff): Use wide-int interfaces.
1042 (get_inner_reference_aff): Take a widest_int pointer
1043 instead of a double_int pointer.
1044 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
1045 * tree-affine.h: Include wide-int.h.
1046 (struct aff_comb_elt): Change type of coef to widest_int.
1047 (struct affine_tree_combination): Change type of offset to widest_int.
1048 (double_int_ext_for_comb): Delete.
1049 (wide_int_ext_for_comb): New.
1050 (aff_combination_const): Use widest_int instead of double_int.
1051 (aff_combination_scale): Likewise.
1052 (aff_combination_add_elt): Likewise.
1053 (aff_combination_constant_multiple_p): Likewise.
1054 (get_inner_reference_aff): Likewise.
1055 (aff_comb_cannot_overlap_p): Likewise.
1056 (aff_combination_zero_p): Use wide-int interfaces.
1057 * tree.c: Include tree.h.
1058 (init_ttree): Use make_int_cst.
1059 (tree_code_size): Removed code for INTEGER_CST case.
1060 (tree_size): Add INTEGER_CST case.
1061 (make_node_stat): Update comment.
1062 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
1063 (build_int_cst_type): Use wide-int interfaces.
1064 (double_int_to_tree): Likewise.
1065 (double_int_fits_to_tree_p): Delete.
1066 (force_fit_type_double): Delete.
1067 (force_fit_type): New.
1068 (int_cst_hash_hash): Use wide-int interfaces.
1069 (int_cst_hash_eq): Likewise.
1070 (build_int_cst_wide): Delete.
1071 (wide_int_to_tree): New.
1072 (cache_integer_cst): Use wide-int interfaces.
1073 (build_low_bits_mask): Likewise.
1074 (cst_and_fits_in_hwi): Likewise.
1075 (real_value_from_int_cst): Likewise.
1076 (make_int_cst_stat): New.
1077 (integer_zerop): Use wide_int interfaces.
1078 (integer_onep): Likewise.
1079 (integer_all_onesp): Likewise.
1080 (integer_pow2p): Likewise.
1081 (integer_nonzerop): Likewise.
1082 (tree_log2): Likewise.
1083 (tree_floor_log2): Likewise.
1084 (tree_ctz): Likewise.
1085 (int_size_in_bytes): Likewise.
1086 (mem_ref_offset): Return an offset_int rather than a double_int.
1087 (build_type_attribute_qual_variant): Use wide_int interfaces.
1088 (type_hash_eq): Likewise
1089 (tree_int_cst_equal): Likewise.
1090 (tree_int_cst_lt): Delete.
1091 (tree_int_cst_compare): Likewise.
1092 (tree_fits_shwi_p): Use wide_int interfaces.
1093 (tree_fits_uhwi_p): Likewise.
1094 (tree_int_cst_sign_bit): Likewise.
1095 (tree_int_cst_sgn): Likewise.
1096 (tree_int_cst_min_precision): Take a signop rather than a bool.
1097 (simple_cst_equal): Use wide_int interfaces.
1098 (compare_tree_int): Likewise.
1099 (iterative_hash_expr): Likewise.
1100 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
1102 (get_type_static_bounds): Use wide_int interfaces.
1103 (tree_int_cst_elt_check_failed): New.
1104 (build_common_tree_nodes): Reordered to set prec before filling in
1106 (int_cst_value): Check cst_and_fits_in_hwi.
1107 (widest_int_cst_value): Use wide_int interfaces.
1108 (upper_bound_in_type): Likewise.
1109 (lower_bound_in_type): Likewise.
1110 (num_ending_zeros): Likewise.
1111 (drop_tree_overflow): Likewise.
1112 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
1113 (gen_conditions_for_pow_cst_base): Likewise.
1114 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
1115 (group_case_labels_stmt): Use wide-int interfaces.
1116 (verify_gimple_assign_binary): Likewise.
1117 (print_loop): Likewise.
1118 * tree-chrec.c (tree_fold_binomial): Likewise.
1119 * tree-core.h (struct tree_base): Add int_length.
1120 (struct tree_int_cst): Change rep of value.
1121 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
1122 (dr_may_alias_p): Likewise.
1123 (max_stmt_executions_tree): Likewise.
1124 * tree.def (INTEGER_CST): Update comment.
1125 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
1126 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
1127 * tree-dump.c: Include wide-int.h and wide-int-print.h.
1128 (dequeue_and_dump): Use wide-int interfaces.
1129 * tree.h: Include wide-int.h.
1130 (NULL_TREE): Moved to earlier loc in file.
1131 (TREE_INT_CST_ELT_CHECK): New.
1132 (tree_int_cst_elt_check_failed): New.
1134 (TREE_INT_CST): Delete.
1135 (TREE_INT_CST_LOW): Use wide-int interfaces.
1136 (TREE_INT_CST_HIGH): Delete.
1137 (TREE_INT_CST_NUNITS): New.
1138 (TREE_INT_CST_EXT_NUNITS): Likewise.
1139 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
1140 (TREE_INT_CST_ELT): Likewise.
1141 (INT_CST_LT): Delete.
1142 (tree_int_cst_elt_check): New (two forms).
1143 (type_code_size): Update comment.
1144 (make_int_cst_stat, make_int_cst): New.
1145 (tree_to_double_int): Delete.
1146 (double_int_fits_to_tree_p): Delete.
1147 (force_fit_type_double): Delete.
1148 (build_int_cstu): Replace with out-of-line function.
1149 (build_int_cst_wide): Delete.
1150 (tree_int_cst_lt): Define inline.
1151 (tree_int_cst_le): New.
1152 (tree_int_cst_compare): Define inline.
1153 (tree_int_cst_min_precision): Take a signop rather than a bool.
1154 (wi::int_traits <const_tree>): New.
1155 (wi::int_traits <tree>): New.
1156 (wi::extended_tree): New.
1157 (wi::int_traits <wi::extended_tree>): New.
1158 (wi::to_widest): New.
1159 (wi::to_offset): New.
1160 (wi::fits_to_tree_p): New.
1161 (wi::min_value): New.
1162 (wi::max_value): New.
1163 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
1164 (copy_tree_body_r): Likewise.
1165 * tree-object-size.c (compute_object_offset): Likewise.
1166 (addr_object_size): Likewise.
1167 * tree-predcom.c: Include wide-int-print.h.
1168 (struct dref_d): Change type of offset to widest_int.
1169 (dump_dref): Call wide-int printer.
1170 (aff_combination_dr_offset): Use wide-int interfaces.
1171 (determine_offset): Take a widest_int pointer rather than a
1173 (split_data_refs_to_components): Use wide-int interfaces.
1174 (suitable_component_p): Likewise.
1175 (order_drefs): Likewise.
1176 (add_ref_to_chain): Likewise.
1177 (valid_initializer_p): Likewise.
1178 (determine_roots_comp): Likewise.
1179 * tree-pretty-print.c: Include wide-int-print.h.
1180 (dump_generic_node): Use wide-int interfaces.
1181 * tree-sra.c (sra_ipa_modify_expr): Likewise.
1182 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
1183 (move_fixed_address_to_symbol): Likewise.
1184 (move_hint_to_base): Likewise.
1185 (move_pointer_to_base): Likewise.
1186 (move_variant_to_index): Likewise.
1187 (most_expensive_mult_to_index): Likewise.
1188 (addr_to_parts): Likewise.
1189 (copy_ref_info): Likewise.
1190 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
1191 (indirect_refs_may_alias_p): Likewise.
1192 (stmt_kills_ref_p_1): Likewise.
1193 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
1194 * tree-ssa-ccp.c: Update comment at top of file. Include
1196 (struct prop_value_d): Change type of mask to widest_int.
1197 (extend_mask): New function.
1198 (dump_lattice_value): Use wide-int interfaces.
1199 (get_default_value): Likewise.
1200 (set_constant_value): Likewise.
1201 (set_value_varying): Likewise.
1202 (valid_lattice_transition): Likewise.
1203 (set_lattice_value): Likewise.
1204 (value_to_double_int): Delete.
1205 (value_to_wide_int): New.
1206 (get_value_from_alignment): Use wide-int interfaces.
1207 (get_value_for_expr): Likewise.
1208 (do_dbg_cnt): Likewise.
1209 (ccp_finalize): Likewise.
1210 (ccp_lattice_meet): Likewise.
1211 (bit_value_unop_1): Use widest_ints rather than double_ints.
1212 (bit_value_binop_1): Likewise.
1213 (bit_value_unop): Use wide-int interfaces.
1214 (bit_value_binop): Likewise.
1215 (bit_value_assume_aligned): Likewise.
1216 (evaluate_stmt): Likewise.
1217 (ccp_fold_stmt): Likewise.
1218 (visit_cond_stmt): Likewise.
1219 (ccp_visit_stmt): Likewise.
1220 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
1221 (constant_pointer_difference): Likewise.
1222 (associate_pointerplus): Likewise.
1223 (combine_conversions): Likewise.
1224 * tree-ssa-loop.h: Include wide-int.h.
1225 (struct tree_niter_desc): Change type of max to widest_int.
1226 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
1227 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
1228 (remove_redundant_iv_tests): Likewise.
1229 (canonicalize_loop_induction_variables): Likewise.
1230 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
1231 (constant_multiple_of): Take a widest_int pointer instead of
1232 a double_int pointer.
1233 (get_computation_aff): Use wide-int interfaces.
1234 (ptr_difference_cost): Likewise.
1235 (difference_cost): Likewise.
1236 (get_loop_invariant_expr_id): Likewise.
1237 (get_computation_cost_at): Likewise.
1238 (iv_elimination_compare_lt): Likewise.
1239 (may_eliminate_iv): Likewise.
1240 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
1241 instead of double_int.
1242 (max_loop_iterations): Likewise.
1243 (max_stmt_executions): Likewise.
1244 (estimated_stmt_executions): Likewise.
1245 * tree-ssa-loop-niter.c: Include wide-int-print.h.
1246 (split_to_var_and_offset): Use wide-int interfaces.
1247 (determine_value_range): Likewise.
1248 (bound_difference_of_offsetted_base): Likewise.
1249 (bounds_add): Take a widest_int instead of a double_int.
1250 (number_of_iterations_ne_max): Use wide-int interfaces.
1251 (number_of_iterations_ne): Likewise.
1252 (number_of_iterations_lt_to_ne): Likewise.
1253 (assert_loop_rolls_lt): Likewise.
1254 (number_of_iterations_lt): Likewise.
1255 (number_of_iterations_le): Likewise.
1256 (number_of_iterations_cond): Likewise.
1257 (number_of_iterations_exit): Likewise.
1258 (finite_loop_p): Likewise.
1259 (derive_constant_upper_bound_assign): Likewise.
1260 (derive_constant_upper_bound): Return a widest_int.
1261 (derive_constant_upper_bound_ops): Likewise.
1262 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
1263 (record_estimate): Take a widest_int rather than a double_int.
1264 (record_nonwrapping_iv): Use wide-int interfaces.
1265 (double_int_cmp): Delete.
1266 (wide_int_cmp): New.
1267 (bound_index): Take a widest_int rather than a double_int.
1268 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
1269 (maybe_lower_iteration_bound): Likewise.
1270 (estimate_numbers_of_iterations_loop): Likewise.
1271 (estimated_loop_iterations): Take a widest_int pointer than than
1272 a double_int pointer.
1273 (estimated_loop_iterations_int): Use wide-int interfaces.
1274 (max_loop_iterations): Take a widest_int pointer than than
1275 a double_int pointer.
1276 (max_loop_iterations_int): Use wide-int interfaces.
1277 (max_stmt_executions): Take a widest_int pointer than than
1278 a double_int pointer.
1279 (estimated_stmt_executions): Likewise.
1280 (n_of_executions_at_most): Use wide-int interfaces.
1281 (scev_probably_wraps_p): Likewise.
1282 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
1284 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
1286 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
1287 double_ints. Adjust for trailing_wide_ints <3> representation.
1288 (set_nonzero_bits): Likewise.
1289 (get_range_info): Return wide_ints rather than double_ints.
1290 Adjust for trailing_wide_ints <3> representation.
1291 (get_nonzero_bits): Likewise.
1292 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
1294 * tree-ssanames.h (struct range_info_def): Replace min, max and
1295 nonzero_bits with a trailing_wide_ints <3>.
1296 (set_range_info): Use wide_int_refs rather than double_ints.
1297 (set_nonzero_bits): Likewise.
1298 (get_range_info): Return wide_ints rather than double_ints.
1299 (get_nonzero_bits): Likewise.
1300 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
1301 * tree-ssa-pre.c (phi_translate_1): Likewise.
1302 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
1303 (acceptable_pow_call): Likewise.
1304 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
1306 (vn_reference_fold_indirect): Likewise.
1307 (vn_reference_maybe_forwprop_address): Likewise.
1308 (valueize_refs_1): Likewise.
1309 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
1310 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
1311 tree_int_cst_lt and tree_int_cst_le.
1312 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
1314 (streamer_alloc_tree): Likewise.
1315 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
1316 (streamer_write_tree_header): Likewise.
1317 (streamer_write_integer_cst): Likewise.
1318 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
1319 (build_constructors): Likewise.
1320 (array_value_type): Likewise.
1321 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
1322 (vect_check_gather): Likewise.
1323 * tree-vect-generic.c (build_replicated_const): Likewise.
1324 (expand_vector_divmod): Likewise.
1325 * tree-vect-loop.c (vect_transform_loop): Likewise.
1326 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
1327 (vect_do_peeling_for_alignment): Likewise.
1328 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
1329 * tree-vrp.c: Include wide-int.h.
1330 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
1331 (extract_range_from_assert): Use wide-int interfaces.
1332 (vrp_int_const_binop): Likewise.
1333 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
1334 double_int pointers.
1335 (ranges_from_anti_range): Use wide-int interfaces.
1336 (quad_int_cmp): Delete.
1337 (quad_int_pair_sort): Likewise.
1338 (extract_range_from_binary_expr_1): Use wide-int interfaces.
1339 (extract_range_from_unary_expr_1): Likewise.
1340 (adjust_range_with_scev): Likewise.
1341 (masked_increment): Take and return wide_ints rather than double_ints.
1342 (register_edge_assert_for_2): Use wide-int interfaces.
1343 (check_array_ref): Likewise.
1344 (search_for_addr_array): Likewise.
1345 (maybe_set_nonzero_bits): Likewise.
1346 (union_ranges): Pass an integer of the correct type instead of
1347 using integer_one_node.
1348 (intersect_ranges): Likewise.
1349 (simplify_truth_ops_using_ranges): Likewise.
1350 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
1351 (range_fits_type_p): Likewise.
1352 (simplify_cond_using_ranges): Likewise. Take a signop rather than
1354 (simplify_conversion_using_ranges): Use wide-int interfaces.
1355 (simplify_float_conversion_using_ranges): Likewise.
1356 (vrp_finalize): Likewise.
1357 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
1358 (gimple_stringops_transform): Likewise.
1359 * varasm.c (decode_addr_const): Likewise.
1360 (const_hash_1): Likewise.
1361 (const_rtx_hash_1): Likewise
1362 (output_constant): Likewise.
1363 (array_size_for_constructor): Likewise.
1364 (output_constructor_regular_field): Likewise.
1365 (output_constructor_bitfield): Likewise.
1366 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
1367 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
1369 * gencheck.c: Define BITS_PER_UNIT.
1372 * wide-int-print.cc: New.
1373 * wide-int-print.h: New.
1375 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1377 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
1379 2014-05-06 Richard Biener <rguenther@suse.de>
1381 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
1382 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
1383 (TODO_verify_all): Adjust.
1384 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
1385 TODO_verify_stmts and TODO_verify_rtl_sharing.
1386 * bb-reorder.c: Likewise.
1387 * cfgexpand.c: Likewise.
1388 * cprop.c: Likewise.
1390 * function.c: Likewise.
1391 * fwprop.c: Likewise.
1393 * gimple-ssa-isolate-paths.c: Likewise.
1394 * gimple-ssa-strength-reduction.c: Likewise.
1395 * ipa-split.c: Likewise.
1396 * loop-init.c: Likewise.
1397 * loop-unroll.c: Likewise.
1398 * lower-subreg.c: Likewise.
1399 * modulo-sched.c: Likewise.
1400 * postreload-gcse.c: Likewise.
1401 * predict.c: Likewise.
1402 * recog.c: Likewise.
1403 * sched-rgn.c: Likewise.
1404 * store-motion.c: Likewise.
1405 * tracer.c: Likewise.
1406 * trans-mem.c: Likewise.
1407 * tree-call-cdce.c: Likewise.
1408 * tree-cfg.c: Likewise.
1409 * tree-cfgcleanup.c: Likewise.
1410 * tree-complex.c: Likewise.
1411 * tree-eh.c: Likewise.
1412 * tree-emutls.c: Likewise.
1413 * tree-if-conv.c: Likewise.
1414 * tree-into-ssa.c: Likewise.
1415 * tree-loop-distribution.c: Likewise.
1416 * tree-object-size.c: Likewise.
1417 * tree-parloops.c: Likewise.
1418 * tree-pass.h: Likewise.
1419 * tree-sra.c: Likewise.
1420 * tree-ssa-ccp.c: Likewise.
1421 * tree-ssa-copy.c: Likewise.
1422 * tree-ssa-copyrename.c: Likewise.
1423 * tree-ssa-dce.c: Likewise.
1424 * tree-ssa-dom.c: Likewise.
1425 * tree-ssa-dse.c: Likewise.
1426 * tree-ssa-forwprop.c: Likewise.
1427 * tree-ssa-ifcombine.c: Likewise.
1428 * tree-ssa-loop-ch.c: Likewise.
1429 * tree-ssa-loop-ivcanon.c: Likewise.
1430 * tree-ssa-loop.c: Likewise.
1431 * tree-ssa-math-opts.c: Likewise.
1432 * tree-ssa-phiopt.c: Likewise.
1433 * tree-ssa-phiprop.c: Likewise.
1434 * tree-ssa-pre.c: Likewise.
1435 * tree-ssa-reassoc.c: Likewise.
1436 * tree-ssa-sink.c: Likewise.
1437 * tree-ssa-strlen.c: Likewise.
1438 * tree-ssa-tail-merge.c: Likewise.
1439 * tree-ssa-uncprop.c: Likewise.
1440 * tree-switch-conversion.c: Likewise.
1441 * tree-tailcall.c: Likewise.
1442 * tree-vect-generic.c: Likewise.
1443 * tree-vectorizer.c: Likewise.
1444 * tree-vrp.c: Likewise.
1446 * var-tracking.c: Likewise.
1447 * bt-load.c: Likewise.
1448 * cfgcleanup.c: Likewise.
1449 * combine-stack-adj.c: Likewise.
1450 * combine.c: Likewise.
1451 * compare-elim.c: Likewise.
1452 * config/epiphany/resolve-sw-modes.c: Likewise.
1453 * config/i386/i386.c: Likewise.
1454 * config/mips/mips.c: Likewise.
1455 * config/s390/s390.c: Likewise.
1456 * config/sh/sh_treg_combine.cc: Likewise.
1457 * config/sparc/sparc.c: Likewise.
1460 * final.c: Likewise.
1461 * ifcvt.c: Likewise.
1462 * mode-switching.c: Likewise.
1463 * passes.c: Likewise.
1464 * postreload.c: Likewise.
1466 * reg-stack.c: Likewise.
1467 * regcprop.c: Likewise.
1468 * regrename.c: Likewise.
1471 2014-05-06 Richard Biener <rguenther@suse.de>
1474 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
1475 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
1477 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
1480 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
1482 2014-05-05 Radovan Obradovic <robradovic@mips.com>
1483 Tom de Vries <tom@codesourcery.com>
1485 * target.def (call_fusage_contains_non_callee_clobbers): New
1487 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
1489 (@node Miscellaneous Register Hooks): New node.
1490 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
1491 * doc/tm.texi: Regenerate.
1493 2014-05-05 Marek Polacek <polacek@redhat.com>
1496 * opts.c (common_handle_option): Call error_at instead of warning_at.
1498 2014-05-05 Richard Biener <rguenther@suse.de>
1500 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
1501 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
1502 under the TODO_verify_il umbrella.
1504 2014-05-05 Richard Biener <rguenther@suse.de>
1506 * passes.c (execute_function_todo): Move TODO_verify_flow under
1507 the TODO_verify_ul umbrella.
1509 2014-05-05 Richard Biener <rguenther@suse.de>
1512 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
1513 X & CST away from a CST that is the mask of a mode.
1515 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1517 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
1518 int argument to enum machine_mode.
1519 (picochip_class_max_nregs): Ditto.
1520 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
1521 (picochip_class_max_nregs): Ditto.
1523 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1525 * target.def: Add new target hook.
1526 * doc/tm.texi: Regenerate.
1527 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
1528 * targhooks.c (default_keep_leaf_when_profiled): New function.
1530 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
1531 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
1533 2014-05-05 Bin Cheng <bin.cheng@arm.com>
1535 PR tree-optimization/60363
1536 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
1537 (copy_phi_args): New parameters. Call get_value_locus_in_path.
1538 (update_destination_phis): New parameter.
1539 (create_edge_and_update_destination_phis): Ditto.
1540 (ssa_fix_duplicate_block_edges): Pass new arguments.
1541 (thread_single_edge): Ditto.
1543 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
1545 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
1546 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
1547 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
1548 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
1549 Use RS6000_BTM_HARD_FLOAT.
1550 (BU_MISC_2): Likewise.
1551 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
1552 RS6000_BTM_HARD_FLOAT.
1553 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
1555 (rs6000_invalid_builtin): Add hard floating builtin support.
1556 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
1557 hard float builtins.
1558 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
1560 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1562 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
1563 Add missing function* argument.
1565 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1567 * lra-constraints.c (valid_address_p): Move earlier in file.
1568 Add a constraint argument to the address_info version.
1569 (satisfies_memory_constraint_p): New function.
1570 (satisfies_address_constraint_p): Likewise.
1571 (process_alt_operands, curr_insn_transform): Use them.
1572 (process_address): Pass the constraint to valid_address_p when
1573 checking address operands.
1575 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1577 * config/mips/mips.c (mips_isa_rev): New variable.
1578 (mips_set_architecture): Set it.
1579 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1581 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1582 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1583 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1584 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1585 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
1586 conditions in terms of mips_isa_rev.
1587 (mips_isa_rev): Declare.
1589 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1591 * config/sh/sh-mem.cc: Use tabs instead of spaces.
1592 (prob_unlikely, prob_likely): Make variables const.
1594 2014-05-03 Denis Chertykov <chertykov@gmail.com>
1596 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
1598 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1600 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
1602 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1604 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
1605 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
1606 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
1608 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
1609 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
1611 Replace usage of ROUND_REG with sh_round_reg.
1612 Use CEIL instead of ROUND_ADVANCE.
1614 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1617 * config/sh/sh.c: Include stdlib headers before everything else.
1619 2014-05-02 Jakub Jelinek <jakub@redhat.com>
1621 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
1622 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
1623 (gimplify_adjust_omp_clauses): Simd region is never
1624 directly nested in combined parallel. Instead, for linear
1625 with copyin/copyout, if in combined for simd loop, make decl
1626 firstprivate/lastprivate on OMP_FOR.
1627 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
1628 expand_omp_for_static_chunk): When setting endvar, also set
1629 fd->loop.v to the same value.
1631 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1633 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
1635 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
1637 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
1640 2014-05-02 Marek Polacek <polacek@redhat.com>
1642 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
1644 2014-05-02 Kito Cheng <kito@0xlab.org>
1646 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
1647 to a C expression marco.
1648 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
1649 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
1650 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
1651 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
1652 HONOR_REG_ALLOC_ORDER.
1653 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
1655 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1657 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
1659 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1661 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
1663 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
1665 * tree-if-conv.c (is_cond_scalar_reduction): New function.
1666 (convert_scalar_cond_reduction): Likewise.
1667 (predicate_scalar_phi): Add recognition and transformation
1668 of simple conditioanl reduction to be vectorizable.
1670 2014-05-01 Marek Polacek <polacek@redhat.com>
1673 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
1675 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
1677 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
1678 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
1679 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
1680 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
1681 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
1682 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
1683 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
1684 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
1686 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
1688 * config/arc/arc.opt (mlra): Move comment above option name
1689 to avoid mis-parsing as language options.
1691 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1693 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
1694 * config/sol2.h: ... here.
1695 * config/sol2-10.h: Remove.
1697 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
1698 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
1699 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
1700 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
1701 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
1702 * config/sol2.h: ... here.
1703 (SECTION_NAME_FORMAT): Don't redefine.
1704 (STARTFILE_ARCH32_SPEC): Rename to ...
1705 (STARTFILE_ARCH_SPEC): ... this.
1706 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
1707 * config/sparc/sol2.h: ... here.
1708 (SECTION_NAME_FORMAT): Don't undef.
1709 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
1710 (SUBTARGET_EXTRA_SPECS): Remove.
1711 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
1713 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
1714 (MD_STARTFILE_PREFIX): Remove.
1715 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
1716 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
1717 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
1718 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
1719 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
1720 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
1721 * config/i386/sol2.h: ... here.
1722 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
1723 * config/i386/sol2-bi.h: Remove.
1724 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
1725 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
1727 * config/i386/t-sol2-64: Rename to ...
1728 * config/i386/t-sol2: ... this.
1729 * config/sparc/t-sol2-64: Rename to ...
1730 * config/sparc/t-sol2: ... this.
1732 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
1733 sol2_tm_file_head, sol2_tm_file_tail.
1734 Include ${cpu_type}/sol2.h before sol2.h.
1736 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
1737 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
1738 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
1739 Reflect i386/t-sol2-64 renaming.
1740 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
1741 Reflect sparc/t-sol2-64 renaming.
1743 2014-04-30 Richard Biener <rguenther@suse.de>
1745 * passes.c (execute_function_todo): Move TODO_verify_stmts
1746 and TODO_verify_ssa under the TODO_verify_il umbrella.
1747 * tree-ssa.h (verify_ssa): Adjust prototype.
1748 * tree-ssa.c (verify_ssa): Add parameter to tell whether
1749 we should verify SSA operands.
1750 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
1751 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
1752 whether we should verify whether not throwing stmts have EH info.
1753 * graphite-scop-detection.c (create_sese_edges): Adjust.
1754 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
1755 * tree-eh.c (lower_try_finally_switch): Do not add the
1756 default case label twice.
1758 2014-04-30 Marek Polacek <polacek@redhat.com>
1760 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
1761 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
1762 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
1763 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
1765 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
1767 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
1768 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
1769 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
1770 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
1771 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
1772 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
1773 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
1774 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
1776 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1778 * tree-cfg.c (dump_function_to_file): Dump the return type of
1779 functions, in a line to itself before the function body, mimicking
1780 the layout of a C function.
1782 2014-04-29 Jakub Jelinek <jakub@redhat.com>
1784 PR tree-optimization/60971
1785 * tree-tailcall.c (process_assignment): Reject conversions which
1788 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
1790 * calls.c (initialize_argument_information): Always treat
1791 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
1792 (expand_call): Likewise.
1793 (emit_library_call_calue_1): Likewise.
1794 * expr.c (PUSH_ARGS_REVERSED): Do not define.
1795 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
1798 2014-04-29 Nick Clifton <nickc@redhat.com>
1800 * config/msp430/msp430.md (umulsidi): Fix typo.
1801 (mulhisi3): Enable even inside interrupt handlers.
1802 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
1803 bigger return address pushed in large mode.
1805 2014-04-29 Nick Clifton <nickc@redhat.com>
1807 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
1808 (arc_init_reg_tables): Use a machine_mode enum to iterate over
1810 * config/m32r/m32r.c (init_reg_tables): Likewise.
1811 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
1812 enum to hold the modes.
1814 2014-04-29 Richard Biener <rguenther@suse.de>
1816 * dominance.c (free_dominance_info): Add overload with
1818 (dom_info_state): Likewise.
1819 (dom_info_available_p): Likewise.
1820 * basic-block.h (free_dominance_info, dom_info_state,
1821 dom_info_available_p): Declare overloads.
1822 * passes.c (execute_function_todo): Verify that verifiers
1823 don't change dominator info state. Drop dominator info
1824 for IPA pass invocations.
1825 * cgraph.c (release_function_body): Restore asserts that
1826 dominator information is released.
1828 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
1830 * doc/invoke.texi: Fix typo.
1831 * tree-vrp.c: Fix typos.
1832 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
1834 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1836 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
1838 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1840 * config/aarch64/aarch64-builtins.c
1841 (aarch64_types_storestruct_lane_qualifiers): New.
1842 (TYPES_STORESTRUCT_LANE): Likewise.
1843 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
1844 (st3_lane): Likewise.
1845 (st4_lane): Likewise.
1846 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
1847 (vec_store_lanesci_lane<mode>): Likewise.
1848 (vec_store_lanesxi_lane<mode>): Likewise.
1849 (aarch64_st2_lane<VQ:mode>): Likewise.
1850 (aarch64_st3_lane<VQ:mode>): Likewise.
1851 (aarch64_st4_lane<VQ:mode>): Likewise.
1852 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
1853 * config/aarch64/arm_neon.h
1854 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
1855 use new macro arguments.
1856 (__ST3_LANE_FUNC): Likewise.
1857 (__ST4_LANE_FUNC): Likewise.
1858 * config/aarch64/iterators.md (V_TWO_ELEM): New.
1859 (V_THREE_ELEM): Likewise.
1860 (V_FOUR_ELEM): Likewise.
1862 2014-04-28 David Malcolm <dmalcolm@redhat.com>
1864 * doc/gimple.texi: Replace the description of the now-defunct
1865 union gimple_statement_d with a diagram showing the
1866 gimple_statement_base class hierarchy and its relationships to
1867 the GSS_ and GIMPLE_ enums.
1869 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1871 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
1872 * config/aarch64/aarch64.c
1873 (aarch64_cannot_change_mode_class): Weaken conditions.
1874 (aarch64_modes_tieable_p): New.
1875 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
1877 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
1879 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
1880 (loadsync_<mode>): Change mode.
1881 (load_quadpti, store_quadpti): New.
1882 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
1883 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
1885 2014-04-28 Martin Jambor <mjambor@suse.cz>
1887 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
1888 same alias type as the original statement.
1889 (subreplacement_assignment_data): New type.
1890 (handle_unscalarized_data_in_subtree): New type of parameter,
1891 generate new memory accesses with same alias type as the original
1893 (load_assign_lhs_subreplacements): Likewise.
1894 (sra_modify_constructor_assign): Generate new memory accesses with
1895 same alias type as the original statement.
1897 2014-04-28 Richard Biener <rguenther@suse.de>
1899 * tree-pass.h (TODO_verify_il): Define.
1900 (TODO_verify_all): Complete properly.
1901 * passes.c (execute_function_todo): Move existing loop-closed
1902 SSA verification under TODO_verify_il.
1903 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
1904 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
1905 Fix tree sharing issue.
1907 2014-04-28 Richard Biener <rguenther@suse.de>
1910 * builtins.def (DEF_C11_BUILTIN): Add.
1911 (BUILT_IN_ALIGNED_ALLOC): Likewise.
1912 * coretypes.h (enum function_class): Add function_c11_misc.
1913 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1914 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
1915 (call_may_clobber_ref_p_1): Likewise.
1916 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
1917 (mark_all_reaching_defs_necessary_1): Likewise.
1918 (propagate_necessity): Likewise.
1919 (eliminate_unnecessary_stmts): Likewise.
1920 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
1922 2014-04-28 Richard Biener <rguenther@suse.de>
1924 * tree-vrp.c (vrp_var_may_overflow): Remove.
1925 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
1926 with overflow immediately bump to one before that value and
1927 let iteration figure out overflow status.
1929 2014-04-28 Richard Biener <rguenther@suse.de>
1931 * configure.ac: Do valgrind header checks unconditionally.
1932 Add --enable-valgrind-annotations.
1933 * system.h: Guard valgrind header inclusion with
1934 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
1935 * alloc-pool.c (pool_alloc, pool_free): Use
1936 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
1937 to guard possibly dead code.
1938 * config.in: Regenerated.
1939 * configure: Likewise.
1941 2014-04-28 Jeff Law <law@redhat.com>
1943 PR tree-optimization/60902
1944 * tree-ssa-threadedge.c
1945 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
1946 over real defs when invalidating outputs from statements that do not
1947 produce useful outputs for threading.
1949 2014-04-28 Richard Biener <rguenther@suse.de>
1951 PR tree-optimization/60979
1952 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
1953 SCOPs that end in a block with a successor with abnormal
1956 2014-04-28 Richard Biener <rguenther@suse.de>
1958 * tree-pass.h (execute_pass_list): Adjust prototype.
1959 * passes.c (pass_manager::execute_early_local_passes): Adjust.
1960 (do_per_function): Change callback signature, push all actual
1961 work to the callbals.
1962 (do_per_function_toporder): Likewise.
1963 (execute_function_dump): Adjust.
1964 (execute_function_todo): Likewise.
1965 (clear_last_verified): Likewise.
1966 (verify_curr_properties): Likewise.
1967 (update_properties_after_pass): Likewise.
1968 (execute_pass_list_1): Split out from ...
1969 (execute_pass_list): ... here. Adjust.
1970 (execute_ipa_pass_list): Likewise.
1971 * cgraphunit.c (cgraph_add_new_function): Adjust.
1972 (analyze_function): Likewise.
1973 (expand_function): Likewise.
1974 * cgraph.c (release_function_body): Free dominance info
1975 here instead of asserting it was magically freed elsewhere.
1977 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
1979 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
1980 * configure: Regenerate.
1981 * config/sparc/sparc.opt (muser-mode): New option.
1982 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
1984 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
1985 * doc/invoke.texi (SPARC options): Document -muser-mode.
1987 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
1989 * cselib.c (find_slot_memmode): Delete.
1990 (cselib_hasher): Change compare_type to a struct.
1991 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
1993 (preserve_constants_and_equivs): Adjust for new compare_type.
1994 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
1995 (wrap_constant): Delete.
1996 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
1998 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
2000 * doc/install.texi (Building with profile feedback): Remove
2003 2014-04-26 Tom de Vries <tom@codesourcery.com>
2005 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
2008 2014-04-25 Cary Coutant <ccoutant@google.com>
2011 * dwarf2out.c (should_move_die_to_comdat): A type definition
2012 can contain a subprogram definition, but don't move it to a
2014 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
2015 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
2017 (clone_tree_hash): Rename to...
2018 (clone_tree_partial): ...this; change callers. Copy
2019 DW_TAG_subprogram DIEs as declarations.
2020 (copy_decls_walk): Don't copy children of a declaration into a
2023 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
2026 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
2029 2014-04-25 Jiong Wang <jiong.wang@arm.com>
2031 * config/arm/predicates.md (call_insn_operand): Add long_call check.
2032 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
2034 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
2037 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2039 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
2041 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2043 PR tree-optimization/60930
2044 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
2045 creating a multiply candidate by folding two constant
2046 multiplicands when the result overflows.
2048 2014-04-25 Jakub Jelinek <jakub@redhat.com>
2050 PR tree-optimization/60960
2051 * tree-vect-generic.c (expand_vector_operation): Only call
2052 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
2054 2014-04-25 Tom de Vries <tom@codesourcery.com>
2056 * expr.c (clobber_reg_mode): New function.
2057 * expr.h (clobber_reg): New function.
2059 2014-04-25 Tom de Vries <tom@codesourcery.com>
2061 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
2064 2014-04-25 Radovan Obradovic <robradovic@mips.com>
2065 Tom de Vries <tom@codesourcery.com>
2067 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
2069 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
2070 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
2071 new argument to find_all_hard_reg_sets call.
2073 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2075 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
2076 Use HOST_WIDE_INT_C for mask literal.
2077 (aarch_rev16_shleft_mask_imm_p): Likewise.
2079 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
2082 * config/sparc/sparc.md (ashlsi3_extend): Delete.
2084 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
2086 PR preprocessor/56540
2087 * config/i386/i386-c.c (ix86_target_macros): Define
2088 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
2090 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2092 * configure.ac (tga_func): Remove.
2093 (LIB_TLS_SPEC): Remove.
2094 * configure: Regenerate.
2095 * config.in: Regenerate.
2096 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
2098 2014-04-25 Richard Biener <rguenther@suse.de>
2101 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
2102 call stmt use/clobber sets during stmt walk instead of
2103 walking the possibly incomplete set of caller edges.
2105 2014-04-25 Richard Biener <rguenther@suse.de>
2108 * passes.c (apply_ipa_transforms): Inline into only caller ...
2109 (execute_one_pass): ... here. Properly bring in function
2110 bodies for nodes we want to apply IPA transforms to.
2112 2014-04-24 Cong Hou <congh@google.com>
2114 PR tree-optimization/60896
2115 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
2116 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
2117 (vect_mark_pattern_stmts): Set the def type of all statements in
2118 PATTERN_DEF_SEQ as vect_internal_def.
2120 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
2122 * doc/extend.texi (PowerPC Built-in Functions): Document new
2123 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
2124 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
2126 * config/rs6000/predicates.md (const_0_to_3_operand): New
2127 predicate to match 0..3 integer constants.
2129 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
2130 to support adding miscellaneous builtin functions.
2131 (BU_DFP_MISC_2): Likewise.
2132 (BU_P7_MISC_1): Likewise.
2133 (BU_P7_MISC_2): Likewise.
2134 (BU_P8V_MISC_3): Likewise.
2135 (BU_MISC_1): Likewise.
2136 (BU_MISC_2): Likewise.
2137 (DIVWE): Add extended divide builtin functions.
2140 (DIVWEUO): Likewise.
2144 (DIVDEUO): Likewise.
2145 (DXEX): Add decimal floating-point builtin functions.
2148 (DDEDPDQ): Likewise.
2150 (DENBCDQ): Likewise.
2157 (CDTBCD): Add new BCD builtin functions.
2161 (BCDADD_LT): Likewise.
2162 (BCDADD_EQ): Likewise.
2163 (BCDADD_GT): Likewise.
2164 (BCDADD_OV): Likewise.
2166 (BCDSUB_LT): Likewise.
2167 (BCDSUB_EQ): Likewise.
2168 (BCDSUB_GT): Likewise.
2169 (BCDSUB_OV): Likewise.
2170 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
2171 (UNPACK_TD): Likewise.
2172 (PACK_TF): Likewise.
2173 (UNPACK_TF): Likewise.
2174 (UNPACK_TF_0): Likewise.
2175 (UNPACK_TF_1): Likewise.
2176 (PACK_V1TI): Likewise.
2177 (UNPACK_V1TI): Likewise.
2179 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
2180 support for decimal floating point builtin functions.
2181 (rs6000_expand_ternop_builtin): Add checks for the new builtin
2182 functions that take constant arguments.
2183 (rs6000_invalid_builtin): Add decimal floating point builtin support.
2184 (rs6000_init_builtins): Setup long double, _Decimal64, and
2185 _Decimal128 types for new builtin functions.
2186 (builtin_function_type): Set the unsigned flags appropriately for
2187 the new builtin functions.
2188 (rs6000_opt_masks): Add support for decimal floating point builtin
2191 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
2192 floating point builtin functions.
2193 (RS6000_BTM_COMMON): Likewise.
2194 (RS6000_BTI_long_double): Likewise.
2195 (RS6000_BTI_dfloat64): Likewise.
2196 (RS6000_BTI_dfloat128): Likewise.
2197 (long_double_type_internal_node): Likewise.
2198 (dfloat64_type_internal_node): Likewise.
2199 (dfloat128_type_internal_node): Likewise.
2201 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
2202 2.07 bcd arithmetic instructions.
2203 (UNSPEC_BCDSUB): Likewise.
2204 (UNSPEC_BCD_OVERFLOW): Likewise.
2205 (UNSPEC_BCD_ADD_SUB): Likewise.
2206 (bcd_add_sub): Likewise.
2207 (BCD_TEST): Likewise.
2208 (bcd<bcd_add_sub>): Likewise.
2209 (bcd<bcd_add_sub>_test): Likewise.
2210 (bcd<bcd_add_sub>_test2): Likewise.
2211 (bcd<bcd_add_sub>_<code>): Likewise.
2212 (peephole2 for combined bcd ops): Likewise.
2214 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
2215 decimal floating point builtin functions.
2216 (UNSPEC_DENBCD): Likewise.
2217 (UNSPEC_DXEX): Likewise.
2218 (UNSPEC_DIEX): Likewise.
2219 (UNSPEC_DSCLI): Likewise.
2220 (UNSPEC_DSCRI): Likewise.
2221 (D64_D128): Likewise.
2222 (dfp_suffix): Likewise.
2223 (dfp_ddedpd_<mode>): Likewise.
2224 (dfp_denbcd_<mode>): Likewise.
2225 (dfp_dxex_<mode>): Likewise.
2226 (dfp_diex_<mode>): Likewise.
2227 (dfp_dscli_<mode>): Likewise.
2228 (dfp_dscri_<mode>): Likewise.
2230 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
2232 (UNSPEC_CDTBCD): Likewise.
2233 (UNSPEC_CBCDTD): Likewise.
2234 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
2235 (UNSPEC_DIVEO): Likewise.
2236 (UNSPEC_DIVEU): Likewise.
2237 (UNSPEC_DIVEUO): Likewise.
2238 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
2239 pack/unpack 128-bit types.
2240 (UNSPEC_PACK_128BIT): Likewise.
2241 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
2242 (udiv<mode>3): Use idiv_ldiv mode attribute.
2243 (div<mode>3): Likewise.
2244 (addg6s): Add new BCD builtin functions.
2247 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
2248 (div_extend): Likewise.
2249 (div<div_extend>_<mode>"): Likewise.
2250 (FP128_64): Add support for new builtin functions to pack/unpack
2252 (unpack<mode>): Likewise.
2253 (unpacktf_0): Likewise.
2254 (unpacktf_1): Likewise.
2255 (unpack<mode>_dm): Likewise.
2256 (unpack<mode>_nodm): Likewise.
2257 (pack<mode>): Likewise.
2258 (unpackv1ti): Likewise.
2259 (packv1ti): Likewise.
2261 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
2263 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
2266 2014-04-24 Jakub Jelinek <jakub@redhat.com>
2268 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
2269 * gimplify.c (omp_is_private): Change last argument's type to int.
2270 Only diagnose lastprivate if the simd argument is 1, only diagnose
2271 linear if the simd argument is 2.
2272 (gimplify_omp_for): Adjust omp_is_private callers. When adding
2273 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
2274 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
2275 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
2276 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
2277 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2278 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
2279 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2280 * tree-nested.c (convert_nonlocal_omp_clauses,
2281 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
2283 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
2286 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
2289 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
2291 * flag-types.h (enum ivar_visibility): Add.
2293 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
2295 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
2296 function * argument.
2298 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
2300 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
2302 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2303 Tom de Vries <tom@codesourcery.com>
2305 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
2306 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
2308 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
2309 * emit-rtl.c (try_split): Same.
2311 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2312 Tom de Vries <tom@codesourcery.com>
2314 * common.opt (fuse-caller-save): New option.
2316 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
2318 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
2319 elements for big-endian.
2321 2014-04-24 Richard Biener <rguenther@suse.de>
2323 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
2324 during TER and instead use the sepops interface for expanding
2325 non-GIMPLE_SINGLE_RHS.
2327 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2329 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
2330 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
2332 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2334 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
2335 assembler 64-bit option.
2336 * configure: Regenerate.
2338 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2340 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
2341 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
2342 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
2343 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
2344 (TARGET_CRYPTO): Take TARGET_SIMD into account.
2346 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2348 * config/aarch64/aarch64-builtins.c
2349 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
2350 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
2351 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
2352 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
2354 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
2355 (Vrevsuff): New mode attribute.
2357 2014-04-24 Terry Guo <terry.guo@arm.com>
2359 * config/arm/arm.h (machine_function): Define variable
2360 after_arm_reorg here.
2361 * config/arm/arm.c (after_arm_reorg): Remove the definition.
2362 (arm_split_constant): Update the way to access variable
2365 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
2367 2014-04-23 Tom de Vries <tom@codesourcery.com>
2369 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
2371 2014-04-23 David Malcolm <dmalcolm@redhat.com>
2373 * is-a.h: Update comments to reflect the following changes to the
2374 "pointerness" of the API, making the template parameter match the
2375 return type, allowing use of is-a.h with typedefs of pointers.
2376 (is_a_helper::cast): Return a T rather then a pointer to a T, so
2377 that the return type matches the parameter to the is_a_helper.
2379 (dyn_cast): Likewise.
2381 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
2382 pointer from the is-a.h API.
2384 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
2385 (is_a_helper <cgraph_node *>::test): ...this, matching change to
2387 (is_a_helper <varpool_node>::test): Likewise, convert to...
2388 (is_a_helper <varpool_node *>::test): ...this.
2390 (varpool_first_variable): Update for removal of implicit pointer
2391 from the is-a.h API.
2392 (varpool_next_variable): Likewise.
2393 (varpool_first_static_initializer): Likewise.
2394 (varpool_next_static_initializer): Likewise.
2395 (varpool_first_defined_variable): Likewise.
2396 (varpool_next_defined_variable): Likewise.
2397 (cgraph_first_defined_function): Likewise.
2398 (cgraph_next_defined_function): Likewise.
2399 (cgraph_first_function): Likewise.
2400 (cgraph_next_function): Likewise.
2401 (cgraph_first_function_with_gimple_body): Likewise.
2402 (cgraph_next_function_with_gimple_body): Likewise.
2403 (cgraph_alias_target): Likewise.
2404 (varpool_alias_target): Likewise.
2405 (cgraph_function_or_thunk_node): Likewise.
2406 (varpool_variable_node): Likewise.
2407 (symtab_real_symbol_p): Likewise.
2408 * cgraphunit.c (referred_to_p): Likewise.
2409 (analyze_functions): Likewise.
2410 (handle_alias_pairs): Likewise.
2411 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2412 * gimple-ssa.h (gimple_vuse_op): Likewise.
2413 (gimple_vdef_op): Likewise.
2414 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
2415 * gimple.c (gimple_build_asm_1): Likewise.
2416 (gimple_build_try): Likewise.
2417 (gimple_build_resx): Likewise.
2418 (gimple_build_eh_dispatch): Likewise.
2419 (gimple_build_omp_for): Likewise.
2420 (gimple_omp_for_set_clauses): Likewise.
2422 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
2423 (is_a_helper <gimple_statement_asm *>::test): ...this.
2424 (is_a_helper <gimple_statement_bind>::test): Convert to...
2425 (is_a_helper <gimple_statement_bind *>::test): ...this.
2426 (is_a_helper <gimple_statement_call>::test): Convert to...
2427 (is_a_helper <gimple_statement_call *>::test): ...this.
2428 (is_a_helper <gimple_statement_catch>::test): Convert to...
2429 (is_a_helper <gimple_statement_catch *>::test): ...this.
2430 (is_a_helper <gimple_statement_resx>::test): Convert to...
2431 (is_a_helper <gimple_statement_resx *>::test): ...this.
2432 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
2433 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
2434 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
2435 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
2436 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
2437 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
2438 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
2439 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
2440 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
2441 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
2442 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
2443 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
2444 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
2445 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
2446 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
2447 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
2448 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
2449 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
2450 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
2451 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
2452 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
2453 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
2454 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
2455 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
2456 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
2457 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
2458 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
2459 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
2460 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
2461 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
2462 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
2463 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
2464 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
2465 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
2466 (is_a_helper <gimple_statement_phi>::test): Convert to...
2467 (is_a_helper <gimple_statement_phi *>::test): ...this.
2468 (is_a_helper <gimple_statement_transaction>::test): Convert to...
2469 (is_a_helper <gimple_statement_transaction *>::test): ...this.
2470 (is_a_helper <gimple_statement_try>::test): Convert to...
2471 (is_a_helper <gimple_statement_try *>::test): ...this.
2472 (is_a_helper <gimple_statement_wce>::test): Convert to...
2473 (is_a_helper <gimple_statement_wce *>::test): ...this.
2474 (is_a_helper <const gimple_statement_asm>::test): Convert to...
2475 (is_a_helper <const gimple_statement_asm *>::test): ...this.
2476 (is_a_helper <const gimple_statement_bind>::test): Convert to...
2477 (is_a_helper <const gimple_statement_bind *>::test): ...this.
2478 (is_a_helper <const gimple_statement_call>::test): Convert to...
2479 (is_a_helper <const gimple_statement_call *>::test): ...this.
2480 (is_a_helper <const gimple_statement_catch>::test): Convert to...
2481 (is_a_helper <const gimple_statement_catch *>::test): ...this.
2482 (is_a_helper <const gimple_statement_resx>::test): Convert to...
2483 (is_a_helper <const gimple_statement_resx *>::test): ...this.
2484 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
2485 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
2486 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
2487 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
2488 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
2490 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
2492 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
2494 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
2496 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
2497 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
2498 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
2500 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
2501 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
2503 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
2504 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
2505 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
2506 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
2507 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
2508 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
2510 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
2511 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
2512 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
2513 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
2515 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
2516 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
2517 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
2518 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
2519 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
2520 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
2521 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
2522 (is_a_helper <const gimple_statement_phi>::test): Convert to...
2523 (is_a_helper <const gimple_statement_phi *>::test): ...this.
2524 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
2525 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
2526 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
2527 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
2528 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
2529 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
2530 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
2532 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
2534 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
2535 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
2537 (gimple_use_ops): Update for removal of implicit pointer from the
2539 (gimple_set_use_ops): Likewise.
2540 (gimple_vuse): Likewise.
2541 (gimple_vdef): Likewise.
2542 (gimple_vuse_ptr): Likewise.
2543 (gimple_vdef_ptr): Likewise.
2544 (gimple_set_vuse): Likewise.
2545 (gimple_set_vdef): Likewise.
2546 (gimple_omp_return_set_lhs): Likewise.
2547 (gimple_omp_return_lhs): Likewise.
2548 (gimple_omp_return_lhs_ptr): Likewise.
2549 (gimple_call_fntype): Likewise.
2550 (gimple_call_set_fntype): Likewise.
2551 (gimple_call_set_internal_fn): Likewise.
2552 (gimple_call_use_set): Likewise.
2553 (gimple_call_clobber_set): Likewise.
2554 (gimple_bind_vars): Likewise.
2555 (gimple_bind_set_vars): Likewise.
2556 (gimple_bind_body_ptr): Likewise.
2557 (gimple_bind_set_body): Likewise.
2558 (gimple_bind_add_stmt): Likewise.
2559 (gimple_bind_block): Likewise.
2560 (gimple_bind_set_block): Likewise.
2561 (gimple_asm_ninputs): Likewise.
2562 (gimple_asm_noutputs): Likewise.
2563 (gimple_asm_nclobbers): Likewise.
2564 (gimple_asm_nlabels): Likewise.
2565 (gimple_asm_input_op): Likewise.
2566 (gimple_asm_input_op_ptr): Likewise.
2567 (gimple_asm_output_op): Likewise.
2568 (gimple_asm_output_op_ptr): Likewise.
2569 (gimple_asm_set_output_op): Likewise.
2570 (gimple_asm_clobber_op): Likewise.
2571 (gimple_asm_set_clobber_op): Likewise.
2572 (gimple_asm_label_op): Likewise.
2573 (gimple_asm_set_label_op): Likewise.
2574 (gimple_asm_string): Likewise.
2575 (gimple_catch_types): Likewise.
2576 (gimple_catch_types_ptr): Likewise.
2577 (gimple_catch_handler_ptr): Likewise.
2578 (gimple_catch_set_types): Likewise.
2579 (gimple_catch_set_handler): Likewise.
2580 (gimple_eh_filter_types): Likewise.
2581 (gimple_eh_filter_types_ptr): Likewise.
2582 (gimple_eh_filter_failure_ptr): Likewise.
2583 (gimple_eh_filter_set_types): Likewise.
2584 (gimple_eh_filter_set_failure): Likewise.
2585 (gimple_eh_must_not_throw_fndecl): Likewise.
2586 (gimple_eh_must_not_throw_set_fndecl): Likewise.
2587 (gimple_eh_else_n_body_ptr): Likewise.
2588 (gimple_eh_else_e_body_ptr): Likewise.
2589 (gimple_eh_else_set_n_body): Likewise.
2590 (gimple_eh_else_set_e_body): Likewise.
2591 (gimple_try_eval_ptr): Likewise.
2592 (gimple_try_cleanup_ptr): Likewise.
2593 (gimple_try_set_eval): Likewise.
2594 (gimple_try_set_cleanup): Likewise.
2595 (gimple_wce_cleanup_ptr): Likewise.
2596 (gimple_wce_set_cleanup): Likewise.
2597 (gimple_phi_capacity): Likewise.
2598 (gimple_phi_num_args): Likewise.
2599 (gimple_phi_result): Likewise.
2600 (gimple_phi_result_ptr): Likewise.
2601 (gimple_phi_set_result): Likewise.
2602 (gimple_phi_arg): Likewise.
2603 (gimple_phi_set_arg): Likewise.
2604 (gimple_resx_region): Likewise.
2605 (gimple_resx_set_region): Likewise.
2606 (gimple_eh_dispatch_region): Likewise.
2607 (gimple_eh_dispatch_set_region): Likewise.
2608 (gimple_omp_critical_name): Likewise.
2609 (gimple_omp_critical_name_ptr): Likewise.
2610 (gimple_omp_critical_set_name): Likewise.
2611 (gimple_omp_for_clauses): Likewise.
2612 (gimple_omp_for_clauses_ptr): Likewise.
2613 (gimple_omp_for_set_clauses): Likewise.
2614 (gimple_omp_for_collapse): Likewise.
2615 (gimple_omp_for_index): Likewise.
2616 (gimple_omp_for_index_ptr): Likewise.
2617 (gimple_omp_for_set_index): Likewise.
2618 (gimple_omp_for_initial): Likewise.
2619 (gimple_omp_for_initial_ptr): Likewise.
2620 (gimple_omp_for_set_initial): Likewise.
2621 (gimple_omp_for_final): Likewise.
2622 (gimple_omp_for_final_ptr): Likewise.
2623 (gimple_omp_for_set_final): Likewise.
2624 (gimple_omp_for_incr): Likewise.
2625 (gimple_omp_for_incr_ptr): Likewise.
2626 (gimple_omp_for_set_incr): Likewise.
2627 (gimple_omp_for_pre_body_ptr): Likewise.
2628 (gimple_omp_for_set_pre_body): Likewise.
2629 (gimple_omp_parallel_clauses): Likewise.
2630 (gimple_omp_parallel_clauses_ptr): Likewise.
2631 (gimple_omp_parallel_set_clauses): Likewise.
2632 (gimple_omp_parallel_child_fn): Likewise.
2633 (gimple_omp_parallel_child_fn_ptr): Likewise.
2634 (gimple_omp_parallel_set_child_fn): Likewise.
2635 (gimple_omp_parallel_data_arg): Likewise.
2636 (gimple_omp_parallel_data_arg_ptr): Likewise.
2637 (gimple_omp_parallel_set_data_arg): Likewise.
2638 (gimple_omp_task_clauses): Likewise.
2639 (gimple_omp_task_clauses_ptr): Likewise.
2640 (gimple_omp_task_set_clauses): Likewise.
2641 (gimple_omp_task_child_fn): Likewise.
2642 (gimple_omp_task_child_fn_ptr): Likewise.
2643 (gimple_omp_task_set_child_fn): Likewise.
2644 (gimple_omp_task_data_arg): Likewise.
2645 (gimple_omp_task_data_arg_ptr): Likewise.
2646 (gimple_omp_task_set_data_arg): Likewise.
2647 (gimple_omp_taskreg_clauses): Likewise.
2648 (gimple_omp_taskreg_clauses_ptr): Likewise.
2649 (gimple_omp_taskreg_set_clauses): Likewise.
2650 (gimple_omp_taskreg_child_fn): Likewise.
2651 (gimple_omp_taskreg_child_fn_ptr): Likewise.
2652 (gimple_omp_taskreg_set_child_fn): Likewise.
2653 (gimple_omp_taskreg_data_arg): Likewise.
2654 (gimple_omp_taskreg_data_arg_ptr): Likewise.
2655 (gimple_omp_taskreg_set_data_arg): Likewise.
2656 (gimple_omp_task_copy_fn): Likewise.
2657 (gimple_omp_task_copy_fn_ptr): Likewise.
2658 (gimple_omp_task_set_copy_fn): Likewise.
2659 (gimple_omp_task_arg_size): Likewise.
2660 (gimple_omp_task_arg_size_ptr): Likewise.
2661 (gimple_omp_task_set_arg_size): Likewise.
2662 (gimple_omp_task_arg_align): Likewise.
2663 (gimple_omp_task_arg_align_ptr): Likewise.
2664 (gimple_omp_task_set_arg_align): Likewise.
2665 (gimple_omp_single_clauses): Likewise.
2666 (gimple_omp_single_clauses_ptr): Likewise.
2667 (gimple_omp_single_set_clauses): Likewise.
2668 (gimple_omp_target_clauses): Likewise.
2669 (gimple_omp_target_clauses_ptr): Likewise.
2670 (gimple_omp_target_set_clauses): Likewise.
2671 (gimple_omp_target_child_fn): Likewise.
2672 (gimple_omp_target_child_fn_ptr): Likewise.
2673 (gimple_omp_target_set_child_fn): Likewise.
2674 (gimple_omp_target_data_arg): Likewise.
2675 (gimple_omp_target_data_arg_ptr): Likewise.
2676 (gimple_omp_target_set_data_arg): Likewise.
2677 (gimple_omp_teams_clauses): Likewise.
2678 (gimple_omp_teams_clauses_ptr): Likewise.
2679 (gimple_omp_teams_set_clauses): Likewise.
2680 (gimple_omp_sections_clauses): Likewise.
2681 (gimple_omp_sections_clauses_ptr): Likewise.
2682 (gimple_omp_sections_set_clauses): Likewise.
2683 (gimple_omp_sections_control): Likewise.
2684 (gimple_omp_sections_control_ptr): Likewise.
2685 (gimple_omp_sections_set_control): Likewise.
2686 (gimple_omp_for_set_cond): Likewise.
2687 (gimple_omp_for_cond): Likewise.
2688 (gimple_omp_atomic_store_set_val): Likewise.
2689 (gimple_omp_atomic_store_val): Likewise.
2690 (gimple_omp_atomic_store_val_ptr): Likewise.
2691 (gimple_omp_atomic_load_set_lhs): Likewise.
2692 (gimple_omp_atomic_load_lhs): Likewise.
2693 (gimple_omp_atomic_load_lhs_ptr): Likewise.
2694 (gimple_omp_atomic_load_set_rhs): Likewise.
2695 (gimple_omp_atomic_load_rhs): Likewise.
2696 (gimple_omp_atomic_load_rhs_ptr): Likewise.
2697 (gimple_omp_continue_control_def): Likewise.
2698 (gimple_omp_continue_control_def_ptr): Likewise.
2699 (gimple_omp_continue_set_control_def): Likewise.
2700 (gimple_omp_continue_control_use): Likewise.
2701 (gimple_omp_continue_control_use_ptr): Likewise.
2702 (gimple_omp_continue_set_control_use): Likewise.
2703 (gimple_transaction_body_ptr): Likewise.
2704 (gimple_transaction_label): Likewise.
2705 (gimple_transaction_label_ptr): Likewise.
2706 (gimple_transaction_set_body): Likewise.
2707 (gimple_transaction_set_label): Likewise.
2709 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
2710 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2711 * ipa-ref.c (ipa_record_reference): Likewise.
2712 * ipa-reference.c (analyze_function): Likewise.
2713 (ipa_reference_write_optimization_summary): Likewise.
2714 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
2715 (address_taken_from_non_vtable_p): Likewise.
2716 (comdat_can_be_unshared_p_1): Likewise.
2717 * lto-cgraph.c (lto_output_ref): Likewise.
2718 (add_references): Likewise.
2719 (compute_ltrans_boundary): Likewise.
2720 (output_symtab): Likewise.
2721 (input_ref): Likewise.
2722 (input_cgraph_1): Likewise.
2723 (output_cgraph_opt_summary): Likewise.
2724 * lto-streamer-out.c (lto_output): Likewise.
2725 (output_symbol_p): Likewise.
2726 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
2727 (lsei_start_function_in_partition): Likewise.
2728 (lsei_next_variable_in_partition): Likewise.
2729 (lsei_start_variable_in_partition): Likewise.
2730 * symtab.c (insert_to_assembler_name_hash): Likewise.
2731 (unlink_from_assembler_name_hash): Likewise.
2732 (symtab_unregister_node): Likewise.
2733 (symtab_remove_node): Likewise.
2734 (dump_symtab_node): Likewise.
2735 (verify_symtab_base): Likewise.
2736 (verify_symtab_node): Likewise.
2737 (symtab_make_decl_local): Likewise.
2738 (symtab_alias_ultimate_target): Likewise.
2739 (symtab_resolve_alias): Likewise.
2740 (symtab_get_symbol_partitioning_class): Likewise.
2741 * tree-phinodes.c (allocate_phi_node): Likewise.
2742 (reserve_phi_args_for_new_edge): Likewise.
2743 (remove_phi_args): Likewise.
2744 * varpool.c (varpool_node_for_asm): Likewise.
2745 (varpool_remove_unreferenced_decls): Likewise.
2747 2014-04-23 Jeff Law <law@redhat.com>
2749 PR tree-optimization/60902
2750 * tree-ssa-threadedge.c
2751 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
2752 invalidate outputs from statements that do not produce useful
2753 outputs for threading.
2755 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2757 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
2758 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
2759 machine descriptions for Stack Smashing Protector.
2761 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
2763 * aarch64.md (<optab>_rol<mode>3): New pattern.
2764 (<optab>_rolsi3_uxtw): Likewise.
2765 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
2767 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
2769 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
2770 (arm_cortex_a12_tune): Likewise.
2772 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2774 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
2776 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2778 * config/arm/arm.md (arm_rev16si2): New pattern.
2779 (arm_rev16si2_alt): Likewise.
2780 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
2782 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2784 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
2785 (rev16<mode>2_alt): Likewise.
2786 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
2787 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
2788 (aarch_rev16_shleft_mask_imm_p): Likewise.
2789 (aarch_rev16_p_1): Likewise.
2790 (aarch_rev16_p): Likewise.
2791 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
2792 (aarch_rev16_shright_mask_imm_p): Likewise.
2793 (aarch_rev16_shleft_mask_imm_p): Likewise.
2795 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2797 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
2798 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
2800 (cortex_a53_extra_costs): Likewise.
2801 (cortex_a57_extra_costs): Likewise.
2802 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
2803 (cortexa7_extra_costs): Likewise.
2804 (cortexa8_extra_costs): Likewise.
2805 (cortexa12_extra_costs): Likewise.
2806 (cortexa15_extra_costs): Likewise.
2807 (v7m_extra_costs): Likewise.
2808 (arm_new_rtx_costs): Handle BSWAP.
2810 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2812 * config/arm/arm.c (cortexa8_extra_costs): New table.
2813 (arm_cortex_a8_tune): New tuning struct.
2814 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
2816 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2818 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
2820 2014-04-23 Richard Biener <rguenther@suse.de>
2822 * Makefile.in (OBJS): Remove loop-unswitch.o.
2823 * tree-pass.h (make_pass_rtl_unswitch): Remove.
2824 * passes.def (pass_rtl_unswitch): Likewise.
2825 * loop-init.c (gate_rtl_unswitch): Likewise.
2826 (rtl_unswitch): Likewise.
2827 (pass_data_rtl_unswitch): Likewise.
2828 (pass_rtl_unswitch): Likewise.
2829 (make_pass_rtl_unswitch): Likewise.
2830 * rtl.h (reversed_condition): Likewise.
2831 (compare_and_jump_seq): Likewise.
2832 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
2834 * loop-unroll.c (compare_and_jump_seq): Likewise.
2836 2014-04-23 Richard Biener <rguenther@suse.de>
2838 PR tree-optimization/60903
2839 * tree-ssa-loop-im.c (analyze_memory_references): Remove
2840 commented code block.
2841 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
2842 loop flags to newly created BBs and edges.
2844 2014-04-23 Nick Clifton <nickc@redhat.com>
2846 * config/msp430/msp430.c (msp430_handle_option): Move function
2848 (msp430_option_override): Simplify mcu and mcpu option handling.
2849 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
2850 support for -mhwmult command line option.
2851 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
2852 -mhwmult command line option.
2853 (msp430_hwmult_enabled): Delete.
2854 (msp43o_output_labelref): Add support for -mhwmult command line option.
2855 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
2856 (umulsidi3): Likewise.
2857 * config/msp430/msp430.opt (mmcu): Add Report attribute.
2858 (mcpu, mlarge, msmall): Likewise.
2859 (mhwmult): New option.
2860 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
2862 (msp430_is_f5_mcu): Remove prototype.
2863 (msp430_use_f5_series_hwmult): Add prototype.
2864 * config/msp430/msp430-opts.h: New file.
2865 * common/config/msp430: New directory.
2866 * common/config/msp430/msp430-common.c: New file.
2867 * config.gcc (msp430): Remove target_has_targetm_common.
2868 * doc/invoke.texi: Document -mhwmult command line option.
2870 2014-04-23 Nick Clifton <nickc@redhat.com>
2872 * config/i386/cygwin.h (ENDFILE_SPEC): Include
2873 default-manifest.o if it can be found in the search path.
2874 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
2876 2014-04-23 Terry Guo <terry.guo@arm.com>
2878 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
2880 2014-04-23 Richard Biener <rguenther@suse.de>
2883 * tree-inline.c (declare_return_variable): Use mark_addressable.
2885 2014-04-23 Richard Biener <rguenther@suse.de>
2888 * loop-init.c (loop_optimizer_init): Make sure to apply
2889 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
2891 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2894 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
2896 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
2897 if flag_sanitize_undefined_trap_on_error.
2898 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
2899 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
2900 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
2901 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
2902 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
2903 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
2904 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
2905 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
2906 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
2907 * ubsan.c (ubsan_instrument_unreachable): Return
2908 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
2909 (ubsan_expand_null_ifn): Emit __builtin_trap ()
2910 if flag_sanitize_undefined_trap_on_error and
2911 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
2912 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
2913 instrument_bool_enum_load): Emit __builtin_trap () if
2914 flag_sanitize_undefined_trap_on_error and
2915 __builtin_handle_*_abort () if !flag_sanitize_recover.
2916 * doc/invoke.texi (-fsanitize-recover,
2917 -fsanitize-undefined-trap-on-error): Document.
2919 2014-04-22 Christian Bruel <christian.bruel@st.com>
2921 * config/sh/sh.md (mov<mode>): Replace movQIHI.
2922 Force immediates to SImode.
2924 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
2926 * config/nios2/nios2.md (UNSPEC_ROUND): New.
2928 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
2929 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
2930 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
2931 (nios2_fpu_insn): Add entry for round.
2932 (N2FPU_NO_ERRNO_P): Define.
2933 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
2935 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
2937 2014-04-22 Richard Henderson <rth@redhat.com>
2939 * config/aarch64/aarch64 (addti3, subti3): New expanders.
2940 (add<GPI>3_compare0): Remove leading * from name.
2941 (add<GPI>3_carryin): Likewise.
2942 (sub<GPI>3_compare0): Likewise.
2943 (sub<GPI>3_carryin): Likewise.
2944 (<su_optab>mulditi3): New expander.
2945 (multi3): New expander.
2946 (madd<GPI>): Remove leading * from name.
2948 2014-04-22 Martin Jambor <mjambor@suse.cz>
2950 * cgraphclones.c (cgraph_function_versioning): Copy
2951 ipa_transforms_to_apply instead of asserting it is empty.
2953 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
2956 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
2957 on count_exp to get mode.
2959 2014-04-22 Andrew Pinski <apinski@cavium.com>
2961 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
2962 Handle TLS for ILP32.
2963 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
2964 (tlsie_small_<mode>): this and handle PTR.
2965 (tlsie_small_sidi): New pattern.
2966 (tlsle_small): Change to an expand to handle ILP32.
2967 (tlsle_small_<mode>): New pattern.
2968 (tlsdesc_small): Rename to ...
2969 (tlsdesc_small_<mode>): this and handle PTR.
2971 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2973 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
2975 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2977 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2978 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
2979 (aarch64_types_signed_poly_qualifiers): Likewise.
2980 (aarch64_types_unsigned_signed_qualifiers): Likewise.
2981 (aarch64_types_poly_signed_qualifiers): Likewise.
2982 (TYPES_REINTERP_SS): Type macro added.
2983 (TYPES_REINTERP_SU): Likewise.
2984 (TYPES_REINTERP_SP): Likewise.
2985 (TYPES_REINTERP_US): Likewise.
2986 (TYPES_REINTERP_PS): Likewise.
2987 (aarch64_fold_builtin): New expression folding added.
2988 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
2989 Declarations removed.
2990 (REINTERP_SS): Declarations added.
2991 (REINTERP_US): Likewise.
2992 (REINTERP_PS): Likewise.
2993 (REINTERP_SU): Likewise.
2994 (REINTERP_SP): Likewise.
2995 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
2996 (vreinterpretq_p8_f64): Likewise.
2997 (vreinterpret_p16_f64): Likewise.
2998 (vreinterpretq_p16_f64): Likewise.
2999 (vreinterpret_f32_f64): Likewise.
3000 (vreinterpretq_f32_f64): Likewise.
3001 (vreinterpret_f64_f32): Likewise.
3002 (vreinterpret_f64_p8): Likewise.
3003 (vreinterpret_f64_p16): Likewise.
3004 (vreinterpret_f64_s8): Likewise.
3005 (vreinterpret_f64_s16): Likewise.
3006 (vreinterpret_f64_s32): Likewise.
3007 (vreinterpret_f64_s64): Likewise.
3008 (vreinterpret_f64_u8): Likewise.
3009 (vreinterpret_f64_u16): Likewise.
3010 (vreinterpret_f64_u32): Likewise.
3011 (vreinterpret_f64_u64): Likewise.
3012 (vreinterpretq_f64_f32): Likewise.
3013 (vreinterpretq_f64_p8): Likewise.
3014 (vreinterpretq_f64_p16): Likewise.
3015 (vreinterpretq_f64_s8): Likewise.
3016 (vreinterpretq_f64_s16): Likewise.
3017 (vreinterpretq_f64_s32): Likewise.
3018 (vreinterpretq_f64_s64): Likewise.
3019 (vreinterpretq_f64_u8): Likewise.
3020 (vreinterpretq_f64_u16): Likewise.
3021 (vreinterpretq_f64_u32): Likewise.
3022 (vreinterpretq_f64_u64): Likewise.
3023 (vreinterpret_s64_f64): Likewise.
3024 (vreinterpretq_s64_f64): Likewise.
3025 (vreinterpret_u64_f64): Likewise.
3026 (vreinterpretq_u64_f64): Likewise.
3027 (vreinterpret_s8_f64): Likewise.
3028 (vreinterpretq_s8_f64): Likewise.
3029 (vreinterpret_s16_f64): Likewise.
3030 (vreinterpretq_s16_f64): Likewise.
3031 (vreinterpret_s32_f64): Likewise.
3032 (vreinterpretq_s32_f64): Likewise.
3033 (vreinterpret_u8_f64): Likewise.
3034 (vreinterpretq_u8_f64): Likewise.
3035 (vreinterpret_u16_f64): Likewise.
3036 (vreinterpretq_u16_f64): Likewise.
3037 (vreinterpret_u32_f64): Likewise.
3038 (vreinterpretq_u32_f64): Likewise.
3040 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3042 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
3043 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
3044 (vreinterpret_p8_s8): Likewise.
3045 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
3046 (vreinterpret_p8_s16): Likewise.
3047 (vreinterpret_p8_s32): Likewise.
3048 (vreinterpret_p8_s64): Likewise.
3049 (vreinterpret_p8_f32): Likewise.
3050 (vreinterpret_p8_u8): Likewise.
3051 (vreinterpret_p8_u16): Likewise.
3052 (vreinterpret_p8_u32): Likewise.
3053 (vreinterpret_p8_u64): Likewise.
3054 (vreinterpret_p8_p16): Likewise.
3055 (vreinterpretq_p8_s8): Likewise.
3056 (vreinterpretq_p8_s16): Likewise.
3057 (vreinterpretq_p8_s32): Likewise.
3058 (vreinterpretq_p8_s64): Likewise.
3059 (vreinterpretq_p8_f32): Likewise.
3060 (vreinterpretq_p8_u8): Likewise.
3061 (vreinterpretq_p8_u16): Likewise.
3062 (vreinterpretq_p8_u32): Likewise.
3063 (vreinterpretq_p8_u64): Likewise.
3064 (vreinterpretq_p8_p16): Likewise.
3065 (vreinterpret_p16_s8): Likewise.
3066 (vreinterpret_p16_s16): Likewise.
3067 (vreinterpret_p16_s32): Likewise.
3068 (vreinterpret_p16_s64): Likewise.
3069 (vreinterpret_p16_f32): Likewise.
3070 (vreinterpret_p16_u8): Likewise.
3071 (vreinterpret_p16_u16): Likewise.
3072 (vreinterpret_p16_u32): Likewise.
3073 (vreinterpret_p16_u64): Likewise.
3074 (vreinterpret_p16_p8): Likewise.
3075 (vreinterpretq_p16_s8): Likewise.
3076 (vreinterpretq_p16_s16): Likewise.
3077 (vreinterpretq_p16_s32): Likewise.
3078 (vreinterpretq_p16_s64): Likewise.
3079 (vreinterpretq_p16_f32): Likewise.
3080 (vreinterpretq_p16_u8): Likewise.
3081 (vreinterpretq_p16_u16): Likewise.
3082 (vreinterpretq_p16_u32): Likewise.
3083 (vreinterpretq_p16_u64): Likewise.
3084 (vreinterpretq_p16_p8): Likewise.
3085 (vreinterpret_f32_s8): Likewise.
3086 (vreinterpret_f32_s16): Likewise.
3087 (vreinterpret_f32_s32): Likewise.
3088 (vreinterpret_f32_s64): Likewise.
3089 (vreinterpret_f32_u8): Likewise.
3090 (vreinterpret_f32_u16): Likewise.
3091 (vreinterpret_f32_u32): Likewise.
3092 (vreinterpret_f32_u64): Likewise.
3093 (vreinterpret_f32_p8): Likewise.
3094 (vreinterpret_f32_p16): Likewise.
3095 (vreinterpretq_f32_s8): Likewise.
3096 (vreinterpretq_f32_s16): Likewise.
3097 (vreinterpretq_f32_s32): Likewise.
3098 (vreinterpretq_f32_s64): Likewise.
3099 (vreinterpretq_f32_u8): Likewise.
3100 (vreinterpretq_f32_u16): Likewise.
3101 (vreinterpretq_f32_u32): Likewise.
3102 (vreinterpretq_f32_u64): Likewise.
3103 (vreinterpretq_f32_p8): Likewise.
3104 (vreinterpretq_f32_p16): Likewise.
3105 (vreinterpret_s64_s8): Likewise.
3106 (vreinterpret_s64_s16): Likewise.
3107 (vreinterpret_s64_s32): Likewise.
3108 (vreinterpret_s64_f32): Likewise.
3109 (vreinterpret_s64_u8): Likewise.
3110 (vreinterpret_s64_u16): Likewise.
3111 (vreinterpret_s64_u32): Likewise.
3112 (vreinterpret_s64_u64): Likewise.
3113 (vreinterpret_s64_p8): Likewise.
3114 (vreinterpret_s64_p16): Likewise.
3115 (vreinterpretq_s64_s8): Likewise.
3116 (vreinterpretq_s64_s16): Likewise.
3117 (vreinterpretq_s64_s32): Likewise.
3118 (vreinterpretq_s64_f32): Likewise.
3119 (vreinterpretq_s64_u8): Likewise.
3120 (vreinterpretq_s64_u16): Likewise.
3121 (vreinterpretq_s64_u32): Likewise.
3122 (vreinterpretq_s64_u64): Likewise.
3123 (vreinterpretq_s64_p8): Likewise.
3124 (vreinterpretq_s64_p16): Likewise.
3125 (vreinterpret_u64_s8): Likewise.
3126 (vreinterpret_u64_s16): Likewise.
3127 (vreinterpret_u64_s32): Likewise.
3128 (vreinterpret_u64_s64): Likewise.
3129 (vreinterpret_u64_f32): Likewise.
3130 (vreinterpret_u64_u8): Likewise.
3131 (vreinterpret_u64_u16): Likewise.
3132 (vreinterpret_u64_u32): Likewise.
3133 (vreinterpret_u64_p8): Likewise.
3134 (vreinterpret_u64_p16): Likewise.
3135 (vreinterpretq_u64_s8): Likewise.
3136 (vreinterpretq_u64_s16): Likewise.
3137 (vreinterpretq_u64_s32): Likewise.
3138 (vreinterpretq_u64_s64): Likewise.
3139 (vreinterpretq_u64_f32): Likewise.
3140 (vreinterpretq_u64_u8): Likewise.
3141 (vreinterpretq_u64_u16): Likewise.
3142 (vreinterpretq_u64_u32): Likewise.
3143 (vreinterpretq_u64_p8): Likewise.
3144 (vreinterpretq_u64_p16): Likewise.
3145 (vreinterpret_s8_s16): Likewise.
3146 (vreinterpret_s8_s32): Likewise.
3147 (vreinterpret_s8_s64): Likewise.
3148 (vreinterpret_s8_f32): Likewise.
3149 (vreinterpret_s8_u8): Likewise.
3150 (vreinterpret_s8_u16): Likewise.
3151 (vreinterpret_s8_u32): Likewise.
3152 (vreinterpret_s8_u64): Likewise.
3153 (vreinterpret_s8_p8): Likewise.
3154 (vreinterpret_s8_p16): Likewise.
3155 (vreinterpretq_s8_s16): Likewise.
3156 (vreinterpretq_s8_s32): Likewise.
3157 (vreinterpretq_s8_s64): Likewise.
3158 (vreinterpretq_s8_f32): Likewise.
3159 (vreinterpretq_s8_u8): Likewise.
3160 (vreinterpretq_s8_u16): Likewise.
3161 (vreinterpretq_s8_u32): Likewise.
3162 (vreinterpretq_s8_u64): Likewise.
3163 (vreinterpretq_s8_p8): Likewise.
3164 (vreinterpretq_s8_p16): Likewise.
3165 (vreinterpret_s16_s8): Likewise.
3166 (vreinterpret_s16_s32): Likewise.
3167 (vreinterpret_s16_s64): Likewise.
3168 (vreinterpret_s16_f32): Likewise.
3169 (vreinterpret_s16_u8): Likewise.
3170 (vreinterpret_s16_u16): Likewise.
3171 (vreinterpret_s16_u32): Likewise.
3172 (vreinterpret_s16_u64): Likewise.
3173 (vreinterpret_s16_p8): Likewise.
3174 (vreinterpret_s16_p16): Likewise.
3175 (vreinterpretq_s16_s8): Likewise.
3176 (vreinterpretq_s16_s32): Likewise.
3177 (vreinterpretq_s16_s64): Likewise.
3178 (vreinterpretq_s16_f32): Likewise.
3179 (vreinterpretq_s16_u8): Likewise.
3180 (vreinterpretq_s16_u16): Likewise.
3181 (vreinterpretq_s16_u32): Likewise.
3182 (vreinterpretq_s16_u64): Likewise.
3183 (vreinterpretq_s16_p8): Likewise.
3184 (vreinterpretq_s16_p16): Likewise.
3185 (vreinterpret_s32_s8): Likewise.
3186 (vreinterpret_s32_s16): Likewise.
3187 (vreinterpret_s32_s64): Likewise.
3188 (vreinterpret_s32_f32): Likewise.
3189 (vreinterpret_s32_u8): Likewise.
3190 (vreinterpret_s32_u16): Likewise.
3191 (vreinterpret_s32_u32): Likewise.
3192 (vreinterpret_s32_u64): Likewise.
3193 (vreinterpret_s32_p8): Likewise.
3194 (vreinterpret_s32_p16): Likewise.
3195 (vreinterpretq_s32_s8): Likewise.
3196 (vreinterpretq_s32_s16): Likewise.
3197 (vreinterpretq_s32_s64): Likewise.
3198 (vreinterpretq_s32_f32): Likewise.
3199 (vreinterpretq_s32_u8): Likewise.
3200 (vreinterpretq_s32_u16): Likewise.
3201 (vreinterpretq_s32_u32): Likewise.
3202 (vreinterpretq_s32_u64): Likewise.
3203 (vreinterpretq_s32_p8): Likewise.
3204 (vreinterpretq_s32_p16): Likewise.
3205 (vreinterpret_u8_s8): Likewise.
3206 (vreinterpret_u8_s16): Likewise.
3207 (vreinterpret_u8_s32): Likewise.
3208 (vreinterpret_u8_s64): Likewise.
3209 (vreinterpret_u8_f32): Likewise.
3210 (vreinterpret_u8_u16): Likewise.
3211 (vreinterpret_u8_u32): Likewise.
3212 (vreinterpret_u8_u64): Likewise.
3213 (vreinterpret_u8_p8): Likewise.
3214 (vreinterpret_u8_p16): Likewise.
3215 (vreinterpretq_u8_s8): Likewise.
3216 (vreinterpretq_u8_s16): Likewise.
3217 (vreinterpretq_u8_s32): Likewise.
3218 (vreinterpretq_u8_s64): Likewise.
3219 (vreinterpretq_u8_f32): Likewise.
3220 (vreinterpretq_u8_u16): Likewise.
3221 (vreinterpretq_u8_u32): Likewise.
3222 (vreinterpretq_u8_u64): Likewise.
3223 (vreinterpretq_u8_p8): Likewise.
3224 (vreinterpretq_u8_p16): Likewise.
3225 (vreinterpret_u16_s8): Likewise.
3226 (vreinterpret_u16_s16): Likewise.
3227 (vreinterpret_u16_s32): Likewise.
3228 (vreinterpret_u16_s64): Likewise.
3229 (vreinterpret_u16_f32): Likewise.
3230 (vreinterpret_u16_u8): Likewise.
3231 (vreinterpret_u16_u32): Likewise.
3232 (vreinterpret_u16_u64): Likewise.
3233 (vreinterpret_u16_p8): Likewise.
3234 (vreinterpret_u16_p16): Likewise.
3235 (vreinterpretq_u16_s8): Likewise.
3236 (vreinterpretq_u16_s16): Likewise.
3237 (vreinterpretq_u16_s32): Likewise.
3238 (vreinterpretq_u16_s64): Likewise.
3239 (vreinterpretq_u16_f32): Likewise.
3240 (vreinterpretq_u16_u8): Likewise.
3241 (vreinterpretq_u16_u32): Likewise.
3242 (vreinterpretq_u16_u64): Likewise.
3243 (vreinterpretq_u16_p8): Likewise.
3244 (vreinterpretq_u16_p16): Likewise.
3245 (vreinterpret_u32_s8): Likewise.
3246 (vreinterpret_u32_s16): Likewise.
3247 (vreinterpret_u32_s32): Likewise.
3248 (vreinterpret_u32_s64): Likewise.
3249 (vreinterpret_u32_f32): Likewise.
3250 (vreinterpret_u32_u8): Likewise.
3251 (vreinterpret_u32_u16): Likewise.
3252 (vreinterpret_u32_u64): Likewise.
3253 (vreinterpret_u32_p8): Likewise.
3254 (vreinterpret_u32_p16): Likewise.
3255 (vreinterpretq_u32_s8): Likewise.
3256 (vreinterpretq_u32_s16): Likewise.
3257 (vreinterpretq_u32_s32): Likewise.
3258 (vreinterpretq_u32_s64): Likewise.
3259 (vreinterpretq_u32_f32): Likewise.
3260 (vreinterpretq_u32_u8): Likewise.
3261 (vreinterpretq_u32_u16): Likewise.
3262 (vreinterpretq_u32_u64): Likewise.
3263 (vreinterpretq_u32_p8): Likewise.
3264 (vreinterpretq_u32_p16): Likewise.
3266 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3268 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
3270 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
3272 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
3273 (vqnegd_s64): Likewise.
3274 (vqabs_s64): Likewise.
3275 (vqabsd_s64): Likewise.
3277 2014-04-22 Richard Henderson <rth@redhat.com>
3279 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
3280 computation to the top of the loop.
3282 2014-04-22 Renlin <renlin.li@arm.com>
3283 Jiong Wang <jiong.wang@arm.com>
3285 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
3286 * config/aarch64/aarch64.c (aarch64_layout_frame)
3287 (aarch64_initial_elimination_offset): Likewise.
3289 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
3291 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
3294 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
3296 * machmode.h (bitwise_mode_for_mode): Declare.
3297 * stor-layout.h (bitwise_type_for_mode): Likewise.
3298 * stor-layout.c (bitwise_mode_for_mode): New function.
3299 (bitwise_type_for_mode): Likewise.
3300 * builtins.c (fold_builtin_memory_op): Use it instead of
3301 int_mode_for_mode and build_nonstandard_integer_type.
3303 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3305 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
3306 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
3307 (*-*-solaris2*): Simplify.
3308 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
3309 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
3310 *-*-solaris2.9* handling.
3312 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
3314 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
3315 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
3317 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
3318 * configure: Regenerate.
3320 * config/i386/sol2-9.h: Remove.
3322 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
3323 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
3324 Remove Solaris 9 references.
3326 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
3328 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
3329 (floatuns<GPI:mode><GPF:mode>2): Remove.
3330 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
3331 and floatuns conversions.
3332 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
3333 and floatuns conversions.
3334 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
3335 (w1,w2): New mode attributes for inequal width conversions.
3337 2014-04-22 Renlin Li <Renlin.Li@arm.com>
3339 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
3340 the output asm format.
3342 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
3344 * config/aarch64/aarch64-simd.md
3345 (aarch64_cm<optab>di): Always split.
3346 (*aarch64_cm<optab>di): New.
3347 (aarch64_cmtstdi): Always split.
3348 (*aarch64_cmtstdi): New.
3350 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3352 PR tree-optimization/60823
3353 * omp-low.c (ipa_simd_modify_function_body): Go through
3354 all SSA_NAMEs and for those refering to vector arguments
3355 which are going to be replaced adjust SSA_NAME_VAR and,
3356 if it is a default definition, change it into a non-default
3357 definition assigned at the beginning of function from new_decl.
3358 (ipa_simd_modify_stmt_ops): Rewritten.
3359 * tree-dfa.c (set_ssa_default_def): When removing default def,
3360 check for NULL loc instead of NULL *loc.
3362 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3364 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
3365 restrictions on core registers for DImode values in Thumb2.
3367 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3369 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
3370 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
3372 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3374 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
3375 (*iordi_notzesidi_di): Likewise.
3376 (*iordi_notsesidi_di): Likewise.
3378 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3380 * config/arm/arm-protos.h (tune_params): New struct members.
3381 * config/arm/arm.c: Initialise tune_params per processor.
3382 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
3383 for speed, based on new tune_params.
3385 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3387 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
3388 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
3389 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
3390 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
3391 * config/aarch64/arm_neon.h (vrnd_f64): Added.
3392 (vrnda_f64): Likewise.
3393 (vrndi_f64): Likewise.
3394 (vrndm_f64): Likewise.
3395 (vrndn_f64): Likewise.
3396 (vrndp_f64): Likewise.
3397 (vrndx_f64): Likewise.
3399 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3401 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
3402 GET_MODE_SIZE argument is enum machine_mode.
3404 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3407 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
3408 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
3410 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
3413 * asan.c (asan_emit_stack_protection): Force the base to align to
3414 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
3415 appropriate bits if STRICT_ALIGNMENT.
3416 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
3418 (expand_used_vars): Leave a space in the stack frame for alignment
3419 if STRICT_ALIGNMENT.
3421 2014-04-21 David Malcolm <dmalcolm@redhat.com>
3423 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
3425 (gimple_store_p): Likewise.
3426 (gimple_assign_load_p): Likewise.
3427 (gimple_assign_cast_p): Likewise.
3428 (gimple_clobber_p): Likewise.
3430 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
3431 rather than a gimple.
3432 (gimple_assign_cast_p): Likewise.
3434 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
3437 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
3438 If mode is DDmode and TARGET_E500_DOUBLE allow move.
3440 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
3441 more debug information for E500 if -mdebug=reg.
3443 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
3446 * config/i386/i386.c (ix86_expand_builtin)
3447 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
3448 register for target RTX.
3449 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
3451 2014-04-18 Cong Hou <congh@google.com>
3453 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
3454 the widen-mult pattern by handling two operands with different sizes,
3455 and operands whose size is smaller than half of the result type.
3457 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3459 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
3460 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
3461 (do_estimate_edge_time): Compute it.
3462 * ipa-inline.c (want_inline_small_function_p): Bypass
3463 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
3465 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3467 * ipa-inline.c (spec_rem): New static variable.
3468 (dump_overall_stats): New function.
3469 (dump_inline_stats): New function.
3471 2014-04-18 Richard Henderson <rth@redhat.com>
3473 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
3474 to GET_MODE_SIZE, not a reg_class_t.
3476 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3478 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
3479 (vsx_xxmrglw_<mode>): Likewise.
3481 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3484 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
3485 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
3486 (rs6000_init_hard_regno_mode_ok): Likewise.
3488 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
3490 * ipa-inline.c (inline_small_functions): Account only non-cold
3492 * doc/invoke.texi (inline-unit-growth): Update documentation.
3494 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
3496 * config/rs6000/rs6000.md (addti3, subti3): New.
3498 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
3501 * config/i386/i386.c (ix86_expand_clear): Remove outdated
3502 comment. Check optimize_insn_for_size_p instead of
3503 optimize_insn_for_speed_p.
3505 2014-04-17 Martin Jambor <mjambor@suse.cz>
3507 * gimple-iterator.c (gsi_start_edge): New function.
3508 * gimple-iterator.h (gsi_start_edge): Declare.
3509 * tree-sra.c (single_non_eh_succ): New function.
3510 (disqualify_ops_if_throwing_stmt): Renamed to
3511 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
3512 having one non-EH successor BB.
3513 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
3514 generate loads into replacements.
3515 (sra_modify_assign): Likewise and and also use the simple path for
3517 (sra_modify_function_body): Commit statements on edges.
3519 2014-04-17 Richard Biener <rguenther@suse.de>
3522 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
3523 comparison results and add clarifying comment.
3525 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3527 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
3528 (blank_mode): Initialize it.
3529 (emit_mode_size_inline, emit_mode_nunits_inline,
3530 emit_mode_inner_inline): New functions.
3531 (emit_insn_modes_h): Call them and surround their output with
3532 #if GCC_VERSION >= 4001 ... #endif.
3533 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
3534 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
3535 mode_* arrays if the argument is __builtin_constant_p.
3536 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
3537 is enum machine_mode.
3539 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3541 * passes.c (opt_pass::execute): Adjust.
3542 (pass_manager::execute_pass_mode_switching): Likewise.
3543 (early_local_passes::execute): Likewise.
3544 (execute_one_pass): Pass cfun to the pass's execute method.
3545 * tree-pass.h (opt_pass::execute): Add function * argument.
3546 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3547 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3548 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3549 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3550 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3551 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
3552 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
3553 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3554 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3555 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3556 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3557 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3558 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3559 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3560 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3561 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3562 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3563 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3564 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3565 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3566 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3567 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3568 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3569 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3570 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3571 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3572 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3573 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3574 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3577 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3579 * passes.c (opt_pass::gate): Take function * argument.
3580 (gate_all_early_local_passes): Merge into
3581 (early_local_passes::gate): this.
3582 (gate_all_early_optimizations): Merge into
3583 (all_early_optimizations::gate): this.
3584 (gate_all_optimizations): Mege into
3585 (all_optimizations::gate): this.
3586 (gate_all_optimizations_g): Merge into
3587 (all_optimizations_g::gate): this.
3588 (gate_rest_of_compilation): Mege into
3589 (rest_of_compilation::gate): this.
3590 (gate_postreload): Merge into
3591 (postreload::gate): this.
3592 (dump_one_pass): Pass cfun to the pass's gate method.
3593 (execute_ipa_summary_passes): Likewise.
3594 (execute_one_pass): Likewise.
3595 (ipa_write_summaries_2): Likewise.
3596 (ipa_write_optimization_summaries_1): Likewise.
3597 (ipa_read_summaries_1): Likewise.
3598 (ipa_read_optimization_summaries_1): Likewise.
3599 (execute_ipa_stmt_fixups): Likewise.
3600 * tree-pass.h (opt_pass::gate): Add function * argument.
3601 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
3602 combine-stack-adj.c, combine.c, compare-elim.c,
3603 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3604 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
3605 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
3606 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
3607 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3608 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3609 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3610 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
3611 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
3612 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
3613 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3614 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3615 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
3616 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3617 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3618 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3619 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3620 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3621 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3622 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3623 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3624 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3625 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3626 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
3627 var-tracking.c, vtable-verify.c, web.c: Adjust.
3629 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3631 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
3632 * configure: Regenerate.
3634 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3636 * passes.c (dump_one_pass): don't check pass->has_gate.
3637 (execute_ipa_summary_passes): Likewise.
3638 (execute_one_pass): Likewise.
3639 (ipa_write_summaries_2): Likewise.
3640 (ipa_write_optimization_summaries_1): Likewise.
3641 (ipa_read_optimization_summaries_1): Likewise.
3642 (execute_ipa_stmt_fixups): Likewise.
3643 * tree-pass.h (pass_data::has_gate): Remove.
3644 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3645 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3646 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3647 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3648 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3649 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
3650 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
3651 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
3652 gimple-low.c, gimple-ssa-isolate-paths.c,
3653 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
3654 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3655 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3656 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
3657 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
3658 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
3659 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
3660 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3661 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
3662 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3663 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3664 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3665 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3666 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3667 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3668 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3669 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3670 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3671 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3672 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3673 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3674 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3677 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3679 * pass_manager.h (pass_manager::register_dump_files_1): Remove
3681 * passes.c (pass_manager::register_dump_files_1): Merge into
3682 (pass_manager::register_dump_files): this, and remove its handling of
3683 properties since the pass always has the properties anyway.
3684 (pass_manager::pass_manager): Adjust.
3686 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3688 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
3689 * passes.c (pass_manager::register_dump_files_1): Remove dead code
3690 dealing with properties.
3691 (pass_manager::register_dump_files): Adjust.
3693 2014-03-20 Mark Wielaard <mjw@redhat.com>
3695 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
3696 then represent the bound as normal constant value.
3698 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3701 Forward port from 4.8 branch
3702 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
3704 * config/i386/bmiintrin.h (_blsi_u32): New.
3708 (_blsmsk_u32): Ditto.
3709 (_blsmsk_u64): Ditto.
3710 (_tzcnt_u32): Ditto.
3711 (_tzcnt_u64): Ditto.
3713 2014-04-17 Kito Cheng <kito@0xlab.org>
3715 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
3717 2014-04-17 Richard Biener <rguenther@suse.de>
3720 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
3721 boolean results for comparisons.
3723 2014-04-17 Richard Biener <rguenther@suse.de>
3725 PR tree-optimization/60836
3726 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
3727 initial PHI args to be gimple values.
3729 2014-04-17 Richard Biener <rguenther@suse.de>
3731 PR tree-optimization/60841
3732 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
3733 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
3734 of stmts to SLP build.
3735 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
3736 (vect_analyze_slp): Likewise.
3737 (vect_analyze_slp_instance): Likewise.
3738 (vect_build_slp_tree): Limit overall SLP tree growth.
3739 * tree-vectorizer.h (vect_analyze_data_refs,
3740 vect_analyze_slp): Adjust prototypes.
3742 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3744 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
3747 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3749 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
3750 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
3751 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
3752 for TARGET_SLOW_PSHUFB
3754 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3756 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
3757 * config/i386/i386.c (intel_cost): Ditto.
3759 2014-04-17 Joey Ye <joey.ye@arm.com>
3761 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
3763 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3765 * opts.c (common_handle_option): Disable -fipa-reference coorectly
3768 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3770 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
3771 (type_all_derivations_known_p): New predicate.
3772 (type_all_ctors_visible_p): New predicate.
3773 (type_possibly_instantiated_p): New predicate.
3774 (get_odr_type): Compute all_derivations_known.
3775 (dump_odr_type): Dump the flag.
3776 (maybe_record_type): Cleanup.
3777 (record_target_from_binfo): Add bases_to_consider array;
3778 record bases for types w/o instances and skip CXX destructor.
3779 (possible_polymorphic_call_targets_1): Add bases_to_consider
3780 and consider_construction parameters; check if type may have instance.
3781 (get_polymorphic_call_info): Set maybe_in_construction to true
3782 when we know nothing.
3783 (record_targets_from_bases): Skip CXX destructors; they are
3784 never called for types in construction.
3785 (possible_polymorphic_call_targets): Do not record target when
3786 type may not have instance.
3788 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3791 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
3792 external aliases alive, too.
3794 2014-04-16 Andrew Pinski <apinski@cavium.com>
3796 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
3799 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3801 * final.c (compute_alignments): Do not apply loop alignment to a block
3802 falling through to the exit.
3804 2014-04-16 Catherine Moore <clm@codesourcery.com>
3806 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
3807 Adjust constraints for microMIPS store patterns.
3809 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3811 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
3813 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3815 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
3816 (append_use): Run at -O0.
3817 (append_vdef): Likewise.
3818 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
3819 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
3821 2014-04-16 Jakub Jelinek <jakub@redhat.com>
3823 PR tree-optimization/60844
3824 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
3825 (propagate_op_to_single_use, remove_visited_stmt_chain,
3826 linearize_expr, repropagate_negates, reassociate_bb): Use it
3827 instead of gsi_remove.
3829 2014-04-16 Martin Jambor <mjambor@suse.cz>
3831 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
3832 ipa_transforms_to_apply.
3833 (cgraph_function_versioning): Assert that old_node has empty
3834 ipa_transforms_to_apply.
3835 * trans-mem.c (ipa_tm_create_version): Likewise.
3836 * tree-inline.c (tree_function_versioning): Do not duplicate
3837 ipa_transforms_to_apply.
3839 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3842 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
3844 Pass necessary as flags on 64-bit Solaris/x86.
3845 Use lowercase relocs for x86_64-*-*.
3846 * configure: Regenerate.
3848 2014-04-15 Jan Hubicka <jh@suse.cz>
3850 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
3851 (maybe_record_node, likely_target_p): Use it.
3853 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3856 Revert following patch
3858 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3861 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3862 software floating point or no floating point registers, do not
3863 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3864 in GPRs that occurs after we tested for GPRs that would never be
3867 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3868 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3869 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3870 specifically allow DDmode, since that does not use the SPE SIMD
3873 2014-03-21 Mark Wielaard <mjw@redhat.com>
3875 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
3876 as unsigned or int depending on type and value used.
3878 2014-04-15 Richard Biener <rguenther@suse.de>
3880 PR rtl-optimization/56965
3881 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
3882 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
3884 * alias.c (true_dependence_1): Do not call
3885 nonoverlapping_component_refs_p.
3886 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
3887 nonoverlapping_component_refs_p.
3888 (indirect_refs_may_alias_p): Likewise.
3890 2014-04-15 Teresa Johnson <tejohnson@google.com>
3892 * cfg.c (dump_bb_info): Fix flags check.
3893 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
3895 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3897 PR rtl-optimization/60663
3898 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
3901 2014-04-15 Richard Biener <rguenther@suse.de>
3903 * lto-streamer.h (LTO_major_version): Bump to 4.
3905 2014-04-15 Richard Biener <rguenther@suse.de>
3907 * common.opt (lto_partition_model): New enum.
3908 (flto-partition=): Merge separate options with a single with argument,
3909 add -flto-partition=one support.
3910 * flag-types.h (enum lto_partition_model): Declare.
3911 * opts.c (finish_options): Remove duplicate -flto-partition=
3913 * lto-wrapper.c (run_gcc): Adjust.
3915 2014-04-15 Richard Biener <rguenther@suse.de>
3917 * alias.c (ncr_compar): New function.
3918 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
3920 2014-04-15 Richard Biener <rguenther@suse.de>
3922 * alias.c (record_component_aliases): Do not walk BINFOs.
3924 2014-04-15 Richard Biener <rguenther@suse.de>
3926 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
3927 Add struct function argument and adjust.
3928 (find_func_aliases_for_call): Likewise.
3929 (find_func_aliases): Likewise.
3930 (find_func_clobbers): Likewise.
3931 (intra_create_variable_infos): Likewise.
3932 (compute_points_to_sets): Likewise.
3933 (ipa_pta_execute): Adjust. Do not push/pop cfun.
3935 2014-04-15 Richard Biener <rguenther@suse.de>
3937 * tree.c (iterative_hash_expr): Use enum tree_code_class
3938 to store TREE_CODE_CLASS.
3939 (tree_block): Likewise.
3940 (tree_set_block): Likewise.
3941 * tree.h (fold_build_pointer_plus_loc): Use
3942 convert_to_ptrofftype_loc.
3944 2014-04-15 Jakub Jelinek <jakub@redhat.com>
3947 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
3950 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
3952 * cfgloop.h (struct loop): Move force_vectorize down.
3953 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
3954 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
3955 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
3956 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
3957 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
3958 * tree-core.h (enum annot_expr_kind): Add new kind values.
3959 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
3960 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
3962 * tree.def (ANNOTATE_EXPR): Tweak comment.
3964 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3966 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
3969 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
3971 * tree.h (TYPE_IDENTIFIER): Declare.
3972 * tree.c (subrange_type_for_debug_p): Use it.
3973 * godump.c (go_format_type): Likewise.
3974 * dwarf2out.c (is_cxx_auto, modified_type_die,
3975 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
3976 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
3978 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3981 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
3983 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
3985 * config/i386/i386.c (examine_argument): Return bool. Return true if
3986 parameter should be passed in memory.
3987 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
3988 (construct_container): Update calls to examine_argument.
3989 (function_arg_advance_64): Ditto.
3990 (return_in_memory_32): Merge with ix86_return_in_memory.
3991 (return_in_memory_64): Ditto.
3992 (return_in_memory_ms_64): Ditto.
3994 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3996 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
3997 * coverage.c (coverage_compute_profile_id): Handle externally visible
4000 2014-04-14 Martin Jambor <mjambor@suse.cz>
4002 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
4003 DECL_DISREGARD_INLINE_LIMITS functions.
4005 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
4008 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
4010 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
4013 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
4014 optimize_insn_for_speed_p instead of
4015 optimize_function_for_speed_p.
4017 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
4019 * doc/invoke.texi (free): Document AArch64.
4021 2014-04-14 Richard Biener <rguenther@suse.de>
4023 PR tree-optimization/60042
4024 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
4025 (insert_into_preds_of_block): Do not prevent PHI insertion
4026 for REFERENCE exprs here ...
4027 (eliminate_dom_walker::before_dom_children): ... but prevent
4028 their use here under similar conditions when applied to the
4029 IL after PRE optimizations.
4031 2014-04-14 Richard Biener <rguenther@suse.de>
4033 * passes.def: Move early points-to after early SRA.
4035 2014-04-14 Richard Biener <rguenther@suse.de>
4037 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
4038 check for which sign-changes we allow when forwarding
4039 a converted value into a switch.
4041 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
4043 * stor-layout.c (place_field): Finalize non-constant offset for the
4046 2014-04-14 Richard Biener <rguenther@suse.de>
4048 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
4050 (expand_switch_using_bit_tests_p): Likewise.
4051 (process_switch): Compute and pass on speed_p based on the
4053 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
4054 optimize_bb_for_speed_p.
4056 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
4058 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
4059 * function.h (struct function): Rename has_force_vect_loops into
4060 has_force_vectorize_loops.
4061 * lto-streamer-in.c (input_cfg): Adjust for renaming.
4062 (input_struct_function_base): Likewise.
4063 * lto-streamer-out.c (output_cfg): Likewise.
4064 (output_struct_function_base): Likewise.
4065 * omp-low.c (expand_omp_simd): Likewise.
4066 * tree-cfg.c (move_sese_region_to_fn): Likewise.
4067 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
4068 (version_loop_for_if_conversion): Likewise.
4069 (tree_if_conversion): Likewise.
4070 (main_tree_if_conversion): Likewise.
4071 (gate_tree_if_conversion): Likewise.
4072 * tree-inline.c (copy_loops): Likewise.
4073 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
4074 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
4075 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
4076 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
4077 * tree-vectorizer.c (vectorize_loops): Likewise.
4078 * tree-vectorizer.h (unlimited_cost_model): Likewise.
4080 2014-04-14 Richard Biener <rguenther@suse.de>
4083 * lto-streamer-out.c (wrap_refs): New function.
4084 (lto_output): Wrap symbol references in global initializes in
4085 type-preserving MEM_REFs.
4087 2014-04-14 Christian Bruel <christian.bruel@st.com>
4089 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
4091 2014-04-14 Christian Bruel <christian.bruel@st.com>
4093 * config/sh/sh.md (setmemqi): New expand pattern.
4094 * config/sh/sh.h (CLEAR_RATIO): Define.
4095 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
4096 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
4098 2014-04-14 Richard Biener <rguenther@suse.de>
4101 * fold-const.c (negate_expr_p): Don't negate directional rounding
4103 (fold_negate_expr): Likewise.
4105 2014-04-14 Richard Biener <rguenther@suse.de>
4107 PR tree-optimization/59817
4108 PR tree-optimization/60453
4109 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
4110 recursion to catch all CHRECs in the scalar evolution and restrict
4111 the predicate for the remains appropriately.
4113 2014-04-12 Catherine Moore <clm@codesourcery.com>
4115 * config/mips/constraints.md: Add new register constraint "kb".
4116 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
4117 (*movhi_internal): Likewise.
4118 (*movqi_internal): Likewise.
4119 * config/mips/mips.h (M16_STORE_REGS): New register class.
4120 (REG_CLASS_NAMES): Add M16_STORE_REGS.
4121 (REG_CLASS_CONTENTS): Likewise.
4122 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
4124 2014-04-11 Tobias Burnus <burnus@net-b.de>
4127 * doc/invoke.texi (-Wformat-signedness): Document it.
4128 (Wformat=2): Mention that this enables -Wformat-signedness.
4130 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4132 * common/config/epiphany/epiphany-common.c
4133 (epiphany_option_optimization_table): Enable section anchors by
4134 default at -O1 or higher.
4135 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
4136 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
4137 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
4138 carries no extra cost.
4139 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
4140 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
4141 * config/epiphany/predicates.md (memclob_operand): New predicate.
4142 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
4143 Use memclob_operand predicate and X constraint for operand 3.
4145 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4147 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
4148 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
4151 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4153 PR rtl-optimization/60651
4154 * mode-switching.c (optimize_mode_switching): Make sure to emit
4155 sets of a lower numbered entity before sets of a higher numbered
4156 entity to a mode of the same or lower priority.
4157 When creating a seginfo for a basic block that starts with a code
4158 label, move the insertion point past the code label.
4159 (new_seginfo): Document and enforce requirement that
4160 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
4161 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
4162 * doc/tm.texi: Regenerate.
4164 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
4167 * config/arc/arc.c (arc_save_restore): Fix assert typo.
4169 2013-04-11 Jakub Jelinek <jakub@redhat.com>
4171 * BASE-VER: Set to 4.10.0.
4173 2014-04-11 Tobias Burnus <burnus@net-b.de>
4176 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
4177 * doc/gcc.texi (Service): Update description in the @menu
4178 * doc/invoke.texi (Option Summary): Remove misplaced and
4181 2014-04-11 Steve Ellcey <sellcey@mips.com>
4182 Jakub Jelinek <jakub@redhat.com>
4185 * expr.c (convert_move): Use emit_store_flag_force instead of
4186 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
4189 2014-04-11 Richard Biener <rguenther@suse.de>
4192 * varasm.c (assemble_alias): Avoid endless error reporting
4193 recursion by setting TREE_ASM_WRITTEN.
4195 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4197 * config/s390/s390.md: Add a splitter for NOT rtx.
4199 2014-04-11 Jakub Jelinek <jakub@redhat.com>
4201 PR rtl-optimization/60663
4202 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
4204 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
4205 Jakub Jelinek <jakub@redhat.com>
4208 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
4209 flag from decl_node to node.
4211 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4214 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
4215 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
4216 ameliorating the cases where it can be.
4218 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
4221 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4223 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4224 (loadsync_<mode>): Change mode.
4225 (load_quadpti, store_quadpti): New.
4226 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4227 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4228 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
4230 2014-04-09 Cong Hou <congh@google.com>
4233 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
4236 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4238 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
4239 instead of vnor to exploit possible fusion opportunity in the
4241 (altivec_expand_vec_perm_const_le): Likewise.
4243 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4245 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4246 (loadsync_<mode>): Change mode.
4247 (load_quadpti, store_quadpti): New.
4248 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4249 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4251 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
4254 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
4255 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
4256 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
4258 2014-04-08 Richard Biener <rguenther@suse.de>
4261 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
4262 a 64bit widest int print double-int similar to on HWI64 hosts.
4264 2014-04-08 Richard Biener <rguenther@suse.de>
4266 PR tree-optimization/60785
4267 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
4268 default defs properly.
4270 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
4272 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
4273 (Weffc++): Likewise.
4275 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
4277 * ipa-devirt.c (maybe_record_node): When node is not recorded,
4278 set completep to false rather than true.
4280 2014-04-07 Douglas B Rupp <rupp@adacore.com>
4283 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
4284 ARM_TARGET2_DWARF_FORMAT.
4286 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
4289 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
4290 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
4293 2014-04-07 Richard Biener <rguenther@suse.de>
4295 PR tree-optimization/60766
4296 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
4297 (may_eliminate_iv): Convert cand_value_at result to desired type.
4299 2014-04-07 Jason Merrill <jason@redhat.com>
4302 * common.opt (-fno-gnu-unique): Add.
4303 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
4305 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4307 * haifa-sched.c: Fix outdated function reference and minor
4308 grammar errors in introductory comment.
4310 2014-04-07 Richard Biener <rguenther@suse.de>
4313 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
4315 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
4317 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
4320 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
4321 size accounting for thunks.
4322 (pa_asm_output_mi_thunk): Use final_start_function() and
4323 final_end_function() to output function start and end directives.
4325 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4327 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
4328 device specific ISA/ feature information. Remove short_sp and
4329 errata_skip ds. Add avr_device_specific_features enum to have device
4331 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
4332 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
4333 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
4334 updated device specific info.
4335 * config/avr/avr-mcus.def: Merge device specific details to
4336 dev_attribute field.
4337 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
4339 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
4340 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
4341 assembler if RMW isa supported by current device.
4342 * config/avr/genmultilib.awk: Update as device info structure changed.
4343 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
4345 2014-04-04 Cong Hou <congh@google.com>
4347 PR tree-optimization/60656
4348 * tree-vect-stmts.c (supportable_widening_operation):
4349 Fix a bug that elements in a vector with vect_used_by_reduction
4350 property are incorrectly reordered when the operation on it is not
4351 consistant with the one in reduction operation.
4353 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
4355 PR rtl-optimization/60155
4356 * gcse.c (record_set_data): New function.
4357 (single_set_gcse): New function.
4358 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
4359 (hoist_code): Likewise.
4360 (get_pressure_class_and_nregs): Likewise.
4362 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
4364 * explow.c (probe_stack_range): Emit a final optimization blockage.
4366 2014-04-04 Anthony Green <green@moxielogic.com>
4368 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
4371 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
4374 * lto-cgraph.c (input_overwrite_node): Check that partitioning
4375 flags are set only during streaming.
4376 * ipa.c (process_references, walk_polymorphic_call_targets,
4377 symtab_remove_unreachable_nodes): Drop bodies of always inline
4378 after early inlining.
4379 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
4381 2014-04-04 Jakub Jelinek <jakub@redhat.com>
4382 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4385 * dwarf2out.c (const_ok_for_output_1): Reject expressions
4388 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4391 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
4393 (cortex_a53_fdivd): Likewise.
4395 2014-04-04 Martin Jambor <mjambor@suse.cz>
4398 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
4400 * cgraph.c (clone_of_p): Also return true if thunks match.
4401 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
4402 cgraph_function_or_thunk_node and an obsolete comment.
4403 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
4405 (build_function_decl_skip_args): Likewise.
4406 (set_new_clone_decl_and_node_flags): New function.
4407 (duplicate_thunk_for_node): Likewise.
4408 (redirect_edge_duplicating_thunks): Likewise.
4409 (cgraph_clone_node): New parameter args_to_skip, pass it to
4410 redirect_edge_duplicating_thunks which is called instead of
4411 cgraph_redirect_edge_callee.
4412 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
4413 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
4415 2014-04-04 Jeff Law <law@redhat.com>
4418 * config/arm/predicates.md (const_int_I_operand): New predicate.
4419 (const_int_M_operand): Similarly.
4420 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
4422 (insv_t2, extv_reg, extzv_t2): Likewise.
4423 (load_multiple_with_writeback): Similarly for const_int_I_operand.
4424 (pop_multiple_with_writeback_and_return): Likewise.
4425 (vfp_pop_multiple_with_writeback): Likewise
4427 2014-04-04 Richard Biener <rguenther@suse.de>
4430 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
4431 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
4433 * gimplify.h (gimple_add_tmp_var_fn): Declare.
4434 * gimplify.c (gimple_add_tmp_var_fn): New function.
4435 * gimple-expr.h (create_tmp_reg_fn): Declare.
4436 * gimple-expr.c (create_tmp_reg_fn): New function.
4437 * gimple-low.c (record_vars_into): Don't change cfun.
4438 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
4439 code generation without cfun.
4441 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
4444 * Makefile.in (install-driver): Fix shell scripting.
4446 2014-04-03 Cong Hou <congh@google.com>
4448 PR tree-optimization/60505
4449 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
4450 threshold of number of iterations below which no vectorization
4452 * tree-vect-loop.c (new_loop_vec_info):
4453 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
4454 * tree-vect-loop.c (vect_analyze_loop_operations):
4455 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
4456 * tree-vect-loop.c (vect_transform_loop):
4457 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
4458 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
4459 of iterations of the loop and see if we should build the epilogue.
4461 2014-04-03 Richard Biener <rguenther@suse.de>
4463 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
4464 (streamer_tree_cache_create): Adjust.
4465 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
4466 to allow optional nodes array.
4467 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
4468 (streamer_tree_cache_append): Likewise.
4469 (streamer_tree_cache_create): Create nodes array optionally
4470 as specified by parameter.
4471 * lto-streamer-out.c (create_output_block): Avoid maintaining
4472 the node array in the writer cache.
4473 (DFS_write_tree): Remove assertion.
4474 (produce_asm_for_decls): Free the out decl state hash table early.
4475 * lto-streamer-in.c (lto_data_in_create): Adjust for
4476 streamer_tree_cache_create prototype change.
4478 2014-04-03 Richard Biener <rguenther@suse.de>
4480 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
4481 set TREE_CHAIN to NULL_TREE.
4483 2014-04-03 Richard Biener <rguenther@suse.de>
4485 PR tree-optimization/60740
4486 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
4487 over all GIMPLE_COND operands.
4489 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
4491 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
4492 (Weffc++): Remove Scott's numbering, merge lists and reference
4495 2014-04-03 Nick Clifton <nickc@redhat.com>
4497 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
4500 2014-04-03 Martin Jambor <mjambor@suse.cz>
4502 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
4503 mention gcc_unreachable before failing.
4504 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
4507 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
4510 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
4511 inconsistent code and instead mark the context inconsistent.
4512 (possible_polymorphic_call_targets): For inconsistent contexts
4513 return empty complete list.
4515 2014-04-02 Anthony Green <green@moxielogic.com>
4517 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
4518 (extendqisi2, extendhisi2): Define.
4519 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
4520 (WCHAR_TYPE): Change to unsigned int.
4522 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4524 PR tree-optimization/60733
4525 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
4526 insertion point for PHI candidates to be the end of the feeding
4527 block for the PHI argument.
4529 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
4531 PR rtl-optimization/60650
4532 * lra-constraints.c (process_alt_operands): Decrease reject for
4533 earlyclobber matching.
4535 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4537 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
4539 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4541 * config/spu/spu.c (pad_bb): Do not crash when the last
4542 insn is CODE_FOR_blockage.
4544 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4546 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
4547 lies outside the target mode.
4549 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4552 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4553 software floating point or no floating point registers, do not
4554 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4555 in GPRs that occurs after we tested for GPRs that would never be
4558 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4559 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4560 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4561 specifically allow DDmode, since that does not use the SPE SIMD
4564 2014-04-02 Richard Biener <rguenther@suse.de>
4567 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4568 MODE_INTs. Properly use negv_optab.
4569 (expand_abs): Likewise.
4571 2014-04-02 Richard Biener <rguenther@suse.de>
4574 * Makefile.in (install-driver): Guard extra installs with special
4577 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4579 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4582 2014-04-01 Richard Henderson <rth@redhat.com>
4585 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
4586 alternative enabled before register allocation.
4588 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
4590 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
4591 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
4593 (nios2_large_got_address): Remove unneeded 'sym' parameter.
4594 (nios2_got_address): Update nios2_large_got_address call site.
4595 (nios2_delegitimize_address): New function.
4596 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
4597 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
4598 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
4600 2014-04-01 Martin Husemann <martin@duskware.de>
4602 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
4605 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
4607 PR rtl-optimization/60604
4608 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
4609 check from register_operand.
4610 (register_operand): Redefine in terms of general_operand.
4611 (nonmemory_operand): Use register_operand for the non-constant cases.
4613 2014-04-01 Richard Biener <rguenther@suse.de>
4615 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
4617 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
4619 * doc/invoke.texi (mapp-regs): Clarify.
4621 2014-03-31 Ulrich Drepper <drepper@gmail.com>
4623 * config/i386/avx512fintrin.h (__v32hi): Define type.
4624 (__v64qi): Likewise.
4625 (_mm512_set1_epi8): Define.
4626 (_mm512_set1_epi16): Define.
4627 (_mm512_set4_epi32): Define.
4628 (_mm512_set4_epi64): Define.
4629 (_mm512_set4_pd): Define.
4630 (_mm512_set4_ps): Define.
4631 (_mm512_setr4_epi64): Define.
4632 (_mm512_setr4_epi32): Define.
4633 (_mm512_setr4_pd): Define.
4634 (_mm512_setr4_ps): Define.
4635 (_mm512_setzero_epi32): Define.
4637 2014-03-31 Martin Jambor <mjambor@suse.cz>
4640 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
4641 callsite_arguments_match_p. Updated all callers. Also check types of
4642 corresponding formal parameters and actual arguments.
4643 (not_all_callers_have_enough_arguments_p) Renamed to
4644 some_callers_have_mismatched_arguments_p.
4646 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
4648 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
4650 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
4653 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
4656 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
4658 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
4659 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
4661 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
4662 Use FMAMODE_NOVF512 mode iterator.
4663 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
4664 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
4665 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
4666 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
4668 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
4669 Use VF_128_256 mode iterator.
4670 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
4673 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4675 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
4676 static chain if needed.
4678 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4681 * lra-constraints.c (index_part_to_reg): New.
4682 (process_address): Use it.
4684 2014-03-27 Jeff Law <law@redhat.com>
4685 Jakub Jelinek <jakub@redhat.com>
4688 * expr.c (do_tablejump): Use simplify_gen_binary rather than
4689 gen_rtx_{PLUS,MULT} to build up the address expression.
4691 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
4692 creating non-canonical RTL.
4694 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4697 * ipa-inline.c (want_inline_small_function_p): Short circuit large
4698 functions; reorganize to make cheap checks first.
4699 (inline_small_functions): Do not estimate growth when dumping;
4701 * ipa-inline.h (inline_summary): Add min_size.
4702 (growth_likely_positive): New function.
4703 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
4704 (set_cond_stmt_execution_predicate): Cleanup.
4705 (estimate_edge_size_and_time): Compute min_size.
4706 (estimate_calls_size_and_time): Likewise.
4707 (estimate_node_size_and_time): Likewise.
4708 (inline_update_overall_summary): Update min_size.
4709 (do_estimate_edge_time): Likewise.
4710 (do_estimate_edge_size): Update.
4711 (do_estimate_edge_hints): Update.
4712 (growth_likely_positive): New function.
4714 2014-03-28 Jakub Jelinek <jakub@redhat.com>
4717 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
4718 also if addr has VOIDmode.
4720 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4722 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
4723 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
4725 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
4726 instructions as well as AdvancedSIMD loads.
4728 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4730 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
4731 Use crypto_aese type.
4732 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
4733 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
4734 crypto_aese, crypto_aesmc. Move to types.md.
4735 * config/arm/types.md (crypto_aes): Split into crypto_aese,
4737 * config/arm/iterators.md (crypto_type): Likewise.
4739 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4741 * cgraph.c: Include expr.h and tree-dfa.h.
4742 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
4745 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4748 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
4749 regs from checking multi-reg pseudos.
4751 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4753 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
4755 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4757 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
4758 if it would clobber the stack pointer, even temporarily.
4760 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
4762 * mode-switching.c: Make small adjustments to the top comment.
4764 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4766 * config/rs6000/constraints.md (wD constraint): New constraint to
4767 match the constant integer to get the top DImode/DFmode out of a
4768 vector in a VSX register.
4770 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
4771 match the constant integer to get the top DImode/DFmode out of a
4772 vector in a VSX register.
4774 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
4777 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
4780 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
4781 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
4783 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
4784 Optimize vec_extract of 64-bit values, where the value being
4785 extracted is in the top word, where we can use scalar
4786 instructions. Add direct move and store support. Combine the big
4787 endian/little endian vector select load support into a single insn.
4788 (vsx_extract_<mode>_internal1): Likewise.
4789 (vsx_extract_<mode>_internal2): Likewise.
4790 (vsx_extract_<mode>_load): Likewise.
4791 (vsx_extract_<mode>_store): Likewise.
4792 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
4793 combined into vsx_extract_<mode>_load.
4794 (vsx_extract_<mode>_one_le): Likewise.
4796 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
4797 define the top 64-bit vector element.
4799 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
4802 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4803 Document vec_vbpermq builtin.
4806 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
4807 enable use of xxsldwi and xxpermdi builtin functions.
4808 (vec_xxpermdi): Likewise.
4810 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4811 Document use of vec_xxsldwi and vec_xxpermdi builtins.
4813 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
4815 PR rtl-optimization/60650
4816 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
4818 (find_spills_for): New.
4819 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
4820 Spill all pseudos on the second iteration.
4822 2014-03-27 Marek Polacek <polacek@redhat.com>
4825 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
4828 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4830 * config/s390/s390.c (s390_can_use_return_insn): Check for
4831 call-saved FPRs on 31 bit.
4833 2014-03-27 Jakub Jelinek <jakub@redhat.com>
4836 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
4837 if they need regimplification, just drop them instead of
4838 calling gimple_regimplify_operands on them.
4840 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
4843 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
4844 (aarch64_frame_pointer_required): Adjust logic.
4845 (aarch64_can_eliminate): Adjust logic.
4846 (aarch64_override_options_after_change): Adjust logic.
4848 2014-03-27 Dehao Chen <dehao@google.com>
4850 * ipa-inline.c (early_inliner): Update node's inline info.
4852 2014-03-26 Dehao Chen <dehao@google.com>
4854 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
4855 compiler inserted conditional jumps for NAN float check.
4857 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4859 * ubsan.h (ubsan_create_data): Change second argument's type
4860 to const location_t *.
4861 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
4863 (ubsan_create_data): Change second argument to const location_t *PLOC.
4864 Create Loc field whenever PLOC is non-NULL.
4865 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
4866 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
4870 * real.c (real_to_integer2): Change type of low to UHWI.
4872 2014-03-26 Tobias Burnus <burnus@net-b.de>
4874 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
4875 (CILK_SELF_SPECS): New define.
4876 (driver_self_specs): Use it.
4878 2014-03-26 Richard Biener <rguenther@suse.de>
4880 * tree-pretty-print.c (percent_K_format): Implement special
4881 case for LTO and its stripped down BLOCK tree.
4883 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4886 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
4888 * tree-vrp.c (simplify_internal_call_using_ranges): If only
4889 one range is range_int_cst_p, but not both, at least optimize
4890 addition/subtraction of 0 and multiplication by 0 or 1.
4891 * gimple-fold.c (gimple_fold_call): Fold
4892 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
4893 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
4894 INTEGER_CSTs, try to fold at least x * 0 and y - y.
4896 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
4898 PR rtl-optimization/60452
4899 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
4900 <case REG>: Return 1 for invalid offsets from the frame pointer.
4902 2014-03-26 Marek Polacek <polacek@redhat.com>
4905 * doc/extend.texi (C Extensions): Mention variable-length arrays in
4908 2014-03-26 Marek Polacek <polacek@redhat.com>
4911 * doc/extend.texi (Designated Inits): Describe what happens to omitted
4914 2014-03-26 Marek Polacek <polacek@redhat.com>
4917 * ira-color.c (update_conflict_hard_regno_costs): Perform the
4918 multiplication in unsigned type.
4920 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4922 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
4924 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4926 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
4928 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4931 * cif-code.def (UNREACHABLE) New code.
4932 * ipa-inline.c (inline_small_functions): Skip edges to
4933 __builtlin_unreachable.
4934 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
4935 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
4936 predicate to __bulitin_unreachable.
4937 (set_cond_stmt_execution_predicate): Fix issue when
4938 invert_tree_comparison returns ERROR_MARK.
4939 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
4940 propagate to inline clones.
4941 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
4943 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
4944 * cgraphclones.c (cgraph_clone_node): If call destination is already
4945 ureachable, do not redirect it back.
4946 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
4949 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4951 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
4952 Do not modify inline clones.
4954 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4956 * config/i386/i386.md (general_sext_operand): New mode attr.
4957 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
4958 don't generate (sign_extend (const_int)).
4959 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
4960 operands[2]. Use We constraint instead of <i> and
4961 <general_sext_operand> predicate instead of <general_operand>.
4962 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
4963 * config/i386/constraints.md (We): New constraint.
4964 * config/i386/predicates.md (x86_64_sext_operand,
4965 sext_operand): New predicates.
4967 2014-03-25 Martin Jambor <mjambor@suse.cz>
4970 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
4971 inconsistent devirtualizations to __builtin_unreachable.
4973 2014-03-25 Marek Polacek <polacek@redhat.com>
4976 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
4978 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
4980 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
4981 order of elements for big-endian.
4983 2014-03-25 Richard Biener <rguenther@suse.de>
4986 * gimplify-me.c (gimple_regimplify_operands): Update the
4989 2014-03-25 Martin Jambor <mjambor@suse.cz>
4992 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
4993 (lto_output_varpool_node): Likewise.
4994 (input_overwrite_node): Likewise.
4995 (input_varpool_node): Likewise.
4997 2014-03-25 Richard Biener <rguenther@suse.de>
4999 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
5000 (run_gcc): Likewise.
5002 2014-03-25 Jakub Jelinek <jakub@redhat.com>
5004 * combine.c (simplify_compare_const): Add MODE argument.
5005 Handle mode_width 0 as very large mode_width.
5006 (try_combine, simplify_comparison): Adjust callers.
5008 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
5009 type to avoid signed integer overflow.
5010 * explow.c (plus_constant): Likewise.
5012 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
5014 * doc/generic.texi: Correct typos.
5016 2014-03-24 Tobias Burnus <burnus@net-b.de>
5018 * doc/invoke.texi (-flto): Expand section about
5019 using static libraries with LTO.
5021 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5023 PR rtl-optimization/60501
5024 * optabs.def (addptr3_optab): New optab.
5025 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
5026 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
5027 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
5029 * lra.c (emit_add3_insn): Use the addptr pattern if available.
5031 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
5033 2014-03-24 Ulrich Drepper <drepper@gmail.com>
5035 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
5038 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
5039 (_mm256_undefined_ps): Define.
5040 (_mm256_undefined_pd): Define.
5041 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
5042 (_mm_undefined_pd): Define.
5043 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
5044 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
5045 (_mm512_undefined_ps): Define.
5046 (_mm512_undefined_pd): Define.
5047 Use _mm*_undefined_*.
5048 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
5050 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
5052 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
5053 (lshr_simd): DI mode added.
5054 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
5055 (aarch64_ushr_simddi): Likewise.
5056 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
5057 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
5058 (vshrd_n_u64): Likewise.
5060 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5062 * Makefile.in (s-macro_list): Depend on cc1.
5064 2014-03-23 Teresa Johnson <tejohnson@google.com>
5066 * ipa-utils.c (ipa_print_order): Use specified dump file.
5068 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
5070 PR rtl-optimization/60601
5071 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
5073 * gcc.c (eval_spec_function): Initialize save_growing_value.
5075 2014-03-22 Jakub Jelinek <jakub@redhat.com>
5078 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
5079 code == MINUS_EXPR, never swap op0 with op1.
5081 * toplev.c (init_local_tick): Avoid signed integer multiplication
5083 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
5084 shift by first operand's bitsize.
5086 2014-03-21 Jakub Jelinek <jakub@redhat.com>
5089 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
5091 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
5092 TARGET_ISA_64BIT_P(x).
5094 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5096 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
5097 pattern for vector nor instead of subtract from splat(-1).
5098 (altivec_expand_vec_perm_const_le): Likewise.
5100 2014-03-21 Richard Henderson <rth@twiddle.net>
5103 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
5104 related insns after epilogue_completed.
5106 2014-03-21 Martin Jambor <mjambor@suse.cz>
5109 * cgraph.h (symtab_node): New flag body_removed.
5110 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
5111 when removing bodies.
5112 * symtab.c (dump_symtab_base): Dump body_removed flag.
5113 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
5114 had their bodies removed.
5116 2014-03-21 Martin Jambor <mjambor@suse.cz>
5119 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
5122 2014-03-21 Richard Biener <rguenther@suse.de>
5124 PR tree-optimization/60577
5125 * tree-core.h (struct tree_base): Document nothrow_flag use
5126 in VAR_DECL_NONALIASED.
5127 * tree.h (VAR_DECL_NONALIASED): New.
5128 (may_be_aliased): Adjust.
5129 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
5131 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5133 * expr.c (expand_expr_real_1): Remove outdated comment.
5135 2014-03-20 Jakub Jelinek <jakub@redhat.com>
5138 * ira.c (adjust_cleared_regs): Call copy_rtx on
5139 *reg_equiv[REGNO (loc)].src_p before passing it to
5140 simplify_replace_fn_rtx.
5143 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
5144 into CONST, put pic register as first operand of PLUS. Use
5145 gen_const_mem for both 32-bit and 64-bit PIC got loads.
5147 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5149 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
5151 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5153 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
5154 around for store forwarding issue in the FPU on the UT699.
5155 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
5156 loads and operations if -mfix-ut699 is specified.
5157 (divtf3_hq): Tweak attribute.
5158 (sqrttf2_hq): Likewise.
5160 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5162 * calls.c (store_one_arg): Remove incorrect const qualification on the
5163 type of the temporary.
5164 * cfgexpand.c (expand_return): Likewise.
5165 * expr.c (expand_constructor): Likewise.
5166 (expand_expr_real_1): Likewise.
5168 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5170 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
5173 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
5176 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
5178 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
5180 * config/arm/aarch-common-protos.h
5181 (alu_cost_table): Fix spelling of "extend".
5182 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
5184 2014-03-19 Richard Biener <rguenther@suse.de>
5187 * tree-core.h (tree_type_common): Re-order pointer members
5188 to reduce recursion depth during GC walks.
5190 2014-03-19 Marek Polacek <polacek@redhat.com>
5193 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
5194 before accessing it.
5196 2014-03-19 Richard Biener <rguenther@suse.de>
5199 * lto-streamer-in.c (input_function): In WPA stage do not drop
5202 2014-03-19 Jakub Jelinek <jakub@redhat.com>
5204 PR tree-optimization/60559
5205 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
5206 with build_zero_cst assignment.
5208 2014-03-18 Kai Tietz <ktietz@redhat.com>
5210 PR rtl-optimization/56356
5211 * sdbout.c (sdbout_parms): Verify that parms'
5212 incoming argument is valid.
5213 (sdbout_reg_parms): Likewise.
5215 2014-03-18 Richard Henderson <rth@redhat.com>
5218 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
5219 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
5220 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
5222 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
5224 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
5225 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
5226 Italicize plugin event names in description. Explain that
5227 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
5228 Remind that no GCC functions should be called after PLUGIN_FINISH.
5229 Explain what pragmas with expansion are.
5231 2014-03-18 Martin Liska <mliska@suse.cz>
5233 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
5234 gimple call statement is update.
5235 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
5236 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
5238 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5241 * ubsan.c (ubsan_instrument_unreachable): Call
5242 initialize_sanitizer_builtins.
5243 (ubsan_pass): Likewise.
5246 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
5247 varpool_finalize_decl instead of rest_of_decl_compilation.
5249 2014-03-18 Richard Biener <rguenther@suse.de>
5251 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
5252 by using bitmap_and_compl instead of bitmap_and_compl_into.
5253 (df_rd_transfer_function): Likewise.
5255 2014-03-18 Richard Biener <rguenther@suse.de>
5257 * doc/lto.texi (fresolution): Fix typo.
5259 2014-03-18 Richard Biener <rguenther@suse.de>
5261 * doc/invoke.texi (flto): Update for changes in 4.9.
5263 2014-03-18 Richard Biener <rguenther@suse.de>
5265 * doc/loop.texi: Remove section on the removed lambda framework.
5266 Update loop docs with recent changes in preserving loop structure.
5268 2014-03-18 Richard Biener <rguenther@suse.de>
5270 * doc/lto.texi (-fresolution): Document.
5272 2014-03-18 Richard Biener <rguenther@suse.de>
5274 * doc/contrib.texi: Adjust my name.
5276 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5279 * internal-fn.c: Include diagnostic-core.h.
5280 (expand_BUILTIN_EXPECT): New function.
5281 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
5282 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
5283 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
5284 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
5286 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
5287 Revert 3 argument __builtin_expect code.
5288 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
5289 * gimple-fold.c (gimple_fold_call): Likewise.
5290 * tree.h (fold_builtin_expect): New prototype.
5291 * builtins.c (build_builtin_expect_predicate): Add predictor
5292 argument, if non-NULL, create 3 argument __builtin_expect.
5293 (fold_builtin_expect): No longer static. Add ARG2 argument,
5294 pass it through to build_builtin_expect_predicate.
5295 (fold_builtin_2): Adjust caller.
5296 (fold_builtin_3): Handle BUILT_IN_EXPECT.
5297 * internal-fn.def (BUILTIN_EXPECT): New.
5299 2014-03-18 Tobias Burnus <burnus@net-b.de>
5302 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
5303 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
5304 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
5306 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
5309 * predict.c (combine_predictions_for_bb): Fix up formatting.
5310 (expr_expected_value_1, expr_expected_value): Add predictor argument,
5311 fill what it points to if non-NULL.
5312 (tree_predict_by_opcode): Adjust caller, use the predictor.
5313 * predict.def (PRED_COMPARE_AND_SWAP): Add.
5315 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
5317 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
5318 proper constant for the store mode.
5320 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
5322 * symtab.c (change_decl_assembler_name): Fix transparent alias
5325 2014-03-16 Renlin Li <Renlin.Li@arm.com>
5327 * config/aarch64/aarch64.c: Correct the comments about the
5328 aarch64 stack layout.
5330 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
5332 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
5333 check for GF_OMP_FOR_KIND_FOR.
5335 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
5337 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
5338 ymm and zmm register names.
5340 2014-03-17 Jakub Jelinek <jakub@redhat.com>
5343 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
5344 note creation for the 2010-08-31 changes.
5346 2014-03-17 Marek Polacek <polacek@redhat.com>
5349 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
5350 as -fno-tree-loop-vectorize.
5351 (expand_omp_simd): Likewise.
5353 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
5355 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
5356 (eligible_for_call_delay): New prototype.
5357 * config/sparc/sparc.c (tls_call_delay): Rename into...
5358 (eligible_for_call_delay): ...this. Return false if the instruction
5359 cannot be put in the delay slot of a branch.
5360 (eligible_for_restore_insn): Simplify.
5361 (eligible_for_return_delay): Return false if the instruction cannot be
5362 put in the delay slot of a branch and simplify.
5363 (eligible_for_sibcall_delay): Return false if the instruction cannot be
5364 put in the delay slot of a branch.
5365 * config/sparc/sparc.md (fix_ut699): New attribute.
5366 (tls_call_delay): Delete.
5367 (in_call_delay): Reimplement.
5368 (eligible_for_sibcall_delay): Rename into...
5369 (in_sibcall_delay): ...this.
5370 (eligible_for_return_delay): Rename into...
5371 (in_return_delay): ...this.
5372 (in_branch_delay): Reimplement.
5373 (in_uncond_branch_delay): Delete.
5374 (in_annul_branch_delay): Delete.
5376 2014-03-14 Richard Henderson <rth@redhat.com>
5379 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
5380 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
5381 (*floathi<X87MODEF>2_i387_with_temp): Remove.
5382 (floathi splitters): Remove.
5383 (float<SWI48x>xf2): New pattern.
5384 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
5385 code that tried to handle DImode for 32-bit, but which was excluded
5386 by the pattern's condition. Drop allocation of stack temporary.
5387 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
5388 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
5389 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
5390 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
5391 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
5392 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
5393 (*float<SWI48><MODEF>2_sse_interunit): Remove.
5394 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
5395 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
5396 (*float<SWI48x><X87MODEF>2_i387): Remove.
5397 (all float _with_temp splitters): Remove.
5398 (*float<SWI48x><MODEF>2_i387): New pattern.
5399 (*float<SWI48><MODEF>2_sse): New pattern.
5400 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
5401 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
5403 2014-03-14 Jakub Jelinek <jakub@redhat.com>
5404 Marek Polacek <polacek@redhat.com>
5407 * common.opt (dump_base_name_prefixed): New Variable.
5408 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
5409 if x_dump_base_name_prefixed is already set, set it at the end.
5411 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
5413 PR rtl-optimization/60508
5414 * lra-constraints.c (get_reload_reg): Add new parameter
5416 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
5417 Pass the new parameter values.
5419 2014-03-14 Richard Biener <rguenther@suse.de>
5421 * common.opt: Revert unintented changes from r205065.
5424 2014-03-14 Richard Biener <rguenther@suse.de>
5427 * cfghooks.c (split_block): Properly adjust all loops the
5428 block was a latch of.
5430 2014-03-14 Martin Jambor <mjambor@suse.cz>
5433 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
5436 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
5439 * config/avr/avr.c (avr_set_current_function): Pass function name
5440 through default_strip_name_encoding before sanity checking instead
5441 of skipping the first char of the assembler name.
5443 2014-03-13 Richard Henderson <rth@redhat.com>
5446 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
5447 (ix86_force_to_memory, ix86_free_from_memory): Remove.
5448 * config/i386/i386-protos.h: Likewise.
5449 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
5450 in the expander instead of a splitter.
5451 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
5452 any possibility of requiring a memory.
5453 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
5454 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
5455 (fp branch splitters): Update for ix86_split_fp_branch.
5456 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
5457 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
5458 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
5459 (*fop_<MODEF>_2_i387): Remove f/r alternative.
5460 (*fop_<MODEF>_3_i387): Likewise.
5461 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
5462 (splitters for the fop_* register patterns): Remove.
5463 (fscalexf4_i387): Rename from *fscalexf4_i387.
5464 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
5466 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5468 PR tree-optimization/59779
5469 * tree-dfa.c (get_ref_base_and_extent): Use double_int
5470 type for bitsize and maxsize instead of HOST_WIDE_INT.
5472 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
5474 PR rtl-optimization/57320
5475 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
5476 the CFG after thread_prologue_and_epilogue_insns.
5478 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
5480 PR rtl-optimization/57189
5481 * lra-constraints.c (process_alt_operands): Disfavor spilling
5484 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
5486 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
5488 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5490 PR tree-optimization/59025
5492 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
5493 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
5495 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
5498 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
5499 calls of avr_out_plus_1.
5501 2014-03-13 Bin Cheng <bin.cheng@arm.com>
5503 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
5504 BB's single pred and update the father loop's latch info later.
5506 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5508 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
5512 (VEC_base): Likewise.
5513 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
5514 registers, we need to swap double words in little endian mode.
5516 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
5517 to be a container mode for 128-bit integer operations added in ISA
5518 2.07. Unlike TImode and PTImode, the preferred register set is
5519 the Altivec/VMX registers for the 128-bit operations.
5521 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
5523 (rs6000_split_128bit_ok_p): Likewise.
5525 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
5526 macros for creating ISA 2.07 normal and overloaded builtin
5527 functions with 3 arguments.
5528 (BU_P8V_OVERLOAD_3): Likewise.
5529 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
5530 for use as overloaded functions.
5531 (VPERM_1TI_UNS): Likewise.
5532 (VSEL_1TI): Likewise.
5533 (VSEL_1TI_UNS): Likewise.
5534 (ST_INTERNAL_1ti): Likewise.
5535 (LD_INTERNAL_1ti): Likewise.
5536 (XXSEL_1TI): Likewise.
5537 (XXSEL_1TI_UNS): Likewise.
5538 (VPERM_1TI): Likewise.
5539 (VPERM_1TI_UNS): Likewise.
5540 (XXPERMDI_1TI): Likewise.
5541 (SET_1TI): Likewise.
5542 (LXVD2X_V1TI): Likewise.
5543 (STXVD2X_V1TI): Likewise.
5544 (VEC_INIT_V1TI): Likewise.
5545 (VEC_SET_V1TI): Likewise.
5546 (VEC_EXT_V1TI): Likewise.
5547 (EQV_V1TI): Likewise.
5548 (NAND_V1TI): Likewise.
5549 (ORC_V1TI): Likewise.
5550 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
5551 added in ISA 2.07. Add both normal 'altivec' builtins, and the
5553 (VADDUQM): Likewise.
5554 (VSUBCUQ): Likewise.
5555 (VADDEUQM): Likewise.
5556 (VADDECUQ): Likewise.
5557 (VSUBEUQM): Likewise.
5558 (VSUBECUQ): Likewise.
5560 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5561 __int128_t and __uint128_t types.
5562 (__uint128_type): Likewise.
5563 (altivec_categorize_keyword): Add support for vector __int128_t,
5564 vector __uint128_t, vector __int128, and vector unsigned __int128
5565 as a container type for TImode operations that need to be done in
5566 VSX/Altivec registers.
5567 (rs6000_macro_to_expand): Likewise.
5568 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5569 to support 128-bit integer instructions vaddcuq, vadduqm,
5570 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5571 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5573 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5574 for V1TImode, and set up preferences to use VSX/Altivec registers.
5575 Setup VSX reload handlers.
5576 (rs6000_debug_reg_global): Likewise.
5577 (rs6000_init_hard_regno_mode_ok): Likewise.
5578 (rs6000_preferred_simd_mode): Likewise.
5579 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5580 (easy_altivec_constant): Likewise.
5581 (output_vec_const_move): Likewise.
5582 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5584 (rs6000_expand_vector_extract): Likewise.
5585 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
5587 (rs6000_const_vec): Add support for V1TImode.
5588 (rs6000_emit_le_vsx_load): Swap double words when loading or
5589 storing TImode/V1TImode.
5590 (rs6000_emit_le_vsx_store): Likewise.
5591 (rs6000_emit_le_vsx_move): Likewise.
5592 (rs6000_emit_move): Add support for V1TImode.
5593 (altivec_expand_ld_builtin): Likewise.
5594 (altivec_expand_st_builtin): Likewise.
5595 (altivec_expand_vec_init_builtin): Likewise.
5596 (altivec_expand_builtin): Likewise.
5597 (rs6000_init_builtins): Add support for V1TImode type. Add
5598 support for ISA 2.07 128-bit integer builtins. Define type names
5599 for the VSX/Altivec vector types.
5600 (altivec_init_builtins): Add support for overloaded vector
5601 functions with V1TImode type.
5602 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
5603 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
5605 (rs6000_split_128bit_ok_p): Likewise.
5606 (rs6000_handle_altivec_attribute): Create V1TImode from vector
5607 __int128_t and vector __uint128_t.
5609 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
5610 and mode attributes.
5617 (VS_scalar): Likewise.
5618 (VS_double): Likewise.
5619 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
5621 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
5622 we support the ISA 2.07 128-bit integer arithmetic instructions.
5623 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
5624 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
5625 and TImode types for use with the builtin functions.
5626 (V1TI_type_node): Likewise.
5627 (unsigned_V1TI_type_node): Likewise.
5628 (intTI_type_internal_node): Likewise.
5629 (uintTI_type_internal_node): Likewise.
5631 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
5632 128-bit builtin functions.
5633 (UNSPEC_VADDEUQM): Likewise.
5634 (UNSPEC_VADDECUQ): Likewise.
5635 (UNSPEC_VSUBCUQ): Likewise.
5636 (UNSPEC_VSUBEUQM): Likewise.
5637 (UNSPEC_VSUBECUQ): Likewise.
5638 (VM): Add V1TImode to vector mode iterators.
5640 (VI_unit): Likewise.
5641 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
5642 (altivec_vaddcuq): Likewise.
5643 (altivec_vsubuqm): Likewise.
5644 (altivec_vsubcuq): Likewise.
5645 (altivec_vaddeuqm): Likewise.
5646 (altivec_vaddecuq): Likewise.
5647 (altivec_vsubeuqm): Likewise.
5648 (altivec_vsubecuq): Likewise.
5650 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
5652 (BOOL_128): Likewise.
5653 (BOOL_REGS_OUTPUT): Likewise.
5654 (BOOL_REGS_OP1): Likewise.
5655 (BOOL_REGS_OP2): Likewise.
5656 (BOOL_REGS_UNARY): Likewise.
5657 (BOOL_REGS_AND_CR0): Likewise.
5659 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
5660 128-bit integer builtin support.
5661 (vec_vadduqm): Likewise.
5662 (vec_vaddecuq): Likewise.
5663 (vec_vaddeuqm): Likewise.
5664 (vec_vsubecuq): Likewise.
5665 (vec_vsubeuqm): Likewise.
5666 (vec_vsubcuq): Likewise.
5667 (vec_vsubuqm): Likewise.
5669 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5670 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
5671 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
5672 128-bit integer add/subtract to ISA 2.07.
5674 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
5676 * config/arc/arc.c (arc_predicate_delay_insns):
5677 Fix third argument passed to conditionalize_nonjump.
5679 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
5681 * config/aarch64/aarch64-builtins.c
5682 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
5683 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
5684 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
5685 instead of __builtin_lfloor.
5686 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
5688 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5690 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
5691 (tree_ssa_ifcombine_bb_1): New function.
5692 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
5693 is an empty forwarder block to then_bb or vice versa and then_bb
5694 and else_bb are effectively swapped.
5696 2014-03-12 Christian Bruel <christian.bruel@st.com>
5699 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
5700 REG_CFA_DEF_CFA note.
5701 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
5702 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
5704 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
5706 PR tree-optimization/60454
5707 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
5709 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5711 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
5712 Do not define target_cpu_default2 to generic.
5713 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
5714 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
5715 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
5717 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5718 Marc Glisse <marc.glisse@inria.fr>
5720 PR tree-optimization/60502
5721 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
5722 instead of build_low_bits_mask.
5724 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5727 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
5728 if there are multiple uses, but op doesn't live on E edge.
5729 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
5730 clobber stmts before __builtin_unreachable.
5732 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
5734 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
5735 hard_frame_pointer_rtx.
5736 * cse.c (cse_insn): Remove volatile check.
5737 * cselib.c (cselib_process_insn): Likewise.
5738 * dse.c (scan_insn): Likewise.
5740 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5742 * config/arc/arc.c (conditionalize_nonjump): New function,
5744 (arc_ifcvt): ... this.
5745 (arc_predicate_delay_insns): Use it.
5747 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5749 * config/arc/predicates.md (extend_operand): During/after reload,
5750 allow const_int_operand.
5751 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
5752 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
5753 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
5755 (umulsi3_highpart_i): Likewise.
5757 2014-03-11 Richard Biener <rguenther@suse.de>
5759 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
5760 Add asserts to guard possible wrong-code bugs.
5762 2014-03-11 Richard Biener <rguenther@suse.de>
5764 PR tree-optimization/60429
5765 PR tree-optimization/60485
5766 * tree-ssa-structalias.c (set_union_with_increment): Properly
5767 take into account all fields that overlap the shifted vars.
5768 (do_sd_constraint): Likewise.
5769 (do_ds_constraint): Likewise.
5770 (get_constraint_for_ptr_offset): Likewise.
5772 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
5774 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
5775 (nios2_compute_frame_layout):
5776 Add calculation of cfun->machine->fp_save_offset.
5777 (nios2_expand_prologue): Correct setting of frame pointer register
5779 (nios2_expand_epilogue): Update recovery of stack pointer from
5780 frame pointer accordingly.
5781 (nios2_initial_elimination_offset): Update calculation of offset
5782 for eliminating to HARD_FRAME_POINTER_REGNUM.
5784 2014-03-10 Jakub Jelinek <jakub@redhat.com>
5787 * ipa.c (symtab_remove_unreachable_nodes): Don't call
5788 cgraph_get_create_node on VAR_DECLs.
5790 2014-03-10 Richard Biener <rguenther@suse.de>
5793 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
5795 2014-03-08 Douglas B Rupp <rupp@gnat.com>
5797 * config/vms/vms.opt (vms_float_format): New variable.
5799 2014-03-08 Tobias Burnus <burnus@net-b.de>
5801 * doc/invoke.texi (-fcilkplus): Update implementation status.
5803 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
5804 Richard Biener <rguenther@suse.de>
5806 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
5807 consistently accross all TUs.
5808 (run_gcc): Enable -fshort-double automatically at link at link-time
5809 and disallow override.
5811 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
5814 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
5815 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
5816 if they can't be used.
5818 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5820 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
5821 for Solaris 11/x86 ld.
5822 * configure: Regenerate.
5824 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5826 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
5827 (LIB_TLS_SPEC): Save as ld_tls_libs.
5828 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
5829 (HAVE_AS_IX86_TLSLDM): New test.
5830 * configure, config.in: Regenerate.
5831 * config/i386/i386.c (legitimize_tls_address): Fall back to
5832 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
5833 cannot support TLS_MODEL_LOCAL_DYNAMIC.
5834 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
5835 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
5837 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
5839 * common.opt (fira-loop-pressure): Mark as optimization.
5841 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
5843 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
5844 an OpenMP mappable type.
5846 2014-03-06 Matthias Klose <doko@ubuntu.com>
5848 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
5849 MULTILIB_OSDIRNAMES is not defined.
5851 2014-03-06 Jakub Jelinek <jakub@redhat.com>
5852 Meador Inge <meadori@codesourcery.com>
5855 * config/arm/arm.c (arm_tls_symbol_p): Remove.
5856 (arm_legitimize_address): Call legitimize_tls_address for any
5857 arm_tls_referenced_p expression, handle constant addend. Call it
5858 before testing for !TARGET_ARM.
5859 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
5861 2014-03-06 Richard Biener <rguenther@suse.de>
5867 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5869 * tree-streamer.c (record_common_node): Assert we don't record
5870 nodes with type double.
5871 (preload_common_node): Skip type double, complex double and double
5872 pointer since it is now frontend dependent due to fshort-double option.
5874 2014-03-06 Richard Biener <rguenther@suse.de>
5876 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
5877 or -fno-lto is specified and the linker has full plugin support.
5878 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
5879 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
5880 * lto-wrapper.c (merge_and_complain): Merge compile-time
5881 optimization levels.
5882 (run_gcc): And pass it through to the link options.
5884 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
5888 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5890 * cselib.c (remove_useless_values): Skip to avoid quadratic
5891 behavior if the condition moved from...
5892 (cselib_process_insn): ... here holds.
5894 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5897 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
5898 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
5901 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
5902 (TM_H): Add x86-tune.def.
5904 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5906 * config/aarch64/aarch64.c (generic_tunings):
5907 Use cortexa57_extra_costs.
5909 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5912 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
5913 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
5914 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
5917 2014-03-04 Heiher <r@hev.cc>
5919 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
5920 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
5922 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
5924 * config/i386/predicates.md (const2356_operand): Change to ...
5925 (const2367_operand): ... this.
5926 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
5928 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5929 (*avx512pf_scatterpf<mode>sf): Ditto.
5930 (avx512pf_scatterpf<mode>df): Ditto.
5931 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5932 (*avx512pf_scatterpf<mode>df): Ditto.
5933 * config/i386/i386.c (ix86_expand_builtin): Update
5934 incorrect hint operand error message.
5936 2014-03-04 Richard Biener <rguenther@suse.de>
5938 * lto-section-in.c (lto_get_section_data): Fix const cast.
5940 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5942 * tree-streamer.c (record_common_node): Assert we don't record
5943 nodes with type double.
5944 (preload_common_node): Skip type double, complex double and double
5945 pointer since it is now frontend dependent due to fshort-double option.
5947 2014-03-04 Richard Biener <rguenther@suse.de>
5950 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
5951 (lto_input_toplevel_asms): Likewise.
5952 * lto-section-in.c (lto_get_section_data): Instead do it here
5955 2014-03-04 Richard Biener <rguenther@suse.de>
5957 PR tree-optimization/60382
5958 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
5959 dead PHIs a reduction.
5961 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
5963 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
5965 (_mm_prefetch): Move out of GCC target("sse") pragma.
5966 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
5967 GCC target("prfchw") pragma.
5968 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
5970 * config/i386/i386.c (ix86_option_override_internal): Enable
5971 -mprfchw with -mprefetchwt1.
5973 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5975 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
5978 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5980 * opts.h (CL_PCH_IGNORE): Define.
5981 * targhooks.c (option_affects_pch_p):
5982 Return false for options that have CL_PCH_IGNORE set.
5983 * opt-functions.awk: Process PchIgnore.
5984 * doc/options.texi: Document PchIgnore.
5986 * config/arc/arc.opt (misize): Add PchIgnore property.
5988 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5990 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
5991 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
5992 constraint on constants to permit them being loaded into
5993 GENERAL_REGS or BASE_REGS.
5995 2014-03-03 Nick Clifton <nickc@redhat.com>
5997 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
5998 anti-cacnonical alternatives.
5999 (negandhi3_real): New pattern.
6000 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
6002 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6004 * config/avr/avr-mcus.def: Remove atxmega16x1.
6005 * config/avr/avr-tables.opt: Regenerate.
6006 * config/avr/t-multilib: Regenerate.
6007 * doc/avr-mmcu.texi: Regenerate.
6009 2014-03-03 Tobias Grosser <tobias@grosser.es>
6010 Mircea Namolaru <mircea.namolaru@inria.fr>
6012 PR tree-optimization/58028
6013 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
6016 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6018 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
6019 not handled by recognizers.
6021 2014-03-03 Jakub Jelinek <jakub@redhat.com>
6024 * function.c (expand_function_end): Don't emit
6025 clobber_return_register sequence if clobber_after is a BARRIER.
6026 * cfgexpand.c (construct_exit_block): Append instructions before
6027 return_label to prev_bb.
6029 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6031 * config/rs6000/constraints.md: Document reserved use of "wc".
6033 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
6036 * ipa.c (function_and_variable_visibility): When dissolving comdat
6037 group, also set all symbols to local.
6039 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
6044 2013-12-14 Jan Hubicka <jh@suse.cz>
6046 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
6048 2014-03-02 Jon Beniston <jon@beniston.com>
6054 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
6055 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
6056 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
6057 (simple_return, *simple_return): New patterns
6058 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
6059 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
6061 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
6063 * dwarf2out.c (gen_subprogram_die): Tidy.
6065 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
6068 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
6069 (*mov_t_msb_neg_negc): ... this new insn.
6071 2014-02-28 Jason Merrill <jason@redhat.com>
6074 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
6077 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
6080 * dwarf2out.c (decltype_auto_die): New static.
6081 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
6082 (gen_type_die_with_usage): Handle 'decltype(auto)'.
6083 (is_cxx_auto): Likewise.
6085 2014-02-28 Ian Bolton <ian.bolton@arm.com>
6087 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
6088 we are not using general regs only.
6090 2014-02-28 Richard Biener <rguenther@suse.de>
6093 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
6094 previous fix and only allow to remove trivial pre-headers
6095 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
6096 (remove_forwarder_block): Properly update the latch of a loop.
6098 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6101 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
6102 (cselib_preserved_hash_table): New.
6103 (preserve_constants_and_equivs): Move preserved vals to it.
6104 (cselib_find_slot): Look it up first.
6105 (cselib_init): Initialize it.
6106 (cselib_finish): Release it.
6107 (dump_cselib_table): Dump it.
6109 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6112 * cselib.c (remove_useless_values): Skip to avoid quadratic
6113 behavior if the condition moved from...
6114 (cselib_process_insn): ... here holds.
6116 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6119 * var-tracking.c (vt_initialize): Apply the same condition to
6120 preserve the CFA base value.
6122 2014-02-28 Joey Ye <joey.ye@arm.com>
6125 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
6126 if reload in progress or completed.
6128 2014-02-28 Tobias Burnus <burnus@net-b.de>
6131 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
6134 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
6136 * doc/tm.texi.in (Condition Code Status): Update documention for
6137 relative locations of cc0-setter and cc0-user.
6139 2014-02-27 Jeff Law <law@redhat.com>
6141 PR rtl-optimization/52714
6142 * combine.c (try_combine): When splitting an unrecognized PARALLEL
6143 into two independent simple sets, if I3 is a jump, ensure the
6144 pattern we place into I3 is a (set (pc) ...).
6146 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
6147 Jeff Law <law@redhat.com>
6149 PR rtl-optimization/49847
6150 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
6151 are in different blocks.
6152 * doc/tm.texi (Condition Code Status): Update documention for
6153 relative locations of cc0-setter and cc0-user.
6155 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
6158 * lra.c (lra_emit_add): Check SUBREG too.
6160 2014-02-27 Andreas Schwab <schwab@suse.de>
6162 * config/m68k/m68k.c (m68k_option_override): Disable
6163 -flive-range-shrinkage for classic m68k.
6164 (m68k_override_options_after_change): Likewise.
6166 2014-02-27 Marek Polacek <polacek@redhat.com>
6169 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
6170 -Wmaybe-uninitialized.
6172 2014-02-27 Alan Modra <amodra@gmail.com>
6175 * reload1.c (emit_input_reload_insns): When reload_override_in,
6176 set old to rl->in_reg when rl->in_reg is a subreg.
6178 2014-02-26 Richard Biener <rguenther@suse.de>
6181 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
6183 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6185 * common/config/i386/predicates.md (const1256_operand): Remove.
6186 (const2356_operand): New.
6187 (const_1_to_2_operand): Remove.
6188 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6189 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6190 (*avx512pf_gatherpf<mode>sf): Ditto.
6191 (avx512pf_gatherpf<mode>df): Ditto.
6192 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6193 (*avx512pf_gatherpf<mode>df): Ditto.
6194 (avx512pf_scatterpf<mode>sf): Ditto.
6195 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6196 (*avx512pf_scatterpf<mode>sf): Ditto.
6197 (avx512pf_scatterpf<mode>df): Ditto.
6198 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6199 (*avx512pf_scatterpf<mode>df): Ditto.
6200 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
6202 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
6204 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
6205 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
6206 (_mm512_mask_testn_epi64_mask): Move to ...
6207 * config/i386/avx512cdintrin.h: Here.
6208 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
6209 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
6210 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
6211 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
6212 TARGET_AVX512F from TARGET_AVX512CD.
6214 2014-02-26 Richard Biener <rguenther@suse.de>
6217 * ipa.c (walk_polymorphic_call_targets): Properly guard
6218 call to inline_update_overall_summary.
6220 2014-02-26 Bin Cheng <bin.cheng@arm.com>
6223 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
6224 and latches only if requested. Fix latch if it is removed.
6225 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
6226 LOOPS_HAVE_PREHEADERS.
6228 2014-02-25 Andrew Pinski <apinski@cavium.com>
6230 * builtins.c (expand_builtin_thread_pointer): Create a new target
6231 when the target is NULL.
6233 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
6235 PR rtl-optimization/60317
6236 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6237 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6238 * lra-assigns.c: Include params.h.
6239 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
6240 other reload pseudos considerations.
6242 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6244 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
6245 to use canonical form for nor<mode>3.
6247 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6250 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
6253 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6255 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
6256 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
6257 (ix86_handle_option): Handle OPT_mprefetchwt1.
6258 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
6259 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
6261 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
6262 OPTION_MASK_ISA_PREFETCHWT1.
6263 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
6264 (PTA_PREFETCHWT1): New.
6265 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
6266 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
6267 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
6268 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
6269 (*prefetch_avx512pf_<mode>_: Change into ...
6270 (*prefetch_prefetchwt1_<mode>: This.
6271 * config/i386/i386.opt (mprefetchwt1): New.
6272 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
6273 (_mm_prefetch): Handle intent to write.
6274 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
6276 2014-02-25 Richard Biener <rguenther@suse.de>
6279 * emit-rtl.c (mem_attrs_htab): Remove.
6280 (mem_attrs_htab_hash): Likewise.
6281 (mem_attrs_htab_eq): Likewise.
6282 (set_mem_attrs): Always allocate new mem-attrs when something changed.
6283 (init_emit_once): Do not allocate mem_attrs_htab.
6285 2014-02-25 Richard Biener <rguenther@suse.de>
6288 * lto-opts.c (lto_write_options): Output non-explicit conservative
6289 -fwrapv, -fno-trapv and -fno-strict-overflow.
6290 * lto-wrapper.c (merge_and_complain): Handle merging those options.
6291 (run_gcc): And pass them through.
6293 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6295 * sel-sched.c (calculate_new_fences): New parameter ptime.
6296 Calculate it as a maximum over all fence cycles.
6297 (sel_sched_region_2): Adjust the call to calculate_new_fences.
6298 Print the final schedule timing when sched_verbose.
6300 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6302 PR rtl-optimization/60292
6303 * sel-sched.c (fill_vec_av_set): Do not reset target availability
6304 bit fot the fence instruction.
6306 2014-02-24 Alangi Derick <alangiderick@gmail.com>
6308 * calls.h: Fix typo in comment.
6310 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
6312 * config/pa/pa.c (pa_output_move_double): Don't valididate when
6313 adjusting offsetable addresses.
6315 2014-02-24 Guozhi Wei <carrot@google.com>
6317 * sparseset.h (sparseset_pop): Fix the wrong index.
6319 2014-02-24 Walter Lee <walt@tilera.com>
6321 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
6322 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
6324 * common/config/tilegx/tilegx-common.c
6325 (TARGET_DEFAULT_TARGET_FLAGS): Define.
6326 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
6328 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
6329 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
6330 (tilegx_gimplify_va_arg_expr): Handle big endian.
6331 (tilegx_expand_unaligned_load): Ditto.
6332 (tilegx_expand_unaligned_store): Ditto.
6333 (TARGET_RETURN_IN_MSB): New.
6334 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
6335 (TARGET_ENDIAN_DEFAULT): New.
6336 (TARGET_BIG_ENDIAN): Handle big endian.
6337 (BYTES_BIG_ENDIAN): Ditto.
6338 (WORDS_BIG_ENDIAN): Ditto.
6339 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
6342 * config/tilegx/tilegx.md (extv): Handle big endian.
6344 (insn_st<n>): Ditto.
6345 (insn_st<n>_add<bitsuffix>): Ditto.
6346 (insn_stnt<n>): Ditto.
6347 (insn_stnt<n>_add<bitsuffix>):Ditto.
6348 (vec_interleave_highv8qi): Handle big endian.
6349 (vec_interleave_highv8qi_be): New.
6350 (vec_interleave_highv8qi_le): New.
6351 (insn_v1int_h): Handle big endian.
6352 (vec_interleave_lowv8qi): Handle big endian.
6353 (vec_interleave_lowv8qi_be): New.
6354 (vec_interleave_lowv8qi_le): New.
6355 (insn_v1int_l): Handle big endian.
6356 (vec_interleave_highv4hi): Handle big endian.
6357 (vec_interleave_highv4hi_be): New.
6358 (vec_interleave_highv4hi_le): New.
6359 (insn_v2int_h): Handle big endian.
6360 (vec_interleave_lowv4hi): Handle big endian.
6361 (vec_interleave_lowv4hi_be): New.
6362 (vec_interleave_lowv4hi_le): New.
6363 (insn_v2int_l): Handle big endian.
6364 (vec_interleave_highv2si): Handle big endian.
6365 (vec_interleave_highv2si_be): New.
6366 (vec_interleave_highv2si_le): New.
6367 (insn_v4int_h): Handle big endian.
6368 (vec_interleave_lowv2si): Handle big endian.
6369 (vec_interleave_lowv2si_be): New.
6370 (vec_interleave_lowv2si_le): New.
6371 (insn_v4int_l): Handle big endian.
6372 * config/tilegx/tilegx.opt (mbig-endian): New option.
6373 (mlittle-endian): New option.
6374 * doc/install.texi: Document tilegxbe-linux.
6375 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
6377 2014-02-24 Martin Jambor <mjambor@suse.cz>
6380 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
6381 there are no parameter descriptors.
6383 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
6385 PR rtl-optimization/60268
6386 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
6387 initialization to ...
6388 (sched_rgn_init): ... here.
6389 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
6391 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6393 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
6396 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
6398 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
6401 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6403 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
6404 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
6406 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6408 * config/microblaze/predicates.md: Add cmp_op predicate.
6409 * config/microblaze/microblaze.md: Add branch_compare instruction
6410 which uses cmp_op predicate and emits cmp insn before branch.
6411 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
6412 to microblaze_expand_conditional_branch and consolidate logic.
6413 (microblaze_expand_conditional_branch): emit branch_compare
6414 insn instead of handling cmp op separate from branch insn.
6416 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6418 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
6421 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6423 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
6424 define_insn with define_expand and new define_insn
6425 *altivec_lve<VI_char>x_internal.
6426 (altivec_stve<VI_char>x): Replace define_insn with define_expand
6427 and new define_insn *altivec_stve<VI_char>x_internal.
6428 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
6430 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
6432 (altivec_expand_stvex_be): New function.
6434 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
6436 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
6437 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
6438 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
6439 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
6441 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
6444 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
6445 instead of emit_move_insn.
6447 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6449 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
6451 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
6452 gen_altivec_vsumsws.
6454 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6456 * config/rs6000/altivec.md (altivec_lvxl): Rename as
6457 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
6458 (altivec_lvxl_<mode>): New define_expand incorporating
6459 -maltivec=be semantics where needed.
6460 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
6461 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
6462 semantics where needed.
6463 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
6464 (altivec_stvx_<mode>): New define_expand incorporating
6465 -maltivec=be semantics where needed.
6466 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
6467 VM2 iterator instead of V4SI.
6468 (altivec_stvxl_<mode>): New define_expand incorporating
6469 -maltivec=be semantics where needed.
6470 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
6471 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
6472 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
6473 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
6474 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
6475 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
6476 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
6477 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
6478 ALTIVEC_BUILTIN_STVXL.
6479 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
6480 (altivec_expand_stvx_be): Likewise.
6481 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
6482 (altivec_expand_lvx_be): Likewise.
6483 (altivec_expand_stvx_be): Likewise.
6484 (altivec_expand_builtin): Add cases for
6485 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
6486 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
6487 (altivec_init_builtins): Add definitions for
6488 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
6489 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
6491 2014-02-21 Catherine Moore <clm@codesourcery.com>
6493 * doc/invoke.texi (mvirt, mno-virt): Document.
6494 * config/mips/mips.opt (mvirt): New option.
6495 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
6497 2014-02-21 Richard Biener <rguenther@suse.de>
6499 PR tree-optimization/60276
6500 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
6501 (STMT_VINFO_MIN_NEG_DIST): New macro.
6502 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
6503 STMT_VINFO_MIN_NEG_DIST.
6504 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
6505 made for negative dependence distances still hold.
6507 2014-02-21 Richard Biener <rguenther@suse.de>
6510 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
6511 DECL_INITIAL for globals not in the current function context.
6513 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6515 PR tree-optimization/56490
6516 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
6517 * tree-ssa-uninit.c: Include params.h.
6518 (compute_control_dep_chain): Add num_calls argument, return false
6519 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
6520 num_calls to recursive call.
6521 (find_predicates): Change dep_chain into normal array,
6522 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
6523 variable and adjust compute_control_dep_chain caller.
6524 (find_def_preds): Likewise.
6526 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
6528 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
6529 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
6531 2014-02-21 Nick Clifton <nickc@redhat.com>
6533 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
6534 (pushhi1): Likewise.
6535 (popqi1): Add mode to pre_dec.
6538 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6540 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
6541 mode for mask of V8SFmode permutation.
6543 2014-02-20 Richard Henderson <rth@redhat.com>
6546 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
6547 a new pseudo for OLDVAL.
6549 2014-02-20 Jakub Jelinek <jakub@redhat.com>
6552 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
6553 gen_reg_rtx if d->testing_p.
6554 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
6555 if d->testing_p and we will certainly return true.
6556 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6559 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6561 * emit-rtl.c (gen_reg_rtx): Assert that
6562 crtl->emit.regno_pointer_align_length is non-zero.
6564 2014-02-20 Richard Henderson <rth@redhat.com>
6567 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6568 on failure the store back into EXPECT.
6570 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6571 Sandra Loosemore <sandra@codesourcery.com>
6573 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6574 * config/nios2/nios2.c (nios2_function_profiler): Add
6575 -fPIC (flag_pic == 2) support.
6576 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6577 (nios2_large_offset_p): New function.
6578 (nios2_unspec_reloc_p): Move up position, update to use
6579 nios2_large_offset_p.
6580 (nios2_unspec_address): Remove function.
6581 (nios2_unspec_offset): New function.
6582 (nios2_large_got_address): New function.
6583 (nios2_got_address): Add large offset support.
6584 (nios2_legitimize_tls_address): Update usage of removed and new
6586 (nios2_symbol_binds_local_p): New function.
6587 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
6588 (nios2_legitimize_address): Update to use nios2_large_offset_p.
6589 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
6590 (nios2_print_operand): Merge H/L processing, add hiadj/lo
6591 processing for (const (unspec ...)).
6592 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
6594 2014-02-20 Richard Biener <rguenther@suse.de>
6596 * tree-cfg.c (replace_uses_by): Mark altered BBs before
6597 doing the substitution.
6598 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
6600 2014-02-20 Martin Jambor <mjambor@suse.cz>
6603 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
6604 info when checking whether lattices are bottom.
6606 2014-02-20 Richard Biener <rguenther@suse.de>
6609 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
6612 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
6615 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
6616 parameter specifying the scaling.
6617 (inline_call): Update.
6618 (want_inline_recursively): Guard division by zero.
6619 (recursive_inlining): Update.
6620 * ipa-inline.h (clone_inlined_nodes): Update.
6622 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6625 * config/i386/i386.c (classify_argument): Pass structures of size
6626 64 bytes or less in register.
6628 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6629 Kirill Yukhin <kirill.yukhin@intel.com>
6631 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
6632 (_mm_rcp28_round_ss): Ditto.
6633 (_mm_rsqrt28_round_sd): Ditto.
6634 (_mm_rsqrt28_round_ss): Ditto.
6635 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
6636 (_mm_rcp14_round_ss): Ditto.
6637 (_mm_rsqrt14_round_sd): Ditto.
6638 (_mm_rsqrt14_round_ss): Ditto.
6639 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
6640 the first input operand, get rid of match_dup.
6641 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
6643 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
6645 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
6646 operand as the first input operand, set type attribute.
6647 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
6649 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
6650 operand as the first input operand, set type attribute.
6652 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6654 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
6657 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
6660 * config/i386/i386.c (construct_container): Remove TFmode check
6661 for X86_64_INTEGER_CLASS.
6663 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
6666 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
6667 only when -Wpsabi is enabled.
6669 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
6672 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
6673 passing arrays in registers are the same as for structs, so remove the
6674 special case for them.
6676 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
6678 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
6679 destination type, extract only the valid bits if the source type is not
6680 integral and has a different mode.
6682 2014-02-19 Richard Biener <rguenther@suse.de>
6685 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
6688 2014-02-19 Richard Biener <rguenther@suse.de>
6691 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
6692 (ipa_modify_call_arguments): Emit an argument load explicitely and
6693 preserve virtual SSA form there and for the replacement call.
6694 Do not update SSA form nor free dominance info.
6696 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6698 * ipa.c (function_and_variable_visibility): Also clear WEAK
6699 flag when disolving COMDAT_GROUP.
6701 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6703 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
6704 * ipa-prop.c (ipa_set_jf_known_type): Return early when
6706 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
6707 do more sanity checks.
6708 (detect_type_change): Return true when giving up early.
6709 (compute_complex_assign_jump_func): Fix type parameter of
6710 ipa_set_ancestor_jf.
6711 (compute_complex_ancestor_jump_func): Likewise.
6712 (update_jump_functions_after_inlining): Fix updating of
6714 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
6716 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6718 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
6719 inline clones when edge disappears.
6721 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6724 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
6725 Split 64-bit moves into 2 patterns. Do not allow the use of
6726 direct move for TDmode in little endian, since the decimal value
6727 has little endian bytes within a word, but the 64-bit pieces are
6728 ordered in a big endian fashion, and normal subreg's of TDmode are
6730 (mov<mode>_64bit_dm): Likewise.
6731 (movtd_64bit_nodm): Likewise.
6733 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6735 PR tree-optimization/60174
6736 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
6737 statement of an SSA_NAME that occurs in an abnormal PHI node.
6739 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6742 * final.c (SEEN_BB): Remove.
6743 (SEEN_NOTE, SEEN_EMITTED): Renumber.
6744 (final_scan_insn): Don't force_source_line on second
6745 NOTE_INSN_BASIC_BLOCK.
6747 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
6750 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
6751 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
6752 (type_natural_mode): Warn ABI change when %zmm register is not
6753 available for AVX512F vector value passing.
6755 2014-02-18 Kai Tietz <ktietz@redhat.com>
6758 * config/i386/i386.c (ix86_expand_prologue): Use value in
6759 rax register as displacement when restoring %r10 or %rax.
6760 Fix wrong offset when restoring both registers.
6762 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6764 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
6765 assertion with conditional return.
6767 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6768 Uros Bizjak <ubizjak@gmail.com>
6771 * config/i386/driver-i386.c (host_detect_local_cpu): If
6772 YMM state is not saved by the OS, also clear has_f16c. Move
6773 CPUID 0x80000001 handling before YMM state saving checking.
6775 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
6777 PR rtl-optimization/58960
6778 * haifa-sched.c (alloc_global_sched_pressure_data): New,
6779 factored out from ...
6780 (sched_init): ... here.
6781 (free_global_sched_pressure_data): New, factored out from ...
6782 (sched_finish): ... here.
6783 * sched-int.h (free_global_sched_pressure_data): Declare.
6784 * sched-rgn.c (nr_regions_initial): New static global.
6785 (haifa_find_rgns): Initialize it.
6786 (schedule_region): Disable sched-pressure for the newly
6789 2014-02-17 Richard Biener <rguenther@suse.de>
6791 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
6792 release SSA defs of pattern stmts.
6794 2014-02-17 Richard Biener <rguenther@suse.de>
6796 * tree-inline.c (expand_call_inline): Release the virtual
6797 operand defined by the call we are about to inline.
6799 2014-02-17 Richard Biener <rguenther@suse.de>
6801 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
6803 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
6804 Ilya Tocar <ilya.tocar@intel.com>
6806 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
6807 arguments order in builtin.
6808 (_mm512_permutexvar_epi64): Ditto.
6809 (_mm512_mask_permutexvar_epi64): Ditto
6810 (_mm512_maskz_permutexvar_epi32): Ditto
6811 (_mm512_permutexvar_epi32): Ditto
6812 (_mm512_mask_permutexvar_epi32): Ditto
6814 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6816 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
6817 (p8_vmrgow): Likewise.
6819 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6821 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
6824 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
6827 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
6828 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
6829 into 64-bit and 32-bit moves. On 64-bit moves, add support for
6830 using direct move instructions on ISA 2.07. Also adjust
6831 instruction length for 64-bit.
6832 (mov<mode>_64bit, TFmode/TDmode): Likewise.
6833 (mov<mode>_32bit, TFmode/TDmode): Likewise.
6835 2014-02-15 Alan Modra <amodra@gmail.com>
6839 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
6840 find_replacement on parts of insn rtl that might be reloaded.
6842 2014-02-15 Richard Biener <rguenther@suse.de>
6844 PR tree-optimization/60183
6845 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
6846 (tree_ssa_phiprop): Calculate and free post-dominators.
6848 2014-02-14 Jeff Law <law@redhat.com>
6850 PR rtl-optimization/60131
6851 * ree.c (get_extended_src_reg): New function.
6852 (combine_reaching_defs): Use it rather than assuming location of REG.
6853 (find_and_remove_re): Verify first operand of extension is
6854 a REG before adding the insns to the copy list.
6856 2014-02-14 Roland McGrath <mcgrathr@google.com>
6858 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
6859 * configure: Regenerated.
6860 * config.in: Regenerated.
6861 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
6862 instead of ASM_SHORT.
6864 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
6865 Richard Earnshaw <rearnsha@arm.com>
6867 PR rtl-optimization/59535
6868 * lra-constraints.c (process_alt_operands): Encourage alternative
6869 when unassigned pseudo class is superset of the alternative class.
6870 (inherit_reload_reg): Don't inherit when optimizing for code size.
6871 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
6872 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
6873 modes not less than 4 for Thumb1.
6875 2014-02-14 Kyle McMartin <kyle@redhat.com>
6878 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
6880 2014-02-14 Richard Biener <rguenther@suse.de>
6882 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
6883 (get_frame_arg): Drop the assert with langhook types_compatible_p.
6884 Do not strip INDIRECT_REFs.
6886 2014-02-14 Richard Biener <rguenther@suse.de>
6889 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
6890 DECL_FUNCTION_SPECIFIC_TARGET.
6891 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
6892 * tree-streamer-out.c (pack_ts_target_option): Remove.
6893 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
6894 (write_ts_function_decl_tree_pointers): Do not stream
6895 DECL_FUNCTION_SPECIFIC_TARGET.
6896 * tree-streamer-in.c (unpack_ts_target_option): Remove.
6897 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
6898 (lto_input_ts_function_decl_tree_pointers): Do not stream
6899 DECL_FUNCTION_SPECIFIC_TARGET.
6901 2014-02-14 Jakub Jelinek <jakub@redhat.com>
6903 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
6904 (get_initial_def_for_induction, vectorizable_induction): Ignore
6905 debug stmts when looking for exit_phi.
6906 (vectorizable_live_operation): Fix up condition.
6908 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6910 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
6911 nreverse() because it changes the content of original tree list.
6913 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6915 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
6916 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
6918 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6920 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
6921 GNU coding standards.
6923 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6926 * dwarf2out.c (gen_subprogram_die): Don't call
6927 add_calling_convention_attribute if subr_die is old_die.
6929 2014-02-13 Sharad Singhai <singhai@google.com>
6931 * doc/optinfo.texi: Fix order of nodes.
6933 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
6935 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
6936 operands[2], not operands[3].
6938 2014-02-13 Richard Biener <rguenther@suse.de>
6941 * doc/install.texi (ISL): Update recommended version to 0.12.2,
6942 mention the possibility of an in-tree build.
6943 (CLooG): Update recommended version to 0.18.1, mention the
6944 possibility of an in-tree build and clarify that the ISL
6945 bundled with CLooG does not work.
6947 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6950 * expr.c (compress_float_constant): If x is a hard register,
6951 extend into a pseudo and then move to x.
6953 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
6955 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
6956 caused by bad second argument to warning_at() with -mhotpatch and
6957 nested functions (e.g. with gfortran).
6959 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
6961 * opts.c (option_name): Remove "enabled by default" rider.
6963 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
6965 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
6967 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
6968 Uros Bizjak <ubizjak@gmail.com>
6971 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
6972 * configure: Regenerated.
6974 2014-02-12 Richard Biener <rguenther@suse.de>
6976 * vec.c (vec_prefix::calculate_allocation): Move as
6977 inline variant to vec.h.
6978 (vec_prefix::calculate_allocation_1): New out-of-line version.
6979 * vec.h (vec_prefix::calculate_allocation_1): Declare.
6980 (vec_prefix::m_has_auto_buf): Rename to ...
6981 (vec_prefix::m_using_auto_storage): ... this.
6982 (vec_prefix::calculate_allocation): Inline the easy cases
6983 and dispatch to calculate_allocation_1 which doesn't need the
6985 (va_heap::reserve): Use gcc_checking_assert.
6986 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
6987 m_using_auto_storage.
6988 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
6990 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
6991 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
6992 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
6994 2014-02-12 Richard Biener <rguenther@suse.de>
6996 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
6997 when we found a dependence.
6999 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
7001 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
7003 (maybe_fold_stmt): ... into this new function.
7004 * omp-low.c (lower_omp): Update comment.
7006 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
7009 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
7012 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
7014 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
7015 identifiers in comments.
7016 (cortexa53_extra_costs): Likewise.
7017 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
7018 (cortexa7_extra_costs): Likewise.
7019 (cortexa12_extra_costs): Likewise.
7020 (cortexa15_extra_costs): Likewise.
7021 (v7m_extra_costs): Likewise.
7023 2014-02-12 Richard Biener <rguenther@suse.de>
7026 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
7027 of posix_memalign being successful.
7028 (lower_stmt): Restrict lowering of posix_memalign to when
7029 -ftree-bit-ccp is enabled.
7031 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
7033 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
7035 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
7037 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
7039 PR rtl-optimization/60116
7040 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
7041 other_insn once the combination has been validated.
7043 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
7046 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
7048 * ipa-devirt.c: Include demangle.h
7049 (odr_violation_reported): New static variable.
7050 (add_type_duplicate): Update odr_violations.
7051 (maybe_record_node): Add completep parameter; update it.
7052 (record_target_from_binfo): Add COMPLETEP parameter;
7053 update it as needed.
7054 (possible_polymorphic_call_targets_1): Likewise.
7055 (struct polymorphic_call_target_d): Add nonconstruction_targets;
7056 rename FINAL to COMPLETE.
7057 (record_targets_from_bases): Sanity check we found the binfo;
7058 fix COMPLETEP updating.
7059 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
7060 parameter, fix computing of COMPLETEP.
7061 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
7062 at LTO time do demangling.
7063 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
7064 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
7066 (gimple_get_virt_method_for_binfo): Likewise.
7067 * gimple-fold.h (gimple_get_virt_method_for_binfo,
7068 gimple_get_virt_method_for_vtable): Update prototypes.
7070 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
7073 * genautomata.c (add_presence_absence): Fix typo with
7074 {final_}presence_list.
7076 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
7079 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
7080 for VSX/Altivec vectors that land in GPR registers.
7082 2014-02-11 Richard Henderson <rth@redhat.com>
7083 Jakub Jelinek <jakub@redhat.com>
7086 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
7087 around drhs if type conversion to lacc->type is not useless.
7089 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7091 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
7093 (cortex-a57.cortex-a53): Likewise.
7094 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
7096 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7098 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
7101 2014-02-11 Renlin Li <Renlin.Li@arm.com>
7103 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
7104 add_options_for_arm_vfp3.
7106 2014-02-11 Jeff Law <law@redhat.com>
7109 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
7110 object with an undesirable mode.
7112 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7115 * config/i386/sol2-9.h: New file.
7116 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
7117 *-*-solaris2.9*): Use it.
7119 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7121 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
7122 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
7124 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7126 * config/microblaze/microblaze.c: Extend mcpu version format
7128 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
7130 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
7132 2014-02-10 Richard Henderson <rth@redhat.com>
7135 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
7136 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
7137 ms-abi vs -mno-accumulate-outgoing-args.
7138 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
7139 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
7142 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
7145 * cfgexpand.c (expand_asm_operands): Attach source location to
7146 ASM_INPUT rtx objects.
7147 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
7149 2014-02-10 Nick Clifton <nickc@redhat.com>
7151 * config/mn10300/mn10300.c (popcount): New function.
7152 (mn10300_expand_prologue): Include saved registers in stack usage
7155 2014-02-10 Jeff Law <law@redhat.com>
7158 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
7159 when changing the SET_DEST of a prior insn to avoid an input reload.
7161 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7163 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
7164 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
7165 -mcall-openbsd, or -mcall-linux.
7166 (CC1_ENDIAN_BIG_SPEC): Remove.
7167 (CC1_ENDIAN_LITTLE_SPEC): Remove.
7168 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7169 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
7170 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
7171 and %cc1_endian_default.
7172 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7174 2014-02-10 Richard Biener <rguenther@suse.de>
7176 PR tree-optimization/60115
7177 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
7178 MEM_REF handling. Properly verify that the accesses are not
7179 out of the objects bound.
7181 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7183 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
7186 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
7188 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
7189 proper constants and fix formatting.
7190 (possible_polymorphic_call_targets): Fix formatting.
7192 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
7193 Ilya Tocar <ilya.tocar@intel.com>
7195 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
7196 (_mm512_loadu_epi32): Renamed into...
7197 (_mm512_loadu_si512): This.
7198 (_mm512_storeu_epi32): Renamed into...
7199 (_mm512_storeu_si512): This.
7200 (_mm512_maskz_ceil_ps): Removed.
7201 (_mm512_maskz_ceil_pd): Ditto.
7202 (_mm512_maskz_floor_ps): Ditto.
7203 (_mm512_maskz_floor_pd): Ditto.
7204 (_mm512_floor_round_ps): Ditto.
7205 (_mm512_floor_round_pd): Ditto.
7206 (_mm512_ceil_round_ps): Ditto.
7207 (_mm512_ceil_round_pd): Ditto.
7208 (_mm512_mask_floor_round_ps): Ditto.
7209 (_mm512_mask_floor_round_pd): Ditto.
7210 (_mm512_mask_ceil_round_ps): Ditto.
7211 (_mm512_mask_ceil_round_pd): Ditto.
7212 (_mm512_maskz_floor_round_ps): Ditto.
7213 (_mm512_maskz_floor_round_pd): Ditto.
7214 (_mm512_maskz_ceil_round_ps): Ditto.
7215 (_mm512_maskz_ceil_round_pd): Ditto.
7216 (_mm512_expand_pd): Ditto.
7217 (_mm512_expand_ps): Ditto.
7218 * config/i386/i386.c (ix86_builtins): Remove
7219 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
7220 (bdesc_args): Ditto.
7221 * config/i386/predicates.md (const1256_operand): New.
7222 (const_1_to_2_operand): Ditto.
7223 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7224 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7225 (*avx512pf_gatherpf<mode>sf): Ditto.
7226 (avx512pf_gatherpf<mode>df): Ditto.
7227 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7228 (*avx512pf_gatherpf<mode>df): Ditto.
7229 (avx512pf_scatterpf<mode>sf): Ditto.
7230 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7231 (*avx512pf_scatterpf<mode>sf): Ditto.
7232 (avx512pf_scatterpf<mode>df): Ditto.
7233 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7234 (*avx512pf_scatterpf<mode>df): Ditto.
7235 (avx512f_expand<mode>): Removed.
7236 (<shift_insn><mode>3<mask_name>): Change predicate type.
7238 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7240 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
7241 not at the end of datarefs vector use ordered_remove to avoid
7242 reordering datarefs vector.
7245 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
7246 mark local addressable non-static vars as GOVD_PRIVATE
7247 instead of GOVD_LOCAL.
7248 * omp-low.c (lower_omp_for): Move gimple_bind_vars
7249 and BLOCK_VARS of gimple_bind_block to new_stmt rather
7253 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
7254 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
7255 assume_aligned or alloc_align attributes.
7256 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
7257 arguments. Handle also assume_aligned and alloc_align attributes.
7258 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
7259 calls to functions with assume_aligned or alloc_align attributes.
7260 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
7262 2014-02-08 Terry Guo <terry.guo@arm.com>
7264 * doc/invoke.texi: Document ARM -march=armv7e-m.
7266 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7268 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
7269 flag on __cilkrts_rethrow builtin.
7272 * ipa-cp.c (determine_versionability): Fail at -O0
7273 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
7274 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
7277 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7280 * tree-inline.c (copy_forbidden): Fail for
7281 __attribute__((optimize (0))) functions.
7283 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7285 * varpool.c: Include pointer-set.h.
7286 (varpool_remove_unreferenced_decls): Variables in other partitions
7287 will not be output; be however careful to not lose information
7290 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7292 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
7293 lookup in the vtable constructor.
7295 2014-02-07 Jeff Law <law@redhat.com>
7298 * config/m68k/m68k.md (ashldi_extsi): Turn into a
7299 define_insn_and_split.
7301 * ipa-inline.c (inline_small_functions): Fix typos.
7303 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7305 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
7306 (s390_can_use_return_insn): Declare.
7307 * config/s390/s390.h (EPILOGUE_USES): Define.
7308 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
7310 (s390_chunkify_start): Handle return JUMP_LABELs.
7311 (s390_early_mach): Emit a main_pool instruction on the entry edge.
7312 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
7313 (s390_can_use_return_insn): New functions.
7314 (s390_fix_long_loop_prediction): Handle conditional returns.
7315 (TARGET_SET_UP_BY_PROLOGUE): Define.
7316 * config/s390/s390.md (ANY_RETURN): New code iterator.
7317 (*creturn, *csimple_return, return, simple_return): New patterns.
7319 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7321 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
7322 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
7323 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
7324 REG_CFA_RESTORE list when deciding not to restore a register.
7326 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7328 * config/s390/s390.c: Include tree-pass.h and context.h.
7329 (s390_early_mach): New function, split out from...
7330 (s390_emit_prologue): ...here.
7331 (pass_data_s390_early_mach): New pass structure.
7332 (pass_s390_early_mach): New class.
7333 (s390_option_override): Create and register early_mach pass.
7334 Move to end of file.
7336 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7338 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
7339 to match for the exit block.
7341 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7343 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
7344 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
7345 Reject misaligned operands.
7347 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7349 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
7351 2014-02-07 Richard Biener <rguenther@suse.de>
7354 * gimple-low.c (lower_builtin_posix_memalign): New function.
7355 (lower_stmt): Call it to lower posix_memalign in a way
7356 to make alignment info accessible.
7358 2014-02-07 Jakub Jelinek <jakub@redhat.com>
7361 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
7362 __builtin_setjmp_receiver.
7364 2014-02-07 Richard Biener <rguenther@suse.de>
7367 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
7368 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
7369 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7370 Handle BUILT_IN_POSIX_MEMALIGN.
7371 (find_func_clobbers): Likewise.
7372 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
7373 (call_may_clobber_ref_p_1): Likewise.
7375 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7378 * ipa-devirt.c (record_target_from_binfo): Remove overactive
7381 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7384 * lto-cgraph.c (lto_output_node): Use
7385 symtab_get_symbol_partitioning_class.
7386 (lto_output_varpool_node): likewise.
7387 (symtab_get_symbol_partitioning_class): Move here from
7389 * cgraph.h (symbol_partitioning_class): Likewise.
7390 (symtab_get_symbol_partitioning_class): Declare.
7392 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7394 * ggc.h (ggc_internal_cleared_alloc): New macro.
7395 * vec.h (vec_safe_copy): Handle memory stats.
7396 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
7397 * target-globals.c (save_target_globals): Likewise.
7399 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7402 * expr.c (emit_move_resolve_push): Export; be bit more selective
7403 on when to clear alias set.
7404 * expr.h (emit_move_resolve_push): Declare.
7405 * function.h (struct function): Add tail_call_marked.
7406 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
7407 * config/i386/i386-protos.h (ix86_expand_push): Remove.
7408 * config/i386/i386.md (TImode move expander): De not call
7410 (FP push expanders): Preserve memory attributes.
7411 * config/i386/sse.md (push<mode>1): Remove.
7412 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
7413 (ix86_expand_push): Remove.
7414 * config/i386/mmx.md (push<mode>1): Remove.
7416 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7418 PR rtl-optimization/60030
7419 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
7420 lopart with paradoxical subreg before shifting it up by hprec.
7422 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7424 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
7425 Remove extra newline at end of file.
7426 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
7427 (arm_issue_rate): Handle cortexa57.
7428 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
7429 (cortex-a57.cortex-a53): Likewise.
7431 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7434 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
7435 don't record in REG_FRAME_RELATED_EXPR registers not set in that
7437 (arm_expand_prologue): Adjust all callers.
7438 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
7439 info, registers also at the lowest numbered registers side. Use
7440 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
7444 * var-tracking.c (adjust_mems): Before adding a SET to
7445 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
7447 2014-02-06 Alan Modra <amodra@gmail.com>
7450 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
7451 change SDmode to DDmode when lra_in_progress.
7453 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7456 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
7457 free_data_ref on the dr first, and before goto again also set dr
7458 to the next dr. For simd_lane_access, free old datarefs[i] before
7459 overwriting it. For get_vectype_for_scalar_type failure, don't
7460 free_data_ref if simd_lane_access.
7462 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
7465 * tree.h (opts_for_fn): New inline function.
7466 (opt_for_fn): Define.
7467 * config/i386/i386.c (ix86_function_regparm): Use
7468 opt_for_fn (decl, optimize) instead of optimize.
7470 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
7472 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
7473 for SYMBOL_REF in large memory model.
7475 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7477 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
7479 (cortex-a57): Likewise.
7480 (cortex-a57.cortex-a53): Likewise.
7482 2014-02-06 Yury Gribov <y.gribov@samsung.com>
7483 Kugan Vivekanandarajah <kuganv@linaro.org>
7485 * config/arm/arm.c (arm_vector_alignment_reachable): Check
7487 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
7489 2014-02-06 Richard Biener <rguenther@suse.de>
7491 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
7492 set_loop_copy and initialize_original_copy_tables.
7494 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
7496 * config/aarch64/aarch64-simd.md
7497 (aarch64_ashr_simddi): Change QI to SI.
7499 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7500 Jakub Jelinek <jakub@redhat.com>
7503 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
7506 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7508 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
7509 CODE_FOR_altivec_vpku[hw]um to
7510 CODE_FOR_altivec_vpku[hw]um_direct.
7511 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
7512 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
7513 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
7514 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
7516 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7518 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
7519 generation for -maltivec=be.
7520 (altivec_vsumsws): Simplify redundant test.
7522 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7524 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
7525 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
7526 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
7527 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
7528 gen_altivec_vpkuwum.
7529 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
7531 (altivec_vpks<VI_char>ss): Likewise.
7532 (altivec_vpks<VI_char>us): Likewise.
7533 (altivec_vpku<VI_char>us): Likewise.
7534 (altivec_vpku<VI_char>um): Likewise.
7535 (altivec_vpku<VI_char>um_direct): New (copy of
7536 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
7538 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
7539 target is little endian and -maltivec=be is not specified.
7540 (*altivec_vupkhs<VU_char>_direct): New (copy of
7541 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
7542 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
7543 target is little endian and -maltivec=be is not specified.
7544 (*altivec_vupkls<VU_char>_direct): New (copy of
7545 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
7546 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
7547 little endian and -maltivec=be is not specified.
7548 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
7549 little endian and -maltivec=be is not specified.
7551 2014-02-05 Richard Henderson <rth@redhat.com>
7554 * combine-stack-adj.c: Revert r206943.
7555 * sched-int.h (struct deps_desc): Add last_args_size.
7556 * sched-deps.c (init_deps): Initialize it.
7557 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7558 contain REG_ARGS_SIZE notes.
7560 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7562 * lto-cgraph.c (asm_nodes_output): Make global.
7563 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7564 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7565 (driver_handle_option): Handle OPT_fwpa.
7567 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7570 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7571 a comment typo and formatting issue. If odr_hash hasn't been
7572 created, return vNULL and set *completep to false.
7575 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7576 bb with no successors.
7578 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7581 * doc/invoke.texi (-march): Clarify documentation for ARM.
7585 2014-02-05 Richard Biener <rguenther@suse.de>
7587 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
7588 when not vectorizing because of too many alias checks.
7589 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
7590 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
7592 2014-02-05 Nick Clifton <nickc@redhat.com>
7594 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
7595 accept extended registers in any mode when compiling for the MN10300.
7597 2014-02-05 Yury Gribov <y.gribov@samsung.com>
7599 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
7600 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
7601 sanitization attributes.
7602 (can_inline_edge_p): Likewise.
7603 (sanitize_attrs_match_for_inline_p): New function.
7605 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7607 * ipa-prop.c (detect_type_change): Shor circuit testing of
7608 type changes on THIS pointer.
7610 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
7613 * config/pa/pa.c (legitimize_tls_address): Return original address
7614 if not passed a SYMBOL_REF rtx.
7615 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
7617 (pa_emit_move_sequence): Simplify TLS source operands.
7618 (pa_legitimate_constant_p): Reject all TLS constants.
7619 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
7620 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
7622 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7624 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
7625 groups when we know they are controlled by LTO.
7626 * varasm.c (default_binds_local_p_1): If object is in other partition,
7627 it will be resolved locally.
7629 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7631 * config/host-linux.c (linux_gt_pch_use_address): Don't
7632 use SSIZE_MAX because it is not always defined.
7634 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
7637 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
7638 threshold for pseudo splitting.
7639 (update_ebb_live_info): Process call argument hard registers and
7640 hard registers from insn definition too.
7641 (max_small_class_regs_num): New constant.
7642 (inherit_in_ebb): Update live hard regs through EBBs. Update
7643 reloads_num only for small register classes. Don't split for
7646 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
7649 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
7652 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7654 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
7657 2014-02-04 Marek Polacek <polacek@redhat.com>
7659 * gdbinit.in (pel): Define.
7661 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7663 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
7666 2014-02-04 Richard Biener <rguenther@suse.de>
7669 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
7670 in function context local.
7671 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
7672 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
7673 similar to LTO_imported_decl_ref.
7675 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7677 PR tree-optimization/60002
7678 * cgraphclones.c (build_function_decl_skip_args): Clear
7681 PR tree-optimization/60023
7682 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
7683 false to gsi_replace.
7684 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
7685 has been in some EH region and vec_stmt could throw, add
7686 vec_stmt into the same EH region.
7687 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
7688 has no lhs, ignore it.
7689 * internal-fn.c (expand_MASK_LOAD): Likewise.
7692 * tree-inline.c (copy_forbidden): Fail for
7693 __attribute__((optimize (0))) functions.
7696 * omp-low.c (simd_clone_struct_copy): If from->inbranch
7697 is set, copy one less argument.
7698 (expand_simd_clones): Don't subtract clone_info->inbranch
7699 from simd_clone_struct_alloc argument.
7701 PR rtl-optimization/57915
7702 * recog.c (simplify_while_replacing): If all unary/binary/relational
7703 operation arguments are constant, attempt to simplify those.
7706 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
7707 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
7709 2014-02-04 Richard Biener <rguenther@suse.de>
7711 PR tree-optimization/60012
7712 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
7713 TBAA disambiguation to all DDRs.
7715 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7718 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
7719 (LINK_SPEC): Use it for -shared, -shared-libgcc.
7721 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7724 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
7726 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7728 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
7729 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
7731 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7734 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
7735 to figure out targets of polymorphic calls with known decl.
7736 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7737 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
7738 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
7739 (get_polymorphic_call_info): ... here.
7740 (get_polymorphic_call_info_from_invariant): New function.
7742 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7744 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
7745 lookup via vtable pointer; check for type consistency
7746 and turn inconsitent facts into UNREACHABLE.
7747 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7748 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
7749 type inconsistent querries; return UNREACHABLE instead.
7751 2014-02-03 Richard Henderson <rth@twiddle.net>
7754 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
7755 already processed this node.
7756 (normalize_one_pred_1): Pass along mark_set.
7757 (normalize_one_pred): Create and destroy a pointer_set_t.
7758 (normalize_one_pred_chain): Likewise.
7760 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
7762 PR gcov-profile/58602
7763 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
7765 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7768 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
7769 -fno-devirtualize; try to devirtualize by the knowledge of
7770 virtual table pointer given by aggregate propagation.
7771 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7772 (ipa_print_node_jump_functions): Dump also offset that
7773 is relevant for polymorphic calls.
7774 (determine_known_aggregate_parts): Add arg_type parameter; use it
7775 instead of determining the type from pointer type.
7776 (ipa_compute_jump_functions_for_edge): Update call of
7777 determine_known_aggregate_parts.
7778 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
7779 (gimple_get_virt_method_for_binfo): ... here; simplify using
7780 vtable_pointer_value_to_vtable.
7781 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
7782 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
7783 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
7784 (vtable_pointer_value_to_vtable): Break out from ...; handle also
7786 (vtable_pointer_value_to_binfo): ... here.
7787 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
7789 2014-02-03 Teresa Johnson <tejohnson@google.com>
7791 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
7792 redef of outer loop index variable.
7794 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
7798 * doc/extend.texi (Function Attributes): Typo.
7800 2014-02-03 Cong Hou <congh@google.com>
7802 PR tree-optimization/60000
7803 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
7804 if the vectorized statement is a store. A store statement can only
7805 appear at the end of pattern statements.
7807 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7809 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
7810 (ix86_option_override_internal): Default long double to 64-bit for
7811 32-bit Bionic and to 128-bit for 64-bit Bionic.
7813 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
7814 TARGET_LONG_DOUBLE_128 is true.
7815 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
7817 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
7818 (mlong-double-64): Negate -mlong-double-128.
7819 (mlong-double-128): New option.
7821 * config/i386/i386-c.c (ix86_target_macros): Define
7822 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
7824 * doc/invoke.texi: Document -mlong-double-128.
7826 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7828 PR rtl-optimization/60024
7829 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
7831 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
7833 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
7835 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7837 PR rtl-optimization/57662
7838 * sel-sched.c (code_motion_path_driver): Do not mark already not
7839 existing blocks in the visiting bitmap.
7841 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7843 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
7844 on the insn being emitted.
7846 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
7847 Will Deacon <will.deacon@arm.com>
7849 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
7851 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7853 * config/arm/arm-tables.opt: Regenerate.
7855 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7857 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
7858 for vector types other than V16QImode.
7859 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
7860 define_expand, and call altivec_expand_vec_perm_le when producing
7861 code with little endian element order.
7862 (*altivec_vperm_<mode>_internal): New insn having previous
7863 behavior of altivec_vperm_<mode>.
7864 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
7865 altivec_expand_vec_perm_le when producing code with little endian
7867 (*altivec_vperm_<mode>_uns_internal): New insn having previous
7868 behavior of altivec_vperm_<mode>_uns.
7870 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7872 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
7873 (altivec_vsumsws): Add handling for -maltivec=be with a little
7875 (altivec_vsumsws_direct): New.
7876 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
7877 gen_altivec_vsumsws.
7879 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7881 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
7882 vtable_pointer_value_to_binfo): New functions.
7883 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
7884 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
7886 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
7888 * config/nios2/nios2.md (load_got_register): Initialize GOT
7889 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
7890 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
7892 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7894 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
7895 preserverd by passthrough, do not propagate the type.
7897 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7899 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
7900 (mips_atomic_assign_expand_fenv): New function.
7901 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
7903 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7905 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
7906 (__builtin_mips_set_fcsr): Likewise.
7907 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
7908 MIPS_USI_FTYPE_VOID.
7909 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
7910 (mips16_expand_set_fcsr): Likewise.
7911 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
7912 (mips16_set_fcsr_stub): Likewise.
7913 (mips16_get_fcsr_one_only_stub): New class.
7914 (mips16_set_fcsr_one_only_stub): Likewise.
7915 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
7916 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
7917 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
7918 (hard_float): New availability predicate.
7919 (mips_builtins): Add get_fcsr and set_fcsr.
7920 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
7921 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
7922 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
7923 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
7924 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
7927 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7929 * config/mips/mips.c (mips_one_only_stub): New class.
7930 (mips_need_mips16_rdhwr_p): Replace with...
7931 (mips16_rdhwr_stub): ...this new variable.
7932 (mips16_stub_call_address): New function.
7933 (mips16_rdhwr_one_only_stub): New class.
7934 (mips_expand_thread_pointer): Use mips16_stub_call_address.
7935 (mips_output_mips16_rdhwr): Delete.
7936 (mips_finish_stub): New function.
7937 (mips_code_end): Use it to handle rdhwr stubs.
7939 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
7942 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
7943 when calculating size of integer atomic types.
7945 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
7947 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
7949 2014-02-01 Jakub Jelinek <jakub@redhat.com>
7951 PR tree-optimization/60003
7952 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
7953 * profile.c (branch_prob): Use gimple_call_builtin_p
7954 to check for BUILT_IN_SETJMP_RECEIVER.
7955 * tree-inline.c (copy_bb): Call notice_special_calls.
7957 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
7960 * lra-constraints.c (process_alt_operands): Update reload_sum only
7963 2014-01-31 Richard Henderson <rth@redhat.com>
7966 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
7967 until after else_eh is processed.
7969 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7971 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
7972 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
7973 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
7974 in smmintrin.h, remove them.
7975 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
7976 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
7977 * config/i386/i386.md (ROUND_SAE): Fix value.
7978 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
7979 (const48_operand): New.
7980 * config/i386/subst.md (round), (round_expand): Use
7981 const_4_or_8_to_11_operand.
7982 (round_saeonly), (round_saeonly_expand): Use const48_operand.
7984 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7986 * config/i386/constraints.md (Yk): Swap meaning with k.
7987 * config/i386/i386.md (movhi_internal): Change Yk to k.
7988 (movqi_internal): Ditto.
7989 (*k<logic><mode>): Ditto.
7992 (kandn<mode>): Ditto.
7993 (*<code>hi_1): Ditto.
7994 (*<code>qi_1): Ditto.
7995 (kxnor<mode>): Ditto.
7996 (kortestzhi): Ditto.
7997 (kortestchi): Ditto.
7999 (*one_cmplhi2_1): Ditto.
8000 (*one_cmplqi2_1): Ditto.
8001 * config/i386/sse.md (): Change k to Yk.
8002 (avx512f_load<mode>_mask): Ditto.
8003 (avx512f_blendm<mode>): Ditto.
8004 (avx512f_store<mode>_mask): Ditto.
8005 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
8006 (avx512f_storedqu<mode>_mask): Ditto.
8007 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
8009 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
8010 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
8011 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
8012 (avx512f_maskcmp<mode>3): Ditto.
8013 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
8014 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
8015 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
8016 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
8017 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
8018 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
8019 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
8020 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
8021 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
8022 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
8023 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
8024 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
8025 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
8026 (vec_extract_lo_<mode>_maskm): Ditto.
8027 (vec_extract_hi_<mode>_maskm): Ditto.
8028 (avx512f_vternlog<mode>_mask): Ditto.
8029 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
8030 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
8031 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
8032 (avx512f_<code>v8div16qi2_mask): Ditto.
8033 (avx512f_<code>v8div16qi2_mask_store): Ditto.
8034 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
8035 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
8036 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
8037 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
8038 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8039 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8040 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8041 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8042 (avx512cd_maskb_vec_dupv8di): Ditto.
8043 (avx512cd_maskw_vec_dupv16si): Ditto.
8044 (avx512f_vpermi2var<mode>3_maskz): Ditto.
8045 (avx512f_vpermi2var<mode>3_mask): Ditto.
8046 (avx512f_vpermi2var<mode>3_mask): Ditto.
8047 (avx512f_vpermt2var<mode>3_maskz): Ditto.
8048 (*avx512f_gathersi<mode>): Ditto.
8049 (*avx512f_gathersi<mode>_2): Ditto.
8050 (*avx512f_gatherdi<mode>): Ditto.
8051 (*avx512f_gatherdi<mode>_2): Ditto.
8052 (*avx512f_scattersi<mode>): Ditto.
8053 (*avx512f_scatterdi<mode>): Ditto.
8054 (avx512f_compress<mode>_mask): Ditto.
8055 (avx512f_compressstore<mode>_mask): Ditto.
8056 (avx512f_expand<mode>_mask): Ditto.
8057 * config/i386/subst.md (mask): Change k to Yk.
8058 (mask_scalar_merge): Ditto.
8061 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
8063 * doc/extend.texi (Vector Extensions): Document ?: in C++.
8065 2014-01-31 Richard Biener <rguenther@suse.de>
8068 * builtins.c (fold_builtin_memory_op): Make sure to not
8069 use a floating-point mode or a boolean or enumeral type for
8072 2014-01-30 DJ Delorie <dj@redhat.com>
8074 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
8075 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
8076 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
8077 whenever main() has an epilogue.
8079 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8081 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
8082 unused variable "field".
8083 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
8084 (vsx_mergeh_<mode>): Likewise.
8085 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
8086 (altivec_vmrghh): Likewise.
8087 (altivec_vmrghw): Likewise.
8088 (altivec_vmrglb): Likewise.
8089 (altivec_vmrglh): Likewise.
8090 (altivec_vmrglw): Likewise.
8091 (altivec_vspltb): Add missing uses.
8092 (altivec_vsplth): Likewise.
8093 (altivec_vspltw): Likewise.
8094 (altivec_vspltsf): Likewise.
8096 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8099 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
8100 frame related instructions.
8102 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
8104 PR rtl-optimization/59959
8105 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
8106 any reload of register whose subreg is invalid.
8108 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8110 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
8111 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
8112 Add missing return type - void.
8114 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8116 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
8117 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
8118 remove element index adjustment for endian (now handled in vsx.md
8120 (altivec_expand_vec_perm_const): Use
8121 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
8122 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
8123 (vsx_xxspltw_<mode>): Adjust element index for little endian.
8124 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
8125 define_expand and a new define_insn *altivec_vspltb_internal;
8126 adjust for -maltivec=be on a little endian target.
8127 (altivec_vspltb_direct): New.
8128 (altivec_vsplth): Divide into a define_expand and a new
8129 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
8130 little endian target.
8131 (altivec_vsplth_direct): New.
8132 (altivec_vspltw): Divide into a define_expand and a new
8133 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
8134 little endian target.
8135 (altivec_vspltw_direct): New.
8136 (altivec_vspltsf): Divide into a define_expand and a new
8137 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
8138 a little endian target.
8140 2014-01-30 Richard Biener <rguenther@suse.de>
8142 PR tree-optimization/59993
8143 * tree-ssa-forwprop.c (associate_pointerplus): Check we
8144 can propagate form the earlier stmt and avoid the transform
8145 when the intermediate result is needed.
8147 2014-01-30 Alangi Derick <alangiderick@gmail.com>
8149 * README.Portability: Fix typo.
8151 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
8153 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
8154 comparison_operator with ordered_comparison_operator.
8156 2014-01-30 Nick Clifton <nickc@redhat.com>
8158 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
8159 Rename to mn10300_store_multiple_regs.
8160 * config/mn10300/mn10300.c: Likewise.
8161 * config/mn10300/mn10300.md (store_movm): Fix typo: call
8162 store_multiple_regs.
8163 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
8164 Call mn10300_store_multiple_regs.
8166 2014-01-30 Nick Clifton <nickc@redhat.com>
8167 DJ Delorie <dj@redhat.com>
8169 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
8170 %fp 2 to keep registers after it properly word-aligned.
8171 (rl78_alloc_physical_registers_umul): Handle the case where both
8172 input operands are the same.
8174 2014-01-30 Richard Biener <rguenther@suse.de>
8176 PR tree-optimization/59903
8177 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
8180 2014-01-30 Jason Merrill <jason@redhat.com>
8183 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
8186 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
8188 2014-01-30 Richard Biener <rguenther@suse.de>
8190 PR tree-optimization/59951
8191 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
8193 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
8196 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
8197 SFmode to DFmode case.
8199 2014-01-29 DJ Delorie <dj@redhat.com>
8201 * config/msp430/msp430.opt (-minrt): New.
8202 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
8204 (ENDFILE_SPEC): Likewise.
8206 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
8208 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
8209 (estimate_function_body_sizes): Use it.
8211 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
8214 * dwarf2out.c (is_cxx_auto): New.
8215 (is_base_type): Use it.
8216 (gen_type_die_with_usage): Likewise.
8218 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8220 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
8221 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
8222 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
8223 -maltivec=be with LE targets.
8224 (vsx_mergeh_<mode>): Likewise.
8225 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
8226 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
8227 (altivec_vmrghb): Replace with define_expand and new
8228 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
8229 (altivec_vmrghb_direct): New define_insn.
8230 (altivec_vmrghh): Replace with define_expand and new
8231 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
8232 (altivec_vmrghh_direct): New define_insn.
8233 (altivec_vmrghw): Replace with define_expand and new
8234 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
8235 (altivec_vmrghw_direct): New define_insn.
8236 (*altivec_vmrghsf): Adjust for endianness.
8237 (altivec_vmrglb): Replace with define_expand and new
8238 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
8239 (altivec_vmrglb_direct): New define_insn.
8240 (altivec_vmrglh): Replace with define_expand and new
8241 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
8242 (altivec_vmrglh_direct): New define_insn.
8243 (altivec_vmrglw): Replace with define_expand and new
8244 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
8245 (altivec_vmrglw_direct): New define_insn.
8246 (*altivec_vmrglsf): Adjust for endianness.
8247 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8248 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8249 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8250 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8251 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8252 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8253 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8254 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8256 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
8258 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
8259 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
8262 2014-01-29 Richard Biener <rguenther@suse.de>
8264 PR tree-optimization/58742
8265 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
8266 associate_pointerplus_align.
8267 (associate_pointerplus_diff): New function.
8268 (associate_pointerplus): Likewise. Call associate_pointerplus_align
8269 and associate_pointerplus_diff.
8271 2014-01-29 Richard Biener <rguenther@suse.de>
8273 * lto-streamer.h (LTO_major_version): Bump to 3.
8274 (LTO_minor_version): Reset to 0.
8276 2014-01-29 Renlin Li <Renlin.Li@arm.com>
8278 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
8279 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
8280 (arm_file_start): Generate correct asm header for armv7ve.
8281 * config/arm/bpabi.h: Add multilib support for armv7ve.
8282 * config/arm/driver-arm.c: Change the architectures of cortex-a7
8283 and cortex-a15 to armv7ve.
8284 * config/arm/t-aprofile: Add multilib support for armv7ve.
8285 * doc/invoke.texi: Document -march=armv7ve.
8287 2014-01-29 Richard Biener <rguenther@suse.de>
8289 PR tree-optimization/58742
8290 * tree-ssa-forwprop.c (associate_plusminus): Return true
8291 if we changed sth, defer EH cleanup to ...
8292 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
8293 (simplify_mult): New function.
8295 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8298 PR tree-optimization/59920
8299 * tree.c (build_common_builtin_nodes): Remove
8300 __builtin_setjmp_dispatcher initialization.
8301 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
8302 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
8303 instead of gsi_after_labels + manually skipping debug stmts.
8304 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
8305 ignore bbs with IFN_ABNORMAL_DISPATCHER.
8306 * tree-inline.c (copy_edges_for_bb): Remove
8307 can_make_abnormal_goto argument, instead add abnormal_goto_dest
8308 argument. Ignore computed_goto_p stmts. Don't call
8309 make_abnormal_goto_edges. If a call might need abnormal edges
8310 for non-local gotos, see if it already has an edge to
8311 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
8312 with true argument, don't do anything then, otherwise add
8313 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
8314 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
8316 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
8317 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
8318 (lower_stmt): Don't set data->calls_builtin_setjmp.
8319 (lower_builtin_setjmp): Adjust comment.
8320 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
8321 * tree-cfg.c (found_computed_goto): Remove.
8322 (factor_computed_gotos): Remove.
8323 (make_goto_expr_edges): Return bool, true for computed gotos.
8324 Don't call make_abnormal_goto_edges.
8325 (build_gimple_cfg): Don't set found_computed_goto, don't call
8326 factor_computed_gotos.
8327 (computed_goto_p): No longer static.
8328 (make_blocks): Don't set found_computed_goto.
8329 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
8330 (make_edges): If make_goto_expr_edges returns true, push bb
8331 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
8332 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
8333 vector. Record mapping between bbs and OpenMP regions if there
8334 are any, adjust make_gimple_omp_edges caller. Call
8335 handle_abnormal_edges.
8336 (make_abnormal_goto_edges): Remove.
8337 * tree-cfg.h (make_abnormal_goto_edges): Remove.
8338 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
8339 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
8340 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
8341 * internal-fn.def (ABNORMAL_DISPATCHER): New.
8342 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
8343 filling *region also set *region_idx to (*region)->entry->index.
8346 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
8347 For REGs set ORIGINAL_REGNO.
8349 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
8351 * doc/md.texi: Mention that a target shouldn't implement
8352 vec_widen_(s|u)mul_even/odd pair if it is less efficient
8355 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8357 PR tree-optimization/59594
8358 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
8359 a copy of the datarefs vector rather than the vector itself.
8361 2014-01-28 Jason Merrill <jason@redhat.com>
8364 * dwarf2out.c (auto_die): New static.
8365 (gen_type_die_with_usage): Handle C++1y 'auto'.
8366 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
8369 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
8372 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
8373 (SPEC_X32): Likewise.
8374 (SPEC_64): Likewise.
8375 * config/i386/i386.c (ix86_option_override_internal): Turn off
8376 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
8378 (x86_file_start): Output .code16gcc for TARGET_16BIT.
8379 * config/i386/i386.h (TARGET_16BIT): New macro.
8380 (TARGET_16BIT_P): Likewise.
8381 * config/i386/i386.opt: Add m16.
8382 * doc/invoke.texi: Document -m16.
8384 2014-01-28 Jakub Jelinek <jakub@redhat.com>
8386 PR preprocessor/59935
8387 * input.c (location_get_source_line): Bail out on when line number
8388 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
8390 2014-01-28 Richard Biener <rguenther@suse.de>
8392 PR tree-optimization/58742
8393 * tree-ssa-forwprop.c (associate_plusminus): Handle
8394 pointer subtraction of the form (T)(P + A) - (T)P.
8396 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8398 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
8401 2014-01-28 Richard Biener <rguenther@suse.de>
8404 2014-01-28 Richard Biener <rguenther@suse.de>
8406 PR rtl-optimization/45364
8407 PR rtl-optimization/59890
8408 * var-tracking.c (local_get_addr_clear_given_value): Handle
8409 already cleared slot.
8410 (val_reset): Handle not allocated local_get_addr_cache.
8411 (vt_find_locations): Use post-order on the inverted CFG.
8413 2014-01-28 Richard Biener <rguenther@suse.de>
8415 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
8417 2014-01-28 Richard Biener <rguenther@suse.de>
8419 PR rtl-optimization/45364
8420 PR rtl-optimization/59890
8421 * var-tracking.c (local_get_addr_clear_given_value): Handle
8422 already cleared slot.
8423 (val_reset): Handle not allocated local_get_addr_cache.
8424 (vt_find_locations): Use post-order on the inverted CFG.
8426 2014-01-28 Alan Modra <amodra@gmail.com>
8428 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
8429 * configure.ac <recursive call for build != host>: Define
8430 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
8431 and LD_FOR_BUILD too.
8432 * configure: Regenerate.
8434 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
8436 * config/i386/i386.c (get_builtin_code_for_version): Separate
8437 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
8438 Broadwell from Haswell.
8440 2014-01-27 Steve Ellcey <sellcey@mips.com>
8442 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
8443 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
8444 * config/mips/mips.c (mips_option_override): Change setting
8446 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
8447 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
8448 Change from Mask to Var.
8450 2014-01-27 Jeff Law <law@redhat.com>
8452 * ipa-inline.c (inline_small_functions): Fix typo.
8454 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
8456 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
8457 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
8458 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
8459 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
8460 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
8461 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
8462 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
8463 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
8464 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
8465 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
8466 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
8467 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
8468 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
8469 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
8470 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
8471 (_mm512_storeu_epi64): Ditto.
8472 (_mm512_cmpge_epi32_mask): Ditto.
8473 (_mm512_cmpge_epu32_mask): Ditto.
8474 (_mm512_cmpge_epi64_mask): Ditto.
8475 (_mm512_cmpge_epu64_mask): Ditto.
8476 (_mm512_cmple_epi32_mask): Ditto.
8477 (_mm512_cmple_epu32_mask): Ditto.
8478 (_mm512_cmple_epi64_mask): Ditto.
8479 (_mm512_cmple_epu64_mask): Ditto.
8480 (_mm512_cmplt_epi32_mask): Ditto.
8481 (_mm512_cmplt_epu32_mask): Ditto.
8482 (_mm512_cmplt_epi64_mask): Ditto.
8483 (_mm512_cmplt_epu64_mask): Ditto.
8484 (_mm512_cmpneq_epi32_mask): Ditto.
8485 (_mm512_cmpneq_epu32_mask): Ditto.
8486 (_mm512_cmpneq_epi64_mask): Ditto.
8487 (_mm512_cmpneq_epu64_mask): Ditto.
8488 (_mm512_expand_pd): Ditto.
8489 (_mm512_expand_ps): Ditto.
8490 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
8491 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
8492 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
8493 * config/i386/i386.c (ix86_builtins): Add
8494 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
8495 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
8496 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
8497 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
8498 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
8499 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
8500 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
8501 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
8502 IX86_BUILTIN_PMOVUSQW512_MEM.
8503 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
8504 __builtin_ia32_pmovsqd512mem_mask,
8505 __builtin_ia32_pmovqd512mem_mask,
8506 __builtin_ia32_pmovusqw512mem_mask,
8507 __builtin_ia32_pmovsqw512mem_mask,
8508 __builtin_ia32_pmovqw512mem_mask,
8509 __builtin_ia32_pmovusdw512mem_mask,
8510 __builtin_ia32_pmovsdw512mem_mask,
8511 __builtin_ia32_pmovdw512mem_mask,
8512 __builtin_ia32_pmovqb512mem_mask,
8513 __builtin_ia32_pmovusqb512mem_mask,
8514 __builtin_ia32_pmovsqb512mem_mask,
8515 __builtin_ia32_pmovusdb512mem_mask,
8516 __builtin_ia32_pmovsdb512mem_mask,
8517 __builtin_ia32_pmovdb512mem_mask.
8518 (bdesc_args): Add __builtin_ia32_expanddf512,
8519 __builtin_ia32_expandsf512.
8520 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
8521 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
8522 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
8523 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
8524 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
8525 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
8526 (avx512f_<code>v8div16qi2_mask_store): This.
8527 (avx512f_expand<mode>): New.
8529 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
8531 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
8533 (_mm512_mask_prefetch_i64gather_pd): Ditto.
8534 (_mm512_prefetch_i32scatter_pd): Ditto.
8535 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
8536 (_mm512_prefetch_i64scatter_pd): Ditto.
8537 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
8538 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
8539 (_mm512_mask_prefetch_i64gather_ps): Ditto.
8540 (_mm512_prefetch_i32scatter_ps): Ditto.
8541 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
8542 (_mm512_prefetch_i64scatter_ps): Ditto.
8543 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
8544 * config/i386/i386-builtin-types.def: Define
8545 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
8546 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
8547 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
8548 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
8549 IX86_BUILTIN_SCATTERPFQPD.
8550 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
8551 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
8552 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
8553 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
8554 __builtin_ia32_scatterpfqps.
8555 (ix86_expand_builtin): Expand new built-ins.
8556 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8557 fix memory access data type.
8558 (*avx512pf_gatherpf<mode>_mask): Ditto.
8559 (*avx512pf_gatherpf<mode>): Ditto.
8560 (avx512pf_scatterpf<mode>): Ditto.
8561 (*avx512pf_scatterpf<mode>_mask): Ditto.
8562 (*avx512pf_scatterpf<mode>): Ditto.
8563 (GATHER_SCATTER_SF_MEM_MODE): New.
8564 (avx512pf_gatherpf<mode>df): Ditto.
8565 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8566 (*avx512pf_scatterpf<mode>df): Ditto.
8568 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8571 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8572 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8575 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8577 * common/config/arm/arm-common.c
8578 (arm_rewrite_mcpu): Handle multiple names.
8580 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8582 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8584 * gimple-builder.h (create_gimple_tmp): Delete.
8586 2014-01-27 Christian Bruel <christian.bruel@st.com>
8588 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
8591 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
8593 * config/pa/pa.md (call): Generate indirect long calls to non-local
8594 functions when outputing 32-bit code.
8595 (call_value): Likewise except for special call to buggy powf function.
8597 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
8598 portable runtime and PIC indirect calls.
8599 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
8600 and PIC call sequences. Use ldo instead of blr to set return register
8601 in PIC call sequence.
8603 2014-01-25 Walter Lee <walt@tilera.com>
8605 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
8606 avoid clobbering a live register.
8608 2014-01-25 Walter Lee <walt@tilera.com>
8610 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
8611 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
8612 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
8613 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
8615 2014-01-25 Walter Lee <walt@tilera.com>
8617 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
8618 arguments on even registers.
8619 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
8621 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
8622 (BIGGEST_ALIGNMENT): Ditto.
8623 (BIGGEST_FIELD_ALIGNMENT): Ditto.
8625 2014-01-25 Walter Lee <walt@tilera.com>
8627 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
8628 insns before bundling.
8629 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8631 2014-01-25 Walter Lee <walt@tilera.com>
8633 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
8634 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
8635 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
8637 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8639 * config/mips/constraints.md (kl): Delete.
8640 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
8641 define expands, using...
8642 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
8643 instructions for MIPS16.
8644 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
8645 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
8647 2014-01-25 Walter Lee <walt@tilera.com>
8649 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
8653 2014-01-25 Walter Lee <walt@tilera.com>
8655 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
8656 (TARGET_EXPAND_TO_RTL_HOOK): Define.
8658 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8660 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
8663 2014-01-25 Jakub Jelinek <jakub@redhat.com>
8665 * print-rtl.c (in_call_function_usage): New var.
8666 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
8667 EXPR_LIST mode as mode and not as reg note name.
8670 * cfgloopmanip.c (copy_loop_info): If
8671 loop->warned_aggressive_loop_optimizations, make sure
8672 the flag is set in target loop too.
8674 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
8676 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
8678 * builtins.def: Likewise.
8679 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
8680 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
8681 * ira.c (ira_setup_eliminable_regset): Likewise.
8682 * omp-low.c (gate_expand_omp): Likewise.
8683 (execute_lower_omp): Likewise.
8684 (diagnose_sb_0): Likewise.
8685 (gate_diagnose_omp_blocks): Likewise.
8686 (simd_clone_clauses_extract): Likewise.
8689 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8691 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
8692 correction for little endian...
8693 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
8696 2014-01-24 Jeff Law <law@redhat.com>
8698 PR tree-optimization/59919
8699 * tree-vrp.c (find_assert_locations_1): Do not register asserts
8700 for non-returning calls.
8702 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
8704 * common/config/aarch64/aarch64-common.c
8705 (aarch64_rewrite_mcpu): Handle multiple names.
8706 * config/aarch64/aarch64.h
8707 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8709 2014-01-24 Dodji Seketeli <dodji@redhat.com>
8711 * input.c (add_file_to_cache_tab): Handle the case where fopen
8714 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
8717 * config/i386/i386.md (pushsf splitter): Get stack adjustment
8718 from push operand if code of push isn't PRE_DEC.
8720 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8723 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
8724 -mquad-memory-atomic. Update -mquad-memory documentation to say
8725 it is only used for non-atomic loads/stores.
8727 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
8728 -mquad-memory or -mquad-memory-atomic switches.
8730 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
8731 -mquad-memory-atomic to ISA 2.07 support.
8733 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
8734 to separate support of normal quad word memory operations (ldq, stq)
8735 from the atomic quad word memory operations.
8737 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
8738 support to separate non-atomic quad word operations from atomic
8739 quad word operations. Disable non-atomic quad word operations in
8740 little endian mode so that we don't have to swap words after the
8741 load and before the store.
8742 (quad_load_store_p): Add comment about atomic quad word support.
8743 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
8744 options printed with -mdebug=reg.
8746 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
8747 -mquad-memory-atomic as the test for whether we have quad word
8748 atomic instructions.
8749 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
8750 or -mp8-vector are used, allow byte/half-word atomic operations.
8752 * config/rs6000/sync.md (load_lockedti): Insure that the address
8753 is a proper indexed or indirect address for the lqarx instruction.
8754 On little endian systems, swap the hi/lo registers after the lqarx
8756 (load_lockedpti): Use indexed_or_indirect_operand predicate to
8757 insure the address is valid for the lqarx instruction.
8758 (store_conditionalti): Insure that the address is a proper indexed
8759 or indirect address for the stqcrx. instruction. On little endian
8760 systems, swap the hi/lo registers before doing the stqcrx.
8762 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
8763 insure the address is valid for the stqcrx. instruction.
8765 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8766 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
8767 type of quad memory support is available.
8769 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
8772 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
8773 there is a danger of looping.
8775 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
8777 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8778 force flag_ira_loop_pressure if set via command line.
8780 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8782 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
8783 (ashr_simd): New builtin handling DI mode.
8784 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
8785 (aarch64_sshr_simddi): New match pattern.
8786 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
8787 (vshrd_n_s64): Likewise.
8788 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
8790 2014-01-23 Nick Clifton <nickc@redhat.com>
8792 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
8793 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
8794 favour of mcu specific scripts.
8795 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
8798 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8799 Alex Velenko <Alex.Velenko@arm.com>
8801 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
8802 (vaddv_s16): Likewise.
8803 (vaddv_s32): Likewise.
8804 (vaddv_u8): Likewise.
8805 (vaddv_u16): Likewise.
8806 (vaddv_u32): Likewise.
8807 (vaddvq_s8): Likewise.
8808 (vaddvq_s16): Likewise.
8809 (vaddvq_s32): Likewise.
8810 (vaddvq_s64): Likewise.
8811 (vaddvq_u8): Likewise.
8812 (vaddvq_u16): Likewise.
8813 (vaddvq_u32): Likewise.
8814 (vaddvq_u64): Likewise.
8815 (vaddv_f32): Likewise.
8816 (vaddvq_f32): Likewise.
8817 (vaddvq_f64): Likewise.
8818 (vmaxv_f32): Likewise.
8819 (vmaxv_s8): Likewise.
8820 (vmaxv_s16): Likewise.
8821 (vmaxv_s32): Likewise.
8822 (vmaxv_u8): Likewise.
8823 (vmaxv_u16): Likewise.
8824 (vmaxv_u32): Likewise.
8825 (vmaxvq_f32): Likewise.
8826 (vmaxvq_f64): Likewise.
8827 (vmaxvq_s8): Likewise.
8828 (vmaxvq_s16): Likewise.
8829 (vmaxvq_s32): Likewise.
8830 (vmaxvq_u8): Likewise.
8831 (vmaxvq_u16): Likewise.
8832 (vmaxvq_u32): Likewise.
8833 (vmaxnmv_f32): Likewise.
8834 (vmaxnmvq_f32): Likewise.
8835 (vmaxnmvq_f64): Likewise.
8836 (vminv_f32): Likewise.
8837 (vminv_s8): Likewise.
8838 (vminv_s16): Likewise.
8839 (vminv_s32): Likewise.
8840 (vminv_u8): Likewise.
8841 (vminv_u16): Likewise.
8842 (vminv_u32): Likewise.
8843 (vminvq_f32): Likewise.
8844 (vminvq_f64): Likewise.
8845 (vminvq_s8): Likewise.
8846 (vminvq_s16): Likewise.
8847 (vminvq_s32): Likewise.
8848 (vminvq_u8): Likewise.
8849 (vminvq_u16): Likewise.
8850 (vminvq_u32): Likewise.
8851 (vminnmv_f32): Likewise.
8852 (vminnmvq_f32): Likewise.
8853 (vminnmvq_f64): Likewise.
8855 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8857 * config/aarch64/aarch64-simd.md
8858 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
8859 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
8860 (*aarch64_mul3_elt<mode>): Likewise.
8861 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
8862 (*aarch64_mul3_elt_to_64v2df): Likewise.
8863 (*aarch64_mla_elt<mode>): Likewise.
8864 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8865 (*aarch64_mls_elt<mode>): Likewise.
8866 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8867 (*aarch64_fma4_elt<mode>): Likewise.
8868 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8869 (*aarch64_fma4_elt_to_64v2df): Likewise.
8870 (*aarch64_fnma4_elt<mode>): Likewise.
8871 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8872 (*aarch64_fnma4_elt_to_64v2df): Likewise.
8873 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
8874 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
8875 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8876 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8877 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
8878 (aarch64_sqdmull_lane<mode>_internal): Likewise.
8879 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
8881 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
8883 * config/aarch64/aarch64-simd.md
8884 (aarch64_be_checked_get_lane<mode>): New define_expand.
8885 * config/aarch64/aarch64-simd-builtins.def
8886 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
8887 New builtin definition.
8888 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
8889 Use new safe be builtin.
8891 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8893 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
8895 (aarch64_be_st1<mode>): Likewise.
8896 (aarch_ld1<VALL:mode>): Define_expand modified.
8897 (aarch_st1<VALL:mode>): Likewise.
8898 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
8899 (UNSPEC_ST1): Likewise.
8901 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
8903 * config/microblaze/microblaze.md: Add trap insn and attribute
8905 2014-01-23 Dodji Seketeli <dodji@redhat.com>
8907 PR preprocessor/58580
8908 * input.h (location_get_source_line): Take an additional line_size
8910 (void diagnostics_file_cache_fini): Declare new function.
8911 * input.c (struct fcache): New type.
8912 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
8913 New static constants.
8914 (diagnostic_file_cache_init, total_lines_num)
8915 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
8916 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
8917 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
8918 (get_next_line, read_next_line, goto_next_line, read_line_num):
8919 New static function definitions.
8920 (diagnostic_file_cache_fini): New function.
8921 (location_get_source_line): Take an additional output line_len
8922 parameter. Re-write using lookup_or_add_file_to_cache_tab and
8924 * diagnostic.c (diagnostic_finish): Call
8925 diagnostic_file_cache_fini.
8926 (adjust_line): Take an additional input parameter for the length
8927 of the line, rather than calculating it with strlen.
8928 (diagnostic_show_locus): Adjust the use of
8929 location_get_source_line and adjust_line with respect to their new
8930 signature. While displaying a line now, do not stop at the first
8931 null byte. Rather, display the zero byte as a space and keep
8932 going until we reach the size of the line.
8933 * Makefile.in: Add vec.o to OBJS-libcommon
8935 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8936 Ilya Tocar <ilya.tocar@intel.com>
8938 * config/i386/avx512fintrin.h (_mm512_kmov): New.
8939 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
8940 (__builtin_ia32_kmov16): Ditto.
8941 * config/i386/i386.md (UNSPEC_KMOV): New.
8944 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8946 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
8947 (_mm512_storeu_si512): Ditto.
8949 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
8952 * rtl.h (get_referenced_operands): Declare.
8953 * recog.c (get_referenced_operands): New function.
8954 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
8955 operands have been referenced when recording LO_SUM references.
8957 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
8959 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
8961 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8963 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
8964 Enable for generic and recent AMD targets.
8966 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8968 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
8969 ARG_SIZE note when adjustment was eliminated.
8971 2014-01-22 Jeff Law <law@redhat.com>
8973 PR tree-optimization/59597
8974 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
8975 in file. Accept new argument REGISTERING and use it to modify
8976 dump output appropriately.
8977 (register_jump_thread): Corresponding changes.
8978 (mark_threaded_blocks): Reinstate code to cancel unprofitable
8979 thread paths involving joiner blocks. Add code to dump cancelled
8980 jump threading paths.
8982 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
8984 PR rtl-optimization/59477
8985 * lra-constraints.c (inherit_in_ebb): Process call for living hard
8986 regs. Update reloads_num and potential_reload_hard_regs for all insns.
8988 2014-01-22 Tom Tromey <tromey@redhat.com>
8990 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
8992 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
8994 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8996 PR rtl-optimization/59896
8997 * lra-constraints.c (process_alt_operands): Check unused note for
8998 matched operands of insn with no output reloads.
9000 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
9002 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
9003 (mips_move_from_gpr_cost): Likewise.
9005 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
9007 PR rtl-optimization/59858
9008 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
9009 ira_class_hard_regs_num.
9010 (process_alt_operands): Increase reject for dying matched operand.
9012 2014-01-21 Jakub Jelinek <jakub@redhat.com>
9015 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
9016 smaller than size, perform several stores or loads and stores
9017 at dst + count - size to store or copy all of size bytes, rather
9018 than just last modesize bytes.
9020 2014-01-20 DJ Delorie <dj@redhat.com>
9022 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
9023 that CLOBBERs are REGs before propogating their values.
9025 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
9028 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
9029 (cgraph_inline_failed_type): New function.
9030 * cgraph.h (DEFCIFCODE): Add type.
9031 (cgraph_inline_failed_type_t): New enum.
9032 (cgraph_inline_failed_type): New prototype.
9033 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
9034 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
9035 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
9036 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
9037 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
9038 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
9039 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
9040 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
9041 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
9042 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
9043 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
9044 OPTIMIZATION_MISMATCH.
9045 * tree-inline.c (expand_call_inline): Emit errors during
9046 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
9048 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9051 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
9052 mode attribute in insn output.
9054 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
9056 * output.h (output_constant): Delete.
9057 * varasm.c (output_constant): Make private.
9059 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
9061 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
9063 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9066 * tree.h (fold_builtin_strcat): New prototype.
9067 * builtins.c (fold_builtin_strcat): No longer static. Add len
9068 argument, if non-NULL, don't call c_strlen. Optimize
9069 directly into __builtin_memcpy instead of __builtin_strcpy.
9070 (fold_builtin_2): Adjust fold_builtin_strcat caller.
9071 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
9073 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9075 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9076 for SImode_address_operand operands, having only a REG argument.
9078 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
9080 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
9081 loader name using mbig-endian.
9082 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
9084 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9086 * doc/invoke.texi (-march): Clarify documentation for AArch64.
9090 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
9092 * config/aarch64/aarch64-protos.h
9093 (aarch64_cannot_change_mode_class_ptr): Declare.
9094 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
9095 aarch64_cannot_change_mode_class_ptr): New.
9096 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
9097 backend hook aarch64_cannot_change_mode_class.
9099 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9101 * common/config/aarch64/aarch64-common.c
9102 (aarch64_handle_option): Don't handle any option order logic here.
9103 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
9104 selected_cpu, warn on architecture version mismatch.
9105 (aarch64_override_options): Fix parsing order for option strings.
9107 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9108 Iain Sandoe <iain@codesourcery.com>
9111 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
9112 warning. Amend comment to reflect current functionality.
9114 2014-01-20 Richard Biener <rguenther@suse.de>
9117 * builtins.c (fold_builtin_strcat): Remove case better handled
9118 by tree-ssa-strlen.c.
9120 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
9122 * config/aarch64/aarch64.opt
9123 (mcpu, march, mtune): Make case-insensitive.
9125 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9128 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9129 if operands[1] is a REG or ZERO_EXTEND of a REG.
9131 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
9133 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
9135 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
9137 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
9138 long non-pic millicode calls.
9140 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9142 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
9144 2014-01-19 Kito Cheng <kito@0xlab.org>
9146 * builtins.c (expand_movstr): Check movstr expand done or fail.
9148 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9149 H.J. Lu <hongjiu.lu@intel.com>
9152 * config/i386/i386.md (*lea<mode>): Zero-extend return register
9153 to DImode for zero-extended addresses.
9155 2014-01-19 Jakub Jelinek <jakub@redhat.com>
9157 PR rtl-optimization/57763
9158 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
9159 on the new indirect jump_insn and increment LABEL_NUSES (label).
9161 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
9165 * config.gcc (x86_archs): New variable.
9166 (x86_64_archs): Likewise.
9167 (x86_cpus): Likewise.
9168 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
9169 --with-arch/--with-cpu= options.
9170 Support --with-arch=/--with-cpu={nehalem,westmere,
9171 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
9173 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9175 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
9176 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
9178 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9180 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
9182 2014-01-18 Jakub Jelinek <jakub@redhat.com>
9185 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
9186 clear cpp_get_options (parse_in)->warn_unused_macros for
9187 ix86_target_macros_internal with cpp_define.
9189 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
9191 * jump.c (delete_related_insns): Keep (use (insn))s.
9192 * reorg.c (redundant_insn): Check for barriers too.
9194 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9196 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
9198 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
9200 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
9201 call to $$dyncall when TARGET_LONG_CALLS is true.
9203 2014-01-17 Jeff Law <law@redhat.com>
9205 * ree.c (combine_set_extension): Temporarily disable test for
9206 changing number of hard registers.
9208 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9211 * ipa-inline-analysis.c (inline_free_summary):
9212 Do not free summary of aliases.
9214 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9217 * gimplify.c (gimplify_expr): Use create_tmp_var
9218 instead of create_tmp_var_raw. If cond doesn't have
9219 integral type, don't add the IFN_ANNOTATE builtin at all.
9221 2014-01-17 Martin Jambor <mjambor@suse.cz>
9224 * ipa-cp.c (prev_edge_clone): New variable.
9225 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
9226 Also resize prev_edge_clone vector.
9227 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
9228 (ipcp_edge_removal_hook): New function.
9229 (ipcp_driver): Register ipcp_edge_removal_hook.
9231 2014-01-17 Andrew Pinski <apinski@cavium.com>
9232 Steve Ellcey <sellcey@mips.com>
9235 * config/mips/mips.c (mips_print_operand): Check operand mode instead
9238 2014-01-17 Jeff Law <law@redhat.com>
9241 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
9242 so that pass_ccp runs first.
9244 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9246 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
9247 (ix86_adjust_cost): Use !TARGET_XXX.
9248 (do_reorder_for_imul): Likewise.
9249 (swap_top_of_ready_list): Likewise.
9250 (ix86_sched_reorder): Likewise.
9252 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9254 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9255 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
9256 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
9257 (intel_memset): New. Duplicate slm_memset.
9258 (intel_cost): New. Duplicate slm_cost.
9259 (m_INTEL): New macro.
9260 (processor_target_table): Add "intel".
9261 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
9262 with PROCESSOR_INTEL for "intel".
9263 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
9264 PROCESSOR_SILVERMONT.
9265 (ix86_issue_rate): Likewise.
9266 (ix86_adjust_cost): Likewise.
9267 (ia32_multipass_dfa_lookahead): Likewise.
9268 (swap_top_of_ready_list): Likewise.
9269 (ix86_sched_reorder): Likewise.
9270 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
9271 instead of TARGET_OPT_AGU.
9272 * config/i386/i386.h (TARGET_INTEL): New.
9273 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
9274 (processor_type): Add PROCESSOR_INTEL.
9275 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
9276 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
9278 2014-01-17 Marek Polacek <polacek@redhat.com>
9281 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
9284 2014-01-17 Richard Biener <rguenther@suse.de>
9286 PR tree-optimization/46590
9287 * opts.c (default_options_table): Add entries for
9288 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
9289 all enabled at -O1 but not for -Og.
9290 * common.opt (fbranch-count-reg): Remove Init(1).
9291 (fmove-loop-invariants): Likewise.
9292 (ftree-pta): Likewise.
9294 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9296 * config/i386/i386.c (ix86_data_alignment): For compatibility with
9297 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
9298 decls to at least the GCC 4.8 used alignments.
9301 * tree-nested.c (convert_nonlocal_reference_stmt,
9302 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
9303 of GIMPLE_BIND stmts, adjust associated decls.
9305 2014-01-17 Richard Biener <rguenther@suse.de>
9307 PR tree-optimization/46590
9308 * vec.h (vec<>::bseach): New member function implementing
9309 binary search according to C89 bsearch.
9310 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
9311 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
9312 bitmap pointer again. Make accesses_in_loop a flat array.
9313 (mem_ref_obstack): New global.
9314 (outermost_indep_loop): Adjust for mem_ref->stored changes.
9315 (mark_ref_stored): Likewise.
9316 (ref_indep_loop_p_2): Likewise.
9317 (set_ref_stored_in_loop): New helper function.
9318 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
9319 (memref_free): Adjust.
9320 (record_mem_ref_loc): Simplify.
9321 (gather_mem_refs_stmt): Adjust.
9322 (sort_locs_in_loop_postorder_cmp): New function.
9323 (analyze_memory_references): Sort accesses_in_loop after
9324 loop postorder number.
9325 (find_ref_loc_in_loop_cmp): New function.
9326 (for_all_locs_in_loop): Find relevant cluster of locs in
9327 accesses_in_loop and iterate without recursion.
9328 (execute_sm): Avoid uninit warning.
9329 (struct ref_always_accessed): Simplify.
9330 (ref_always_accessed::operator ()): Likewise.
9331 (ref_always_accessed_p): Likewise.
9332 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
9333 loop postorder numbers here.
9334 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
9337 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9340 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
9341 on decls for which assemble_alias has been called.
9343 2014-01-17 Nick Clifton <nickc@redhat.com>
9345 * config/msp430/msp430.opt: (mcpu): New option.
9346 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
9347 (msp430_option_override): Parse target_cpu. If the MCU name
9348 matches a generic string, clear target_mcu.
9349 (msp430_attr): Allow numeric interrupt values up to 63.
9350 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
9351 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
9353 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
9355 * config/msp430/msp430.md (popm): Use %J rather than %I.
9356 (addsi3): Use msp430_nonimmediate_operand for operand 2.
9357 (addhi_cy_i): Use immediate_operand for operand 2.
9358 * doc/invoke.texi: Document -mcpu option.
9360 2014-01-17 Richard Biener <rguenther@suse.de>
9362 PR rtl-optimization/38518
9363 * df.h (df_analyze_loop): Declare.
9364 * df-core.c: Include cfgloop.h.
9365 (df_analyze_1): Split out main part of df_analyze.
9366 (df_analyze): Adjust.
9367 (loop_inverted_post_order_compute): New function.
9368 (loop_post_order_compute): Likewise.
9369 (df_analyze_loop): New function avoiding whole-function
9371 * loop-invariant.c (find_defs): Use df_analyze_loop.
9372 (find_invariants): Adjust.
9373 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
9375 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
9377 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
9378 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
9380 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
9382 * ipa-ref.c (ipa_remove_stmt_references): Fix references
9383 traversal when removing references.
9385 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
9388 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
9390 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
9393 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
9394 pushing a reload for an autoinc when we had previously reloaded an
9395 inner part of the address.
9397 2014-01-16 Jakub Jelinek <jakub@redhat.com>
9399 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
9401 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
9402 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
9403 when not giving up or versioning for alias only because of
9405 (vect_analyze_data_ref_dependences): Set to true.
9406 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
9408 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
9409 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
9413 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
9416 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
9417 operand 0 predicate for gathers, use a new pseudo as subtarget.
9419 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9422 * lra-constraints.c (process_alt_operands): Add printing debug info.
9423 Check absence of input/output reloads for matched operands too.
9425 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9427 PR rtl-optimization/59835
9428 * ira.c (ira_init_register_move_cost): Increase cost for
9431 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
9433 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
9435 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
9438 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
9439 non-register objects. Use gen_(high/low)part more consistently.
9442 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9445 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
9446 endian support, remove tests for WORDS_BIG_ENDIAN.
9447 (p8_mfvsrd_3_<mode>): Likewise.
9448 (reload_gpr_from_vsx<mode>): Likewise.
9449 (reload_gpr_from_vsxsf): Likewise.
9450 (p8_mfvsrd_4_disf): Likewise.
9452 2014-01-16 Richard Biener <rguenther@suse.de>
9454 PR rtl-optimization/46590
9455 * lcm.c (compute_antinout_edge): Use postorder iteration.
9456 (compute_laterin): Use inverted postorder iteration.
9458 2014-01-16 Nick Clifton <nickc@redhat.com>
9461 * varasm.c (output_constant): Return the number of bytes actually
9463 (output_constructor_array_range): Update the field size with the
9464 number of bytes emitted by output_constant.
9465 (output_constructor_regular_field): Likewise. Also do not
9466 complain if the total number of bytes emitted is now greater
9467 than the expected fieldpos.
9468 * output.h (output_constant): Update prototype and descriptive comment.
9470 2014-01-16 Marek Polacek <polacek@redhat.com>
9473 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
9474 it is error_mark_node.
9476 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
9478 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
9479 VALID_AVX256_REG_OR_OI_MODE.
9481 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
9483 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
9484 current procedure should be profiled.
9486 2014-01-15 Andrew Pinski <apinski@cavium.com>
9488 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
9489 of moving from/to the STACK_REG register class.
9491 2014-01-15 Richard Henderson <rth@redhat.com>
9494 * reginfo.c (global_regs_decl): Globalize.
9495 * rtl.h (global_regs_decl): Declare.
9496 * ira.c (do_reload): Diagnose frame_pointer_needed and it
9497 reserved via global_regs.
9499 2014-01-15 Teresa Johnson <tejohnson@google.com>
9501 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
9503 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9505 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
9506 and vmulosh rather than call gen_vec_widen_smult_*.
9507 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
9508 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
9509 (vec_widen_smult_even_v16qi): Likewise.
9510 (vec_widen_umult_even_v8hi): Likewise.
9511 (vec_widen_smult_even_v8hi): Likewise.
9512 (vec_widen_umult_odd_v16qi): Likewise.
9513 (vec_widen_smult_odd_v16qi): Likewise.
9514 (vec_widen_umult_odd_v8hi): Likewise.
9515 (vec_widen_smult_odd_v8hi): Likewise.
9516 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
9517 vmuloub rather than call gen_vec_widen_umult_*.
9518 (vec_widen_umult_lo_v16qi): Likewise.
9519 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
9520 vmulosb rather than call gen_vec_widen_smult_*.
9521 (vec_widen_smult_lo_v16qi): Likewise.
9522 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
9523 rather than call gen_vec_widen_umult_*.
9524 (vec_widen_umult_lo_v8hi): Likewise.
9525 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
9526 rather than call gen_vec_widen_smult_*.
9527 (vec_widen_smult_lo_v8hi): Likewise.
9529 2014-01-15 Jeff Law <law@redhat.com>
9531 PR tree-optimization/59747
9532 * ree.c (find_and_remove_re): Properly handle case where a second
9533 eliminated extension requires widening a copy created for elimination
9534 of a prior extension.
9535 (combine_set_extension): Ensure that the number of hard regs needed
9536 for a destination register does not change when we widen it.
9538 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
9540 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
9541 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
9542 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
9543 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
9544 (avr-*-rtems*): Likewise.
9545 (bfin*-rtems*): Likewise.
9546 (moxie-*-rtems*): Likewise.
9547 (h8300-*-rtems*): Likewise.
9548 (i[34567]86-*-rtems*): Likewise.
9549 (lm32-*-rtems*): Likewise.
9550 (m32r-*-rtems*): Likewise.
9551 (m68k-*-rtems*): Likewise.
9552 (microblaze*-*-rtems*): Likewise.
9553 (mips*-*-rtems*): Likewise.
9554 (powerpc-*-rtems*): Likewise.
9555 (sh-*-rtems*): Likewise.
9556 (sparc-*-rtems*): Likewise.
9557 (sparc64-*-rtems*): Likewise.
9558 (v850-*-rtems*): Likewise.
9559 (m32c-*-rtems*): Likewise.
9561 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9563 PR rtl-optimization/59511
9564 * ira.c (ira_init_register_move_cost): Use memory costs for some
9565 cases of register move cost calculations.
9566 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9567 instead of BB frequency.
9568 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9569 * lra-assigns.c (find_hard_regno_for): Ditto.
9571 2014-01-15 Richard Biener <rguenther@suse.de>
9573 PR tree-optimization/59822
9574 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9575 (vectorizable_load): Use it to hoist defs of uses of invariant
9576 loads out of the loop.
9578 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9579 Kugan Vivekanandarajah <kuganv@linaro.org>
9582 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9585 2014-01-15 Richard Biener <rguenther@suse.de>
9587 PR rtl-optimization/59802
9588 * lcm.c (compute_available): Use inverted postorder to seed
9589 the initial worklist.
9591 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9594 * config/s390/s390.c (s390_preferred_reload_class): Don't return
9595 ADDR_REGS for invalid symrefs in non-PIC code.
9597 2014-01-15 Jakub Jelinek <jakub@redhat.com>
9600 * builtins.c (determine_block_size): Initialize *probable_max_size
9601 even if len_rtx is CONST_INT.
9603 2014-01-14 Andrew Pinski <apinski@cavium.com>
9605 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
9606 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
9607 (cortexa53_tunings): Likewise.
9608 (aarch64_sched_issue_rate): New function.
9609 (TARGET_SCHED_ISSUE_RATE): Define.
9611 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9613 * ira-costs.c (find_costs_and_classes): Add missed
9614 ira_init_register_move_cost_if_necessary.
9616 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9619 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
9621 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
9624 * config/i386/i386.c (type_natural_mode): Add a bool parameter
9625 to indicate if type is used for function return value. Warn ABI
9626 change if the vector mode isn't available for function return value.
9627 (ix86_function_arg_advance): Pass false to type_natural_mode.
9628 (ix86_function_arg): Likewise.
9629 (ix86_gimplify_va_arg): Likewise.
9630 (function_arg_32): Don't warn ABI change.
9631 (ix86_function_value): Pass true to type_natural_mode.
9632 (ix86_return_in_memory): Likewise.
9633 (ix86_struct_value_rtx): Removed.
9634 (TARGET_STRUCT_VALUE_RTX): Likewise.
9636 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9638 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
9639 converting a conditional jump into a conditional return.
9641 2014-01-14 Richard Biener <rguenther@suse.de>
9643 PR tree-optimization/58921
9644 PR tree-optimization/59006
9645 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
9646 hoisting invariant stmts.
9647 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
9648 invariant loads on the preheader edge if possible.
9650 2014-01-14 Joey Ye <joey.ye@arm.com>
9652 * doc/plugin.texi (Building GCC plugins): Update to C++.
9654 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
9656 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
9657 (_mm_rcp28_round_ss): Ditto.
9658 (_mm_rsqrt28_round_sd): Ditto.
9659 (_mm_rsqrt28_round_ss): Ditto.
9660 (_mm_rcp28_sd): Ditto.
9661 (_mm_rcp28_ss): Ditto.
9662 (_mm_rsqrt28_sd): Ditto.
9663 (_mm_rsqrt28_ss): Ditto.
9664 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
9665 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
9666 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
9667 (IX86_BUILTIN_RCP28SD): Ditto.
9668 (IX86_BUILTIN_RCP28SS): Ditto.
9669 (IX86_BUILTIN_RSQRT28SD): Ditto.
9670 (IX86_BUILTIN_RSQRT28SS): Ditto.
9671 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
9672 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
9673 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
9674 (ix86_expand_special_args_builtin): Expand new FTYPE.
9675 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
9676 (srcp14<mode>): Make insn unary.
9677 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
9678 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
9679 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
9680 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
9681 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
9682 Fix rounding: make it SAE only.
9683 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9685 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9687 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
9688 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
9689 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
9690 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
9691 (round_saeonly_mask_scalar_operand4): Ditto.
9692 (round_saeonly_mask_scalar_op3): Ditto.
9693 (round_saeonly_mask_scalar_op4): Ditto.
9695 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9697 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9698 Implement -maltivec=be for vec_insert and vec_extract.
9700 2014-01-10 DJ Delorie <dj@redhat.com>
9702 * config/msp430/msp430.md (call_internal): Don't allow memory
9703 references with SP as the base register.
9704 (call_value_internal): Likewise.
9705 * config/msp430/constraints.md (Yc): New. For memory references
9706 that don't use SP as a base register.
9708 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
9709 "an integer without a # prefix"
9710 * config/msp430/msp430.md (epilogue_helper): Use it.
9712 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9715 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
9716 AVX512F gather builtins.
9717 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
9718 on gather decls with INTEGER_TYPE masktype.
9719 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
9720 directly into the builtin rather than hoisting it before loop.
9722 PR tree-optimization/59387
9723 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
9724 (scev_const_prop): If folded_casts and type has undefined overflow,
9725 use force_gimple_operand instead of force_gimple_operand_gsi and
9726 for each added stmt if it is assign with
9727 arith_code_with_undefined_signed_overflow, call
9728 rewrite_to_defined_overflow.
9729 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
9730 gimple-fold.h instead.
9731 (arith_code_with_undefined_signed_overflow,
9732 rewrite_to_defined_overflow): Moved to ...
9733 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
9734 rewrite_to_defined_overflow): ... here. No longer static.
9735 Include gimplify-me.h.
9736 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
9737 rewrite_to_defined_overflow): New prototypes.
9739 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9741 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
9743 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
9745 * builtins.c (get_object_alignment_2): Minor tweak.
9746 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
9748 2014-01-13 Christian Bruel <christian.bruel@st.com>
9750 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
9751 optimized non constant lengths.
9753 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9756 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
9757 load as __atomic_load_N if possible.
9759 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9761 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
9763 (rs6000_expand_builtin): Adjust call.
9765 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9768 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
9769 * config/rs6000/rs6000.c: Include target-globals.h.
9770 (rs6000_set_current_function): Instead of doing target_reinit
9771 unconditionally, use save_target_globals_default_opts and
9772 restore_target_globals.
9774 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
9776 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
9777 (rs6000_expand_builtin): Handle mffs and mtfsf.
9778 (rs6000_init_builtins): Define mffs and mtfsf.
9779 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
9780 (rs6000_mffs): New pattern.
9781 (rs6000_mtfsf): New pattern.
9783 2014-01-11 Bin Cheng <bin.cheng@arm.com>
9785 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
9786 Start narrowing with START. Apply candidate-use pair
9787 and check overall cost in narrowing.
9788 (iv_ca_prune): Pass new argument.
9790 2014-01-10 Jeff Law <law@redhat.com>
9793 * ree.c (combine_reaching_defs): Ensure the defining statement
9794 occurs before the extension when optimizing extensions with
9795 different source and destination hard registers.
9797 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9800 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
9801 vtables into the type inheritance graph.
9803 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9805 PR rtl-optimization/59754
9806 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
9807 modes in the REGNO != REGNO case.
9809 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9811 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
9813 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9815 PR tree-optimization/59745
9816 * tree-predcom.c (tree_predictive_commoning_loop): Call
9817 free_affine_expand_cache if giving up because components is NULL.
9819 * target-globals.c (save_target_globals): Allocate < 4KB structs using
9820 GC in payload of target_globals struct instead of allocating them on
9821 the heap and the larger structs separately using GC.
9822 * target-globals.h (struct target_globals): Make regs, hard_regs,
9823 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
9824 of GTY((skip)) and change type to void *.
9825 (reset_target_globals): Cast loads from those fields to corresponding
9828 2014-01-10 Steve Ellcey <sellcey@mips.com>
9831 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
9832 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
9833 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
9835 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
9838 * aarch64-modes.def (CC_Zmode): New flags mode.
9839 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
9840 represents an equality.
9841 (aarch64_get_condition_code): Handle CC_Zmode.
9842 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
9844 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9846 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
9847 extraction in good case.
9849 2014-01-10 Richard Biener <rguenther@suse.de>
9851 PR tree-optimization/59374
9852 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
9853 checking after SLP discovery. Mark stmts not participating
9854 in any SLP instance properly.
9856 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9858 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
9859 when handling a SET rtx.
9861 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9863 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
9864 (cortex-a57): Likewise.
9865 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
9867 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9869 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
9870 non-iwmmxt builtins.
9872 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9876 * ipa-devirt.c record_target_from_binfo): Take as argument
9877 stack of binfos and lookup matching one for virtual inheritance.
9878 (possible_polymorphic_call_targets_1): Update.
9880 2014-01-10 Huacai Chen <chenhc@lemote.com>
9882 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
9883 kernel strings for Loongson-2E/2F/3A.
9885 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9888 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
9889 is_gimple_call before calling gimple_call_internal_p.
9891 2014-01-09 Steve Ellcey <sellcey@mips.com>
9893 * Makefile.in (TREE_FLOW_H): Remove.
9894 (TREE_SSA_H): Add file names from tree-flow.h.
9895 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
9896 * tree.h: Remove tree-flow.h reference.
9897 * hash-table.h: Remove tree-flow.h reference.
9898 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
9899 reference with tree-ssa-loop.h.
9901 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9903 * doc/invoke.texi: Add -maltivec={be,le} options, and document
9904 default element-order behavior for -maltivec.
9905 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
9906 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
9907 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
9908 when targeting big endian, at least for now.
9909 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
9911 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9914 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
9915 var satisfies use_register_for_decl, just take into account type
9916 alignment, rather than decl alignment.
9918 PR tree-optimization/59622
9919 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
9920 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
9921 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
9922 Don't devirtualize for inplace at all. For targets.length () == 1,
9923 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
9925 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9927 * config/i386/i386.md (cpu): Remove the unused btver1.
9929 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9931 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
9933 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9936 * tree-core.h (struct target_globals): New forward declaration.
9937 (struct tree_target_option): Add globals field.
9938 * tree.h (TREE_TARGET_GLOBALS): Define.
9939 (prepare_target_option_nodes_for_pch): New prototype.
9940 * target-globals.h (struct target_globals): Define even if
9942 * tree.c (prepare_target_option_node_for_pch,
9943 prepare_target_option_nodes_for_pch): New functions.
9944 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
9945 * config/i386/i386.c: Include target-globals.h.
9946 (ix86_set_current_function): Instead of doing target_reinit
9947 unconditionally, use save_target_globals_default_opts and
9948 restore_target_globals.
9950 2014-01-09 Richard Biener <rguenther@suse.de>
9952 PR tree-optimization/59715
9953 * tree-cfg.h (split_critical_edges): Declare.
9954 * tree-cfg.c (split_critical_edges): Export.
9955 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
9957 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
9959 * cfgexpand.c (expand_stack_vars): Optionally disable
9960 asan stack protection.
9961 (expand_used_vars): Likewise.
9962 (partition_stack_vars): Likewise.
9963 * asan.c (asan_emit_stack_protection): Optionally disable
9964 after return stack usage.
9965 (instrument_derefs): Optionally disable memory access instrumentation.
9966 (instrument_builtin_call): Likewise.
9967 (instrument_strlen_call): Likewise.
9968 (asan_protect_global): Optionally disable global variables protection.
9969 * doc/invoke.texi: Added doc for new options.
9970 * params.def: Added new options.
9971 * params.h: Likewise.
9973 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9975 PR rtl-optimization/59724
9976 * ifcvt.c (cond_exec_process_if_block): Don't call
9977 flow_find_head_matching_sequence with 0 longest_match.
9978 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
9979 non-active insns if !stop_after.
9980 (try_head_merge_bb): Revert 2014-01-07 changes.
9982 2014-01-08 Jeff Law <law@redhat.com>
9984 * ree.c (get_sub_rtx): New function, extracted from...
9985 (merge_def_and_ext): Here.
9986 (combine_reaching_defs): Use get_sub_rtx.
9988 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
9990 * cgraph.h (varpool_variable_node): Do not choke on null node.
9992 2014-01-08 Catherine Moore <clm@codesourcery.com>
9994 * config/mips/mips.md (simple_return): Attempt to use JRC
9996 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
9998 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
10000 PR rtl-optimization/59137
10001 * reorg.c (steal_delay_list_from_target): Call update_block for
10003 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
10005 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10007 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
10008 two duplicate entries.
10010 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
10013 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
10015 * config/mips/mips.c (mips_truncated_op_cost): New function.
10016 (mips_rtx_costs): Adjust test for BADDU.
10017 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
10019 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
10021 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
10022 (*baddu_si): ...this new pattern.
10024 2014-01-08 Jakub Jelinek <jakub@redhat.com>
10027 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
10029 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
10031 PR middle-end/57748
10032 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
10034 (expand_expr, expand_normal): Adjust.
10035 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
10036 inner_reference_p. Use inner_reference_p to expand inner references.
10037 (store_expr): Adjust.
10038 * cfgexpand.c (expand_call_stmt): Adjust.
10040 2014-01-08 Rong Xu <xur@google.com>
10042 * gcov-io.c (gcov_var): Move from gcov-io.h.
10043 (gcov_position): Ditto.
10044 (gcov_is_error): Ditto.
10045 (gcov_rewrite): Ditto.
10046 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
10047 only part to libgcc/libgcov.h.
10049 2014-01-08 Marek Polacek <polacek@redhat.com>
10051 PR middle-end/59669
10052 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
10054 2014-01-08 Marek Polacek <polacek@redhat.com>
10057 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
10059 2014-01-08 Jakub Jelinek <jakub@redhat.com>
10061 PR rtl-optimization/59649
10062 * stor-layout.c (get_mode_bounds): For BImode return
10063 0 and STORE_FLAG_VALUE.
10065 2014-01-08 Richard Biener <rguenther@suse.de>
10067 PR middle-end/59630
10068 * gimple.h (is_gimple_builtin_call): Remove.
10069 (gimple_builtin_call_types_compatible_p): New.
10070 (gimple_call_builtin_p): New overload.
10071 * gimple.c (is_gimple_builtin_call): Remove.
10072 (validate_call): Rename to ...
10073 (gimple_builtin_call_types_compatible_p): ... this and export. Also
10074 check return types.
10075 (validate_type): New static function.
10076 (gimple_call_builtin_p): New overload and adjust.
10077 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
10078 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
10079 (gimple_fold_stmt_to_constant_1): Likewise.
10080 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
10082 2014-01-08 Richard Biener <rguenther@suse.de>
10084 PR middle-end/59471
10085 * gimplify.c (gimplify_expr): Gimplify register-register type
10086 VIEW_CONVERT_EXPRs to separate stmts.
10088 2014-01-07 Jeff Law <law@redhat.com>
10090 PR middle-end/53623
10091 * ree.c (combine_set_extension): Handle case where source
10092 and destination registers in an extension insn are different.
10093 (combine_reaching_defs): Allow source and destination registers
10094 in extension to be different under limited circumstances.
10095 (add_removable_extension): Remove restriction that the
10096 source and destination registers in the extension are the same.
10097 (find_and_remove_re): Emit a copy from the extension's
10098 destination to its source after the defining insn if
10099 the source and destination registers are different.
10101 PR middle-end/59285
10102 * ifcvt.c (merge_if_block): If we are merging a block with more than
10103 one successor with a block with no successors, remove any BARRIER
10104 after the second block.
10106 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
10108 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
10110 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
10113 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
10114 for 14-bit register offsets when INT14_OK_STRICT is false.
10116 2014-01-07 Roland Stigge <stigge@antcom.de>
10117 Michael Meissner <meissner@linux.vnet.ibm.com>
10120 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
10121 Only check TFmode for SPE constants. Don't check TImode or TDmode.
10123 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
10125 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
10128 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
10130 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
10131 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
10132 rtx is const0_rtx or not.
10134 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
10137 * target-globals.c (save_target_globals): Remove this_fn_optab
10139 * toplev.c: Include optabs.h.
10140 (target_reinit): Temporarily restore the global options if another
10141 set of options are in force.
10143 2014-01-07 Jakub Jelinek <jakub@redhat.com>
10145 PR rtl-optimization/58668
10146 * cfgcleanup.c (flow_find_cross_jump): Don't count
10147 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
10148 to determine what is counted.
10149 (flow_find_head_matching_sequence): Use active_insn_p to determine
10151 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
10153 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
10154 determine what is counted.
10156 PR tree-optimization/59643
10157 * tree-predcom.c (split_data_refs_to_components): If one dr is
10158 read and one write, determine_offset fails and the write isn't
10159 in the bad component, just put the read into the bad component.
10161 2014-01-07 Mike Stump <mikestump@comcast.net>
10162 Jakub Jelinek <jakub@redhat.com>
10165 * tree-core.h (struct tree_optimization_option): Change optabs
10166 type from unsigned char * to void *.
10167 * optabs.c (init_tree_optimization_optabs): Adjust
10168 TREE_OPTIMIZATION_OPTABS initialization.
10170 2014-01-06 Jakub Jelinek <jakub@redhat.com>
10173 * config/i386/i386.h (struct machine_function): Add
10174 no_drap_save_restore field.
10175 * config/i386/i386.c (ix86_save_reg): Use
10176 !cfun->machine->no_drap_save_restore instead of
10177 crtl->stack_realign_needed.
10178 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
10179 this function clears frame_pointer_needed. Set
10180 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
10181 and DRAP reg is needed.
10183 2014-01-06 Marek Polacek <polacek@redhat.com>
10186 * doc/implement-c.texi: Mention that other integer types are
10187 permitted as bit-field types in strictly conforming mode.
10189 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
10191 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
10192 is newly allocated.
10194 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
10196 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
10198 2014-01-06 Martin Jambor <mjambor@suse.cz>
10201 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
10203 * ipa-prop.c (ipa_print_node_params): Fix indentation.
10205 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
10209 * var-tracking.c (add_stores): Preserve the value of the source even if
10210 we don't record the store.
10212 2014-01-06 Terry Guo <terry.guo@arm.com>
10214 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
10216 2014-01-05 Iain Sandoe <iain@codesourcery.com>
10219 * config/darwin.c (darwin_function_section): Adjust return values to
10220 correspond to optimisation changes made in r206070.
10222 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
10224 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
10225 from prefetch_block tune setting.
10226 (nocona_cost): Correct size of prefetch block to 64.
10228 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
10230 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
10231 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
10232 used to save the static chain register in the computation of the offset
10233 from which the FP registers need to be restored.
10235 2014-01-04 Jakub Jelinek <jakub@redhat.com>
10237 PR tree-optimization/59519
10238 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
10239 ICE if get_current_def (current_new_name) is already non-NULL, as long
10240 as it is a phi result of some other phi in *new_exit_bb that has
10243 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
10244 or vmovdqu* for misaligned_operand.
10245 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
10246 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
10247 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
10248 aligned_mem for AVX512F masked aligned load and store builtins and for
10249 non-temporal moves.
10251 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
10253 PR tree-optimization/59651
10254 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
10255 Address range for negative step should be added by TYPE_SIZE_UNIT.
10257 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
10259 * config/m68k/m68k.c (handle_move_double): Handle pushes with
10260 overlapping registers also for registers other than the stack pointer.
10262 2014-01-03 Marek Polacek <polacek@redhat.com>
10265 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
10268 2014-01-03 Jakub Jelinek <jakub@redhat.com>
10271 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
10274 * config/i386/i386.md (MODE_SIZE): New mode attribute.
10275 (push splitter): Use <P:MODE_SIZE> instead of
10276 GET_MODE_SIZE (<P:MODE>mode).
10277 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
10278 (mov -1, reg peephole2): Likewise.
10279 * config/i386/sse.md (*mov<mode>_internal,
10280 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
10281 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
10282 *<code><mode>3, *andnot<mode>3<mask_name>,
10283 <mask_codefor><code><mode>3<mask_name>): Likewise.
10284 * config/i386/subst.md (mask_mode512bit_condition,
10285 sd_mask_mode512bit_condition): Likewise.
10287 2014-01-02 Xinliang David Li <davidxl@google.com>
10289 PR tree-optimization/59303
10290 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
10291 (dump_predicates): Better output format.
10292 (pred_equal_p): New function.
10293 (is_neq_relop_p): Ditto.
10294 (is_neq_zero_form_p): Ditto.
10295 (pred_expr_equal_p): Ditto.
10296 (pred_neg_p): Ditto.
10297 (simplify_pred): Ditto.
10298 (simplify_preds_2): Ditto.
10299 (simplify_preds_3): Ditto.
10300 (simplify_preds_4): Ditto.
10301 (simplify_preds): Ditto.
10302 (push_pred): Ditto.
10303 (push_to_worklist): Ditto.
10304 (get_pred_info_from_cmp): Ditto.
10305 (is_degenerated_phi): Ditto.
10306 (normalize_one_pred_1): Ditto.
10307 (normalize_one_pred): Ditto.
10308 (normalize_one_pred_chain): Ditto.
10309 (normalize_preds): Ditto.
10310 (normalize_cond_1): Remove function.
10311 (normalize_cond): Ditto.
10312 (is_gcond_subset_of): Ditto.
10313 (is_subset_of_any): Ditto.
10314 (is_or_set_subset_of): Ditto.
10315 (is_and_set_subset_of): Ditto.
10316 (is_norm_cond_subset_of): Ditto.
10317 (pred_chain_length_cmp): Ditto.
10318 (convert_control_dep_chain_into_preds): Type change.
10319 (find_predicates): Ditto.
10320 (find_def_preds): Ditto.
10321 (destroy_predicates_vecs): Ditto.
10322 (find_matching_predicates_in_rest_chains): Ditto.
10323 (use_pred_not_overlap_with_undef_path_pred): Ditto.
10324 (is_pred_expr_subset): Ditto.
10325 (is_pred_chain_subset_of): Ditto.
10326 (is_included_in): Ditto.
10327 (is_superset_of): Ditto.
10329 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10331 Update copyright years.
10333 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10335 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
10336 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
10337 config/arc/arc.md, config/arc/arc.opt,
10338 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
10339 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
10340 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
10341 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
10342 config/linux-protos.h, config/linux.c, config/winnt-c.c,
10343 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
10344 vtable-verify.c, vtable-verify.h: Use the standard form for the
10347 2014-01-02 Tobias Burnus <burnus@net-b.de>
10349 * gcc.c (process_command): Update copyright notice dates.
10350 * gcov-dump.c: Ditto.
10352 * doc/cpp.texi: Bump @copying's copyright year.
10353 * doc/cppinternals.texi: Ditto.
10354 * doc/gcc.texi: Ditto.
10355 * doc/gccint.texi: Ditto.
10356 * doc/gcov.texi: Ditto.
10357 * doc/install.texi: Ditto.
10358 * doc/invoke.texi: Ditto.
10360 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10362 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
10364 2014-01-01 Jakub Jelinek <jakub@redhat.com>
10366 * config/i386/sse.md (*mov<mode>_internal): Guard
10367 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
10369 PR rtl-optimization/59647
10370 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
10371 new_rtx into UNSIGNED_FLOAT rtxes.
10373 Copyright (C) 2014 Free Software Foundation, Inc.
10375 Copying and distribution of this file, with or without modification,
10376 are permitted in any medium without royalty provided the copyright
10377 notice and this notice are preserved.