1 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
3 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
4 * config/arm/aout.h (License): Same.
5 * config/arm/bpabi.h (License): Same.
6 * config/arm/elf.h (License): Same.
7 * config/arm/linux-elf.h (License): Same.
8 * config/arm/linux-gas.h (License): Same.
9 * config/arm/netbsd-elf.h (License): Same.
10 * config/arm/uclinux-eabi.h (License): Same.
11 * config/arm/uclinux-elf.h (License): Same.
12 * config/arm/vxworks.h (License): Same.
14 2014-05-11 Jakub Jelinek <jakub@redhat.com>
16 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
17 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
18 number of operands to 3.
19 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
20 * tree-nested.c (convert_nonlocal_omp_clauses,
21 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
22 * gimplify.c (gimplify_scan_omp_clauses): Handle
23 OMP_CLAUSE_LINEAR_STMT.
24 * omp-low.c (lower_rec_input_clauses): Fix typo.
25 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
26 cast between Fortran boolean_type_node and C _Bool if
29 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
31 PR tree-optimization/61136
32 * wide-int.h (multiple_of_p): Define a version that doesn't return
34 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
35 integer_zerop/const_binop pair.
36 (multiple_of_p): Likewise, converting both operands to widest_int
39 2014-05-09 Teresa Johnson <tejohnson@google.com>
41 * cgraphunit.c (analyze_functions): Use correct dump file.
43 2014-05-09 Florian Weimer <fweimer@redhat.com>
45 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
47 (stack_protect_return_slot_p): New function.
48 (expand_used_vars): Call stack_protect_decl_p and
49 stack_protect_return_slot_p for -fstack-protector-strong.
51 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
54 * fold-const.c (fold_binary_loc): Changed width of mask.
56 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
58 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
59 unsigned int initializers for regno_in, regno_out.
61 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
64 * config/avr/avr.md (cc): Add new attribute set_vzn.
65 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
66 Set cc insn attribute to set_vzn instead of set_zn for alternatives
68 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
69 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
70 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
72 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
75 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
77 * wide-int.cc (UTItype): Define.
78 (UDWtype): Define for appropriate W_TYPE_SIZE.
80 2014-05-09 Richard Biener <rguenther@suse.de>
82 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
83 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
84 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
85 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
86 ssa_propagate): Adjust.
88 2014-05-08 Jeff Law <law@redhat.com>
90 PR tree-optimization/61009
91 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
92 tri-state rather than a boolean. When a block is too big to
93 thread through, inform caller via negative return value.
94 (thread_across_edge): If a block was too big for normal threading,
95 then it's too big for a joiner too, so remove temporary equivalences
96 and return immediately.
98 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
99 Matthias Klose <doko@ubuntu.com>
102 * optc-gen.awk: Fix option handling for -Wunused-parameter.
104 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
107 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
109 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
112 * config/alpha/alpha.c: Include gimple-iterator.h.
113 (alpha_gimple_fold_builtin): New function. Move
114 ALPHA_BUILTIN_UMULH folding from ...
115 (alpha_fold_builtin): ... here.
116 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
118 2014-05-08 Wei Mi <wmi@google.com>
121 * config/i386/i386.c (ix86_compute_frame_layout): Update
122 preferred_stack_boundary for call, expanded from tls descriptor.
123 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
124 to depend on SP register.
125 (*tls_local_dynamic_base_32_gnu): Ditto.
126 (*tls_local_dynamic_32_once): Ditto.
127 (tls_global_dynamic_64_<mode>): Set
128 ix86_tls_descriptor_calls_expanded_in_cfun.
129 (tls_local_dynamic_base_64_<mode>): Ditto.
130 (tls_global_dynamic_32): Set
131 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
132 to depend on SP register.
133 (tls_local_dynamic_base_32): Ditto.
135 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
137 * config/arm/arm_neon.h: Update comment.
138 * config/arm/neon-docgen.ml: Delete.
139 * config/arm/neon-gen.ml: Delete.
140 * doc/arm-neon-intrinsics.texi: Update comment.
142 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
144 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
146 (vand, vorr, veor, vorn, vbic): Remove.
147 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
149 (neon_vsub_unspec): Likewise.
150 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
152 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
154 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
155 (vadd_s16): Likewise.
156 (vadd_s32): Likewise.
157 (vadd_f32): Likewise.
159 (vadd_u16): Likewise.
160 (vadd_u32): Likewise.
161 (vadd_s64): Likewise.
162 (vadd_u64): Likewise.
163 (vaddq_s8): Likewise.
164 (vaddq_s16): Likewise.
165 (vaddq_s32): Likewise.
166 (vaddq_s64): Likewise.
167 (vaddq_f32): Likewise.
168 (vaddq_u8): Likewise.
169 (vaddq_u16): Likewise.
170 (vaddq_u32): Likewise.
171 (vaddq_u64): Likewise.
173 (vmul_s16): Likewise.
174 (vmul_s32): Likewise.
175 (vmul_f32): Likewise.
177 (vmul_u16): Likewise.
178 (vmul_u32): Likewise.
180 (vmulq_s8): Likewise.
181 (vmulq_s16): Likewise.
182 (vmulq_s32): Likewise.
183 (vmulq_f32): Likewise.
184 (vmulq_u8): Likewise.
185 (vmulq_u16): Likewise.
186 (vmulq_u32): Likewise.
188 (vsub_s16): Likewise.
189 (vsub_s32): Likewise.
190 (vsub_f32): Likewise.
192 (vsub_u16): Likewise.
193 (vsub_u32): Likewise.
194 (vsub_s64): Likewise.
195 (vsub_u64): Likewise.
196 (vsubq_s8): Likewise.
197 (vsubq_s16): Likewise.
198 (vsubq_s32): Likewise.
199 (vsubq_s64): Likewise.
200 (vsubq_f32): Likewise.
201 (vsubq_u8): Likewise.
202 (vsubq_u16): Likewise.
203 (vsubq_u32): Likewise.
204 (vsubq_u64): Likewise.
206 (vand_s16): Likewise.
207 (vand_s32): Likewise.
209 (vand_u16): Likewise.
210 (vand_u32): Likewise.
211 (vand_s64): Likewise.
212 (vand_u64): Likewise.
213 (vandq_s8): Likewise.
214 (vandq_s16): Likewise.
215 (vandq_s32): Likewise.
216 (vandq_s64): Likewise.
217 (vandq_u8): Likewise.
218 (vandq_u16): Likewise.
219 (vandq_u32): Likewise.
220 (vandq_u64): Likewise.
222 (vorr_s16): Likewise.
223 (vorr_s32): Likewise.
225 (vorr_u16): Likewise.
226 (vorr_u32): Likewise.
227 (vorr_s64): Likewise.
228 (vorr_u64): Likewise.
229 (vorrq_s8): Likewise.
230 (vorrq_s16): Likewise.
231 (vorrq_s32): Likewise.
232 (vorrq_s64): Likewise.
233 (vorrq_u8): Likewise.
234 (vorrq_u16): Likewise.
235 (vorrq_u32): Likewise.
236 (vorrq_u64): Likewise.
238 (veor_s16): Likewise.
239 (veor_s32): Likewise.
241 (veor_u16): Likewise.
242 (veor_u32): Likewise.
243 (veor_s64): Likewise.
244 (veor_u64): Likewise.
245 (veorq_s8): Likewise.
246 (veorq_s16): Likewise.
247 (veorq_s32): Likewise.
248 (veorq_s64): Likewise.
249 (veorq_u8): Likewise.
250 (veorq_u16): Likewise.
251 (veorq_u32): Likewise.
252 (veorq_u64): Likewise.
254 (vbic_s16): Likewise.
255 (vbic_s32): Likewise.
257 (vbic_u16): Likewise.
258 (vbic_u32): Likewise.
259 (vbic_s64): Likewise.
260 (vbic_u64): Likewise.
261 (vbicq_s8): Likewise.
262 (vbicq_s16): Likewise.
263 (vbicq_s32): Likewise.
264 (vbicq_s64): Likewise.
265 (vbicq_u8): Likewise.
266 (vbicq_u16): Likewise.
267 (vbicq_u32): Likewise.
268 (vbicq_u64): Likewise.
270 (vorn_s16): Likewise.
271 (vorn_s32): Likewise.
273 (vorn_u16): Likewise.
274 (vorn_u32): Likewise.
275 (vorn_s64): Likewise.
276 (vorn_u64): Likewise.
277 (vornq_s8): Likewise.
278 (vornq_s16): Likewise.
279 (vornq_s32): Likewise.
280 (vornq_s64): Likewise.
281 (vornq_u8): Likewise.
282 (vornq_u16): Likewise.
283 (vornq_u32): Likewise.
284 (vornq_u64): Likewise.
286 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
288 * wide-int.cc (UTItype): Define.
289 (UDWtype): Define for appropriate W_TYPE_SIZE.
291 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
293 PR tree-optimization/59100
294 * tree-ssa-phiopt.c: Include tree-inline.h.
295 (neutral_element_p, absorbing_element_p): New functions.
296 (value_replacement): Handle conditional binary operations with a
297 neutral or absorbing element.
299 2014-05-08 Richard Biener <rguenther@suse.de>
301 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
302 folding the expression.
303 (valueize_expr): Remove.
304 (visit_reference_op_load): Do not valueize the result of
306 (simplify_binary_expression): Likewise.
307 (simplify_unary_expression): Likewise.
309 2014-05-08 Richard Biener <rguenther@suse.de>
311 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
312 looking at TYPE_ARG_TYPES.
314 2014-05-08 Richard Biener <rguenther@suse.de>
316 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
317 pointer propagation special-case.
319 2014-05-08 Bin Cheng <bin.cheng@arm.com>
321 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
322 core part of address expressions.
324 2014-05-08 Alan Modra <amodra@gmail.com>
327 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
328 loads and stores when -mno-strict-align at any alignment.
329 (expand_block_clear): Similarly. Also correct calculation of
332 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
335 * tree-complex.c (expand_complex_move): Keep line info when expanding
337 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
338 of complex expression. Use new argument to display correct location
339 for values coming from phi statement.
340 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
341 (warn_uninitialized_phi): Pass location of phi argument to
343 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
344 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
346 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
348 * config/rs6000/predicates.md (indexed_address_mem): New.
349 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
350 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
351 fpstore_ux, fpstore_u.
352 (sign_extend, indexed, update): New.
353 (cell_micro): Adjust.
354 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
355 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
356 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
357 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
358 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
359 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
360 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
361 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
362 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
363 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
364 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
365 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
366 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
367 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
368 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
370 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
371 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
372 *vsx_extract_<mode>_store): Adjust.
373 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
374 is_cracked_insn, insn_must_be_first_in_group,
375 insn_must_be_last_in_group): Adjust.
377 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
379 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
380 ppc440-fpstore): Adjust.
381 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
382 ppc476-fpstore): Adjust.
383 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
384 ppc601-fpstore): Adjust.
385 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
387 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
389 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
390 ppc7450-fpstore): Adjust.
391 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
392 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
393 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
395 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
396 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
397 cell-fpstore, cell-fpstore-update): Adjust.
398 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
399 ppce300c3_store, ppce300c3_fpstore): Adjust.
400 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
401 e500mc_fpstore): Adjust.
402 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
403 e500mc64_store, e500mc64_fpstore): Adjust.
404 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
405 e5500_fpstore): Adjust.
406 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
407 e6500_fpstore): Adjust.
408 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
410 * config/rs6000/power4.md (power4-load, power4-load-ext,
411 power4-load-ext-update, power4-load-ext-update-indexed,
412 power4-load-update-indexed, power4-load-update, power4-fpload,
413 power4-fpload-update, power4-store, power4-store-update,
414 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
416 * config/rs6000/power5.md (power5-load, power5-load-ext,
417 power5-load-ext-update, power5-load-ext-update-indexed,
418 power5-load-update-indexed, power5-load-update, power5-fpload,
419 power5-fpload-update, power5-store, power5-store-update,
420 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
422 * config/rs6000/power6.md (power6-load, power6-load-ext,
423 power6-load-update, power6-load-update-indexed,
424 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
425 power6-fpload-update, power6-store, power6-store-update,
426 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
428 * config/rs6000/power7.md (power7-load, power7-load-ext,
429 power7-load-update, power7-load-update-indexed,
430 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
431 power7-fpload-update, power7-store, power7-store-update,
432 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
434 * config/rs6000/power8.md (power8-load, power8-load-update,
435 power8-load-ext, power8-load-ext-update, power8-fpload,
436 power8-fpload-update, power8-store, power8-store-update-indexed,
437 power8-fpstore, power8-fpstore-update): Adjust.
438 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
440 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
441 titan_lsu_store, titan_lsu_fpstore): Adjust.
442 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
444 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
447 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
448 unrolled byte insns. Emit address increments after move insns.
450 2014-05-07 David Malcolm <dmalcolm@redhat.com>
452 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
453 const_gimple, rather than a gimple.
454 (gimple_call_builtin_p): Likewise, for the three variants.
456 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
457 (gimple_call_builtin_p): Likewise, for the three variants.
459 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
461 PR tree-optimization/61095
462 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
464 2014-05-07 Richard Biener <rguenther@suse.de>
466 PR tree-optimization/61034
467 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
468 (maybe_skip_until): Use translate to take into account
469 lattices when trying to do disambiguations.
470 (get_continuation_for_phi_1): Likewise.
471 (get_continuation_for_phi): Adjust for added translate arguments.
472 (walk_non_aliased_vuses): Likewise.
473 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
474 (walk_non_aliased_vuses): Likewise.
475 (call_may_clobber_ref_p_1): Declare.
476 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
477 calls. Stop early if we are only supposed to disambiguate.
478 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
480 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
482 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
483 Emit an error when the function has arguments.
485 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
487 * cfgloop.h (unswitch_loops): Remove.
488 * doc/passes.texi: Remove references to loop-unswitch.c
489 * timevar.def (TV_LOOP_UNSWITCH): Remove.
491 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
493 * tree-vect-data-refs.c (vect_grouped_load_supported): New
494 check for loads group of length 3.
495 (vect_permute_load_chain): New permutations for loads group of
497 * tree-vect-stmts.c (vect_model_load_cost): Change cost
498 of vec_perm_shuffle for the new permutations.
500 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
502 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
503 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
504 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
505 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
506 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
507 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
508 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
509 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
511 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
513 * loop-unswitch.c: Delete.
515 2014-05-07 Richard Biener <rguenther@suse.de>
517 * config.gcc: Always set need_64bit_hwint to yes.
519 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
521 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
522 of using optimize_size.
524 2014-05-06 Mike Stump <mikestump@comcast.net>
526 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
528 2014-05-06 Joseph Myers <joseph@codesourcery.com>
530 * config/i386/sse.md (*mov<mode>_internal)
531 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
532 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
533 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
534 (*<code><mode>3, *andnot<mode>3<mask_name>)
535 (<mask_codefor><code><mode>3<mask_name>): Only consider
536 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
538 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
541 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
543 * lra-constraints.c (valid_address_p): Move earlier in file.
544 Add a constraint argument to the address_info version.
545 (satisfies_memory_constraint_p): New function.
546 (satisfies_address_constraint_p): Likewise.
547 (process_alt_operands, curr_insn_transform): Use them.
548 (process_address): Pass the constraint to valid_address_p when
549 checking address operands.
551 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
553 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
554 to their respective blocks. Fix inadvertent use of "node".
556 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
558 * emit-rtl.c (init_derived_machine_modes): New functionm, split
560 (init_emit_once): ...here.
561 * rtl.h (init_derived_machine_modes): Declare.
562 * toplev.c (do_compile): Call it even if no_backend.
564 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
565 Mike Stump <mikestump@comcast.net>
566 Richard Sandiford <rdsandiford@googlemail.com>
567 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
569 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
570 (rtx_equal_for_memref_p): Update comment.
571 (adjust_offset_for_component_ref): Use wide-int interfaces.
572 * builtins.c (get_object_alignment_2): Likewise.
573 (c_readstr): Likewise.
574 (target_char_cast): Add comment.
575 (determine_block_size): Use wide-int interfaces.
576 (expand_builtin_signbit): Likewise.
577 (fold_builtin_int_roundingfn): Likewise.
578 (fold_builtin_bitop): Likewise.
579 (fold_builtin_bswap): Likewise.
580 (fold_builtin_logarithm): Use signop.
581 (fold_builtin_pow): Likewise.
582 (fold_builtin_memory_op): Use wide-int interfaces.
583 (fold_builtin_object_size): Likewise.
584 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
585 nb_iterations_estimate.
586 (record_niter_bound): Use wide-int interfaces.
587 (get_estimated_loop_iterations_int): Likewise.
588 (get_estimated_loop_iterations): Likewise.
589 (get_max_loop_iterations): Likewise.
590 * cfgloop.h: Include wide-int.h.
591 (struct nb_iter_bound): Change bound to widest_int.
592 (struct loop): Change nb_iterations_upper_bound and
593 nb_iterations_estimate to widest_int.
594 (record_niter_bound): Switch to use widest_int.
595 (get_estimated_loop_iterations): Likewise.
596 (get_max_loop_iterations): Likewise.
597 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
599 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
600 * combine.c (try_combine): Likewise.
601 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
602 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
604 (aarch64_float_const_representable_p): Likewise.
605 * config/arc/arc.c: Include wide-int.h.
606 (arc_can_use_doloop_p): Use wide-int interfaces.
607 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
608 (vfp3_const_double_index): Likewise.
609 * config/avr/avr.c (avr_out_round): Likewise.
610 (avr_fold_builtin): Likewise.
611 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
612 (bfin_can_use_doloop_p): Likewise.
613 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
614 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
615 * config/i386/i386.c: Include wide-int.h.
616 (ix86_data_alignment): Use wide-int interfaces.
617 (ix86_local_alignment): Likewise.
618 (ix86_emit_swsqrtsf): Update real_from_integer.
619 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
620 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
621 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
622 (zero_constant): Likewise.
623 (input_operand): Likewise.
624 (splat_input_operand): Likewise.
625 (non_logical_cint_operand): Change const_double to const_wide_int.
626 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
627 (easy_altivec_constant): Remove comment.
628 (paired_expand_vector_init): Use CONSTANT_P.
629 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
630 (rs6000_emit_move): Update checks.
631 (rs6000_aggregate_candidate): Use wide-int interfaces.
632 (rs6000_expand_ternop_builtin): Likewise.
633 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
634 (rs6000_assemble_integer): Likewise.
635 (rs6000_hash_constant): Likewise.
636 (output_toc): Likewise.
637 (rs6000_rtx_costs): Likewise.
638 (rs6000_emit_swrsqrt); Update call to real_from_integer.
639 * config/rs6000/rs6000-c.c: Include wide-int.h.
640 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
641 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
642 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
643 Handle CONST_WIDE_INT.
644 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
645 Use tree_fits_uhwi_p.
646 * config/sparc/sparc.c: Include wide-int.h.
647 (sparc_fold_builtin): Use wide-int interfaces.
648 * config/vax/vax.c: Include wide-int.h.
649 (vax_float_literal): Use real_from_integer.
650 * coretypes.h (struct hwivec_def): New.
653 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
654 (equiv_constant): Handle CONST_WIDE_INT.
655 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
656 (cselib_hash_rtx): Handle CONST_WIDE_INT.
657 * dbxout.c (stabstr_U): Use wide-int interfaces.
658 (dbxout_type): Update to use cst_fits_shwi_p.
659 * defaults.h (LOG2_BITS_PER_UNIT): Define.
660 (TARGET_SUPPORTS_WIDE_INT): Add default.
661 * dfp.c: Include wide-int.h.
662 (decimal_real_to_integer2): Use wide-int interfaces and rename to
663 decimal_real_to_integer.
664 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
665 decimal_real_to_integer.
666 * doc/generic.texi (Constant expressions): Update for wide_int.
667 * doc/rtl.texi (const_double): Likewise.
668 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
669 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
670 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
671 (REAL_VALUE_FROM_INT): Remove.
672 (TARGET_SUPPORTS_WIDE_INT): New.
673 * doc/tm.texi: Regenerate.
674 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
675 * double-int.h: Include wide-int.h.
676 (struct wi::int_traits): New.
677 * dwarf2out.c (get_full_len): New.
678 (dw_val_equal_p): Add case dw_val_class_wide_int.
679 (size_of_loc_descr): Likewise.
680 (output_loc_operands): Likewise.
681 (insert_double): Remove.
682 (insert_wide_int): New.
684 (print_die): Add case dw_val_class_wide_int.
685 (attr_checksum): Likewise.
686 (attr_checksum_ordered): Likewise.
687 (same_dw_val_p): Likewise.
688 (size_of_die): Likewise.
689 (value_format): Likewise.
690 (output_die): Likewise.
691 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
693 (clz_loc_descriptor): Use wide-int interfaces.
694 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
695 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
696 (round_up_to_align): Use wide-int interfaces.
697 (field_byte_offset): Likewise.
698 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
699 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
700 CONST_DOUBLE handling. Use wide-int interfaces.
701 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
702 (gen_enumeration_type_die): Use add_AT_wide.
703 (hash_loc_operands): Add case dw_val_class_wide_int.
704 (compare_loc_operands): Likewise.
705 * dwarf2out.h: Include wide-int.h.
707 (enum dw_val_class): Add dw_val_class_wide_int.
708 (struct dw_val_struct): Add val_wide.
709 * emit-rtl.c (const_wide_int_htab): New.
710 (const_wide_int_htab_hash): New.
711 (const_wide_int_htab_eq): New.
712 (lookup_const_wide_int): New.
713 (const_double_htab_hash): Use wide-int interfaces.
714 (const_double_htab_eq): Likewise.
715 (rtx_to_double_int): Conditionally compile for wide-int.
716 (immed_double_int_const): Rename to immed_wide_int_const and
718 (immed_double_const): Conditionally compile for wide-int.
719 (init_emit_once): Use wide-int interfaces.
720 * explow.c (plus_constant): Likewise.
721 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
722 (lshift_value): Use wide-int interfaces.
723 (expand_mult): Likewise.
724 (choose_multiplier): Likewise.
725 (expand_smod_pow2): Likewise.
726 (make_tree): Likewise.
727 * expr.c (convert_modes): Consolidate handling of constants.
728 Use wide-int interfaces.
729 (emit_group_load_1): Add note.
730 (store_expr): Update comment.
731 (get_inner_reference): Use wide-int interfaces.
732 (expand_constructor): Update comment.
733 (expand_expr_real_2): Use wide-int interfaces.
734 (expand_expr_real_1): Likewise.
735 (reduce_to_bit_field_precision): Likewise.
736 (const_vector_from_tree): Likewise.
737 * final.c: Include wide-int-print.h.
738 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
739 * fixed-value.c: Include wide-int.h.
740 (fixed_from_string): Use wide-int interfaces.
741 (fixed_to_decimal): Likewise.
742 (fixed_convert_from_real): Likewise.
743 (real_convert_from_fixed): Likewise.
744 * fold-const.h (mem_ref_offset): Return an offset_int.
745 (div_if_zero_remainder): Remove code parameter.
746 * fold-const.c (div_if_zero_remainder): Remove code parameter.
747 Use wide-int interfaces.
748 (may_negate_without_overflow_p): Use wide-int interfaces.
749 (negate_expr_p): Likewise.
750 (fold_negate_expr): Likewise.
751 (int_const_binop_1): Likewise.
752 (const_binop): Likewise.
753 (fold_convert_const_int_from_int): Likewise.
754 (fold_convert_const_int_from_real): Likewise.
755 (fold_convert_const_int_from_fixed): Likewise.
756 (fold_convert_const_fixed_from_int): Likewise.
757 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
758 (sign_bit_p): Use wide-int interfaces.
759 (make_range_step): Likewise.
760 (build_range_check): Likewise. Pass an integer of the correct type
761 instead of using integer_one_node.
762 (range_predecessor): Pass an integer of the correct type instead
763 of using integer_one_node.
764 (range_successor): Likewise.
765 (merge_ranges): Likewise.
766 (unextend): Use wide-int interfaces.
767 (extract_muldiv_1): Likewise.
768 (fold_div_compare): Likewise.
769 (fold_single_bit_test): Likewise.
770 (fold_sign_changed_comparison): Likewise.
771 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
772 (fold_plusminus_mult_expr): Use wide-int interfaces.
773 (native_encode_int): Likewise.
774 (native_interpret_int): Likewise.
775 (fold_unary_loc): Likewise.
776 (pointer_may_wrap_p): Likewise.
777 (size_low_cst): Likewise.
778 (mask_with_tz): Likewise.
779 (fold_binary_loc): Likewise.
780 (fold_ternary_loc): Likewise.
781 (multiple_of_p): Likewise.
782 (tree_call_nonnegative_warnv_p): Update calls to
783 tree_int_cst_min_precision and real_from_integer.
784 (fold_negate_const): Use wide-int interfaces.
785 (fold_abs_const): Likewise.
786 (fold_relational_const): Use tree_int_cst_lt.
787 (round_up_loc): Use wide-int interfaces.
788 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
789 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
790 * gengtype.c: Remove include of double-int.h.
791 (do_typedef): Use wide-int interfaces.
792 (open_base_files): Add wide-int.h.
793 (main): Add offset_int and widest_int typedefs.
794 * gengtype-lex.l: Handle "^".
795 (CXX_KEYWORD): Add "static".
796 * gengtype-parse.c (require3): New.
797 (require_template_declaration): Handle constant template arguments
798 and nested templates.
799 * gengtype-state.c: Don't include "double-int.h".
800 * genpreds.c (write_one_predicate_function): Update comment.
801 (write_tm_constrs_h): Add check for hval and lval use in
803 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
804 (add_to_sequence): Likewise.
805 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
806 and const_double_operand.
807 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
809 * gimple-fold.c (get_base_constructor): Likewise.
810 (fold_array_ctor_reference): Likewise.
811 (fold_nonarray_ctor_reference): Likewise.
812 (fold_const_aggregate_ref_1): Likewise.
813 (gimple_val_nonnegative_real_p): Likewise.
814 (gimple_fold_indirect_ref): Likewise.
815 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
816 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
817 (struct slsr_cand_d): Change index to be widest_int.
818 (struct incr_info_d): Change incr to be widest_int.
819 (alloc_cand_and_find_basis): Use wide-int interfaces.
820 (slsr_process_phi): Likewise.
821 (backtrace_base_for_ref): Likewise. Return a widest_int.
822 (restructure_reference): Take a widest_int instead of a double_int.
823 (slsr_process_ref): Use wide-int interfaces.
824 (create_mul_ssa_cand): Likewise.
825 (create_mul_imm_cand): Likewise.
826 (create_add_ssa_cand): Likewise.
827 (create_add_imm_cand): Take a widest_int instead of a double_int.
828 (slsr_process_add): Use wide-int interfaces.
829 (slsr_process_cast): Likewise.
830 (slsr_process_copy): Likewise.
831 (dump_candidate): Likewise.
832 (dump_incr_vec): Likewise.
833 (replace_ref): Likewise.
834 (cand_increment): Likewise. Return a widest_int.
835 (cand_abs_increment): Likewise.
836 (replace_mult_candidate): Take a widest_int instead of a double_int.
837 (replace_unconditional_candidate): Use wide-int interfaces.
838 (incr_vec_index): Take a widest_int instead of a double_int.
839 (create_add_on_incoming_edge): Likewise.
840 (create_phi_basis): Use wide-int interfaces.
841 (replace_conditional_candidate): Likewise.
842 (record_increment): Take a widest_int instead of a double_int.
843 (record_phi_increments): Use wide-int interfaces.
844 (phi_incr_cost): Take a widest_int instead of a double_int.
845 (lowest_cost_path): Likewise.
846 (total_savings): Likewise.
847 (analyze_increments): Use wide-int interfaces.
848 (ncd_with_phi): Take a widest_int instead of a double_int.
849 (ncd_of_cand_and_phis): Likewise.
850 (nearest_common_dominator_for_cands): Likewise.
851 (insert_initializers): Use wide-int interfaces.
852 (all_phi_incrs_profitable): Likewise.
853 (replace_one_candidate): Likewise.
854 (replace_profitable_candidates): Likewise.
855 * godump.c: Include wide-int-print.h.
856 (go_output_typedef): Use wide-int interfaces.
857 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
858 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
859 (build_loop_iteration_domains): Likewise.
860 * hooks.h: Include wide-int.h rather than double-int.h.
861 (hook_bool_dint_dint_uint_bool_true): Delete.
862 (hook_bool_wint_wint_uint_bool_true): Declare.
863 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
864 (hook_bool_wint_wint_uint_bool_true): New.
865 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
867 (ubsan_expand_si_overflow_mul_check): Likewise.
868 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
869 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
870 (get_ancestor_addr_info): Likewise.
871 (ipa_modify_call_arguments): Likewise.
872 * loop-doloop.c (doloop_modify): Likewise.
873 (doloop_optimize): Likewise.
874 * loop-iv.c (iv_number_of_iterations): Likewise.
875 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
876 (unroll_loop_constant_iterations): Likewise.
877 (decide_unroll_runtime_iterations): Likewise.
878 (unroll_loop_runtime_iterations): Likewise.
879 (decide_peel_simple): Likewise.
880 (decide_unroll_stupid): Likewise.
881 * lto-streamer-in.c (streamer_read_wi): Add.
882 (input_cfg): Use wide-int interfaces.
883 (lto_input_tree_1): Likewise.
884 * lto-streamer-out.c (streamer_write_wi): Add.
885 (hash_tree): Use wide-int interfaces.
886 (output_cfg): Likewise.
887 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
888 (GTFILES): Add wide-int.h and signop.h.
889 (TAGS): Look for .cc files too.
890 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
891 * optabs.c (expand_subword_shift): Likewise.
892 (expand_doubleword_shift): Likewise.
893 (expand_absneg_bit): Likewise.
894 (expand_copysign_absneg): Likewise.
895 (expand_copysign_bit): Likewise.
896 * postreload.c (reload_cse_simplify_set): Likewise.
897 * predict.c (predict_iv_comparison): Likewise.
898 * pretty-print.h: Include wide-int-print.h.
900 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
901 * print-tree.c: Include wide-int-print.h.
902 (print_node_brief): Use wide-int interfaces.
903 (print_node): Likewise.
904 * read-rtl.c (validate_const_wide_int): New.
905 (read_rtx_code): Add CONST_WIDE_INT case.
906 * real.c: Include wide-int.h.
907 (real_to_integer2): Delete.
908 (real_to_integer): New function, returning a wide_int.
909 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
910 (ten_to_ptwo): Update call to real_from_integer.
911 (real_digit): Likewise.
912 * real.h: Include signop.h, wide-int.h and insn-modes.h.
913 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
914 (REAL_VALUE_TO_INT): Delete.
915 (real_to_integer): Declare a wide-int form.
916 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
917 * recog.c (const_int_operand): Improve comment.
918 (const_scalar_int_operand): New.
919 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
920 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
921 (split_double): Likewise.
922 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
923 (rtx_size): Likewise.
924 (rtx_alloc_stat_v): New.
925 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
926 (cwi_output_hex): New.
927 (iterative_hash_rtx): Handle CONST_WIDE_INT.
928 (cwi_check_failed_bounds): New.
929 * rtl.def (CONST_WIDE_INT): New.
930 * rtl.h: Include <utility> and wide-int.h.
931 (struct hwivec_def): New.
932 (CWI_GET_NUM_ELEM): New.
933 (CWI_PUT_NUM_ELEM): New.
934 (struct rtx_def): Add num_elem and hwiv.
935 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
936 (CASE_CONST_UNIQUE): Likewise.
937 (CASE_CONST_ANY): Likewise.
938 (CONST_SCALAR_INT_P): Likewise.
939 (CONST_WIDE_INT_P): New.
942 (cwi_check_failed_bounds): New.
945 (CONST_WIDE_INT_VEC) New.
946 (CONST_WIDE_INT_NUNITS) New.
947 (CONST_WIDE_INT_ELT) New.
948 (rtx_mode_t): New type.
949 (wi::int_traits <rtx_mode_t>): New.
951 (wi::min_value): New.
952 (wi::max_value): New.
954 (const_wide_int_alloc): New.
955 (immed_wide_int_const): New.
956 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
957 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
958 * signop.h: New file.
959 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
960 (simplify_const_unary_operation): Use wide-int interfaces.
961 (simplify_binary_operation_1): Likewise.
962 (simplify_const_binary_operation): Likewise.
963 (simplify_const_relational_operation): Likewise.
964 (simplify_immed_subreg): Likewise.
965 * stmt.c (expand_case): Likewise.
966 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
967 signop rather than a bool.
968 * stor-layout.c (layout_type): Use wide-int interfaces.
969 (initialize_sizetypes): Update calls to
970 set_min_and_max_values_for_integral_type.
971 (set_min_and_max_values_for_integral_type): Take a signop rather
972 than a bool. Use wide-int interfaces.
973 (fixup_signed_type): Update accordingly. Remove
974 HOST_BITS_PER_DOUBLE_INT limit.
975 (fixup_unsigned_type): Likewise.
976 * system.h (STATIC_CONSTANT_P): New.
977 (STATIC_ASSERT): New.
978 * target.def (can_use_doloop_p): Take widest_ints rather than
980 * target.h: Include wide-int.h rather than double-int.h.
981 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
983 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
984 rather than INT_CST_LT_UNSIGNED.
985 (can_use_doloop_if_innermost): Take widest_ints rather than
987 * tree-affine.c: Include wide-int-print.h.
988 (double_int_ext_for_comb): Delete.
989 (wide_int_ext_for_comb): New.
990 (aff_combination_zero): Use wide-int interfaces.
991 (aff_combination_const): Take a widest_int instead of a double_int.
992 (aff_combination_elt): Use wide-int interfaces.
993 (aff_combination_scale): Take a widest_int instead of a double_int.
994 (aff_combination_add_elt): Likewise.
995 (aff_combination_add_cst): Likewise.
996 (aff_combination_add): Use wide-int interfaces.
997 (aff_combination_convert): Likewise.
998 (tree_to_aff_combination): Likewise.
999 (add_elt_to_tree): Take a widest_int instead of a double_int.
1000 (aff_combination_to_tree): Use wide-int interfaces.
1001 (aff_combination_remove_elt): Likewise.
1002 (aff_combination_add_product): Take a widest_int instead of
1004 (aff_combination_mult): Use wide-int interfaces.
1005 (aff_combination_expand): Likewise.
1006 (double_int_constant_multiple_p): Delete.
1007 (wide_int_constant_multiple_p): New.
1008 (aff_combination_constant_multiple_p): Take a widest_int pointer
1009 instead of a double_int pointer.
1010 (print_aff): Use wide-int interfaces.
1011 (get_inner_reference_aff): Take a widest_int pointer
1012 instead of a double_int pointer.
1013 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
1014 * tree-affine.h: Include wide-int.h.
1015 (struct aff_comb_elt): Change type of coef to widest_int.
1016 (struct affine_tree_combination): Change type of offset to widest_int.
1017 (double_int_ext_for_comb): Delete.
1018 (wide_int_ext_for_comb): New.
1019 (aff_combination_const): Use widest_int instead of double_int.
1020 (aff_combination_scale): Likewise.
1021 (aff_combination_add_elt): Likewise.
1022 (aff_combination_constant_multiple_p): Likewise.
1023 (get_inner_reference_aff): Likewise.
1024 (aff_comb_cannot_overlap_p): Likewise.
1025 (aff_combination_zero_p): Use wide-int interfaces.
1026 * tree.c: Include tree.h.
1027 (init_ttree): Use make_int_cst.
1028 (tree_code_size): Removed code for INTEGER_CST case.
1029 (tree_size): Add INTEGER_CST case.
1030 (make_node_stat): Update comment.
1031 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
1032 (build_int_cst_type): Use wide-int interfaces.
1033 (double_int_to_tree): Likewise.
1034 (double_int_fits_to_tree_p): Delete.
1035 (force_fit_type_double): Delete.
1036 (force_fit_type): New.
1037 (int_cst_hash_hash): Use wide-int interfaces.
1038 (int_cst_hash_eq): Likewise.
1039 (build_int_cst_wide): Delete.
1040 (wide_int_to_tree): New.
1041 (cache_integer_cst): Use wide-int interfaces.
1042 (build_low_bits_mask): Likewise.
1043 (cst_and_fits_in_hwi): Likewise.
1044 (real_value_from_int_cst): Likewise.
1045 (make_int_cst_stat): New.
1046 (integer_zerop): Use wide_int interfaces.
1047 (integer_onep): Likewise.
1048 (integer_all_onesp): Likewise.
1049 (integer_pow2p): Likewise.
1050 (integer_nonzerop): Likewise.
1051 (tree_log2): Likewise.
1052 (tree_floor_log2): Likewise.
1053 (tree_ctz): Likewise.
1054 (int_size_in_bytes): Likewise.
1055 (mem_ref_offset): Return an offset_int rather than a double_int.
1056 (build_type_attribute_qual_variant): Use wide_int interfaces.
1057 (type_hash_eq): Likewise
1058 (tree_int_cst_equal): Likewise.
1059 (tree_int_cst_lt): Delete.
1060 (tree_int_cst_compare): Likewise.
1061 (tree_fits_shwi_p): Use wide_int interfaces.
1062 (tree_fits_uhwi_p): Likewise.
1063 (tree_int_cst_sign_bit): Likewise.
1064 (tree_int_cst_sgn): Likewise.
1065 (tree_int_cst_min_precision): Take a signop rather than a bool.
1066 (simple_cst_equal): Use wide_int interfaces.
1067 (compare_tree_int): Likewise.
1068 (iterative_hash_expr): Likewise.
1069 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
1071 (get_type_static_bounds): Use wide_int interfaces.
1072 (tree_int_cst_elt_check_failed): New.
1073 (build_common_tree_nodes): Reordered to set prec before filling in
1075 (int_cst_value): Check cst_and_fits_in_hwi.
1076 (widest_int_cst_value): Use wide_int interfaces.
1077 (upper_bound_in_type): Likewise.
1078 (lower_bound_in_type): Likewise.
1079 (num_ending_zeros): Likewise.
1080 (drop_tree_overflow): Likewise.
1081 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
1082 (gen_conditions_for_pow_cst_base): Likewise.
1083 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
1084 (group_case_labels_stmt): Use wide-int interfaces.
1085 (verify_gimple_assign_binary): Likewise.
1086 (print_loop): Likewise.
1087 * tree-chrec.c (tree_fold_binomial): Likewise.
1088 * tree-core.h (struct tree_base): Add int_length.
1089 (struct tree_int_cst): Change rep of value.
1090 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
1091 (dr_may_alias_p): Likewise.
1092 (max_stmt_executions_tree): Likewise.
1093 * tree.def (INTEGER_CST): Update comment.
1094 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
1095 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
1096 * tree-dump.c: Include wide-int.h and wide-int-print.h.
1097 (dequeue_and_dump): Use wide-int interfaces.
1098 * tree.h: Include wide-int.h.
1099 (NULL_TREE): Moved to earlier loc in file.
1100 (TREE_INT_CST_ELT_CHECK): New.
1101 (tree_int_cst_elt_check_failed): New.
1103 (TREE_INT_CST): Delete.
1104 (TREE_INT_CST_LOW): Use wide-int interfaces.
1105 (TREE_INT_CST_HIGH): Delete.
1106 (TREE_INT_CST_NUNITS): New.
1107 (TREE_INT_CST_EXT_NUNITS): Likewise.
1108 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
1109 (TREE_INT_CST_ELT): Likewise.
1110 (INT_CST_LT): Delete.
1111 (tree_int_cst_elt_check): New (two forms).
1112 (type_code_size): Update comment.
1113 (make_int_cst_stat, make_int_cst): New.
1114 (tree_to_double_int): Delete.
1115 (double_int_fits_to_tree_p): Delete.
1116 (force_fit_type_double): Delete.
1117 (build_int_cstu): Replace with out-of-line function.
1118 (build_int_cst_wide): Delete.
1119 (tree_int_cst_lt): Define inline.
1120 (tree_int_cst_le): New.
1121 (tree_int_cst_compare): Define inline.
1122 (tree_int_cst_min_precision): Take a signop rather than a bool.
1123 (wi::int_traits <const_tree>): New.
1124 (wi::int_traits <tree>): New.
1125 (wi::extended_tree): New.
1126 (wi::int_traits <wi::extended_tree>): New.
1127 (wi::to_widest): New.
1128 (wi::to_offset): New.
1129 (wi::fits_to_tree_p): New.
1130 (wi::min_value): New.
1131 (wi::max_value): New.
1132 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
1133 (copy_tree_body_r): Likewise.
1134 * tree-object-size.c (compute_object_offset): Likewise.
1135 (addr_object_size): Likewise.
1136 * tree-predcom.c: Include wide-int-print.h.
1137 (struct dref_d): Change type of offset to widest_int.
1138 (dump_dref): Call wide-int printer.
1139 (aff_combination_dr_offset): Use wide-int interfaces.
1140 (determine_offset): Take a widest_int pointer rather than a
1142 (split_data_refs_to_components): Use wide-int interfaces.
1143 (suitable_component_p): Likewise.
1144 (order_drefs): Likewise.
1145 (add_ref_to_chain): Likewise.
1146 (valid_initializer_p): Likewise.
1147 (determine_roots_comp): Likewise.
1148 * tree-pretty-print.c: Include wide-int-print.h.
1149 (dump_generic_node): Use wide-int interfaces.
1150 * tree-sra.c (sra_ipa_modify_expr): Likewise.
1151 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
1152 (move_fixed_address_to_symbol): Likewise.
1153 (move_hint_to_base): Likewise.
1154 (move_pointer_to_base): Likewise.
1155 (move_variant_to_index): Likewise.
1156 (most_expensive_mult_to_index): Likewise.
1157 (addr_to_parts): Likewise.
1158 (copy_ref_info): Likewise.
1159 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
1160 (indirect_refs_may_alias_p): Likewise.
1161 (stmt_kills_ref_p_1): Likewise.
1162 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
1163 * tree-ssa-ccp.c: Update comment at top of file. Include
1165 (struct prop_value_d): Change type of mask to widest_int.
1166 (extend_mask): New function.
1167 (dump_lattice_value): Use wide-int interfaces.
1168 (get_default_value): Likewise.
1169 (set_constant_value): Likewise.
1170 (set_value_varying): Likewise.
1171 (valid_lattice_transition): Likewise.
1172 (set_lattice_value): Likewise.
1173 (value_to_double_int): Delete.
1174 (value_to_wide_int): New.
1175 (get_value_from_alignment): Use wide-int interfaces.
1176 (get_value_for_expr): Likewise.
1177 (do_dbg_cnt): Likewise.
1178 (ccp_finalize): Likewise.
1179 (ccp_lattice_meet): Likewise.
1180 (bit_value_unop_1): Use widest_ints rather than double_ints.
1181 (bit_value_binop_1): Likewise.
1182 (bit_value_unop): Use wide-int interfaces.
1183 (bit_value_binop): Likewise.
1184 (bit_value_assume_aligned): Likewise.
1185 (evaluate_stmt): Likewise.
1186 (ccp_fold_stmt): Likewise.
1187 (visit_cond_stmt): Likewise.
1188 (ccp_visit_stmt): Likewise.
1189 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
1190 (constant_pointer_difference): Likewise.
1191 (associate_pointerplus): Likewise.
1192 (combine_conversions): Likewise.
1193 * tree-ssa-loop.h: Include wide-int.h.
1194 (struct tree_niter_desc): Change type of max to widest_int.
1195 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
1196 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
1197 (remove_redundant_iv_tests): Likewise.
1198 (canonicalize_loop_induction_variables): Likewise.
1199 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
1200 (constant_multiple_of): Take a widest_int pointer instead of
1201 a double_int pointer.
1202 (get_computation_aff): Use wide-int interfaces.
1203 (ptr_difference_cost): Likewise.
1204 (difference_cost): Likewise.
1205 (get_loop_invariant_expr_id): Likewise.
1206 (get_computation_cost_at): Likewise.
1207 (iv_elimination_compare_lt): Likewise.
1208 (may_eliminate_iv): Likewise.
1209 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
1210 instead of double_int.
1211 (max_loop_iterations): Likewise.
1212 (max_stmt_executions): Likewise.
1213 (estimated_stmt_executions): Likewise.
1214 * tree-ssa-loop-niter.c: Include wide-int-print.h.
1215 (split_to_var_and_offset): Use wide-int interfaces.
1216 (determine_value_range): Likewise.
1217 (bound_difference_of_offsetted_base): Likewise.
1218 (bounds_add): Take a widest_int instead of a double_int.
1219 (number_of_iterations_ne_max): Use wide-int interfaces.
1220 (number_of_iterations_ne): Likewise.
1221 (number_of_iterations_lt_to_ne): Likewise.
1222 (assert_loop_rolls_lt): Likewise.
1223 (number_of_iterations_lt): Likewise.
1224 (number_of_iterations_le): Likewise.
1225 (number_of_iterations_cond): Likewise.
1226 (number_of_iterations_exit): Likewise.
1227 (finite_loop_p): Likewise.
1228 (derive_constant_upper_bound_assign): Likewise.
1229 (derive_constant_upper_bound): Return a widest_int.
1230 (derive_constant_upper_bound_ops): Likewise.
1231 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
1232 (record_estimate): Take a widest_int rather than a double_int.
1233 (record_nonwrapping_iv): Use wide-int interfaces.
1234 (double_int_cmp): Delete.
1235 (wide_int_cmp): New.
1236 (bound_index): Take a widest_int rather than a double_int.
1237 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
1238 (maybe_lower_iteration_bound): Likewise.
1239 (estimate_numbers_of_iterations_loop): Likewise.
1240 (estimated_loop_iterations): Take a widest_int pointer than than
1241 a double_int pointer.
1242 (estimated_loop_iterations_int): Use wide-int interfaces.
1243 (max_loop_iterations): Take a widest_int pointer than than
1244 a double_int pointer.
1245 (max_loop_iterations_int): Use wide-int interfaces.
1246 (max_stmt_executions): Take a widest_int pointer than than
1247 a double_int pointer.
1248 (estimated_stmt_executions): Likewise.
1249 (n_of_executions_at_most): Use wide-int interfaces.
1250 (scev_probably_wraps_p): Likewise.
1251 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
1253 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
1255 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
1256 double_ints. Adjust for trailing_wide_ints <3> representation.
1257 (set_nonzero_bits): Likewise.
1258 (get_range_info): Return wide_ints rather than double_ints.
1259 Adjust for trailing_wide_ints <3> representation.
1260 (get_nonzero_bits): Likewise.
1261 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
1263 * tree-ssanames.h (struct range_info_def): Replace min, max and
1264 nonzero_bits with a trailing_wide_ints <3>.
1265 (set_range_info): Use wide_int_refs rather than double_ints.
1266 (set_nonzero_bits): Likewise.
1267 (get_range_info): Return wide_ints rather than double_ints.
1268 (get_nonzero_bits): Likewise.
1269 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
1270 * tree-ssa-pre.c (phi_translate_1): Likewise.
1271 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
1272 (acceptable_pow_call): Likewise.
1273 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
1275 (vn_reference_fold_indirect): Likewise.
1276 (vn_reference_maybe_forwprop_address): Likewise.
1277 (valueize_refs_1): Likewise.
1278 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
1279 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
1280 tree_int_cst_lt and tree_int_cst_le.
1281 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
1283 (streamer_alloc_tree): Likewise.
1284 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
1285 (streamer_write_tree_header): Likewise.
1286 (streamer_write_integer_cst): Likewise.
1287 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
1288 (build_constructors): Likewise.
1289 (array_value_type): Likewise.
1290 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
1291 (vect_check_gather): Likewise.
1292 * tree-vect-generic.c (build_replicated_const): Likewise.
1293 (expand_vector_divmod): Likewise.
1294 * tree-vect-loop.c (vect_transform_loop): Likewise.
1295 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
1296 (vect_do_peeling_for_alignment): Likewise.
1297 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
1298 * tree-vrp.c: Include wide-int.h.
1299 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
1300 (extract_range_from_assert): Use wide-int interfaces.
1301 (vrp_int_const_binop): Likewise.
1302 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
1303 double_int pointers.
1304 (ranges_from_anti_range): Use wide-int interfaces.
1305 (quad_int_cmp): Delete.
1306 (quad_int_pair_sort): Likewise.
1307 (extract_range_from_binary_expr_1): Use wide-int interfaces.
1308 (extract_range_from_unary_expr_1): Likewise.
1309 (adjust_range_with_scev): Likewise.
1310 (masked_increment): Take and return wide_ints rather than double_ints.
1311 (register_edge_assert_for_2): Use wide-int interfaces.
1312 (check_array_ref): Likewise.
1313 (search_for_addr_array): Likewise.
1314 (maybe_set_nonzero_bits): Likewise.
1315 (union_ranges): Pass an integer of the correct type instead of
1316 using integer_one_node.
1317 (intersect_ranges): Likewise.
1318 (simplify_truth_ops_using_ranges): Likewise.
1319 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
1320 (range_fits_type_p): Likewise.
1321 (simplify_cond_using_ranges): Likewise. Take a signop rather than
1323 (simplify_conversion_using_ranges): Use wide-int interfaces.
1324 (simplify_float_conversion_using_ranges): Likewise.
1325 (vrp_finalize): Likewise.
1326 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
1327 (gimple_stringops_transform): Likewise.
1328 * varasm.c (decode_addr_const): Likewise.
1329 (const_hash_1): Likewise.
1330 (const_rtx_hash_1): Likewise
1331 (output_constant): Likewise.
1332 (array_size_for_constructor): Likewise.
1333 (output_constructor_regular_field): Likewise.
1334 (output_constructor_bitfield): Likewise.
1335 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
1336 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
1338 * gencheck.c: Define BITS_PER_UNIT.
1341 * wide-int-print.cc: New.
1342 * wide-int-print.h: New.
1344 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1346 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
1348 2014-05-06 Richard Biener <rguenther@suse.de>
1350 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
1351 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
1352 (TODO_verify_all): Adjust.
1353 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
1354 TODO_verify_stmts and TODO_verify_rtl_sharing.
1355 * bb-reorder.c: Likewise.
1356 * cfgexpand.c: Likewise.
1357 * cprop.c: Likewise.
1359 * function.c: Likewise.
1360 * fwprop.c: Likewise.
1362 * gimple-ssa-isolate-paths.c: Likewise.
1363 * gimple-ssa-strength-reduction.c: Likewise.
1364 * ipa-split.c: Likewise.
1365 * loop-init.c: Likewise.
1366 * loop-unroll.c: Likewise.
1367 * lower-subreg.c: Likewise.
1368 * modulo-sched.c: Likewise.
1369 * postreload-gcse.c: Likewise.
1370 * predict.c: Likewise.
1371 * recog.c: Likewise.
1372 * sched-rgn.c: Likewise.
1373 * store-motion.c: Likewise.
1374 * tracer.c: Likewise.
1375 * trans-mem.c: Likewise.
1376 * tree-call-cdce.c: Likewise.
1377 * tree-cfg.c: Likewise.
1378 * tree-cfgcleanup.c: Likewise.
1379 * tree-complex.c: Likewise.
1380 * tree-eh.c: Likewise.
1381 * tree-emutls.c: Likewise.
1382 * tree-if-conv.c: Likewise.
1383 * tree-into-ssa.c: Likewise.
1384 * tree-loop-distribution.c: Likewise.
1385 * tree-object-size.c: Likewise.
1386 * tree-parloops.c: Likewise.
1387 * tree-pass.h: Likewise.
1388 * tree-sra.c: Likewise.
1389 * tree-ssa-ccp.c: Likewise.
1390 * tree-ssa-copy.c: Likewise.
1391 * tree-ssa-copyrename.c: Likewise.
1392 * tree-ssa-dce.c: Likewise.
1393 * tree-ssa-dom.c: Likewise.
1394 * tree-ssa-dse.c: Likewise.
1395 * tree-ssa-forwprop.c: Likewise.
1396 * tree-ssa-ifcombine.c: Likewise.
1397 * tree-ssa-loop-ch.c: Likewise.
1398 * tree-ssa-loop-ivcanon.c: Likewise.
1399 * tree-ssa-loop.c: Likewise.
1400 * tree-ssa-math-opts.c: Likewise.
1401 * tree-ssa-phiopt.c: Likewise.
1402 * tree-ssa-phiprop.c: Likewise.
1403 * tree-ssa-pre.c: Likewise.
1404 * tree-ssa-reassoc.c: Likewise.
1405 * tree-ssa-sink.c: Likewise.
1406 * tree-ssa-strlen.c: Likewise.
1407 * tree-ssa-tail-merge.c: Likewise.
1408 * tree-ssa-uncprop.c: Likewise.
1409 * tree-switch-conversion.c: Likewise.
1410 * tree-tailcall.c: Likewise.
1411 * tree-vect-generic.c: Likewise.
1412 * tree-vectorizer.c: Likewise.
1413 * tree-vrp.c: Likewise.
1415 * var-tracking.c: Likewise.
1416 * bt-load.c: Likewise.
1417 * cfgcleanup.c: Likewise.
1418 * combine-stack-adj.c: Likewise.
1419 * combine.c: Likewise.
1420 * compare-elim.c: Likewise.
1421 * config/epiphany/resolve-sw-modes.c: Likewise.
1422 * config/i386/i386.c: Likewise.
1423 * config/mips/mips.c: Likewise.
1424 * config/s390/s390.c: Likewise.
1425 * config/sh/sh_treg_combine.cc: Likewise.
1426 * config/sparc/sparc.c: Likewise.
1429 * final.c: Likewise.
1430 * ifcvt.c: Likewise.
1431 * mode-switching.c: Likewise.
1432 * passes.c: Likewise.
1433 * postreload.c: Likewise.
1435 * reg-stack.c: Likewise.
1436 * regcprop.c: Likewise.
1437 * regrename.c: Likewise.
1440 2014-05-06 Richard Biener <rguenther@suse.de>
1443 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
1444 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
1446 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
1449 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
1451 2014-05-05 Radovan Obradovic <robradovic@mips.com>
1452 Tom de Vries <tom@codesourcery.com>
1454 * target.def (call_fusage_contains_non_callee_clobbers): New
1456 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
1458 (@node Miscellaneous Register Hooks): New node.
1459 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
1460 * doc/tm.texi: Regenerate.
1462 2014-05-05 Marek Polacek <polacek@redhat.com>
1465 * opts.c (common_handle_option): Call error_at instead of warning_at.
1467 2014-05-05 Richard Biener <rguenther@suse.de>
1469 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
1470 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
1471 under the TODO_verify_il umbrella.
1473 2014-05-05 Richard Biener <rguenther@suse.de>
1475 * passes.c (execute_function_todo): Move TODO_verify_flow under
1476 the TODO_verify_ul umbrella.
1478 2014-05-05 Richard Biener <rguenther@suse.de>
1481 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
1482 X & CST away from a CST that is the mask of a mode.
1484 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1486 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
1487 int argument to enum machine_mode.
1488 (picochip_class_max_nregs): Ditto.
1489 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
1490 (picochip_class_max_nregs): Ditto.
1492 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1494 * target.def: Add new target hook.
1495 * doc/tm.texi: Regenerate.
1496 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
1497 * targhooks.c (default_keep_leaf_when_profiled): New function.
1499 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
1500 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
1502 2014-05-05 Bin Cheng <bin.cheng@arm.com>
1504 PR tree-optimization/60363
1505 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
1506 (copy_phi_args): New parameters. Call get_value_locus_in_path.
1507 (update_destination_phis): New parameter.
1508 (create_edge_and_update_destination_phis): Ditto.
1509 (ssa_fix_duplicate_block_edges): Pass new arguments.
1510 (thread_single_edge): Ditto.
1512 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
1514 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
1515 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
1516 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
1517 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
1518 Use RS6000_BTM_HARD_FLOAT.
1519 (BU_MISC_2): Likewise.
1520 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
1521 RS6000_BTM_HARD_FLOAT.
1522 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
1524 (rs6000_invalid_builtin): Add hard floating builtin support.
1525 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
1526 hard float builtins.
1527 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
1529 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1531 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
1532 Add missing function* argument.
1534 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1536 * lra-constraints.c (valid_address_p): Move earlier in file.
1537 Add a constraint argument to the address_info version.
1538 (satisfies_memory_constraint_p): New function.
1539 (satisfies_address_constraint_p): Likewise.
1540 (process_alt_operands, curr_insn_transform): Use them.
1541 (process_address): Pass the constraint to valid_address_p when
1542 checking address operands.
1544 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1546 * config/mips/mips.c (mips_isa_rev): New variable.
1547 (mips_set_architecture): Set it.
1548 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1550 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1551 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1552 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1553 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1554 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
1555 conditions in terms of mips_isa_rev.
1556 (mips_isa_rev): Declare.
1558 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1560 * config/sh/sh-mem.cc: Use tabs instead of spaces.
1561 (prob_unlikely, prob_likely): Make variables const.
1563 2014-05-03 Denis Chertykov <chertykov@gmail.com>
1565 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
1567 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1569 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
1571 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1573 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
1574 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
1575 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
1577 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
1578 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
1580 Replace usage of ROUND_REG with sh_round_reg.
1581 Use CEIL instead of ROUND_ADVANCE.
1583 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1586 * config/sh/sh.c: Include stdlib headers before everything else.
1588 2014-05-02 Jakub Jelinek <jakub@redhat.com>
1590 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
1591 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
1592 (gimplify_adjust_omp_clauses): Simd region is never
1593 directly nested in combined parallel. Instead, for linear
1594 with copyin/copyout, if in combined for simd loop, make decl
1595 firstprivate/lastprivate on OMP_FOR.
1596 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
1597 expand_omp_for_static_chunk): When setting endvar, also set
1598 fd->loop.v to the same value.
1600 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1602 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
1604 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
1606 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
1609 2014-05-02 Marek Polacek <polacek@redhat.com>
1611 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
1613 2014-05-02 Kito Cheng <kito@0xlab.org>
1615 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
1616 to a C expression marco.
1617 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
1618 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
1619 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
1620 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
1621 HONOR_REG_ALLOC_ORDER.
1622 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
1624 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1626 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
1628 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1630 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
1632 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
1634 * tree-if-conv.c (is_cond_scalar_reduction): New function.
1635 (convert_scalar_cond_reduction): Likewise.
1636 (predicate_scalar_phi): Add recognition and transformation
1637 of simple conditioanl reduction to be vectorizable.
1639 2014-05-01 Marek Polacek <polacek@redhat.com>
1642 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
1644 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
1646 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
1647 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
1648 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
1649 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
1650 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
1651 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
1652 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
1653 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
1655 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
1657 * config/arc/arc.opt (mlra): Move comment above option name
1658 to avoid mis-parsing as language options.
1660 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1662 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
1663 * config/sol2.h: ... here.
1664 * config/sol2-10.h: Remove.
1666 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
1667 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
1668 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
1669 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
1670 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
1671 * config/sol2.h: ... here.
1672 (SECTION_NAME_FORMAT): Don't redefine.
1673 (STARTFILE_ARCH32_SPEC): Rename to ...
1674 (STARTFILE_ARCH_SPEC): ... this.
1675 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
1676 * config/sparc/sol2.h: ... here.
1677 (SECTION_NAME_FORMAT): Don't undef.
1678 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
1679 (SUBTARGET_EXTRA_SPECS): Remove.
1680 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
1682 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
1683 (MD_STARTFILE_PREFIX): Remove.
1684 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
1685 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
1686 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
1687 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
1688 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
1689 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
1690 * config/i386/sol2.h: ... here.
1691 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
1692 * config/i386/sol2-bi.h: Remove.
1693 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
1694 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
1696 * config/i386/t-sol2-64: Rename to ...
1697 * config/i386/t-sol2: ... this.
1698 * config/sparc/t-sol2-64: Rename to ...
1699 * config/sparc/t-sol2: ... this.
1701 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
1702 sol2_tm_file_head, sol2_tm_file_tail.
1703 Include ${cpu_type}/sol2.h before sol2.h.
1705 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
1706 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
1707 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
1708 Reflect i386/t-sol2-64 renaming.
1709 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
1710 Reflect sparc/t-sol2-64 renaming.
1712 2014-04-30 Richard Biener <rguenther@suse.de>
1714 * passes.c (execute_function_todo): Move TODO_verify_stmts
1715 and TODO_verify_ssa under the TODO_verify_il umbrella.
1716 * tree-ssa.h (verify_ssa): Adjust prototype.
1717 * tree-ssa.c (verify_ssa): Add parameter to tell whether
1718 we should verify SSA operands.
1719 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
1720 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
1721 whether we should verify whether not throwing stmts have EH info.
1722 * graphite-scop-detection.c (create_sese_edges): Adjust.
1723 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
1724 * tree-eh.c (lower_try_finally_switch): Do not add the
1725 default case label twice.
1727 2014-04-30 Marek Polacek <polacek@redhat.com>
1729 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
1730 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
1731 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
1732 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
1734 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
1736 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
1737 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
1738 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
1739 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
1740 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
1741 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
1742 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
1743 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
1745 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1747 * tree-cfg.c (dump_function_to_file): Dump the return type of
1748 functions, in a line to itself before the function body, mimicking
1749 the layout of a C function.
1751 2014-04-29 Jakub Jelinek <jakub@redhat.com>
1753 PR tree-optimization/60971
1754 * tree-tailcall.c (process_assignment): Reject conversions which
1757 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
1759 * calls.c (initialize_argument_information): Always treat
1760 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
1761 (expand_call): Likewise.
1762 (emit_library_call_calue_1): Likewise.
1763 * expr.c (PUSH_ARGS_REVERSED): Do not define.
1764 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
1767 2014-04-29 Nick Clifton <nickc@redhat.com>
1769 * config/msp430/msp430.md (umulsidi): Fix typo.
1770 (mulhisi3): Enable even inside interrupt handlers.
1771 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
1772 bigger return address pushed in large mode.
1774 2014-04-29 Nick Clifton <nickc@redhat.com>
1776 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
1777 (arc_init_reg_tables): Use a machine_mode enum to iterate over
1779 * config/m32r/m32r.c (init_reg_tables): Likewise.
1780 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
1781 enum to hold the modes.
1783 2014-04-29 Richard Biener <rguenther@suse.de>
1785 * dominance.c (free_dominance_info): Add overload with
1787 (dom_info_state): Likewise.
1788 (dom_info_available_p): Likewise.
1789 * basic-block.h (free_dominance_info, dom_info_state,
1790 dom_info_available_p): Declare overloads.
1791 * passes.c (execute_function_todo): Verify that verifiers
1792 don't change dominator info state. Drop dominator info
1793 for IPA pass invocations.
1794 * cgraph.c (release_function_body): Restore asserts that
1795 dominator information is released.
1797 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
1799 * doc/invoke.texi: Fix typo.
1800 * tree-vrp.c: Fix typos.
1801 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
1803 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1805 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
1807 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1809 * config/aarch64/aarch64-builtins.c
1810 (aarch64_types_storestruct_lane_qualifiers): New.
1811 (TYPES_STORESTRUCT_LANE): Likewise.
1812 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
1813 (st3_lane): Likewise.
1814 (st4_lane): Likewise.
1815 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
1816 (vec_store_lanesci_lane<mode>): Likewise.
1817 (vec_store_lanesxi_lane<mode>): Likewise.
1818 (aarch64_st2_lane<VQ:mode>): Likewise.
1819 (aarch64_st3_lane<VQ:mode>): Likewise.
1820 (aarch64_st4_lane<VQ:mode>): Likewise.
1821 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
1822 * config/aarch64/arm_neon.h
1823 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
1824 use new macro arguments.
1825 (__ST3_LANE_FUNC): Likewise.
1826 (__ST4_LANE_FUNC): Likewise.
1827 * config/aarch64/iterators.md (V_TWO_ELEM): New.
1828 (V_THREE_ELEM): Likewise.
1829 (V_FOUR_ELEM): Likewise.
1831 2014-04-28 David Malcolm <dmalcolm@redhat.com>
1833 * doc/gimple.texi: Replace the description of the now-defunct
1834 union gimple_statement_d with a diagram showing the
1835 gimple_statement_base class hierarchy and its relationships to
1836 the GSS_ and GIMPLE_ enums.
1838 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
1840 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
1841 * config/aarch64/aarch64.c
1842 (aarch64_cannot_change_mode_class): Weaken conditions.
1843 (aarch64_modes_tieable_p): New.
1844 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
1846 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
1848 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
1849 (loadsync_<mode>): Change mode.
1850 (load_quadpti, store_quadpti): New.
1851 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
1852 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
1854 2014-04-28 Martin Jambor <mjambor@suse.cz>
1856 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
1857 same alias type as the original statement.
1858 (subreplacement_assignment_data): New type.
1859 (handle_unscalarized_data_in_subtree): New type of parameter,
1860 generate new memory accesses with same alias type as the original
1862 (load_assign_lhs_subreplacements): Likewise.
1863 (sra_modify_constructor_assign): Generate new memory accesses with
1864 same alias type as the original statement.
1866 2014-04-28 Richard Biener <rguenther@suse.de>
1868 * tree-pass.h (TODO_verify_il): Define.
1869 (TODO_verify_all): Complete properly.
1870 * passes.c (execute_function_todo): Move existing loop-closed
1871 SSA verification under TODO_verify_il.
1872 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
1873 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
1874 Fix tree sharing issue.
1876 2014-04-28 Richard Biener <rguenther@suse.de>
1879 * builtins.def (DEF_C11_BUILTIN): Add.
1880 (BUILT_IN_ALIGNED_ALLOC): Likewise.
1881 * coretypes.h (enum function_class): Add function_c11_misc.
1882 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1883 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
1884 (call_may_clobber_ref_p_1): Likewise.
1885 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
1886 (mark_all_reaching_defs_necessary_1): Likewise.
1887 (propagate_necessity): Likewise.
1888 (eliminate_unnecessary_stmts): Likewise.
1889 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
1891 2014-04-28 Richard Biener <rguenther@suse.de>
1893 * tree-vrp.c (vrp_var_may_overflow): Remove.
1894 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
1895 with overflow immediately bump to one before that value and
1896 let iteration figure out overflow status.
1898 2014-04-28 Richard Biener <rguenther@suse.de>
1900 * configure.ac: Do valgrind header checks unconditionally.
1901 Add --enable-valgrind-annotations.
1902 * system.h: Guard valgrind header inclusion with
1903 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
1904 * alloc-pool.c (pool_alloc, pool_free): Use
1905 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
1906 to guard possibly dead code.
1907 * config.in: Regenerated.
1908 * configure: Likewise.
1910 2014-04-28 Jeff Law <law@redhat.com>
1912 PR tree-optimization/60902
1913 * tree-ssa-threadedge.c
1914 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
1915 over real defs when invalidating outputs from statements that do not
1916 produce useful outputs for threading.
1918 2014-04-28 Richard Biener <rguenther@suse.de>
1920 PR tree-optimization/60979
1921 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
1922 SCOPs that end in a block with a successor with abnormal
1925 2014-04-28 Richard Biener <rguenther@suse.de>
1927 * tree-pass.h (execute_pass_list): Adjust prototype.
1928 * passes.c (pass_manager::execute_early_local_passes): Adjust.
1929 (do_per_function): Change callback signature, push all actual
1930 work to the callbals.
1931 (do_per_function_toporder): Likewise.
1932 (execute_function_dump): Adjust.
1933 (execute_function_todo): Likewise.
1934 (clear_last_verified): Likewise.
1935 (verify_curr_properties): Likewise.
1936 (update_properties_after_pass): Likewise.
1937 (execute_pass_list_1): Split out from ...
1938 (execute_pass_list): ... here. Adjust.
1939 (execute_ipa_pass_list): Likewise.
1940 * cgraphunit.c (cgraph_add_new_function): Adjust.
1941 (analyze_function): Likewise.
1942 (expand_function): Likewise.
1943 * cgraph.c (release_function_body): Free dominance info
1944 here instead of asserting it was magically freed elsewhere.
1946 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
1948 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
1949 * configure: Regenerate.
1950 * config/sparc/sparc.opt (muser-mode): New option.
1951 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
1953 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
1954 * doc/invoke.texi (SPARC options): Document -muser-mode.
1956 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
1958 * cselib.c (find_slot_memmode): Delete.
1959 (cselib_hasher): Change compare_type to a struct.
1960 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
1962 (preserve_constants_and_equivs): Adjust for new compare_type.
1963 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
1964 (wrap_constant): Delete.
1965 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
1967 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
1969 * doc/install.texi (Building with profile feedback): Remove
1972 2014-04-26 Tom de Vries <tom@codesourcery.com>
1974 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
1977 2014-04-25 Cary Coutant <ccoutant@google.com>
1980 * dwarf2out.c (should_move_die_to_comdat): A type definition
1981 can contain a subprogram definition, but don't move it to a
1983 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
1984 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
1986 (clone_tree_hash): Rename to...
1987 (clone_tree_partial): ...this; change callers. Copy
1988 DW_TAG_subprogram DIEs as declarations.
1989 (copy_decls_walk): Don't copy children of a declaration into a
1992 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
1995 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
1998 2014-04-25 Jiong Wang <jiong.wang@arm.com>
2000 * config/arm/predicates.md (call_insn_operand): Add long_call check.
2001 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
2003 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
2006 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2008 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
2010 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2012 PR tree-optimization/60930
2013 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
2014 creating a multiply candidate by folding two constant
2015 multiplicands when the result overflows.
2017 2014-04-25 Jakub Jelinek <jakub@redhat.com>
2019 PR tree-optimization/60960
2020 * tree-vect-generic.c (expand_vector_operation): Only call
2021 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
2023 2014-04-25 Tom de Vries <tom@codesourcery.com>
2025 * expr.c (clobber_reg_mode): New function.
2026 * expr.h (clobber_reg): New function.
2028 2014-04-25 Tom de Vries <tom@codesourcery.com>
2030 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
2033 2014-04-25 Radovan Obradovic <robradovic@mips.com>
2034 Tom de Vries <tom@codesourcery.com>
2036 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
2038 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
2039 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
2040 new argument to find_all_hard_reg_sets call.
2042 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2044 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
2045 Use HOST_WIDE_INT_C for mask literal.
2046 (aarch_rev16_shleft_mask_imm_p): Likewise.
2048 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
2051 * config/sparc/sparc.md (ashlsi3_extend): Delete.
2053 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
2055 PR preprocessor/56540
2056 * config/i386/i386-c.c (ix86_target_macros): Define
2057 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
2059 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2061 * configure.ac (tga_func): Remove.
2062 (LIB_TLS_SPEC): Remove.
2063 * configure: Regenerate.
2064 * config.in: Regenerate.
2065 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
2067 2014-04-25 Richard Biener <rguenther@suse.de>
2070 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
2071 call stmt use/clobber sets during stmt walk instead of
2072 walking the possibly incomplete set of caller edges.
2074 2014-04-25 Richard Biener <rguenther@suse.de>
2077 * passes.c (apply_ipa_transforms): Inline into only caller ...
2078 (execute_one_pass): ... here. Properly bring in function
2079 bodies for nodes we want to apply IPA transforms to.
2081 2014-04-24 Cong Hou <congh@google.com>
2083 PR tree-optimization/60896
2084 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
2085 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
2086 (vect_mark_pattern_stmts): Set the def type of all statements in
2087 PATTERN_DEF_SEQ as vect_internal_def.
2089 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
2091 * doc/extend.texi (PowerPC Built-in Functions): Document new
2092 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
2093 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
2095 * config/rs6000/predicates.md (const_0_to_3_operand): New
2096 predicate to match 0..3 integer constants.
2098 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
2099 to support adding miscellaneous builtin functions.
2100 (BU_DFP_MISC_2): Likewise.
2101 (BU_P7_MISC_1): Likewise.
2102 (BU_P7_MISC_2): Likewise.
2103 (BU_P8V_MISC_3): Likewise.
2104 (BU_MISC_1): Likewise.
2105 (BU_MISC_2): Likewise.
2106 (DIVWE): Add extended divide builtin functions.
2109 (DIVWEUO): Likewise.
2113 (DIVDEUO): Likewise.
2114 (DXEX): Add decimal floating-point builtin functions.
2117 (DDEDPDQ): Likewise.
2119 (DENBCDQ): Likewise.
2126 (CDTBCD): Add new BCD builtin functions.
2130 (BCDADD_LT): Likewise.
2131 (BCDADD_EQ): Likewise.
2132 (BCDADD_GT): Likewise.
2133 (BCDADD_OV): Likewise.
2135 (BCDSUB_LT): Likewise.
2136 (BCDSUB_EQ): Likewise.
2137 (BCDSUB_GT): Likewise.
2138 (BCDSUB_OV): Likewise.
2139 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
2140 (UNPACK_TD): Likewise.
2141 (PACK_TF): Likewise.
2142 (UNPACK_TF): Likewise.
2143 (UNPACK_TF_0): Likewise.
2144 (UNPACK_TF_1): Likewise.
2145 (PACK_V1TI): Likewise.
2146 (UNPACK_V1TI): Likewise.
2148 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
2149 support for decimal floating point builtin functions.
2150 (rs6000_expand_ternop_builtin): Add checks for the new builtin
2151 functions that take constant arguments.
2152 (rs6000_invalid_builtin): Add decimal floating point builtin support.
2153 (rs6000_init_builtins): Setup long double, _Decimal64, and
2154 _Decimal128 types for new builtin functions.
2155 (builtin_function_type): Set the unsigned flags appropriately for
2156 the new builtin functions.
2157 (rs6000_opt_masks): Add support for decimal floating point builtin
2160 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
2161 floating point builtin functions.
2162 (RS6000_BTM_COMMON): Likewise.
2163 (RS6000_BTI_long_double): Likewise.
2164 (RS6000_BTI_dfloat64): Likewise.
2165 (RS6000_BTI_dfloat128): Likewise.
2166 (long_double_type_internal_node): Likewise.
2167 (dfloat64_type_internal_node): Likewise.
2168 (dfloat128_type_internal_node): Likewise.
2170 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
2171 2.07 bcd arithmetic instructions.
2172 (UNSPEC_BCDSUB): Likewise.
2173 (UNSPEC_BCD_OVERFLOW): Likewise.
2174 (UNSPEC_BCD_ADD_SUB): Likewise.
2175 (bcd_add_sub): Likewise.
2176 (BCD_TEST): Likewise.
2177 (bcd<bcd_add_sub>): Likewise.
2178 (bcd<bcd_add_sub>_test): Likewise.
2179 (bcd<bcd_add_sub>_test2): Likewise.
2180 (bcd<bcd_add_sub>_<code>): Likewise.
2181 (peephole2 for combined bcd ops): Likewise.
2183 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
2184 decimal floating point builtin functions.
2185 (UNSPEC_DENBCD): Likewise.
2186 (UNSPEC_DXEX): Likewise.
2187 (UNSPEC_DIEX): Likewise.
2188 (UNSPEC_DSCLI): Likewise.
2189 (UNSPEC_DSCRI): Likewise.
2190 (D64_D128): Likewise.
2191 (dfp_suffix): Likewise.
2192 (dfp_ddedpd_<mode>): Likewise.
2193 (dfp_denbcd_<mode>): Likewise.
2194 (dfp_dxex_<mode>): Likewise.
2195 (dfp_diex_<mode>): Likewise.
2196 (dfp_dscli_<mode>): Likewise.
2197 (dfp_dscri_<mode>): Likewise.
2199 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
2201 (UNSPEC_CDTBCD): Likewise.
2202 (UNSPEC_CBCDTD): Likewise.
2203 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
2204 (UNSPEC_DIVEO): Likewise.
2205 (UNSPEC_DIVEU): Likewise.
2206 (UNSPEC_DIVEUO): Likewise.
2207 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
2208 pack/unpack 128-bit types.
2209 (UNSPEC_PACK_128BIT): Likewise.
2210 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
2211 (udiv<mode>3): Use idiv_ldiv mode attribute.
2212 (div<mode>3): Likewise.
2213 (addg6s): Add new BCD builtin functions.
2216 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
2217 (div_extend): Likewise.
2218 (div<div_extend>_<mode>"): Likewise.
2219 (FP128_64): Add support for new builtin functions to pack/unpack
2221 (unpack<mode>): Likewise.
2222 (unpacktf_0): Likewise.
2223 (unpacktf_1): Likewise.
2224 (unpack<mode>_dm): Likewise.
2225 (unpack<mode>_nodm): Likewise.
2226 (pack<mode>): Likewise.
2227 (unpackv1ti): Likewise.
2228 (packv1ti): Likewise.
2230 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
2232 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
2235 2014-04-24 Jakub Jelinek <jakub@redhat.com>
2237 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
2238 * gimplify.c (omp_is_private): Change last argument's type to int.
2239 Only diagnose lastprivate if the simd argument is 1, only diagnose
2240 linear if the simd argument is 2.
2241 (gimplify_omp_for): Adjust omp_is_private callers. When adding
2242 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
2243 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
2244 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
2245 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
2246 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2247 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
2248 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2249 * tree-nested.c (convert_nonlocal_omp_clauses,
2250 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
2252 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
2255 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
2258 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
2260 * flag-types.h (enum ivar_visibility): Add.
2262 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
2264 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
2265 function * argument.
2267 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
2269 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
2271 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2272 Tom de Vries <tom@codesourcery.com>
2274 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
2275 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
2277 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
2278 * emit-rtl.c (try_split): Same.
2280 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2281 Tom de Vries <tom@codesourcery.com>
2283 * common.opt (fuse-caller-save): New option.
2285 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
2287 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
2288 elements for big-endian.
2290 2014-04-24 Richard Biener <rguenther@suse.de>
2292 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
2293 during TER and instead use the sepops interface for expanding
2294 non-GIMPLE_SINGLE_RHS.
2296 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2298 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
2299 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
2301 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2303 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
2304 assembler 64-bit option.
2305 * configure: Regenerate.
2307 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2309 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
2310 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
2311 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
2312 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
2313 (TARGET_CRYPTO): Take TARGET_SIMD into account.
2315 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2317 * config/aarch64/aarch64-builtins.c
2318 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
2319 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
2320 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
2321 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
2323 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
2324 (Vrevsuff): New mode attribute.
2326 2014-04-24 Terry Guo <terry.guo@arm.com>
2328 * config/arm/arm.h (machine_function): Define variable
2329 after_arm_reorg here.
2330 * config/arm/arm.c (after_arm_reorg): Remove the definition.
2331 (arm_split_constant): Update the way to access variable
2334 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
2336 2014-04-23 Tom de Vries <tom@codesourcery.com>
2338 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
2340 2014-04-23 David Malcolm <dmalcolm@redhat.com>
2342 * is-a.h: Update comments to reflect the following changes to the
2343 "pointerness" of the API, making the template parameter match the
2344 return type, allowing use of is-a.h with typedefs of pointers.
2345 (is_a_helper::cast): Return a T rather then a pointer to a T, so
2346 that the return type matches the parameter to the is_a_helper.
2348 (dyn_cast): Likewise.
2350 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
2351 pointer from the is-a.h API.
2353 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
2354 (is_a_helper <cgraph_node *>::test): ...this, matching change to
2356 (is_a_helper <varpool_node>::test): Likewise, convert to...
2357 (is_a_helper <varpool_node *>::test): ...this.
2359 (varpool_first_variable): Update for removal of implicit pointer
2360 from the is-a.h API.
2361 (varpool_next_variable): Likewise.
2362 (varpool_first_static_initializer): Likewise.
2363 (varpool_next_static_initializer): Likewise.
2364 (varpool_first_defined_variable): Likewise.
2365 (varpool_next_defined_variable): Likewise.
2366 (cgraph_first_defined_function): Likewise.
2367 (cgraph_next_defined_function): Likewise.
2368 (cgraph_first_function): Likewise.
2369 (cgraph_next_function): Likewise.
2370 (cgraph_first_function_with_gimple_body): Likewise.
2371 (cgraph_next_function_with_gimple_body): Likewise.
2372 (cgraph_alias_target): Likewise.
2373 (varpool_alias_target): Likewise.
2374 (cgraph_function_or_thunk_node): Likewise.
2375 (varpool_variable_node): Likewise.
2376 (symtab_real_symbol_p): Likewise.
2377 * cgraphunit.c (referred_to_p): Likewise.
2378 (analyze_functions): Likewise.
2379 (handle_alias_pairs): Likewise.
2380 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2381 * gimple-ssa.h (gimple_vuse_op): Likewise.
2382 (gimple_vdef_op): Likewise.
2383 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
2384 * gimple.c (gimple_build_asm_1): Likewise.
2385 (gimple_build_try): Likewise.
2386 (gimple_build_resx): Likewise.
2387 (gimple_build_eh_dispatch): Likewise.
2388 (gimple_build_omp_for): Likewise.
2389 (gimple_omp_for_set_clauses): Likewise.
2391 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
2392 (is_a_helper <gimple_statement_asm *>::test): ...this.
2393 (is_a_helper <gimple_statement_bind>::test): Convert to...
2394 (is_a_helper <gimple_statement_bind *>::test): ...this.
2395 (is_a_helper <gimple_statement_call>::test): Convert to...
2396 (is_a_helper <gimple_statement_call *>::test): ...this.
2397 (is_a_helper <gimple_statement_catch>::test): Convert to...
2398 (is_a_helper <gimple_statement_catch *>::test): ...this.
2399 (is_a_helper <gimple_statement_resx>::test): Convert to...
2400 (is_a_helper <gimple_statement_resx *>::test): ...this.
2401 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
2402 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
2403 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
2404 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
2405 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
2406 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
2407 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
2408 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
2409 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
2410 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
2411 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
2412 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
2413 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
2414 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
2415 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
2416 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
2417 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
2418 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
2419 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
2420 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
2421 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
2422 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
2423 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
2424 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
2425 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
2426 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
2427 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
2428 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
2429 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
2430 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
2431 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
2432 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
2433 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
2434 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
2435 (is_a_helper <gimple_statement_phi>::test): Convert to...
2436 (is_a_helper <gimple_statement_phi *>::test): ...this.
2437 (is_a_helper <gimple_statement_transaction>::test): Convert to...
2438 (is_a_helper <gimple_statement_transaction *>::test): ...this.
2439 (is_a_helper <gimple_statement_try>::test): Convert to...
2440 (is_a_helper <gimple_statement_try *>::test): ...this.
2441 (is_a_helper <gimple_statement_wce>::test): Convert to...
2442 (is_a_helper <gimple_statement_wce *>::test): ...this.
2443 (is_a_helper <const gimple_statement_asm>::test): Convert to...
2444 (is_a_helper <const gimple_statement_asm *>::test): ...this.
2445 (is_a_helper <const gimple_statement_bind>::test): Convert to...
2446 (is_a_helper <const gimple_statement_bind *>::test): ...this.
2447 (is_a_helper <const gimple_statement_call>::test): Convert to...
2448 (is_a_helper <const gimple_statement_call *>::test): ...this.
2449 (is_a_helper <const gimple_statement_catch>::test): Convert to...
2450 (is_a_helper <const gimple_statement_catch *>::test): ...this.
2451 (is_a_helper <const gimple_statement_resx>::test): Convert to...
2452 (is_a_helper <const gimple_statement_resx *>::test): ...this.
2453 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
2454 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
2455 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
2456 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
2457 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
2459 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
2461 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
2463 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
2465 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
2466 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
2467 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
2469 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
2470 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
2472 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
2473 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
2474 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
2475 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
2476 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
2477 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
2479 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
2480 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
2481 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
2482 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
2484 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
2485 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
2486 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
2487 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
2488 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
2489 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
2490 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
2491 (is_a_helper <const gimple_statement_phi>::test): Convert to...
2492 (is_a_helper <const gimple_statement_phi *>::test): ...this.
2493 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
2494 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
2495 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
2496 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
2497 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
2498 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
2499 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
2501 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
2503 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
2504 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
2506 (gimple_use_ops): Update for removal of implicit pointer from the
2508 (gimple_set_use_ops): Likewise.
2509 (gimple_vuse): Likewise.
2510 (gimple_vdef): Likewise.
2511 (gimple_vuse_ptr): Likewise.
2512 (gimple_vdef_ptr): Likewise.
2513 (gimple_set_vuse): Likewise.
2514 (gimple_set_vdef): Likewise.
2515 (gimple_omp_return_set_lhs): Likewise.
2516 (gimple_omp_return_lhs): Likewise.
2517 (gimple_omp_return_lhs_ptr): Likewise.
2518 (gimple_call_fntype): Likewise.
2519 (gimple_call_set_fntype): Likewise.
2520 (gimple_call_set_internal_fn): Likewise.
2521 (gimple_call_use_set): Likewise.
2522 (gimple_call_clobber_set): Likewise.
2523 (gimple_bind_vars): Likewise.
2524 (gimple_bind_set_vars): Likewise.
2525 (gimple_bind_body_ptr): Likewise.
2526 (gimple_bind_set_body): Likewise.
2527 (gimple_bind_add_stmt): Likewise.
2528 (gimple_bind_block): Likewise.
2529 (gimple_bind_set_block): Likewise.
2530 (gimple_asm_ninputs): Likewise.
2531 (gimple_asm_noutputs): Likewise.
2532 (gimple_asm_nclobbers): Likewise.
2533 (gimple_asm_nlabels): Likewise.
2534 (gimple_asm_input_op): Likewise.
2535 (gimple_asm_input_op_ptr): Likewise.
2536 (gimple_asm_output_op): Likewise.
2537 (gimple_asm_output_op_ptr): Likewise.
2538 (gimple_asm_set_output_op): Likewise.
2539 (gimple_asm_clobber_op): Likewise.
2540 (gimple_asm_set_clobber_op): Likewise.
2541 (gimple_asm_label_op): Likewise.
2542 (gimple_asm_set_label_op): Likewise.
2543 (gimple_asm_string): Likewise.
2544 (gimple_catch_types): Likewise.
2545 (gimple_catch_types_ptr): Likewise.
2546 (gimple_catch_handler_ptr): Likewise.
2547 (gimple_catch_set_types): Likewise.
2548 (gimple_catch_set_handler): Likewise.
2549 (gimple_eh_filter_types): Likewise.
2550 (gimple_eh_filter_types_ptr): Likewise.
2551 (gimple_eh_filter_failure_ptr): Likewise.
2552 (gimple_eh_filter_set_types): Likewise.
2553 (gimple_eh_filter_set_failure): Likewise.
2554 (gimple_eh_must_not_throw_fndecl): Likewise.
2555 (gimple_eh_must_not_throw_set_fndecl): Likewise.
2556 (gimple_eh_else_n_body_ptr): Likewise.
2557 (gimple_eh_else_e_body_ptr): Likewise.
2558 (gimple_eh_else_set_n_body): Likewise.
2559 (gimple_eh_else_set_e_body): Likewise.
2560 (gimple_try_eval_ptr): Likewise.
2561 (gimple_try_cleanup_ptr): Likewise.
2562 (gimple_try_set_eval): Likewise.
2563 (gimple_try_set_cleanup): Likewise.
2564 (gimple_wce_cleanup_ptr): Likewise.
2565 (gimple_wce_set_cleanup): Likewise.
2566 (gimple_phi_capacity): Likewise.
2567 (gimple_phi_num_args): Likewise.
2568 (gimple_phi_result): Likewise.
2569 (gimple_phi_result_ptr): Likewise.
2570 (gimple_phi_set_result): Likewise.
2571 (gimple_phi_arg): Likewise.
2572 (gimple_phi_set_arg): Likewise.
2573 (gimple_resx_region): Likewise.
2574 (gimple_resx_set_region): Likewise.
2575 (gimple_eh_dispatch_region): Likewise.
2576 (gimple_eh_dispatch_set_region): Likewise.
2577 (gimple_omp_critical_name): Likewise.
2578 (gimple_omp_critical_name_ptr): Likewise.
2579 (gimple_omp_critical_set_name): Likewise.
2580 (gimple_omp_for_clauses): Likewise.
2581 (gimple_omp_for_clauses_ptr): Likewise.
2582 (gimple_omp_for_set_clauses): Likewise.
2583 (gimple_omp_for_collapse): Likewise.
2584 (gimple_omp_for_index): Likewise.
2585 (gimple_omp_for_index_ptr): Likewise.
2586 (gimple_omp_for_set_index): Likewise.
2587 (gimple_omp_for_initial): Likewise.
2588 (gimple_omp_for_initial_ptr): Likewise.
2589 (gimple_omp_for_set_initial): Likewise.
2590 (gimple_omp_for_final): Likewise.
2591 (gimple_omp_for_final_ptr): Likewise.
2592 (gimple_omp_for_set_final): Likewise.
2593 (gimple_omp_for_incr): Likewise.
2594 (gimple_omp_for_incr_ptr): Likewise.
2595 (gimple_omp_for_set_incr): Likewise.
2596 (gimple_omp_for_pre_body_ptr): Likewise.
2597 (gimple_omp_for_set_pre_body): Likewise.
2598 (gimple_omp_parallel_clauses): Likewise.
2599 (gimple_omp_parallel_clauses_ptr): Likewise.
2600 (gimple_omp_parallel_set_clauses): Likewise.
2601 (gimple_omp_parallel_child_fn): Likewise.
2602 (gimple_omp_parallel_child_fn_ptr): Likewise.
2603 (gimple_omp_parallel_set_child_fn): Likewise.
2604 (gimple_omp_parallel_data_arg): Likewise.
2605 (gimple_omp_parallel_data_arg_ptr): Likewise.
2606 (gimple_omp_parallel_set_data_arg): Likewise.
2607 (gimple_omp_task_clauses): Likewise.
2608 (gimple_omp_task_clauses_ptr): Likewise.
2609 (gimple_omp_task_set_clauses): Likewise.
2610 (gimple_omp_task_child_fn): Likewise.
2611 (gimple_omp_task_child_fn_ptr): Likewise.
2612 (gimple_omp_task_set_child_fn): Likewise.
2613 (gimple_omp_task_data_arg): Likewise.
2614 (gimple_omp_task_data_arg_ptr): Likewise.
2615 (gimple_omp_task_set_data_arg): Likewise.
2616 (gimple_omp_taskreg_clauses): Likewise.
2617 (gimple_omp_taskreg_clauses_ptr): Likewise.
2618 (gimple_omp_taskreg_set_clauses): Likewise.
2619 (gimple_omp_taskreg_child_fn): Likewise.
2620 (gimple_omp_taskreg_child_fn_ptr): Likewise.
2621 (gimple_omp_taskreg_set_child_fn): Likewise.
2622 (gimple_omp_taskreg_data_arg): Likewise.
2623 (gimple_omp_taskreg_data_arg_ptr): Likewise.
2624 (gimple_omp_taskreg_set_data_arg): Likewise.
2625 (gimple_omp_task_copy_fn): Likewise.
2626 (gimple_omp_task_copy_fn_ptr): Likewise.
2627 (gimple_omp_task_set_copy_fn): Likewise.
2628 (gimple_omp_task_arg_size): Likewise.
2629 (gimple_omp_task_arg_size_ptr): Likewise.
2630 (gimple_omp_task_set_arg_size): Likewise.
2631 (gimple_omp_task_arg_align): Likewise.
2632 (gimple_omp_task_arg_align_ptr): Likewise.
2633 (gimple_omp_task_set_arg_align): Likewise.
2634 (gimple_omp_single_clauses): Likewise.
2635 (gimple_omp_single_clauses_ptr): Likewise.
2636 (gimple_omp_single_set_clauses): Likewise.
2637 (gimple_omp_target_clauses): Likewise.
2638 (gimple_omp_target_clauses_ptr): Likewise.
2639 (gimple_omp_target_set_clauses): Likewise.
2640 (gimple_omp_target_child_fn): Likewise.
2641 (gimple_omp_target_child_fn_ptr): Likewise.
2642 (gimple_omp_target_set_child_fn): Likewise.
2643 (gimple_omp_target_data_arg): Likewise.
2644 (gimple_omp_target_data_arg_ptr): Likewise.
2645 (gimple_omp_target_set_data_arg): Likewise.
2646 (gimple_omp_teams_clauses): Likewise.
2647 (gimple_omp_teams_clauses_ptr): Likewise.
2648 (gimple_omp_teams_set_clauses): Likewise.
2649 (gimple_omp_sections_clauses): Likewise.
2650 (gimple_omp_sections_clauses_ptr): Likewise.
2651 (gimple_omp_sections_set_clauses): Likewise.
2652 (gimple_omp_sections_control): Likewise.
2653 (gimple_omp_sections_control_ptr): Likewise.
2654 (gimple_omp_sections_set_control): Likewise.
2655 (gimple_omp_for_set_cond): Likewise.
2656 (gimple_omp_for_cond): Likewise.
2657 (gimple_omp_atomic_store_set_val): Likewise.
2658 (gimple_omp_atomic_store_val): Likewise.
2659 (gimple_omp_atomic_store_val_ptr): Likewise.
2660 (gimple_omp_atomic_load_set_lhs): Likewise.
2661 (gimple_omp_atomic_load_lhs): Likewise.
2662 (gimple_omp_atomic_load_lhs_ptr): Likewise.
2663 (gimple_omp_atomic_load_set_rhs): Likewise.
2664 (gimple_omp_atomic_load_rhs): Likewise.
2665 (gimple_omp_atomic_load_rhs_ptr): Likewise.
2666 (gimple_omp_continue_control_def): Likewise.
2667 (gimple_omp_continue_control_def_ptr): Likewise.
2668 (gimple_omp_continue_set_control_def): Likewise.
2669 (gimple_omp_continue_control_use): Likewise.
2670 (gimple_omp_continue_control_use_ptr): Likewise.
2671 (gimple_omp_continue_set_control_use): Likewise.
2672 (gimple_transaction_body_ptr): Likewise.
2673 (gimple_transaction_label): Likewise.
2674 (gimple_transaction_label_ptr): Likewise.
2675 (gimple_transaction_set_body): Likewise.
2676 (gimple_transaction_set_label): Likewise.
2678 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
2679 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2680 * ipa-ref.c (ipa_record_reference): Likewise.
2681 * ipa-reference.c (analyze_function): Likewise.
2682 (ipa_reference_write_optimization_summary): Likewise.
2683 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
2684 (address_taken_from_non_vtable_p): Likewise.
2685 (comdat_can_be_unshared_p_1): Likewise.
2686 * lto-cgraph.c (lto_output_ref): Likewise.
2687 (add_references): Likewise.
2688 (compute_ltrans_boundary): Likewise.
2689 (output_symtab): Likewise.
2690 (input_ref): Likewise.
2691 (input_cgraph_1): Likewise.
2692 (output_cgraph_opt_summary): Likewise.
2693 * lto-streamer-out.c (lto_output): Likewise.
2694 (output_symbol_p): Likewise.
2695 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
2696 (lsei_start_function_in_partition): Likewise.
2697 (lsei_next_variable_in_partition): Likewise.
2698 (lsei_start_variable_in_partition): Likewise.
2699 * symtab.c (insert_to_assembler_name_hash): Likewise.
2700 (unlink_from_assembler_name_hash): Likewise.
2701 (symtab_unregister_node): Likewise.
2702 (symtab_remove_node): Likewise.
2703 (dump_symtab_node): Likewise.
2704 (verify_symtab_base): Likewise.
2705 (verify_symtab_node): Likewise.
2706 (symtab_make_decl_local): Likewise.
2707 (symtab_alias_ultimate_target): Likewise.
2708 (symtab_resolve_alias): Likewise.
2709 (symtab_get_symbol_partitioning_class): Likewise.
2710 * tree-phinodes.c (allocate_phi_node): Likewise.
2711 (reserve_phi_args_for_new_edge): Likewise.
2712 (remove_phi_args): Likewise.
2713 * varpool.c (varpool_node_for_asm): Likewise.
2714 (varpool_remove_unreferenced_decls): Likewise.
2716 2014-04-23 Jeff Law <law@redhat.com>
2718 PR tree-optimization/60902
2719 * tree-ssa-threadedge.c
2720 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
2721 invalidate outputs from statements that do not produce useful
2722 outputs for threading.
2724 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2726 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
2727 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
2728 machine descriptions for Stack Smashing Protector.
2730 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
2732 * aarch64.md (<optab>_rol<mode>3): New pattern.
2733 (<optab>_rolsi3_uxtw): Likewise.
2734 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
2736 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
2738 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
2739 (arm_cortex_a12_tune): Likewise.
2741 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2743 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
2745 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2747 * config/arm/arm.md (arm_rev16si2): New pattern.
2748 (arm_rev16si2_alt): Likewise.
2749 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
2751 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2753 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
2754 (rev16<mode>2_alt): Likewise.
2755 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
2756 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
2757 (aarch_rev16_shleft_mask_imm_p): Likewise.
2758 (aarch_rev16_p_1): Likewise.
2759 (aarch_rev16_p): Likewise.
2760 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
2761 (aarch_rev16_shright_mask_imm_p): Likewise.
2762 (aarch_rev16_shleft_mask_imm_p): Likewise.
2764 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2766 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
2767 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
2769 (cortex_a53_extra_costs): Likewise.
2770 (cortex_a57_extra_costs): Likewise.
2771 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
2772 (cortexa7_extra_costs): Likewise.
2773 (cortexa8_extra_costs): Likewise.
2774 (cortexa12_extra_costs): Likewise.
2775 (cortexa15_extra_costs): Likewise.
2776 (v7m_extra_costs): Likewise.
2777 (arm_new_rtx_costs): Handle BSWAP.
2779 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2781 * config/arm/arm.c (cortexa8_extra_costs): New table.
2782 (arm_cortex_a8_tune): New tuning struct.
2783 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
2785 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2787 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
2789 2014-04-23 Richard Biener <rguenther@suse.de>
2791 * Makefile.in (OBJS): Remove loop-unswitch.o.
2792 * tree-pass.h (make_pass_rtl_unswitch): Remove.
2793 * passes.def (pass_rtl_unswitch): Likewise.
2794 * loop-init.c (gate_rtl_unswitch): Likewise.
2795 (rtl_unswitch): Likewise.
2796 (pass_data_rtl_unswitch): Likewise.
2797 (pass_rtl_unswitch): Likewise.
2798 (make_pass_rtl_unswitch): Likewise.
2799 * rtl.h (reversed_condition): Likewise.
2800 (compare_and_jump_seq): Likewise.
2801 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
2803 * loop-unroll.c (compare_and_jump_seq): Likewise.
2805 2014-04-23 Richard Biener <rguenther@suse.de>
2807 PR tree-optimization/60903
2808 * tree-ssa-loop-im.c (analyze_memory_references): Remove
2809 commented code block.
2810 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
2811 loop flags to newly created BBs and edges.
2813 2014-04-23 Nick Clifton <nickc@redhat.com>
2815 * config/msp430/msp430.c (msp430_handle_option): Move function
2817 (msp430_option_override): Simplify mcu and mcpu option handling.
2818 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
2819 support for -mhwmult command line option.
2820 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
2821 -mhwmult command line option.
2822 (msp430_hwmult_enabled): Delete.
2823 (msp43o_output_labelref): Add support for -mhwmult command line option.
2824 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
2825 (umulsidi3): Likewise.
2826 * config/msp430/msp430.opt (mmcu): Add Report attribute.
2827 (mcpu, mlarge, msmall): Likewise.
2828 (mhwmult): New option.
2829 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
2831 (msp430_is_f5_mcu): Remove prototype.
2832 (msp430_use_f5_series_hwmult): Add prototype.
2833 * config/msp430/msp430-opts.h: New file.
2834 * common/config/msp430: New directory.
2835 * common/config/msp430/msp430-common.c: New file.
2836 * config.gcc (msp430): Remove target_has_targetm_common.
2837 * doc/invoke.texi: Document -mhwmult command line option.
2839 2014-04-23 Nick Clifton <nickc@redhat.com>
2841 * config/i386/cygwin.h (ENDFILE_SPEC): Include
2842 default-manifest.o if it can be found in the search path.
2843 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
2845 2014-04-23 Terry Guo <terry.guo@arm.com>
2847 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
2849 2014-04-23 Richard Biener <rguenther@suse.de>
2852 * tree-inline.c (declare_return_variable): Use mark_addressable.
2854 2014-04-23 Richard Biener <rguenther@suse.de>
2857 * loop-init.c (loop_optimizer_init): Make sure to apply
2858 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
2860 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2863 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
2865 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
2866 if flag_sanitize_undefined_trap_on_error.
2867 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
2868 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
2869 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
2870 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
2871 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
2872 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
2873 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
2874 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
2875 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
2876 * ubsan.c (ubsan_instrument_unreachable): Return
2877 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
2878 (ubsan_expand_null_ifn): Emit __builtin_trap ()
2879 if flag_sanitize_undefined_trap_on_error and
2880 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
2881 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
2882 instrument_bool_enum_load): Emit __builtin_trap () if
2883 flag_sanitize_undefined_trap_on_error and
2884 __builtin_handle_*_abort () if !flag_sanitize_recover.
2885 * doc/invoke.texi (-fsanitize-recover,
2886 -fsanitize-undefined-trap-on-error): Document.
2888 2014-04-22 Christian Bruel <christian.bruel@st.com>
2890 * config/sh/sh.md (mov<mode>): Replace movQIHI.
2891 Force immediates to SImode.
2893 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
2895 * config/nios2/nios2.md (UNSPEC_ROUND): New.
2897 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
2898 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
2899 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
2900 (nios2_fpu_insn): Add entry for round.
2901 (N2FPU_NO_ERRNO_P): Define.
2902 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
2904 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
2906 2014-04-22 Richard Henderson <rth@redhat.com>
2908 * config/aarch64/aarch64 (addti3, subti3): New expanders.
2909 (add<GPI>3_compare0): Remove leading * from name.
2910 (add<GPI>3_carryin): Likewise.
2911 (sub<GPI>3_compare0): Likewise.
2912 (sub<GPI>3_carryin): Likewise.
2913 (<su_optab>mulditi3): New expander.
2914 (multi3): New expander.
2915 (madd<GPI>): Remove leading * from name.
2917 2014-04-22 Martin Jambor <mjambor@suse.cz>
2919 * cgraphclones.c (cgraph_function_versioning): Copy
2920 ipa_transforms_to_apply instead of asserting it is empty.
2922 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
2925 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
2926 on count_exp to get mode.
2928 2014-04-22 Andrew Pinski <apinski@cavium.com>
2930 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
2931 Handle TLS for ILP32.
2932 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
2933 (tlsie_small_<mode>): this and handle PTR.
2934 (tlsie_small_sidi): New pattern.
2935 (tlsle_small): Change to an expand to handle ILP32.
2936 (tlsle_small_<mode>): New pattern.
2937 (tlsdesc_small): Rename to ...
2938 (tlsdesc_small_<mode>): this and handle PTR.
2940 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2942 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
2944 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2946 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
2947 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
2948 (aarch64_types_signed_poly_qualifiers): Likewise.
2949 (aarch64_types_unsigned_signed_qualifiers): Likewise.
2950 (aarch64_types_poly_signed_qualifiers): Likewise.
2951 (TYPES_REINTERP_SS): Type macro added.
2952 (TYPES_REINTERP_SU): Likewise.
2953 (TYPES_REINTERP_SP): Likewise.
2954 (TYPES_REINTERP_US): Likewise.
2955 (TYPES_REINTERP_PS): Likewise.
2956 (aarch64_fold_builtin): New expression folding added.
2957 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
2958 Declarations removed.
2959 (REINTERP_SS): Declarations added.
2960 (REINTERP_US): Likewise.
2961 (REINTERP_PS): Likewise.
2962 (REINTERP_SU): Likewise.
2963 (REINTERP_SP): Likewise.
2964 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
2965 (vreinterpretq_p8_f64): Likewise.
2966 (vreinterpret_p16_f64): Likewise.
2967 (vreinterpretq_p16_f64): Likewise.
2968 (vreinterpret_f32_f64): Likewise.
2969 (vreinterpretq_f32_f64): Likewise.
2970 (vreinterpret_f64_f32): Likewise.
2971 (vreinterpret_f64_p8): Likewise.
2972 (vreinterpret_f64_p16): Likewise.
2973 (vreinterpret_f64_s8): Likewise.
2974 (vreinterpret_f64_s16): Likewise.
2975 (vreinterpret_f64_s32): Likewise.
2976 (vreinterpret_f64_s64): Likewise.
2977 (vreinterpret_f64_u8): Likewise.
2978 (vreinterpret_f64_u16): Likewise.
2979 (vreinterpret_f64_u32): Likewise.
2980 (vreinterpret_f64_u64): Likewise.
2981 (vreinterpretq_f64_f32): Likewise.
2982 (vreinterpretq_f64_p8): Likewise.
2983 (vreinterpretq_f64_p16): Likewise.
2984 (vreinterpretq_f64_s8): Likewise.
2985 (vreinterpretq_f64_s16): Likewise.
2986 (vreinterpretq_f64_s32): Likewise.
2987 (vreinterpretq_f64_s64): Likewise.
2988 (vreinterpretq_f64_u8): Likewise.
2989 (vreinterpretq_f64_u16): Likewise.
2990 (vreinterpretq_f64_u32): Likewise.
2991 (vreinterpretq_f64_u64): Likewise.
2992 (vreinterpret_s64_f64): Likewise.
2993 (vreinterpretq_s64_f64): Likewise.
2994 (vreinterpret_u64_f64): Likewise.
2995 (vreinterpretq_u64_f64): Likewise.
2996 (vreinterpret_s8_f64): Likewise.
2997 (vreinterpretq_s8_f64): Likewise.
2998 (vreinterpret_s16_f64): Likewise.
2999 (vreinterpretq_s16_f64): Likewise.
3000 (vreinterpret_s32_f64): Likewise.
3001 (vreinterpretq_s32_f64): Likewise.
3002 (vreinterpret_u8_f64): Likewise.
3003 (vreinterpretq_u8_f64): Likewise.
3004 (vreinterpret_u16_f64): Likewise.
3005 (vreinterpretq_u16_f64): Likewise.
3006 (vreinterpret_u32_f64): Likewise.
3007 (vreinterpretq_u32_f64): Likewise.
3009 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3011 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
3012 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
3013 (vreinterpret_p8_s8): Likewise.
3014 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
3015 (vreinterpret_p8_s16): Likewise.
3016 (vreinterpret_p8_s32): Likewise.
3017 (vreinterpret_p8_s64): Likewise.
3018 (vreinterpret_p8_f32): Likewise.
3019 (vreinterpret_p8_u8): Likewise.
3020 (vreinterpret_p8_u16): Likewise.
3021 (vreinterpret_p8_u32): Likewise.
3022 (vreinterpret_p8_u64): Likewise.
3023 (vreinterpret_p8_p16): Likewise.
3024 (vreinterpretq_p8_s8): Likewise.
3025 (vreinterpretq_p8_s16): Likewise.
3026 (vreinterpretq_p8_s32): Likewise.
3027 (vreinterpretq_p8_s64): Likewise.
3028 (vreinterpretq_p8_f32): Likewise.
3029 (vreinterpretq_p8_u8): Likewise.
3030 (vreinterpretq_p8_u16): Likewise.
3031 (vreinterpretq_p8_u32): Likewise.
3032 (vreinterpretq_p8_u64): Likewise.
3033 (vreinterpretq_p8_p16): Likewise.
3034 (vreinterpret_p16_s8): Likewise.
3035 (vreinterpret_p16_s16): Likewise.
3036 (vreinterpret_p16_s32): Likewise.
3037 (vreinterpret_p16_s64): Likewise.
3038 (vreinterpret_p16_f32): Likewise.
3039 (vreinterpret_p16_u8): Likewise.
3040 (vreinterpret_p16_u16): Likewise.
3041 (vreinterpret_p16_u32): Likewise.
3042 (vreinterpret_p16_u64): Likewise.
3043 (vreinterpret_p16_p8): Likewise.
3044 (vreinterpretq_p16_s8): Likewise.
3045 (vreinterpretq_p16_s16): Likewise.
3046 (vreinterpretq_p16_s32): Likewise.
3047 (vreinterpretq_p16_s64): Likewise.
3048 (vreinterpretq_p16_f32): Likewise.
3049 (vreinterpretq_p16_u8): Likewise.
3050 (vreinterpretq_p16_u16): Likewise.
3051 (vreinterpretq_p16_u32): Likewise.
3052 (vreinterpretq_p16_u64): Likewise.
3053 (vreinterpretq_p16_p8): Likewise.
3054 (vreinterpret_f32_s8): Likewise.
3055 (vreinterpret_f32_s16): Likewise.
3056 (vreinterpret_f32_s32): Likewise.
3057 (vreinterpret_f32_s64): Likewise.
3058 (vreinterpret_f32_u8): Likewise.
3059 (vreinterpret_f32_u16): Likewise.
3060 (vreinterpret_f32_u32): Likewise.
3061 (vreinterpret_f32_u64): Likewise.
3062 (vreinterpret_f32_p8): Likewise.
3063 (vreinterpret_f32_p16): Likewise.
3064 (vreinterpretq_f32_s8): Likewise.
3065 (vreinterpretq_f32_s16): Likewise.
3066 (vreinterpretq_f32_s32): Likewise.
3067 (vreinterpretq_f32_s64): Likewise.
3068 (vreinterpretq_f32_u8): Likewise.
3069 (vreinterpretq_f32_u16): Likewise.
3070 (vreinterpretq_f32_u32): Likewise.
3071 (vreinterpretq_f32_u64): Likewise.
3072 (vreinterpretq_f32_p8): Likewise.
3073 (vreinterpretq_f32_p16): Likewise.
3074 (vreinterpret_s64_s8): Likewise.
3075 (vreinterpret_s64_s16): Likewise.
3076 (vreinterpret_s64_s32): Likewise.
3077 (vreinterpret_s64_f32): Likewise.
3078 (vreinterpret_s64_u8): Likewise.
3079 (vreinterpret_s64_u16): Likewise.
3080 (vreinterpret_s64_u32): Likewise.
3081 (vreinterpret_s64_u64): Likewise.
3082 (vreinterpret_s64_p8): Likewise.
3083 (vreinterpret_s64_p16): Likewise.
3084 (vreinterpretq_s64_s8): Likewise.
3085 (vreinterpretq_s64_s16): Likewise.
3086 (vreinterpretq_s64_s32): Likewise.
3087 (vreinterpretq_s64_f32): Likewise.
3088 (vreinterpretq_s64_u8): Likewise.
3089 (vreinterpretq_s64_u16): Likewise.
3090 (vreinterpretq_s64_u32): Likewise.
3091 (vreinterpretq_s64_u64): Likewise.
3092 (vreinterpretq_s64_p8): Likewise.
3093 (vreinterpretq_s64_p16): Likewise.
3094 (vreinterpret_u64_s8): Likewise.
3095 (vreinterpret_u64_s16): Likewise.
3096 (vreinterpret_u64_s32): Likewise.
3097 (vreinterpret_u64_s64): Likewise.
3098 (vreinterpret_u64_f32): Likewise.
3099 (vreinterpret_u64_u8): Likewise.
3100 (vreinterpret_u64_u16): Likewise.
3101 (vreinterpret_u64_u32): Likewise.
3102 (vreinterpret_u64_p8): Likewise.
3103 (vreinterpret_u64_p16): Likewise.
3104 (vreinterpretq_u64_s8): Likewise.
3105 (vreinterpretq_u64_s16): Likewise.
3106 (vreinterpretq_u64_s32): Likewise.
3107 (vreinterpretq_u64_s64): Likewise.
3108 (vreinterpretq_u64_f32): Likewise.
3109 (vreinterpretq_u64_u8): Likewise.
3110 (vreinterpretq_u64_u16): Likewise.
3111 (vreinterpretq_u64_u32): Likewise.
3112 (vreinterpretq_u64_p8): Likewise.
3113 (vreinterpretq_u64_p16): Likewise.
3114 (vreinterpret_s8_s16): Likewise.
3115 (vreinterpret_s8_s32): Likewise.
3116 (vreinterpret_s8_s64): Likewise.
3117 (vreinterpret_s8_f32): Likewise.
3118 (vreinterpret_s8_u8): Likewise.
3119 (vreinterpret_s8_u16): Likewise.
3120 (vreinterpret_s8_u32): Likewise.
3121 (vreinterpret_s8_u64): Likewise.
3122 (vreinterpret_s8_p8): Likewise.
3123 (vreinterpret_s8_p16): Likewise.
3124 (vreinterpretq_s8_s16): Likewise.
3125 (vreinterpretq_s8_s32): Likewise.
3126 (vreinterpretq_s8_s64): Likewise.
3127 (vreinterpretq_s8_f32): Likewise.
3128 (vreinterpretq_s8_u8): Likewise.
3129 (vreinterpretq_s8_u16): Likewise.
3130 (vreinterpretq_s8_u32): Likewise.
3131 (vreinterpretq_s8_u64): Likewise.
3132 (vreinterpretq_s8_p8): Likewise.
3133 (vreinterpretq_s8_p16): Likewise.
3134 (vreinterpret_s16_s8): Likewise.
3135 (vreinterpret_s16_s32): Likewise.
3136 (vreinterpret_s16_s64): Likewise.
3137 (vreinterpret_s16_f32): Likewise.
3138 (vreinterpret_s16_u8): Likewise.
3139 (vreinterpret_s16_u16): Likewise.
3140 (vreinterpret_s16_u32): Likewise.
3141 (vreinterpret_s16_u64): Likewise.
3142 (vreinterpret_s16_p8): Likewise.
3143 (vreinterpret_s16_p16): Likewise.
3144 (vreinterpretq_s16_s8): Likewise.
3145 (vreinterpretq_s16_s32): Likewise.
3146 (vreinterpretq_s16_s64): Likewise.
3147 (vreinterpretq_s16_f32): Likewise.
3148 (vreinterpretq_s16_u8): Likewise.
3149 (vreinterpretq_s16_u16): Likewise.
3150 (vreinterpretq_s16_u32): Likewise.
3151 (vreinterpretq_s16_u64): Likewise.
3152 (vreinterpretq_s16_p8): Likewise.
3153 (vreinterpretq_s16_p16): Likewise.
3154 (vreinterpret_s32_s8): Likewise.
3155 (vreinterpret_s32_s16): Likewise.
3156 (vreinterpret_s32_s64): Likewise.
3157 (vreinterpret_s32_f32): Likewise.
3158 (vreinterpret_s32_u8): Likewise.
3159 (vreinterpret_s32_u16): Likewise.
3160 (vreinterpret_s32_u32): Likewise.
3161 (vreinterpret_s32_u64): Likewise.
3162 (vreinterpret_s32_p8): Likewise.
3163 (vreinterpret_s32_p16): Likewise.
3164 (vreinterpretq_s32_s8): Likewise.
3165 (vreinterpretq_s32_s16): Likewise.
3166 (vreinterpretq_s32_s64): Likewise.
3167 (vreinterpretq_s32_f32): Likewise.
3168 (vreinterpretq_s32_u8): Likewise.
3169 (vreinterpretq_s32_u16): Likewise.
3170 (vreinterpretq_s32_u32): Likewise.
3171 (vreinterpretq_s32_u64): Likewise.
3172 (vreinterpretq_s32_p8): Likewise.
3173 (vreinterpretq_s32_p16): Likewise.
3174 (vreinterpret_u8_s8): Likewise.
3175 (vreinterpret_u8_s16): Likewise.
3176 (vreinterpret_u8_s32): Likewise.
3177 (vreinterpret_u8_s64): Likewise.
3178 (vreinterpret_u8_f32): Likewise.
3179 (vreinterpret_u8_u16): Likewise.
3180 (vreinterpret_u8_u32): Likewise.
3181 (vreinterpret_u8_u64): Likewise.
3182 (vreinterpret_u8_p8): Likewise.
3183 (vreinterpret_u8_p16): Likewise.
3184 (vreinterpretq_u8_s8): Likewise.
3185 (vreinterpretq_u8_s16): Likewise.
3186 (vreinterpretq_u8_s32): Likewise.
3187 (vreinterpretq_u8_s64): Likewise.
3188 (vreinterpretq_u8_f32): Likewise.
3189 (vreinterpretq_u8_u16): Likewise.
3190 (vreinterpretq_u8_u32): Likewise.
3191 (vreinterpretq_u8_u64): Likewise.
3192 (vreinterpretq_u8_p8): Likewise.
3193 (vreinterpretq_u8_p16): Likewise.
3194 (vreinterpret_u16_s8): Likewise.
3195 (vreinterpret_u16_s16): Likewise.
3196 (vreinterpret_u16_s32): Likewise.
3197 (vreinterpret_u16_s64): Likewise.
3198 (vreinterpret_u16_f32): Likewise.
3199 (vreinterpret_u16_u8): Likewise.
3200 (vreinterpret_u16_u32): Likewise.
3201 (vreinterpret_u16_u64): Likewise.
3202 (vreinterpret_u16_p8): Likewise.
3203 (vreinterpret_u16_p16): Likewise.
3204 (vreinterpretq_u16_s8): Likewise.
3205 (vreinterpretq_u16_s16): Likewise.
3206 (vreinterpretq_u16_s32): Likewise.
3207 (vreinterpretq_u16_s64): Likewise.
3208 (vreinterpretq_u16_f32): Likewise.
3209 (vreinterpretq_u16_u8): Likewise.
3210 (vreinterpretq_u16_u32): Likewise.
3211 (vreinterpretq_u16_u64): Likewise.
3212 (vreinterpretq_u16_p8): Likewise.
3213 (vreinterpretq_u16_p16): Likewise.
3214 (vreinterpret_u32_s8): Likewise.
3215 (vreinterpret_u32_s16): Likewise.
3216 (vreinterpret_u32_s32): Likewise.
3217 (vreinterpret_u32_s64): Likewise.
3218 (vreinterpret_u32_f32): Likewise.
3219 (vreinterpret_u32_u8): Likewise.
3220 (vreinterpret_u32_u16): Likewise.
3221 (vreinterpret_u32_u64): Likewise.
3222 (vreinterpret_u32_p8): Likewise.
3223 (vreinterpret_u32_p16): Likewise.
3224 (vreinterpretq_u32_s8): Likewise.
3225 (vreinterpretq_u32_s16): Likewise.
3226 (vreinterpretq_u32_s32): Likewise.
3227 (vreinterpretq_u32_s64): Likewise.
3228 (vreinterpretq_u32_f32): Likewise.
3229 (vreinterpretq_u32_u8): Likewise.
3230 (vreinterpretq_u32_u16): Likewise.
3231 (vreinterpretq_u32_u64): Likewise.
3232 (vreinterpretq_u32_p8): Likewise.
3233 (vreinterpretq_u32_p16): Likewise.
3235 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3237 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
3239 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
3241 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
3242 (vqnegd_s64): Likewise.
3243 (vqabs_s64): Likewise.
3244 (vqabsd_s64): Likewise.
3246 2014-04-22 Richard Henderson <rth@redhat.com>
3248 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
3249 computation to the top of the loop.
3251 2014-04-22 Renlin <renlin.li@arm.com>
3252 Jiong Wang <jiong.wang@arm.com>
3254 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
3255 * config/aarch64/aarch64.c (aarch64_layout_frame)
3256 (aarch64_initial_elimination_offset): Likewise.
3258 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
3260 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
3263 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
3265 * machmode.h (bitwise_mode_for_mode): Declare.
3266 * stor-layout.h (bitwise_type_for_mode): Likewise.
3267 * stor-layout.c (bitwise_mode_for_mode): New function.
3268 (bitwise_type_for_mode): Likewise.
3269 * builtins.c (fold_builtin_memory_op): Use it instead of
3270 int_mode_for_mode and build_nonstandard_integer_type.
3272 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3274 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
3275 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
3276 (*-*-solaris2*): Simplify.
3277 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
3278 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
3279 *-*-solaris2.9* handling.
3281 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
3283 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
3284 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
3286 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
3287 * configure: Regenerate.
3289 * config/i386/sol2-9.h: Remove.
3291 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
3292 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
3293 Remove Solaris 9 references.
3295 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
3297 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
3298 (floatuns<GPI:mode><GPF:mode>2): Remove.
3299 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
3300 and floatuns conversions.
3301 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
3302 and floatuns conversions.
3303 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
3304 (w1,w2): New mode attributes for inequal width conversions.
3306 2014-04-22 Renlin Li <Renlin.Li@arm.com>
3308 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
3309 the output asm format.
3311 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
3313 * config/aarch64/aarch64-simd.md
3314 (aarch64_cm<optab>di): Always split.
3315 (*aarch64_cm<optab>di): New.
3316 (aarch64_cmtstdi): Always split.
3317 (*aarch64_cmtstdi): New.
3319 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3321 PR tree-optimization/60823
3322 * omp-low.c (ipa_simd_modify_function_body): Go through
3323 all SSA_NAMEs and for those refering to vector arguments
3324 which are going to be replaced adjust SSA_NAME_VAR and,
3325 if it is a default definition, change it into a non-default
3326 definition assigned at the beginning of function from new_decl.
3327 (ipa_simd_modify_stmt_ops): Rewritten.
3328 * tree-dfa.c (set_ssa_default_def): When removing default def,
3329 check for NULL loc instead of NULL *loc.
3331 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3333 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
3334 restrictions on core registers for DImode values in Thumb2.
3336 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3338 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
3339 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
3341 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3343 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
3344 (*iordi_notzesidi_di): Likewise.
3345 (*iordi_notsesidi_di): Likewise.
3347 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3349 * config/arm/arm-protos.h (tune_params): New struct members.
3350 * config/arm/arm.c: Initialise tune_params per processor.
3351 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
3352 for speed, based on new tune_params.
3354 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3356 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
3357 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
3358 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
3359 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
3360 * config/aarch64/arm_neon.h (vrnd_f64): Added.
3361 (vrnda_f64): Likewise.
3362 (vrndi_f64): Likewise.
3363 (vrndm_f64): Likewise.
3364 (vrndn_f64): Likewise.
3365 (vrndp_f64): Likewise.
3366 (vrndx_f64): Likewise.
3368 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3370 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
3371 GET_MODE_SIZE argument is enum machine_mode.
3373 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3376 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
3377 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
3379 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
3382 * asan.c (asan_emit_stack_protection): Force the base to align to
3383 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
3384 appropriate bits if STRICT_ALIGNMENT.
3385 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
3387 (expand_used_vars): Leave a space in the stack frame for alignment
3388 if STRICT_ALIGNMENT.
3390 2014-04-21 David Malcolm <dmalcolm@redhat.com>
3392 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
3394 (gimple_store_p): Likewise.
3395 (gimple_assign_load_p): Likewise.
3396 (gimple_assign_cast_p): Likewise.
3397 (gimple_clobber_p): Likewise.
3399 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
3400 rather than a gimple.
3401 (gimple_assign_cast_p): Likewise.
3403 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
3406 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
3407 If mode is DDmode and TARGET_E500_DOUBLE allow move.
3409 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
3410 more debug information for E500 if -mdebug=reg.
3412 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
3415 * config/i386/i386.c (ix86_expand_builtin)
3416 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
3417 register for target RTX.
3418 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
3420 2014-04-18 Cong Hou <congh@google.com>
3422 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
3423 the widen-mult pattern by handling two operands with different sizes,
3424 and operands whose size is smaller than half of the result type.
3426 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3428 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
3429 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
3430 (do_estimate_edge_time): Compute it.
3431 * ipa-inline.c (want_inline_small_function_p): Bypass
3432 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
3434 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3436 * ipa-inline.c (spec_rem): New static variable.
3437 (dump_overall_stats): New function.
3438 (dump_inline_stats): New function.
3440 2014-04-18 Richard Henderson <rth@redhat.com>
3442 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
3443 to GET_MODE_SIZE, not a reg_class_t.
3445 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3447 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
3448 (vsx_xxmrglw_<mode>): Likewise.
3450 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3453 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
3454 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
3455 (rs6000_init_hard_regno_mode_ok): Likewise.
3457 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
3459 * ipa-inline.c (inline_small_functions): Account only non-cold
3461 * doc/invoke.texi (inline-unit-growth): Update documentation.
3463 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
3465 * config/rs6000/rs6000.md (addti3, subti3): New.
3467 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
3470 * config/i386/i386.c (ix86_expand_clear): Remove outdated
3471 comment. Check optimize_insn_for_size_p instead of
3472 optimize_insn_for_speed_p.
3474 2014-04-17 Martin Jambor <mjambor@suse.cz>
3476 * gimple-iterator.c (gsi_start_edge): New function.
3477 * gimple-iterator.h (gsi_start_edge): Declare.
3478 * tree-sra.c (single_non_eh_succ): New function.
3479 (disqualify_ops_if_throwing_stmt): Renamed to
3480 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
3481 having one non-EH successor BB.
3482 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
3483 generate loads into replacements.
3484 (sra_modify_assign): Likewise and and also use the simple path for
3486 (sra_modify_function_body): Commit statements on edges.
3488 2014-04-17 Richard Biener <rguenther@suse.de>
3491 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
3492 comparison results and add clarifying comment.
3494 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3496 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
3497 (blank_mode): Initialize it.
3498 (emit_mode_size_inline, emit_mode_nunits_inline,
3499 emit_mode_inner_inline): New functions.
3500 (emit_insn_modes_h): Call them and surround their output with
3501 #if GCC_VERSION >= 4001 ... #endif.
3502 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
3503 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
3504 mode_* arrays if the argument is __builtin_constant_p.
3505 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
3506 is enum machine_mode.
3508 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3510 * passes.c (opt_pass::execute): Adjust.
3511 (pass_manager::execute_pass_mode_switching): Likewise.
3512 (early_local_passes::execute): Likewise.
3513 (execute_one_pass): Pass cfun to the pass's execute method.
3514 * tree-pass.h (opt_pass::execute): Add function * argument.
3515 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3516 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3517 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3518 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3519 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3520 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
3521 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
3522 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3523 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3524 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3525 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3526 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3527 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3528 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3529 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3530 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3531 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3532 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3533 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3534 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3535 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3536 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3537 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3538 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3539 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3540 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3541 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3542 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3543 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3546 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3548 * passes.c (opt_pass::gate): Take function * argument.
3549 (gate_all_early_local_passes): Merge into
3550 (early_local_passes::gate): this.
3551 (gate_all_early_optimizations): Merge into
3552 (all_early_optimizations::gate): this.
3553 (gate_all_optimizations): Mege into
3554 (all_optimizations::gate): this.
3555 (gate_all_optimizations_g): Merge into
3556 (all_optimizations_g::gate): this.
3557 (gate_rest_of_compilation): Mege into
3558 (rest_of_compilation::gate): this.
3559 (gate_postreload): Merge into
3560 (postreload::gate): this.
3561 (dump_one_pass): Pass cfun to the pass's gate method.
3562 (execute_ipa_summary_passes): Likewise.
3563 (execute_one_pass): Likewise.
3564 (ipa_write_summaries_2): Likewise.
3565 (ipa_write_optimization_summaries_1): Likewise.
3566 (ipa_read_summaries_1): Likewise.
3567 (ipa_read_optimization_summaries_1): Likewise.
3568 (execute_ipa_stmt_fixups): Likewise.
3569 * tree-pass.h (opt_pass::gate): Add function * argument.
3570 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
3571 combine-stack-adj.c, combine.c, compare-elim.c,
3572 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3573 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
3574 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
3575 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
3576 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3577 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3578 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3579 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
3580 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
3581 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
3582 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3583 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3584 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
3585 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3586 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3587 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3588 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3589 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3590 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3591 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3592 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3593 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3594 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3595 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
3596 var-tracking.c, vtable-verify.c, web.c: Adjust.
3598 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3600 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
3601 * configure: Regenerate.
3603 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3605 * passes.c (dump_one_pass): don't check pass->has_gate.
3606 (execute_ipa_summary_passes): Likewise.
3607 (execute_one_pass): Likewise.
3608 (ipa_write_summaries_2): Likewise.
3609 (ipa_write_optimization_summaries_1): Likewise.
3610 (ipa_read_optimization_summaries_1): Likewise.
3611 (execute_ipa_stmt_fixups): Likewise.
3612 * tree-pass.h (pass_data::has_gate): Remove.
3613 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3614 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3615 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3616 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3617 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3618 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
3619 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
3620 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
3621 gimple-low.c, gimple-ssa-isolate-paths.c,
3622 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
3623 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3624 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
3625 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
3626 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
3627 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
3628 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
3629 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3630 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
3631 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3632 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
3633 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3634 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3635 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3636 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3637 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3638 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3639 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3640 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3641 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3642 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3643 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3646 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3648 * pass_manager.h (pass_manager::register_dump_files_1): Remove
3650 * passes.c (pass_manager::register_dump_files_1): Merge into
3651 (pass_manager::register_dump_files): this, and remove its handling of
3652 properties since the pass always has the properties anyway.
3653 (pass_manager::pass_manager): Adjust.
3655 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3657 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
3658 * passes.c (pass_manager::register_dump_files_1): Remove dead code
3659 dealing with properties.
3660 (pass_manager::register_dump_files): Adjust.
3662 2014-03-20 Mark Wielaard <mjw@redhat.com>
3664 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
3665 then represent the bound as normal constant value.
3667 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3670 Forward port from 4.8 branch
3671 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
3673 * config/i386/bmiintrin.h (_blsi_u32): New.
3677 (_blsmsk_u32): Ditto.
3678 (_blsmsk_u64): Ditto.
3679 (_tzcnt_u32): Ditto.
3680 (_tzcnt_u64): Ditto.
3682 2014-04-17 Kito Cheng <kito@0xlab.org>
3684 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
3686 2014-04-17 Richard Biener <rguenther@suse.de>
3689 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
3690 boolean results for comparisons.
3692 2014-04-17 Richard Biener <rguenther@suse.de>
3694 PR tree-optimization/60836
3695 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
3696 initial PHI args to be gimple values.
3698 2014-04-17 Richard Biener <rguenther@suse.de>
3700 PR tree-optimization/60841
3701 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
3702 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
3703 of stmts to SLP build.
3704 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
3705 (vect_analyze_slp): Likewise.
3706 (vect_analyze_slp_instance): Likewise.
3707 (vect_build_slp_tree): Limit overall SLP tree growth.
3708 * tree-vectorizer.h (vect_analyze_data_refs,
3709 vect_analyze_slp): Adjust prototypes.
3711 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3713 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
3716 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3718 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
3719 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
3720 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
3721 for TARGET_SLOW_PSHUFB
3723 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
3725 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
3726 * config/i386/i386.c (intel_cost): Ditto.
3728 2014-04-17 Joey Ye <joey.ye@arm.com>
3730 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
3732 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3734 * opts.c (common_handle_option): Disable -fipa-reference coorectly
3737 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3739 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
3740 (type_all_derivations_known_p): New predicate.
3741 (type_all_ctors_visible_p): New predicate.
3742 (type_possibly_instantiated_p): New predicate.
3743 (get_odr_type): Compute all_derivations_known.
3744 (dump_odr_type): Dump the flag.
3745 (maybe_record_type): Cleanup.
3746 (record_target_from_binfo): Add bases_to_consider array;
3747 record bases for types w/o instances and skip CXX destructor.
3748 (possible_polymorphic_call_targets_1): Add bases_to_consider
3749 and consider_construction parameters; check if type may have instance.
3750 (get_polymorphic_call_info): Set maybe_in_construction to true
3751 when we know nothing.
3752 (record_targets_from_bases): Skip CXX destructors; they are
3753 never called for types in construction.
3754 (possible_polymorphic_call_targets): Do not record target when
3755 type may not have instance.
3757 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
3760 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
3761 external aliases alive, too.
3763 2014-04-16 Andrew Pinski <apinski@cavium.com>
3765 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
3768 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3770 * final.c (compute_alignments): Do not apply loop alignment to a block
3771 falling through to the exit.
3773 2014-04-16 Catherine Moore <clm@codesourcery.com>
3775 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
3776 Adjust constraints for microMIPS store patterns.
3778 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
3780 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
3782 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
3784 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
3785 (append_use): Run at -O0.
3786 (append_vdef): Likewise.
3787 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
3788 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
3790 2014-04-16 Jakub Jelinek <jakub@redhat.com>
3792 PR tree-optimization/60844
3793 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
3794 (propagate_op_to_single_use, remove_visited_stmt_chain,
3795 linearize_expr, repropagate_negates, reassociate_bb): Use it
3796 instead of gsi_remove.
3798 2014-04-16 Martin Jambor <mjambor@suse.cz>
3800 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
3801 ipa_transforms_to_apply.
3802 (cgraph_function_versioning): Assert that old_node has empty
3803 ipa_transforms_to_apply.
3804 * trans-mem.c (ipa_tm_create_version): Likewise.
3805 * tree-inline.c (tree_function_versioning): Do not duplicate
3806 ipa_transforms_to_apply.
3808 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3811 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
3813 Pass necessary as flags on 64-bit Solaris/x86.
3814 Use lowercase relocs for x86_64-*-*.
3815 * configure: Regenerate.
3817 2014-04-15 Jan Hubicka <jh@suse.cz>
3819 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
3820 (maybe_record_node, likely_target_p): Use it.
3822 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3825 Revert following patch
3827 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3830 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3831 software floating point or no floating point registers, do not
3832 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3833 in GPRs that occurs after we tested for GPRs that would never be
3836 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3837 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3838 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3839 specifically allow DDmode, since that does not use the SPE SIMD
3842 2014-03-21 Mark Wielaard <mjw@redhat.com>
3844 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
3845 as unsigned or int depending on type and value used.
3847 2014-04-15 Richard Biener <rguenther@suse.de>
3849 PR rtl-optimization/56965
3850 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
3851 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
3853 * alias.c (true_dependence_1): Do not call
3854 nonoverlapping_component_refs_p.
3855 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
3856 nonoverlapping_component_refs_p.
3857 (indirect_refs_may_alias_p): Likewise.
3859 2014-04-15 Teresa Johnson <tejohnson@google.com>
3861 * cfg.c (dump_bb_info): Fix flags check.
3862 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
3864 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3866 PR rtl-optimization/60663
3867 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
3870 2014-04-15 Richard Biener <rguenther@suse.de>
3872 * lto-streamer.h (LTO_major_version): Bump to 4.
3874 2014-04-15 Richard Biener <rguenther@suse.de>
3876 * common.opt (lto_partition_model): New enum.
3877 (flto-partition=): Merge separate options with a single with argument,
3878 add -flto-partition=one support.
3879 * flag-types.h (enum lto_partition_model): Declare.
3880 * opts.c (finish_options): Remove duplicate -flto-partition=
3882 * lto-wrapper.c (run_gcc): Adjust.
3884 2014-04-15 Richard Biener <rguenther@suse.de>
3886 * alias.c (ncr_compar): New function.
3887 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
3889 2014-04-15 Richard Biener <rguenther@suse.de>
3891 * alias.c (record_component_aliases): Do not walk BINFOs.
3893 2014-04-15 Richard Biener <rguenther@suse.de>
3895 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
3896 Add struct function argument and adjust.
3897 (find_func_aliases_for_call): Likewise.
3898 (find_func_aliases): Likewise.
3899 (find_func_clobbers): Likewise.
3900 (intra_create_variable_infos): Likewise.
3901 (compute_points_to_sets): Likewise.
3902 (ipa_pta_execute): Adjust. Do not push/pop cfun.
3904 2014-04-15 Richard Biener <rguenther@suse.de>
3906 * tree.c (iterative_hash_expr): Use enum tree_code_class
3907 to store TREE_CODE_CLASS.
3908 (tree_block): Likewise.
3909 (tree_set_block): Likewise.
3910 * tree.h (fold_build_pointer_plus_loc): Use
3911 convert_to_ptrofftype_loc.
3913 2014-04-15 Jakub Jelinek <jakub@redhat.com>
3916 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
3919 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
3921 * cfgloop.h (struct loop): Move force_vectorize down.
3922 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
3923 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
3924 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
3925 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
3926 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
3927 * tree-core.h (enum annot_expr_kind): Add new kind values.
3928 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
3929 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
3931 * tree.def (ANNOTATE_EXPR): Tweak comment.
3933 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3935 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
3938 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
3940 * tree.h (TYPE_IDENTIFIER): Declare.
3941 * tree.c (subrange_type_for_debug_p): Use it.
3942 * godump.c (go_format_type): Likewise.
3943 * dwarf2out.c (is_cxx_auto, modified_type_die,
3944 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
3945 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
3947 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3950 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
3952 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
3954 * config/i386/i386.c (examine_argument): Return bool. Return true if
3955 parameter should be passed in memory.
3956 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
3957 (construct_container): Update calls to examine_argument.
3958 (function_arg_advance_64): Ditto.
3959 (return_in_memory_32): Merge with ix86_return_in_memory.
3960 (return_in_memory_64): Ditto.
3961 (return_in_memory_ms_64): Ditto.
3963 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
3965 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
3966 * coverage.c (coverage_compute_profile_id): Handle externally visible
3969 2014-04-14 Martin Jambor <mjambor@suse.cz>
3971 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
3972 DECL_DISREGARD_INLINE_LIMITS functions.
3974 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3977 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
3979 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
3982 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
3983 optimize_insn_for_speed_p instead of
3984 optimize_function_for_speed_p.
3986 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
3988 * doc/invoke.texi (free): Document AArch64.
3990 2014-04-14 Richard Biener <rguenther@suse.de>
3992 PR tree-optimization/60042
3993 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
3994 (insert_into_preds_of_block): Do not prevent PHI insertion
3995 for REFERENCE exprs here ...
3996 (eliminate_dom_walker::before_dom_children): ... but prevent
3997 their use here under similar conditions when applied to the
3998 IL after PRE optimizations.
4000 2014-04-14 Richard Biener <rguenther@suse.de>
4002 * passes.def: Move early points-to after early SRA.
4004 2014-04-14 Richard Biener <rguenther@suse.de>
4006 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
4007 check for which sign-changes we allow when forwarding
4008 a converted value into a switch.
4010 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
4012 * stor-layout.c (place_field): Finalize non-constant offset for the
4015 2014-04-14 Richard Biener <rguenther@suse.de>
4017 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
4019 (expand_switch_using_bit_tests_p): Likewise.
4020 (process_switch): Compute and pass on speed_p based on the
4022 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
4023 optimize_bb_for_speed_p.
4025 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
4027 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
4028 * function.h (struct function): Rename has_force_vect_loops into
4029 has_force_vectorize_loops.
4030 * lto-streamer-in.c (input_cfg): Adjust for renaming.
4031 (input_struct_function_base): Likewise.
4032 * lto-streamer-out.c (output_cfg): Likewise.
4033 (output_struct_function_base): Likewise.
4034 * omp-low.c (expand_omp_simd): Likewise.
4035 * tree-cfg.c (move_sese_region_to_fn): Likewise.
4036 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
4037 (version_loop_for_if_conversion): Likewise.
4038 (tree_if_conversion): Likewise.
4039 (main_tree_if_conversion): Likewise.
4040 (gate_tree_if_conversion): Likewise.
4041 * tree-inline.c (copy_loops): Likewise.
4042 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
4043 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
4044 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
4045 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
4046 * tree-vectorizer.c (vectorize_loops): Likewise.
4047 * tree-vectorizer.h (unlimited_cost_model): Likewise.
4049 2014-04-14 Richard Biener <rguenther@suse.de>
4052 * lto-streamer-out.c (wrap_refs): New function.
4053 (lto_output): Wrap symbol references in global initializes in
4054 type-preserving MEM_REFs.
4056 2014-04-14 Christian Bruel <christian.bruel@st.com>
4058 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
4060 2014-04-14 Christian Bruel <christian.bruel@st.com>
4062 * config/sh/sh.md (setmemqi): New expand pattern.
4063 * config/sh/sh.h (CLEAR_RATIO): Define.
4064 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
4065 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
4067 2014-04-14 Richard Biener <rguenther@suse.de>
4070 * fold-const.c (negate_expr_p): Don't negate directional rounding
4072 (fold_negate_expr): Likewise.
4074 2014-04-14 Richard Biener <rguenther@suse.de>
4076 PR tree-optimization/59817
4077 PR tree-optimization/60453
4078 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
4079 recursion to catch all CHRECs in the scalar evolution and restrict
4080 the predicate for the remains appropriately.
4082 2014-04-12 Catherine Moore <clm@codesourcery.com>
4084 * config/mips/constraints.md: Add new register constraint "kb".
4085 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
4086 (*movhi_internal): Likewise.
4087 (*movqi_internal): Likewise.
4088 * config/mips/mips.h (M16_STORE_REGS): New register class.
4089 (REG_CLASS_NAMES): Add M16_STORE_REGS.
4090 (REG_CLASS_CONTENTS): Likewise.
4091 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
4093 2014-04-11 Tobias Burnus <burnus@net-b.de>
4096 * doc/invoke.texi (-Wformat-signedness): Document it.
4097 (Wformat=2): Mention that this enables -Wformat-signedness.
4099 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4101 * common/config/epiphany/epiphany-common.c
4102 (epiphany_option_optimization_table): Enable section anchors by
4103 default at -O1 or higher.
4104 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
4105 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
4106 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
4107 carries no extra cost.
4108 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
4109 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
4110 * config/epiphany/predicates.md (memclob_operand): New predicate.
4111 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
4112 Use memclob_operand predicate and X constraint for operand 3.
4114 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4116 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
4117 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
4120 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4122 PR rtl-optimization/60651
4123 * mode-switching.c (optimize_mode_switching): Make sure to emit
4124 sets of a lower numbered entity before sets of a higher numbered
4125 entity to a mode of the same or lower priority.
4126 When creating a seginfo for a basic block that starts with a code
4127 label, move the insertion point past the code label.
4128 (new_seginfo): Document and enforce requirement that
4129 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
4130 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
4131 * doc/tm.texi: Regenerate.
4133 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
4136 * config/arc/arc.c (arc_save_restore): Fix assert typo.
4138 2013-04-11 Jakub Jelinek <jakub@redhat.com>
4140 * BASE-VER: Set to 4.10.0.
4142 2014-04-11 Tobias Burnus <burnus@net-b.de>
4145 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
4146 * doc/gcc.texi (Service): Update description in the @menu
4147 * doc/invoke.texi (Option Summary): Remove misplaced and
4150 2014-04-11 Steve Ellcey <sellcey@mips.com>
4151 Jakub Jelinek <jakub@redhat.com>
4154 * expr.c (convert_move): Use emit_store_flag_force instead of
4155 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
4158 2014-04-11 Richard Biener <rguenther@suse.de>
4161 * varasm.c (assemble_alias): Avoid endless error reporting
4162 recursion by setting TREE_ASM_WRITTEN.
4164 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4166 * config/s390/s390.md: Add a splitter for NOT rtx.
4168 2014-04-11 Jakub Jelinek <jakub@redhat.com>
4170 PR rtl-optimization/60663
4171 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
4173 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
4174 Jakub Jelinek <jakub@redhat.com>
4177 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
4178 flag from decl_node to node.
4180 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4183 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
4184 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
4185 ameliorating the cases where it can be.
4187 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
4190 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4192 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4193 (loadsync_<mode>): Change mode.
4194 (load_quadpti, store_quadpti): New.
4195 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4196 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4197 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
4199 2014-04-09 Cong Hou <congh@google.com>
4202 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
4205 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4207 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
4208 instead of vnor to exploit possible fusion opportunity in the
4210 (altivec_expand_vec_perm_const_le): Likewise.
4212 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4214 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4215 (loadsync_<mode>): Change mode.
4216 (load_quadpti, store_quadpti): New.
4217 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4218 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4220 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
4223 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
4224 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
4225 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
4227 2014-04-08 Richard Biener <rguenther@suse.de>
4230 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
4231 a 64bit widest int print double-int similar to on HWI64 hosts.
4233 2014-04-08 Richard Biener <rguenther@suse.de>
4235 PR tree-optimization/60785
4236 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
4237 default defs properly.
4239 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
4241 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
4242 (Weffc++): Likewise.
4244 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
4246 * ipa-devirt.c (maybe_record_node): When node is not recorded,
4247 set completep to false rather than true.
4249 2014-04-07 Douglas B Rupp <rupp@adacore.com>
4252 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
4253 ARM_TARGET2_DWARF_FORMAT.
4255 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
4258 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
4259 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
4262 2014-04-07 Richard Biener <rguenther@suse.de>
4264 PR tree-optimization/60766
4265 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
4266 (may_eliminate_iv): Convert cand_value_at result to desired type.
4268 2014-04-07 Jason Merrill <jason@redhat.com>
4271 * common.opt (-fno-gnu-unique): Add.
4272 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
4274 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4276 * haifa-sched.c: Fix outdated function reference and minor
4277 grammar errors in introductory comment.
4279 2014-04-07 Richard Biener <rguenther@suse.de>
4282 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
4284 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
4286 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
4289 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
4290 size accounting for thunks.
4291 (pa_asm_output_mi_thunk): Use final_start_function() and
4292 final_end_function() to output function start and end directives.
4294 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4296 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
4297 device specific ISA/ feature information. Remove short_sp and
4298 errata_skip ds. Add avr_device_specific_features enum to have device
4300 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
4301 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
4302 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
4303 updated device specific info.
4304 * config/avr/avr-mcus.def: Merge device specific details to
4305 dev_attribute field.
4306 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
4308 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
4309 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
4310 assembler if RMW isa supported by current device.
4311 * config/avr/genmultilib.awk: Update as device info structure changed.
4312 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
4314 2014-04-04 Cong Hou <congh@google.com>
4316 PR tree-optimization/60656
4317 * tree-vect-stmts.c (supportable_widening_operation):
4318 Fix a bug that elements in a vector with vect_used_by_reduction
4319 property are incorrectly reordered when the operation on it is not
4320 consistant with the one in reduction operation.
4322 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
4324 PR rtl-optimization/60155
4325 * gcse.c (record_set_data): New function.
4326 (single_set_gcse): New function.
4327 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
4328 (hoist_code): Likewise.
4329 (get_pressure_class_and_nregs): Likewise.
4331 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
4333 * explow.c (probe_stack_range): Emit a final optimization blockage.
4335 2014-04-04 Anthony Green <green@moxielogic.com>
4337 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
4340 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
4343 * lto-cgraph.c (input_overwrite_node): Check that partitioning
4344 flags are set only during streaming.
4345 * ipa.c (process_references, walk_polymorphic_call_targets,
4346 symtab_remove_unreachable_nodes): Drop bodies of always inline
4347 after early inlining.
4348 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
4350 2014-04-04 Jakub Jelinek <jakub@redhat.com>
4351 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4354 * dwarf2out.c (const_ok_for_output_1): Reject expressions
4357 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4360 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
4362 (cortex_a53_fdivd): Likewise.
4364 2014-04-04 Martin Jambor <mjambor@suse.cz>
4367 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
4369 * cgraph.c (clone_of_p): Also return true if thunks match.
4370 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
4371 cgraph_function_or_thunk_node and an obsolete comment.
4372 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
4374 (build_function_decl_skip_args): Likewise.
4375 (set_new_clone_decl_and_node_flags): New function.
4376 (duplicate_thunk_for_node): Likewise.
4377 (redirect_edge_duplicating_thunks): Likewise.
4378 (cgraph_clone_node): New parameter args_to_skip, pass it to
4379 redirect_edge_duplicating_thunks which is called instead of
4380 cgraph_redirect_edge_callee.
4381 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
4382 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
4384 2014-04-04 Jeff Law <law@redhat.com>
4387 * config/arm/predicates.md (const_int_I_operand): New predicate.
4388 (const_int_M_operand): Similarly.
4389 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
4391 (insv_t2, extv_reg, extzv_t2): Likewise.
4392 (load_multiple_with_writeback): Similarly for const_int_I_operand.
4393 (pop_multiple_with_writeback_and_return): Likewise.
4394 (vfp_pop_multiple_with_writeback): Likewise
4396 2014-04-04 Richard Biener <rguenther@suse.de>
4399 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
4400 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
4402 * gimplify.h (gimple_add_tmp_var_fn): Declare.
4403 * gimplify.c (gimple_add_tmp_var_fn): New function.
4404 * gimple-expr.h (create_tmp_reg_fn): Declare.
4405 * gimple-expr.c (create_tmp_reg_fn): New function.
4406 * gimple-low.c (record_vars_into): Don't change cfun.
4407 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
4408 code generation without cfun.
4410 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
4413 * Makefile.in (install-driver): Fix shell scripting.
4415 2014-04-03 Cong Hou <congh@google.com>
4417 PR tree-optimization/60505
4418 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
4419 threshold of number of iterations below which no vectorization
4421 * tree-vect-loop.c (new_loop_vec_info):
4422 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
4423 * tree-vect-loop.c (vect_analyze_loop_operations):
4424 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
4425 * tree-vect-loop.c (vect_transform_loop):
4426 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
4427 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
4428 of iterations of the loop and see if we should build the epilogue.
4430 2014-04-03 Richard Biener <rguenther@suse.de>
4432 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
4433 (streamer_tree_cache_create): Adjust.
4434 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
4435 to allow optional nodes array.
4436 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
4437 (streamer_tree_cache_append): Likewise.
4438 (streamer_tree_cache_create): Create nodes array optionally
4439 as specified by parameter.
4440 * lto-streamer-out.c (create_output_block): Avoid maintaining
4441 the node array in the writer cache.
4442 (DFS_write_tree): Remove assertion.
4443 (produce_asm_for_decls): Free the out decl state hash table early.
4444 * lto-streamer-in.c (lto_data_in_create): Adjust for
4445 streamer_tree_cache_create prototype change.
4447 2014-04-03 Richard Biener <rguenther@suse.de>
4449 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
4450 set TREE_CHAIN to NULL_TREE.
4452 2014-04-03 Richard Biener <rguenther@suse.de>
4454 PR tree-optimization/60740
4455 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
4456 over all GIMPLE_COND operands.
4458 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
4460 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
4461 (Weffc++): Remove Scott's numbering, merge lists and reference
4464 2014-04-03 Nick Clifton <nickc@redhat.com>
4466 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
4469 2014-04-03 Martin Jambor <mjambor@suse.cz>
4471 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
4472 mention gcc_unreachable before failing.
4473 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
4476 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
4479 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
4480 inconsistent code and instead mark the context inconsistent.
4481 (possible_polymorphic_call_targets): For inconsistent contexts
4482 return empty complete list.
4484 2014-04-02 Anthony Green <green@moxielogic.com>
4486 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
4487 (extendqisi2, extendhisi2): Define.
4488 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
4489 (WCHAR_TYPE): Change to unsigned int.
4491 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4493 PR tree-optimization/60733
4494 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
4495 insertion point for PHI candidates to be the end of the feeding
4496 block for the PHI argument.
4498 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
4500 PR rtl-optimization/60650
4501 * lra-constraints.c (process_alt_operands): Decrease reject for
4502 earlyclobber matching.
4504 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4506 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
4508 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4510 * config/spu/spu.c (pad_bb): Do not crash when the last
4511 insn is CODE_FOR_blockage.
4513 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4515 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
4516 lies outside the target mode.
4518 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4521 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4522 software floating point or no floating point registers, do not
4523 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4524 in GPRs that occurs after we tested for GPRs that would never be
4527 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4528 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4529 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4530 specifically allow DDmode, since that does not use the SPE SIMD
4533 2014-04-02 Richard Biener <rguenther@suse.de>
4536 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4537 MODE_INTs. Properly use negv_optab.
4538 (expand_abs): Likewise.
4540 2014-04-02 Richard Biener <rguenther@suse.de>
4543 * Makefile.in (install-driver): Guard extra installs with special
4546 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4548 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4551 2014-04-01 Richard Henderson <rth@redhat.com>
4554 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
4555 alternative enabled before register allocation.
4557 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
4559 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
4560 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
4562 (nios2_large_got_address): Remove unneeded 'sym' parameter.
4563 (nios2_got_address): Update nios2_large_got_address call site.
4564 (nios2_delegitimize_address): New function.
4565 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
4566 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
4567 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
4569 2014-04-01 Martin Husemann <martin@duskware.de>
4571 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
4574 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
4576 PR rtl-optimization/60604
4577 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
4578 check from register_operand.
4579 (register_operand): Redefine in terms of general_operand.
4580 (nonmemory_operand): Use register_operand for the non-constant cases.
4582 2014-04-01 Richard Biener <rguenther@suse.de>
4584 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
4586 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
4588 * doc/invoke.texi (mapp-regs): Clarify.
4590 2014-03-31 Ulrich Drepper <drepper@gmail.com>
4592 * config/i386/avx512fintrin.h (__v32hi): Define type.
4593 (__v64qi): Likewise.
4594 (_mm512_set1_epi8): Define.
4595 (_mm512_set1_epi16): Define.
4596 (_mm512_set4_epi32): Define.
4597 (_mm512_set4_epi64): Define.
4598 (_mm512_set4_pd): Define.
4599 (_mm512_set4_ps): Define.
4600 (_mm512_setr4_epi64): Define.
4601 (_mm512_setr4_epi32): Define.
4602 (_mm512_setr4_pd): Define.
4603 (_mm512_setr4_ps): Define.
4604 (_mm512_setzero_epi32): Define.
4606 2014-03-31 Martin Jambor <mjambor@suse.cz>
4609 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
4610 callsite_arguments_match_p. Updated all callers. Also check types of
4611 corresponding formal parameters and actual arguments.
4612 (not_all_callers_have_enough_arguments_p) Renamed to
4613 some_callers_have_mismatched_arguments_p.
4615 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
4617 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
4619 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
4622 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
4625 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
4627 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
4628 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
4630 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
4631 Use FMAMODE_NOVF512 mode iterator.
4632 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
4633 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
4634 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
4635 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
4637 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
4638 Use VF_128_256 mode iterator.
4639 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
4642 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4644 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
4645 static chain if needed.
4647 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4650 * lra-constraints.c (index_part_to_reg): New.
4651 (process_address): Use it.
4653 2014-03-27 Jeff Law <law@redhat.com>
4654 Jakub Jelinek <jakub@redhat.com>
4657 * expr.c (do_tablejump): Use simplify_gen_binary rather than
4658 gen_rtx_{PLUS,MULT} to build up the address expression.
4660 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
4661 creating non-canonical RTL.
4663 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4666 * ipa-inline.c (want_inline_small_function_p): Short circuit large
4667 functions; reorganize to make cheap checks first.
4668 (inline_small_functions): Do not estimate growth when dumping;
4670 * ipa-inline.h (inline_summary): Add min_size.
4671 (growth_likely_positive): New function.
4672 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
4673 (set_cond_stmt_execution_predicate): Cleanup.
4674 (estimate_edge_size_and_time): Compute min_size.
4675 (estimate_calls_size_and_time): Likewise.
4676 (estimate_node_size_and_time): Likewise.
4677 (inline_update_overall_summary): Update min_size.
4678 (do_estimate_edge_time): Likewise.
4679 (do_estimate_edge_size): Update.
4680 (do_estimate_edge_hints): Update.
4681 (growth_likely_positive): New function.
4683 2014-03-28 Jakub Jelinek <jakub@redhat.com>
4686 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
4687 also if addr has VOIDmode.
4689 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4691 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
4692 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
4694 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
4695 instructions as well as AdvancedSIMD loads.
4697 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4699 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
4700 Use crypto_aese type.
4701 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
4702 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
4703 crypto_aese, crypto_aesmc. Move to types.md.
4704 * config/arm/types.md (crypto_aes): Split into crypto_aese,
4706 * config/arm/iterators.md (crypto_type): Likewise.
4708 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
4710 * cgraph.c: Include expr.h and tree-dfa.h.
4711 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
4714 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
4717 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
4718 regs from checking multi-reg pseudos.
4720 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4722 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
4724 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4726 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
4727 if it would clobber the stack pointer, even temporarily.
4729 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
4731 * mode-switching.c: Make small adjustments to the top comment.
4733 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4735 * config/rs6000/constraints.md (wD constraint): New constraint to
4736 match the constant integer to get the top DImode/DFmode out of a
4737 vector in a VSX register.
4739 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
4740 match the constant integer to get the top DImode/DFmode out of a
4741 vector in a VSX register.
4743 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
4746 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
4749 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
4750 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
4752 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
4753 Optimize vec_extract of 64-bit values, where the value being
4754 extracted is in the top word, where we can use scalar
4755 instructions. Add direct move and store support. Combine the big
4756 endian/little endian vector select load support into a single insn.
4757 (vsx_extract_<mode>_internal1): Likewise.
4758 (vsx_extract_<mode>_internal2): Likewise.
4759 (vsx_extract_<mode>_load): Likewise.
4760 (vsx_extract_<mode>_store): Likewise.
4761 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
4762 combined into vsx_extract_<mode>_load.
4763 (vsx_extract_<mode>_one_le): Likewise.
4765 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
4766 define the top 64-bit vector element.
4768 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
4771 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4772 Document vec_vbpermq builtin.
4775 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
4776 enable use of xxsldwi and xxpermdi builtin functions.
4777 (vec_xxpermdi): Likewise.
4779 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4780 Document use of vec_xxsldwi and vec_xxpermdi builtins.
4782 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
4784 PR rtl-optimization/60650
4785 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
4787 (find_spills_for): New.
4788 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
4789 Spill all pseudos on the second iteration.
4791 2014-03-27 Marek Polacek <polacek@redhat.com>
4794 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
4797 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4799 * config/s390/s390.c (s390_can_use_return_insn): Check for
4800 call-saved FPRs on 31 bit.
4802 2014-03-27 Jakub Jelinek <jakub@redhat.com>
4805 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
4806 if they need regimplification, just drop them instead of
4807 calling gimple_regimplify_operands on them.
4809 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
4812 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
4813 (aarch64_frame_pointer_required): Adjust logic.
4814 (aarch64_can_eliminate): Adjust logic.
4815 (aarch64_override_options_after_change): Adjust logic.
4817 2014-03-27 Dehao Chen <dehao@google.com>
4819 * ipa-inline.c (early_inliner): Update node's inline info.
4821 2014-03-26 Dehao Chen <dehao@google.com>
4823 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
4824 compiler inserted conditional jumps for NAN float check.
4826 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4828 * ubsan.h (ubsan_create_data): Change second argument's type
4829 to const location_t *.
4830 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
4832 (ubsan_create_data): Change second argument to const location_t *PLOC.
4833 Create Loc field whenever PLOC is non-NULL.
4834 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
4835 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
4839 * real.c (real_to_integer2): Change type of low to UHWI.
4841 2014-03-26 Tobias Burnus <burnus@net-b.de>
4843 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
4844 (CILK_SELF_SPECS): New define.
4845 (driver_self_specs): Use it.
4847 2014-03-26 Richard Biener <rguenther@suse.de>
4849 * tree-pretty-print.c (percent_K_format): Implement special
4850 case for LTO and its stripped down BLOCK tree.
4852 2014-03-26 Jakub Jelinek <jakub@redhat.com>
4855 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
4857 * tree-vrp.c (simplify_internal_call_using_ranges): If only
4858 one range is range_int_cst_p, but not both, at least optimize
4859 addition/subtraction of 0 and multiplication by 0 or 1.
4860 * gimple-fold.c (gimple_fold_call): Fold
4861 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
4862 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
4863 INTEGER_CSTs, try to fold at least x * 0 and y - y.
4865 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
4867 PR rtl-optimization/60452
4868 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
4869 <case REG>: Return 1 for invalid offsets from the frame pointer.
4871 2014-03-26 Marek Polacek <polacek@redhat.com>
4874 * doc/extend.texi (C Extensions): Mention variable-length arrays in
4877 2014-03-26 Marek Polacek <polacek@redhat.com>
4880 * doc/extend.texi (Designated Inits): Describe what happens to omitted
4883 2014-03-26 Marek Polacek <polacek@redhat.com>
4886 * ira-color.c (update_conflict_hard_regno_costs): Perform the
4887 multiplication in unsigned type.
4889 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4891 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
4893 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
4895 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
4897 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4900 * cif-code.def (UNREACHABLE) New code.
4901 * ipa-inline.c (inline_small_functions): Skip edges to
4902 __builtlin_unreachable.
4903 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
4904 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
4905 predicate to __bulitin_unreachable.
4906 (set_cond_stmt_execution_predicate): Fix issue when
4907 invert_tree_comparison returns ERROR_MARK.
4908 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
4909 propagate to inline clones.
4910 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
4912 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
4913 * cgraphclones.c (cgraph_clone_node): If call destination is already
4914 ureachable, do not redirect it back.
4915 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
4918 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
4920 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
4921 Do not modify inline clones.
4923 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4925 * config/i386/i386.md (general_sext_operand): New mode attr.
4926 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
4927 don't generate (sign_extend (const_int)).
4928 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
4929 operands[2]. Use We constraint instead of <i> and
4930 <general_sext_operand> predicate instead of <general_operand>.
4931 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
4932 * config/i386/constraints.md (We): New constraint.
4933 * config/i386/predicates.md (x86_64_sext_operand,
4934 sext_operand): New predicates.
4936 2014-03-25 Martin Jambor <mjambor@suse.cz>
4939 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
4940 inconsistent devirtualizations to __builtin_unreachable.
4942 2014-03-25 Marek Polacek <polacek@redhat.com>
4945 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
4947 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
4949 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
4950 order of elements for big-endian.
4952 2014-03-25 Richard Biener <rguenther@suse.de>
4955 * gimplify-me.c (gimple_regimplify_operands): Update the
4958 2014-03-25 Martin Jambor <mjambor@suse.cz>
4961 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
4962 (lto_output_varpool_node): Likewise.
4963 (input_overwrite_node): Likewise.
4964 (input_varpool_node): Likewise.
4966 2014-03-25 Richard Biener <rguenther@suse.de>
4968 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
4969 (run_gcc): Likewise.
4971 2014-03-25 Jakub Jelinek <jakub@redhat.com>
4973 * combine.c (simplify_compare_const): Add MODE argument.
4974 Handle mode_width 0 as very large mode_width.
4975 (try_combine, simplify_comparison): Adjust callers.
4977 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
4978 type to avoid signed integer overflow.
4979 * explow.c (plus_constant): Likewise.
4981 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
4983 * doc/generic.texi: Correct typos.
4985 2014-03-24 Tobias Burnus <burnus@net-b.de>
4987 * doc/invoke.texi (-flto): Expand section about
4988 using static libraries with LTO.
4990 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4992 PR rtl-optimization/60501
4993 * optabs.def (addptr3_optab): New optab.
4994 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
4995 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
4996 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
4998 * lra.c (emit_add3_insn): Use the addptr pattern if available.
5000 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
5002 2014-03-24 Ulrich Drepper <drepper@gmail.com>
5004 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
5007 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
5008 (_mm256_undefined_ps): Define.
5009 (_mm256_undefined_pd): Define.
5010 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
5011 (_mm_undefined_pd): Define.
5012 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
5013 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
5014 (_mm512_undefined_ps): Define.
5015 (_mm512_undefined_pd): Define.
5016 Use _mm*_undefined_*.
5017 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
5019 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
5021 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
5022 (lshr_simd): DI mode added.
5023 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
5024 (aarch64_ushr_simddi): Likewise.
5025 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
5026 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
5027 (vshrd_n_u64): Likewise.
5029 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5031 * Makefile.in (s-macro_list): Depend on cc1.
5033 2014-03-23 Teresa Johnson <tejohnson@google.com>
5035 * ipa-utils.c (ipa_print_order): Use specified dump file.
5037 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
5039 PR rtl-optimization/60601
5040 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
5042 * gcc.c (eval_spec_function): Initialize save_growing_value.
5044 2014-03-22 Jakub Jelinek <jakub@redhat.com>
5047 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
5048 code == MINUS_EXPR, never swap op0 with op1.
5050 * toplev.c (init_local_tick): Avoid signed integer multiplication
5052 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
5053 shift by first operand's bitsize.
5055 2014-03-21 Jakub Jelinek <jakub@redhat.com>
5058 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
5060 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
5061 TARGET_ISA_64BIT_P(x).
5063 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5065 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
5066 pattern for vector nor instead of subtract from splat(-1).
5067 (altivec_expand_vec_perm_const_le): Likewise.
5069 2014-03-21 Richard Henderson <rth@twiddle.net>
5072 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
5073 related insns after epilogue_completed.
5075 2014-03-21 Martin Jambor <mjambor@suse.cz>
5078 * cgraph.h (symtab_node): New flag body_removed.
5079 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
5080 when removing bodies.
5081 * symtab.c (dump_symtab_base): Dump body_removed flag.
5082 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
5083 had their bodies removed.
5085 2014-03-21 Martin Jambor <mjambor@suse.cz>
5088 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
5091 2014-03-21 Richard Biener <rguenther@suse.de>
5093 PR tree-optimization/60577
5094 * tree-core.h (struct tree_base): Document nothrow_flag use
5095 in VAR_DECL_NONALIASED.
5096 * tree.h (VAR_DECL_NONALIASED): New.
5097 (may_be_aliased): Adjust.
5098 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
5100 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5102 * expr.c (expand_expr_real_1): Remove outdated comment.
5104 2014-03-20 Jakub Jelinek <jakub@redhat.com>
5107 * ira.c (adjust_cleared_regs): Call copy_rtx on
5108 *reg_equiv[REGNO (loc)].src_p before passing it to
5109 simplify_replace_fn_rtx.
5112 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
5113 into CONST, put pic register as first operand of PLUS. Use
5114 gen_const_mem for both 32-bit and 64-bit PIC got loads.
5116 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5118 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
5120 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5122 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
5123 around for store forwarding issue in the FPU on the UT699.
5124 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
5125 loads and operations if -mfix-ut699 is specified.
5126 (divtf3_hq): Tweak attribute.
5127 (sqrttf2_hq): Likewise.
5129 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5131 * calls.c (store_one_arg): Remove incorrect const qualification on the
5132 type of the temporary.
5133 * cfgexpand.c (expand_return): Likewise.
5134 * expr.c (expand_constructor): Likewise.
5135 (expand_expr_real_1): Likewise.
5137 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5139 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
5142 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
5145 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
5147 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
5149 * config/arm/aarch-common-protos.h
5150 (alu_cost_table): Fix spelling of "extend".
5151 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
5153 2014-03-19 Richard Biener <rguenther@suse.de>
5156 * tree-core.h (tree_type_common): Re-order pointer members
5157 to reduce recursion depth during GC walks.
5159 2014-03-19 Marek Polacek <polacek@redhat.com>
5162 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
5163 before accessing it.
5165 2014-03-19 Richard Biener <rguenther@suse.de>
5168 * lto-streamer-in.c (input_function): In WPA stage do not drop
5171 2014-03-19 Jakub Jelinek <jakub@redhat.com>
5173 PR tree-optimization/60559
5174 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
5175 with build_zero_cst assignment.
5177 2014-03-18 Kai Tietz <ktietz@redhat.com>
5179 PR rtl-optimization/56356
5180 * sdbout.c (sdbout_parms): Verify that parms'
5181 incoming argument is valid.
5182 (sdbout_reg_parms): Likewise.
5184 2014-03-18 Richard Henderson <rth@redhat.com>
5187 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
5188 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
5189 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
5191 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
5193 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
5194 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
5195 Italicize plugin event names in description. Explain that
5196 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
5197 Remind that no GCC functions should be called after PLUGIN_FINISH.
5198 Explain what pragmas with expansion are.
5200 2014-03-18 Martin Liska <mliska@suse.cz>
5202 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
5203 gimple call statement is update.
5204 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
5205 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
5207 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5210 * ubsan.c (ubsan_instrument_unreachable): Call
5211 initialize_sanitizer_builtins.
5212 (ubsan_pass): Likewise.
5215 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
5216 varpool_finalize_decl instead of rest_of_decl_compilation.
5218 2014-03-18 Richard Biener <rguenther@suse.de>
5220 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
5221 by using bitmap_and_compl instead of bitmap_and_compl_into.
5222 (df_rd_transfer_function): Likewise.
5224 2014-03-18 Richard Biener <rguenther@suse.de>
5226 * doc/lto.texi (fresolution): Fix typo.
5228 2014-03-18 Richard Biener <rguenther@suse.de>
5230 * doc/invoke.texi (flto): Update for changes in 4.9.
5232 2014-03-18 Richard Biener <rguenther@suse.de>
5234 * doc/loop.texi: Remove section on the removed lambda framework.
5235 Update loop docs with recent changes in preserving loop structure.
5237 2014-03-18 Richard Biener <rguenther@suse.de>
5239 * doc/lto.texi (-fresolution): Document.
5241 2014-03-18 Richard Biener <rguenther@suse.de>
5243 * doc/contrib.texi: Adjust my name.
5245 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5248 * internal-fn.c: Include diagnostic-core.h.
5249 (expand_BUILTIN_EXPECT): New function.
5250 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
5251 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
5252 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
5253 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
5255 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
5256 Revert 3 argument __builtin_expect code.
5257 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
5258 * gimple-fold.c (gimple_fold_call): Likewise.
5259 * tree.h (fold_builtin_expect): New prototype.
5260 * builtins.c (build_builtin_expect_predicate): Add predictor
5261 argument, if non-NULL, create 3 argument __builtin_expect.
5262 (fold_builtin_expect): No longer static. Add ARG2 argument,
5263 pass it through to build_builtin_expect_predicate.
5264 (fold_builtin_2): Adjust caller.
5265 (fold_builtin_3): Handle BUILT_IN_EXPECT.
5266 * internal-fn.def (BUILTIN_EXPECT): New.
5268 2014-03-18 Tobias Burnus <burnus@net-b.de>
5271 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
5272 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
5273 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
5275 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
5278 * predict.c (combine_predictions_for_bb): Fix up formatting.
5279 (expr_expected_value_1, expr_expected_value): Add predictor argument,
5280 fill what it points to if non-NULL.
5281 (tree_predict_by_opcode): Adjust caller, use the predictor.
5282 * predict.def (PRED_COMPARE_AND_SWAP): Add.
5284 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
5286 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
5287 proper constant for the store mode.
5289 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
5291 * symtab.c (change_decl_assembler_name): Fix transparent alias
5294 2014-03-16 Renlin Li <Renlin.Li@arm.com>
5296 * config/aarch64/aarch64.c: Correct the comments about the
5297 aarch64 stack layout.
5299 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
5301 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
5302 check for GF_OMP_FOR_KIND_FOR.
5304 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
5306 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
5307 ymm and zmm register names.
5309 2014-03-17 Jakub Jelinek <jakub@redhat.com>
5312 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
5313 note creation for the 2010-08-31 changes.
5315 2014-03-17 Marek Polacek <polacek@redhat.com>
5318 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
5319 as -fno-tree-loop-vectorize.
5320 (expand_omp_simd): Likewise.
5322 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
5324 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
5325 (eligible_for_call_delay): New prototype.
5326 * config/sparc/sparc.c (tls_call_delay): Rename into...
5327 (eligible_for_call_delay): ...this. Return false if the instruction
5328 cannot be put in the delay slot of a branch.
5329 (eligible_for_restore_insn): Simplify.
5330 (eligible_for_return_delay): Return false if the instruction cannot be
5331 put in the delay slot of a branch and simplify.
5332 (eligible_for_sibcall_delay): Return false if the instruction cannot be
5333 put in the delay slot of a branch.
5334 * config/sparc/sparc.md (fix_ut699): New attribute.
5335 (tls_call_delay): Delete.
5336 (in_call_delay): Reimplement.
5337 (eligible_for_sibcall_delay): Rename into...
5338 (in_sibcall_delay): ...this.
5339 (eligible_for_return_delay): Rename into...
5340 (in_return_delay): ...this.
5341 (in_branch_delay): Reimplement.
5342 (in_uncond_branch_delay): Delete.
5343 (in_annul_branch_delay): Delete.
5345 2014-03-14 Richard Henderson <rth@redhat.com>
5348 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
5349 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
5350 (*floathi<X87MODEF>2_i387_with_temp): Remove.
5351 (floathi splitters): Remove.
5352 (float<SWI48x>xf2): New pattern.
5353 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
5354 code that tried to handle DImode for 32-bit, but which was excluded
5355 by the pattern's condition. Drop allocation of stack temporary.
5356 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
5357 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
5358 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
5359 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
5360 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
5361 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
5362 (*float<SWI48><MODEF>2_sse_interunit): Remove.
5363 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
5364 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
5365 (*float<SWI48x><X87MODEF>2_i387): Remove.
5366 (all float _with_temp splitters): Remove.
5367 (*float<SWI48x><MODEF>2_i387): New pattern.
5368 (*float<SWI48><MODEF>2_sse): New pattern.
5369 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
5370 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
5372 2014-03-14 Jakub Jelinek <jakub@redhat.com>
5373 Marek Polacek <polacek@redhat.com>
5376 * common.opt (dump_base_name_prefixed): New Variable.
5377 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
5378 if x_dump_base_name_prefixed is already set, set it at the end.
5380 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
5382 PR rtl-optimization/60508
5383 * lra-constraints.c (get_reload_reg): Add new parameter
5385 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
5386 Pass the new parameter values.
5388 2014-03-14 Richard Biener <rguenther@suse.de>
5390 * common.opt: Revert unintented changes from r205065.
5393 2014-03-14 Richard Biener <rguenther@suse.de>
5396 * cfghooks.c (split_block): Properly adjust all loops the
5397 block was a latch of.
5399 2014-03-14 Martin Jambor <mjambor@suse.cz>
5402 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
5405 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
5408 * config/avr/avr.c (avr_set_current_function): Pass function name
5409 through default_strip_name_encoding before sanity checking instead
5410 of skipping the first char of the assembler name.
5412 2014-03-13 Richard Henderson <rth@redhat.com>
5415 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
5416 (ix86_force_to_memory, ix86_free_from_memory): Remove.
5417 * config/i386/i386-protos.h: Likewise.
5418 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
5419 in the expander instead of a splitter.
5420 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
5421 any possibility of requiring a memory.
5422 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
5423 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
5424 (fp branch splitters): Update for ix86_split_fp_branch.
5425 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
5426 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
5427 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
5428 (*fop_<MODEF>_2_i387): Remove f/r alternative.
5429 (*fop_<MODEF>_3_i387): Likewise.
5430 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
5431 (splitters for the fop_* register patterns): Remove.
5432 (fscalexf4_i387): Rename from *fscalexf4_i387.
5433 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
5435 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5437 PR tree-optimization/59779
5438 * tree-dfa.c (get_ref_base_and_extent): Use double_int
5439 type for bitsize and maxsize instead of HOST_WIDE_INT.
5441 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
5443 PR rtl-optimization/57320
5444 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
5445 the CFG after thread_prologue_and_epilogue_insns.
5447 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
5449 PR rtl-optimization/57189
5450 * lra-constraints.c (process_alt_operands): Disfavor spilling
5453 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
5455 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
5457 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5459 PR tree-optimization/59025
5461 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
5462 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
5464 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
5467 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
5468 calls of avr_out_plus_1.
5470 2014-03-13 Bin Cheng <bin.cheng@arm.com>
5472 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
5473 BB's single pred and update the father loop's latch info later.
5475 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5477 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
5481 (VEC_base): Likewise.
5482 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
5483 registers, we need to swap double words in little endian mode.
5485 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
5486 to be a container mode for 128-bit integer operations added in ISA
5487 2.07. Unlike TImode and PTImode, the preferred register set is
5488 the Altivec/VMX registers for the 128-bit operations.
5490 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
5492 (rs6000_split_128bit_ok_p): Likewise.
5494 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
5495 macros for creating ISA 2.07 normal and overloaded builtin
5496 functions with 3 arguments.
5497 (BU_P8V_OVERLOAD_3): Likewise.
5498 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
5499 for use as overloaded functions.
5500 (VPERM_1TI_UNS): Likewise.
5501 (VSEL_1TI): Likewise.
5502 (VSEL_1TI_UNS): Likewise.
5503 (ST_INTERNAL_1ti): Likewise.
5504 (LD_INTERNAL_1ti): Likewise.
5505 (XXSEL_1TI): Likewise.
5506 (XXSEL_1TI_UNS): Likewise.
5507 (VPERM_1TI): Likewise.
5508 (VPERM_1TI_UNS): Likewise.
5509 (XXPERMDI_1TI): Likewise.
5510 (SET_1TI): Likewise.
5511 (LXVD2X_V1TI): Likewise.
5512 (STXVD2X_V1TI): Likewise.
5513 (VEC_INIT_V1TI): Likewise.
5514 (VEC_SET_V1TI): Likewise.
5515 (VEC_EXT_V1TI): Likewise.
5516 (EQV_V1TI): Likewise.
5517 (NAND_V1TI): Likewise.
5518 (ORC_V1TI): Likewise.
5519 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
5520 added in ISA 2.07. Add both normal 'altivec' builtins, and the
5522 (VADDUQM): Likewise.
5523 (VSUBCUQ): Likewise.
5524 (VADDEUQM): Likewise.
5525 (VADDECUQ): Likewise.
5526 (VSUBEUQM): Likewise.
5527 (VSUBECUQ): Likewise.
5529 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5530 __int128_t and __uint128_t types.
5531 (__uint128_type): Likewise.
5532 (altivec_categorize_keyword): Add support for vector __int128_t,
5533 vector __uint128_t, vector __int128, and vector unsigned __int128
5534 as a container type for TImode operations that need to be done in
5535 VSX/Altivec registers.
5536 (rs6000_macro_to_expand): Likewise.
5537 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5538 to support 128-bit integer instructions vaddcuq, vadduqm,
5539 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5540 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5542 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5543 for V1TImode, and set up preferences to use VSX/Altivec registers.
5544 Setup VSX reload handlers.
5545 (rs6000_debug_reg_global): Likewise.
5546 (rs6000_init_hard_regno_mode_ok): Likewise.
5547 (rs6000_preferred_simd_mode): Likewise.
5548 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5549 (easy_altivec_constant): Likewise.
5550 (output_vec_const_move): Likewise.
5551 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5553 (rs6000_expand_vector_extract): Likewise.
5554 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
5556 (rs6000_const_vec): Add support for V1TImode.
5557 (rs6000_emit_le_vsx_load): Swap double words when loading or
5558 storing TImode/V1TImode.
5559 (rs6000_emit_le_vsx_store): Likewise.
5560 (rs6000_emit_le_vsx_move): Likewise.
5561 (rs6000_emit_move): Add support for V1TImode.
5562 (altivec_expand_ld_builtin): Likewise.
5563 (altivec_expand_st_builtin): Likewise.
5564 (altivec_expand_vec_init_builtin): Likewise.
5565 (altivec_expand_builtin): Likewise.
5566 (rs6000_init_builtins): Add support for V1TImode type. Add
5567 support for ISA 2.07 128-bit integer builtins. Define type names
5568 for the VSX/Altivec vector types.
5569 (altivec_init_builtins): Add support for overloaded vector
5570 functions with V1TImode type.
5571 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
5572 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
5574 (rs6000_split_128bit_ok_p): Likewise.
5575 (rs6000_handle_altivec_attribute): Create V1TImode from vector
5576 __int128_t and vector __uint128_t.
5578 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
5579 and mode attributes.
5586 (VS_scalar): Likewise.
5587 (VS_double): Likewise.
5588 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
5590 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
5591 we support the ISA 2.07 128-bit integer arithmetic instructions.
5592 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
5593 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
5594 and TImode types for use with the builtin functions.
5595 (V1TI_type_node): Likewise.
5596 (unsigned_V1TI_type_node): Likewise.
5597 (intTI_type_internal_node): Likewise.
5598 (uintTI_type_internal_node): Likewise.
5600 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
5601 128-bit builtin functions.
5602 (UNSPEC_VADDEUQM): Likewise.
5603 (UNSPEC_VADDECUQ): Likewise.
5604 (UNSPEC_VSUBCUQ): Likewise.
5605 (UNSPEC_VSUBEUQM): Likewise.
5606 (UNSPEC_VSUBECUQ): Likewise.
5607 (VM): Add V1TImode to vector mode iterators.
5609 (VI_unit): Likewise.
5610 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
5611 (altivec_vaddcuq): Likewise.
5612 (altivec_vsubuqm): Likewise.
5613 (altivec_vsubcuq): Likewise.
5614 (altivec_vaddeuqm): Likewise.
5615 (altivec_vaddecuq): Likewise.
5616 (altivec_vsubeuqm): Likewise.
5617 (altivec_vsubecuq): Likewise.
5619 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
5621 (BOOL_128): Likewise.
5622 (BOOL_REGS_OUTPUT): Likewise.
5623 (BOOL_REGS_OP1): Likewise.
5624 (BOOL_REGS_OP2): Likewise.
5625 (BOOL_REGS_UNARY): Likewise.
5626 (BOOL_REGS_AND_CR0): Likewise.
5628 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
5629 128-bit integer builtin support.
5630 (vec_vadduqm): Likewise.
5631 (vec_vaddecuq): Likewise.
5632 (vec_vaddeuqm): Likewise.
5633 (vec_vsubecuq): Likewise.
5634 (vec_vsubeuqm): Likewise.
5635 (vec_vsubcuq): Likewise.
5636 (vec_vsubuqm): Likewise.
5638 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5639 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
5640 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
5641 128-bit integer add/subtract to ISA 2.07.
5643 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
5645 * config/arc/arc.c (arc_predicate_delay_insns):
5646 Fix third argument passed to conditionalize_nonjump.
5648 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
5650 * config/aarch64/aarch64-builtins.c
5651 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
5652 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
5653 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
5654 instead of __builtin_lfloor.
5655 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
5657 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5659 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
5660 (tree_ssa_ifcombine_bb_1): New function.
5661 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
5662 is an empty forwarder block to then_bb or vice versa and then_bb
5663 and else_bb are effectively swapped.
5665 2014-03-12 Christian Bruel <christian.bruel@st.com>
5668 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
5669 REG_CFA_DEF_CFA note.
5670 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
5671 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
5673 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
5675 PR tree-optimization/60454
5676 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
5678 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5680 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
5681 Do not define target_cpu_default2 to generic.
5682 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
5683 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
5684 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
5686 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5687 Marc Glisse <marc.glisse@inria.fr>
5689 PR tree-optimization/60502
5690 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
5691 instead of build_low_bits_mask.
5693 2014-03-12 Jakub Jelinek <jakub@redhat.com>
5696 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
5697 if there are multiple uses, but op doesn't live on E edge.
5698 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
5699 clobber stmts before __builtin_unreachable.
5701 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
5703 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
5704 hard_frame_pointer_rtx.
5705 * cse.c (cse_insn): Remove volatile check.
5706 * cselib.c (cselib_process_insn): Likewise.
5707 * dse.c (scan_insn): Likewise.
5709 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5711 * config/arc/arc.c (conditionalize_nonjump): New function,
5713 (arc_ifcvt): ... this.
5714 (arc_predicate_delay_insns): Use it.
5716 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
5718 * config/arc/predicates.md (extend_operand): During/after reload,
5719 allow const_int_operand.
5720 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
5721 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
5722 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
5724 (umulsi3_highpart_i): Likewise.
5726 2014-03-11 Richard Biener <rguenther@suse.de>
5728 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
5729 Add asserts to guard possible wrong-code bugs.
5731 2014-03-11 Richard Biener <rguenther@suse.de>
5733 PR tree-optimization/60429
5734 PR tree-optimization/60485
5735 * tree-ssa-structalias.c (set_union_with_increment): Properly
5736 take into account all fields that overlap the shifted vars.
5737 (do_sd_constraint): Likewise.
5738 (do_ds_constraint): Likewise.
5739 (get_constraint_for_ptr_offset): Likewise.
5741 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
5743 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
5744 (nios2_compute_frame_layout):
5745 Add calculation of cfun->machine->fp_save_offset.
5746 (nios2_expand_prologue): Correct setting of frame pointer register
5748 (nios2_expand_epilogue): Update recovery of stack pointer from
5749 frame pointer accordingly.
5750 (nios2_initial_elimination_offset): Update calculation of offset
5751 for eliminating to HARD_FRAME_POINTER_REGNUM.
5753 2014-03-10 Jakub Jelinek <jakub@redhat.com>
5756 * ipa.c (symtab_remove_unreachable_nodes): Don't call
5757 cgraph_get_create_node on VAR_DECLs.
5759 2014-03-10 Richard Biener <rguenther@suse.de>
5762 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
5764 2014-03-08 Douglas B Rupp <rupp@gnat.com>
5766 * config/vms/vms.opt (vms_float_format): New variable.
5768 2014-03-08 Tobias Burnus <burnus@net-b.de>
5770 * doc/invoke.texi (-fcilkplus): Update implementation status.
5772 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
5773 Richard Biener <rguenther@suse.de>
5775 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
5776 consistently accross all TUs.
5777 (run_gcc): Enable -fshort-double automatically at link at link-time
5778 and disallow override.
5780 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
5783 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
5784 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
5785 if they can't be used.
5787 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5789 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
5790 for Solaris 11/x86 ld.
5791 * configure: Regenerate.
5793 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5795 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
5796 (LIB_TLS_SPEC): Save as ld_tls_libs.
5797 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
5798 (HAVE_AS_IX86_TLSLDM): New test.
5799 * configure, config.in: Regenerate.
5800 * config/i386/i386.c (legitimize_tls_address): Fall back to
5801 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
5802 cannot support TLS_MODEL_LOCAL_DYNAMIC.
5803 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
5804 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
5806 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
5808 * common.opt (fira-loop-pressure): Mark as optimization.
5810 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
5812 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
5813 an OpenMP mappable type.
5815 2014-03-06 Matthias Klose <doko@ubuntu.com>
5817 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
5818 MULTILIB_OSDIRNAMES is not defined.
5820 2014-03-06 Jakub Jelinek <jakub@redhat.com>
5821 Meador Inge <meadori@codesourcery.com>
5824 * config/arm/arm.c (arm_tls_symbol_p): Remove.
5825 (arm_legitimize_address): Call legitimize_tls_address for any
5826 arm_tls_referenced_p expression, handle constant addend. Call it
5827 before testing for !TARGET_ARM.
5828 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
5830 2014-03-06 Richard Biener <rguenther@suse.de>
5836 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5838 * tree-streamer.c (record_common_node): Assert we don't record
5839 nodes with type double.
5840 (preload_common_node): Skip type double, complex double and double
5841 pointer since it is now frontend dependent due to fshort-double option.
5843 2014-03-06 Richard Biener <rguenther@suse.de>
5845 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
5846 or -fno-lto is specified and the linker has full plugin support.
5847 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
5848 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
5849 * lto-wrapper.c (merge_and_complain): Merge compile-time
5850 optimization levels.
5851 (run_gcc): And pass it through to the link options.
5853 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
5857 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
5859 * cselib.c (remove_useless_values): Skip to avoid quadratic
5860 behavior if the condition moved from...
5861 (cselib_process_insn): ... here holds.
5863 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5866 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
5867 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
5870 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
5871 (TM_H): Add x86-tune.def.
5873 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5875 * config/aarch64/aarch64.c (generic_tunings):
5876 Use cortexa57_extra_costs.
5878 2014-03-05 Jakub Jelinek <jakub@redhat.com>
5881 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
5882 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
5883 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
5886 2014-03-04 Heiher <r@hev.cc>
5888 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
5889 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
5891 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
5893 * config/i386/predicates.md (const2356_operand): Change to ...
5894 (const2367_operand): ... this.
5895 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
5897 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5898 (*avx512pf_scatterpf<mode>sf): Ditto.
5899 (avx512pf_scatterpf<mode>df): Ditto.
5900 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5901 (*avx512pf_scatterpf<mode>df): Ditto.
5902 * config/i386/i386.c (ix86_expand_builtin): Update
5903 incorrect hint operand error message.
5905 2014-03-04 Richard Biener <rguenther@suse.de>
5907 * lto-section-in.c (lto_get_section_data): Fix const cast.
5909 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
5911 * tree-streamer.c (record_common_node): Assert we don't record
5912 nodes with type double.
5913 (preload_common_node): Skip type double, complex double and double
5914 pointer since it is now frontend dependent due to fshort-double option.
5916 2014-03-04 Richard Biener <rguenther@suse.de>
5919 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
5920 (lto_input_toplevel_asms): Likewise.
5921 * lto-section-in.c (lto_get_section_data): Instead do it here
5924 2014-03-04 Richard Biener <rguenther@suse.de>
5926 PR tree-optimization/60382
5927 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
5928 dead PHIs a reduction.
5930 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
5932 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
5934 (_mm_prefetch): Move out of GCC target("sse") pragma.
5935 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
5936 GCC target("prfchw") pragma.
5937 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
5939 * config/i386/i386.c (ix86_option_override_internal): Enable
5940 -mprfchw with -mprefetchwt1.
5942 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5944 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
5947 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
5949 * opts.h (CL_PCH_IGNORE): Define.
5950 * targhooks.c (option_affects_pch_p):
5951 Return false for options that have CL_PCH_IGNORE set.
5952 * opt-functions.awk: Process PchIgnore.
5953 * doc/options.texi: Document PchIgnore.
5955 * config/arc/arc.opt (misize): Add PchIgnore property.
5957 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5959 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
5960 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
5961 constraint on constants to permit them being loaded into
5962 GENERAL_REGS or BASE_REGS.
5964 2014-03-03 Nick Clifton <nickc@redhat.com>
5966 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
5967 anti-cacnonical alternatives.
5968 (negandhi3_real): New pattern.
5969 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
5971 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5973 * config/avr/avr-mcus.def: Remove atxmega16x1.
5974 * config/avr/avr-tables.opt: Regenerate.
5975 * config/avr/t-multilib: Regenerate.
5976 * doc/avr-mmcu.texi: Regenerate.
5978 2014-03-03 Tobias Grosser <tobias@grosser.es>
5979 Mircea Namolaru <mircea.namolaru@inria.fr>
5981 PR tree-optimization/58028
5982 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
5985 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5987 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
5988 not handled by recognizers.
5990 2014-03-03 Jakub Jelinek <jakub@redhat.com>
5993 * function.c (expand_function_end): Don't emit
5994 clobber_return_register sequence if clobber_after is a BARRIER.
5995 * cfgexpand.c (construct_exit_block): Append instructions before
5996 return_label to prev_bb.
5998 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6000 * config/rs6000/constraints.md: Document reserved use of "wc".
6002 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
6005 * ipa.c (function_and_variable_visibility): When dissolving comdat
6006 group, also set all symbols to local.
6008 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
6013 2013-12-14 Jan Hubicka <jh@suse.cz>
6015 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
6017 2014-03-02 Jon Beniston <jon@beniston.com>
6023 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
6024 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
6025 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
6026 (simple_return, *simple_return): New patterns
6027 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
6028 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
6030 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
6032 * dwarf2out.c (gen_subprogram_die): Tidy.
6034 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
6037 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
6038 (*mov_t_msb_neg_negc): ... this new insn.
6040 2014-02-28 Jason Merrill <jason@redhat.com>
6043 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
6046 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
6049 * dwarf2out.c (decltype_auto_die): New static.
6050 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
6051 (gen_type_die_with_usage): Handle 'decltype(auto)'.
6052 (is_cxx_auto): Likewise.
6054 2014-02-28 Ian Bolton <ian.bolton@arm.com>
6056 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
6057 we are not using general regs only.
6059 2014-02-28 Richard Biener <rguenther@suse.de>
6062 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
6063 previous fix and only allow to remove trivial pre-headers
6064 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
6065 (remove_forwarder_block): Properly update the latch of a loop.
6067 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6070 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
6071 (cselib_preserved_hash_table): New.
6072 (preserve_constants_and_equivs): Move preserved vals to it.
6073 (cselib_find_slot): Look it up first.
6074 (cselib_init): Initialize it.
6075 (cselib_finish): Release it.
6076 (dump_cselib_table): Dump it.
6078 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6081 * cselib.c (remove_useless_values): Skip to avoid quadratic
6082 behavior if the condition moved from...
6083 (cselib_process_insn): ... here holds.
6085 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6088 * var-tracking.c (vt_initialize): Apply the same condition to
6089 preserve the CFA base value.
6091 2014-02-28 Joey Ye <joey.ye@arm.com>
6094 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
6095 if reload in progress or completed.
6097 2014-02-28 Tobias Burnus <burnus@net-b.de>
6100 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
6103 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
6105 * doc/tm.texi.in (Condition Code Status): Update documention for
6106 relative locations of cc0-setter and cc0-user.
6108 2014-02-27 Jeff Law <law@redhat.com>
6110 PR rtl-optimization/52714
6111 * combine.c (try_combine): When splitting an unrecognized PARALLEL
6112 into two independent simple sets, if I3 is a jump, ensure the
6113 pattern we place into I3 is a (set (pc) ...).
6115 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
6116 Jeff Law <law@redhat.com>
6118 PR rtl-optimization/49847
6119 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
6120 are in different blocks.
6121 * doc/tm.texi (Condition Code Status): Update documention for
6122 relative locations of cc0-setter and cc0-user.
6124 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
6127 * lra.c (lra_emit_add): Check SUBREG too.
6129 2014-02-27 Andreas Schwab <schwab@suse.de>
6131 * config/m68k/m68k.c (m68k_option_override): Disable
6132 -flive-range-shrinkage for classic m68k.
6133 (m68k_override_options_after_change): Likewise.
6135 2014-02-27 Marek Polacek <polacek@redhat.com>
6138 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
6139 -Wmaybe-uninitialized.
6141 2014-02-27 Alan Modra <amodra@gmail.com>
6144 * reload1.c (emit_input_reload_insns): When reload_override_in,
6145 set old to rl->in_reg when rl->in_reg is a subreg.
6147 2014-02-26 Richard Biener <rguenther@suse.de>
6150 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
6152 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6154 * common/config/i386/predicates.md (const1256_operand): Remove.
6155 (const2356_operand): New.
6156 (const_1_to_2_operand): Remove.
6157 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6158 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6159 (*avx512pf_gatherpf<mode>sf): Ditto.
6160 (avx512pf_gatherpf<mode>df): Ditto.
6161 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6162 (*avx512pf_gatherpf<mode>df): Ditto.
6163 (avx512pf_scatterpf<mode>sf): Ditto.
6164 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6165 (*avx512pf_scatterpf<mode>sf): Ditto.
6166 (avx512pf_scatterpf<mode>df): Ditto.
6167 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6168 (*avx512pf_scatterpf<mode>df): Ditto.
6169 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
6171 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
6173 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
6174 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
6175 (_mm512_mask_testn_epi64_mask): Move to ...
6176 * config/i386/avx512cdintrin.h: Here.
6177 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
6178 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
6179 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
6180 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
6181 TARGET_AVX512F from TARGET_AVX512CD.
6183 2014-02-26 Richard Biener <rguenther@suse.de>
6186 * ipa.c (walk_polymorphic_call_targets): Properly guard
6187 call to inline_update_overall_summary.
6189 2014-02-26 Bin Cheng <bin.cheng@arm.com>
6192 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
6193 and latches only if requested. Fix latch if it is removed.
6194 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
6195 LOOPS_HAVE_PREHEADERS.
6197 2014-02-25 Andrew Pinski <apinski@cavium.com>
6199 * builtins.c (expand_builtin_thread_pointer): Create a new target
6200 when the target is NULL.
6202 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
6204 PR rtl-optimization/60317
6205 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6206 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6207 * lra-assigns.c: Include params.h.
6208 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
6209 other reload pseudos considerations.
6211 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6213 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
6214 to use canonical form for nor<mode>3.
6216 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6219 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
6222 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6224 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
6225 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
6226 (ix86_handle_option): Handle OPT_mprefetchwt1.
6227 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
6228 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
6230 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
6231 OPTION_MASK_ISA_PREFETCHWT1.
6232 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
6233 (PTA_PREFETCHWT1): New.
6234 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
6235 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
6236 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
6237 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
6238 (*prefetch_avx512pf_<mode>_: Change into ...
6239 (*prefetch_prefetchwt1_<mode>: This.
6240 * config/i386/i386.opt (mprefetchwt1): New.
6241 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
6242 (_mm_prefetch): Handle intent to write.
6243 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
6245 2014-02-25 Richard Biener <rguenther@suse.de>
6248 * emit-rtl.c (mem_attrs_htab): Remove.
6249 (mem_attrs_htab_hash): Likewise.
6250 (mem_attrs_htab_eq): Likewise.
6251 (set_mem_attrs): Always allocate new mem-attrs when something changed.
6252 (init_emit_once): Do not allocate mem_attrs_htab.
6254 2014-02-25 Richard Biener <rguenther@suse.de>
6257 * lto-opts.c (lto_write_options): Output non-explicit conservative
6258 -fwrapv, -fno-trapv and -fno-strict-overflow.
6259 * lto-wrapper.c (merge_and_complain): Handle merging those options.
6260 (run_gcc): And pass them through.
6262 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6264 * sel-sched.c (calculate_new_fences): New parameter ptime.
6265 Calculate it as a maximum over all fence cycles.
6266 (sel_sched_region_2): Adjust the call to calculate_new_fences.
6267 Print the final schedule timing when sched_verbose.
6269 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6271 PR rtl-optimization/60292
6272 * sel-sched.c (fill_vec_av_set): Do not reset target availability
6273 bit fot the fence instruction.
6275 2014-02-24 Alangi Derick <alangiderick@gmail.com>
6277 * calls.h: Fix typo in comment.
6279 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
6281 * config/pa/pa.c (pa_output_move_double): Don't valididate when
6282 adjusting offsetable addresses.
6284 2014-02-24 Guozhi Wei <carrot@google.com>
6286 * sparseset.h (sparseset_pop): Fix the wrong index.
6288 2014-02-24 Walter Lee <walt@tilera.com>
6290 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
6291 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
6293 * common/config/tilegx/tilegx-common.c
6294 (TARGET_DEFAULT_TARGET_FLAGS): Define.
6295 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
6297 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
6298 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
6299 (tilegx_gimplify_va_arg_expr): Handle big endian.
6300 (tilegx_expand_unaligned_load): Ditto.
6301 (tilegx_expand_unaligned_store): Ditto.
6302 (TARGET_RETURN_IN_MSB): New.
6303 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
6304 (TARGET_ENDIAN_DEFAULT): New.
6305 (TARGET_BIG_ENDIAN): Handle big endian.
6306 (BYTES_BIG_ENDIAN): Ditto.
6307 (WORDS_BIG_ENDIAN): Ditto.
6308 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
6311 * config/tilegx/tilegx.md (extv): Handle big endian.
6313 (insn_st<n>): Ditto.
6314 (insn_st<n>_add<bitsuffix>): Ditto.
6315 (insn_stnt<n>): Ditto.
6316 (insn_stnt<n>_add<bitsuffix>):Ditto.
6317 (vec_interleave_highv8qi): Handle big endian.
6318 (vec_interleave_highv8qi_be): New.
6319 (vec_interleave_highv8qi_le): New.
6320 (insn_v1int_h): Handle big endian.
6321 (vec_interleave_lowv8qi): Handle big endian.
6322 (vec_interleave_lowv8qi_be): New.
6323 (vec_interleave_lowv8qi_le): New.
6324 (insn_v1int_l): Handle big endian.
6325 (vec_interleave_highv4hi): Handle big endian.
6326 (vec_interleave_highv4hi_be): New.
6327 (vec_interleave_highv4hi_le): New.
6328 (insn_v2int_h): Handle big endian.
6329 (vec_interleave_lowv4hi): Handle big endian.
6330 (vec_interleave_lowv4hi_be): New.
6331 (vec_interleave_lowv4hi_le): New.
6332 (insn_v2int_l): Handle big endian.
6333 (vec_interleave_highv2si): Handle big endian.
6334 (vec_interleave_highv2si_be): New.
6335 (vec_interleave_highv2si_le): New.
6336 (insn_v4int_h): Handle big endian.
6337 (vec_interleave_lowv2si): Handle big endian.
6338 (vec_interleave_lowv2si_be): New.
6339 (vec_interleave_lowv2si_le): New.
6340 (insn_v4int_l): Handle big endian.
6341 * config/tilegx/tilegx.opt (mbig-endian): New option.
6342 (mlittle-endian): New option.
6343 * doc/install.texi: Document tilegxbe-linux.
6344 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
6346 2014-02-24 Martin Jambor <mjambor@suse.cz>
6349 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
6350 there are no parameter descriptors.
6352 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
6354 PR rtl-optimization/60268
6355 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
6356 initialization to ...
6357 (sched_rgn_init): ... here.
6358 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
6360 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6362 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
6365 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
6367 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
6370 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6372 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
6373 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
6375 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6377 * config/microblaze/predicates.md: Add cmp_op predicate.
6378 * config/microblaze/microblaze.md: Add branch_compare instruction
6379 which uses cmp_op predicate and emits cmp insn before branch.
6380 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
6381 to microblaze_expand_conditional_branch and consolidate logic.
6382 (microblaze_expand_conditional_branch): emit branch_compare
6383 insn instead of handling cmp op separate from branch insn.
6385 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6387 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
6390 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6392 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
6393 define_insn with define_expand and new define_insn
6394 *altivec_lve<VI_char>x_internal.
6395 (altivec_stve<VI_char>x): Replace define_insn with define_expand
6396 and new define_insn *altivec_stve<VI_char>x_internal.
6397 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
6399 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
6401 (altivec_expand_stvex_be): New function.
6403 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
6405 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
6406 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
6407 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
6408 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
6410 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
6413 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
6414 instead of emit_move_insn.
6416 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6418 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
6420 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
6421 gen_altivec_vsumsws.
6423 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6425 * config/rs6000/altivec.md (altivec_lvxl): Rename as
6426 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
6427 (altivec_lvxl_<mode>): New define_expand incorporating
6428 -maltivec=be semantics where needed.
6429 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
6430 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
6431 semantics where needed.
6432 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
6433 (altivec_stvx_<mode>): New define_expand incorporating
6434 -maltivec=be semantics where needed.
6435 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
6436 VM2 iterator instead of V4SI.
6437 (altivec_stvxl_<mode>): New define_expand incorporating
6438 -maltivec=be semantics where needed.
6439 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
6440 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
6441 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
6442 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
6443 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
6444 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
6445 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
6446 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
6447 ALTIVEC_BUILTIN_STVXL.
6448 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
6449 (altivec_expand_stvx_be): Likewise.
6450 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
6451 (altivec_expand_lvx_be): Likewise.
6452 (altivec_expand_stvx_be): Likewise.
6453 (altivec_expand_builtin): Add cases for
6454 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
6455 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
6456 (altivec_init_builtins): Add definitions for
6457 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
6458 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
6460 2014-02-21 Catherine Moore <clm@codesourcery.com>
6462 * doc/invoke.texi (mvirt, mno-virt): Document.
6463 * config/mips/mips.opt (mvirt): New option.
6464 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
6466 2014-02-21 Richard Biener <rguenther@suse.de>
6468 PR tree-optimization/60276
6469 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
6470 (STMT_VINFO_MIN_NEG_DIST): New macro.
6471 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
6472 STMT_VINFO_MIN_NEG_DIST.
6473 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
6474 made for negative dependence distances still hold.
6476 2014-02-21 Richard Biener <rguenther@suse.de>
6479 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
6480 DECL_INITIAL for globals not in the current function context.
6482 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6484 PR tree-optimization/56490
6485 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
6486 * tree-ssa-uninit.c: Include params.h.
6487 (compute_control_dep_chain): Add num_calls argument, return false
6488 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
6489 num_calls to recursive call.
6490 (find_predicates): Change dep_chain into normal array,
6491 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
6492 variable and adjust compute_control_dep_chain caller.
6493 (find_def_preds): Likewise.
6495 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
6497 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
6498 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
6500 2014-02-21 Nick Clifton <nickc@redhat.com>
6502 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
6503 (pushhi1): Likewise.
6504 (popqi1): Add mode to pre_dec.
6507 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6509 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
6510 mode for mask of V8SFmode permutation.
6512 2014-02-20 Richard Henderson <rth@redhat.com>
6515 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
6516 a new pseudo for OLDVAL.
6518 2014-02-20 Jakub Jelinek <jakub@redhat.com>
6521 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
6522 gen_reg_rtx if d->testing_p.
6523 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
6524 if d->testing_p and we will certainly return true.
6525 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6528 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6530 * emit-rtl.c (gen_reg_rtx): Assert that
6531 crtl->emit.regno_pointer_align_length is non-zero.
6533 2014-02-20 Richard Henderson <rth@redhat.com>
6536 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6537 on failure the store back into EXPECT.
6539 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6540 Sandra Loosemore <sandra@codesourcery.com>
6542 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6543 * config/nios2/nios2.c (nios2_function_profiler): Add
6544 -fPIC (flag_pic == 2) support.
6545 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6546 (nios2_large_offset_p): New function.
6547 (nios2_unspec_reloc_p): Move up position, update to use
6548 nios2_large_offset_p.
6549 (nios2_unspec_address): Remove function.
6550 (nios2_unspec_offset): New function.
6551 (nios2_large_got_address): New function.
6552 (nios2_got_address): Add large offset support.
6553 (nios2_legitimize_tls_address): Update usage of removed and new
6555 (nios2_symbol_binds_local_p): New function.
6556 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
6557 (nios2_legitimize_address): Update to use nios2_large_offset_p.
6558 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
6559 (nios2_print_operand): Merge H/L processing, add hiadj/lo
6560 processing for (const (unspec ...)).
6561 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
6563 2014-02-20 Richard Biener <rguenther@suse.de>
6565 * tree-cfg.c (replace_uses_by): Mark altered BBs before
6566 doing the substitution.
6567 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
6569 2014-02-20 Martin Jambor <mjambor@suse.cz>
6572 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
6573 info when checking whether lattices are bottom.
6575 2014-02-20 Richard Biener <rguenther@suse.de>
6578 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
6581 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
6584 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
6585 parameter specifying the scaling.
6586 (inline_call): Update.
6587 (want_inline_recursively): Guard division by zero.
6588 (recursive_inlining): Update.
6589 * ipa-inline.h (clone_inlined_nodes): Update.
6591 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6594 * config/i386/i386.c (classify_argument): Pass structures of size
6595 64 bytes or less in register.
6597 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
6598 Kirill Yukhin <kirill.yukhin@intel.com>
6600 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
6601 (_mm_rcp28_round_ss): Ditto.
6602 (_mm_rsqrt28_round_sd): Ditto.
6603 (_mm_rsqrt28_round_ss): Ditto.
6604 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
6605 (_mm_rcp14_round_ss): Ditto.
6606 (_mm_rsqrt14_round_sd): Ditto.
6607 (_mm_rsqrt14_round_ss): Ditto.
6608 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
6609 the first input operand, get rid of match_dup.
6610 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
6612 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
6614 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
6615 operand as the first input operand, set type attribute.
6616 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
6618 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
6619 operand as the first input operand, set type attribute.
6621 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6623 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
6626 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
6629 * config/i386/i386.c (construct_container): Remove TFmode check
6630 for X86_64_INTEGER_CLASS.
6632 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
6635 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
6636 only when -Wpsabi is enabled.
6638 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
6641 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
6642 passing arrays in registers are the same as for structs, so remove the
6643 special case for them.
6645 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
6647 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
6648 destination type, extract only the valid bits if the source type is not
6649 integral and has a different mode.
6651 2014-02-19 Richard Biener <rguenther@suse.de>
6654 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
6657 2014-02-19 Richard Biener <rguenther@suse.de>
6660 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
6661 (ipa_modify_call_arguments): Emit an argument load explicitely and
6662 preserve virtual SSA form there and for the replacement call.
6663 Do not update SSA form nor free dominance info.
6665 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6667 * ipa.c (function_and_variable_visibility): Also clear WEAK
6668 flag when disolving COMDAT_GROUP.
6670 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6672 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
6673 * ipa-prop.c (ipa_set_jf_known_type): Return early when
6675 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
6676 do more sanity checks.
6677 (detect_type_change): Return true when giving up early.
6678 (compute_complex_assign_jump_func): Fix type parameter of
6679 ipa_set_ancestor_jf.
6680 (compute_complex_ancestor_jump_func): Likewise.
6681 (update_jump_functions_after_inlining): Fix updating of
6683 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
6685 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
6687 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
6688 inline clones when edge disappears.
6690 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
6693 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
6694 Split 64-bit moves into 2 patterns. Do not allow the use of
6695 direct move for TDmode in little endian, since the decimal value
6696 has little endian bytes within a word, but the 64-bit pieces are
6697 ordered in a big endian fashion, and normal subreg's of TDmode are
6699 (mov<mode>_64bit_dm): Likewise.
6700 (movtd_64bit_nodm): Likewise.
6702 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6704 PR tree-optimization/60174
6705 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
6706 statement of an SSA_NAME that occurs in an abnormal PHI node.
6708 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6711 * final.c (SEEN_BB): Remove.
6712 (SEEN_NOTE, SEEN_EMITTED): Renumber.
6713 (final_scan_insn): Don't force_source_line on second
6714 NOTE_INSN_BASIC_BLOCK.
6716 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
6719 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
6720 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
6721 (type_natural_mode): Warn ABI change when %zmm register is not
6722 available for AVX512F vector value passing.
6724 2014-02-18 Kai Tietz <ktietz@redhat.com>
6727 * config/i386/i386.c (ix86_expand_prologue): Use value in
6728 rax register as displacement when restoring %r10 or %rax.
6729 Fix wrong offset when restoring both registers.
6731 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
6733 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
6734 assertion with conditional return.
6736 2014-02-18 Jakub Jelinek <jakub@redhat.com>
6737 Uros Bizjak <ubizjak@gmail.com>
6740 * config/i386/driver-i386.c (host_detect_local_cpu): If
6741 YMM state is not saved by the OS, also clear has_f16c. Move
6742 CPUID 0x80000001 handling before YMM state saving checking.
6744 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
6746 PR rtl-optimization/58960
6747 * haifa-sched.c (alloc_global_sched_pressure_data): New,
6748 factored out from ...
6749 (sched_init): ... here.
6750 (free_global_sched_pressure_data): New, factored out from ...
6751 (sched_finish): ... here.
6752 * sched-int.h (free_global_sched_pressure_data): Declare.
6753 * sched-rgn.c (nr_regions_initial): New static global.
6754 (haifa_find_rgns): Initialize it.
6755 (schedule_region): Disable sched-pressure for the newly
6758 2014-02-17 Richard Biener <rguenther@suse.de>
6760 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
6761 release SSA defs of pattern stmts.
6763 2014-02-17 Richard Biener <rguenther@suse.de>
6765 * tree-inline.c (expand_call_inline): Release the virtual
6766 operand defined by the call we are about to inline.
6768 2014-02-17 Richard Biener <rguenther@suse.de>
6770 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
6772 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
6773 Ilya Tocar <ilya.tocar@intel.com>
6775 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
6776 arguments order in builtin.
6777 (_mm512_permutexvar_epi64): Ditto.
6778 (_mm512_mask_permutexvar_epi64): Ditto
6779 (_mm512_maskz_permutexvar_epi32): Ditto
6780 (_mm512_permutexvar_epi32): Ditto
6781 (_mm512_mask_permutexvar_epi32): Ditto
6783 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6785 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
6786 (p8_vmrgow): Likewise.
6788 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6790 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
6793 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
6796 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
6797 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
6798 into 64-bit and 32-bit moves. On 64-bit moves, add support for
6799 using direct move instructions on ISA 2.07. Also adjust
6800 instruction length for 64-bit.
6801 (mov<mode>_64bit, TFmode/TDmode): Likewise.
6802 (mov<mode>_32bit, TFmode/TDmode): Likewise.
6804 2014-02-15 Alan Modra <amodra@gmail.com>
6808 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
6809 find_replacement on parts of insn rtl that might be reloaded.
6811 2014-02-15 Richard Biener <rguenther@suse.de>
6813 PR tree-optimization/60183
6814 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
6815 (tree_ssa_phiprop): Calculate and free post-dominators.
6817 2014-02-14 Jeff Law <law@redhat.com>
6819 PR rtl-optimization/60131
6820 * ree.c (get_extended_src_reg): New function.
6821 (combine_reaching_defs): Use it rather than assuming location of REG.
6822 (find_and_remove_re): Verify first operand of extension is
6823 a REG before adding the insns to the copy list.
6825 2014-02-14 Roland McGrath <mcgrathr@google.com>
6827 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
6828 * configure: Regenerated.
6829 * config.in: Regenerated.
6830 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
6831 instead of ASM_SHORT.
6833 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
6834 Richard Earnshaw <rearnsha@arm.com>
6836 PR rtl-optimization/59535
6837 * lra-constraints.c (process_alt_operands): Encourage alternative
6838 when unassigned pseudo class is superset of the alternative class.
6839 (inherit_reload_reg): Don't inherit when optimizing for code size.
6840 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
6841 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
6842 modes not less than 4 for Thumb1.
6844 2014-02-14 Kyle McMartin <kyle@redhat.com>
6847 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
6849 2014-02-14 Richard Biener <rguenther@suse.de>
6851 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
6852 (get_frame_arg): Drop the assert with langhook types_compatible_p.
6853 Do not strip INDIRECT_REFs.
6855 2014-02-14 Richard Biener <rguenther@suse.de>
6858 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
6859 DECL_FUNCTION_SPECIFIC_TARGET.
6860 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
6861 * tree-streamer-out.c (pack_ts_target_option): Remove.
6862 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
6863 (write_ts_function_decl_tree_pointers): Do not stream
6864 DECL_FUNCTION_SPECIFIC_TARGET.
6865 * tree-streamer-in.c (unpack_ts_target_option): Remove.
6866 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
6867 (lto_input_ts_function_decl_tree_pointers): Do not stream
6868 DECL_FUNCTION_SPECIFIC_TARGET.
6870 2014-02-14 Jakub Jelinek <jakub@redhat.com>
6872 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
6873 (get_initial_def_for_induction, vectorizable_induction): Ignore
6874 debug stmts when looking for exit_phi.
6875 (vectorizable_live_operation): Fix up condition.
6877 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6879 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
6880 nreverse() because it changes the content of original tree list.
6882 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6884 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
6885 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
6887 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
6889 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
6890 GNU coding standards.
6892 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6895 * dwarf2out.c (gen_subprogram_die): Don't call
6896 add_calling_convention_attribute if subr_die is old_die.
6898 2014-02-13 Sharad Singhai <singhai@google.com>
6900 * doc/optinfo.texi: Fix order of nodes.
6902 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
6904 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
6905 operands[2], not operands[3].
6907 2014-02-13 Richard Biener <rguenther@suse.de>
6910 * doc/install.texi (ISL): Update recommended version to 0.12.2,
6911 mention the possibility of an in-tree build.
6912 (CLooG): Update recommended version to 0.18.1, mention the
6913 possibility of an in-tree build and clarify that the ISL
6914 bundled with CLooG does not work.
6916 2014-02-13 Jakub Jelinek <jakub@redhat.com>
6919 * expr.c (compress_float_constant): If x is a hard register,
6920 extend into a pseudo and then move to x.
6922 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
6924 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
6925 caused by bad second argument to warning_at() with -mhotpatch and
6926 nested functions (e.g. with gfortran).
6928 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
6930 * opts.c (option_name): Remove "enabled by default" rider.
6932 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
6934 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
6936 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
6937 Uros Bizjak <ubizjak@gmail.com>
6940 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
6941 * configure: Regenerated.
6943 2014-02-12 Richard Biener <rguenther@suse.de>
6945 * vec.c (vec_prefix::calculate_allocation): Move as
6946 inline variant to vec.h.
6947 (vec_prefix::calculate_allocation_1): New out-of-line version.
6948 * vec.h (vec_prefix::calculate_allocation_1): Declare.
6949 (vec_prefix::m_has_auto_buf): Rename to ...
6950 (vec_prefix::m_using_auto_storage): ... this.
6951 (vec_prefix::calculate_allocation): Inline the easy cases
6952 and dispatch to calculate_allocation_1 which doesn't need the
6954 (va_heap::reserve): Use gcc_checking_assert.
6955 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
6956 m_using_auto_storage.
6957 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
6959 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
6960 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
6961 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
6963 2014-02-12 Richard Biener <rguenther@suse.de>
6965 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
6966 when we found a dependence.
6968 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
6970 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
6972 (maybe_fold_stmt): ... into this new function.
6973 * omp-low.c (lower_omp): Update comment.
6975 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
6978 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
6981 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
6983 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
6984 identifiers in comments.
6985 (cortexa53_extra_costs): Likewise.
6986 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
6987 (cortexa7_extra_costs): Likewise.
6988 (cortexa12_extra_costs): Likewise.
6989 (cortexa15_extra_costs): Likewise.
6990 (v7m_extra_costs): Likewise.
6992 2014-02-12 Richard Biener <rguenther@suse.de>
6995 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
6996 of posix_memalign being successful.
6997 (lower_stmt): Restrict lowering of posix_memalign to when
6998 -ftree-bit-ccp is enabled.
7000 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
7002 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
7004 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
7006 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
7008 PR rtl-optimization/60116
7009 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
7010 other_insn once the combination has been validated.
7012 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
7015 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
7017 * ipa-devirt.c: Include demangle.h
7018 (odr_violation_reported): New static variable.
7019 (add_type_duplicate): Update odr_violations.
7020 (maybe_record_node): Add completep parameter; update it.
7021 (record_target_from_binfo): Add COMPLETEP parameter;
7022 update it as needed.
7023 (possible_polymorphic_call_targets_1): Likewise.
7024 (struct polymorphic_call_target_d): Add nonconstruction_targets;
7025 rename FINAL to COMPLETE.
7026 (record_targets_from_bases): Sanity check we found the binfo;
7027 fix COMPLETEP updating.
7028 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
7029 parameter, fix computing of COMPLETEP.
7030 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
7031 at LTO time do demangling.
7032 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
7033 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
7035 (gimple_get_virt_method_for_binfo): Likewise.
7036 * gimple-fold.h (gimple_get_virt_method_for_binfo,
7037 gimple_get_virt_method_for_vtable): Update prototypes.
7039 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
7042 * genautomata.c (add_presence_absence): Fix typo with
7043 {final_}presence_list.
7045 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
7048 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
7049 for VSX/Altivec vectors that land in GPR registers.
7051 2014-02-11 Richard Henderson <rth@redhat.com>
7052 Jakub Jelinek <jakub@redhat.com>
7055 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
7056 around drhs if type conversion to lacc->type is not useless.
7058 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7060 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
7062 (cortex-a57.cortex-a53): Likewise.
7063 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
7065 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7067 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
7070 2014-02-11 Renlin Li <Renlin.Li@arm.com>
7072 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
7073 add_options_for_arm_vfp3.
7075 2014-02-11 Jeff Law <law@redhat.com>
7078 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
7079 object with an undesirable mode.
7081 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7084 * config/i386/sol2-9.h: New file.
7085 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
7086 *-*-solaris2.9*): Use it.
7088 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7090 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
7091 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
7093 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7095 * config/microblaze/microblaze.c: Extend mcpu version format
7097 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
7099 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
7101 2014-02-10 Richard Henderson <rth@redhat.com>
7104 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
7105 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
7106 ms-abi vs -mno-accumulate-outgoing-args.
7107 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
7108 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
7111 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
7114 * cfgexpand.c (expand_asm_operands): Attach source location to
7115 ASM_INPUT rtx objects.
7116 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
7118 2014-02-10 Nick Clifton <nickc@redhat.com>
7120 * config/mn10300/mn10300.c (popcount): New function.
7121 (mn10300_expand_prologue): Include saved registers in stack usage
7124 2014-02-10 Jeff Law <law@redhat.com>
7127 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
7128 when changing the SET_DEST of a prior insn to avoid an input reload.
7130 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7132 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
7133 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
7134 -mcall-openbsd, or -mcall-linux.
7135 (CC1_ENDIAN_BIG_SPEC): Remove.
7136 (CC1_ENDIAN_LITTLE_SPEC): Remove.
7137 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7138 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
7139 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
7140 and %cc1_endian_default.
7141 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7143 2014-02-10 Richard Biener <rguenther@suse.de>
7145 PR tree-optimization/60115
7146 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
7147 MEM_REF handling. Properly verify that the accesses are not
7148 out of the objects bound.
7150 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7152 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
7155 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
7157 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
7158 proper constants and fix formatting.
7159 (possible_polymorphic_call_targets): Fix formatting.
7161 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
7162 Ilya Tocar <ilya.tocar@intel.com>
7164 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
7165 (_mm512_loadu_epi32): Renamed into...
7166 (_mm512_loadu_si512): This.
7167 (_mm512_storeu_epi32): Renamed into...
7168 (_mm512_storeu_si512): This.
7169 (_mm512_maskz_ceil_ps): Removed.
7170 (_mm512_maskz_ceil_pd): Ditto.
7171 (_mm512_maskz_floor_ps): Ditto.
7172 (_mm512_maskz_floor_pd): Ditto.
7173 (_mm512_floor_round_ps): Ditto.
7174 (_mm512_floor_round_pd): Ditto.
7175 (_mm512_ceil_round_ps): Ditto.
7176 (_mm512_ceil_round_pd): Ditto.
7177 (_mm512_mask_floor_round_ps): Ditto.
7178 (_mm512_mask_floor_round_pd): Ditto.
7179 (_mm512_mask_ceil_round_ps): Ditto.
7180 (_mm512_mask_ceil_round_pd): Ditto.
7181 (_mm512_maskz_floor_round_ps): Ditto.
7182 (_mm512_maskz_floor_round_pd): Ditto.
7183 (_mm512_maskz_ceil_round_ps): Ditto.
7184 (_mm512_maskz_ceil_round_pd): Ditto.
7185 (_mm512_expand_pd): Ditto.
7186 (_mm512_expand_ps): Ditto.
7187 * config/i386/i386.c (ix86_builtins): Remove
7188 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
7189 (bdesc_args): Ditto.
7190 * config/i386/predicates.md (const1256_operand): New.
7191 (const_1_to_2_operand): Ditto.
7192 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7193 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7194 (*avx512pf_gatherpf<mode>sf): Ditto.
7195 (avx512pf_gatherpf<mode>df): Ditto.
7196 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7197 (*avx512pf_gatherpf<mode>df): Ditto.
7198 (avx512pf_scatterpf<mode>sf): Ditto.
7199 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7200 (*avx512pf_scatterpf<mode>sf): Ditto.
7201 (avx512pf_scatterpf<mode>df): Ditto.
7202 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7203 (*avx512pf_scatterpf<mode>df): Ditto.
7204 (avx512f_expand<mode>): Removed.
7205 (<shift_insn><mode>3<mask_name>): Change predicate type.
7207 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7209 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
7210 not at the end of datarefs vector use ordered_remove to avoid
7211 reordering datarefs vector.
7214 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
7215 mark local addressable non-static vars as GOVD_PRIVATE
7216 instead of GOVD_LOCAL.
7217 * omp-low.c (lower_omp_for): Move gimple_bind_vars
7218 and BLOCK_VARS of gimple_bind_block to new_stmt rather
7222 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
7223 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
7224 assume_aligned or alloc_align attributes.
7225 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
7226 arguments. Handle also assume_aligned and alloc_align attributes.
7227 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
7228 calls to functions with assume_aligned or alloc_align attributes.
7229 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
7231 2014-02-08 Terry Guo <terry.guo@arm.com>
7233 * doc/invoke.texi: Document ARM -march=armv7e-m.
7235 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7237 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
7238 flag on __cilkrts_rethrow builtin.
7241 * ipa-cp.c (determine_versionability): Fail at -O0
7242 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
7243 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
7246 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7249 * tree-inline.c (copy_forbidden): Fail for
7250 __attribute__((optimize (0))) functions.
7252 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7254 * varpool.c: Include pointer-set.h.
7255 (varpool_remove_unreferenced_decls): Variables in other partitions
7256 will not be output; be however careful to not lose information
7259 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7261 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
7262 lookup in the vtable constructor.
7264 2014-02-07 Jeff Law <law@redhat.com>
7267 * config/m68k/m68k.md (ashldi_extsi): Turn into a
7268 define_insn_and_split.
7270 * ipa-inline.c (inline_small_functions): Fix typos.
7272 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7274 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
7275 (s390_can_use_return_insn): Declare.
7276 * config/s390/s390.h (EPILOGUE_USES): Define.
7277 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
7279 (s390_chunkify_start): Handle return JUMP_LABELs.
7280 (s390_early_mach): Emit a main_pool instruction on the entry edge.
7281 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
7282 (s390_can_use_return_insn): New functions.
7283 (s390_fix_long_loop_prediction): Handle conditional returns.
7284 (TARGET_SET_UP_BY_PROLOGUE): Define.
7285 * config/s390/s390.md (ANY_RETURN): New code iterator.
7286 (*creturn, *csimple_return, return, simple_return): New patterns.
7288 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7290 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
7291 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
7292 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
7293 REG_CFA_RESTORE list when deciding not to restore a register.
7295 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7297 * config/s390/s390.c: Include tree-pass.h and context.h.
7298 (s390_early_mach): New function, split out from...
7299 (s390_emit_prologue): ...here.
7300 (pass_data_s390_early_mach): New pass structure.
7301 (pass_s390_early_mach): New class.
7302 (s390_option_override): Create and register early_mach pass.
7303 Move to end of file.
7305 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7307 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
7308 to match for the exit block.
7310 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7312 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
7313 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
7314 Reject misaligned operands.
7316 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7318 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
7320 2014-02-07 Richard Biener <rguenther@suse.de>
7323 * gimple-low.c (lower_builtin_posix_memalign): New function.
7324 (lower_stmt): Call it to lower posix_memalign in a way
7325 to make alignment info accessible.
7327 2014-02-07 Jakub Jelinek <jakub@redhat.com>
7330 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
7331 __builtin_setjmp_receiver.
7333 2014-02-07 Richard Biener <rguenther@suse.de>
7336 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
7337 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
7338 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7339 Handle BUILT_IN_POSIX_MEMALIGN.
7340 (find_func_clobbers): Likewise.
7341 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
7342 (call_may_clobber_ref_p_1): Likewise.
7344 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7347 * ipa-devirt.c (record_target_from_binfo): Remove overactive
7350 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7353 * lto-cgraph.c (lto_output_node): Use
7354 symtab_get_symbol_partitioning_class.
7355 (lto_output_varpool_node): likewise.
7356 (symtab_get_symbol_partitioning_class): Move here from
7358 * cgraph.h (symbol_partitioning_class): Likewise.
7359 (symtab_get_symbol_partitioning_class): Declare.
7361 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7363 * ggc.h (ggc_internal_cleared_alloc): New macro.
7364 * vec.h (vec_safe_copy): Handle memory stats.
7365 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
7366 * target-globals.c (save_target_globals): Likewise.
7368 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7371 * expr.c (emit_move_resolve_push): Export; be bit more selective
7372 on when to clear alias set.
7373 * expr.h (emit_move_resolve_push): Declare.
7374 * function.h (struct function): Add tail_call_marked.
7375 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
7376 * config/i386/i386-protos.h (ix86_expand_push): Remove.
7377 * config/i386/i386.md (TImode move expander): De not call
7379 (FP push expanders): Preserve memory attributes.
7380 * config/i386/sse.md (push<mode>1): Remove.
7381 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
7382 (ix86_expand_push): Remove.
7383 * config/i386/mmx.md (push<mode>1): Remove.
7385 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7387 PR rtl-optimization/60030
7388 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
7389 lopart with paradoxical subreg before shifting it up by hprec.
7391 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7393 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
7394 Remove extra newline at end of file.
7395 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
7396 (arm_issue_rate): Handle cortexa57.
7397 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
7398 (cortex-a57.cortex-a53): Likewise.
7400 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7403 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
7404 don't record in REG_FRAME_RELATED_EXPR registers not set in that
7406 (arm_expand_prologue): Adjust all callers.
7407 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
7408 info, registers also at the lowest numbered registers side. Use
7409 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
7413 * var-tracking.c (adjust_mems): Before adding a SET to
7414 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
7416 2014-02-06 Alan Modra <amodra@gmail.com>
7419 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
7420 change SDmode to DDmode when lra_in_progress.
7422 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7425 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
7426 free_data_ref on the dr first, and before goto again also set dr
7427 to the next dr. For simd_lane_access, free old datarefs[i] before
7428 overwriting it. For get_vectype_for_scalar_type failure, don't
7429 free_data_ref if simd_lane_access.
7431 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
7434 * tree.h (opts_for_fn): New inline function.
7435 (opt_for_fn): Define.
7436 * config/i386/i386.c (ix86_function_regparm): Use
7437 opt_for_fn (decl, optimize) instead of optimize.
7439 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
7441 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
7442 for SYMBOL_REF in large memory model.
7444 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7446 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
7448 (cortex-a57): Likewise.
7449 (cortex-a57.cortex-a53): Likewise.
7451 2014-02-06 Yury Gribov <y.gribov@samsung.com>
7452 Kugan Vivekanandarajah <kuganv@linaro.org>
7454 * config/arm/arm.c (arm_vector_alignment_reachable): Check
7456 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
7458 2014-02-06 Richard Biener <rguenther@suse.de>
7460 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
7461 set_loop_copy and initialize_original_copy_tables.
7463 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
7465 * config/aarch64/aarch64-simd.md
7466 (aarch64_ashr_simddi): Change QI to SI.
7468 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7469 Jakub Jelinek <jakub@redhat.com>
7472 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
7475 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7477 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
7478 CODE_FOR_altivec_vpku[hw]um to
7479 CODE_FOR_altivec_vpku[hw]um_direct.
7480 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
7481 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
7482 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
7483 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
7485 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7487 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
7488 generation for -maltivec=be.
7489 (altivec_vsumsws): Simplify redundant test.
7491 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7493 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
7494 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
7495 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
7496 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
7497 gen_altivec_vpkuwum.
7498 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
7500 (altivec_vpks<VI_char>ss): Likewise.
7501 (altivec_vpks<VI_char>us): Likewise.
7502 (altivec_vpku<VI_char>us): Likewise.
7503 (altivec_vpku<VI_char>um): Likewise.
7504 (altivec_vpku<VI_char>um_direct): New (copy of
7505 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
7507 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
7508 target is little endian and -maltivec=be is not specified.
7509 (*altivec_vupkhs<VU_char>_direct): New (copy of
7510 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
7511 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
7512 target is little endian and -maltivec=be is not specified.
7513 (*altivec_vupkls<VU_char>_direct): New (copy of
7514 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
7515 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
7516 little endian and -maltivec=be is not specified.
7517 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
7518 little endian and -maltivec=be is not specified.
7520 2014-02-05 Richard Henderson <rth@redhat.com>
7523 * combine-stack-adj.c: Revert r206943.
7524 * sched-int.h (struct deps_desc): Add last_args_size.
7525 * sched-deps.c (init_deps): Initialize it.
7526 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7527 contain REG_ARGS_SIZE notes.
7529 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7531 * lto-cgraph.c (asm_nodes_output): Make global.
7532 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7533 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7534 (driver_handle_option): Handle OPT_fwpa.
7536 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7539 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7540 a comment typo and formatting issue. If odr_hash hasn't been
7541 created, return vNULL and set *completep to false.
7544 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7545 bb with no successors.
7547 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7550 * doc/invoke.texi (-march): Clarify documentation for ARM.
7554 2014-02-05 Richard Biener <rguenther@suse.de>
7556 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
7557 when not vectorizing because of too many alias checks.
7558 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
7559 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
7561 2014-02-05 Nick Clifton <nickc@redhat.com>
7563 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
7564 accept extended registers in any mode when compiling for the MN10300.
7566 2014-02-05 Yury Gribov <y.gribov@samsung.com>
7568 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
7569 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
7570 sanitization attributes.
7571 (can_inline_edge_p): Likewise.
7572 (sanitize_attrs_match_for_inline_p): New function.
7574 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7576 * ipa-prop.c (detect_type_change): Shor circuit testing of
7577 type changes on THIS pointer.
7579 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
7582 * config/pa/pa.c (legitimize_tls_address): Return original address
7583 if not passed a SYMBOL_REF rtx.
7584 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
7586 (pa_emit_move_sequence): Simplify TLS source operands.
7587 (pa_legitimate_constant_p): Reject all TLS constants.
7588 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
7589 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
7591 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7593 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
7594 groups when we know they are controlled by LTO.
7595 * varasm.c (default_binds_local_p_1): If object is in other partition,
7596 it will be resolved locally.
7598 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7600 * config/host-linux.c (linux_gt_pch_use_address): Don't
7601 use SSIZE_MAX because it is not always defined.
7603 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
7606 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
7607 threshold for pseudo splitting.
7608 (update_ebb_live_info): Process call argument hard registers and
7609 hard registers from insn definition too.
7610 (max_small_class_regs_num): New constant.
7611 (inherit_in_ebb): Update live hard regs through EBBs. Update
7612 reloads_num only for small register classes. Don't split for
7615 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
7618 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
7621 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
7623 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
7626 2014-02-04 Marek Polacek <polacek@redhat.com>
7628 * gdbinit.in (pel): Define.
7630 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
7632 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
7635 2014-02-04 Richard Biener <rguenther@suse.de>
7638 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
7639 in function context local.
7640 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
7641 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
7642 similar to LTO_imported_decl_ref.
7644 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7646 PR tree-optimization/60002
7647 * cgraphclones.c (build_function_decl_skip_args): Clear
7650 PR tree-optimization/60023
7651 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
7652 false to gsi_replace.
7653 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
7654 has been in some EH region and vec_stmt could throw, add
7655 vec_stmt into the same EH region.
7656 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
7657 has no lhs, ignore it.
7658 * internal-fn.c (expand_MASK_LOAD): Likewise.
7661 * tree-inline.c (copy_forbidden): Fail for
7662 __attribute__((optimize (0))) functions.
7665 * omp-low.c (simd_clone_struct_copy): If from->inbranch
7666 is set, copy one less argument.
7667 (expand_simd_clones): Don't subtract clone_info->inbranch
7668 from simd_clone_struct_alloc argument.
7670 PR rtl-optimization/57915
7671 * recog.c (simplify_while_replacing): If all unary/binary/relational
7672 operation arguments are constant, attempt to simplify those.
7675 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
7676 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
7678 2014-02-04 Richard Biener <rguenther@suse.de>
7680 PR tree-optimization/60012
7681 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
7682 TBAA disambiguation to all DDRs.
7684 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7687 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
7688 (LINK_SPEC): Use it for -shared, -shared-libgcc.
7690 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7693 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
7695 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7697 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
7698 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
7700 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7703 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
7704 to figure out targets of polymorphic calls with known decl.
7705 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7706 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
7707 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
7708 (get_polymorphic_call_info): ... here.
7709 (get_polymorphic_call_info_from_invariant): New function.
7711 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7713 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
7714 lookup via vtable pointer; check for type consistency
7715 and turn inconsitent facts into UNREACHABLE.
7716 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7717 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
7718 type inconsistent querries; return UNREACHABLE instead.
7720 2014-02-03 Richard Henderson <rth@twiddle.net>
7723 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
7724 already processed this node.
7725 (normalize_one_pred_1): Pass along mark_set.
7726 (normalize_one_pred): Create and destroy a pointer_set_t.
7727 (normalize_one_pred_chain): Likewise.
7729 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
7731 PR gcov-profile/58602
7732 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
7734 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
7737 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
7738 -fno-devirtualize; try to devirtualize by the knowledge of
7739 virtual table pointer given by aggregate propagation.
7740 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
7741 (ipa_print_node_jump_functions): Dump also offset that
7742 is relevant for polymorphic calls.
7743 (determine_known_aggregate_parts): Add arg_type parameter; use it
7744 instead of determining the type from pointer type.
7745 (ipa_compute_jump_functions_for_edge): Update call of
7746 determine_known_aggregate_parts.
7747 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
7748 (gimple_get_virt_method_for_binfo): ... here; simplify using
7749 vtable_pointer_value_to_vtable.
7750 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
7751 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
7752 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
7753 (vtable_pointer_value_to_vtable): Break out from ...; handle also
7755 (vtable_pointer_value_to_binfo): ... here.
7756 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
7758 2014-02-03 Teresa Johnson <tejohnson@google.com>
7760 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
7761 redef of outer loop index variable.
7763 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
7767 * doc/extend.texi (Function Attributes): Typo.
7769 2014-02-03 Cong Hou <congh@google.com>
7771 PR tree-optimization/60000
7772 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
7773 if the vectorized statement is a store. A store statement can only
7774 appear at the end of pattern statements.
7776 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7778 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
7779 (ix86_option_override_internal): Default long double to 64-bit for
7780 32-bit Bionic and to 128-bit for 64-bit Bionic.
7782 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
7783 TARGET_LONG_DOUBLE_128 is true.
7784 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
7786 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
7787 (mlong-double-64): Negate -mlong-double-128.
7788 (mlong-double-128): New option.
7790 * config/i386/i386-c.c (ix86_target_macros): Define
7791 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
7793 * doc/invoke.texi: Document -mlong-double-128.
7795 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
7797 PR rtl-optimization/60024
7798 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
7800 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
7802 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
7804 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7806 PR rtl-optimization/57662
7807 * sel-sched.c (code_motion_path_driver): Do not mark already not
7808 existing blocks in the visiting bitmap.
7810 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
7812 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
7813 on the insn being emitted.
7815 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
7816 Will Deacon <will.deacon@arm.com>
7818 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
7820 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7822 * config/arm/arm-tables.opt: Regenerate.
7824 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7826 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
7827 for vector types other than V16QImode.
7828 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
7829 define_expand, and call altivec_expand_vec_perm_le when producing
7830 code with little endian element order.
7831 (*altivec_vperm_<mode>_internal): New insn having previous
7832 behavior of altivec_vperm_<mode>.
7833 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
7834 altivec_expand_vec_perm_le when producing code with little endian
7836 (*altivec_vperm_<mode>_uns_internal): New insn having previous
7837 behavior of altivec_vperm_<mode>_uns.
7839 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7841 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
7842 (altivec_vsumsws): Add handling for -maltivec=be with a little
7844 (altivec_vsumsws_direct): New.
7845 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
7846 gen_altivec_vsumsws.
7848 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7850 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
7851 vtable_pointer_value_to_binfo): New functions.
7852 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
7853 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
7855 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
7857 * config/nios2/nios2.md (load_got_register): Initialize GOT
7858 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
7859 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
7861 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
7863 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
7864 preserverd by passthrough, do not propagate the type.
7866 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7868 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
7869 (mips_atomic_assign_expand_fenv): New function.
7870 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
7872 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7874 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
7875 (__builtin_mips_set_fcsr): Likewise.
7876 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
7877 MIPS_USI_FTYPE_VOID.
7878 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
7879 (mips16_expand_set_fcsr): Likewise.
7880 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
7881 (mips16_set_fcsr_stub): Likewise.
7882 (mips16_get_fcsr_one_only_stub): New class.
7883 (mips16_set_fcsr_one_only_stub): Likewise.
7884 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
7885 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
7886 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
7887 (hard_float): New availability predicate.
7888 (mips_builtins): Add get_fcsr and set_fcsr.
7889 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
7890 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
7891 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
7892 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
7893 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
7896 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
7898 * config/mips/mips.c (mips_one_only_stub): New class.
7899 (mips_need_mips16_rdhwr_p): Replace with...
7900 (mips16_rdhwr_stub): ...this new variable.
7901 (mips16_stub_call_address): New function.
7902 (mips16_rdhwr_one_only_stub): New class.
7903 (mips_expand_thread_pointer): Use mips16_stub_call_address.
7904 (mips_output_mips16_rdhwr): Delete.
7905 (mips_finish_stub): New function.
7906 (mips_code_end): Use it to handle rdhwr stubs.
7908 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
7911 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
7912 when calculating size of integer atomic types.
7914 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
7916 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
7918 2014-02-01 Jakub Jelinek <jakub@redhat.com>
7920 PR tree-optimization/60003
7921 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
7922 * profile.c (branch_prob): Use gimple_call_builtin_p
7923 to check for BUILT_IN_SETJMP_RECEIVER.
7924 * tree-inline.c (copy_bb): Call notice_special_calls.
7926 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
7929 * lra-constraints.c (process_alt_operands): Update reload_sum only
7932 2014-01-31 Richard Henderson <rth@redhat.com>
7935 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
7936 until after else_eh is processed.
7938 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7940 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
7941 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
7942 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
7943 in smmintrin.h, remove them.
7944 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
7945 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
7946 * config/i386/i386.md (ROUND_SAE): Fix value.
7947 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
7948 (const48_operand): New.
7949 * config/i386/subst.md (round), (round_expand): Use
7950 const_4_or_8_to_11_operand.
7951 (round_saeonly), (round_saeonly_expand): Use const48_operand.
7953 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
7955 * config/i386/constraints.md (Yk): Swap meaning with k.
7956 * config/i386/i386.md (movhi_internal): Change Yk to k.
7957 (movqi_internal): Ditto.
7958 (*k<logic><mode>): Ditto.
7961 (kandn<mode>): Ditto.
7962 (*<code>hi_1): Ditto.
7963 (*<code>qi_1): Ditto.
7964 (kxnor<mode>): Ditto.
7965 (kortestzhi): Ditto.
7966 (kortestchi): Ditto.
7968 (*one_cmplhi2_1): Ditto.
7969 (*one_cmplqi2_1): Ditto.
7970 * config/i386/sse.md (): Change k to Yk.
7971 (avx512f_load<mode>_mask): Ditto.
7972 (avx512f_blendm<mode>): Ditto.
7973 (avx512f_store<mode>_mask): Ditto.
7974 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
7975 (avx512f_storedqu<mode>_mask): Ditto.
7976 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
7978 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
7979 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
7980 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
7981 (avx512f_maskcmp<mode>3): Ditto.
7982 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
7983 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
7984 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
7985 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
7986 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
7987 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
7988 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
7989 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
7990 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
7991 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
7992 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
7993 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
7994 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
7995 (vec_extract_lo_<mode>_maskm): Ditto.
7996 (vec_extract_hi_<mode>_maskm): Ditto.
7997 (avx512f_vternlog<mode>_mask): Ditto.
7998 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
7999 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
8000 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
8001 (avx512f_<code>v8div16qi2_mask): Ditto.
8002 (avx512f_<code>v8div16qi2_mask_store): Ditto.
8003 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
8004 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
8005 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
8006 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
8007 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8008 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8009 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8010 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8011 (avx512cd_maskb_vec_dupv8di): Ditto.
8012 (avx512cd_maskw_vec_dupv16si): Ditto.
8013 (avx512f_vpermi2var<mode>3_maskz): Ditto.
8014 (avx512f_vpermi2var<mode>3_mask): Ditto.
8015 (avx512f_vpermi2var<mode>3_mask): Ditto.
8016 (avx512f_vpermt2var<mode>3_maskz): Ditto.
8017 (*avx512f_gathersi<mode>): Ditto.
8018 (*avx512f_gathersi<mode>_2): Ditto.
8019 (*avx512f_gatherdi<mode>): Ditto.
8020 (*avx512f_gatherdi<mode>_2): Ditto.
8021 (*avx512f_scattersi<mode>): Ditto.
8022 (*avx512f_scatterdi<mode>): Ditto.
8023 (avx512f_compress<mode>_mask): Ditto.
8024 (avx512f_compressstore<mode>_mask): Ditto.
8025 (avx512f_expand<mode>_mask): Ditto.
8026 * config/i386/subst.md (mask): Change k to Yk.
8027 (mask_scalar_merge): Ditto.
8030 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
8032 * doc/extend.texi (Vector Extensions): Document ?: in C++.
8034 2014-01-31 Richard Biener <rguenther@suse.de>
8037 * builtins.c (fold_builtin_memory_op): Make sure to not
8038 use a floating-point mode or a boolean or enumeral type for
8041 2014-01-30 DJ Delorie <dj@redhat.com>
8043 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
8044 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
8045 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
8046 whenever main() has an epilogue.
8048 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8050 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
8051 unused variable "field".
8052 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
8053 (vsx_mergeh_<mode>): Likewise.
8054 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
8055 (altivec_vmrghh): Likewise.
8056 (altivec_vmrghw): Likewise.
8057 (altivec_vmrglb): Likewise.
8058 (altivec_vmrglh): Likewise.
8059 (altivec_vmrglw): Likewise.
8060 (altivec_vspltb): Add missing uses.
8061 (altivec_vsplth): Likewise.
8062 (altivec_vspltw): Likewise.
8063 (altivec_vspltsf): Likewise.
8065 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8068 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
8069 frame related instructions.
8071 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
8073 PR rtl-optimization/59959
8074 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
8075 any reload of register whose subreg is invalid.
8077 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8079 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
8080 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
8081 Add missing return type - void.
8083 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8085 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
8086 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
8087 remove element index adjustment for endian (now handled in vsx.md
8089 (altivec_expand_vec_perm_const): Use
8090 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
8091 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
8092 (vsx_xxspltw_<mode>): Adjust element index for little endian.
8093 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
8094 define_expand and a new define_insn *altivec_vspltb_internal;
8095 adjust for -maltivec=be on a little endian target.
8096 (altivec_vspltb_direct): New.
8097 (altivec_vsplth): Divide into a define_expand and a new
8098 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
8099 little endian target.
8100 (altivec_vsplth_direct): New.
8101 (altivec_vspltw): Divide into a define_expand and a new
8102 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
8103 little endian target.
8104 (altivec_vspltw_direct): New.
8105 (altivec_vspltsf): Divide into a define_expand and a new
8106 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
8107 a little endian target.
8109 2014-01-30 Richard Biener <rguenther@suse.de>
8111 PR tree-optimization/59993
8112 * tree-ssa-forwprop.c (associate_pointerplus): Check we
8113 can propagate form the earlier stmt and avoid the transform
8114 when the intermediate result is needed.
8116 2014-01-30 Alangi Derick <alangiderick@gmail.com>
8118 * README.Portability: Fix typo.
8120 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
8122 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
8123 comparison_operator with ordered_comparison_operator.
8125 2014-01-30 Nick Clifton <nickc@redhat.com>
8127 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
8128 Rename to mn10300_store_multiple_regs.
8129 * config/mn10300/mn10300.c: Likewise.
8130 * config/mn10300/mn10300.md (store_movm): Fix typo: call
8131 store_multiple_regs.
8132 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
8133 Call mn10300_store_multiple_regs.
8135 2014-01-30 Nick Clifton <nickc@redhat.com>
8136 DJ Delorie <dj@redhat.com>
8138 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
8139 %fp 2 to keep registers after it properly word-aligned.
8140 (rl78_alloc_physical_registers_umul): Handle the case where both
8141 input operands are the same.
8143 2014-01-30 Richard Biener <rguenther@suse.de>
8145 PR tree-optimization/59903
8146 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
8149 2014-01-30 Jason Merrill <jason@redhat.com>
8152 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
8155 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
8157 2014-01-30 Richard Biener <rguenther@suse.de>
8159 PR tree-optimization/59951
8160 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
8162 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
8165 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
8166 SFmode to DFmode case.
8168 2014-01-29 DJ Delorie <dj@redhat.com>
8170 * config/msp430/msp430.opt (-minrt): New.
8171 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
8173 (ENDFILE_SPEC): Likewise.
8175 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
8177 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
8178 (estimate_function_body_sizes): Use it.
8180 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
8183 * dwarf2out.c (is_cxx_auto): New.
8184 (is_base_type): Use it.
8185 (gen_type_die_with_usage): Likewise.
8187 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8189 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
8190 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
8191 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
8192 -maltivec=be with LE targets.
8193 (vsx_mergeh_<mode>): Likewise.
8194 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
8195 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
8196 (altivec_vmrghb): Replace with define_expand and new
8197 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
8198 (altivec_vmrghb_direct): New define_insn.
8199 (altivec_vmrghh): Replace with define_expand and new
8200 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
8201 (altivec_vmrghh_direct): New define_insn.
8202 (altivec_vmrghw): Replace with define_expand and new
8203 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
8204 (altivec_vmrghw_direct): New define_insn.
8205 (*altivec_vmrghsf): Adjust for endianness.
8206 (altivec_vmrglb): Replace with define_expand and new
8207 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
8208 (altivec_vmrglb_direct): New define_insn.
8209 (altivec_vmrglh): Replace with define_expand and new
8210 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
8211 (altivec_vmrglh_direct): New define_insn.
8212 (altivec_vmrglw): Replace with define_expand and new
8213 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
8214 (altivec_vmrglw_direct): New define_insn.
8215 (*altivec_vmrglsf): Adjust for endianness.
8216 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8217 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8218 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8219 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8220 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8221 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8222 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8223 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8225 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
8227 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
8228 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
8231 2014-01-29 Richard Biener <rguenther@suse.de>
8233 PR tree-optimization/58742
8234 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
8235 associate_pointerplus_align.
8236 (associate_pointerplus_diff): New function.
8237 (associate_pointerplus): Likewise. Call associate_pointerplus_align
8238 and associate_pointerplus_diff.
8240 2014-01-29 Richard Biener <rguenther@suse.de>
8242 * lto-streamer.h (LTO_major_version): Bump to 3.
8243 (LTO_minor_version): Reset to 0.
8245 2014-01-29 Renlin Li <Renlin.Li@arm.com>
8247 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
8248 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
8249 (arm_file_start): Generate correct asm header for armv7ve.
8250 * config/arm/bpabi.h: Add multilib support for armv7ve.
8251 * config/arm/driver-arm.c: Change the architectures of cortex-a7
8252 and cortex-a15 to armv7ve.
8253 * config/arm/t-aprofile: Add multilib support for armv7ve.
8254 * doc/invoke.texi: Document -march=armv7ve.
8256 2014-01-29 Richard Biener <rguenther@suse.de>
8258 PR tree-optimization/58742
8259 * tree-ssa-forwprop.c (associate_plusminus): Return true
8260 if we changed sth, defer EH cleanup to ...
8261 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
8262 (simplify_mult): New function.
8264 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8267 PR tree-optimization/59920
8268 * tree.c (build_common_builtin_nodes): Remove
8269 __builtin_setjmp_dispatcher initialization.
8270 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
8271 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
8272 instead of gsi_after_labels + manually skipping debug stmts.
8273 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
8274 ignore bbs with IFN_ABNORMAL_DISPATCHER.
8275 * tree-inline.c (copy_edges_for_bb): Remove
8276 can_make_abnormal_goto argument, instead add abnormal_goto_dest
8277 argument. Ignore computed_goto_p stmts. Don't call
8278 make_abnormal_goto_edges. If a call might need abnormal edges
8279 for non-local gotos, see if it already has an edge to
8280 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
8281 with true argument, don't do anything then, otherwise add
8282 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
8283 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
8285 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
8286 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
8287 (lower_stmt): Don't set data->calls_builtin_setjmp.
8288 (lower_builtin_setjmp): Adjust comment.
8289 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
8290 * tree-cfg.c (found_computed_goto): Remove.
8291 (factor_computed_gotos): Remove.
8292 (make_goto_expr_edges): Return bool, true for computed gotos.
8293 Don't call make_abnormal_goto_edges.
8294 (build_gimple_cfg): Don't set found_computed_goto, don't call
8295 factor_computed_gotos.
8296 (computed_goto_p): No longer static.
8297 (make_blocks): Don't set found_computed_goto.
8298 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
8299 (make_edges): If make_goto_expr_edges returns true, push bb
8300 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
8301 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
8302 vector. Record mapping between bbs and OpenMP regions if there
8303 are any, adjust make_gimple_omp_edges caller. Call
8304 handle_abnormal_edges.
8305 (make_abnormal_goto_edges): Remove.
8306 * tree-cfg.h (make_abnormal_goto_edges): Remove.
8307 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
8308 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
8309 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
8310 * internal-fn.def (ABNORMAL_DISPATCHER): New.
8311 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
8312 filling *region also set *region_idx to (*region)->entry->index.
8315 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
8316 For REGs set ORIGINAL_REGNO.
8318 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
8320 * doc/md.texi: Mention that a target shouldn't implement
8321 vec_widen_(s|u)mul_even/odd pair if it is less efficient
8324 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8326 PR tree-optimization/59594
8327 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
8328 a copy of the datarefs vector rather than the vector itself.
8330 2014-01-28 Jason Merrill <jason@redhat.com>
8333 * dwarf2out.c (auto_die): New static.
8334 (gen_type_die_with_usage): Handle C++1y 'auto'.
8335 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
8338 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
8341 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
8342 (SPEC_X32): Likewise.
8343 (SPEC_64): Likewise.
8344 * config/i386/i386.c (ix86_option_override_internal): Turn off
8345 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
8347 (x86_file_start): Output .code16gcc for TARGET_16BIT.
8348 * config/i386/i386.h (TARGET_16BIT): New macro.
8349 (TARGET_16BIT_P): Likewise.
8350 * config/i386/i386.opt: Add m16.
8351 * doc/invoke.texi: Document -m16.
8353 2014-01-28 Jakub Jelinek <jakub@redhat.com>
8355 PR preprocessor/59935
8356 * input.c (location_get_source_line): Bail out on when line number
8357 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
8359 2014-01-28 Richard Biener <rguenther@suse.de>
8361 PR tree-optimization/58742
8362 * tree-ssa-forwprop.c (associate_plusminus): Handle
8363 pointer subtraction of the form (T)(P + A) - (T)P.
8365 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8367 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
8370 2014-01-28 Richard Biener <rguenther@suse.de>
8373 2014-01-28 Richard Biener <rguenther@suse.de>
8375 PR rtl-optimization/45364
8376 PR rtl-optimization/59890
8377 * var-tracking.c (local_get_addr_clear_given_value): Handle
8378 already cleared slot.
8379 (val_reset): Handle not allocated local_get_addr_cache.
8380 (vt_find_locations): Use post-order on the inverted CFG.
8382 2014-01-28 Richard Biener <rguenther@suse.de>
8384 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
8386 2014-01-28 Richard Biener <rguenther@suse.de>
8388 PR rtl-optimization/45364
8389 PR rtl-optimization/59890
8390 * var-tracking.c (local_get_addr_clear_given_value): Handle
8391 already cleared slot.
8392 (val_reset): Handle not allocated local_get_addr_cache.
8393 (vt_find_locations): Use post-order on the inverted CFG.
8395 2014-01-28 Alan Modra <amodra@gmail.com>
8397 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
8398 * configure.ac <recursive call for build != host>: Define
8399 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
8400 and LD_FOR_BUILD too.
8401 * configure: Regenerate.
8403 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
8405 * config/i386/i386.c (get_builtin_code_for_version): Separate
8406 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
8407 Broadwell from Haswell.
8409 2014-01-27 Steve Ellcey <sellcey@mips.com>
8411 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
8412 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
8413 * config/mips/mips.c (mips_option_override): Change setting
8415 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
8416 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
8417 Change from Mask to Var.
8419 2014-01-27 Jeff Law <law@redhat.com>
8421 * ipa-inline.c (inline_small_functions): Fix typo.
8423 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
8425 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
8426 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
8427 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
8428 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
8429 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
8430 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
8431 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
8432 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
8433 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
8434 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
8435 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
8436 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
8437 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
8438 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
8439 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
8440 (_mm512_storeu_epi64): Ditto.
8441 (_mm512_cmpge_epi32_mask): Ditto.
8442 (_mm512_cmpge_epu32_mask): Ditto.
8443 (_mm512_cmpge_epi64_mask): Ditto.
8444 (_mm512_cmpge_epu64_mask): Ditto.
8445 (_mm512_cmple_epi32_mask): Ditto.
8446 (_mm512_cmple_epu32_mask): Ditto.
8447 (_mm512_cmple_epi64_mask): Ditto.
8448 (_mm512_cmple_epu64_mask): Ditto.
8449 (_mm512_cmplt_epi32_mask): Ditto.
8450 (_mm512_cmplt_epu32_mask): Ditto.
8451 (_mm512_cmplt_epi64_mask): Ditto.
8452 (_mm512_cmplt_epu64_mask): Ditto.
8453 (_mm512_cmpneq_epi32_mask): Ditto.
8454 (_mm512_cmpneq_epu32_mask): Ditto.
8455 (_mm512_cmpneq_epi64_mask): Ditto.
8456 (_mm512_cmpneq_epu64_mask): Ditto.
8457 (_mm512_expand_pd): Ditto.
8458 (_mm512_expand_ps): Ditto.
8459 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
8460 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
8461 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
8462 * config/i386/i386.c (ix86_builtins): Add
8463 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
8464 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
8465 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
8466 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
8467 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
8468 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
8469 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
8470 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
8471 IX86_BUILTIN_PMOVUSQW512_MEM.
8472 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
8473 __builtin_ia32_pmovsqd512mem_mask,
8474 __builtin_ia32_pmovqd512mem_mask,
8475 __builtin_ia32_pmovusqw512mem_mask,
8476 __builtin_ia32_pmovsqw512mem_mask,
8477 __builtin_ia32_pmovqw512mem_mask,
8478 __builtin_ia32_pmovusdw512mem_mask,
8479 __builtin_ia32_pmovsdw512mem_mask,
8480 __builtin_ia32_pmovdw512mem_mask,
8481 __builtin_ia32_pmovqb512mem_mask,
8482 __builtin_ia32_pmovusqb512mem_mask,
8483 __builtin_ia32_pmovsqb512mem_mask,
8484 __builtin_ia32_pmovusdb512mem_mask,
8485 __builtin_ia32_pmovsdb512mem_mask,
8486 __builtin_ia32_pmovdb512mem_mask.
8487 (bdesc_args): Add __builtin_ia32_expanddf512,
8488 __builtin_ia32_expandsf512.
8489 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
8490 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
8491 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
8492 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
8493 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
8494 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
8495 (avx512f_<code>v8div16qi2_mask_store): This.
8496 (avx512f_expand<mode>): New.
8498 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
8500 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
8502 (_mm512_mask_prefetch_i64gather_pd): Ditto.
8503 (_mm512_prefetch_i32scatter_pd): Ditto.
8504 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
8505 (_mm512_prefetch_i64scatter_pd): Ditto.
8506 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
8507 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
8508 (_mm512_mask_prefetch_i64gather_ps): Ditto.
8509 (_mm512_prefetch_i32scatter_ps): Ditto.
8510 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
8511 (_mm512_prefetch_i64scatter_ps): Ditto.
8512 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
8513 * config/i386/i386-builtin-types.def: Define
8514 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
8515 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
8516 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
8517 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
8518 IX86_BUILTIN_SCATTERPFQPD.
8519 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
8520 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
8521 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
8522 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
8523 __builtin_ia32_scatterpfqps.
8524 (ix86_expand_builtin): Expand new built-ins.
8525 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8526 fix memory access data type.
8527 (*avx512pf_gatherpf<mode>_mask): Ditto.
8528 (*avx512pf_gatherpf<mode>): Ditto.
8529 (avx512pf_scatterpf<mode>): Ditto.
8530 (*avx512pf_scatterpf<mode>_mask): Ditto.
8531 (*avx512pf_scatterpf<mode>): Ditto.
8532 (GATHER_SCATTER_SF_MEM_MODE): New.
8533 (avx512pf_gatherpf<mode>df): Ditto.
8534 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8535 (*avx512pf_scatterpf<mode>df): Ditto.
8537 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8540 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8541 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8544 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8546 * common/config/arm/arm-common.c
8547 (arm_rewrite_mcpu): Handle multiple names.
8549 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8551 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8553 * gimple-builder.h (create_gimple_tmp): Delete.
8555 2014-01-27 Christian Bruel <christian.bruel@st.com>
8557 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
8560 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
8562 * config/pa/pa.md (call): Generate indirect long calls to non-local
8563 functions when outputing 32-bit code.
8564 (call_value): Likewise except for special call to buggy powf function.
8566 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
8567 portable runtime and PIC indirect calls.
8568 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
8569 and PIC call sequences. Use ldo instead of blr to set return register
8570 in PIC call sequence.
8572 2014-01-25 Walter Lee <walt@tilera.com>
8574 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
8575 avoid clobbering a live register.
8577 2014-01-25 Walter Lee <walt@tilera.com>
8579 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
8580 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
8581 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
8582 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
8584 2014-01-25 Walter Lee <walt@tilera.com>
8586 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
8587 arguments on even registers.
8588 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
8590 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
8591 (BIGGEST_ALIGNMENT): Ditto.
8592 (BIGGEST_FIELD_ALIGNMENT): Ditto.
8594 2014-01-25 Walter Lee <walt@tilera.com>
8596 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
8597 insns before bundling.
8598 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8600 2014-01-25 Walter Lee <walt@tilera.com>
8602 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
8603 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
8604 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
8606 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8608 * config/mips/constraints.md (kl): Delete.
8609 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
8610 define expands, using...
8611 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
8612 instructions for MIPS16.
8613 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
8614 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
8616 2014-01-25 Walter Lee <walt@tilera.com>
8618 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
8622 2014-01-25 Walter Lee <walt@tilera.com>
8624 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
8625 (TARGET_EXPAND_TO_RTL_HOOK): Define.
8627 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
8629 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
8632 2014-01-25 Jakub Jelinek <jakub@redhat.com>
8634 * print-rtl.c (in_call_function_usage): New var.
8635 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
8636 EXPR_LIST mode as mode and not as reg note name.
8639 * cfgloopmanip.c (copy_loop_info): If
8640 loop->warned_aggressive_loop_optimizations, make sure
8641 the flag is set in target loop too.
8643 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
8645 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
8647 * builtins.def: Likewise.
8648 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
8649 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
8650 * ira.c (ira_setup_eliminable_regset): Likewise.
8651 * omp-low.c (gate_expand_omp): Likewise.
8652 (execute_lower_omp): Likewise.
8653 (diagnose_sb_0): Likewise.
8654 (gate_diagnose_omp_blocks): Likewise.
8655 (simd_clone_clauses_extract): Likewise.
8658 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8660 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
8661 correction for little endian...
8662 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
8665 2014-01-24 Jeff Law <law@redhat.com>
8667 PR tree-optimization/59919
8668 * tree-vrp.c (find_assert_locations_1): Do not register asserts
8669 for non-returning calls.
8671 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
8673 * common/config/aarch64/aarch64-common.c
8674 (aarch64_rewrite_mcpu): Handle multiple names.
8675 * config/aarch64/aarch64.h
8676 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8678 2014-01-24 Dodji Seketeli <dodji@redhat.com>
8680 * input.c (add_file_to_cache_tab): Handle the case where fopen
8683 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
8686 * config/i386/i386.md (pushsf splitter): Get stack adjustment
8687 from push operand if code of push isn't PRE_DEC.
8689 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8692 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
8693 -mquad-memory-atomic. Update -mquad-memory documentation to say
8694 it is only used for non-atomic loads/stores.
8696 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
8697 -mquad-memory or -mquad-memory-atomic switches.
8699 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
8700 -mquad-memory-atomic to ISA 2.07 support.
8702 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
8703 to separate support of normal quad word memory operations (ldq, stq)
8704 from the atomic quad word memory operations.
8706 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
8707 support to separate non-atomic quad word operations from atomic
8708 quad word operations. Disable non-atomic quad word operations in
8709 little endian mode so that we don't have to swap words after the
8710 load and before the store.
8711 (quad_load_store_p): Add comment about atomic quad word support.
8712 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
8713 options printed with -mdebug=reg.
8715 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
8716 -mquad-memory-atomic as the test for whether we have quad word
8717 atomic instructions.
8718 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
8719 or -mp8-vector are used, allow byte/half-word atomic operations.
8721 * config/rs6000/sync.md (load_lockedti): Insure that the address
8722 is a proper indexed or indirect address for the lqarx instruction.
8723 On little endian systems, swap the hi/lo registers after the lqarx
8725 (load_lockedpti): Use indexed_or_indirect_operand predicate to
8726 insure the address is valid for the lqarx instruction.
8727 (store_conditionalti): Insure that the address is a proper indexed
8728 or indirect address for the stqcrx. instruction. On little endian
8729 systems, swap the hi/lo registers before doing the stqcrx.
8731 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
8732 insure the address is valid for the stqcrx. instruction.
8734 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8735 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
8736 type of quad memory support is available.
8738 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
8741 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
8742 there is a danger of looping.
8744 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
8746 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8747 force flag_ira_loop_pressure if set via command line.
8749 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8751 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
8752 (ashr_simd): New builtin handling DI mode.
8753 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
8754 (aarch64_sshr_simddi): New match pattern.
8755 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
8756 (vshrd_n_s64): Likewise.
8757 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
8759 2014-01-23 Nick Clifton <nickc@redhat.com>
8761 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
8762 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
8763 favour of mcu specific scripts.
8764 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
8767 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8768 Alex Velenko <Alex.Velenko@arm.com>
8770 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
8771 (vaddv_s16): Likewise.
8772 (vaddv_s32): Likewise.
8773 (vaddv_u8): Likewise.
8774 (vaddv_u16): Likewise.
8775 (vaddv_u32): Likewise.
8776 (vaddvq_s8): Likewise.
8777 (vaddvq_s16): Likewise.
8778 (vaddvq_s32): Likewise.
8779 (vaddvq_s64): Likewise.
8780 (vaddvq_u8): Likewise.
8781 (vaddvq_u16): Likewise.
8782 (vaddvq_u32): Likewise.
8783 (vaddvq_u64): Likewise.
8784 (vaddv_f32): Likewise.
8785 (vaddvq_f32): Likewise.
8786 (vaddvq_f64): Likewise.
8787 (vmaxv_f32): Likewise.
8788 (vmaxv_s8): Likewise.
8789 (vmaxv_s16): Likewise.
8790 (vmaxv_s32): Likewise.
8791 (vmaxv_u8): Likewise.
8792 (vmaxv_u16): Likewise.
8793 (vmaxv_u32): Likewise.
8794 (vmaxvq_f32): Likewise.
8795 (vmaxvq_f64): Likewise.
8796 (vmaxvq_s8): Likewise.
8797 (vmaxvq_s16): Likewise.
8798 (vmaxvq_s32): Likewise.
8799 (vmaxvq_u8): Likewise.
8800 (vmaxvq_u16): Likewise.
8801 (vmaxvq_u32): Likewise.
8802 (vmaxnmv_f32): Likewise.
8803 (vmaxnmvq_f32): Likewise.
8804 (vmaxnmvq_f64): Likewise.
8805 (vminv_f32): Likewise.
8806 (vminv_s8): Likewise.
8807 (vminv_s16): Likewise.
8808 (vminv_s32): Likewise.
8809 (vminv_u8): Likewise.
8810 (vminv_u16): Likewise.
8811 (vminv_u32): Likewise.
8812 (vminvq_f32): Likewise.
8813 (vminvq_f64): Likewise.
8814 (vminvq_s8): Likewise.
8815 (vminvq_s16): Likewise.
8816 (vminvq_s32): Likewise.
8817 (vminvq_u8): Likewise.
8818 (vminvq_u16): Likewise.
8819 (vminvq_u32): Likewise.
8820 (vminnmv_f32): Likewise.
8821 (vminnmvq_f32): Likewise.
8822 (vminnmvq_f64): Likewise.
8824 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
8826 * config/aarch64/aarch64-simd.md
8827 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
8828 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
8829 (*aarch64_mul3_elt<mode>): Likewise.
8830 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
8831 (*aarch64_mul3_elt_to_64v2df): Likewise.
8832 (*aarch64_mla_elt<mode>): Likewise.
8833 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8834 (*aarch64_mls_elt<mode>): Likewise.
8835 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8836 (*aarch64_fma4_elt<mode>): Likewise.
8837 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8838 (*aarch64_fma4_elt_to_64v2df): Likewise.
8839 (*aarch64_fnma4_elt<mode>): Likewise.
8840 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8841 (*aarch64_fnma4_elt_to_64v2df): Likewise.
8842 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
8843 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
8844 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8845 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
8846 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
8847 (aarch64_sqdmull_lane<mode>_internal): Likewise.
8848 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
8850 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
8852 * config/aarch64/aarch64-simd.md
8853 (aarch64_be_checked_get_lane<mode>): New define_expand.
8854 * config/aarch64/aarch64-simd-builtins.def
8855 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
8856 New builtin definition.
8857 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
8858 Use new safe be builtin.
8860 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
8862 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
8864 (aarch64_be_st1<mode>): Likewise.
8865 (aarch_ld1<VALL:mode>): Define_expand modified.
8866 (aarch_st1<VALL:mode>): Likewise.
8867 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
8868 (UNSPEC_ST1): Likewise.
8870 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
8872 * config/microblaze/microblaze.md: Add trap insn and attribute
8874 2014-01-23 Dodji Seketeli <dodji@redhat.com>
8876 PR preprocessor/58580
8877 * input.h (location_get_source_line): Take an additional line_size
8879 (void diagnostics_file_cache_fini): Declare new function.
8880 * input.c (struct fcache): New type.
8881 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
8882 New static constants.
8883 (diagnostic_file_cache_init, total_lines_num)
8884 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
8885 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
8886 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
8887 (get_next_line, read_next_line, goto_next_line, read_line_num):
8888 New static function definitions.
8889 (diagnostic_file_cache_fini): New function.
8890 (location_get_source_line): Take an additional output line_len
8891 parameter. Re-write using lookup_or_add_file_to_cache_tab and
8893 * diagnostic.c (diagnostic_finish): Call
8894 diagnostic_file_cache_fini.
8895 (adjust_line): Take an additional input parameter for the length
8896 of the line, rather than calculating it with strlen.
8897 (diagnostic_show_locus): Adjust the use of
8898 location_get_source_line and adjust_line with respect to their new
8899 signature. While displaying a line now, do not stop at the first
8900 null byte. Rather, display the zero byte as a space and keep
8901 going until we reach the size of the line.
8902 * Makefile.in: Add vec.o to OBJS-libcommon
8904 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8905 Ilya Tocar <ilya.tocar@intel.com>
8907 * config/i386/avx512fintrin.h (_mm512_kmov): New.
8908 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
8909 (__builtin_ia32_kmov16): Ditto.
8910 * config/i386/i386.md (UNSPEC_KMOV): New.
8913 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
8915 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
8916 (_mm512_storeu_si512): Ditto.
8918 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
8921 * rtl.h (get_referenced_operands): Declare.
8922 * recog.c (get_referenced_operands): New function.
8923 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
8924 operands have been referenced when recording LO_SUM references.
8926 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
8928 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
8930 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8932 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
8933 Enable for generic and recent AMD targets.
8935 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
8937 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
8938 ARG_SIZE note when adjustment was eliminated.
8940 2014-01-22 Jeff Law <law@redhat.com>
8942 PR tree-optimization/59597
8943 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
8944 in file. Accept new argument REGISTERING and use it to modify
8945 dump output appropriately.
8946 (register_jump_thread): Corresponding changes.
8947 (mark_threaded_blocks): Reinstate code to cancel unprofitable
8948 thread paths involving joiner blocks. Add code to dump cancelled
8949 jump threading paths.
8951 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
8953 PR rtl-optimization/59477
8954 * lra-constraints.c (inherit_in_ebb): Process call for living hard
8955 regs. Update reloads_num and potential_reload_hard_regs for all insns.
8957 2014-01-22 Tom Tromey <tromey@redhat.com>
8959 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
8961 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
8963 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8965 PR rtl-optimization/59896
8966 * lra-constraints.c (process_alt_operands): Check unused note for
8967 matched operands of insn with no output reloads.
8969 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
8971 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
8972 (mips_move_from_gpr_cost): Likewise.
8974 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
8976 PR rtl-optimization/59858
8977 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
8978 ira_class_hard_regs_num.
8979 (process_alt_operands): Increase reject for dying matched operand.
8981 2014-01-21 Jakub Jelinek <jakub@redhat.com>
8984 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
8985 smaller than size, perform several stores or loads and stores
8986 at dst + count - size to store or copy all of size bytes, rather
8987 than just last modesize bytes.
8989 2014-01-20 DJ Delorie <dj@redhat.com>
8991 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
8992 that CLOBBERs are REGs before propogating their values.
8994 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
8997 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
8998 (cgraph_inline_failed_type): New function.
8999 * cgraph.h (DEFCIFCODE): Add type.
9000 (cgraph_inline_failed_type_t): New enum.
9001 (cgraph_inline_failed_type): New prototype.
9002 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
9003 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
9004 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
9005 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
9006 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
9007 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
9008 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
9009 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
9010 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
9011 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
9012 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
9013 OPTIMIZATION_MISMATCH.
9014 * tree-inline.c (expand_call_inline): Emit errors during
9015 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
9017 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9020 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
9021 mode attribute in insn output.
9023 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
9025 * output.h (output_constant): Delete.
9026 * varasm.c (output_constant): Make private.
9028 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
9030 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
9032 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9035 * tree.h (fold_builtin_strcat): New prototype.
9036 * builtins.c (fold_builtin_strcat): No longer static. Add len
9037 argument, if non-NULL, don't call c_strlen. Optimize
9038 directly into __builtin_memcpy instead of __builtin_strcpy.
9039 (fold_builtin_2): Adjust fold_builtin_strcat caller.
9040 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
9042 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9044 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9045 for SImode_address_operand operands, having only a REG argument.
9047 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
9049 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
9050 loader name using mbig-endian.
9051 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
9053 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9055 * doc/invoke.texi (-march): Clarify documentation for AArch64.
9059 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
9061 * config/aarch64/aarch64-protos.h
9062 (aarch64_cannot_change_mode_class_ptr): Declare.
9063 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
9064 aarch64_cannot_change_mode_class_ptr): New.
9065 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
9066 backend hook aarch64_cannot_change_mode_class.
9068 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9070 * common/config/aarch64/aarch64-common.c
9071 (aarch64_handle_option): Don't handle any option order logic here.
9072 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
9073 selected_cpu, warn on architecture version mismatch.
9074 (aarch64_override_options): Fix parsing order for option strings.
9076 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9077 Iain Sandoe <iain@codesourcery.com>
9080 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
9081 warning. Amend comment to reflect current functionality.
9083 2014-01-20 Richard Biener <rguenther@suse.de>
9086 * builtins.c (fold_builtin_strcat): Remove case better handled
9087 by tree-ssa-strlen.c.
9089 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
9091 * config/aarch64/aarch64.opt
9092 (mcpu, march, mtune): Make case-insensitive.
9094 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9097 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9098 if operands[1] is a REG or ZERO_EXTEND of a REG.
9100 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
9102 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
9104 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
9106 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
9107 long non-pic millicode calls.
9109 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9111 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
9113 2014-01-19 Kito Cheng <kito@0xlab.org>
9115 * builtins.c (expand_movstr): Check movstr expand done or fail.
9117 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9118 H.J. Lu <hongjiu.lu@intel.com>
9121 * config/i386/i386.md (*lea<mode>): Zero-extend return register
9122 to DImode for zero-extended addresses.
9124 2014-01-19 Jakub Jelinek <jakub@redhat.com>
9126 PR rtl-optimization/57763
9127 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
9128 on the new indirect jump_insn and increment LABEL_NUSES (label).
9130 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
9134 * config.gcc (x86_archs): New variable.
9135 (x86_64_archs): Likewise.
9136 (x86_cpus): Likewise.
9137 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
9138 --with-arch/--with-cpu= options.
9139 Support --with-arch=/--with-cpu={nehalem,westmere,
9140 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
9142 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9144 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
9145 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
9147 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9149 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
9151 2014-01-18 Jakub Jelinek <jakub@redhat.com>
9154 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
9155 clear cpp_get_options (parse_in)->warn_unused_macros for
9156 ix86_target_macros_internal with cpp_define.
9158 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
9160 * jump.c (delete_related_insns): Keep (use (insn))s.
9161 * reorg.c (redundant_insn): Check for barriers too.
9163 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9165 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
9167 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
9169 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
9170 call to $$dyncall when TARGET_LONG_CALLS is true.
9172 2014-01-17 Jeff Law <law@redhat.com>
9174 * ree.c (combine_set_extension): Temporarily disable test for
9175 changing number of hard registers.
9177 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9180 * ipa-inline-analysis.c (inline_free_summary):
9181 Do not free summary of aliases.
9183 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9186 * gimplify.c (gimplify_expr): Use create_tmp_var
9187 instead of create_tmp_var_raw. If cond doesn't have
9188 integral type, don't add the IFN_ANNOTATE builtin at all.
9190 2014-01-17 Martin Jambor <mjambor@suse.cz>
9193 * ipa-cp.c (prev_edge_clone): New variable.
9194 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
9195 Also resize prev_edge_clone vector.
9196 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
9197 (ipcp_edge_removal_hook): New function.
9198 (ipcp_driver): Register ipcp_edge_removal_hook.
9200 2014-01-17 Andrew Pinski <apinski@cavium.com>
9201 Steve Ellcey <sellcey@mips.com>
9204 * config/mips/mips.c (mips_print_operand): Check operand mode instead
9207 2014-01-17 Jeff Law <law@redhat.com>
9210 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
9211 so that pass_ccp runs first.
9213 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9215 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
9216 (ix86_adjust_cost): Use !TARGET_XXX.
9217 (do_reorder_for_imul): Likewise.
9218 (swap_top_of_ready_list): Likewise.
9219 (ix86_sched_reorder): Likewise.
9221 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9223 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9224 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
9225 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
9226 (intel_memset): New. Duplicate slm_memset.
9227 (intel_cost): New. Duplicate slm_cost.
9228 (m_INTEL): New macro.
9229 (processor_target_table): Add "intel".
9230 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
9231 with PROCESSOR_INTEL for "intel".
9232 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
9233 PROCESSOR_SILVERMONT.
9234 (ix86_issue_rate): Likewise.
9235 (ix86_adjust_cost): Likewise.
9236 (ia32_multipass_dfa_lookahead): Likewise.
9237 (swap_top_of_ready_list): Likewise.
9238 (ix86_sched_reorder): Likewise.
9239 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
9240 instead of TARGET_OPT_AGU.
9241 * config/i386/i386.h (TARGET_INTEL): New.
9242 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
9243 (processor_type): Add PROCESSOR_INTEL.
9244 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
9245 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
9247 2014-01-17 Marek Polacek <polacek@redhat.com>
9250 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
9253 2014-01-17 Richard Biener <rguenther@suse.de>
9255 PR tree-optimization/46590
9256 * opts.c (default_options_table): Add entries for
9257 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
9258 all enabled at -O1 but not for -Og.
9259 * common.opt (fbranch-count-reg): Remove Init(1).
9260 (fmove-loop-invariants): Likewise.
9261 (ftree-pta): Likewise.
9263 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9265 * config/i386/i386.c (ix86_data_alignment): For compatibility with
9266 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
9267 decls to at least the GCC 4.8 used alignments.
9270 * tree-nested.c (convert_nonlocal_reference_stmt,
9271 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
9272 of GIMPLE_BIND stmts, adjust associated decls.
9274 2014-01-17 Richard Biener <rguenther@suse.de>
9276 PR tree-optimization/46590
9277 * vec.h (vec<>::bseach): New member function implementing
9278 binary search according to C89 bsearch.
9279 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
9280 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
9281 bitmap pointer again. Make accesses_in_loop a flat array.
9282 (mem_ref_obstack): New global.
9283 (outermost_indep_loop): Adjust for mem_ref->stored changes.
9284 (mark_ref_stored): Likewise.
9285 (ref_indep_loop_p_2): Likewise.
9286 (set_ref_stored_in_loop): New helper function.
9287 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
9288 (memref_free): Adjust.
9289 (record_mem_ref_loc): Simplify.
9290 (gather_mem_refs_stmt): Adjust.
9291 (sort_locs_in_loop_postorder_cmp): New function.
9292 (analyze_memory_references): Sort accesses_in_loop after
9293 loop postorder number.
9294 (find_ref_loc_in_loop_cmp): New function.
9295 (for_all_locs_in_loop): Find relevant cluster of locs in
9296 accesses_in_loop and iterate without recursion.
9297 (execute_sm): Avoid uninit warning.
9298 (struct ref_always_accessed): Simplify.
9299 (ref_always_accessed::operator ()): Likewise.
9300 (ref_always_accessed_p): Likewise.
9301 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
9302 loop postorder numbers here.
9303 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
9306 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9309 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
9310 on decls for which assemble_alias has been called.
9312 2014-01-17 Nick Clifton <nickc@redhat.com>
9314 * config/msp430/msp430.opt: (mcpu): New option.
9315 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
9316 (msp430_option_override): Parse target_cpu. If the MCU name
9317 matches a generic string, clear target_mcu.
9318 (msp430_attr): Allow numeric interrupt values up to 63.
9319 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
9320 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
9322 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
9324 * config/msp430/msp430.md (popm): Use %J rather than %I.
9325 (addsi3): Use msp430_nonimmediate_operand for operand 2.
9326 (addhi_cy_i): Use immediate_operand for operand 2.
9327 * doc/invoke.texi: Document -mcpu option.
9329 2014-01-17 Richard Biener <rguenther@suse.de>
9331 PR rtl-optimization/38518
9332 * df.h (df_analyze_loop): Declare.
9333 * df-core.c: Include cfgloop.h.
9334 (df_analyze_1): Split out main part of df_analyze.
9335 (df_analyze): Adjust.
9336 (loop_inverted_post_order_compute): New function.
9337 (loop_post_order_compute): Likewise.
9338 (df_analyze_loop): New function avoiding whole-function
9340 * loop-invariant.c (find_defs): Use df_analyze_loop.
9341 (find_invariants): Adjust.
9342 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
9344 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
9346 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
9347 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
9349 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
9351 * ipa-ref.c (ipa_remove_stmt_references): Fix references
9352 traversal when removing references.
9354 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
9357 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
9359 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
9362 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
9363 pushing a reload for an autoinc when we had previously reloaded an
9364 inner part of the address.
9366 2014-01-16 Jakub Jelinek <jakub@redhat.com>
9368 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
9370 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
9371 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
9372 when not giving up or versioning for alias only because of
9374 (vect_analyze_data_ref_dependences): Set to true.
9375 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
9377 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
9378 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
9382 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
9385 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
9386 operand 0 predicate for gathers, use a new pseudo as subtarget.
9388 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9391 * lra-constraints.c (process_alt_operands): Add printing debug info.
9392 Check absence of input/output reloads for matched operands too.
9394 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9396 PR rtl-optimization/59835
9397 * ira.c (ira_init_register_move_cost): Increase cost for
9400 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
9402 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
9404 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
9407 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
9408 non-register objects. Use gen_(high/low)part more consistently.
9411 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9414 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
9415 endian support, remove tests for WORDS_BIG_ENDIAN.
9416 (p8_mfvsrd_3_<mode>): Likewise.
9417 (reload_gpr_from_vsx<mode>): Likewise.
9418 (reload_gpr_from_vsxsf): Likewise.
9419 (p8_mfvsrd_4_disf): Likewise.
9421 2014-01-16 Richard Biener <rguenther@suse.de>
9423 PR rtl-optimization/46590
9424 * lcm.c (compute_antinout_edge): Use postorder iteration.
9425 (compute_laterin): Use inverted postorder iteration.
9427 2014-01-16 Nick Clifton <nickc@redhat.com>
9430 * varasm.c (output_constant): Return the number of bytes actually
9432 (output_constructor_array_range): Update the field size with the
9433 number of bytes emitted by output_constant.
9434 (output_constructor_regular_field): Likewise. Also do not
9435 complain if the total number of bytes emitted is now greater
9436 than the expected fieldpos.
9437 * output.h (output_constant): Update prototype and descriptive comment.
9439 2014-01-16 Marek Polacek <polacek@redhat.com>
9442 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
9443 it is error_mark_node.
9445 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
9447 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
9448 VALID_AVX256_REG_OR_OI_MODE.
9450 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
9452 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
9453 current procedure should be profiled.
9455 2014-01-15 Andrew Pinski <apinski@cavium.com>
9457 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
9458 of moving from/to the STACK_REG register class.
9460 2014-01-15 Richard Henderson <rth@redhat.com>
9463 * reginfo.c (global_regs_decl): Globalize.
9464 * rtl.h (global_regs_decl): Declare.
9465 * ira.c (do_reload): Diagnose frame_pointer_needed and it
9466 reserved via global_regs.
9468 2014-01-15 Teresa Johnson <tejohnson@google.com>
9470 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
9472 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9474 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
9475 and vmulosh rather than call gen_vec_widen_smult_*.
9476 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
9477 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
9478 (vec_widen_smult_even_v16qi): Likewise.
9479 (vec_widen_umult_even_v8hi): Likewise.
9480 (vec_widen_smult_even_v8hi): Likewise.
9481 (vec_widen_umult_odd_v16qi): Likewise.
9482 (vec_widen_smult_odd_v16qi): Likewise.
9483 (vec_widen_umult_odd_v8hi): Likewise.
9484 (vec_widen_smult_odd_v8hi): Likewise.
9485 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
9486 vmuloub rather than call gen_vec_widen_umult_*.
9487 (vec_widen_umult_lo_v16qi): Likewise.
9488 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
9489 vmulosb rather than call gen_vec_widen_smult_*.
9490 (vec_widen_smult_lo_v16qi): Likewise.
9491 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
9492 rather than call gen_vec_widen_umult_*.
9493 (vec_widen_umult_lo_v8hi): Likewise.
9494 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
9495 rather than call gen_vec_widen_smult_*.
9496 (vec_widen_smult_lo_v8hi): Likewise.
9498 2014-01-15 Jeff Law <law@redhat.com>
9500 PR tree-optimization/59747
9501 * ree.c (find_and_remove_re): Properly handle case where a second
9502 eliminated extension requires widening a copy created for elimination
9503 of a prior extension.
9504 (combine_set_extension): Ensure that the number of hard regs needed
9505 for a destination register does not change when we widen it.
9507 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
9509 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
9510 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
9511 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
9512 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
9513 (avr-*-rtems*): Likewise.
9514 (bfin*-rtems*): Likewise.
9515 (moxie-*-rtems*): Likewise.
9516 (h8300-*-rtems*): Likewise.
9517 (i[34567]86-*-rtems*): Likewise.
9518 (lm32-*-rtems*): Likewise.
9519 (m32r-*-rtems*): Likewise.
9520 (m68k-*-rtems*): Likewise.
9521 (microblaze*-*-rtems*): Likewise.
9522 (mips*-*-rtems*): Likewise.
9523 (powerpc-*-rtems*): Likewise.
9524 (sh-*-rtems*): Likewise.
9525 (sparc-*-rtems*): Likewise.
9526 (sparc64-*-rtems*): Likewise.
9527 (v850-*-rtems*): Likewise.
9528 (m32c-*-rtems*): Likewise.
9530 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9532 PR rtl-optimization/59511
9533 * ira.c (ira_init_register_move_cost): Use memory costs for some
9534 cases of register move cost calculations.
9535 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9536 instead of BB frequency.
9537 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9538 * lra-assigns.c (find_hard_regno_for): Ditto.
9540 2014-01-15 Richard Biener <rguenther@suse.de>
9542 PR tree-optimization/59822
9543 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9544 (vectorizable_load): Use it to hoist defs of uses of invariant
9545 loads out of the loop.
9547 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9548 Kugan Vivekanandarajah <kuganv@linaro.org>
9551 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9554 2014-01-15 Richard Biener <rguenther@suse.de>
9556 PR rtl-optimization/59802
9557 * lcm.c (compute_available): Use inverted postorder to seed
9558 the initial worklist.
9560 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9563 * config/s390/s390.c (s390_preferred_reload_class): Don't return
9564 ADDR_REGS for invalid symrefs in non-PIC code.
9566 2014-01-15 Jakub Jelinek <jakub@redhat.com>
9569 * builtins.c (determine_block_size): Initialize *probable_max_size
9570 even if len_rtx is CONST_INT.
9572 2014-01-14 Andrew Pinski <apinski@cavium.com>
9574 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
9575 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
9576 (cortexa53_tunings): Likewise.
9577 (aarch64_sched_issue_rate): New function.
9578 (TARGET_SCHED_ISSUE_RATE): Define.
9580 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9582 * ira-costs.c (find_costs_and_classes): Add missed
9583 ira_init_register_move_cost_if_necessary.
9585 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
9588 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
9590 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
9593 * config/i386/i386.c (type_natural_mode): Add a bool parameter
9594 to indicate if type is used for function return value. Warn ABI
9595 change if the vector mode isn't available for function return value.
9596 (ix86_function_arg_advance): Pass false to type_natural_mode.
9597 (ix86_function_arg): Likewise.
9598 (ix86_gimplify_va_arg): Likewise.
9599 (function_arg_32): Don't warn ABI change.
9600 (ix86_function_value): Pass true to type_natural_mode.
9601 (ix86_return_in_memory): Likewise.
9602 (ix86_struct_value_rtx): Removed.
9603 (TARGET_STRUCT_VALUE_RTX): Likewise.
9605 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9607 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
9608 converting a conditional jump into a conditional return.
9610 2014-01-14 Richard Biener <rguenther@suse.de>
9612 PR tree-optimization/58921
9613 PR tree-optimization/59006
9614 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
9615 hoisting invariant stmts.
9616 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
9617 invariant loads on the preheader edge if possible.
9619 2014-01-14 Joey Ye <joey.ye@arm.com>
9621 * doc/plugin.texi (Building GCC plugins): Update to C++.
9623 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
9625 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
9626 (_mm_rcp28_round_ss): Ditto.
9627 (_mm_rsqrt28_round_sd): Ditto.
9628 (_mm_rsqrt28_round_ss): Ditto.
9629 (_mm_rcp28_sd): Ditto.
9630 (_mm_rcp28_ss): Ditto.
9631 (_mm_rsqrt28_sd): Ditto.
9632 (_mm_rsqrt28_ss): Ditto.
9633 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
9634 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
9635 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
9636 (IX86_BUILTIN_RCP28SD): Ditto.
9637 (IX86_BUILTIN_RCP28SS): Ditto.
9638 (IX86_BUILTIN_RSQRT28SD): Ditto.
9639 (IX86_BUILTIN_RSQRT28SS): Ditto.
9640 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
9641 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
9642 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
9643 (ix86_expand_special_args_builtin): Expand new FTYPE.
9644 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
9645 (srcp14<mode>): Make insn unary.
9646 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
9647 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
9648 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
9649 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
9650 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
9651 Fix rounding: make it SAE only.
9652 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9654 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9656 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
9657 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
9658 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
9659 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
9660 (round_saeonly_mask_scalar_operand4): Ditto.
9661 (round_saeonly_mask_scalar_op3): Ditto.
9662 (round_saeonly_mask_scalar_op4): Ditto.
9664 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9666 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
9667 Implement -maltivec=be for vec_insert and vec_extract.
9669 2014-01-10 DJ Delorie <dj@redhat.com>
9671 * config/msp430/msp430.md (call_internal): Don't allow memory
9672 references with SP as the base register.
9673 (call_value_internal): Likewise.
9674 * config/msp430/constraints.md (Yc): New. For memory references
9675 that don't use SP as a base register.
9677 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
9678 "an integer without a # prefix"
9679 * config/msp430/msp430.md (epilogue_helper): Use it.
9681 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9684 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
9685 AVX512F gather builtins.
9686 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
9687 on gather decls with INTEGER_TYPE masktype.
9688 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
9689 directly into the builtin rather than hoisting it before loop.
9691 PR tree-optimization/59387
9692 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
9693 (scev_const_prop): If folded_casts and type has undefined overflow,
9694 use force_gimple_operand instead of force_gimple_operand_gsi and
9695 for each added stmt if it is assign with
9696 arith_code_with_undefined_signed_overflow, call
9697 rewrite_to_defined_overflow.
9698 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
9699 gimple-fold.h instead.
9700 (arith_code_with_undefined_signed_overflow,
9701 rewrite_to_defined_overflow): Moved to ...
9702 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
9703 rewrite_to_defined_overflow): ... here. No longer static.
9704 Include gimplify-me.h.
9705 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
9706 rewrite_to_defined_overflow): New prototypes.
9708 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9710 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
9712 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
9714 * builtins.c (get_object_alignment_2): Minor tweak.
9715 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
9717 2014-01-13 Christian Bruel <christian.bruel@st.com>
9719 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
9720 optimized non constant lengths.
9722 2014-01-13 Jakub Jelinek <jakub@redhat.com>
9725 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
9726 load as __atomic_load_N if possible.
9728 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9730 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
9732 (rs6000_expand_builtin): Adjust call.
9734 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
9737 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
9738 * config/rs6000/rs6000.c: Include target-globals.h.
9739 (rs6000_set_current_function): Instead of doing target_reinit
9740 unconditionally, use save_target_globals_default_opts and
9741 restore_target_globals.
9743 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
9745 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
9746 (rs6000_expand_builtin): Handle mffs and mtfsf.
9747 (rs6000_init_builtins): Define mffs and mtfsf.
9748 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
9749 (rs6000_mffs): New pattern.
9750 (rs6000_mtfsf): New pattern.
9752 2014-01-11 Bin Cheng <bin.cheng@arm.com>
9754 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
9755 Start narrowing with START. Apply candidate-use pair
9756 and check overall cost in narrowing.
9757 (iv_ca_prune): Pass new argument.
9759 2014-01-10 Jeff Law <law@redhat.com>
9762 * ree.c (combine_reaching_defs): Ensure the defining statement
9763 occurs before the extension when optimizing extensions with
9764 different source and destination hard registers.
9766 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9769 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
9770 vtables into the type inheritance graph.
9772 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9774 PR rtl-optimization/59754
9775 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
9776 modes in the REGNO != REGNO case.
9778 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9780 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
9782 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9784 PR tree-optimization/59745
9785 * tree-predcom.c (tree_predictive_commoning_loop): Call
9786 free_affine_expand_cache if giving up because components is NULL.
9788 * target-globals.c (save_target_globals): Allocate < 4KB structs using
9789 GC in payload of target_globals struct instead of allocating them on
9790 the heap and the larger structs separately using GC.
9791 * target-globals.h (struct target_globals): Make regs, hard_regs,
9792 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
9793 of GTY((skip)) and change type to void *.
9794 (reset_target_globals): Cast loads from those fields to corresponding
9797 2014-01-10 Steve Ellcey <sellcey@mips.com>
9800 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
9801 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
9802 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
9804 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
9807 * aarch64-modes.def (CC_Zmode): New flags mode.
9808 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
9809 represents an equality.
9810 (aarch64_get_condition_code): Handle CC_Zmode.
9811 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
9813 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9815 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
9816 extraction in good case.
9818 2014-01-10 Richard Biener <rguenther@suse.de>
9820 PR tree-optimization/59374
9821 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
9822 checking after SLP discovery. Mark stmts not participating
9823 in any SLP instance properly.
9825 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9827 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
9828 when handling a SET rtx.
9830 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9832 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
9833 (cortex-a57): Likewise.
9834 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
9836 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9838 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
9839 non-iwmmxt builtins.
9841 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
9845 * ipa-devirt.c record_target_from_binfo): Take as argument
9846 stack of binfos and lookup matching one for virtual inheritance.
9847 (possible_polymorphic_call_targets_1): Update.
9849 2014-01-10 Huacai Chen <chenhc@lemote.com>
9851 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
9852 kernel strings for Loongson-2E/2F/3A.
9854 2014-01-10 Jakub Jelinek <jakub@redhat.com>
9857 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
9858 is_gimple_call before calling gimple_call_internal_p.
9860 2014-01-09 Steve Ellcey <sellcey@mips.com>
9862 * Makefile.in (TREE_FLOW_H): Remove.
9863 (TREE_SSA_H): Add file names from tree-flow.h.
9864 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
9865 * tree.h: Remove tree-flow.h reference.
9866 * hash-table.h: Remove tree-flow.h reference.
9867 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
9868 reference with tree-ssa-loop.h.
9870 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9872 * doc/invoke.texi: Add -maltivec={be,le} options, and document
9873 default element-order behavior for -maltivec.
9874 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
9875 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
9876 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
9877 when targeting big endian, at least for now.
9878 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
9880 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9883 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
9884 var satisfies use_register_for_decl, just take into account type
9885 alignment, rather than decl alignment.
9887 PR tree-optimization/59622
9888 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
9889 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
9890 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
9891 Don't devirtualize for inplace at all. For targets.length () == 1,
9892 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
9894 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9896 * config/i386/i386.md (cpu): Remove the unused btver1.
9898 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
9900 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
9902 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9905 * tree-core.h (struct target_globals): New forward declaration.
9906 (struct tree_target_option): Add globals field.
9907 * tree.h (TREE_TARGET_GLOBALS): Define.
9908 (prepare_target_option_nodes_for_pch): New prototype.
9909 * target-globals.h (struct target_globals): Define even if
9911 * tree.c (prepare_target_option_node_for_pch,
9912 prepare_target_option_nodes_for_pch): New functions.
9913 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
9914 * config/i386/i386.c: Include target-globals.h.
9915 (ix86_set_current_function): Instead of doing target_reinit
9916 unconditionally, use save_target_globals_default_opts and
9917 restore_target_globals.
9919 2014-01-09 Richard Biener <rguenther@suse.de>
9921 PR tree-optimization/59715
9922 * tree-cfg.h (split_critical_edges): Declare.
9923 * tree-cfg.c (split_critical_edges): Export.
9924 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
9926 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
9928 * cfgexpand.c (expand_stack_vars): Optionally disable
9929 asan stack protection.
9930 (expand_used_vars): Likewise.
9931 (partition_stack_vars): Likewise.
9932 * asan.c (asan_emit_stack_protection): Optionally disable
9933 after return stack usage.
9934 (instrument_derefs): Optionally disable memory access instrumentation.
9935 (instrument_builtin_call): Likewise.
9936 (instrument_strlen_call): Likewise.
9937 (asan_protect_global): Optionally disable global variables protection.
9938 * doc/invoke.texi: Added doc for new options.
9939 * params.def: Added new options.
9940 * params.h: Likewise.
9942 2014-01-09 Jakub Jelinek <jakub@redhat.com>
9944 PR rtl-optimization/59724
9945 * ifcvt.c (cond_exec_process_if_block): Don't call
9946 flow_find_head_matching_sequence with 0 longest_match.
9947 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
9948 non-active insns if !stop_after.
9949 (try_head_merge_bb): Revert 2014-01-07 changes.
9951 2014-01-08 Jeff Law <law@redhat.com>
9953 * ree.c (get_sub_rtx): New function, extracted from...
9954 (merge_def_and_ext): Here.
9955 (combine_reaching_defs): Use get_sub_rtx.
9957 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
9959 * cgraph.h (varpool_variable_node): Do not choke on null node.
9961 2014-01-08 Catherine Moore <clm@codesourcery.com>
9963 * config/mips/mips.md (simple_return): Attempt to use JRC
9965 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
9967 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9969 PR rtl-optimization/59137
9970 * reorg.c (steal_delay_list_from_target): Call update_block for
9972 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
9974 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9976 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
9977 two duplicate entries.
9979 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
9982 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
9984 * config/mips/mips.c (mips_truncated_op_cost): New function.
9985 (mips_rtx_costs): Adjust test for BADDU.
9986 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
9988 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
9990 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
9991 (*baddu_si): ...this new pattern.
9993 2014-01-08 Jakub Jelinek <jakub@redhat.com>
9996 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
9998 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
10000 PR middle-end/57748
10001 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
10003 (expand_expr, expand_normal): Adjust.
10004 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
10005 inner_reference_p. Use inner_reference_p to expand inner references.
10006 (store_expr): Adjust.
10007 * cfgexpand.c (expand_call_stmt): Adjust.
10009 2014-01-08 Rong Xu <xur@google.com>
10011 * gcov-io.c (gcov_var): Move from gcov-io.h.
10012 (gcov_position): Ditto.
10013 (gcov_is_error): Ditto.
10014 (gcov_rewrite): Ditto.
10015 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
10016 only part to libgcc/libgcov.h.
10018 2014-01-08 Marek Polacek <polacek@redhat.com>
10020 PR middle-end/59669
10021 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
10023 2014-01-08 Marek Polacek <polacek@redhat.com>
10026 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
10028 2014-01-08 Jakub Jelinek <jakub@redhat.com>
10030 PR rtl-optimization/59649
10031 * stor-layout.c (get_mode_bounds): For BImode return
10032 0 and STORE_FLAG_VALUE.
10034 2014-01-08 Richard Biener <rguenther@suse.de>
10036 PR middle-end/59630
10037 * gimple.h (is_gimple_builtin_call): Remove.
10038 (gimple_builtin_call_types_compatible_p): New.
10039 (gimple_call_builtin_p): New overload.
10040 * gimple.c (is_gimple_builtin_call): Remove.
10041 (validate_call): Rename to ...
10042 (gimple_builtin_call_types_compatible_p): ... this and export. Also
10043 check return types.
10044 (validate_type): New static function.
10045 (gimple_call_builtin_p): New overload and adjust.
10046 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
10047 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
10048 (gimple_fold_stmt_to_constant_1): Likewise.
10049 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
10051 2014-01-08 Richard Biener <rguenther@suse.de>
10053 PR middle-end/59471
10054 * gimplify.c (gimplify_expr): Gimplify register-register type
10055 VIEW_CONVERT_EXPRs to separate stmts.
10057 2014-01-07 Jeff Law <law@redhat.com>
10059 PR middle-end/53623
10060 * ree.c (combine_set_extension): Handle case where source
10061 and destination registers in an extension insn are different.
10062 (combine_reaching_defs): Allow source and destination registers
10063 in extension to be different under limited circumstances.
10064 (add_removable_extension): Remove restriction that the
10065 source and destination registers in the extension are the same.
10066 (find_and_remove_re): Emit a copy from the extension's
10067 destination to its source after the defining insn if
10068 the source and destination registers are different.
10070 PR middle-end/59285
10071 * ifcvt.c (merge_if_block): If we are merging a block with more than
10072 one successor with a block with no successors, remove any BARRIER
10073 after the second block.
10075 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
10077 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
10079 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
10082 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
10083 for 14-bit register offsets when INT14_OK_STRICT is false.
10085 2014-01-07 Roland Stigge <stigge@antcom.de>
10086 Michael Meissner <meissner@linux.vnet.ibm.com>
10089 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
10090 Only check TFmode for SPE constants. Don't check TImode or TDmode.
10092 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
10094 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
10097 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
10099 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
10100 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
10101 rtx is const0_rtx or not.
10103 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
10106 * target-globals.c (save_target_globals): Remove this_fn_optab
10108 * toplev.c: Include optabs.h.
10109 (target_reinit): Temporarily restore the global options if another
10110 set of options are in force.
10112 2014-01-07 Jakub Jelinek <jakub@redhat.com>
10114 PR rtl-optimization/58668
10115 * cfgcleanup.c (flow_find_cross_jump): Don't count
10116 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
10117 to determine what is counted.
10118 (flow_find_head_matching_sequence): Use active_insn_p to determine
10120 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
10122 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
10123 determine what is counted.
10125 PR tree-optimization/59643
10126 * tree-predcom.c (split_data_refs_to_components): If one dr is
10127 read and one write, determine_offset fails and the write isn't
10128 in the bad component, just put the read into the bad component.
10130 2014-01-07 Mike Stump <mikestump@comcast.net>
10131 Jakub Jelinek <jakub@redhat.com>
10134 * tree-core.h (struct tree_optimization_option): Change optabs
10135 type from unsigned char * to void *.
10136 * optabs.c (init_tree_optimization_optabs): Adjust
10137 TREE_OPTIMIZATION_OPTABS initialization.
10139 2014-01-06 Jakub Jelinek <jakub@redhat.com>
10142 * config/i386/i386.h (struct machine_function): Add
10143 no_drap_save_restore field.
10144 * config/i386/i386.c (ix86_save_reg): Use
10145 !cfun->machine->no_drap_save_restore instead of
10146 crtl->stack_realign_needed.
10147 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
10148 this function clears frame_pointer_needed. Set
10149 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
10150 and DRAP reg is needed.
10152 2014-01-06 Marek Polacek <polacek@redhat.com>
10155 * doc/implement-c.texi: Mention that other integer types are
10156 permitted as bit-field types in strictly conforming mode.
10158 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
10160 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
10161 is newly allocated.
10163 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
10165 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
10167 2014-01-06 Martin Jambor <mjambor@suse.cz>
10170 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
10172 * ipa-prop.c (ipa_print_node_params): Fix indentation.
10174 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
10178 * var-tracking.c (add_stores): Preserve the value of the source even if
10179 we don't record the store.
10181 2014-01-06 Terry Guo <terry.guo@arm.com>
10183 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
10185 2014-01-05 Iain Sandoe <iain@codesourcery.com>
10188 * config/darwin.c (darwin_function_section): Adjust return values to
10189 correspond to optimisation changes made in r206070.
10191 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
10193 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
10194 from prefetch_block tune setting.
10195 (nocona_cost): Correct size of prefetch block to 64.
10197 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
10199 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
10200 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
10201 used to save the static chain register in the computation of the offset
10202 from which the FP registers need to be restored.
10204 2014-01-04 Jakub Jelinek <jakub@redhat.com>
10206 PR tree-optimization/59519
10207 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
10208 ICE if get_current_def (current_new_name) is already non-NULL, as long
10209 as it is a phi result of some other phi in *new_exit_bb that has
10212 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
10213 or vmovdqu* for misaligned_operand.
10214 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
10215 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
10216 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
10217 aligned_mem for AVX512F masked aligned load and store builtins and for
10218 non-temporal moves.
10220 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
10222 PR tree-optimization/59651
10223 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
10224 Address range for negative step should be added by TYPE_SIZE_UNIT.
10226 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
10228 * config/m68k/m68k.c (handle_move_double): Handle pushes with
10229 overlapping registers also for registers other than the stack pointer.
10231 2014-01-03 Marek Polacek <polacek@redhat.com>
10234 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
10237 2014-01-03 Jakub Jelinek <jakub@redhat.com>
10240 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
10243 * config/i386/i386.md (MODE_SIZE): New mode attribute.
10244 (push splitter): Use <P:MODE_SIZE> instead of
10245 GET_MODE_SIZE (<P:MODE>mode).
10246 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
10247 (mov -1, reg peephole2): Likewise.
10248 * config/i386/sse.md (*mov<mode>_internal,
10249 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
10250 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
10251 *<code><mode>3, *andnot<mode>3<mask_name>,
10252 <mask_codefor><code><mode>3<mask_name>): Likewise.
10253 * config/i386/subst.md (mask_mode512bit_condition,
10254 sd_mask_mode512bit_condition): Likewise.
10256 2014-01-02 Xinliang David Li <davidxl@google.com>
10258 PR tree-optimization/59303
10259 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
10260 (dump_predicates): Better output format.
10261 (pred_equal_p): New function.
10262 (is_neq_relop_p): Ditto.
10263 (is_neq_zero_form_p): Ditto.
10264 (pred_expr_equal_p): Ditto.
10265 (pred_neg_p): Ditto.
10266 (simplify_pred): Ditto.
10267 (simplify_preds_2): Ditto.
10268 (simplify_preds_3): Ditto.
10269 (simplify_preds_4): Ditto.
10270 (simplify_preds): Ditto.
10271 (push_pred): Ditto.
10272 (push_to_worklist): Ditto.
10273 (get_pred_info_from_cmp): Ditto.
10274 (is_degenerated_phi): Ditto.
10275 (normalize_one_pred_1): Ditto.
10276 (normalize_one_pred): Ditto.
10277 (normalize_one_pred_chain): Ditto.
10278 (normalize_preds): Ditto.
10279 (normalize_cond_1): Remove function.
10280 (normalize_cond): Ditto.
10281 (is_gcond_subset_of): Ditto.
10282 (is_subset_of_any): Ditto.
10283 (is_or_set_subset_of): Ditto.
10284 (is_and_set_subset_of): Ditto.
10285 (is_norm_cond_subset_of): Ditto.
10286 (pred_chain_length_cmp): Ditto.
10287 (convert_control_dep_chain_into_preds): Type change.
10288 (find_predicates): Ditto.
10289 (find_def_preds): Ditto.
10290 (destroy_predicates_vecs): Ditto.
10291 (find_matching_predicates_in_rest_chains): Ditto.
10292 (use_pred_not_overlap_with_undef_path_pred): Ditto.
10293 (is_pred_expr_subset): Ditto.
10294 (is_pred_chain_subset_of): Ditto.
10295 (is_included_in): Ditto.
10296 (is_superset_of): Ditto.
10298 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10300 Update copyright years.
10302 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10304 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
10305 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
10306 config/arc/arc.md, config/arc/arc.opt,
10307 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
10308 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
10309 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
10310 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
10311 config/linux-protos.h, config/linux.c, config/winnt-c.c,
10312 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
10313 vtable-verify.c, vtable-verify.h: Use the standard form for the
10316 2014-01-02 Tobias Burnus <burnus@net-b.de>
10318 * gcc.c (process_command): Update copyright notice dates.
10319 * gcov-dump.c: Ditto.
10321 * doc/cpp.texi: Bump @copying's copyright year.
10322 * doc/cppinternals.texi: Ditto.
10323 * doc/gcc.texi: Ditto.
10324 * doc/gccint.texi: Ditto.
10325 * doc/gcov.texi: Ditto.
10326 * doc/install.texi: Ditto.
10327 * doc/invoke.texi: Ditto.
10329 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10331 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
10333 2014-01-01 Jakub Jelinek <jakub@redhat.com>
10335 * config/i386/sse.md (*mov<mode>_internal): Guard
10336 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
10338 PR rtl-optimization/59647
10339 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
10340 new_rtx into UNSIGNED_FLOAT rtxes.
10342 Copyright (C) 2014 Free Software Foundation, Inc.
10344 Copying and distribution of this file, with or without modification,
10345 are permitted in any medium without royalty provided the copyright
10346 notice and this notice are preserved.