1 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
3 * cselib.c (find_slot_memmode): Delete.
4 (cselib_hasher): Change compare_type to a struct.
5 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
7 (preserve_constants_and_equivs): Adjust for new compare_type.
8 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
9 (wrap_constant): Delete.
10 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
12 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
14 * doc/install.texi (Building with profile feedback): Remove
17 2014-04-26 Tom de Vries <tom@codesourcery.com>
19 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
22 2014-04-25 Cary Coutant <ccoutant@google.com>
25 * dwarf2out.c (should_move_die_to_comdat): A type definition
26 can contain a subprogram definition, but don't move it to a
28 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
29 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
31 (clone_tree_hash): Rename to...
32 (clone_tree_partial): ...this; change callers. Copy
33 DW_TAG_subprogram DIEs as declarations.
34 (copy_decls_walk): Don't copy children of a declaration into a
37 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
40 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
43 2014-04-25 Jiong Wang <jiong.wang@arm.com>
45 * config/arm/predicates.md (call_insn_operand): Add long_call check.
46 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
48 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
51 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
53 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
56 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
58 PR tree-optimization/60930
59 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
60 creating a multiply candidate by folding two constant
61 multiplicands when the result overflows.
63 2014-04-25 Jakub Jelinek <jakub@redhat.com>
65 PR tree-optimization/60960
66 * tree-vect-generic.c (expand_vector_operation): Only call
67 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
69 2014-04-25 Tom de Vries <tom@codesourcery.com>
71 * expr.c (clobber_reg_mode): New function.
72 * expr.h (clobber_reg): New function.
74 2014-04-25 Tom de Vries <tom@codesourcery.com>
76 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
79 2014-04-25 Radovan Obradovic <robradovic@mips.com>
80 Tom de Vries <tom@codesourcery.com>
82 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
84 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
85 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
86 new argument to find_all_hard_reg_sets call.
88 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
90 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
91 Use HOST_WIDE_INT_C for mask literal.
92 (aarch_rev16_shleft_mask_imm_p): Likewise.
94 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
97 * config/sparc/sparc.md (ashlsi3_extend): Delete.
99 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
101 PR preprocessor/56540
102 * config/i386/i386-c.c (ix86_target_macros): Define
103 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
105 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
107 * configure.ac (tga_func): Remove.
108 (LIB_TLS_SPEC): Remove.
109 * configure: Regenerate.
110 * config.in: Regenerate.
111 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
113 2014-04-25 Richard Biener <rguenther@suse.de>
116 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
117 call stmt use/clobber sets during stmt walk instead of
118 walking the possibly incomplete set of caller edges.
120 2014-04-25 Richard Biener <rguenther@suse.de>
123 * passes.c (apply_ipa_transforms): Inline into only caller ...
124 (execute_one_pass): ... here. Properly bring in function
125 bodies for nodes we want to apply IPA transforms to.
127 2014-04-24 Cong Hou <congh@google.com>
129 PR tree-optimization/60896
130 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
131 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
132 (vect_mark_pattern_stmts): Set the def type of all statements in
133 PATTERN_DEF_SEQ as vect_internal_def.
135 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
137 * doc/extend.texi (PowerPC Built-in Functions): Document new
138 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
140 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
142 * config/rs6000/predicates.md (const_0_to_3_operand): New
143 predicate to match 0..3 integer constants.
145 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
146 to support adding miscellaneous builtin functions.
147 (BU_DFP_MISC_2): Likewise.
148 (BU_P7_MISC_1): Likewise.
149 (BU_P7_MISC_2): Likewise.
150 (BU_P8V_MISC_3): Likewise.
151 (BU_MISC_1): Likewise.
152 (BU_MISC_2): Likewise.
153 (DIVWE): Add extended divide builtin functions.
161 (DXEX): Add decimal floating-point builtin functions.
173 (CDTBCD): Add new BCD builtin functions.
177 (BCDADD_LT): Likewise.
178 (BCDADD_EQ): Likewise.
179 (BCDADD_GT): Likewise.
180 (BCDADD_OV): Likewise.
182 (BCDSUB_LT): Likewise.
183 (BCDSUB_EQ): Likewise.
184 (BCDSUB_GT): Likewise.
185 (BCDSUB_OV): Likewise.
186 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
187 (UNPACK_TD): Likewise.
189 (UNPACK_TF): Likewise.
190 (UNPACK_TF_0): Likewise.
191 (UNPACK_TF_1): Likewise.
192 (PACK_V1TI): Likewise.
193 (UNPACK_V1TI): Likewise.
195 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
196 support for decimal floating point builtin functions.
197 (rs6000_expand_ternop_builtin): Add checks for the new builtin
198 functions that take constant arguments.
199 (rs6000_invalid_builtin): Add decimal floating point builtin
201 (rs6000_init_builtins): Setup long double, _Decimal64, and
202 _Decimal128 types for new builtin functions.
203 (builtin_function_type): Set the unsigned flags appropriately for
204 the new builtin functions.
205 (rs6000_opt_masks): Add support for decimal floating point builtin
208 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
209 floating point builtin functions.
210 (RS6000_BTM_COMMON): Likewise.
211 (RS6000_BTI_long_double): Likewise.
212 (RS6000_BTI_dfloat64): Likewise.
213 (RS6000_BTI_dfloat128): Likewise.
214 (long_double_type_internal_node): Likewise.
215 (dfloat64_type_internal_node): Likewise.
216 (dfloat128_type_internal_node): Likewise.
218 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
219 2.07 bcd arithmetic instructions.
220 (UNSPEC_BCDSUB): Likewise.
221 (UNSPEC_BCD_OVERFLOW): Likewise.
222 (UNSPEC_BCD_ADD_SUB): Likewise.
223 (bcd_add_sub): Likewise.
224 (BCD_TEST): Likewise.
225 (bcd<bcd_add_sub>): Likewise.
226 (bcd<bcd_add_sub>_test): Likewise.
227 (bcd<bcd_add_sub>_test2): Likewise.
228 (bcd<bcd_add_sub>_<code>): Likewise.
229 (peephole2 for combined bcd ops): Likewise.
231 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
232 decimal floating point builtin functions.
233 (UNSPEC_DENBCD): Likewise.
234 (UNSPEC_DXEX): Likewise.
235 (UNSPEC_DIEX): Likewise.
236 (UNSPEC_DSCLI): Likewise.
237 (UNSPEC_DSCRI): Likewise.
238 (D64_D128): Likewise.
239 (dfp_suffix): Likewise.
240 (dfp_ddedpd_<mode>): Likewise.
241 (dfp_denbcd_<mode>): Likewise.
242 (dfp_dxex_<mode>): Likewise.
243 (dfp_diex_<mode>): Likewise.
244 (dfp_dscli_<mode>): Likewise.
245 (dfp_dscri_<mode>): Likewise.
247 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
249 (UNSPEC_CDTBCD): Likewise.
250 (UNSPEC_CBCDTD): Likewise.
251 (UNSPEC_DIVE): Add support for new extended divide builtin
253 (UNSPEC_DIVEO): Likewise.
254 (UNSPEC_DIVEU): Likewise.
255 (UNSPEC_DIVEUO): Likewise.
256 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
257 pack/unpack 128-bit types.
258 (UNSPEC_PACK_128BIT): Likewise.
259 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
260 (udiv<mode>3): Use idiv_ldiv mode attribute.
261 (div<mode>3): Likewise.
262 (addg6s): Add new BCD builtin functions.
265 (UNSPEC_DIV_EXTEND): Add support for new extended divide
267 (div_extend): Likewise.
268 (div<div_extend>_<mode>"): Likewise.
269 (FP128_64): Add support for new builtin functions to pack/unpack
271 (unpack<mode>): Likewise.
272 (unpacktf_0): Likewise.
273 (unpacktf_1): Likewise.
274 (unpack<mode>_dm): Likewise.
275 (unpack<mode>_nodm): Likewise.
276 (pack<mode>): Likewise.
277 (unpackv1ti): Likewise.
278 (packv1ti): Likewise.
280 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
282 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
285 2014-04-24 Jakub Jelinek <jakub@redhat.com>
287 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
288 * gimplify.c (omp_is_private): Change last argument's type to int.
289 Only diagnose lastprivate if the simd argument is 1, only diagnose
290 linear if the simd argument is 2.
291 (gimplify_omp_for): Adjust omp_is_private callers. When adding
292 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
293 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
294 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
295 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
296 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
297 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
298 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
299 * tree-nested.c (convert_nonlocal_omp_clauses,
300 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
302 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
305 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
308 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
310 * flag-types.h (enum ivar_visibility): Add.
312 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
314 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
317 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
319 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
321 2014-04-24 Radovan Obradovic <robradovic@mips.com>
322 Tom de Vries <tom@codesourcery.com>
324 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
325 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
327 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
328 * emit-rtl.c (try_split): Same.
330 2014-04-24 Radovan Obradovic <robradovic@mips.com>
331 Tom de Vries <tom@codesourcery.com>
333 * common.opt (fuse-caller-save): New option.
335 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
337 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
340 2014-04-24 Richard Biener <rguenther@suse.de>
342 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
343 during TER and instead use the sepops interface for expanding
344 non-GIMPLE_SINGLE_RHS.
346 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
348 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
349 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
351 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
353 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
354 assembler 64-bit option.
355 * configure: Regenerate.
357 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
359 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
360 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
361 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
362 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
363 (TARGET_CRYPTO): Take TARGET_SIMD into account.
365 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
367 * config/aarch64/aarch64-builtins.c
368 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
369 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
370 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
371 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
373 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
374 (Vrevsuff): New mode attribute.
376 2014-04-24 Terry Guo <terry.guo@arm.com>
378 * config/arm/arm.h (machine_function): Define variable
379 after_arm_reorg here.
380 * config/arm/arm.c (after_arm_reorg): Remove the definition.
381 (arm_split_constant): Update the way to access variable
384 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
386 2014-04-23 Tom de Vries <tom@codesourcery.com>
388 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
390 2014-04-23 David Malcolm <dmalcolm@redhat.com>
392 * is-a.h: Update comments to reflect the following changes to the
393 "pointerness" of the API, making the template parameter match the
394 return type, allowing use of is-a.h with typedefs of pointers.
395 (is_a_helper::cast): Return a T rather then a pointer to a T, so
396 that the return type matches the parameter to the is_a_helper.
398 (dyn_cast): Likewise.
400 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
401 pointer from the is-a.h API.
403 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
404 (is_a_helper <cgraph_node *>::test): ...this, matching change to
406 (is_a_helper <varpool_node>::test): Likewise, convert to...
407 (is_a_helper <varpool_node *>::test): ...this.
409 (varpool_first_variable): Update for removal of implicit pointer
411 (varpool_next_variable): Likewise.
412 (varpool_first_static_initializer): Likewise.
413 (varpool_next_static_initializer): Likewise.
414 (varpool_first_defined_variable): Likewise.
415 (varpool_next_defined_variable): Likewise.
416 (cgraph_first_defined_function): Likewise.
417 (cgraph_next_defined_function): Likewise.
418 (cgraph_first_function): Likewise.
419 (cgraph_next_function): Likewise.
420 (cgraph_first_function_with_gimple_body): Likewise.
421 (cgraph_next_function_with_gimple_body): Likewise.
422 (cgraph_alias_target): Likewise.
423 (varpool_alias_target): Likewise.
424 (cgraph_function_or_thunk_node): Likewise.
425 (varpool_variable_node): Likewise.
426 (symtab_real_symbol_p): Likewise.
427 * cgraphunit.c (referred_to_p): Likewise.
428 (analyze_functions): Likewise.
429 (handle_alias_pairs): Likewise.
430 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
431 * gimple-ssa.h (gimple_vuse_op): Likewise.
432 (gimple_vdef_op): Likewise.
433 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
434 * gimple.c (gimple_build_asm_1): Likewise.
435 (gimple_build_try): Likewise.
436 (gimple_build_resx): Likewise.
437 (gimple_build_eh_dispatch): Likewise.
438 (gimple_build_omp_for): Likewise.
439 (gimple_omp_for_set_clauses): Likewise.
441 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
442 (is_a_helper <gimple_statement_asm *>::test): ...this.
443 (is_a_helper <gimple_statement_bind>::test): Convert to...
444 (is_a_helper <gimple_statement_bind *>::test): ...this.
445 (is_a_helper <gimple_statement_call>::test): Convert to...
446 (is_a_helper <gimple_statement_call *>::test): ...this.
447 (is_a_helper <gimple_statement_catch>::test): Convert to...
448 (is_a_helper <gimple_statement_catch *>::test): ...this.
449 (is_a_helper <gimple_statement_resx>::test): Convert to...
450 (is_a_helper <gimple_statement_resx *>::test): ...this.
451 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
452 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
453 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
454 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
455 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
456 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
457 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
458 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
459 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
460 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
461 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
462 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
463 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
464 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
465 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
466 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
467 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
468 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
469 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
470 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
471 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
472 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
473 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
474 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
475 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
476 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
477 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
478 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
479 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
480 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
481 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
482 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
483 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
484 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
485 (is_a_helper <gimple_statement_phi>::test): Convert to...
486 (is_a_helper <gimple_statement_phi *>::test): ...this.
487 (is_a_helper <gimple_statement_transaction>::test): Convert to...
488 (is_a_helper <gimple_statement_transaction *>::test): ...this.
489 (is_a_helper <gimple_statement_try>::test): Convert to...
490 (is_a_helper <gimple_statement_try *>::test): ...this.
491 (is_a_helper <gimple_statement_wce>::test): Convert to...
492 (is_a_helper <gimple_statement_wce *>::test): ...this.
493 (is_a_helper <const gimple_statement_asm>::test): Convert to...
494 (is_a_helper <const gimple_statement_asm *>::test): ...this.
495 (is_a_helper <const gimple_statement_bind>::test): Convert to...
496 (is_a_helper <const gimple_statement_bind *>::test): ...this.
497 (is_a_helper <const gimple_statement_call>::test): Convert to...
498 (is_a_helper <const gimple_statement_call *>::test): ...this.
499 (is_a_helper <const gimple_statement_catch>::test): Convert to...
500 (is_a_helper <const gimple_statement_catch *>::test): ...this.
501 (is_a_helper <const gimple_statement_resx>::test): Convert to...
502 (is_a_helper <const gimple_statement_resx *>::test): ...this.
503 (is_a_helper <const gimple_statement_eh_dispatch>::test):
505 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
507 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
509 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
510 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
512 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
514 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
516 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
518 (is_a_helper <const gimple_statement_omp_return>::test): Convert
520 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
521 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
523 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
524 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
526 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
527 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
528 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
529 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
531 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
532 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
534 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
535 (is_a_helper <const gimple_statement_omp_target>::test): Convert
537 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
538 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
540 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
541 (is_a_helper <const gimple_statement_omp_single>::test): Convert
543 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
544 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
546 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
547 (is_a_helper <const gimple_statement_omp_task>::test): Convert
549 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
550 (is_a_helper <const gimple_statement_phi>::test): Convert to...
551 (is_a_helper <const gimple_statement_phi *>::test): ...this.
552 (is_a_helper <const gimple_statement_transaction>::test): Convert
554 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
555 (is_a_helper <const gimple_statement_with_ops>::test): Convert
557 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
558 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
559 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
560 (is_a_helper <const gimple_statement_with_memory_ops>::test):
562 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
564 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
566 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
568 (gimple_use_ops): Update for removal of implicit pointer from the
570 (gimple_set_use_ops): Likewise.
571 (gimple_vuse): Likewise.
572 (gimple_vdef): Likewise.
573 (gimple_vuse_ptr): Likewise.
574 (gimple_vdef_ptr): Likewise.
575 (gimple_set_vuse): Likewise.
576 (gimple_set_vdef): Likewise.
577 (gimple_omp_return_set_lhs): Likewise.
578 (gimple_omp_return_lhs): Likewise.
579 (gimple_omp_return_lhs_ptr): Likewise.
580 (gimple_call_fntype): Likewise.
581 (gimple_call_set_fntype): Likewise.
582 (gimple_call_set_internal_fn): Likewise.
583 (gimple_call_use_set): Likewise.
584 (gimple_call_clobber_set): Likewise.
585 (gimple_bind_vars): Likewise.
586 (gimple_bind_set_vars): Likewise.
587 (gimple_bind_body_ptr): Likewise.
588 (gimple_bind_set_body): Likewise.
589 (gimple_bind_add_stmt): Likewise.
590 (gimple_bind_block): Likewise.
591 (gimple_bind_set_block): Likewise.
592 (gimple_asm_ninputs): Likewise.
593 (gimple_asm_noutputs): Likewise.
594 (gimple_asm_nclobbers): Likewise.
595 (gimple_asm_nlabels): Likewise.
596 (gimple_asm_input_op): Likewise.
597 (gimple_asm_input_op_ptr): Likewise.
598 (gimple_asm_output_op): Likewise.
599 (gimple_asm_output_op_ptr): Likewise.
600 (gimple_asm_set_output_op): Likewise.
601 (gimple_asm_clobber_op): Likewise.
602 (gimple_asm_set_clobber_op): Likewise.
603 (gimple_asm_label_op): Likewise.
604 (gimple_asm_set_label_op): Likewise.
605 (gimple_asm_string): Likewise.
606 (gimple_catch_types): Likewise.
607 (gimple_catch_types_ptr): Likewise.
608 (gimple_catch_handler_ptr): Likewise.
609 (gimple_catch_set_types): Likewise.
610 (gimple_catch_set_handler): Likewise.
611 (gimple_eh_filter_types): Likewise.
612 (gimple_eh_filter_types_ptr): Likewise.
613 (gimple_eh_filter_failure_ptr): Likewise.
614 (gimple_eh_filter_set_types): Likewise.
615 (gimple_eh_filter_set_failure): Likewise.
616 (gimple_eh_must_not_throw_fndecl): Likewise.
617 (gimple_eh_must_not_throw_set_fndecl): Likewise.
618 (gimple_eh_else_n_body_ptr): Likewise.
619 (gimple_eh_else_e_body_ptr): Likewise.
620 (gimple_eh_else_set_n_body): Likewise.
621 (gimple_eh_else_set_e_body): Likewise.
622 (gimple_try_eval_ptr): Likewise.
623 (gimple_try_cleanup_ptr): Likewise.
624 (gimple_try_set_eval): Likewise.
625 (gimple_try_set_cleanup): Likewise.
626 (gimple_wce_cleanup_ptr): Likewise.
627 (gimple_wce_set_cleanup): Likewise.
628 (gimple_phi_capacity): Likewise.
629 (gimple_phi_num_args): Likewise.
630 (gimple_phi_result): Likewise.
631 (gimple_phi_result_ptr): Likewise.
632 (gimple_phi_set_result): Likewise.
633 (gimple_phi_arg): Likewise.
634 (gimple_phi_set_arg): Likewise.
635 (gimple_resx_region): Likewise.
636 (gimple_resx_set_region): Likewise.
637 (gimple_eh_dispatch_region): Likewise.
638 (gimple_eh_dispatch_set_region): Likewise.
639 (gimple_omp_critical_name): Likewise.
640 (gimple_omp_critical_name_ptr): Likewise.
641 (gimple_omp_critical_set_name): Likewise.
642 (gimple_omp_for_clauses): Likewise.
643 (gimple_omp_for_clauses_ptr): Likewise.
644 (gimple_omp_for_set_clauses): Likewise.
645 (gimple_omp_for_collapse): Likewise.
646 (gimple_omp_for_index): Likewise.
647 (gimple_omp_for_index_ptr): Likewise.
648 (gimple_omp_for_set_index): Likewise.
649 (gimple_omp_for_initial): Likewise.
650 (gimple_omp_for_initial_ptr): Likewise.
651 (gimple_omp_for_set_initial): Likewise.
652 (gimple_omp_for_final): Likewise.
653 (gimple_omp_for_final_ptr): Likewise.
654 (gimple_omp_for_set_final): Likewise.
655 (gimple_omp_for_incr): Likewise.
656 (gimple_omp_for_incr_ptr): Likewise.
657 (gimple_omp_for_set_incr): Likewise.
658 (gimple_omp_for_pre_body_ptr): Likewise.
659 (gimple_omp_for_set_pre_body): Likewise.
660 (gimple_omp_parallel_clauses): Likewise.
661 (gimple_omp_parallel_clauses_ptr): Likewise.
662 (gimple_omp_parallel_set_clauses): Likewise.
663 (gimple_omp_parallel_child_fn): Likewise.
664 (gimple_omp_parallel_child_fn_ptr): Likewise.
665 (gimple_omp_parallel_set_child_fn): Likewise.
666 (gimple_omp_parallel_data_arg): Likewise.
667 (gimple_omp_parallel_data_arg_ptr): Likewise.
668 (gimple_omp_parallel_set_data_arg): Likewise.
669 (gimple_omp_task_clauses): Likewise.
670 (gimple_omp_task_clauses_ptr): Likewise.
671 (gimple_omp_task_set_clauses): Likewise.
672 (gimple_omp_task_child_fn): Likewise.
673 (gimple_omp_task_child_fn_ptr): Likewise.
674 (gimple_omp_task_set_child_fn): Likewise.
675 (gimple_omp_task_data_arg): Likewise.
676 (gimple_omp_task_data_arg_ptr): Likewise.
677 (gimple_omp_task_set_data_arg): Likewise.
678 (gimple_omp_taskreg_clauses): Likewise.
679 (gimple_omp_taskreg_clauses_ptr): Likewise.
680 (gimple_omp_taskreg_set_clauses): Likewise.
681 (gimple_omp_taskreg_child_fn): Likewise.
682 (gimple_omp_taskreg_child_fn_ptr): Likewise.
683 (gimple_omp_taskreg_set_child_fn): Likewise.
684 (gimple_omp_taskreg_data_arg): Likewise.
685 (gimple_omp_taskreg_data_arg_ptr): Likewise.
686 (gimple_omp_taskreg_set_data_arg): Likewise.
687 (gimple_omp_task_copy_fn): Likewise.
688 (gimple_omp_task_copy_fn_ptr): Likewise.
689 (gimple_omp_task_set_copy_fn): Likewise.
690 (gimple_omp_task_arg_size): Likewise.
691 (gimple_omp_task_arg_size_ptr): Likewise.
692 (gimple_omp_task_set_arg_size): Likewise.
693 (gimple_omp_task_arg_align): Likewise.
694 (gimple_omp_task_arg_align_ptr): Likewise.
695 (gimple_omp_task_set_arg_align): Likewise.
696 (gimple_omp_single_clauses): Likewise.
697 (gimple_omp_single_clauses_ptr): Likewise.
698 (gimple_omp_single_set_clauses): Likewise.
699 (gimple_omp_target_clauses): Likewise.
700 (gimple_omp_target_clauses_ptr): Likewise.
701 (gimple_omp_target_set_clauses): Likewise.
702 (gimple_omp_target_child_fn): Likewise.
703 (gimple_omp_target_child_fn_ptr): Likewise.
704 (gimple_omp_target_set_child_fn): Likewise.
705 (gimple_omp_target_data_arg): Likewise.
706 (gimple_omp_target_data_arg_ptr): Likewise.
707 (gimple_omp_target_set_data_arg): Likewise.
708 (gimple_omp_teams_clauses): Likewise.
709 (gimple_omp_teams_clauses_ptr): Likewise.
710 (gimple_omp_teams_set_clauses): Likewise.
711 (gimple_omp_sections_clauses): Likewise.
712 (gimple_omp_sections_clauses_ptr): Likewise.
713 (gimple_omp_sections_set_clauses): Likewise.
714 (gimple_omp_sections_control): Likewise.
715 (gimple_omp_sections_control_ptr): Likewise.
716 (gimple_omp_sections_set_control): Likewise.
717 (gimple_omp_for_set_cond): Likewise.
718 (gimple_omp_for_cond): Likewise.
719 (gimple_omp_atomic_store_set_val): Likewise.
720 (gimple_omp_atomic_store_val): Likewise.
721 (gimple_omp_atomic_store_val_ptr): Likewise.
722 (gimple_omp_atomic_load_set_lhs): Likewise.
723 (gimple_omp_atomic_load_lhs): Likewise.
724 (gimple_omp_atomic_load_lhs_ptr): Likewise.
725 (gimple_omp_atomic_load_set_rhs): Likewise.
726 (gimple_omp_atomic_load_rhs): Likewise.
727 (gimple_omp_atomic_load_rhs_ptr): Likewise.
728 (gimple_omp_continue_control_def): Likewise.
729 (gimple_omp_continue_control_def_ptr): Likewise.
730 (gimple_omp_continue_set_control_def): Likewise.
731 (gimple_omp_continue_control_use): Likewise.
732 (gimple_omp_continue_control_use_ptr): Likewise.
733 (gimple_omp_continue_set_control_use): Likewise.
734 (gimple_transaction_body_ptr): Likewise.
735 (gimple_transaction_label): Likewise.
736 (gimple_transaction_label_ptr): Likewise.
737 (gimple_transaction_set_body): Likewise.
738 (gimple_transaction_set_label): Likewise.
740 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
741 * ipa-inline-analysis.c (inline_write_summary): Likewise.
742 * ipa-ref.c (ipa_record_reference): Likewise.
743 * ipa-reference.c (analyze_function): Likewise.
744 (ipa_reference_write_optimization_summary): Likewise.
745 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
746 (address_taken_from_non_vtable_p): Likewise.
747 (comdat_can_be_unshared_p_1): Likewise.
748 * lto-cgraph.c (lto_output_ref): Likewise.
749 (add_references): Likewise.
750 (compute_ltrans_boundary): Likewise.
751 (output_symtab): Likewise.
752 (input_ref): Likewise.
753 (input_cgraph_1): Likewise.
754 (output_cgraph_opt_summary): Likewise.
755 * lto-streamer-out.c (lto_output): Likewise.
756 (output_symbol_p): Likewise.
757 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
758 (lsei_start_function_in_partition): Likewise.
759 (lsei_next_variable_in_partition): Likewise.
760 (lsei_start_variable_in_partition): Likewise.
761 * symtab.c (insert_to_assembler_name_hash): Likewise.
762 (unlink_from_assembler_name_hash): Likewise.
763 (symtab_unregister_node): Likewise.
764 (symtab_remove_node): Likewise.
765 (dump_symtab_node): Likewise.
766 (verify_symtab_base): Likewise.
767 (verify_symtab_node): Likewise.
768 (symtab_make_decl_local): Likewise.
769 (symtab_alias_ultimate_target): Likewise.
770 (symtab_resolve_alias): Likewise.
771 (symtab_get_symbol_partitioning_class): Likewise.
772 * tree-phinodes.c (allocate_phi_node): Likewise.
773 (reserve_phi_args_for_new_edge): Likewise.
774 (remove_phi_args): Likewise.
775 * varpool.c (varpool_node_for_asm): Likewise.
776 (varpool_remove_unreferenced_decls): Likewise.
778 2014-04-23 Jeff Law <law@redhat.com>
780 PR tree-optimization/60902
781 * tree-ssa-threadedge.c
782 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
783 invalidate outputs from statements that do not produce useful
784 outputs for threading.
786 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
788 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
789 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
790 machine descriptions for Stack Smashing Protector.
792 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
794 * aarch64.md (<optab>_rol<mode>3): New pattern.
795 (<optab>_rolsi3_uxtw): Likewise.
796 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
798 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
800 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
801 (arm_cortex_a12_tune): Likewise.
803 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
805 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
807 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
809 * config/arm/arm.md (arm_rev16si2): New pattern.
810 (arm_rev16si2_alt): Likewise.
811 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
813 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
815 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
816 (rev16<mode>2_alt): Likewise.
817 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
818 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
819 (aarch_rev16_shleft_mask_imm_p): Likewise.
820 (aarch_rev16_p_1): Likewise.
821 (aarch_rev16_p): Likewise.
822 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
823 (aarch_rev16_shright_mask_imm_p): Likewise.
824 (aarch_rev16_shleft_mask_imm_p): Likewise.
826 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
828 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
829 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
831 (cortex_a53_extra_costs): Likewise.
832 (cortex_a57_extra_costs): Likewise.
833 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
834 (cortexa7_extra_costs): Likewise.
835 (cortexa8_extra_costs): Likewise.
836 (cortexa12_extra_costs): Likewise.
837 (cortexa15_extra_costs): Likewise.
838 (v7m_extra_costs): Likewise.
839 (arm_new_rtx_costs): Handle BSWAP.
841 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
843 * config/arm/arm.c (cortexa8_extra_costs): New table.
844 (arm_cortex_a8_tune): New tuning struct.
845 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
847 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
849 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
851 2014-04-23 Richard Biener <rguenther@suse.de>
853 * Makefile.in (OBJS): Remove loop-unswitch.o.
854 * loop-unswitch.c: Delete.
855 * tree-pass.h (make_pass_rtl_unswitch): Remove.
856 * passes.def (pass_rtl_unswitch): Likewise.
857 * loop-init.c (gate_rtl_unswitch): Likewise.
858 (rtl_unswitch): Likewise.
859 (pass_data_rtl_unswitch): Likewise.
860 (pass_rtl_unswitch): Likewise.
861 (make_pass_rtl_unswitch): Likewise.
862 * rtl.h (reversed_condition): Likewise.
863 (compare_and_jump_seq): Likewise.
864 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
866 * loop-unroll.c (compare_and_jump_seq): Likewise.
868 2014-04-23 Richard Biener <rguenther@suse.de>
870 PR tree-optimization/60903
871 * tree-ssa-loop-im.c (analyze_memory_references): Remove
872 commented code block.
873 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
874 loop flags to newly created BBs and edges.
876 2014-04-23 Nick Clifton <nickc@redhat.com>
878 * config/msp430/msp430.c (msp430_handle_option): Move function
880 (msp430_option_override): Simplify mcu and mcpu option handling.
881 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
882 support for -mhwmult command line option.
883 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
884 -mhwmult command line option.
885 (msp430_hwmult_enabled): Delete.
886 (msp43o_output_labelref): Add support for -mhwmult command line
888 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
889 (umulsidi3): Likewise.
890 * config/msp430/msp430.opt (mmcu): Add Report attribute.
891 (mcpu, mlarge, msmall): Likewise.
892 (mhwmult): New option.
893 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
895 (msp430_is_f5_mcu): Remove prototype.
896 (msp430_use_f5_series_hwmult): Add prototype.
897 * config/msp430/msp430-opts.h: New file.
898 * common/config/msp430: New directory.
899 * common/config/msp430/msp430-common.c: New file.
900 * config.gcc (msp430): Remove target_has_targetm_common.
901 * doc/invoke.texi: Document -mhwmult command line option.
903 2014-04-23 Nick Clifton <nickc@redhat.com>
905 * config/i386/cygwin.h (ENDFILE_SPEC): Include
906 default-manifest.o if it can be found in the search path.
907 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
909 2014-04-23 Terry Guo <terry.guo@arm.com>
911 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
913 2014-04-23 Richard Biener <rguenther@suse.de>
916 * tree-inline.c (declare_return_variable): Use mark_addressable.
918 2014-04-23 Richard Biener <rguenther@suse.de>
921 * loop-init.c (loop_optimizer_init): Make sure to apply
922 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
924 2014-04-22 Jakub Jelinek <jakub@redhat.com>
927 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
929 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
930 if flag_sanitize_undefined_trap_on_error.
931 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
932 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
933 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
934 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
935 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
936 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
937 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
938 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
939 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
940 * ubsan.c (ubsan_instrument_unreachable): Return
941 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
942 (ubsan_expand_null_ifn): Emit __builtin_trap ()
943 if flag_sanitize_undefined_trap_on_error and
944 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
945 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
946 instrument_bool_enum_load): Emit __builtin_trap () if
947 flag_sanitize_undefined_trap_on_error and
948 __builtin_handle_*_abort () if !flag_sanitize_recover.
949 * doc/invoke.texi (-fsanitize-recover,
950 -fsanitize-undefined-trap-on-error): Document.
952 2014-04-22 Christian Bruel <christian.bruel@st.com>
954 * config/sh/sh.md (mov<mode>): Replace movQIHI.
955 Force immediates to SImode.
957 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
959 * config/nios2/nios2.md (UNSPEC_ROUND): New.
961 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
962 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
963 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
964 (nios2_fpu_insn): Add entry for round.
965 (N2FPU_NO_ERRNO_P): Define.
966 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
968 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
970 2014-04-22 Richard Henderson <rth@redhat.com>
972 * config/aarch64/aarch64 (addti3, subti3): New expanders.
973 (add<GPI>3_compare0): Remove leading * from name.
974 (add<GPI>3_carryin): Likewise.
975 (sub<GPI>3_compare0): Likewise.
976 (sub<GPI>3_carryin): Likewise.
977 (<su_optab>mulditi3): New expander.
978 (multi3): New expander.
979 (madd<GPI>): Remove leading * from name.
981 2014-04-22 Martin Jambor <mjambor@suse.cz>
983 * cgraphclones.c (cgraph_function_versioning): Copy
984 ipa_transforms_to_apply instead of asserting it is empty.
986 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
989 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
990 on count_exp to get mode.
992 2014-04-22 Andrew Pinski <apinski@cavium.com>
994 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
995 Handle TLS for ILP32.
996 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
997 (tlsie_small_<mode>): this and handle PTR.
998 (tlsie_small_sidi): New pattern.
999 (tlsle_small): Change to an expand to handle ILP32.
1000 (tlsle_small_<mode>): New pattern.
1001 (tlsdesc_small): Rename to ...
1002 (tlsdesc_small_<mode>): this and handle PTR.
1004 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1006 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1008 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1010 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1011 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1012 (aarch64_types_signed_poly_qualifiers): Likewise.
1013 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1014 (aarch64_types_poly_signed_qualifiers): Likewise.
1015 (TYPES_REINTERP_SS): Type macro added.
1016 (TYPES_REINTERP_SU): Likewise.
1017 (TYPES_REINTERP_SP): Likewise.
1018 (TYPES_REINTERP_US): Likewise.
1019 (TYPES_REINTERP_PS): Likewise.
1020 (aarch64_fold_builtin): New expression folding added.
1021 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1022 Declarations removed.
1023 (REINTERP_SS): Declarations added.
1024 (REINTERP_US): Likewise.
1025 (REINTERP_PS): Likewise.
1026 (REINTERP_SU): Likewise.
1027 (REINTERP_SP): Likewise.
1028 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1029 (vreinterpretq_p8_f64): Likewise.
1030 (vreinterpret_p16_f64): Likewise.
1031 (vreinterpretq_p16_f64): Likewise.
1032 (vreinterpret_f32_f64): Likewise.
1033 (vreinterpretq_f32_f64): Likewise.
1034 (vreinterpret_f64_f32): Likewise.
1035 (vreinterpret_f64_p8): Likewise.
1036 (vreinterpret_f64_p16): Likewise.
1037 (vreinterpret_f64_s8): Likewise.
1038 (vreinterpret_f64_s16): Likewise.
1039 (vreinterpret_f64_s32): Likewise.
1040 (vreinterpret_f64_s64): Likewise.
1041 (vreinterpret_f64_u8): Likewise.
1042 (vreinterpret_f64_u16): Likewise.
1043 (vreinterpret_f64_u32): Likewise.
1044 (vreinterpret_f64_u64): Likewise.
1045 (vreinterpretq_f64_f32): Likewise.
1046 (vreinterpretq_f64_p8): Likewise.
1047 (vreinterpretq_f64_p16): Likewise.
1048 (vreinterpretq_f64_s8): Likewise.
1049 (vreinterpretq_f64_s16): Likewise.
1050 (vreinterpretq_f64_s32): Likewise.
1051 (vreinterpretq_f64_s64): Likewise.
1052 (vreinterpretq_f64_u8): Likewise.
1053 (vreinterpretq_f64_u16): Likewise.
1054 (vreinterpretq_f64_u32): Likewise.
1055 (vreinterpretq_f64_u64): Likewise.
1056 (vreinterpret_s64_f64): Likewise.
1057 (vreinterpretq_s64_f64): Likewise.
1058 (vreinterpret_u64_f64): Likewise.
1059 (vreinterpretq_u64_f64): Likewise.
1060 (vreinterpret_s8_f64): Likewise.
1061 (vreinterpretq_s8_f64): Likewise.
1062 (vreinterpret_s16_f64): Likewise.
1063 (vreinterpretq_s16_f64): Likewise.
1064 (vreinterpret_s32_f64): Likewise.
1065 (vreinterpretq_s32_f64): Likewise.
1066 (vreinterpret_u8_f64): Likewise.
1067 (vreinterpretq_u8_f64): Likewise.
1068 (vreinterpret_u16_f64): Likewise.
1069 (vreinterpretq_u16_f64): Likewise.
1070 (vreinterpret_u32_f64): Likewise.
1071 (vreinterpretq_u32_f64): Likewise.
1073 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1075 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1076 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1077 (vreinterpret_p8_s8): Likewise.
1078 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1079 (vreinterpret_p8_s16): Likewise.
1080 (vreinterpret_p8_s32): Likewise.
1081 (vreinterpret_p8_s64): Likewise.
1082 (vreinterpret_p8_f32): Likewise.
1083 (vreinterpret_p8_u8): Likewise.
1084 (vreinterpret_p8_u16): Likewise.
1085 (vreinterpret_p8_u32): Likewise.
1086 (vreinterpret_p8_u64): Likewise.
1087 (vreinterpret_p8_p16): Likewise.
1088 (vreinterpretq_p8_s8): Likewise.
1089 (vreinterpretq_p8_s16): Likewise.
1090 (vreinterpretq_p8_s32): Likewise.
1091 (vreinterpretq_p8_s64): Likewise.
1092 (vreinterpretq_p8_f32): Likewise.
1093 (vreinterpretq_p8_u8): Likewise.
1094 (vreinterpretq_p8_u16): Likewise.
1095 (vreinterpretq_p8_u32): Likewise.
1096 (vreinterpretq_p8_u64): Likewise.
1097 (vreinterpretq_p8_p16): Likewise.
1098 (vreinterpret_p16_s8): Likewise.
1099 (vreinterpret_p16_s16): Likewise.
1100 (vreinterpret_p16_s32): Likewise.
1101 (vreinterpret_p16_s64): Likewise.
1102 (vreinterpret_p16_f32): Likewise.
1103 (vreinterpret_p16_u8): Likewise.
1104 (vreinterpret_p16_u16): Likewise.
1105 (vreinterpret_p16_u32): Likewise.
1106 (vreinterpret_p16_u64): Likewise.
1107 (vreinterpret_p16_p8): Likewise.
1108 (vreinterpretq_p16_s8): Likewise.
1109 (vreinterpretq_p16_s16): Likewise.
1110 (vreinterpretq_p16_s32): Likewise.
1111 (vreinterpretq_p16_s64): Likewise.
1112 (vreinterpretq_p16_f32): Likewise.
1113 (vreinterpretq_p16_u8): Likewise.
1114 (vreinterpretq_p16_u16): Likewise.
1115 (vreinterpretq_p16_u32): Likewise.
1116 (vreinterpretq_p16_u64): Likewise.
1117 (vreinterpretq_p16_p8): Likewise.
1118 (vreinterpret_f32_s8): Likewise.
1119 (vreinterpret_f32_s16): Likewise.
1120 (vreinterpret_f32_s32): Likewise.
1121 (vreinterpret_f32_s64): Likewise.
1122 (vreinterpret_f32_u8): Likewise.
1123 (vreinterpret_f32_u16): Likewise.
1124 (vreinterpret_f32_u32): Likewise.
1125 (vreinterpret_f32_u64): Likewise.
1126 (vreinterpret_f32_p8): Likewise.
1127 (vreinterpret_f32_p16): Likewise.
1128 (vreinterpretq_f32_s8): Likewise.
1129 (vreinterpretq_f32_s16): Likewise.
1130 (vreinterpretq_f32_s32): Likewise.
1131 (vreinterpretq_f32_s64): Likewise.
1132 (vreinterpretq_f32_u8): Likewise.
1133 (vreinterpretq_f32_u16): Likewise.
1134 (vreinterpretq_f32_u32): Likewise.
1135 (vreinterpretq_f32_u64): Likewise.
1136 (vreinterpretq_f32_p8): Likewise.
1137 (vreinterpretq_f32_p16): Likewise.
1138 (vreinterpret_s64_s8): Likewise.
1139 (vreinterpret_s64_s16): Likewise.
1140 (vreinterpret_s64_s32): Likewise.
1141 (vreinterpret_s64_f32): Likewise.
1142 (vreinterpret_s64_u8): Likewise.
1143 (vreinterpret_s64_u16): Likewise.
1144 (vreinterpret_s64_u32): Likewise.
1145 (vreinterpret_s64_u64): Likewise.
1146 (vreinterpret_s64_p8): Likewise.
1147 (vreinterpret_s64_p16): Likewise.
1148 (vreinterpretq_s64_s8): Likewise.
1149 (vreinterpretq_s64_s16): Likewise.
1150 (vreinterpretq_s64_s32): Likewise.
1151 (vreinterpretq_s64_f32): Likewise.
1152 (vreinterpretq_s64_u8): Likewise.
1153 (vreinterpretq_s64_u16): Likewise.
1154 (vreinterpretq_s64_u32): Likewise.
1155 (vreinterpretq_s64_u64): Likewise.
1156 (vreinterpretq_s64_p8): Likewise.
1157 (vreinterpretq_s64_p16): Likewise.
1158 (vreinterpret_u64_s8): Likewise.
1159 (vreinterpret_u64_s16): Likewise.
1160 (vreinterpret_u64_s32): Likewise.
1161 (vreinterpret_u64_s64): Likewise.
1162 (vreinterpret_u64_f32): Likewise.
1163 (vreinterpret_u64_u8): Likewise.
1164 (vreinterpret_u64_u16): Likewise.
1165 (vreinterpret_u64_u32): Likewise.
1166 (vreinterpret_u64_p8): Likewise.
1167 (vreinterpret_u64_p16): Likewise.
1168 (vreinterpretq_u64_s8): Likewise.
1169 (vreinterpretq_u64_s16): Likewise.
1170 (vreinterpretq_u64_s32): Likewise.
1171 (vreinterpretq_u64_s64): Likewise.
1172 (vreinterpretq_u64_f32): Likewise.
1173 (vreinterpretq_u64_u8): Likewise.
1174 (vreinterpretq_u64_u16): Likewise.
1175 (vreinterpretq_u64_u32): Likewise.
1176 (vreinterpretq_u64_p8): Likewise.
1177 (vreinterpretq_u64_p16): Likewise.
1178 (vreinterpret_s8_s16): Likewise.
1179 (vreinterpret_s8_s32): Likewise.
1180 (vreinterpret_s8_s64): Likewise.
1181 (vreinterpret_s8_f32): Likewise.
1182 (vreinterpret_s8_u8): Likewise.
1183 (vreinterpret_s8_u16): Likewise.
1184 (vreinterpret_s8_u32): Likewise.
1185 (vreinterpret_s8_u64): Likewise.
1186 (vreinterpret_s8_p8): Likewise.
1187 (vreinterpret_s8_p16): Likewise.
1188 (vreinterpretq_s8_s16): Likewise.
1189 (vreinterpretq_s8_s32): Likewise.
1190 (vreinterpretq_s8_s64): Likewise.
1191 (vreinterpretq_s8_f32): Likewise.
1192 (vreinterpretq_s8_u8): Likewise.
1193 (vreinterpretq_s8_u16): Likewise.
1194 (vreinterpretq_s8_u32): Likewise.
1195 (vreinterpretq_s8_u64): Likewise.
1196 (vreinterpretq_s8_p8): Likewise.
1197 (vreinterpretq_s8_p16): Likewise.
1198 (vreinterpret_s16_s8): Likewise.
1199 (vreinterpret_s16_s32): Likewise.
1200 (vreinterpret_s16_s64): Likewise.
1201 (vreinterpret_s16_f32): Likewise.
1202 (vreinterpret_s16_u8): Likewise.
1203 (vreinterpret_s16_u16): Likewise.
1204 (vreinterpret_s16_u32): Likewise.
1205 (vreinterpret_s16_u64): Likewise.
1206 (vreinterpret_s16_p8): Likewise.
1207 (vreinterpret_s16_p16): Likewise.
1208 (vreinterpretq_s16_s8): Likewise.
1209 (vreinterpretq_s16_s32): Likewise.
1210 (vreinterpretq_s16_s64): Likewise.
1211 (vreinterpretq_s16_f32): Likewise.
1212 (vreinterpretq_s16_u8): Likewise.
1213 (vreinterpretq_s16_u16): Likewise.
1214 (vreinterpretq_s16_u32): Likewise.
1215 (vreinterpretq_s16_u64): Likewise.
1216 (vreinterpretq_s16_p8): Likewise.
1217 (vreinterpretq_s16_p16): Likewise.
1218 (vreinterpret_s32_s8): Likewise.
1219 (vreinterpret_s32_s16): Likewise.
1220 (vreinterpret_s32_s64): Likewise.
1221 (vreinterpret_s32_f32): Likewise.
1222 (vreinterpret_s32_u8): Likewise.
1223 (vreinterpret_s32_u16): Likewise.
1224 (vreinterpret_s32_u32): Likewise.
1225 (vreinterpret_s32_u64): Likewise.
1226 (vreinterpret_s32_p8): Likewise.
1227 (vreinterpret_s32_p16): Likewise.
1228 (vreinterpretq_s32_s8): Likewise.
1229 (vreinterpretq_s32_s16): Likewise.
1230 (vreinterpretq_s32_s64): Likewise.
1231 (vreinterpretq_s32_f32): Likewise.
1232 (vreinterpretq_s32_u8): Likewise.
1233 (vreinterpretq_s32_u16): Likewise.
1234 (vreinterpretq_s32_u32): Likewise.
1235 (vreinterpretq_s32_u64): Likewise.
1236 (vreinterpretq_s32_p8): Likewise.
1237 (vreinterpretq_s32_p16): Likewise.
1238 (vreinterpret_u8_s8): Likewise.
1239 (vreinterpret_u8_s16): Likewise.
1240 (vreinterpret_u8_s32): Likewise.
1241 (vreinterpret_u8_s64): Likewise.
1242 (vreinterpret_u8_f32): Likewise.
1243 (vreinterpret_u8_u16): Likewise.
1244 (vreinterpret_u8_u32): Likewise.
1245 (vreinterpret_u8_u64): Likewise.
1246 (vreinterpret_u8_p8): Likewise.
1247 (vreinterpret_u8_p16): Likewise.
1248 (vreinterpretq_u8_s8): Likewise.
1249 (vreinterpretq_u8_s16): Likewise.
1250 (vreinterpretq_u8_s32): Likewise.
1251 (vreinterpretq_u8_s64): Likewise.
1252 (vreinterpretq_u8_f32): Likewise.
1253 (vreinterpretq_u8_u16): Likewise.
1254 (vreinterpretq_u8_u32): Likewise.
1255 (vreinterpretq_u8_u64): Likewise.
1256 (vreinterpretq_u8_p8): Likewise.
1257 (vreinterpretq_u8_p16): Likewise.
1258 (vreinterpret_u16_s8): Likewise.
1259 (vreinterpret_u16_s16): Likewise.
1260 (vreinterpret_u16_s32): Likewise.
1261 (vreinterpret_u16_s64): Likewise.
1262 (vreinterpret_u16_f32): Likewise.
1263 (vreinterpret_u16_u8): Likewise.
1264 (vreinterpret_u16_u32): Likewise.
1265 (vreinterpret_u16_u64): Likewise.
1266 (vreinterpret_u16_p8): Likewise.
1267 (vreinterpret_u16_p16): Likewise.
1268 (vreinterpretq_u16_s8): Likewise.
1269 (vreinterpretq_u16_s16): Likewise.
1270 (vreinterpretq_u16_s32): Likewise.
1271 (vreinterpretq_u16_s64): Likewise.
1272 (vreinterpretq_u16_f32): Likewise.
1273 (vreinterpretq_u16_u8): Likewise.
1274 (vreinterpretq_u16_u32): Likewise.
1275 (vreinterpretq_u16_u64): Likewise.
1276 (vreinterpretq_u16_p8): Likewise.
1277 (vreinterpretq_u16_p16): Likewise.
1278 (vreinterpret_u32_s8): Likewise.
1279 (vreinterpret_u32_s16): Likewise.
1280 (vreinterpret_u32_s32): Likewise.
1281 (vreinterpret_u32_s64): Likewise.
1282 (vreinterpret_u32_f32): Likewise.
1283 (vreinterpret_u32_u8): Likewise.
1284 (vreinterpret_u32_u16): Likewise.
1285 (vreinterpret_u32_u64): Likewise.
1286 (vreinterpret_u32_p8): Likewise.
1287 (vreinterpret_u32_p16): Likewise.
1288 (vreinterpretq_u32_s8): Likewise.
1289 (vreinterpretq_u32_s16): Likewise.
1290 (vreinterpretq_u32_s32): Likewise.
1291 (vreinterpretq_u32_s64): Likewise.
1292 (vreinterpretq_u32_f32): Likewise.
1293 (vreinterpretq_u32_u8): Likewise.
1294 (vreinterpretq_u32_u16): Likewise.
1295 (vreinterpretq_u32_u64): Likewise.
1296 (vreinterpretq_u32_p8): Likewise.
1297 (vreinterpretq_u32_p16): Likewise.
1299 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1301 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1303 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1306 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1307 (vqnegd_s64): Likewise.
1308 (vqabs_s64): Likewise.
1309 (vqabsd_s64): Likewise.
1311 2014-04-22 Richard Henderson <rth@redhat.com>
1313 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1314 computation to the top of the loop.
1316 2014-04-22 Renlin <renlin.li@arm.com>
1317 Jiong Wang <jiong.wang@arm.com>
1319 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1320 * config/aarch64/aarch64.c (aarch64_layout_frame)
1321 (aarch64_initial_elimination_offset): Likewise.
1323 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1325 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1328 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1330 * machmode.h (bitwise_mode_for_mode): Declare.
1331 * stor-layout.h (bitwise_type_for_mode): Likewise.
1332 * stor-layout.c (bitwise_mode_for_mode): New function.
1333 (bitwise_type_for_mode): Likewise.
1334 * builtins.c (fold_builtin_memory_op): Use it instead of
1335 int_mode_for_mode and build_nonstandard_integer_type.
1337 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1339 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1340 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1341 (*-*-solaris2*): Simplify.
1342 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1343 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1344 *-*-solaris2.9* handling.
1346 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1348 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1349 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1351 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1352 * configure: Regenerate.
1354 * config/i386/sol2-9.h: Remove.
1356 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1357 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1358 Remove Solaris 9 references.
1360 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1362 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1363 (floatuns<GPI:mode><GPF:mode>2): Remove.
1364 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1365 and floatuns conversions.
1366 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1367 and floatuns conversions.
1368 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1369 (w1,w2): New mode attributes for inequal width conversions.
1371 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1373 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1374 the output asm format.
1376 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1378 * config/aarch64/aarch64-simd.md
1379 (aarch64_cm<optab>di): Always split.
1380 (*aarch64_cm<optab>di): New.
1381 (aarch64_cmtstdi): Always split.
1382 (*aarch64_cmtstdi): New.
1384 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1386 PR tree-optimization/60823
1387 * omp-low.c (ipa_simd_modify_function_body): Go through
1388 all SSA_NAMEs and for those refering to vector arguments
1389 which are going to be replaced adjust SSA_NAME_VAR and,
1390 if it is a default definition, change it into a non-default
1391 definition assigned at the beginning of function from new_decl.
1392 (ipa_simd_modify_stmt_ops): Rewritten.
1393 * tree-dfa.c (set_ssa_default_def): When removing default def,
1394 check for NULL loc instead of NULL *loc.
1396 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1398 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1399 restrictions on core registers for DImode values in Thumb2.
1401 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1403 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1404 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1406 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1408 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1409 (*iordi_notzesidi_di): Likewise.
1410 (*iordi_notsesidi_di): Likewise.
1412 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1414 * config/arm/arm-protos.h (tune_params): New struct members.
1415 * config/arm/arm.c: Initialise tune_params per processor.
1416 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1417 for speed, based on new tune_params.
1419 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1421 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1423 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1425 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1427 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1428 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1429 (vrnda_f64): Likewise.
1430 (vrndi_f64): Likewise.
1431 (vrndm_f64): Likewise.
1432 (vrndn_f64): Likewise.
1433 (vrndp_f64): Likewise.
1434 (vrndx_f64): Likewise.
1436 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1438 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1439 GET_MODE_SIZE argument is enum machine_mode.
1441 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1444 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1445 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1447 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1450 * asan.c (asan_emit_stack_protection): Force the base to align to
1451 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1452 appropriate bits if STRICT_ALIGNMENT.
1453 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1455 (expand_used_vars): Leave a space in the stack frame for alignment
1456 if STRICT_ALIGNMENT.
1458 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1460 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1462 (gimple_store_p): Likewise.
1463 (gimple_assign_load_p): Likewise.
1464 (gimple_assign_cast_p): Likewise.
1465 (gimple_clobber_p): Likewise.
1467 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1468 rather than a gimple.
1469 (gimple_assign_cast_p): Likewise.
1471 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1474 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1475 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1477 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1478 more debug information for E500 if -mdebug=reg.
1480 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1483 * config/i386/i386.c (ix86_expand_builtin)
1484 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1485 register for target RTX.
1486 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1488 2014-04-18 Cong Hou <congh@google.com>
1490 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1491 the widen-mult pattern by handling two operands with different sizes,
1492 and operands whose size is smaller than half of the result type.
1494 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1496 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1497 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1498 (do_estimate_edge_time): Compute it.
1499 * ipa-inline.c (want_inline_small_function_p): Bypass
1500 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1502 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1504 * ipa-inline.c (spec_rem): New static variable.
1505 (dump_overall_stats): New function.
1506 (dump_inline_stats): New function.
1508 2014-04-18 Richard Henderson <rth@redhat.com>
1510 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1511 to GET_MODE_SIZE, not a reg_class_t.
1513 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1515 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1516 (vsx_xxmrglw_<mode>): Likewise.
1518 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1521 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1522 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1523 (rs6000_init_hard_regno_mode_ok): Likewise.
1525 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1527 * ipa-inline.c (inline_small_functions): Account only non-cold
1529 * doc/invoke.texi (inline-unit-growth): Update documentation.
1531 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1533 * config/rs6000/rs6000.md (addti3, subti3): New.
1535 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1538 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1539 comment. Check optimize_insn_for_size_p instead of
1540 optimize_insn_for_speed_p.
1542 2014-04-17 Martin Jambor <mjambor@suse.cz>
1544 * gimple-iterator.c (gsi_start_edge): New function.
1545 * gimple-iterator.h (gsi_start_edge): Declare.
1546 * tree-sra.c (single_non_eh_succ): New function.
1547 (disqualify_ops_if_throwing_stmt): Renamed to
1548 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1549 having one non-EH successor BB.
1550 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1551 generate loads into replacements.
1552 (sra_modify_assign): Likewise and and also use the simple path for
1554 (sra_modify_function_body): Commit statements on edges.
1556 2014-04-17 Richard Biener <rguenther@suse.de>
1559 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1560 comparison results and add clarifying comment.
1562 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1564 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1565 (blank_mode): Initialize it.
1566 (emit_mode_size_inline, emit_mode_nunits_inline,
1567 emit_mode_inner_inline): New functions.
1568 (emit_insn_modes_h): Call them and surround their output with
1569 #if GCC_VERSION >= 4001 ... #endif.
1570 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1571 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1572 mode_* arrays if the argument is __builtin_constant_p.
1573 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1574 is enum machine_mode.
1576 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1578 * passes.c (opt_pass::execute): Adjust.
1579 (pass_manager::execute_pass_mode_switching): Likewise.
1580 (early_local_passes::execute): Likewise.
1581 (execute_one_pass): Pass cfun to the pass's execute method.
1582 * tree-pass.h (opt_pass::execute): Add function * argument.
1583 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1584 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1585 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1586 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1587 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1588 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1589 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1590 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1591 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1592 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1593 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1594 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1595 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1596 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1597 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1598 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1599 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1600 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1601 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1602 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1603 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1604 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1605 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1606 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1607 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1608 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1609 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1610 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1611 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1614 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1616 * passes.c (opt_pass::gate): Take function * argument.
1617 (gate_all_early_local_passes): Merge into
1618 (early_local_passes::gate): this.
1619 (gate_all_early_optimizations): Merge into
1620 (all_early_optimizations::gate): this.
1621 (gate_all_optimizations): Mege into
1622 (all_optimizations::gate): this.
1623 (gate_all_optimizations_g): Merge into
1624 (all_optimizations_g::gate): this.
1625 (gate_rest_of_compilation): Mege into
1626 (rest_of_compilation::gate): this.
1627 (gate_postreload): Merge into
1628 (postreload::gate): this.
1629 (dump_one_pass): Pass cfun to the pass's gate method.
1630 (execute_ipa_summary_passes): Likewise.
1631 (execute_one_pass): Likewise.
1632 (ipa_write_summaries_2): Likewise.
1633 (ipa_write_optimization_summaries_1): Likewise.
1634 (ipa_read_summaries_1): Likewise.
1635 (ipa_read_optimization_summaries_1): Likewise.
1636 (execute_ipa_stmt_fixups): Likewise.
1637 * tree-pass.h (opt_pass::gate): Add function * argument.
1638 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
1639 combine-stack-adj.c, combine.c, compare-elim.c,
1640 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1641 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
1642 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
1643 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
1644 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1645 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1646 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1647 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
1648 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
1649 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
1650 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1651 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1652 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
1653 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1654 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1655 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1656 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1657 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1658 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1659 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1660 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1661 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1662 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1663 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
1664 var-tracking.c, vtable-verify.c, web.c: Adjust.
1666 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1668 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
1669 * configure: Regenerate.
1671 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1673 * passes.c (dump_one_pass): don't check pass->has_gate.
1674 (execute_ipa_summary_passes): Likewise.
1675 (execute_one_pass): Likewise.
1676 (ipa_write_summaries_2): Likewise.
1677 (ipa_write_optimization_summaries_1): Likewise.
1678 (ipa_read_optimization_summaries_1): Likewise.
1679 (execute_ipa_stmt_fixups): Likewise.
1680 * tree-pass.h (pass_data::has_gate): Remove.
1681 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1682 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1683 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1684 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1685 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1686 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
1687 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
1688 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
1689 gimple-low.c, gimple-ssa-isolate-paths.c,
1690 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
1691 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
1692 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1693 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
1694 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
1695 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
1696 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
1697 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1698 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
1699 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1700 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1701 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1702 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1703 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1704 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1705 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1706 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1707 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1708 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1709 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1710 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1711 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1714 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1716 * pass_manager.h (pass_manager::register_dump_files_1): Remove
1718 * passes.c (pass_manager::register_dump_files_1): Merge into
1719 (pass_manager::register_dump_files): this, and remove its handling of
1720 properties since the pass always has the properties anyway.
1721 (pass_manager::pass_manager): Adjust.
1723 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1725 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
1726 * passes.c (pass_manager::register_dump_files_1): Remove dead code
1727 dealing with properties.
1728 (pass_manager::register_dump_files): Adjust.
1730 2014-03-20 Mark Wielaard <mjw@redhat.com>
1732 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
1733 then represent the bound as normal constant value.
1735 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1738 Forward port from 4.8 branch
1739 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
1741 * config/i386/bmiintrin.h (_blsi_u32): New.
1745 (_blsmsk_u32): Ditto.
1746 (_blsmsk_u64): Ditto.
1747 (_tzcnt_u32): Ditto.
1748 (_tzcnt_u64): Ditto.
1750 2014-04-17 Kito Cheng <kito@0xlab.org>
1752 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
1754 2014-04-17 Richard Biener <rguenther@suse.de>
1757 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
1758 boolean results for comparisons.
1760 2014-04-17 Richard Biener <rguenther@suse.de>
1762 PR tree-optimization/60836
1763 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
1764 initial PHI args to be gimple values.
1766 2014-04-17 Richard Biener <rguenther@suse.de>
1768 PR tree-optimization/60841
1769 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
1770 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
1771 of stmts to SLP build.
1772 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
1773 (vect_analyze_slp): Likewise.
1774 (vect_analyze_slp_instance): Likewise.
1775 (vect_build_slp_tree): Limit overall SLP tree growth.
1776 * tree-vectorizer.h (vect_analyze_data_refs,
1777 vect_analyze_slp): Adjust prototypes.
1779 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1781 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
1784 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1786 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
1787 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
1788 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
1789 for TARGET_SLOW_PSHUFB
1791 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1793 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
1794 * config/i386/i386.c (intel_cost): Ditto.
1796 2014-04-17 Joey Ye <joey.ye@arm.com>
1798 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
1800 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1802 * opts.c (common_handle_option): Disable -fipa-reference coorectly
1805 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1807 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
1808 (type_all_derivations_known_p): New predicate.
1809 (type_all_ctors_visible_p): New predicate.
1810 (type_possibly_instantiated_p): New predicate.
1811 (get_odr_type): Compute all_derivations_known.
1812 (dump_odr_type): Dump the flag.
1813 (maybe_record_type): Cleanup.
1814 (record_target_from_binfo): Add bases_to_consider array;
1815 record bases for types w/o instances and skip CXX destructor.
1816 (possible_polymorphic_call_targets_1): Add bases_to_consider
1817 and consider_construction parameters; check if type may have instance.
1818 (get_polymorphic_call_info): Set maybe_in_construction to true
1819 when we know nothing.
1820 (record_targets_from_bases): Skip CXX destructors; they are
1821 never called for types in construction.
1822 (possible_polymorphic_call_targets): Do not record target when
1823 type may not have instance.
1825 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1828 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
1829 external aliases alive, too.
1831 2014-04-16 Andrew Pinski <apinski@cavium.com>
1833 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
1836 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1838 * final.c (compute_alignments): Do not apply loop alignment to a block
1839 falling through to the exit.
1841 2014-04-16 Catherine Moore <clm@codesourcery.com>
1843 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
1844 Adjust constraints for microMIPS store patterns.
1846 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
1848 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
1850 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1852 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
1853 (append_use): Run at -O0.
1854 (append_vdef): Likewise.
1855 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
1856 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
1858 2014-04-16 Jakub Jelinek <jakub@redhat.com>
1860 PR tree-optimization/60844
1861 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
1862 (propagate_op_to_single_use, remove_visited_stmt_chain,
1863 linearize_expr, repropagate_negates, reassociate_bb): Use it
1864 instead of gsi_remove.
1866 2014-04-16 Martin Jambor <mjambor@suse.cz>
1868 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
1869 ipa_transforms_to_apply.
1870 (cgraph_function_versioning): Assert that old_node has empty
1871 ipa_transforms_to_apply.
1872 * trans-mem.c (ipa_tm_create_version): Likewise.
1873 * tree-inline.c (tree_function_versioning): Do not duplicate
1874 ipa_transforms_to_apply.
1876 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1879 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
1881 Pass necessary as flags on 64-bit Solaris/x86.
1882 Use lowercase relocs for x86_64-*-*.
1883 * configure: Regenerate.
1885 2014-04-15 Jan Hubicka <jh@suse.cz>
1887 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
1888 (maybe_record_node, likely_target_p): Use it.
1890 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1893 Revert following patch
1895 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
1898 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
1899 software floating point or no floating point registers, do not
1900 allow any type in the FPRs. Eliminate a test for SPE SIMD types
1901 in GPRs that occurs after we tested for GPRs that would never be
1904 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
1905 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
1906 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
1907 specifically allow DDmode, since that does not use the SPE SIMD
1910 2014-03-21 Mark Wielaard <mjw@redhat.com>
1912 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
1913 as unsigned or int depending on type and value used.
1915 2014-04-15 Richard Biener <rguenther@suse.de>
1917 PR rtl-optimization/56965
1918 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
1919 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
1921 * alias.c (true_dependence_1): Do not call
1922 nonoverlapping_component_refs_p.
1923 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
1924 nonoverlapping_component_refs_p.
1925 (indirect_refs_may_alias_p): Likewise.
1927 2014-04-15 Teresa Johnson <tejohnson@google.com>
1929 * cfg.c (dump_bb_info): Fix flags check.
1930 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
1932 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1934 PR rtl-optimization/60663
1935 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
1938 2014-04-15 Richard Biener <rguenther@suse.de>
1940 * lto-streamer.h (LTO_major_version): Bump to 4.
1942 2014-04-15 Richard Biener <rguenther@suse.de>
1944 * common.opt (lto_partition_model): New enum.
1945 (flto-partition=): Merge separate options with a single with argument,
1946 add -flto-partition=one support.
1947 * flag-types.h (enum lto_partition_model): Declare.
1948 * opts.c (finish_options): Remove duplicate -flto-partition=
1950 * lto-wrapper.c (run_gcc): Adjust.
1952 2014-04-15 Richard Biener <rguenther@suse.de>
1954 * alias.c (ncr_compar): New function.
1955 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
1957 2014-04-15 Richard Biener <rguenther@suse.de>
1959 * alias.c (record_component_aliases): Do not walk BINFOs.
1961 2014-04-15 Richard Biener <rguenther@suse.de>
1963 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
1964 Add struct function argument and adjust.
1965 (find_func_aliases_for_call): Likewise.
1966 (find_func_aliases): Likewise.
1967 (find_func_clobbers): Likewise.
1968 (intra_create_variable_infos): Likewise.
1969 (compute_points_to_sets): Likewise.
1970 (ipa_pta_execute): Adjust. Do not push/pop cfun.
1972 2014-04-15 Richard Biener <rguenther@suse.de>
1974 * tree.c (iterative_hash_expr): Use enum tree_code_class
1975 to store TREE_CODE_CLASS.
1976 (tree_block): Likewise.
1977 (tree_set_block): Likewise.
1978 * tree.h (fold_build_pointer_plus_loc): Use
1979 convert_to_ptrofftype_loc.
1981 2014-04-15 Jakub Jelinek <jakub@redhat.com>
1984 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
1987 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
1989 * cfgloop.h (struct loop): Move force_vectorize down.
1990 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
1991 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
1992 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
1993 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
1994 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
1995 * tree-core.h (enum annot_expr_kind): Add new kind values.
1996 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
1997 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
1999 * tree.def (ANNOTATE_EXPR): Tweak comment.
2001 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2003 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2006 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2008 * tree.h (TYPE_IDENTIFIER): Declare.
2009 * tree.c (subrange_type_for_debug_p): Use it.
2010 * godump.c (go_format_type): Likewise.
2011 * dwarf2out.c (is_cxx_auto, modified_type_die,
2012 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2013 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2015 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2018 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2020 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2022 * config/i386/i386.c (examine_argument): Return bool. Return true if
2023 parameter should be passed in memory.
2024 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2025 (construct_container): Update calls to examine_argument.
2026 (function_arg_advance_64): Ditto.
2027 (return_in_memory_32): Merge with ix86_return_in_memory.
2028 (return_in_memory_64): Ditto.
2029 (return_in_memory_ms_64): Ditto.
2031 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2033 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2034 * coverage.c (coverage_compute_profile_id): Handle externally visible
2037 2014-04-14 Martin Jambor <mjambor@suse.cz>
2039 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2040 DECL_DISREGARD_INLINE_LIMITS functions.
2042 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2045 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2047 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2050 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2051 optimize_insn_for_speed_p instead of
2052 optimize_function_for_speed_p.
2054 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2056 * doc/invoke.texi (free): Document AArch64.
2058 2014-04-14 Richard Biener <rguenther@suse.de>
2060 PR tree-optimization/60042
2061 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2062 (insert_into_preds_of_block): Do not prevent PHI insertion
2063 for REFERENCE exprs here ...
2064 (eliminate_dom_walker::before_dom_children): ... but prevent
2065 their use here under similar conditions when applied to the
2066 IL after PRE optimizations.
2068 2014-04-14 Richard Biener <rguenther@suse.de>
2070 * passes.def: Move early points-to after early SRA.
2072 2014-04-14 Richard Biener <rguenther@suse.de>
2074 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2075 check for which sign-changes we allow when forwarding
2076 a converted value into a switch.
2078 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2080 * stor-layout.c (place_field): Finalize non-constant offset for the
2083 2014-04-14 Richard Biener <rguenther@suse.de>
2085 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2087 (expand_switch_using_bit_tests_p): Likewise.
2088 (process_switch): Compute and pass on speed_p based on the
2090 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2091 optimize_bb_for_speed_p.
2093 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2095 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2096 * function.h (struct function): Rename has_force_vect_loops into
2097 has_force_vectorize_loops.
2098 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2099 (input_struct_function_base): Likewise.
2100 * lto-streamer-out.c (output_cfg): Likewise.
2101 (output_struct_function_base): Likewise.
2102 * omp-low.c (expand_omp_simd): Likewise.
2103 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2104 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2105 (version_loop_for_if_conversion): Likewise.
2106 (tree_if_conversion): Likewise.
2107 (main_tree_if_conversion): Likewise.
2108 (gate_tree_if_conversion): Likewise.
2109 * tree-inline.c (copy_loops): Likewise.
2110 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2111 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2112 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2113 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2114 * tree-vectorizer.c (vectorize_loops): Likewise.
2115 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2117 2014-04-14 Richard Biener <rguenther@suse.de>
2120 * lto-streamer-out.c (wrap_refs): New function.
2121 (lto_output): Wrap symbol references in global initializes in
2122 type-preserving MEM_REFs.
2124 2014-04-14 Christian Bruel <christian.bruel@st.com>
2126 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2128 2014-04-14 Christian Bruel <christian.bruel@st.com>
2130 * config/sh/sh.md (setmemqi): New expand pattern.
2131 * config/sh/sh.h (CLEAR_RATIO): Define.
2132 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2133 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2135 2014-04-14 Richard Biener <rguenther@suse.de>
2138 * fold-const.c (negate_expr_p): Don't negate directional rounding
2140 (fold_negate_expr): Likewise.
2142 2014-04-14 Richard Biener <rguenther@suse.de>
2144 PR tree-optimization/59817
2145 PR tree-optimization/60453
2146 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2147 recursion to catch all CHRECs in the scalar evolution and restrict
2148 the predicate for the remains appropriately.
2150 2014-04-12 Catherine Moore <clm@codesourcery.com>
2152 * config/mips/constraints.md: Add new register constraint "kb".
2153 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2154 (*movhi_internal): Likewise.
2155 (*movqi_internal): Likewise.
2156 * config/mips/mips.h (M16_STORE_REGS): New register class.
2157 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2158 (REG_CLASS_CONTENTS): Likewise.
2159 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2161 2014-04-11 Tobias Burnus <burnus@net-b.de>
2164 * doc/invoke.texi (-Wformat-signedness): Document it.
2165 (Wformat=2): Mention that this enables -Wformat-signedness.
2167 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2169 * common/config/epiphany/epiphany-common.c
2170 (epiphany_option_optimization_table): Enable section anchors by
2171 default at -O1 or higher.
2172 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2173 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2174 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2175 carries no extra cost.
2176 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2177 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2178 * config/epiphany/predicates.md (memclob_operand): New predicate.
2179 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2180 Use memclob_operand predicate and X constraint for operand 3.
2182 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2184 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2185 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2188 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2190 PR rtl-optimization/60651
2191 * mode-switching.c (optimize_mode_switching): Make sure to emit
2192 sets of a lower numbered entity before sets of a higher numbered
2193 entity to a mode of the same or lower priority.
2194 When creating a seginfo for a basic block that starts with a code
2195 label, move the insertion point past the code label.
2196 (new_seginfo): Document and enforce requirement that
2197 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2198 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2199 * doc/tm.texi: Regenerate.
2201 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2204 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2206 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2208 * BASE-VER: Set to 4.10.0.
2210 2014-04-11 Tobias Burnus <burnus@net-b.de>
2213 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2214 * doc/gcc.texi (Service): Update description in the @menu
2215 * doc/invoke.texi (Option Summary): Remove misplaced and
2218 2014-04-11 Steve Ellcey <sellcey@mips.com>
2219 Jakub Jelinek <jakub@redhat.com>
2222 * expr.c (convert_move): Use emit_store_flag_force instead of
2223 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2226 2014-04-11 Richard Biener <rguenther@suse.de>
2229 * varasm.c (assemble_alias): Avoid endless error reporting
2230 recursion by setting TREE_ASM_WRITTEN.
2232 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2234 * config/s390/s390.md: Add a splitter for NOT rtx.
2236 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2238 PR rtl-optimization/60663
2239 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2241 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2242 Jakub Jelinek <jakub@redhat.com>
2245 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2246 flag from decl_node to node.
2248 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2251 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2252 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2253 ameliorating the cases where it can be.
2255 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2258 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2260 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2261 (loadsync_<mode>): Change mode.
2262 (load_quadpti, store_quadpti): New.
2263 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2264 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2265 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2267 2014-04-09 Cong Hou <congh@google.com>
2270 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2273 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2275 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2276 instead of vnor to exploit possible fusion opportunity in the
2278 (altivec_expand_vec_perm_const_le): Likewise.
2280 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2282 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2283 (loadsync_<mode>): Change mode.
2284 (load_quadpti, store_quadpti): New.
2285 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2286 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2288 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2291 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2292 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2293 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2295 2014-04-08 Richard Biener <rguenther@suse.de>
2298 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2299 a 64bit widest int print double-int similar to on HWI64 hosts.
2301 2014-04-08 Richard Biener <rguenther@suse.de>
2303 PR tree-optimization/60785
2304 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2305 default defs properly.
2307 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2309 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2310 (Weffc++): Likewise.
2312 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2314 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2315 set completep to false rather than true.
2317 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2320 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2321 ARM_TARGET2_DWARF_FORMAT.
2323 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2326 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2327 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2330 2014-04-07 Richard Biener <rguenther@suse.de>
2332 PR tree-optimization/60766
2333 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2334 (may_eliminate_iv): Convert cand_value_at result to desired type.
2336 2014-04-07 Jason Merrill <jason@redhat.com>
2339 * common.opt (-fno-gnu-unique): Add.
2340 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2342 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2344 * haifa-sched.c: Fix outdated function reference and minor
2345 grammar errors in introductory comment.
2347 2014-04-07 Richard Biener <rguenther@suse.de>
2350 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2352 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2354 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2357 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2358 size accounting for thunks.
2359 (pa_asm_output_mi_thunk): Use final_start_function() and
2360 final_end_function() to output function start and end directives.
2362 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2364 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2365 device specific ISA/ feature information. Remove short_sp and
2366 errata_skip ds. Add avr_device_specific_features enum to have device
2368 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2369 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2370 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2371 updated device specific info.
2372 * config/avr/avr-mcus.def: Merge device specific details to
2373 dev_attribute field.
2374 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2376 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2377 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2378 assembler if RMW isa supported by current device.
2379 * config/avr/genmultilib.awk: Update as device info structure changed.
2380 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2382 2014-04-04 Cong Hou <congh@google.com>
2384 PR tree-optimization/60656
2385 * tree-vect-stmts.c (supportable_widening_operation):
2386 Fix a bug that elements in a vector with vect_used_by_reduction
2387 property are incorrectly reordered when the operation on it is not
2388 consistant with the one in reduction operation.
2390 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2392 PR rtl-optimization/60155
2393 * gcse.c (record_set_data): New function.
2394 (single_set_gcse): New function.
2395 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2396 (hoist_code): Likewise.
2397 (get_pressure_class_and_nregs): Likewise.
2399 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2401 * explow.c (probe_stack_range): Emit a final optimization blockage.
2403 2014-04-04 Anthony Green <green@moxielogic.com>
2405 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2408 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2411 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2412 flags are set only during streaming.
2413 * ipa.c (process_references, walk_polymorphic_call_targets,
2414 symtab_remove_unreachable_nodes): Drop bodies of always inline
2415 after early inlining.
2416 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2418 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2419 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2422 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2425 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2428 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2430 (cortex_a53_fdivd): Likewise.
2432 2014-04-04 Martin Jambor <mjambor@suse.cz>
2435 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2437 * cgraph.c (clone_of_p): Also return true if thunks match.
2438 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2439 cgraph_function_or_thunk_node and an obsolete comment.
2440 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2442 (build_function_decl_skip_args): Likewise.
2443 (set_new_clone_decl_and_node_flags): New function.
2444 (duplicate_thunk_for_node): Likewise.
2445 (redirect_edge_duplicating_thunks): Likewise.
2446 (cgraph_clone_node): New parameter args_to_skip, pass it to
2447 redirect_edge_duplicating_thunks which is called instead of
2448 cgraph_redirect_edge_callee.
2449 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2450 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2452 2014-04-04 Jeff Law <law@redhat.com>
2455 * config/arm/predicates.md (const_int_I_operand): New predicate.
2456 (const_int_M_operand): Similarly.
2457 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2459 (insv_t2, extv_reg, extzv_t2): Likewise.
2460 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2461 (pop_multiple_with_writeback_and_return): Likewise.
2462 (vfp_pop_multiple_with_writeback): Likewise
2464 2014-04-04 Richard Biener <rguenther@suse.de>
2467 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2468 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2470 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2471 * gimplify.c (gimple_add_tmp_var_fn): New function.
2472 * gimple-expr.h (create_tmp_reg_fn): Declare.
2473 * gimple-expr.c (create_tmp_reg_fn): New function.
2474 * gimple-low.c (record_vars_into): Don't change cfun.
2475 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2476 code generation without cfun.
2478 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2481 * Makefile.in (install-driver): Fix shell scripting.
2483 2014-04-03 Cong Hou <congh@google.com>
2485 PR tree-optimization/60505
2486 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2487 threshold of number of iterations below which no vectorization
2489 * tree-vect-loop.c (new_loop_vec_info):
2490 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2491 * tree-vect-loop.c (vect_analyze_loop_operations):
2492 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2493 * tree-vect-loop.c (vect_transform_loop):
2494 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2495 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2496 of iterations of the loop and see if we should build the epilogue.
2498 2014-04-03 Richard Biener <rguenther@suse.de>
2500 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2501 (streamer_tree_cache_create): Adjust.
2502 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2503 to allow optional nodes array.
2504 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2505 (streamer_tree_cache_append): Likewise.
2506 (streamer_tree_cache_create): Create nodes array optionally
2507 as specified by parameter.
2508 * lto-streamer-out.c (create_output_block): Avoid maintaining
2509 the node array in the writer cache.
2510 (DFS_write_tree): Remove assertion.
2511 (produce_asm_for_decls): Free the out decl state hash table early.
2512 * lto-streamer-in.c (lto_data_in_create): Adjust for
2513 streamer_tree_cache_create prototype change.
2515 2014-04-03 Richard Biener <rguenther@suse.de>
2517 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2518 set TREE_CHAIN to NULL_TREE.
2520 2014-04-03 Richard Biener <rguenther@suse.de>
2522 PR tree-optimization/60740
2523 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2524 over all GIMPLE_COND operands.
2526 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2528 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2529 (Weffc++): Remove Scott's numbering, merge lists and reference
2532 2014-04-03 Nick Clifton <nickc@redhat.com>
2534 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2537 2014-04-03 Martin Jambor <mjambor@suse.cz>
2539 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2540 mention gcc_unreachable before failing.
2541 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2544 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
2547 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2548 inconsistent code and instead mark the context inconsistent.
2549 (possible_polymorphic_call_targets): For inconsistent contexts
2550 return empty complete list.
2552 2014-04-02 Anthony Green <green@moxielogic.com>
2554 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2555 (extendqisi2, extendhisi2): Define.
2556 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2557 (WCHAR_TYPE): Change to unsigned int.
2559 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2561 PR tree-optimization/60733
2562 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2563 insertion point for PHI candidates to be the end of the feeding
2564 block for the PHI argument.
2566 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2568 PR rtl-optimization/60650
2569 * lra-constraints.c (process_alt_operands): Decrease reject for
2570 earlyclobber matching.
2572 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2574 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2576 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2578 * config/spu/spu.c (pad_bb): Do not crash when the last
2579 insn is CODE_FOR_blockage.
2581 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2583 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2584 lies outside the target mode.
2586 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2589 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2590 software floating point or no floating point registers, do not
2591 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2592 in GPRs that occurs after we tested for GPRs that would never be
2595 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2596 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2597 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2598 specifically allow DDmode, since that does not use the SPE SIMD
2601 2014-04-02 Richard Biener <rguenther@suse.de>
2604 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2605 MODE_INTs. Properly use negv_optab.
2606 (expand_abs): Likewise.
2608 2014-04-02 Richard Biener <rguenther@suse.de>
2611 * Makefile.in (install-driver): Guard extra installs with special
2614 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
2616 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2619 2014-04-01 Richard Henderson <rth@redhat.com>
2622 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
2623 alternative enabled before register allocation.
2625 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
2627 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
2628 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
2630 (nios2_large_got_address): Remove unneeded 'sym' parameter.
2631 (nios2_got_address): Update nios2_large_got_address call site.
2632 (nios2_delegitimize_address): New function.
2633 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
2634 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
2635 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
2637 2014-04-01 Martin Husemann <martin@duskware.de>
2639 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
2642 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
2644 PR rtl-optimization/60604
2645 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
2646 check from register_operand.
2647 (register_operand): Redefine in terms of general_operand.
2648 (nonmemory_operand): Use register_operand for the non-constant cases.
2650 2014-04-01 Richard Biener <rguenther@suse.de>
2652 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
2654 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
2656 * doc/invoke.texi (mapp-regs): Clarify.
2658 2014-03-31 Ulrich Drepper <drepper@gmail.com>
2660 * config/i386/avx512fintrin.h (__v32hi): Define type.
2661 (__v64qi): Likewise.
2662 (_mm512_set1_epi8): Define.
2663 (_mm512_set1_epi16): Define.
2664 (_mm512_set4_epi32): Define.
2665 (_mm512_set4_epi64): Define.
2666 (_mm512_set4_pd): Define.
2667 (_mm512_set4_ps): Define.
2668 (_mm512_setr4_epi64): Define.
2669 (_mm512_setr4_epi32): Define.
2670 (_mm512_setr4_pd): Define.
2671 (_mm512_setr4_ps): Define.
2672 (_mm512_setzero_epi32): Define.
2674 2014-03-31 Martin Jambor <mjambor@suse.cz>
2677 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
2678 callsite_arguments_match_p. Updated all callers. Also check types of
2679 corresponding formal parameters and actual arguments.
2680 (not_all_callers_have_enough_arguments_p) Renamed to
2681 some_callers_have_mismatched_arguments_p.
2683 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
2685 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
2687 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
2690 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
2693 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
2695 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
2696 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
2698 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
2699 Use FMAMODE_NOVF512 mode iterator.
2700 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
2701 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
2702 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
2703 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
2705 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
2706 Use VF_128_256 mode iterator.
2707 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
2710 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2712 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
2713 static chain if needed.
2715 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2718 * lra-constraints.c (index_part_to_reg): New.
2719 (process_address): Use it.
2721 2014-03-27 Jeff Law <law@redhat.com>
2722 Jakub Jelinek <jakub@redhat.com>
2725 * expr.c (do_tablejump): Use simplify_gen_binary rather than
2726 gen_rtx_{PLUS,MULT} to build up the address expression.
2728 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
2729 creating non-canonical RTL.
2731 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2734 * ipa-inline.c (want_inline_small_function_p): Short circuit large
2735 functions; reorganize to make cheap checks first.
2736 (inline_small_functions): Do not estimate growth when dumping;
2738 * ipa-inline.h (inline_summary): Add min_size.
2739 (growth_likely_positive): New function.
2740 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
2741 (set_cond_stmt_execution_predicate): Cleanup.
2742 (estimate_edge_size_and_time): Compute min_size.
2743 (estimate_calls_size_and_time): Likewise.
2744 (estimate_node_size_and_time): Likewise.
2745 (inline_update_overall_summary): Update min_size.
2746 (do_estimate_edge_time): Likewise.
2747 (do_estimate_edge_size): Update.
2748 (do_estimate_edge_hints): Update.
2749 (growth_likely_positive): New function.
2751 2014-03-28 Jakub Jelinek <jakub@redhat.com>
2754 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
2755 also if addr has VOIDmode.
2757 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2759 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
2760 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
2762 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
2763 instructions as well as AdvancedSIMD loads.
2765 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2767 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
2768 Use crypto_aese type.
2769 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
2770 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
2771 crypto_aese, crypto_aesmc. Move to types.md.
2772 * config/arm/types.md (crypto_aes): Split into crypto_aese,
2774 * config/arm/iterators.md (crypto_type): Likewise.
2776 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2778 * cgraph.c: Include expr.h and tree-dfa.h.
2779 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
2782 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2785 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
2786 regs from checking multi-reg pseudos.
2788 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2790 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
2792 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2794 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
2795 if it would clobber the stack pointer, even temporarily.
2797 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
2799 * mode-switching.c: Make small adjustments to the top comment.
2801 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
2803 * config/rs6000/constraints.md (wD constraint): New constraint to
2804 match the constant integer to get the top DImode/DFmode out of a
2805 vector in a VSX register.
2807 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
2808 match the constant integer to get the top DImode/DFmode out of a
2809 vector in a VSX register.
2811 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
2814 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
2817 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
2818 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
2820 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
2821 Optimize vec_extract of 64-bit values, where the value being
2822 extracted is in the top word, where we can use scalar
2823 instructions. Add direct move and store support. Combine the big
2824 endian/little endian vector select load support into a single insn.
2825 (vsx_extract_<mode>_internal1): Likewise.
2826 (vsx_extract_<mode>_internal2): Likewise.
2827 (vsx_extract_<mode>_load): Likewise.
2828 (vsx_extract_<mode>_store): Likewise.
2829 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
2830 combined into vsx_extract_<mode>_load.
2831 (vsx_extract_<mode>_one_le): Likewise.
2833 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
2834 define the top 64-bit vector element.
2836 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
2839 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2840 Document vec_vbpermq builtin.
2843 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
2844 enable use of xxsldwi and xxpermdi builtin functions.
2845 (vec_xxpermdi): Likewise.
2847 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2848 Document use of vec_xxsldwi and vec_xxpermdi builtins.
2850 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
2852 PR rtl-optimization/60650
2853 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
2855 (find_spills_for): New.
2856 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
2857 Spill all pseudos on the second iteration.
2859 2014-03-27 Marek Polacek <polacek@redhat.com>
2862 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
2865 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2867 * config/s390/s390.c (s390_can_use_return_insn): Check for
2868 call-saved FPRs on 31 bit.
2870 2014-03-27 Jakub Jelinek <jakub@redhat.com>
2873 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
2874 if they need regimplification, just drop them instead of
2875 calling gimple_regimplify_operands on them.
2877 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
2880 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
2881 (aarch64_frame_pointer_required): Adjust logic.
2882 (aarch64_can_eliminate): Adjust logic.
2883 (aarch64_override_options_after_change): Adjust logic.
2885 2014-03-27 Dehao Chen <dehao@google.com>
2887 * ipa-inline.c (early_inliner): Update node's inline info.
2889 2014-03-26 Dehao Chen <dehao@google.com>
2891 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
2892 compiler inserted conditional jumps for NAN float check.
2894 2014-03-26 Jakub Jelinek <jakub@redhat.com>
2896 * ubsan.h (ubsan_create_data): Change second argument's type
2897 to const location_t *.
2898 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
2900 (ubsan_create_data): Change second argument to const location_t *PLOC.
2901 Create Loc field whenever PLOC is non-NULL.
2902 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
2903 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
2907 * real.c (real_to_integer2): Change type of low to UHWI.
2909 2014-03-26 Tobias Burnus <burnus@net-b.de>
2911 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
2912 (CILK_SELF_SPECS): New define.
2913 (driver_self_specs): Use it.
2915 2014-03-26 Richard Biener <rguenther@suse.de>
2917 * tree-pretty-print.c (percent_K_format): Implement special
2918 case for LTO and its stripped down BLOCK tree.
2920 2014-03-26 Jakub Jelinek <jakub@redhat.com>
2923 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
2925 * tree-vrp.c (simplify_internal_call_using_ranges): If only
2926 one range is range_int_cst_p, but not both, at least optimize
2927 addition/subtraction of 0 and multiplication by 0 or 1.
2928 * gimple-fold.c (gimple_fold_call): Fold
2929 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
2930 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
2931 INTEGER_CSTs, try to fold at least x * 0 and y - y.
2933 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
2935 PR rtl-optimization/60452
2936 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
2937 <case REG>: Return 1 for invalid offsets from the frame pointer.
2939 2014-03-26 Marek Polacek <polacek@redhat.com>
2942 * doc/extend.texi (C Extensions): Mention variable-length arrays in
2945 2014-03-26 Marek Polacek <polacek@redhat.com>
2948 * doc/extend.texi (Designated Inits): Describe what happens to omitted
2951 2014-03-26 Marek Polacek <polacek@redhat.com>
2954 * ira-color.c (update_conflict_hard_regno_costs): Perform the
2955 multiplication in unsigned type.
2957 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
2959 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
2961 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
2963 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
2965 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
2968 * cif-code.def (UNREACHABLE) New code.
2969 * ipa-inline.c (inline_small_functions): Skip edges to
2970 __builtlin_unreachable.
2971 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
2972 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
2973 predicate to __bulitin_unreachable.
2974 (set_cond_stmt_execution_predicate): Fix issue when
2975 invert_tree_comparison returns ERROR_MARK.
2976 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
2977 propagate to inline clones.
2978 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
2980 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
2981 * cgraphclones.c (cgraph_clone_node): If call destination is already
2982 ureachable, do not redirect it back.
2983 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
2986 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
2988 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
2989 Do not modify inline clones.
2991 2014-03-25 Jakub Jelinek <jakub@redhat.com>
2993 * config/i386/i386.md (general_sext_operand): New mode attr.
2994 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
2995 don't generate (sign_extend (const_int)).
2996 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
2997 operands[2]. Use We constraint instead of <i> and
2998 <general_sext_operand> predicate instead of <general_operand>.
2999 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3000 * config/i386/constraints.md (We): New constraint.
3001 * config/i386/predicates.md (x86_64_sext_operand,
3002 sext_operand): New predicates.
3004 2014-03-25 Martin Jambor <mjambor@suse.cz>
3007 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3008 inconsistent devirtualizations to __builtin_unreachable.
3010 2014-03-25 Marek Polacek <polacek@redhat.com>
3013 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3015 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3017 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3018 order of elements for big-endian.
3020 2014-03-25 Richard Biener <rguenther@suse.de>
3023 * gimplify-me.c (gimple_regimplify_operands): Update the
3026 2014-03-25 Martin Jambor <mjambor@suse.cz>
3029 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3030 (lto_output_varpool_node): Likewise.
3031 (input_overwrite_node): Likewise.
3032 (input_varpool_node): Likewise.
3034 2014-03-25 Richard Biener <rguenther@suse.de>
3036 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3037 (run_gcc): Likewise.
3039 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3041 * combine.c (simplify_compare_const): Add MODE argument.
3042 Handle mode_width 0 as very large mode_width.
3043 (try_combine, simplify_comparison): Adjust callers.
3045 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3046 type to avoid signed integer overflow.
3047 * explow.c (plus_constant): Likewise.
3049 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3051 * doc/generic.texi: Correct typos.
3053 2014-03-24 Tobias Burnus <burnus@net-b.de>
3055 * doc/invoke.texi (-flto): Expand section about
3056 using static libraries with LTO.
3058 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3060 PR rtl-optimization/60501
3061 * optabs.def (addptr3_optab): New optab.
3062 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3063 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3064 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3066 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3068 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3070 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3072 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3075 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3076 (_mm256_undefined_ps): Define.
3077 (_mm256_undefined_pd): Define.
3078 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3079 (_mm_undefined_pd): Define.
3080 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3081 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3082 (_mm512_undefined_ps): Define.
3083 (_mm512_undefined_pd): Define.
3084 Use _mm*_undefined_*.
3085 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3087 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3089 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3090 (lshr_simd): DI mode added.
3091 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3092 (aarch64_ushr_simddi): Likewise.
3093 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3094 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3095 (vshrd_n_u64): Likewise.
3097 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3099 * Makefile.in (s-macro_list): Depend on cc1.
3101 2014-03-23 Teresa Johnson <tejohnson@google.com>
3103 * ipa-utils.c (ipa_print_order): Use specified dump file.
3105 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3107 PR rtl-optimization/60601
3108 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3110 * gcc.c (eval_spec_function): Initialize save_growing_value.
3112 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3115 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3116 code == MINUS_EXPR, never swap op0 with op1.
3118 * toplev.c (init_local_tick): Avoid signed integer multiplication
3120 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3121 shift by first operand's bitsize.
3123 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3126 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3128 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3129 TARGET_ISA_64BIT_P(x).
3131 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3133 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3134 pattern for vector nor instead of subtract from splat(-1).
3135 (altivec_expand_vec_perm_const_le): Likewise.
3137 2014-03-21 Richard Henderson <rth@twiddle.net>
3140 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3141 related insns after epilogue_completed.
3143 2014-03-21 Martin Jambor <mjambor@suse.cz>
3146 * cgraph.h (symtab_node): New flag body_removed.
3147 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3148 when removing bodies.
3149 * symtab.c (dump_symtab_base): Dump body_removed flag.
3150 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3151 had their bodies removed.
3153 2014-03-21 Martin Jambor <mjambor@suse.cz>
3156 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3159 2014-03-21 Richard Biener <rguenther@suse.de>
3161 PR tree-optimization/60577
3162 * tree-core.h (struct tree_base): Document nothrow_flag use
3163 in VAR_DECL_NONALIASED.
3164 * tree.h (VAR_DECL_NONALIASED): New.
3165 (may_be_aliased): Adjust.
3166 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3168 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3170 * expr.c (expand_expr_real_1): Remove outdated comment.
3172 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3175 * ira.c (adjust_cleared_regs): Call copy_rtx on
3176 *reg_equiv[REGNO (loc)].src_p before passing it to
3177 simplify_replace_fn_rtx.
3180 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3181 into CONST, put pic register as first operand of PLUS. Use
3182 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3184 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3186 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3188 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3190 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3191 around for store forwarding issue in the FPU on the UT699.
3192 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3193 loads and operations if -mfix-ut699 is specified.
3194 (divtf3_hq): Tweak attribute.
3195 (sqrttf2_hq): Likewise.
3197 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3199 * calls.c (store_one_arg): Remove incorrect const qualification on the
3200 type of the temporary.
3201 * cfgexpand.c (expand_return): Likewise.
3202 * expr.c (expand_constructor): Likewise.
3203 (expand_expr_real_1): Likewise.
3205 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3207 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3210 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3213 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3215 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3217 * config/arm/aarch-common-protos.h
3218 (alu_cost_table): Fix spelling of "extend".
3219 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3221 2014-03-19 Richard Biener <rguenther@suse.de>
3224 * tree-core.h (tree_type_common): Re-order pointer members
3225 to reduce recursion depth during GC walks.
3227 2014-03-19 Marek Polacek <polacek@redhat.com>
3230 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3231 before accessing it.
3233 2014-03-19 Richard Biener <rguenther@suse.de>
3236 * lto-streamer-in.c (input_function): In WPA stage do not drop
3239 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3241 PR tree-optimization/60559
3242 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3243 with build_zero_cst assignment.
3245 2014-03-18 Kai Tietz <ktietz@redhat.com>
3247 PR rtl-optimization/56356
3248 * sdbout.c (sdbout_parms): Verify that parms'
3249 incoming argument is valid.
3250 (sdbout_reg_parms): Likewise.
3252 2014-03-18 Richard Henderson <rth@redhat.com>
3255 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3256 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3257 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3259 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3261 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3262 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3263 Italicize plugin event names in description. Explain that
3264 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3265 Remind that no GCC functions should be called after PLUGIN_FINISH.
3266 Explain what pragmas with expansion are.
3268 2014-03-18 Martin Liska <mliska@suse.cz>
3270 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3271 gimple call statement is update.
3272 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3273 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3275 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3278 * ubsan.c (ubsan_instrument_unreachable): Call
3279 initialize_sanitizer_builtins.
3280 (ubsan_pass): Likewise.
3283 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3284 varpool_finalize_decl instead of rest_of_decl_compilation.
3286 2014-03-18 Richard Biener <rguenther@suse.de>
3288 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3289 by using bitmap_and_compl instead of bitmap_and_compl_into.
3290 (df_rd_transfer_function): Likewise.
3292 2014-03-18 Richard Biener <rguenther@suse.de>
3294 * doc/lto.texi (fresolution): Fix typo.
3296 2014-03-18 Richard Biener <rguenther@suse.de>
3298 * doc/invoke.texi (flto): Update for changes in 4.9.
3300 2014-03-18 Richard Biener <rguenther@suse.de>
3302 * doc/loop.texi: Remove section on the removed lambda framework.
3303 Update loop docs with recent changes in preserving loop structure.
3305 2014-03-18 Richard Biener <rguenther@suse.de>
3307 * doc/lto.texi (-fresolution): Document.
3309 2014-03-18 Richard Biener <rguenther@suse.de>
3311 * doc/contrib.texi: Adjust my name.
3313 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3316 * internal-fn.c: Include diagnostic-core.h.
3317 (expand_BUILTIN_EXPECT): New function.
3318 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3319 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3320 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3321 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3323 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3324 Revert 3 argument __builtin_expect code.
3325 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3326 * gimple-fold.c (gimple_fold_call): Likewise.
3327 * tree.h (fold_builtin_expect): New prototype.
3328 * builtins.c (build_builtin_expect_predicate): Add predictor
3329 argument, if non-NULL, create 3 argument __builtin_expect.
3330 (fold_builtin_expect): No longer static. Add ARG2 argument,
3331 pass it through to build_builtin_expect_predicate.
3332 (fold_builtin_2): Adjust caller.
3333 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3334 * internal-fn.def (BUILTIN_EXPECT): New.
3336 2014-03-18 Tobias Burnus <burnus@net-b.de>
3339 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3340 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3341 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3343 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3346 * predict.c (combine_predictions_for_bb): Fix up formatting.
3347 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3348 fill what it points to if non-NULL.
3349 (tree_predict_by_opcode): Adjust caller, use the predictor.
3350 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3352 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3354 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3355 proper constant for the store mode.
3357 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3359 * symtab.c (change_decl_assembler_name): Fix transparent alias
3362 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3364 * config/aarch64/aarch64.c: Correct the comments about the
3365 aarch64 stack layout.
3367 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3369 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3370 check for GF_OMP_FOR_KIND_FOR.
3372 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3374 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3375 ymm and zmm register names.
3377 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3380 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3381 note creation for the 2010-08-31 changes.
3383 2014-03-17 Marek Polacek <polacek@redhat.com>
3386 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3387 as -fno-tree-loop-vectorize.
3388 (expand_omp_simd): Likewise.
3390 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3392 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3393 (eligible_for_call_delay): New prototype.
3394 * config/sparc/sparc.c (tls_call_delay): Rename into...
3395 (eligible_for_call_delay): ...this. Return false if the instruction
3396 cannot be put in the delay slot of a branch.
3397 (eligible_for_restore_insn): Simplify.
3398 (eligible_for_return_delay): Return false if the instruction cannot be
3399 put in the delay slot of a branch and simplify.
3400 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3401 put in the delay slot of a branch.
3402 * config/sparc/sparc.md (fix_ut699): New attribute.
3403 (tls_call_delay): Delete.
3404 (in_call_delay): Reimplement.
3405 (eligible_for_sibcall_delay): Rename into...
3406 (in_sibcall_delay): ...this.
3407 (eligible_for_return_delay): Rename into...
3408 (in_return_delay): ...this.
3409 (in_branch_delay): Reimplement.
3410 (in_uncond_branch_delay): Delete.
3411 (in_annul_branch_delay): Delete.
3413 2014-03-14 Richard Henderson <rth@redhat.com>
3416 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3417 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3418 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3419 (floathi splitters): Remove.
3420 (float<SWI48x>xf2): New pattern.
3421 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3422 code that tried to handle DImode for 32-bit, but which was excluded
3423 by the pattern's condition. Drop allocation of stack temporary.
3424 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3425 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3426 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3427 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3428 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3429 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3430 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3431 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3432 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3433 (*float<SWI48x><X87MODEF>2_i387): Remove.
3434 (all float _with_temp splitters): Remove.
3435 (*float<SWI48x><MODEF>2_i387): New pattern.
3436 (*float<SWI48><MODEF>2_sse): New pattern.
3437 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3438 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3440 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3441 Marek Polacek <polacek@redhat.com>
3444 * common.opt (dump_base_name_prefixed): New Variable.
3445 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3446 if x_dump_base_name_prefixed is already set, set it at the end.
3448 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3450 PR rtl-optimization/60508
3451 * lra-constraints.c (get_reload_reg): Add new parameter
3453 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3454 Pass the new parameter values.
3456 2014-03-14 Richard Biener <rguenther@suse.de>
3458 * common.opt: Revert unintented changes from r205065.
3461 2014-03-14 Richard Biener <rguenther@suse.de>
3464 * cfghooks.c (split_block): Properly adjust all loops the
3465 block was a latch of.
3467 2014-03-14 Martin Jambor <mjambor@suse.cz>
3470 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3473 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3476 * config/avr/avr.c (avr_set_current_function): Pass function name
3477 through default_strip_name_encoding before sanity checking instead
3478 of skipping the first char of the assembler name.
3480 2014-03-13 Richard Henderson <rth@redhat.com>
3483 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3484 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3485 * config/i386/i386-protos.h: Likewise.
3486 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3487 in the expander instead of a splitter.
3488 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3489 any possibility of requiring a memory.
3490 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3491 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3492 (fp branch splitters): Update for ix86_split_fp_branch.
3493 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3494 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3495 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3496 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3497 (*fop_<MODEF>_3_i387): Likewise.
3498 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3499 (splitters for the fop_* register patterns): Remove.
3500 (fscalexf4_i387): Rename from *fscalexf4_i387.
3501 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3503 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3505 PR tree-optimization/59779
3506 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3507 type for bitsize and maxsize instead of HOST_WIDE_INT.
3509 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3511 PR rtl-optimization/57320
3512 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3513 the CFG after thread_prologue_and_epilogue_insns.
3515 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3517 PR rtl-optimization/57189
3518 * lra-constraints.c (process_alt_operands): Disfavor spilling
3521 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3523 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3525 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3527 PR tree-optimization/59025
3529 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3530 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3532 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3535 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3536 calls of avr_out_plus_1.
3538 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3540 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3541 BB's single pred and update the father loop's latch info later.
3543 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3545 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3549 (VEC_base): Likewise.
3550 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3551 registers, we need to swap double words in little endian mode.
3553 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3554 to be a container mode for 128-bit integer operations added in ISA
3555 2.07. Unlike TImode and PTImode, the preferred register set is
3556 the Altivec/VMX registers for the 128-bit operations.
3558 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3560 (rs6000_split_128bit_ok_p): Likewise.
3562 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3563 macros for creating ISA 2.07 normal and overloaded builtin
3564 functions with 3 arguments.
3565 (BU_P8V_OVERLOAD_3): Likewise.
3566 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3567 for use as overloaded functions.
3568 (VPERM_1TI_UNS): Likewise.
3569 (VSEL_1TI): Likewise.
3570 (VSEL_1TI_UNS): Likewise.
3571 (ST_INTERNAL_1ti): Likewise.
3572 (LD_INTERNAL_1ti): Likewise.
3573 (XXSEL_1TI): Likewise.
3574 (XXSEL_1TI_UNS): Likewise.
3575 (VPERM_1TI): Likewise.
3576 (VPERM_1TI_UNS): Likewise.
3577 (XXPERMDI_1TI): Likewise.
3578 (SET_1TI): Likewise.
3579 (LXVD2X_V1TI): Likewise.
3580 (STXVD2X_V1TI): Likewise.
3581 (VEC_INIT_V1TI): Likewise.
3582 (VEC_SET_V1TI): Likewise.
3583 (VEC_EXT_V1TI): Likewise.
3584 (EQV_V1TI): Likewise.
3585 (NAND_V1TI): Likewise.
3586 (ORC_V1TI): Likewise.
3587 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3588 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3590 (VADDUQM): Likewise.
3591 (VSUBCUQ): Likewise.
3592 (VADDEUQM): Likewise.
3593 (VADDECUQ): Likewise.
3594 (VSUBEUQM): Likewise.
3595 (VSUBECUQ): Likewise.
3597 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3598 __int128_t and __uint128_t types.
3599 (__uint128_type): Likewise.
3600 (altivec_categorize_keyword): Add support for vector __int128_t,
3601 vector __uint128_t, vector __int128, and vector unsigned __int128
3602 as a container type for TImode operations that need to be done in
3603 VSX/Altivec registers.
3604 (rs6000_macro_to_expand): Likewise.
3605 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3606 to support 128-bit integer instructions vaddcuq, vadduqm,
3607 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3608 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
3610 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
3611 for V1TImode, and set up preferences to use VSX/Altivec registers.
3612 Setup VSX reload handlers.
3613 (rs6000_debug_reg_global): Likewise.
3614 (rs6000_init_hard_regno_mode_ok): Likewise.
3615 (rs6000_preferred_simd_mode): Likewise.
3616 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
3617 (easy_altivec_constant): Likewise.
3618 (output_vec_const_move): Likewise.
3619 (rs6000_expand_vector_set): Convert V1TImode set and extract to
3621 (rs6000_expand_vector_extract): Likewise.
3622 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
3624 (rs6000_const_vec): Add support for V1TImode.
3625 (rs6000_emit_le_vsx_load): Swap double words when loading or
3626 storing TImode/V1TImode.
3627 (rs6000_emit_le_vsx_store): Likewise.
3628 (rs6000_emit_le_vsx_move): Likewise.
3629 (rs6000_emit_move): Add support for V1TImode.
3630 (altivec_expand_ld_builtin): Likewise.
3631 (altivec_expand_st_builtin): Likewise.
3632 (altivec_expand_vec_init_builtin): Likewise.
3633 (altivec_expand_builtin): Likewise.
3634 (rs6000_init_builtins): Add support for V1TImode type. Add
3635 support for ISA 2.07 128-bit integer builtins. Define type names
3636 for the VSX/Altivec vector types.
3637 (altivec_init_builtins): Add support for overloaded vector
3638 functions with V1TImode type.
3639 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
3640 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
3642 (rs6000_split_128bit_ok_p): Likewise.
3643 (rs6000_handle_altivec_attribute): Create V1TImode from vector
3644 __int128_t and vector __uint128_t.
3646 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
3647 and mode attributes.
3654 (VS_scalar): Likewise.
3655 (VS_double): Likewise.
3656 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
3658 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
3659 we support the ISA 2.07 128-bit integer arithmetic instructions.
3660 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
3661 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
3662 and TImode types for use with the builtin functions.
3663 (V1TI_type_node): Likewise.
3664 (unsigned_V1TI_type_node): Likewise.
3665 (intTI_type_internal_node): Likewise.
3666 (uintTI_type_internal_node): Likewise.
3668 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
3669 128-bit builtin functions.
3670 (UNSPEC_VADDEUQM): Likewise.
3671 (UNSPEC_VADDECUQ): Likewise.
3672 (UNSPEC_VSUBCUQ): Likewise.
3673 (UNSPEC_VSUBEUQM): Likewise.
3674 (UNSPEC_VSUBECUQ): Likewise.
3675 (VM): Add V1TImode to vector mode iterators.
3677 (VI_unit): Likewise.
3678 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
3679 (altivec_vaddcuq): Likewise.
3680 (altivec_vsubuqm): Likewise.
3681 (altivec_vsubcuq): Likewise.
3682 (altivec_vaddeuqm): Likewise.
3683 (altivec_vaddecuq): Likewise.
3684 (altivec_vsubeuqm): Likewise.
3685 (altivec_vsubecuq): Likewise.
3687 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
3689 (BOOL_128): Likewise.
3690 (BOOL_REGS_OUTPUT): Likewise.
3691 (BOOL_REGS_OP1): Likewise.
3692 (BOOL_REGS_OP2): Likewise.
3693 (BOOL_REGS_UNARY): Likewise.
3694 (BOOL_REGS_AND_CR0): Likewise.
3696 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
3697 128-bit integer builtin support.
3698 (vec_vadduqm): Likewise.
3699 (vec_vaddecuq): Likewise.
3700 (vec_vaddeuqm): Likewise.
3701 (vec_vsubecuq): Likewise.
3702 (vec_vsubeuqm): Likewise.
3703 (vec_vsubcuq): Likewise.
3704 (vec_vsubuqm): Likewise.
3706 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3707 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
3708 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
3709 128-bit integer add/subtract to ISA 2.07.
3711 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
3713 * config/arc/arc.c (arc_predicate_delay_insns):
3714 Fix third argument passed to conditionalize_nonjump.
3716 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
3718 * config/aarch64/aarch64-builtins.c
3719 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
3720 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
3721 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
3722 instead of __builtin_lfloor.
3723 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
3725 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3727 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
3728 (tree_ssa_ifcombine_bb_1): New function.
3729 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
3730 is an empty forwarder block to then_bb or vice versa and then_bb
3731 and else_bb are effectively swapped.
3733 2014-03-12 Christian Bruel <christian.bruel@st.com>
3736 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
3737 REG_CFA_DEF_CFA note.
3738 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
3739 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
3741 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
3743 PR tree-optimization/60454
3744 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
3746 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3748 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
3749 Do not define target_cpu_default2 to generic.
3750 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
3751 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
3752 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
3754 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3755 Marc Glisse <marc.glisse@inria.fr>
3757 PR tree-optimization/60502
3758 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
3759 instead of build_low_bits_mask.
3761 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3764 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
3765 if there are multiple uses, but op doesn't live on E edge.
3766 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
3767 clobber stmts before __builtin_unreachable.
3769 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
3771 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
3772 hard_frame_pointer_rtx.
3773 * cse.c (cse_insn): Remove volatile check.
3774 * cselib.c (cselib_process_insn): Likewise.
3775 * dse.c (scan_insn): Likewise.
3777 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3779 * config/arc/arc.c (conditionalize_nonjump): New function,
3781 (arc_ifcvt): ... this.
3782 (arc_predicate_delay_insns): Use it.
3784 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3786 * config/arc/predicates.md (extend_operand): During/after reload,
3787 allow const_int_operand.
3788 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
3789 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
3790 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
3792 (umulsi3_highpart_i): Likewise.
3794 2014-03-11 Richard Biener <rguenther@suse.de>
3796 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
3797 Add asserts to guard possible wrong-code bugs.
3799 2014-03-11 Richard Biener <rguenther@suse.de>
3801 PR tree-optimization/60429
3802 PR tree-optimization/60485
3803 * tree-ssa-structalias.c (set_union_with_increment): Properly
3804 take into account all fields that overlap the shifted vars.
3805 (do_sd_constraint): Likewise.
3806 (do_ds_constraint): Likewise.
3807 (get_constraint_for_ptr_offset): Likewise.
3809 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
3811 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
3812 (nios2_compute_frame_layout):
3813 Add calculation of cfun->machine->fp_save_offset.
3814 (nios2_expand_prologue): Correct setting of frame pointer register
3816 (nios2_expand_epilogue): Update recovery of stack pointer from
3817 frame pointer accordingly.
3818 (nios2_initial_elimination_offset): Update calculation of offset
3819 for eliminating to HARD_FRAME_POINTER_REGNUM.
3821 2014-03-10 Jakub Jelinek <jakub@redhat.com>
3824 * ipa.c (symtab_remove_unreachable_nodes): Don't call
3825 cgraph_get_create_node on VAR_DECLs.
3827 2014-03-10 Richard Biener <rguenther@suse.de>
3830 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
3832 2014-03-08 Douglas B Rupp <rupp@gnat.com>
3834 * config/vms/vms.opt (vms_float_format): New variable.
3836 2014-03-08 Tobias Burnus <burnus@net-b.de>
3838 * doc/invoke.texi (-fcilkplus): Update implementation status.
3840 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
3841 Richard Biener <rguenther@suse.de>
3843 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
3844 consistently accross all TUs.
3845 (run_gcc): Enable -fshort-double automatically at link at link-time
3846 and disallow override.
3848 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
3851 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
3852 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
3853 if they can't be used.
3855 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3857 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
3858 for Solaris 11/x86 ld.
3859 * configure: Regenerate.
3861 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3863 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
3864 (LIB_TLS_SPEC): Save as ld_tls_libs.
3865 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
3866 (HAVE_AS_IX86_TLSLDM): New test.
3867 * configure, config.in: Regenerate.
3868 * config/i386/i386.c (legitimize_tls_address): Fall back to
3869 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
3870 cannot support TLS_MODEL_LOCAL_DYNAMIC.
3871 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
3872 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
3874 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
3876 * common.opt (fira-loop-pressure): Mark as optimization.
3878 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
3880 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
3881 an OpenMP mappable type.
3883 2014-03-06 Matthias Klose <doko@ubuntu.com>
3885 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
3886 MULTILIB_OSDIRNAMES is not defined.
3888 2014-03-06 Jakub Jelinek <jakub@redhat.com>
3889 Meador Inge <meadori@codesourcery.com>
3892 * config/arm/arm.c (arm_tls_symbol_p): Remove.
3893 (arm_legitimize_address): Call legitimize_tls_address for any
3894 arm_tls_referenced_p expression, handle constant addend. Call it
3895 before testing for !TARGET_ARM.
3896 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
3898 2014-03-06 Richard Biener <rguenther@suse.de>
3904 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
3906 * tree-streamer.c (record_common_node): Assert we don't record
3907 nodes with type double.
3908 (preload_common_node): Skip type double, complex double and double
3909 pointer since it is now frontend dependent due to fshort-double option.
3911 2014-03-06 Richard Biener <rguenther@suse.de>
3913 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
3914 or -fno-lto is specified and the linker has full plugin support.
3915 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
3916 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
3917 * lto-wrapper.c (merge_and_complain): Merge compile-time
3918 optimization levels.
3919 (run_gcc): And pass it through to the link options.
3921 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
3925 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
3927 * cselib.c (remove_useless_values): Skip to avoid quadratic
3928 behavior if the condition moved from...
3929 (cselib_process_insn): ... here holds.
3931 2014-03-05 Jakub Jelinek <jakub@redhat.com>
3934 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
3935 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
3938 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
3939 (TM_H): Add x86-tune.def.
3941 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3943 * config/aarch64/aarch64.c (generic_tunings):
3944 Use cortexa57_extra_costs.
3946 2014-03-05 Jakub Jelinek <jakub@redhat.com>
3949 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
3950 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
3951 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
3954 2014-03-04 Heiher <r@hev.cc>
3956 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
3957 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
3959 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
3961 * config/i386/predicates.md (const2356_operand): Change to ...
3962 (const2367_operand): ... this.
3963 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
3965 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
3966 (*avx512pf_scatterpf<mode>sf): Ditto.
3967 (avx512pf_scatterpf<mode>df): Ditto.
3968 (*avx512pf_scatterpf<mode>df_mask): Ditto.
3969 (*avx512pf_scatterpf<mode>df): Ditto.
3970 * config/i386/i386.c (ix86_expand_builtin): Update
3971 incorrect hint operand error message.
3973 2014-03-04 Richard Biener <rguenther@suse.de>
3975 * lto-section-in.c (lto_get_section_data): Fix const cast.
3977 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
3979 * tree-streamer.c (record_common_node): Assert we don't record
3980 nodes with type double.
3981 (preload_common_node): Skip type double, complex double and double
3982 pointer since it is now frontend dependent due to fshort-double option.
3984 2014-03-04 Richard Biener <rguenther@suse.de>
3987 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
3988 (lto_input_toplevel_asms): Likewise.
3989 * lto-section-in.c (lto_get_section_data): Instead do it here
3992 2014-03-04 Richard Biener <rguenther@suse.de>
3994 PR tree-optimization/60382
3995 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
3996 dead PHIs a reduction.
3998 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4000 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4002 (_mm_prefetch): Move out of GCC target("sse") pragma.
4003 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4004 GCC target("prfchw") pragma.
4005 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4007 * config/i386/i386.c (ix86_option_override_internal): Enable
4008 -mprfchw with -mprefetchwt1.
4010 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4012 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4015 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4017 * opts.h (CL_PCH_IGNORE): Define.
4018 * targhooks.c (option_affects_pch_p):
4019 Return false for options that have CL_PCH_IGNORE set.
4020 * opt-functions.awk: Process PchIgnore.
4021 * doc/options.texi: Document PchIgnore.
4023 * config/arc/arc.opt (misize): Add PchIgnore property.
4025 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4027 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4028 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4029 constraint on constants to permit them being loaded into
4030 GENERAL_REGS or BASE_REGS.
4032 2014-03-03 Nick Clifton <nickc@redhat.com>
4034 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4035 anti-cacnonical alternatives.
4036 (negandhi3_real): New pattern.
4037 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4039 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4041 * config/avr/avr-mcus.def: Remove atxmega16x1.
4042 * config/avr/avr-tables.opt: Regenerate.
4043 * config/avr/t-multilib: Regenerate.
4044 * doc/avr-mmcu.texi: Regenerate.
4046 2014-03-03 Tobias Grosser <tobias@grosser.es>
4047 Mircea Namolaru <mircea.namolaru@inria.fr>
4049 PR tree-optimization/58028
4050 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4053 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4055 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4056 not handled by recognizers.
4058 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4061 * function.c (expand_function_end): Don't emit
4062 clobber_return_register sequence if clobber_after is a BARRIER.
4063 * cfgexpand.c (construct_exit_block): Append instructions before
4064 return_label to prev_bb.
4066 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4068 * config/rs6000/constraints.md: Document reserved use of "wc".
4070 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4073 * ipa.c (function_and_variable_visibility): When dissolving comdat
4074 group, also set all symbols to local.
4076 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4081 2013-12-14 Jan Hubicka <jh@suse.cz>
4083 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4085 2014-03-02 Jon Beniston <jon@beniston.com>
4091 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4092 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4093 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4094 (simple_return, *simple_return): New patterns
4095 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4096 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4098 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4100 * dwarf2out.c (gen_subprogram_die): Tidy.
4102 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4105 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4106 (*mov_t_msb_neg_negc): ... this new insn.
4108 2014-02-28 Jason Merrill <jason@redhat.com>
4111 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4114 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4117 * dwarf2out.c (decltype_auto_die): New static.
4118 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4119 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4120 (is_cxx_auto): Likewise.
4122 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4124 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4125 we are not using general regs only.
4127 2014-02-28 Richard Biener <rguenther@suse.de>
4130 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4131 previous fix and only allow to remove trivial pre-headers
4132 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4133 (remove_forwarder_block): Properly update the latch of a loop.
4135 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4138 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4139 (cselib_preserved_hash_table): New.
4140 (preserve_constants_and_equivs): Move preserved vals to it.
4141 (cselib_find_slot): Look it up first.
4142 (cselib_init): Initialize it.
4143 (cselib_finish): Release it.
4144 (dump_cselib_table): Dump it.
4146 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4149 * cselib.c (remove_useless_values): Skip to avoid quadratic
4150 behavior if the condition moved from...
4151 (cselib_process_insn): ... here holds.
4153 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4156 * var-tracking.c (vt_initialize): Apply the same condition to
4157 preserve the CFA base value.
4159 2014-02-28 Joey Ye <joey.ye@arm.com>
4162 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4163 if reload in progress or completed.
4165 2014-02-28 Tobias Burnus <burnus@net-b.de>
4168 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4171 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4173 * doc/tm.texi.in (Condition Code Status): Update documention for
4174 relative locations of cc0-setter and cc0-user.
4176 2014-02-27 Jeff Law <law@redhat.com>
4178 PR rtl-optimization/52714
4179 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4180 into two independent simple sets, if I3 is a jump, ensure the
4181 pattern we place into I3 is a (set (pc) ...).
4183 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4184 Jeff Law <law@redhat.com>
4186 PR rtl-optimization/49847
4187 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4188 are in different blocks.
4189 * doc/tm.texi (Condition Code Status): Update documention for
4190 relative locations of cc0-setter and cc0-user.
4192 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4195 * lra.c (lra_emit_add): Check SUBREG too.
4197 2014-02-27 Andreas Schwab <schwab@suse.de>
4199 * config/m68k/m68k.c (m68k_option_override): Disable
4200 -flive-range-shrinkage for classic m68k.
4201 (m68k_override_options_after_change): Likewise.
4203 2014-02-27 Marek Polacek <polacek@redhat.com>
4206 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4207 -Wmaybe-uninitialized.
4209 2014-02-27 Alan Modra <amodra@gmail.com>
4212 * reload1.c (emit_input_reload_insns): When reload_override_in,
4213 set old to rl->in_reg when rl->in_reg is a subreg.
4215 2014-02-26 Richard Biener <rguenther@suse.de>
4218 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4220 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4222 * common/config/i386/predicates.md (const1256_operand): Remove.
4223 (const2356_operand): New.
4224 (const_1_to_2_operand): Remove.
4225 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4226 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4227 (*avx512pf_gatherpf<mode>sf): Ditto.
4228 (avx512pf_gatherpf<mode>df): Ditto.
4229 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4230 (*avx512pf_gatherpf<mode>df): Ditto.
4231 (avx512pf_scatterpf<mode>sf): Ditto.
4232 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4233 (*avx512pf_scatterpf<mode>sf): Ditto.
4234 (avx512pf_scatterpf<mode>df): Ditto.
4235 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4236 (*avx512pf_scatterpf<mode>df): Ditto.
4237 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4239 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4241 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4242 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4243 (_mm512_mask_testn_epi64_mask): Move to ...
4244 * config/i386/avx512cdintrin.h: Here.
4245 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4246 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4247 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4248 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4249 TARGET_AVX512F from TARGET_AVX512CD.
4251 2014-02-26 Richard Biener <rguenther@suse.de>
4254 * ipa.c (walk_polymorphic_call_targets): Properly guard
4255 call to inline_update_overall_summary.
4257 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4260 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4261 and latches only if requested. Fix latch if it is removed.
4262 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4263 LOOPS_HAVE_PREHEADERS.
4265 2014-02-25 Andrew Pinski <apinski@cavium.com>
4267 * builtins.c (expand_builtin_thread_pointer): Create a new target
4268 when the target is NULL.
4270 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4272 PR rtl-optimization/60317
4273 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4274 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4275 * lra-assigns.c: Include params.h.
4276 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4277 other reload pseudos considerations.
4279 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4281 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4282 to use canonical form for nor<mode>3.
4284 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4287 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4290 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4292 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4293 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4294 (ix86_handle_option): Handle OPT_mprefetchwt1.
4295 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4296 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4298 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4299 OPTION_MASK_ISA_PREFETCHWT1.
4300 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4301 (PTA_PREFETCHWT1): New.
4302 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4303 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4304 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4305 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4306 (*prefetch_avx512pf_<mode>_: Change into ...
4307 (*prefetch_prefetchwt1_<mode>: This.
4308 * config/i386/i386.opt (mprefetchwt1): New.
4309 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4310 (_mm_prefetch): Handle intent to write.
4311 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4313 2014-02-25 Richard Biener <rguenther@suse.de>
4316 * emit-rtl.c (mem_attrs_htab): Remove.
4317 (mem_attrs_htab_hash): Likewise.
4318 (mem_attrs_htab_eq): Likewise.
4319 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4320 (init_emit_once): Do not allocate mem_attrs_htab.
4322 2014-02-25 Richard Biener <rguenther@suse.de>
4325 * lto-opts.c (lto_write_options): Output non-explicit conservative
4326 -fwrapv, -fno-trapv and -fno-strict-overflow.
4327 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4328 (run_gcc): And pass them through.
4330 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4332 * sel-sched.c (calculate_new_fences): New parameter ptime.
4333 Calculate it as a maximum over all fence cycles.
4334 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4335 Print the final schedule timing when sched_verbose.
4337 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4339 PR rtl-optimization/60292
4340 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4341 bit fot the fence instruction.
4343 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4345 * calls.h: Fix typo in comment.
4347 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4349 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4350 adjusting offsetable addresses.
4352 2014-02-24 Guozhi Wei <carrot@google.com>
4354 * sparseset.h (sparseset_pop): Fix the wrong index.
4356 2014-02-24 Walter Lee <walt@tilera.com>
4358 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4359 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4361 * common/config/tilegx/tilegx-common.c
4362 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4363 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4365 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4366 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4367 (tilegx_gimplify_va_arg_expr): Handle big endian.
4368 (tilegx_expand_unaligned_load): Ditto.
4369 (tilegx_expand_unaligned_store): Ditto.
4370 (TARGET_RETURN_IN_MSB): New.
4371 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4372 (TARGET_ENDIAN_DEFAULT): New.
4373 (TARGET_BIG_ENDIAN): Handle big endian.
4374 (BYTES_BIG_ENDIAN): Ditto.
4375 (WORDS_BIG_ENDIAN): Ditto.
4376 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4379 * config/tilegx/tilegx.md (extv): Handle big endian.
4381 (insn_st<n>): Ditto.
4382 (insn_st<n>_add<bitsuffix>): Ditto.
4383 (insn_stnt<n>): Ditto.
4384 (insn_stnt<n>_add<bitsuffix>):Ditto.
4385 (vec_interleave_highv8qi): Handle big endian.
4386 (vec_interleave_highv8qi_be): New.
4387 (vec_interleave_highv8qi_le): New.
4388 (insn_v1int_h): Handle big endian.
4389 (vec_interleave_lowv8qi): Handle big endian.
4390 (vec_interleave_lowv8qi_be): New.
4391 (vec_interleave_lowv8qi_le): New.
4392 (insn_v1int_l): Handle big endian.
4393 (vec_interleave_highv4hi): Handle big endian.
4394 (vec_interleave_highv4hi_be): New.
4395 (vec_interleave_highv4hi_le): New.
4396 (insn_v2int_h): Handle big endian.
4397 (vec_interleave_lowv4hi): Handle big endian.
4398 (vec_interleave_lowv4hi_be): New.
4399 (vec_interleave_lowv4hi_le): New.
4400 (insn_v2int_l): Handle big endian.
4401 (vec_interleave_highv2si): Handle big endian.
4402 (vec_interleave_highv2si_be): New.
4403 (vec_interleave_highv2si_le): New.
4404 (insn_v4int_h): Handle big endian.
4405 (vec_interleave_lowv2si): Handle big endian.
4406 (vec_interleave_lowv2si_be): New.
4407 (vec_interleave_lowv2si_le): New.
4408 (insn_v4int_l): Handle big endian.
4409 * config/tilegx/tilegx.opt (mbig-endian): New option.
4410 (mlittle-endian): New option.
4411 * doc/install.texi: Document tilegxbe-linux.
4412 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4414 2014-02-24 Martin Jambor <mjambor@suse.cz>
4417 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4418 there are no parameter descriptors.
4420 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4422 PR rtl-optimization/60268
4423 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4424 initialization to ...
4425 (sched_rgn_init): ... here.
4426 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4428 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4430 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4433 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4435 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4438 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4440 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4441 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4443 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4445 * config/microblaze/predicates.md: Add cmp_op predicate.
4446 * config/microblaze/microblaze.md: Add branch_compare instruction
4447 which uses cmp_op predicate and emits cmp insn before branch.
4448 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4449 to microblaze_expand_conditional_branch and consolidate logic.
4450 (microblaze_expand_conditional_branch): emit branch_compare
4451 insn instead of handling cmp op separate from branch insn.
4453 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4455 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4458 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4460 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4461 define_insn with define_expand and new define_insn
4462 *altivec_lve<VI_char>x_internal.
4463 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4464 and new define_insn *altivec_stve<VI_char>x_internal.
4465 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4467 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4469 (altivec_expand_stvex_be): New function.
4471 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4473 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4474 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4475 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4476 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4478 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4481 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4482 instead of emit_move_insn.
4484 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4486 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4488 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4489 gen_altivec_vsumsws.
4491 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4493 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4494 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4495 (altivec_lvxl_<mode>): New define_expand incorporating
4496 -maltivec=be semantics where needed.
4497 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4498 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4499 semantics where needed.
4500 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4501 (altivec_stvx_<mode>): New define_expand incorporating
4502 -maltivec=be semantics where needed.
4503 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4504 VM2 iterator instead of V4SI.
4505 (altivec_stvxl_<mode>): New define_expand incorporating
4506 -maltivec=be semantics where needed.
4507 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4508 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4509 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4510 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4511 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4512 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4513 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4514 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4515 ALTIVEC_BUILTIN_STVXL.
4516 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4517 (altivec_expand_stvx_be): Likewise.
4518 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4519 (altivec_expand_lvx_be): Likewise.
4520 (altivec_expand_stvx_be): Likewise.
4521 (altivec_expand_builtin): Add cases for
4522 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4523 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4524 (altivec_init_builtins): Add definitions for
4525 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4526 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4528 2014-02-21 Catherine Moore <clm@codesourcery.com>
4530 * doc/invoke.texi (mvirt, mno-virt): Document.
4531 * config/mips/mips.opt (mvirt): New option.
4532 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4534 2014-02-21 Richard Biener <rguenther@suse.de>
4536 PR tree-optimization/60276
4537 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4538 (STMT_VINFO_MIN_NEG_DIST): New macro.
4539 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4540 STMT_VINFO_MIN_NEG_DIST.
4541 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4542 made for negative dependence distances still hold.
4544 2014-02-21 Richard Biener <rguenther@suse.de>
4547 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4548 DECL_INITIAL for globals not in the current function context.
4550 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4552 PR tree-optimization/56490
4553 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4554 * tree-ssa-uninit.c: Include params.h.
4555 (compute_control_dep_chain): Add num_calls argument, return false
4556 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4557 num_calls to recursive call.
4558 (find_predicates): Change dep_chain into normal array,
4559 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4560 variable and adjust compute_control_dep_chain caller.
4561 (find_def_preds): Likewise.
4563 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4565 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4566 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4568 2014-02-21 Nick Clifton <nickc@redhat.com>
4570 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4571 (pushhi1): Likewise.
4572 (popqi1): Add mode to pre_dec.
4575 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4577 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4578 mode for mask of V8SFmode permutation.
4580 2014-02-20 Richard Henderson <rth@redhat.com>
4583 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4584 a new pseudo for OLDVAL.
4586 2014-02-20 Jakub Jelinek <jakub@redhat.com>
4589 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4590 gen_reg_rtx if d->testing_p.
4591 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4592 if d->testing_p and we will certainly return true.
4593 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4596 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
4598 * emit-rtl.c (gen_reg_rtx): Assert that
4599 crtl->emit.regno_pointer_align_length is non-zero.
4601 2014-02-20 Richard Henderson <rth@redhat.com>
4604 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4605 on failure the store back into EXPECT.
4607 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4608 Sandra Loosemore <sandra@codesourcery.com>
4610 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
4611 * config/nios2/nios2.c (nios2_function_profiler): Add
4612 -fPIC (flag_pic == 2) support.
4613 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
4614 (nios2_large_offset_p): New function.
4615 (nios2_unspec_reloc_p): Move up position, update to use
4616 nios2_large_offset_p.
4617 (nios2_unspec_address): Remove function.
4618 (nios2_unspec_offset): New function.
4619 (nios2_large_got_address): New function.
4620 (nios2_got_address): Add large offset support.
4621 (nios2_legitimize_tls_address): Update usage of removed and new
4623 (nios2_symbol_binds_local_p): New function.
4624 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
4625 (nios2_legitimize_address): Update to use nios2_large_offset_p.
4626 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
4627 (nios2_print_operand): Merge H/L processing, add hiadj/lo
4628 processing for (const (unspec ...)).
4629 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
4631 2014-02-20 Richard Biener <rguenther@suse.de>
4633 * tree-cfg.c (replace_uses_by): Mark altered BBs before
4634 doing the substitution.
4635 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
4637 2014-02-20 Martin Jambor <mjambor@suse.cz>
4640 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
4641 info when checking whether lattices are bottom.
4643 2014-02-20 Richard Biener <rguenther@suse.de>
4646 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
4649 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
4652 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
4653 parameter specifying the scaling.
4654 (inline_call): Update.
4655 (want_inline_recursively): Guard division by zero.
4656 (recursive_inlining): Update.
4657 * ipa-inline.h (clone_inlined_nodes): Update.
4659 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4662 * config/i386/i386.c (classify_argument): Pass structures of size
4663 64 bytes or less in register.
4665 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4666 Kirill Yukhin <kirill.yukhin@intel.com>
4668 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
4669 (_mm_rcp28_round_ss): Ditto.
4670 (_mm_rsqrt28_round_sd): Ditto.
4671 (_mm_rsqrt28_round_ss): Ditto.
4672 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
4673 (_mm_rcp14_round_ss): Ditto.
4674 (_mm_rsqrt14_round_sd): Ditto.
4675 (_mm_rsqrt14_round_ss): Ditto.
4676 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
4677 the first input operand, get rid of match_dup.
4678 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
4680 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
4682 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
4683 operand as the first input operand, set type attribute.
4684 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
4686 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
4687 operand as the first input operand, set type attribute.
4689 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4691 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
4694 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
4697 * config/i386/i386.c (construct_container): Remove TFmode check
4698 for X86_64_INTEGER_CLASS.
4700 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
4703 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
4704 only when -Wpsabi is enabled.
4706 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
4709 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
4710 passing arrays in registers are the same as for structs, so remove the
4711 special case for them.
4713 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
4715 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
4716 destination type, extract only the valid bits if the source type is not
4717 integral and has a different mode.
4719 2014-02-19 Richard Biener <rguenther@suse.de>
4722 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
4725 2014-02-19 Richard Biener <rguenther@suse.de>
4728 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
4729 (ipa_modify_call_arguments): Emit an argument load explicitely and
4730 preserve virtual SSA form there and for the replacement call.
4731 Do not update SSA form nor free dominance info.
4733 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4735 * ipa.c (function_and_variable_visibility): Also clear WEAK
4736 flag when disolving COMDAT_GROUP.
4738 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4740 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
4741 * ipa-prop.c (ipa_set_jf_known_type): Return early when
4743 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
4744 do more sanity checks.
4745 (detect_type_change): Return true when giving up early.
4746 (compute_complex_assign_jump_func): Fix type parameter of
4747 ipa_set_ancestor_jf.
4748 (compute_complex_ancestor_jump_func): Likewise.
4749 (update_jump_functions_after_inlining): Fix updating of
4751 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
4753 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4755 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
4756 inline clones when edge disappears.
4758 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4761 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
4762 Split 64-bit moves into 2 patterns. Do not allow the use of
4763 direct move for TDmode in little endian, since the decimal value
4764 has little endian bytes within a word, but the 64-bit pieces are
4765 ordered in a big endian fashion, and normal subreg's of TDmode are
4767 (mov<mode>_64bit_dm): Likewise.
4768 (movtd_64bit_nodm): Likewise.
4770 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4772 PR tree-optimization/60174
4773 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
4774 statement of an SSA_NAME that occurs in an abnormal PHI node.
4776 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4779 * final.c (SEEN_BB): Remove.
4780 (SEEN_NOTE, SEEN_EMITTED): Renumber.
4781 (final_scan_insn): Don't force_source_line on second
4782 NOTE_INSN_BASIC_BLOCK.
4784 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
4787 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
4788 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
4789 (type_natural_mode): Warn ABI change when %zmm register is not
4790 available for AVX512F vector value passing.
4792 2014-02-18 Kai Tietz <ktietz@redhat.com>
4795 * config/i386/i386.c (ix86_expand_prologue): Use value in
4796 rax register as displacement when restoring %r10 or %rax.
4797 Fix wrong offset when restoring both registers.
4799 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4801 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
4802 assertion with conditional return.
4804 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4805 Uros Bizjak <ubizjak@gmail.com>
4808 * config/i386/driver-i386.c (host_detect_local_cpu): If
4809 YMM state is not saved by the OS, also clear has_f16c. Move
4810 CPUID 0x80000001 handling before YMM state saving checking.
4812 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
4814 PR rtl-optimization/58960
4815 * haifa-sched.c (alloc_global_sched_pressure_data): New,
4816 factored out from ...
4817 (sched_init): ... here.
4818 (free_global_sched_pressure_data): New, factored out from ...
4819 (sched_finish): ... here.
4820 * sched-int.h (free_global_sched_pressure_data): Declare.
4821 * sched-rgn.c (nr_regions_initial): New static global.
4822 (haifa_find_rgns): Initialize it.
4823 (schedule_region): Disable sched-pressure for the newly
4826 2014-02-17 Richard Biener <rguenther@suse.de>
4828 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
4829 release SSA defs of pattern stmts.
4831 2014-02-17 Richard Biener <rguenther@suse.de>
4833 * tree-inline.c (expand_call_inline): Release the virtual
4834 operand defined by the call we are about to inline.
4836 2014-02-17 Richard Biener <rguenther@suse.de>
4838 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
4840 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
4841 Ilya Tocar <ilya.tocar@intel.com>
4843 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
4844 arguments order in builtin.
4845 (_mm512_permutexvar_epi64): Ditto.
4846 (_mm512_mask_permutexvar_epi64): Ditto
4847 (_mm512_maskz_permutexvar_epi32): Ditto
4848 (_mm512_permutexvar_epi32): Ditto
4849 (_mm512_mask_permutexvar_epi32): Ditto
4851 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4853 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
4854 (p8_vmrgow): Likewise.
4856 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4858 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
4861 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
4864 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
4865 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
4866 into 64-bit and 32-bit moves. On 64-bit moves, add support for
4867 using direct move instructions on ISA 2.07. Also adjust
4868 instruction length for 64-bit.
4869 (mov<mode>_64bit, TFmode/TDmode): Likewise.
4870 (mov<mode>_32bit, TFmode/TDmode): Likewise.
4872 2014-02-15 Alan Modra <amodra@gmail.com>
4876 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
4877 find_replacement on parts of insn rtl that might be reloaded.
4879 2014-02-15 Richard Biener <rguenther@suse.de>
4881 PR tree-optimization/60183
4882 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
4883 (tree_ssa_phiprop): Calculate and free post-dominators.
4885 2014-02-14 Jeff Law <law@redhat.com>
4887 PR rtl-optimization/60131
4888 * ree.c (get_extended_src_reg): New function.
4889 (combine_reaching_defs): Use it rather than assuming location of REG.
4890 (find_and_remove_re): Verify first operand of extension is
4891 a REG before adding the insns to the copy list.
4893 2014-02-14 Roland McGrath <mcgrathr@google.com>
4895 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
4896 * configure: Regenerated.
4897 * config.in: Regenerated.
4898 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
4899 instead of ASM_SHORT.
4901 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
4902 Richard Earnshaw <rearnsha@arm.com>
4904 PR rtl-optimization/59535
4905 * lra-constraints.c (process_alt_operands): Encourage alternative
4906 when unassigned pseudo class is superset of the alternative class.
4907 (inherit_reload_reg): Don't inherit when optimizing for code size.
4908 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
4909 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
4910 modes not less than 4 for Thumb1.
4912 2014-02-14 Kyle McMartin <kyle@redhat.com>
4915 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
4917 2014-02-14 Richard Biener <rguenther@suse.de>
4919 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
4920 (get_frame_arg): Drop the assert with langhook types_compatible_p.
4921 Do not strip INDIRECT_REFs.
4923 2014-02-14 Richard Biener <rguenther@suse.de>
4926 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
4927 DECL_FUNCTION_SPECIFIC_TARGET.
4928 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
4929 * tree-streamer-out.c (pack_ts_target_option): Remove.
4930 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
4931 (write_ts_function_decl_tree_pointers): Do not stream
4932 DECL_FUNCTION_SPECIFIC_TARGET.
4933 * tree-streamer-in.c (unpack_ts_target_option): Remove.
4934 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
4935 (lto_input_ts_function_decl_tree_pointers): Do not stream
4936 DECL_FUNCTION_SPECIFIC_TARGET.
4938 2014-02-14 Jakub Jelinek <jakub@redhat.com>
4940 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
4941 (get_initial_def_for_induction, vectorizable_induction): Ignore
4942 debug stmts when looking for exit_phi.
4943 (vectorizable_live_operation): Fix up condition.
4945 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
4947 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
4948 nreverse() because it changes the content of original tree list.
4950 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
4952 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
4953 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
4955 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
4957 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
4958 GNU coding standards.
4960 2014-02-13 Jakub Jelinek <jakub@redhat.com>
4963 * dwarf2out.c (gen_subprogram_die): Don't call
4964 add_calling_convention_attribute if subr_die is old_die.
4966 2014-02-13 Sharad Singhai <singhai@google.com>
4968 * doc/optinfo.texi: Fix order of nodes.
4970 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
4972 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
4973 operands[2], not operands[3].
4975 2014-02-13 Richard Biener <rguenther@suse.de>
4978 * doc/install.texi (ISL): Update recommended version to 0.12.2,
4979 mention the possibility of an in-tree build.
4980 (CLooG): Update recommended version to 0.18.1, mention the
4981 possibility of an in-tree build and clarify that the ISL
4982 bundled with CLooG does not work.
4984 2014-02-13 Jakub Jelinek <jakub@redhat.com>
4987 * expr.c (compress_float_constant): If x is a hard register,
4988 extend into a pseudo and then move to x.
4990 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
4992 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
4993 caused by bad second argument to warning_at() with -mhotpatch and
4994 nested functions (e.g. with gfortran).
4996 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
4998 * opts.c (option_name): Remove "enabled by default" rider.
5000 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5002 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5004 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5005 Uros Bizjak <ubizjak@gmail.com>
5008 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5009 * configure: Regenerated.
5011 2014-02-12 Richard Biener <rguenther@suse.de>
5013 * vec.c (vec_prefix::calculate_allocation): Move as
5014 inline variant to vec.h.
5015 (vec_prefix::calculate_allocation_1): New out-of-line version.
5016 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5017 (vec_prefix::m_has_auto_buf): Rename to ...
5018 (vec_prefix::m_using_auto_storage): ... this.
5019 (vec_prefix::calculate_allocation): Inline the easy cases
5020 and dispatch to calculate_allocation_1 which doesn't need the
5022 (va_heap::reserve): Use gcc_checking_assert.
5023 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5024 m_using_auto_storage.
5025 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5027 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5028 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5029 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5031 2014-02-12 Richard Biener <rguenther@suse.de>
5033 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5034 when we found a dependence.
5036 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5038 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5040 (maybe_fold_stmt): ... into this new function.
5041 * omp-low.c (lower_omp): Update comment.
5043 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5046 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5049 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5051 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5052 identifiers in comments.
5053 (cortexa53_extra_costs): Likewise.
5054 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5055 (cortexa7_extra_costs): Likewise.
5056 (cortexa12_extra_costs): Likewise.
5057 (cortexa15_extra_costs): Likewise.
5058 (v7m_extra_costs): Likewise.
5060 2014-02-12 Richard Biener <rguenther@suse.de>
5063 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5064 of posix_memalign being successful.
5065 (lower_stmt): Restrict lowering of posix_memalign to when
5066 -ftree-bit-ccp is enabled.
5068 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5070 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5072 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5074 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5076 PR rtl-optimization/60116
5077 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5078 other_insn once the combination has been validated.
5080 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5083 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5085 * ipa-devirt.c: Include demangle.h
5086 (odr_violation_reported): New static variable.
5087 (add_type_duplicate): Update odr_violations.
5088 (maybe_record_node): Add completep parameter; update it.
5089 (record_target_from_binfo): Add COMPLETEP parameter;
5090 update it as needed.
5091 (possible_polymorphic_call_targets_1): Likewise.
5092 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5093 rename FINAL to COMPLETE.
5094 (record_targets_from_bases): Sanity check we found the binfo;
5095 fix COMPLETEP updating.
5096 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5097 parameter, fix computing of COMPLETEP.
5098 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5099 at LTO time do demangling.
5100 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5101 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5103 (gimple_get_virt_method_for_binfo): Likewise.
5104 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5105 gimple_get_virt_method_for_vtable): Update prototypes.
5107 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5110 * genautomata.c (add_presence_absence): Fix typo with
5111 {final_}presence_list.
5113 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5116 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5117 for VSX/Altivec vectors that land in GPR registers.
5119 2014-02-11 Richard Henderson <rth@redhat.com>
5120 Jakub Jelinek <jakub@redhat.com>
5123 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5124 around drhs if type conversion to lacc->type is not useless.
5126 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5128 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5130 (cortex-a57.cortex-a53): Likewise.
5131 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5133 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5135 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5138 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5140 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5141 add_options_for_arm_vfp3.
5143 2014-02-11 Jeff Law <law@redhat.com>
5146 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5147 object with an undesirable mode.
5149 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5152 * config/i386/sol2-9.h: New file.
5153 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5154 *-*-solaris2.9*): Use it.
5156 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5158 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5159 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5161 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5163 * config/microblaze/microblaze.c: Extend mcpu version format
5165 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5167 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5169 2014-02-10 Richard Henderson <rth@redhat.com>
5172 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5173 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5174 ms-abi vs -mno-accumulate-outgoing-args.
5175 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5176 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5179 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5182 * cfgexpand.c (expand_asm_operands): Attach source location to
5183 ASM_INPUT rtx objects.
5184 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5186 2014-02-10 Nick Clifton <nickc@redhat.com>
5188 * config/mn10300/mn10300.c (popcount): New function.
5189 (mn10300_expand_prologue): Include saved registers in stack usage
5192 2014-02-10 Jeff Law <law@redhat.com>
5195 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5196 when changing the SET_DEST of a prior insn to avoid an input reload.
5198 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5200 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5201 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5202 -mcall-openbsd, or -mcall-linux.
5203 (CC1_ENDIAN_BIG_SPEC): Remove.
5204 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5205 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5206 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5207 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5208 and %cc1_endian_default.
5209 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5211 2014-02-10 Richard Biener <rguenther@suse.de>
5213 PR tree-optimization/60115
5214 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5215 MEM_REF handling. Properly verify that the accesses are not
5216 out of the objects bound.
5218 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5220 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5223 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5225 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5226 proper constants and fix formatting.
5227 (possible_polymorphic_call_targets): Fix formatting.
5229 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5230 Ilya Tocar <ilya.tocar@intel.com>
5232 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5233 (_mm512_loadu_epi32): Renamed into...
5234 (_mm512_loadu_si512): This.
5235 (_mm512_storeu_epi32): Renamed into...
5236 (_mm512_storeu_si512): This.
5237 (_mm512_maskz_ceil_ps): Removed.
5238 (_mm512_maskz_ceil_pd): Ditto.
5239 (_mm512_maskz_floor_ps): Ditto.
5240 (_mm512_maskz_floor_pd): Ditto.
5241 (_mm512_floor_round_ps): Ditto.
5242 (_mm512_floor_round_pd): Ditto.
5243 (_mm512_ceil_round_ps): Ditto.
5244 (_mm512_ceil_round_pd): Ditto.
5245 (_mm512_mask_floor_round_ps): Ditto.
5246 (_mm512_mask_floor_round_pd): Ditto.
5247 (_mm512_mask_ceil_round_ps): Ditto.
5248 (_mm512_mask_ceil_round_pd): Ditto.
5249 (_mm512_maskz_floor_round_ps): Ditto.
5250 (_mm512_maskz_floor_round_pd): Ditto.
5251 (_mm512_maskz_ceil_round_ps): Ditto.
5252 (_mm512_maskz_ceil_round_pd): Ditto.
5253 (_mm512_expand_pd): Ditto.
5254 (_mm512_expand_ps): Ditto.
5255 * config/i386/i386.c (ix86_builtins): Remove
5256 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5257 (bdesc_args): Ditto.
5258 * config/i386/predicates.md (const1256_operand): New.
5259 (const_1_to_2_operand): Ditto.
5260 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5261 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5262 (*avx512pf_gatherpf<mode>sf): Ditto.
5263 (avx512pf_gatherpf<mode>df): Ditto.
5264 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5265 (*avx512pf_gatherpf<mode>df): Ditto.
5266 (avx512pf_scatterpf<mode>sf): Ditto.
5267 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5268 (*avx512pf_scatterpf<mode>sf): Ditto.
5269 (avx512pf_scatterpf<mode>df): Ditto.
5270 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5271 (*avx512pf_scatterpf<mode>df): Ditto.
5272 (avx512f_expand<mode>): Removed.
5273 (<shift_insn><mode>3<mask_name>): Change predicate type.
5275 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5277 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5278 not at the end of datarefs vector use ordered_remove to avoid
5279 reordering datarefs vector.
5282 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5283 mark local addressable non-static vars as GOVD_PRIVATE
5284 instead of GOVD_LOCAL.
5285 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5286 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5290 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5291 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5292 assume_aligned or alloc_align attributes.
5293 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5294 arguments. Handle also assume_aligned and alloc_align attributes.
5295 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5296 calls to functions with assume_aligned or alloc_align attributes.
5297 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5299 2014-02-08 Terry Guo <terry.guo@arm.com>
5301 * doc/invoke.texi: Document ARM -march=armv7e-m.
5303 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5305 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5306 flag on __cilkrts_rethrow builtin.
5309 * ipa-cp.c (determine_versionability): Fail at -O0
5310 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5311 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5314 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5317 * tree-inline.c (copy_forbidden): Fail for
5318 __attribute__((optimize (0))) functions.
5320 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5322 * varpool.c: Include pointer-set.h.
5323 (varpool_remove_unreferenced_decls): Variables in other partitions
5324 will not be output; be however careful to not lose information
5327 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5329 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5330 lookup in the vtable constructor.
5332 2014-02-07 Jeff Law <law@redhat.com>
5335 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5336 define_insn_and_split.
5338 * ipa-inline.c (inline_small_functions): Fix typos.
5340 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5342 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5343 (s390_can_use_return_insn): Declare.
5344 * config/s390/s390.h (EPILOGUE_USES): Define.
5345 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5347 (s390_chunkify_start): Handle return JUMP_LABELs.
5348 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5349 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5350 (s390_can_use_return_insn): New functions.
5351 (s390_fix_long_loop_prediction): Handle conditional returns.
5352 (TARGET_SET_UP_BY_PROLOGUE): Define.
5353 * config/s390/s390.md (ANY_RETURN): New code iterator.
5354 (*creturn, *csimple_return, return, simple_return): New patterns.
5356 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5358 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5359 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5360 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5361 REG_CFA_RESTORE list when deciding not to restore a register.
5363 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5365 * config/s390/s390.c: Include tree-pass.h and context.h.
5366 (s390_early_mach): New function, split out from...
5367 (s390_emit_prologue): ...here.
5368 (pass_data_s390_early_mach): New pass structure.
5369 (pass_s390_early_mach): New class.
5370 (s390_option_override): Create and register early_mach pass.
5371 Move to end of file.
5373 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5375 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5376 to match for the exit block.
5378 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5380 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5381 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5382 Reject misaligned operands.
5384 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5386 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5388 2014-02-07 Richard Biener <rguenther@suse.de>
5391 * gimple-low.c (lower_builtin_posix_memalign): New function.
5392 (lower_stmt): Call it to lower posix_memalign in a way
5393 to make alignment info accessible.
5395 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5398 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5399 __builtin_setjmp_receiver.
5401 2014-02-07 Richard Biener <rguenther@suse.de>
5404 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5405 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5406 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5407 Handle BUILT_IN_POSIX_MEMALIGN.
5408 (find_func_clobbers): Likewise.
5409 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5410 (call_may_clobber_ref_p_1): Likewise.
5412 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5415 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5418 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5421 * lto-cgraph.c (lto_output_node): Use
5422 symtab_get_symbol_partitioning_class.
5423 (lto_output_varpool_node): likewise.
5424 (symtab_get_symbol_partitioning_class): Move here from
5426 * cgraph.h (symbol_partitioning_class): Likewise.
5427 (symtab_get_symbol_partitioning_class): Declare.
5429 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5431 * ggc.h (ggc_internal_cleared_alloc): New macro.
5432 * vec.h (vec_safe_copy): Handle memory stats.
5433 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5434 * target-globals.c (save_target_globals): Likewise.
5436 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5439 * expr.c (emit_move_resolve_push): Export; be bit more selective
5440 on when to clear alias set.
5441 * expr.h (emit_move_resolve_push): Declare.
5442 * function.h (struct function): Add tail_call_marked.
5443 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5444 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5445 * config/i386/i386.md (TImode move expander): De not call
5447 (FP push expanders): Preserve memory attributes.
5448 * config/i386/sse.md (push<mode>1): Remove.
5449 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5450 (ix86_expand_push): Remove.
5451 * config/i386/mmx.md (push<mode>1): Remove.
5453 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5455 PR rtl-optimization/60030
5456 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5457 lopart with paradoxical subreg before shifting it up by hprec.
5459 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5461 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5462 Remove extra newline at end of file.
5463 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5464 (arm_issue_rate): Handle cortexa57.
5465 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5466 (cortex-a57.cortex-a53): Likewise.
5468 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5471 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5472 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5474 (arm_expand_prologue): Adjust all callers.
5475 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5476 info, registers also at the lowest numbered registers side. Use
5477 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5481 * var-tracking.c (adjust_mems): Before adding a SET to
5482 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5484 2014-02-06 Alan Modra <amodra@gmail.com>
5487 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5488 change SDmode to DDmode when lra_in_progress.
5490 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5493 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5494 free_data_ref on the dr first, and before goto again also set dr
5495 to the next dr. For simd_lane_access, free old datarefs[i] before
5496 overwriting it. For get_vectype_for_scalar_type failure, don't
5497 free_data_ref if simd_lane_access.
5499 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5502 * tree.h (opts_for_fn): New inline function.
5503 (opt_for_fn): Define.
5504 * config/i386/i386.c (ix86_function_regparm): Use
5505 opt_for_fn (decl, optimize) instead of optimize.
5507 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5509 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5510 for SYMBOL_REF in large memory model.
5512 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5514 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5516 (cortex-a57): Likewise.
5517 (cortex-a57.cortex-a53): Likewise.
5519 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5520 Kugan Vivekanandarajah <kuganv@linaro.org>
5522 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5524 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5526 2014-02-06 Richard Biener <rguenther@suse.de>
5528 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5529 set_loop_copy and initialize_original_copy_tables.
5531 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5533 * config/aarch64/aarch64-simd.md
5534 (aarch64_ashr_simddi): Change QI to SI.
5536 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5537 Jakub Jelinek <jakub@redhat.com>
5540 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5543 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5545 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5546 CODE_FOR_altivec_vpku[hw]um to
5547 CODE_FOR_altivec_vpku[hw]um_direct.
5548 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5549 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5550 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5551 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5553 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5555 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5556 generation for -maltivec=be.
5557 (altivec_vsumsws): Simplify redundant test.
5559 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5561 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5562 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5563 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5564 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5565 gen_altivec_vpkuwum.
5566 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5568 (altivec_vpks<VI_char>ss): Likewise.
5569 (altivec_vpks<VI_char>us): Likewise.
5570 (altivec_vpku<VI_char>us): Likewise.
5571 (altivec_vpku<VI_char>um): Likewise.
5572 (altivec_vpku<VI_char>um_direct): New (copy of
5573 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5575 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5576 target is little endian and -maltivec=be is not specified.
5577 (*altivec_vupkhs<VU_char>_direct): New (copy of
5578 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
5579 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5580 target is little endian and -maltivec=be is not specified.
5581 (*altivec_vupkls<VU_char>_direct): New (copy of
5582 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
5583 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5584 little endian and -maltivec=be is not specified.
5585 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5586 little endian and -maltivec=be is not specified.
5588 2014-02-05 Richard Henderson <rth@redhat.com>
5591 * combine-stack-adj.c: Revert r206943.
5592 * sched-int.h (struct deps_desc): Add last_args_size.
5593 * sched-deps.c (init_deps): Initialize it.
5594 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5595 contain REG_ARGS_SIZE notes.
5597 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5599 * lto-cgraph.c (asm_nodes_output): Make global.
5600 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
5601 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
5602 (driver_handle_option): Handle OPT_fwpa.
5604 2014-02-05 Jakub Jelinek <jakub@redhat.com>
5607 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5608 a comment typo and formatting issue. If odr_hash hasn't been
5609 created, return vNULL and set *completep to false.
5612 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
5613 bb with no successors.
5615 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
5618 * doc/invoke.texi (-march): Clarify documentation for ARM.
5622 2014-02-05 Richard Biener <rguenther@suse.de>
5624 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
5625 when not vectorizing because of too many alias checks.
5626 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
5627 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
5629 2014-02-05 Nick Clifton <nickc@redhat.com>
5631 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
5632 accept extended registers in any mode when compiling for the MN10300.
5634 2014-02-05 Yury Gribov <y.gribov@samsung.com>
5636 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
5637 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
5638 sanitization attributes.
5639 (can_inline_edge_p): Likewise.
5640 (sanitize_attrs_match_for_inline_p): New function.
5642 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5644 * ipa-prop.c (detect_type_change): Shor circuit testing of
5645 type changes on THIS pointer.
5647 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
5650 * config/pa/pa.c (legitimize_tls_address): Return original address
5651 if not passed a SYMBOL_REF rtx.
5652 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
5654 (pa_emit_move_sequence): Simplify TLS source operands.
5655 (pa_legitimate_constant_p): Reject all TLS constants.
5656 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
5657 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
5659 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5661 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
5662 groups when we know they are controlled by LTO.
5663 * varasm.c (default_binds_local_p_1): If object is in other partition,
5664 it will be resolved locally.
5666 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5668 * config/host-linux.c (linux_gt_pch_use_address): Don't
5669 use SSIZE_MAX because it is not always defined.
5671 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
5674 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
5675 threshold for pseudo splitting.
5676 (update_ebb_live_info): Process call argument hard registers and
5677 hard registers from insn definition too.
5678 (max_small_class_regs_num): New constant.
5679 (inherit_in_ebb): Update live hard regs through EBBs. Update
5680 reloads_num only for small register classes. Don't split for
5683 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
5686 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
5689 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5691 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
5694 2014-02-04 Marek Polacek <polacek@redhat.com>
5696 * gdbinit.in (pel): Define.
5698 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5700 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
5703 2014-02-04 Richard Biener <rguenther@suse.de>
5706 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
5707 in function context local.
5708 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
5709 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
5710 similar to LTO_imported_decl_ref.
5712 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5714 PR tree-optimization/60002
5715 * cgraphclones.c (build_function_decl_skip_args): Clear
5718 PR tree-optimization/60023
5719 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
5720 false to gsi_replace.
5721 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
5722 has been in some EH region and vec_stmt could throw, add
5723 vec_stmt into the same EH region.
5724 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
5725 has no lhs, ignore it.
5726 * internal-fn.c (expand_MASK_LOAD): Likewise.
5729 * tree-inline.c (copy_forbidden): Fail for
5730 __attribute__((optimize (0))) functions.
5733 * omp-low.c (simd_clone_struct_copy): If from->inbranch
5734 is set, copy one less argument.
5735 (expand_simd_clones): Don't subtract clone_info->inbranch
5736 from simd_clone_struct_alloc argument.
5738 PR rtl-optimization/57915
5739 * recog.c (simplify_while_replacing): If all unary/binary/relational
5740 operation arguments are constant, attempt to simplify those.
5743 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
5744 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
5746 2014-02-04 Richard Biener <rguenther@suse.de>
5748 PR tree-optimization/60012
5749 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
5750 TBAA disambiguation to all DDRs.
5752 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5755 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
5756 (LINK_SPEC): Use it for -shared, -shared-libgcc.
5758 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5761 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
5763 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5765 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
5766 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
5768 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5771 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
5772 to figure out targets of polymorphic calls with known decl.
5773 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5774 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
5775 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
5776 (get_polymorphic_call_info): ... here.
5777 (get_polymorphic_call_info_from_invariant): New function.
5779 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5781 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
5782 lookup via vtable pointer; check for type consistency
5783 and turn inconsitent facts into UNREACHABLE.
5784 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5785 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
5786 type inconsistent querries; return UNREACHABLE instead.
5788 2014-02-03 Richard Henderson <rth@twiddle.net>
5791 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
5792 already processed this node.
5793 (normalize_one_pred_1): Pass along mark_set.
5794 (normalize_one_pred): Create and destroy a pointer_set_t.
5795 (normalize_one_pred_chain): Likewise.
5797 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
5799 PR gcov-profile/58602
5800 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
5802 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5805 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
5806 -fno-devirtualize; try to devirtualize by the knowledge of
5807 virtual table pointer given by aggregate propagation.
5808 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5809 (ipa_print_node_jump_functions): Dump also offset that
5810 is relevant for polymorphic calls.
5811 (determine_known_aggregate_parts): Add arg_type parameter; use it
5812 instead of determining the type from pointer type.
5813 (ipa_compute_jump_functions_for_edge): Update call of
5814 determine_known_aggregate_parts.
5815 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
5816 (gimple_get_virt_method_for_binfo): ... here; simplify using
5817 vtable_pointer_value_to_vtable.
5818 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
5819 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
5820 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
5821 (vtable_pointer_value_to_vtable): Break out from ...; handle also
5823 (vtable_pointer_value_to_binfo): ... here.
5824 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
5826 2014-02-03 Teresa Johnson <tejohnson@google.com>
5828 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
5829 redef of outer loop index variable.
5831 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
5835 * doc/extend.texi (Function Attributes): Typo.
5837 2014-02-03 Cong Hou <congh@google.com>
5839 PR tree-optimization/60000
5840 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
5841 if the vectorized statement is a store. A store statement can only
5842 appear at the end of pattern statements.
5844 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5846 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
5847 (ix86_option_override_internal): Default long double to 64-bit for
5848 32-bit Bionic and to 128-bit for 64-bit Bionic.
5850 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
5851 TARGET_LONG_DOUBLE_128 is true.
5852 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
5854 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
5855 (mlong-double-64): Negate -mlong-double-128.
5856 (mlong-double-128): New option.
5858 * config/i386/i386-c.c (ix86_target_macros): Define
5859 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
5861 * doc/invoke.texi: Document -mlong-double-128.
5863 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5865 PR rtl-optimization/60024
5866 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
5868 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
5870 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
5872 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
5874 PR rtl-optimization/57662
5875 * sel-sched.c (code_motion_path_driver): Do not mark already not
5876 existing blocks in the visiting bitmap.
5878 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
5880 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
5881 on the insn being emitted.
5883 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
5884 Will Deacon <will.deacon@arm.com>
5886 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
5888 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5890 * config/arm/arm-tables.opt: Regenerate.
5892 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5894 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
5895 for vector types other than V16QImode.
5896 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
5897 define_expand, and call altivec_expand_vec_perm_le when producing
5898 code with little endian element order.
5899 (*altivec_vperm_<mode>_internal): New insn having previous
5900 behavior of altivec_vperm_<mode>.
5901 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
5902 altivec_expand_vec_perm_le when producing code with little endian
5904 (*altivec_vperm_<mode>_uns_internal): New insn having previous
5905 behavior of altivec_vperm_<mode>_uns.
5907 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5909 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
5910 (altivec_vsumsws): Add handling for -maltivec=be with a little
5912 (altivec_vsumsws_direct): New.
5913 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
5914 gen_altivec_vsumsws.
5916 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
5918 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
5919 vtable_pointer_value_to_binfo): New functions.
5920 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
5921 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
5923 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
5925 * config/nios2/nios2.md (load_got_register): Initialize GOT
5926 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
5927 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
5929 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
5931 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
5932 preserverd by passthrough, do not propagate the type.
5934 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
5936 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
5937 (mips_atomic_assign_expand_fenv): New function.
5938 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
5940 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
5942 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
5943 (__builtin_mips_set_fcsr): Likewise.
5944 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
5945 MIPS_USI_FTYPE_VOID.
5946 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
5947 (mips16_expand_set_fcsr): Likewise.
5948 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
5949 (mips16_set_fcsr_stub): Likewise.
5950 (mips16_get_fcsr_one_only_stub): New class.
5951 (mips16_set_fcsr_one_only_stub): Likewise.
5952 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
5953 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
5954 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
5955 (hard_float): New availability predicate.
5956 (mips_builtins): Add get_fcsr and set_fcsr.
5957 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
5958 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
5959 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
5960 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
5961 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
5964 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
5966 * config/mips/mips.c (mips_one_only_stub): New class.
5967 (mips_need_mips16_rdhwr_p): Replace with...
5968 (mips16_rdhwr_stub): ...this new variable.
5969 (mips16_stub_call_address): New function.
5970 (mips16_rdhwr_one_only_stub): New class.
5971 (mips_expand_thread_pointer): Use mips16_stub_call_address.
5972 (mips_output_mips16_rdhwr): Delete.
5973 (mips_finish_stub): New function.
5974 (mips_code_end): Use it to handle rdhwr stubs.
5976 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
5979 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
5980 when calculating size of integer atomic types.
5982 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
5984 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
5986 2014-02-01 Jakub Jelinek <jakub@redhat.com>
5988 PR tree-optimization/60003
5989 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
5990 * profile.c (branch_prob): Use gimple_call_builtin_p
5991 to check for BUILT_IN_SETJMP_RECEIVER.
5992 * tree-inline.c (copy_bb): Call notice_special_calls.
5994 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
5997 * lra-constraints.c (process_alt_operands): Update reload_sum only
6000 2014-01-31 Richard Henderson <rth@redhat.com>
6003 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6004 until after else_eh is processed.
6006 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6008 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6009 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6010 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6011 in smmintrin.h, remove them.
6012 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6013 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6014 * config/i386/i386.md (ROUND_SAE): Fix value.
6015 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6016 (const48_operand): New.
6017 * config/i386/subst.md (round), (round_expand): Use
6018 const_4_or_8_to_11_operand.
6019 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6021 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6023 * config/i386/constraints.md (Yk): Swap meaning with k.
6024 * config/i386/i386.md (movhi_internal): Change Yk to k.
6025 (movqi_internal): Ditto.
6026 (*k<logic><mode>): Ditto.
6029 (kandn<mode>): Ditto.
6030 (*<code>hi_1): Ditto.
6031 (*<code>qi_1): Ditto.
6032 (kxnor<mode>): Ditto.
6033 (kortestzhi): Ditto.
6034 (kortestchi): Ditto.
6036 (*one_cmplhi2_1): Ditto.
6037 (*one_cmplqi2_1): Ditto.
6038 * config/i386/sse.md (): Change k to Yk.
6039 (avx512f_load<mode>_mask): Ditto.
6040 (avx512f_blendm<mode>): Ditto.
6041 (avx512f_store<mode>_mask): Ditto.
6042 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6043 (avx512f_storedqu<mode>_mask): Ditto.
6044 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6046 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6047 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6048 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6049 (avx512f_maskcmp<mode>3): Ditto.
6050 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6051 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6052 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6053 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6054 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6055 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6056 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6057 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6058 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6059 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6060 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6061 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6062 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6063 (vec_extract_lo_<mode>_maskm): Ditto.
6064 (vec_extract_hi_<mode>_maskm): Ditto.
6065 (avx512f_vternlog<mode>_mask): Ditto.
6066 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6067 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6068 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6069 (avx512f_<code>v8div16qi2_mask): Ditto.
6070 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6071 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6072 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6073 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6074 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6075 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6076 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6077 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6078 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6079 (avx512cd_maskb_vec_dupv8di): Ditto.
6080 (avx512cd_maskw_vec_dupv16si): Ditto.
6081 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6082 (avx512f_vpermi2var<mode>3_mask): Ditto.
6083 (avx512f_vpermi2var<mode>3_mask): Ditto.
6084 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6085 (*avx512f_gathersi<mode>): Ditto.
6086 (*avx512f_gathersi<mode>_2): Ditto.
6087 (*avx512f_gatherdi<mode>): Ditto.
6088 (*avx512f_gatherdi<mode>_2): Ditto.
6089 (*avx512f_scattersi<mode>): Ditto.
6090 (*avx512f_scatterdi<mode>): Ditto.
6091 (avx512f_compress<mode>_mask): Ditto.
6092 (avx512f_compressstore<mode>_mask): Ditto.
6093 (avx512f_expand<mode>_mask): Ditto.
6094 * config/i386/subst.md (mask): Change k to Yk.
6095 (mask_scalar_merge): Ditto.
6098 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6100 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6102 2014-01-31 Richard Biener <rguenther@suse.de>
6105 * builtins.c (fold_builtin_memory_op): Make sure to not
6106 use a floating-point mode or a boolean or enumeral type for
6109 2014-01-30 DJ Delorie <dj@redhat.com>
6111 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6112 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6113 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6114 whenever main() has an epilogue.
6116 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6118 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6119 unused variable "field".
6120 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6121 (vsx_mergeh_<mode>): Likewise.
6122 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6123 (altivec_vmrghh): Likewise.
6124 (altivec_vmrghw): Likewise.
6125 (altivec_vmrglb): Likewise.
6126 (altivec_vmrglh): Likewise.
6127 (altivec_vmrglw): Likewise.
6128 (altivec_vspltb): Add missing uses.
6129 (altivec_vsplth): Likewise.
6130 (altivec_vspltw): Likewise.
6131 (altivec_vspltsf): Likewise.
6133 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6136 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6137 frame related instructions.
6139 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6141 PR rtl-optimization/59959
6142 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6143 any reload of register whose subreg is invalid.
6145 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6147 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6148 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6149 Add missing return type - void.
6151 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6153 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6154 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6155 remove element index adjustment for endian (now handled in vsx.md
6157 (altivec_expand_vec_perm_const): Use
6158 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6159 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6160 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6161 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6162 define_expand and a new define_insn *altivec_vspltb_internal;
6163 adjust for -maltivec=be on a little endian target.
6164 (altivec_vspltb_direct): New.
6165 (altivec_vsplth): Divide into a define_expand and a new
6166 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6167 little endian target.
6168 (altivec_vsplth_direct): New.
6169 (altivec_vspltw): Divide into a define_expand and a new
6170 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6171 little endian target.
6172 (altivec_vspltw_direct): New.
6173 (altivec_vspltsf): Divide into a define_expand and a new
6174 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6175 a little endian target.
6177 2014-01-30 Richard Biener <rguenther@suse.de>
6179 PR tree-optimization/59993
6180 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6181 can propagate form the earlier stmt and avoid the transform
6182 when the intermediate result is needed.
6184 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6186 * README.Portability: Fix typo.
6188 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6190 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6191 comparison_operator with ordered_comparison_operator.
6193 2014-01-30 Nick Clifton <nickc@redhat.com>
6195 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6196 Rename to mn10300_store_multiple_regs.
6197 * config/mn10300/mn10300.c: Likewise.
6198 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6199 store_multiple_regs.
6200 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6201 Call mn10300_store_multiple_regs.
6203 2014-01-30 Nick Clifton <nickc@redhat.com>
6204 DJ Delorie <dj@redhat.com>
6206 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6207 %fp 2 to keep registers after it properly word-aligned.
6208 (rl78_alloc_physical_registers_umul): Handle the case where both
6209 input operands are the same.
6211 2014-01-30 Richard Biener <rguenther@suse.de>
6213 PR tree-optimization/59903
6214 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6217 2014-01-30 Jason Merrill <jason@redhat.com>
6220 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6223 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6225 2014-01-30 Richard Biener <rguenther@suse.de>
6227 PR tree-optimization/59951
6228 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6230 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6233 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6234 SFmode to DFmode case.
6236 2014-01-29 DJ Delorie <dj@redhat.com>
6238 * config/msp430/msp430.opt (-minrt): New.
6239 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6241 (ENDFILE_SPEC): Likewise.
6243 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6245 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6246 (estimate_function_body_sizes): Use it.
6248 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6251 * dwarf2out.c (is_cxx_auto): New.
6252 (is_base_type): Use it.
6253 (gen_type_die_with_usage): Likewise.
6255 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6257 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6258 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6259 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6260 -maltivec=be with LE targets.
6261 (vsx_mergeh_<mode>): Likewise.
6262 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6263 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6264 (altivec_vmrghb): Replace with define_expand and new
6265 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6266 (altivec_vmrghb_direct): New define_insn.
6267 (altivec_vmrghh): Replace with define_expand and new
6268 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6269 (altivec_vmrghh_direct): New define_insn.
6270 (altivec_vmrghw): Replace with define_expand and new
6271 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6272 (altivec_vmrghw_direct): New define_insn.
6273 (*altivec_vmrghsf): Adjust for endianness.
6274 (altivec_vmrglb): Replace with define_expand and new
6275 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6276 (altivec_vmrglb_direct): New define_insn.
6277 (altivec_vmrglh): Replace with define_expand and new
6278 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6279 (altivec_vmrglh_direct): New define_insn.
6280 (altivec_vmrglw): Replace with define_expand and new
6281 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6282 (altivec_vmrglw_direct): New define_insn.
6283 (*altivec_vmrglsf): Adjust for endianness.
6284 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6285 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6286 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6287 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6288 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6289 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6290 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6291 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6293 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6295 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6296 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6299 2014-01-29 Richard Biener <rguenther@suse.de>
6301 PR tree-optimization/58742
6302 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6303 associate_pointerplus_align.
6304 (associate_pointerplus_diff): New function.
6305 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6306 and associate_pointerplus_diff.
6308 2014-01-29 Richard Biener <rguenther@suse.de>
6310 * lto-streamer.h (LTO_major_version): Bump to 3.
6311 (LTO_minor_version): Reset to 0.
6313 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6315 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6316 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6317 (arm_file_start): Generate correct asm header for armv7ve.
6318 * config/arm/bpabi.h: Add multilib support for armv7ve.
6319 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6320 and cortex-a15 to armv7ve.
6321 * config/arm/t-aprofile: Add multilib support for armv7ve.
6322 * doc/invoke.texi: Document -march=armv7ve.
6324 2014-01-29 Richard Biener <rguenther@suse.de>
6326 PR tree-optimization/58742
6327 * tree-ssa-forwprop.c (associate_plusminus): Return true
6328 if we changed sth, defer EH cleanup to ...
6329 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6330 (simplify_mult): New function.
6332 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6335 PR tree-optimization/59920
6336 * tree.c (build_common_builtin_nodes): Remove
6337 __builtin_setjmp_dispatcher initialization.
6338 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6339 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6340 instead of gsi_after_labels + manually skipping debug stmts.
6341 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6342 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6343 * tree-inline.c (copy_edges_for_bb): Remove
6344 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6345 argument. Ignore computed_goto_p stmts. Don't call
6346 make_abnormal_goto_edges. If a call might need abnormal edges
6347 for non-local gotos, see if it already has an edge to
6348 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6349 with true argument, don't do anything then, otherwise add
6350 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6351 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6353 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6354 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6355 (lower_stmt): Don't set data->calls_builtin_setjmp.
6356 (lower_builtin_setjmp): Adjust comment.
6357 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6358 * tree-cfg.c (found_computed_goto): Remove.
6359 (factor_computed_gotos): Remove.
6360 (make_goto_expr_edges): Return bool, true for computed gotos.
6361 Don't call make_abnormal_goto_edges.
6362 (build_gimple_cfg): Don't set found_computed_goto, don't call
6363 factor_computed_gotos.
6364 (computed_goto_p): No longer static.
6365 (make_blocks): Don't set found_computed_goto.
6366 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6367 (make_edges): If make_goto_expr_edges returns true, push bb
6368 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6369 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6370 vector. Record mapping between bbs and OpenMP regions if there
6371 are any, adjust make_gimple_omp_edges caller. Call
6372 handle_abnormal_edges.
6373 (make_abnormal_goto_edges): Remove.
6374 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6375 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6376 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6377 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6378 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6379 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6380 filling *region also set *region_idx to (*region)->entry->index.
6383 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6384 For REGs set ORIGINAL_REGNO.
6386 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6388 * doc/md.texi: Mention that a target shouldn't implement
6389 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6392 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6394 PR tree-optimization/59594
6395 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6396 a copy of the datarefs vector rather than the vector itself.
6398 2014-01-28 Jason Merrill <jason@redhat.com>
6401 * dwarf2out.c (auto_die): New static.
6402 (gen_type_die_with_usage): Handle C++1y 'auto'.
6403 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6406 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6409 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6410 (SPEC_X32): Likewise.
6411 (SPEC_64): Likewise.
6412 * config/i386/i386.c (ix86_option_override_internal): Turn off
6413 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6415 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6416 * config/i386/i386.h (TARGET_16BIT): New macro.
6417 (TARGET_16BIT_P): Likewise.
6418 * config/i386/i386.opt: Add m16.
6419 * doc/invoke.texi: Document -m16.
6421 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6423 PR preprocessor/59935
6424 * input.c (location_get_source_line): Bail out on when line number
6425 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6427 2014-01-28 Richard Biener <rguenther@suse.de>
6429 PR tree-optimization/58742
6430 * tree-ssa-forwprop.c (associate_plusminus): Handle
6431 pointer subtraction of the form (T)(P + A) - (T)P.
6433 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6435 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6438 2014-01-28 Richard Biener <rguenther@suse.de>
6441 2014-01-28 Richard Biener <rguenther@suse.de>
6443 PR rtl-optimization/45364
6444 PR rtl-optimization/59890
6445 * var-tracking.c (local_get_addr_clear_given_value): Handle
6446 already cleared slot.
6447 (val_reset): Handle not allocated local_get_addr_cache.
6448 (vt_find_locations): Use post-order on the inverted CFG.
6450 2014-01-28 Richard Biener <rguenther@suse.de>
6452 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6454 2014-01-28 Richard Biener <rguenther@suse.de>
6456 PR rtl-optimization/45364
6457 PR rtl-optimization/59890
6458 * var-tracking.c (local_get_addr_clear_given_value): Handle
6459 already cleared slot.
6460 (val_reset): Handle not allocated local_get_addr_cache.
6461 (vt_find_locations): Use post-order on the inverted CFG.
6463 2014-01-28 Alan Modra <amodra@gmail.com>
6465 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6466 * configure.ac <recursive call for build != host>: Define
6467 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6468 and LD_FOR_BUILD too.
6469 * configure: Regenerate.
6471 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6473 * config/i386/i386.c (get_builtin_code_for_version): Separate
6474 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6475 Broadwell from Haswell.
6477 2014-01-27 Steve Ellcey <sellcey@mips.com>
6479 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6480 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6481 * config/mips/mips.c (mips_option_override): Change setting
6483 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6484 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6485 Change from Mask to Var.
6487 2014-01-27 Jeff Law <law@redhat.com>
6489 * ipa-inline.c (inline_small_functions): Fix typo.
6491 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6493 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6494 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6495 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6496 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6497 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6498 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6499 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6500 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6501 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6502 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6503 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6504 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6505 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6506 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6507 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6508 (_mm512_storeu_epi64): Ditto.
6509 (_mm512_cmpge_epi32_mask): Ditto.
6510 (_mm512_cmpge_epu32_mask): Ditto.
6511 (_mm512_cmpge_epi64_mask): Ditto.
6512 (_mm512_cmpge_epu64_mask): Ditto.
6513 (_mm512_cmple_epi32_mask): Ditto.
6514 (_mm512_cmple_epu32_mask): Ditto.
6515 (_mm512_cmple_epi64_mask): Ditto.
6516 (_mm512_cmple_epu64_mask): Ditto.
6517 (_mm512_cmplt_epi32_mask): Ditto.
6518 (_mm512_cmplt_epu32_mask): Ditto.
6519 (_mm512_cmplt_epi64_mask): Ditto.
6520 (_mm512_cmplt_epu64_mask): Ditto.
6521 (_mm512_cmpneq_epi32_mask): Ditto.
6522 (_mm512_cmpneq_epu32_mask): Ditto.
6523 (_mm512_cmpneq_epi64_mask): Ditto.
6524 (_mm512_cmpneq_epu64_mask): Ditto.
6525 (_mm512_expand_pd): Ditto.
6526 (_mm512_expand_ps): Ditto.
6527 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6528 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6529 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6530 * config/i386/i386.c (ix86_builtins): Add
6531 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6532 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6533 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6534 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6535 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6536 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6537 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6538 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6539 IX86_BUILTIN_PMOVUSQW512_MEM.
6540 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6541 __builtin_ia32_pmovsqd512mem_mask,
6542 __builtin_ia32_pmovqd512mem_mask,
6543 __builtin_ia32_pmovusqw512mem_mask,
6544 __builtin_ia32_pmovsqw512mem_mask,
6545 __builtin_ia32_pmovqw512mem_mask,
6546 __builtin_ia32_pmovusdw512mem_mask,
6547 __builtin_ia32_pmovsdw512mem_mask,
6548 __builtin_ia32_pmovdw512mem_mask,
6549 __builtin_ia32_pmovqb512mem_mask,
6550 __builtin_ia32_pmovusqb512mem_mask,
6551 __builtin_ia32_pmovsqb512mem_mask,
6552 __builtin_ia32_pmovusdb512mem_mask,
6553 __builtin_ia32_pmovsdb512mem_mask,
6554 __builtin_ia32_pmovdb512mem_mask.
6555 (bdesc_args): Add __builtin_ia32_expanddf512,
6556 __builtin_ia32_expandsf512.
6557 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6558 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6559 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6560 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6561 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6562 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6563 (avx512f_<code>v8div16qi2_mask_store): This.
6564 (avx512f_expand<mode>): New.
6566 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6568 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
6570 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6571 (_mm512_prefetch_i32scatter_pd): Ditto.
6572 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6573 (_mm512_prefetch_i64scatter_pd): Ditto.
6574 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6575 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6576 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6577 (_mm512_prefetch_i32scatter_ps): Ditto.
6578 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6579 (_mm512_prefetch_i64scatter_ps): Ditto.
6580 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6581 * config/i386/i386-builtin-types.def: Define
6582 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6583 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
6584 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6585 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6586 IX86_BUILTIN_SCATTERPFQPD.
6587 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6588 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6589 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6590 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6591 __builtin_ia32_scatterpfqps.
6592 (ix86_expand_builtin): Expand new built-ins.
6593 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6594 fix memory access data type.
6595 (*avx512pf_gatherpf<mode>_mask): Ditto.
6596 (*avx512pf_gatherpf<mode>): Ditto.
6597 (avx512pf_scatterpf<mode>): Ditto.
6598 (*avx512pf_scatterpf<mode>_mask): Ditto.
6599 (*avx512pf_scatterpf<mode>): Ditto.
6600 (GATHER_SCATTER_SF_MEM_MODE): New.
6601 (avx512pf_gatherpf<mode>df): Ditto.
6602 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6603 (*avx512pf_scatterpf<mode>df): Ditto.
6605 2014-01-27 Jakub Jelinek <jakub@redhat.com>
6608 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
6609 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
6612 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6614 * common/config/arm/arm-common.c
6615 (arm_rewrite_mcpu): Handle multiple names.
6617 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6619 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6621 * gimple-builder.h (create_gimple_tmp): Delete.
6623 2014-01-27 Christian Bruel <christian.bruel@st.com>
6625 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
6628 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
6630 * config/pa/pa.md (call): Generate indirect long calls to non-local
6631 functions when outputing 32-bit code.
6632 (call_value): Likewise except for special call to buggy powf function.
6634 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
6635 portable runtime and PIC indirect calls.
6636 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
6637 and PIC call sequences. Use ldo instead of blr to set return register
6638 in PIC call sequence.
6640 2014-01-25 Walter Lee <walt@tilera.com>
6642 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
6643 avoid clobbering a live register.
6645 2014-01-25 Walter Lee <walt@tilera.com>
6647 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
6648 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
6649 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
6650 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
6652 2014-01-25 Walter Lee <walt@tilera.com>
6654 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
6655 arguments on even registers.
6656 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
6658 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
6659 (BIGGEST_ALIGNMENT): Ditto.
6660 (BIGGEST_FIELD_ALIGNMENT): Ditto.
6662 2014-01-25 Walter Lee <walt@tilera.com>
6664 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
6665 insns before bundling.
6666 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
6668 2014-01-25 Walter Lee <walt@tilera.com>
6670 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
6671 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
6672 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
6674 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6676 * config/mips/constraints.md (kl): Delete.
6677 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
6678 define expands, using...
6679 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
6680 instructions for MIPS16.
6681 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
6682 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
6684 2014-01-25 Walter Lee <walt@tilera.com>
6686 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
6690 2014-01-25 Walter Lee <walt@tilera.com>
6692 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
6693 (TARGET_EXPAND_TO_RTL_HOOK): Define.
6695 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6697 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
6700 2014-01-25 Jakub Jelinek <jakub@redhat.com>
6702 * print-rtl.c (in_call_function_usage): New var.
6703 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
6704 EXPR_LIST mode as mode and not as reg note name.
6707 * cfgloopmanip.c (copy_loop_info): If
6708 loop->warned_aggressive_loop_optimizations, make sure
6709 the flag is set in target loop too.
6711 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
6713 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
6715 * builtins.def: Likewise.
6716 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
6717 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
6718 * ira.c (ira_setup_eliminable_regset): Likewise.
6719 * omp-low.c (gate_expand_omp): Likewise.
6720 (execute_lower_omp): Likewise.
6721 (diagnose_sb_0): Likewise.
6722 (gate_diagnose_omp_blocks): Likewise.
6723 (simd_clone_clauses_extract): Likewise.
6726 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6728 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
6729 correction for little endian...
6730 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
6733 2014-01-24 Jeff Law <law@redhat.com>
6735 PR tree-optimization/59919
6736 * tree-vrp.c (find_assert_locations_1): Do not register asserts
6737 for non-returning calls.
6739 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
6741 * common/config/aarch64/aarch64-common.c
6742 (aarch64_rewrite_mcpu): Handle multiple names.
6743 * config/aarch64/aarch64.h
6744 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6746 2014-01-24 Dodji Seketeli <dodji@redhat.com>
6748 * input.c (add_file_to_cache_tab): Handle the case where fopen
6751 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
6754 * config/i386/i386.md (pushsf splitter): Get stack adjustment
6755 from push operand if code of push isn't PRE_DEC.
6757 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6760 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
6761 -mquad-memory-atomic. Update -mquad-memory documentation to say
6762 it is only used for non-atomic loads/stores.
6764 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
6765 -mquad-memory or -mquad-memory-atomic switches.
6767 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
6768 -mquad-memory-atomic to ISA 2.07 support.
6770 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
6771 to separate support of normal quad word memory operations (ldq, stq)
6772 from the atomic quad word memory operations.
6774 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
6775 support to separate non-atomic quad word operations from atomic
6776 quad word operations. Disable non-atomic quad word operations in
6777 little endian mode so that we don't have to swap words after the
6778 load and before the store.
6779 (quad_load_store_p): Add comment about atomic quad word support.
6780 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
6781 options printed with -mdebug=reg.
6783 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
6784 -mquad-memory-atomic as the test for whether we have quad word
6785 atomic instructions.
6786 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
6787 or -mp8-vector are used, allow byte/half-word atomic operations.
6789 * config/rs6000/sync.md (load_lockedti): Insure that the address
6790 is a proper indexed or indirect address for the lqarx instruction.
6791 On little endian systems, swap the hi/lo registers after the lqarx
6793 (load_lockedpti): Use indexed_or_indirect_operand predicate to
6794 insure the address is valid for the lqarx instruction.
6795 (store_conditionalti): Insure that the address is a proper indexed
6796 or indirect address for the stqcrx. instruction. On little endian
6797 systems, swap the hi/lo registers before doing the stqcrx.
6799 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
6800 insure the address is valid for the stqcrx. instruction.
6802 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
6803 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
6804 type of quad memory support is available.
6806 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
6809 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
6810 there is a danger of looping.
6812 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
6814 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
6815 force flag_ira_loop_pressure if set via command line.
6817 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
6819 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
6820 (ashr_simd): New builtin handling DI mode.
6821 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
6822 (aarch64_sshr_simddi): New match pattern.
6823 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
6824 (vshrd_n_s64): Likewise.
6825 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
6827 2014-01-23 Nick Clifton <nickc@redhat.com>
6829 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
6830 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
6831 favour of mcu specific scripts.
6832 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
6835 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
6836 Alex Velenko <Alex.Velenko@arm.com>
6838 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
6839 (vaddv_s16): Likewise.
6840 (vaddv_s32): Likewise.
6841 (vaddv_u8): Likewise.
6842 (vaddv_u16): Likewise.
6843 (vaddv_u32): Likewise.
6844 (vaddvq_s8): Likewise.
6845 (vaddvq_s16): Likewise.
6846 (vaddvq_s32): Likewise.
6847 (vaddvq_s64): Likewise.
6848 (vaddvq_u8): Likewise.
6849 (vaddvq_u16): Likewise.
6850 (vaddvq_u32): Likewise.
6851 (vaddvq_u64): Likewise.
6852 (vaddv_f32): Likewise.
6853 (vaddvq_f32): Likewise.
6854 (vaddvq_f64): Likewise.
6855 (vmaxv_f32): Likewise.
6856 (vmaxv_s8): Likewise.
6857 (vmaxv_s16): Likewise.
6858 (vmaxv_s32): Likewise.
6859 (vmaxv_u8): Likewise.
6860 (vmaxv_u16): Likewise.
6861 (vmaxv_u32): Likewise.
6862 (vmaxvq_f32): Likewise.
6863 (vmaxvq_f64): Likewise.
6864 (vmaxvq_s8): Likewise.
6865 (vmaxvq_s16): Likewise.
6866 (vmaxvq_s32): Likewise.
6867 (vmaxvq_u8): Likewise.
6868 (vmaxvq_u16): Likewise.
6869 (vmaxvq_u32): Likewise.
6870 (vmaxnmv_f32): Likewise.
6871 (vmaxnmvq_f32): Likewise.
6872 (vmaxnmvq_f64): Likewise.
6873 (vminv_f32): Likewise.
6874 (vminv_s8): Likewise.
6875 (vminv_s16): Likewise.
6876 (vminv_s32): Likewise.
6877 (vminv_u8): Likewise.
6878 (vminv_u16): Likewise.
6879 (vminv_u32): Likewise.
6880 (vminvq_f32): Likewise.
6881 (vminvq_f64): Likewise.
6882 (vminvq_s8): Likewise.
6883 (vminvq_s16): Likewise.
6884 (vminvq_s32): Likewise.
6885 (vminvq_u8): Likewise.
6886 (vminvq_u16): Likewise.
6887 (vminvq_u32): Likewise.
6888 (vminnmv_f32): Likewise.
6889 (vminnmvq_f32): Likewise.
6890 (vminnmvq_f64): Likewise.
6892 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
6894 * config/aarch64/aarch64-simd.md
6895 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
6896 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
6897 (*aarch64_mul3_elt<mode>): Likewise.
6898 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
6899 (*aarch64_mul3_elt_to_64v2df): Likewise.
6900 (*aarch64_mla_elt<mode>): Likewise.
6901 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
6902 (*aarch64_mls_elt<mode>): Likewise.
6903 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
6904 (*aarch64_fma4_elt<mode>): Likewise.
6905 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
6906 (*aarch64_fma4_elt_to_64v2df): Likewise.
6907 (*aarch64_fnma4_elt<mode>): Likewise.
6908 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
6909 (*aarch64_fnma4_elt_to_64v2df): Likewise.
6910 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
6911 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
6912 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6913 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6914 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
6915 (aarch64_sqdmull_lane<mode>_internal): Likewise.
6916 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
6918 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
6920 * config/aarch64/aarch64-simd.md
6921 (aarch64_be_checked_get_lane<mode>): New define_expand.
6922 * config/aarch64/aarch64-simd-builtins.def
6923 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
6924 New builtin definition.
6925 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
6926 Use new safe be builtin.
6928 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
6930 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
6932 (aarch64_be_st1<mode>): Likewise.
6933 (aarch_ld1<VALL:mode>): Define_expand modified.
6934 (aarch_st1<VALL:mode>): Likewise.
6935 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
6936 (UNSPEC_ST1): Likewise.
6938 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
6940 * config/microblaze/microblaze.md: Add trap insn and attribute
6942 2014-01-23 Dodji Seketeli <dodji@redhat.com>
6944 PR preprocessor/58580
6945 * input.h (location_get_source_line): Take an additional line_size
6947 (void diagnostics_file_cache_fini): Declare new function.
6948 * input.c (struct fcache): New type.
6949 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
6950 New static constants.
6951 (diagnostic_file_cache_init, total_lines_num)
6952 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
6953 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
6954 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
6955 (get_next_line, read_next_line, goto_next_line, read_line_num):
6956 New static function definitions.
6957 (diagnostic_file_cache_fini): New function.
6958 (location_get_source_line): Take an additional output line_len
6959 parameter. Re-write using lookup_or_add_file_to_cache_tab and
6961 * diagnostic.c (diagnostic_finish): Call
6962 diagnostic_file_cache_fini.
6963 (adjust_line): Take an additional input parameter for the length
6964 of the line, rather than calculating it with strlen.
6965 (diagnostic_show_locus): Adjust the use of
6966 location_get_source_line and adjust_line with respect to their new
6967 signature. While displaying a line now, do not stop at the first
6968 null byte. Rather, display the zero byte as a space and keep
6969 going until we reach the size of the line.
6970 * Makefile.in: Add vec.o to OBJS-libcommon
6972 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
6973 Ilya Tocar <ilya.tocar@intel.com>
6975 * config/i386/avx512fintrin.h (_mm512_kmov): New.
6976 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
6977 (__builtin_ia32_kmov16): Ditto.
6978 * config/i386/i386.md (UNSPEC_KMOV): New.
6981 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
6983 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
6984 (_mm512_storeu_si512): Ditto.
6986 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
6989 * rtl.h (get_referenced_operands): Declare.
6990 * recog.c (get_referenced_operands): New function.
6991 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
6992 operands have been referenced when recording LO_SUM references.
6994 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
6996 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
6998 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7000 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7001 Enable for generic and recent AMD targets.
7003 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7005 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7006 ARG_SIZE note when adjustment was eliminated.
7008 2014-01-22 Jeff Law <law@redhat.com>
7010 PR tree-optimization/59597
7011 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7012 in file. Accept new argument REGISTERING and use it to modify
7013 dump output appropriately.
7014 (register_jump_thread): Corresponding changes.
7015 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7016 thread paths involving joiner blocks. Add code to dump cancelled
7017 jump threading paths.
7019 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7021 PR rtl-optimization/59477
7022 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7023 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7025 2014-01-22 Tom Tromey <tromey@redhat.com>
7027 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7029 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7031 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7033 PR rtl-optimization/59896
7034 * lra-constraints.c (process_alt_operands): Check unused note for
7035 matched operands of insn with no output reloads.
7037 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7039 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7040 (mips_move_from_gpr_cost): Likewise.
7042 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7044 PR rtl-optimization/59858
7045 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7046 ira_class_hard_regs_num.
7047 (process_alt_operands): Increase reject for dying matched operand.
7049 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7052 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7053 smaller than size, perform several stores or loads and stores
7054 at dst + count - size to store or copy all of size bytes, rather
7055 than just last modesize bytes.
7057 2014-01-20 DJ Delorie <dj@redhat.com>
7059 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7060 that CLOBBERs are REGs before propogating their values.
7062 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7065 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7066 (cgraph_inline_failed_type): New function.
7067 * cgraph.h (DEFCIFCODE): Add type.
7068 (cgraph_inline_failed_type_t): New enum.
7069 (cgraph_inline_failed_type): New prototype.
7070 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7071 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7072 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7073 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7074 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7075 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7076 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7077 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7078 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7079 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7080 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7081 OPTIMIZATION_MISMATCH.
7082 * tree-inline.c (expand_call_inline): Emit errors during
7083 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7085 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7088 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7089 mode attribute in insn output.
7091 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7093 * output.h (output_constant): Delete.
7094 * varasm.c (output_constant): Make private.
7096 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7098 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7100 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7103 * tree.h (fold_builtin_strcat): New prototype.
7104 * builtins.c (fold_builtin_strcat): No longer static. Add len
7105 argument, if non-NULL, don't call c_strlen. Optimize
7106 directly into __builtin_memcpy instead of __builtin_strcpy.
7107 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7108 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7110 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7112 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7113 for SImode_address_operand operands, having only a REG argument.
7115 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7117 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7118 loader name using mbig-endian.
7119 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7121 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7123 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7127 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7129 * config/aarch64/aarch64-protos.h
7130 (aarch64_cannot_change_mode_class_ptr): Declare.
7131 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7132 aarch64_cannot_change_mode_class_ptr): New.
7133 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7134 backend hook aarch64_cannot_change_mode_class.
7136 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7138 * common/config/aarch64/aarch64-common.c
7139 (aarch64_handle_option): Don't handle any option order logic here.
7140 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7141 selected_cpu, warn on architecture version mismatch.
7142 (aarch64_override_options): Fix parsing order for option strings.
7144 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7145 Iain Sandoe <iain@codesourcery.com>
7148 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7149 warning. Amend comment to reflect current functionality.
7151 2014-01-20 Richard Biener <rguenther@suse.de>
7154 * builtins.c (fold_builtin_strcat): Remove case better handled
7155 by tree-ssa-strlen.c.
7157 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7159 * config/aarch64/aarch64.opt
7160 (mcpu, march, mtune): Make case-insensitive.
7162 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7165 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7166 if operands[1] is a REG or ZERO_EXTEND of a REG.
7168 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7170 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7172 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7174 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7175 long non-pic millicode calls.
7177 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7179 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7181 2014-01-19 Kito Cheng <kito@0xlab.org>
7183 * builtins.c (expand_movstr): Check movstr expand done or fail.
7185 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7186 H.J. Lu <hongjiu.lu@intel.com>
7189 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7190 to DImode for zero-extended addresses.
7192 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7194 PR rtl-optimization/57763
7195 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7196 on the new indirect jump_insn and increment LABEL_NUSES (label).
7198 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7202 * config.gcc (x86_archs): New variable.
7203 (x86_64_archs): Likewise.
7204 (x86_cpus): Likewise.
7205 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7206 --with-arch/--with-cpu= options.
7207 Support --with-arch=/--with-cpu={nehalem,westmere,
7208 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7210 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7212 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7213 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7215 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7217 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7219 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7222 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7223 clear cpp_get_options (parse_in)->warn_unused_macros for
7224 ix86_target_macros_internal with cpp_define.
7226 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7228 * jump.c (delete_related_insns): Keep (use (insn))s.
7229 * reorg.c (redundant_insn): Check for barriers too.
7231 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7233 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7235 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7237 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7238 call to $$dyncall when TARGET_LONG_CALLS is true.
7240 2014-01-17 Jeff Law <law@redhat.com>
7242 * ree.c (combine_set_extension): Temporarily disable test for
7243 changing number of hard registers.
7245 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7248 * ipa-inline-analysis.c (inline_free_summary):
7249 Do not free summary of aliases.
7251 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7254 * gimplify.c (gimplify_expr): Use create_tmp_var
7255 instead of create_tmp_var_raw. If cond doesn't have
7256 integral type, don't add the IFN_ANNOTATE builtin at all.
7258 2014-01-17 Martin Jambor <mjambor@suse.cz>
7261 * ipa-cp.c (prev_edge_clone): New variable.
7262 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7263 Also resize prev_edge_clone vector.
7264 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7265 (ipcp_edge_removal_hook): New function.
7266 (ipcp_driver): Register ipcp_edge_removal_hook.
7268 2014-01-17 Andrew Pinski <apinski@cavium.com>
7269 Steve Ellcey <sellcey@mips.com>
7272 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7275 2014-01-17 Jeff Law <law@redhat.com>
7278 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7279 so that pass_ccp runs first.
7281 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7283 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7284 (ix86_adjust_cost): Use !TARGET_XXX.
7285 (do_reorder_for_imul): Likewise.
7286 (swap_top_of_ready_list): Likewise.
7287 (ix86_sched_reorder): Likewise.
7289 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7291 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7292 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7293 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7294 (intel_memset): New. Duplicate slm_memset.
7295 (intel_cost): New. Duplicate slm_cost.
7296 (m_INTEL): New macro.
7297 (processor_target_table): Add "intel".
7298 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7299 with PROCESSOR_INTEL for "intel".
7300 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7301 PROCESSOR_SILVERMONT.
7302 (ix86_issue_rate): Likewise.
7303 (ix86_adjust_cost): Likewise.
7304 (ia32_multipass_dfa_lookahead): Likewise.
7305 (swap_top_of_ready_list): Likewise.
7306 (ix86_sched_reorder): Likewise.
7307 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7308 instead of TARGET_OPT_AGU.
7309 * config/i386/i386.h (TARGET_INTEL): New.
7310 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7311 (processor_type): Add PROCESSOR_INTEL.
7312 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7313 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7315 2014-01-17 Marek Polacek <polacek@redhat.com>
7318 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7321 2014-01-17 Richard Biener <rguenther@suse.de>
7323 PR tree-optimization/46590
7324 * opts.c (default_options_table): Add entries for
7325 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7326 all enabled at -O1 but not for -Og.
7327 * common.opt (fbranch-count-reg): Remove Init(1).
7328 (fmove-loop-invariants): Likewise.
7329 (ftree-pta): Likewise.
7331 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7333 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7334 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7335 decls to at least the GCC 4.8 used alignments.
7338 * tree-nested.c (convert_nonlocal_reference_stmt,
7339 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7340 of GIMPLE_BIND stmts, adjust associated decls.
7342 2014-01-17 Richard Biener <rguenther@suse.de>
7344 PR tree-optimization/46590
7345 * vec.h (vec<>::bseach): New member function implementing
7346 binary search according to C89 bsearch.
7347 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7348 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7349 bitmap pointer again. Make accesses_in_loop a flat array.
7350 (mem_ref_obstack): New global.
7351 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7352 (mark_ref_stored): Likewise.
7353 (ref_indep_loop_p_2): Likewise.
7354 (set_ref_stored_in_loop): New helper function.
7355 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7356 (memref_free): Adjust.
7357 (record_mem_ref_loc): Simplify.
7358 (gather_mem_refs_stmt): Adjust.
7359 (sort_locs_in_loop_postorder_cmp): New function.
7360 (analyze_memory_references): Sort accesses_in_loop after
7361 loop postorder number.
7362 (find_ref_loc_in_loop_cmp): New function.
7363 (for_all_locs_in_loop): Find relevant cluster of locs in
7364 accesses_in_loop and iterate without recursion.
7365 (execute_sm): Avoid uninit warning.
7366 (struct ref_always_accessed): Simplify.
7367 (ref_always_accessed::operator ()): Likewise.
7368 (ref_always_accessed_p): Likewise.
7369 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7370 loop postorder numbers here.
7371 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7374 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7377 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7378 on decls for which assemble_alias has been called.
7380 2014-01-17 Nick Clifton <nickc@redhat.com>
7382 * config/msp430/msp430.opt: (mcpu): New option.
7383 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7384 (msp430_option_override): Parse target_cpu. If the MCU name
7385 matches a generic string, clear target_mcu.
7386 (msp430_attr): Allow numeric interrupt values up to 63.
7387 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7388 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7390 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7392 * config/msp430/msp430.md (popm): Use %J rather than %I.
7393 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7394 (addhi_cy_i): Use immediate_operand for operand 2.
7395 * doc/invoke.texi: Document -mcpu option.
7397 2014-01-17 Richard Biener <rguenther@suse.de>
7399 PR rtl-optimization/38518
7400 * df.h (df_analyze_loop): Declare.
7401 * df-core.c: Include cfgloop.h.
7402 (df_analyze_1): Split out main part of df_analyze.
7403 (df_analyze): Adjust.
7404 (loop_inverted_post_order_compute): New function.
7405 (loop_post_order_compute): Likewise.
7406 (df_analyze_loop): New function avoiding whole-function
7408 * loop-invariant.c (find_defs): Use df_analyze_loop.
7409 (find_invariants): Adjust.
7410 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7412 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7414 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7415 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7417 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7419 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7420 traversal when removing references.
7422 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7425 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7427 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7430 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7431 pushing a reload for an autoinc when we had previously reloaded an
7432 inner part of the address.
7434 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7436 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7438 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7439 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7440 when not giving up or versioning for alias only because of
7442 (vect_analyze_data_ref_dependences): Set to true.
7443 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7445 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7446 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7450 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7453 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7454 operand 0 predicate for gathers, use a new pseudo as subtarget.
7456 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7459 * lra-constraints.c (process_alt_operands): Add printing debug info.
7460 Check absence of input/output reloads for matched operands too.
7462 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7464 PR rtl-optimization/59835
7465 * ira.c (ira_init_register_move_cost): Increase cost for
7468 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7470 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7472 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7475 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7476 non-register objects. Use gen_(high/low)part more consistently.
7479 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7482 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7483 endian support, remove tests for WORDS_BIG_ENDIAN.
7484 (p8_mfvsrd_3_<mode>): Likewise.
7485 (reload_gpr_from_vsx<mode>): Likewise.
7486 (reload_gpr_from_vsxsf): Likewise.
7487 (p8_mfvsrd_4_disf): Likewise.
7489 2014-01-16 Richard Biener <rguenther@suse.de>
7491 PR rtl-optimization/46590
7492 * lcm.c (compute_antinout_edge): Use postorder iteration.
7493 (compute_laterin): Use inverted postorder iteration.
7495 2014-01-16 Nick Clifton <nickc@redhat.com>
7498 * varasm.c (output_constant): Return the number of bytes actually
7500 (output_constructor_array_range): Update the field size with the
7501 number of bytes emitted by output_constant.
7502 (output_constructor_regular_field): Likewise. Also do not
7503 complain if the total number of bytes emitted is now greater
7504 than the expected fieldpos.
7505 * output.h (output_constant): Update prototype and descriptive comment.
7507 2014-01-16 Marek Polacek <polacek@redhat.com>
7510 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7511 it is error_mark_node.
7513 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7515 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7516 VALID_AVX256_REG_OR_OI_MODE.
7518 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7520 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7521 current procedure should be profiled.
7523 2014-01-15 Andrew Pinski <apinski@cavium.com>
7525 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7526 of moving from/to the STACK_REG register class.
7528 2014-01-15 Richard Henderson <rth@redhat.com>
7531 * reginfo.c (global_regs_decl): Globalize.
7532 * rtl.h (global_regs_decl): Declare.
7533 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7534 reserved via global_regs.
7536 2014-01-15 Teresa Johnson <tejohnson@google.com>
7538 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7540 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7542 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7543 and vmulosh rather than call gen_vec_widen_smult_*.
7544 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7545 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7546 (vec_widen_smult_even_v16qi): Likewise.
7547 (vec_widen_umult_even_v8hi): Likewise.
7548 (vec_widen_smult_even_v8hi): Likewise.
7549 (vec_widen_umult_odd_v16qi): Likewise.
7550 (vec_widen_smult_odd_v16qi): Likewise.
7551 (vec_widen_umult_odd_v8hi): Likewise.
7552 (vec_widen_smult_odd_v8hi): Likewise.
7553 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7554 vmuloub rather than call gen_vec_widen_umult_*.
7555 (vec_widen_umult_lo_v16qi): Likewise.
7556 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7557 vmulosb rather than call gen_vec_widen_smult_*.
7558 (vec_widen_smult_lo_v16qi): Likewise.
7559 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7560 rather than call gen_vec_widen_umult_*.
7561 (vec_widen_umult_lo_v8hi): Likewise.
7562 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7563 rather than call gen_vec_widen_smult_*.
7564 (vec_widen_smult_lo_v8hi): Likewise.
7566 2014-01-15 Jeff Law <law@redhat.com>
7568 PR tree-optimization/59747
7569 * ree.c (find_and_remove_re): Properly handle case where a second
7570 eliminated extension requires widening a copy created for elimination
7571 of a prior extension.
7572 (combine_set_extension): Ensure that the number of hard regs needed
7573 for a destination register does not change when we widen it.
7575 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7577 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7578 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7579 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7580 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7581 (avr-*-rtems*): Likewise.
7582 (bfin*-rtems*): Likewise.
7583 (moxie-*-rtems*): Likewise.
7584 (h8300-*-rtems*): Likewise.
7585 (i[34567]86-*-rtems*): Likewise.
7586 (lm32-*-rtems*): Likewise.
7587 (m32r-*-rtems*): Likewise.
7588 (m68k-*-rtems*): Likewise.
7589 (microblaze*-*-rtems*): Likewise.
7590 (mips*-*-rtems*): Likewise.
7591 (powerpc-*-rtems*): Likewise.
7592 (sh-*-rtems*): Likewise.
7593 (sparc-*-rtems*): Likewise.
7594 (sparc64-*-rtems*): Likewise.
7595 (v850-*-rtems*): Likewise.
7596 (m32c-*-rtems*): Likewise.
7598 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7600 PR rtl-optimization/59511
7601 * ira.c (ira_init_register_move_cost): Use memory costs for some
7602 cases of register move cost calculations.
7603 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7604 instead of BB frequency.
7605 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7606 * lra-assigns.c (find_hard_regno_for): Ditto.
7608 2014-01-15 Richard Biener <rguenther@suse.de>
7610 PR tree-optimization/59822
7611 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
7612 (vectorizable_load): Use it to hoist defs of uses of invariant
7613 loads out of the loop.
7615 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7616 Kugan Vivekanandarajah <kuganv@linaro.org>
7619 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
7622 2014-01-15 Richard Biener <rguenther@suse.de>
7624 PR rtl-optimization/59802
7625 * lcm.c (compute_available): Use inverted postorder to seed
7626 the initial worklist.
7628 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7631 * config/s390/s390.c (s390_preferred_reload_class): Don't return
7632 ADDR_REGS for invalid symrefs in non-PIC code.
7634 2014-01-15 Jakub Jelinek <jakub@redhat.com>
7637 * builtins.c (determine_block_size): Initialize *probable_max_size
7638 even if len_rtx is CONST_INT.
7640 2014-01-14 Andrew Pinski <apinski@cavium.com>
7642 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
7643 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
7644 (cortexa53_tunings): Likewise.
7645 (aarch64_sched_issue_rate): New function.
7646 (TARGET_SCHED_ISSUE_RATE): Define.
7648 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7650 * ira-costs.c (find_costs_and_classes): Add missed
7651 ira_init_register_move_cost_if_necessary.
7653 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7656 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
7658 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
7661 * config/i386/i386.c (type_natural_mode): Add a bool parameter
7662 to indicate if type is used for function return value. Warn ABI
7663 change if the vector mode isn't available for function return value.
7664 (ix86_function_arg_advance): Pass false to type_natural_mode.
7665 (ix86_function_arg): Likewise.
7666 (ix86_gimplify_va_arg): Likewise.
7667 (function_arg_32): Don't warn ABI change.
7668 (ix86_function_value): Pass true to type_natural_mode.
7669 (ix86_return_in_memory): Likewise.
7670 (ix86_struct_value_rtx): Removed.
7671 (TARGET_STRUCT_VALUE_RTX): Likewise.
7673 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7675 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
7676 converting a conditional jump into a conditional return.
7678 2014-01-14 Richard Biener <rguenther@suse.de>
7680 PR tree-optimization/58921
7681 PR tree-optimization/59006
7682 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
7683 hoisting invariant stmts.
7684 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
7685 invariant loads on the preheader edge if possible.
7687 2014-01-14 Joey Ye <joey.ye@arm.com>
7689 * doc/plugin.texi (Building GCC plugins): Update to C++.
7691 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
7693 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
7694 (_mm_rcp28_round_ss): Ditto.
7695 (_mm_rsqrt28_round_sd): Ditto.
7696 (_mm_rsqrt28_round_ss): Ditto.
7697 (_mm_rcp28_sd): Ditto.
7698 (_mm_rcp28_ss): Ditto.
7699 (_mm_rsqrt28_sd): Ditto.
7700 (_mm_rsqrt28_ss): Ditto.
7701 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
7702 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
7703 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
7704 (IX86_BUILTIN_RCP28SD): Ditto.
7705 (IX86_BUILTIN_RCP28SS): Ditto.
7706 (IX86_BUILTIN_RSQRT28SD): Ditto.
7707 (IX86_BUILTIN_RSQRT28SS): Ditto.
7708 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
7709 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
7710 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
7711 (ix86_expand_special_args_builtin): Expand new FTYPE.
7712 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
7713 (srcp14<mode>): Make insn unary.
7714 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
7715 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
7716 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
7717 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
7718 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
7719 Fix rounding: make it SAE only.
7720 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7722 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7724 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
7725 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
7726 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
7727 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
7728 (round_saeonly_mask_scalar_operand4): Ditto.
7729 (round_saeonly_mask_scalar_op3): Ditto.
7730 (round_saeonly_mask_scalar_op4): Ditto.
7732 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7734 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7735 Implement -maltivec=be for vec_insert and vec_extract.
7737 2014-01-10 DJ Delorie <dj@redhat.com>
7739 * config/msp430/msp430.md (call_internal): Don't allow memory
7740 references with SP as the base register.
7741 (call_value_internal): Likewise.
7742 * config/msp430/constraints.md (Yc): New. For memory references
7743 that don't use SP as a base register.
7745 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
7746 "an integer without a # prefix"
7747 * config/msp430/msp430.md (epilogue_helper): Use it.
7749 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7752 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
7753 AVX512F gather builtins.
7754 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
7755 on gather decls with INTEGER_TYPE masktype.
7756 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
7757 directly into the builtin rather than hoisting it before loop.
7759 PR tree-optimization/59387
7760 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
7761 (scev_const_prop): If folded_casts and type has undefined overflow,
7762 use force_gimple_operand instead of force_gimple_operand_gsi and
7763 for each added stmt if it is assign with
7764 arith_code_with_undefined_signed_overflow, call
7765 rewrite_to_defined_overflow.
7766 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
7767 gimple-fold.h instead.
7768 (arith_code_with_undefined_signed_overflow,
7769 rewrite_to_defined_overflow): Moved to ...
7770 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
7771 rewrite_to_defined_overflow): ... here. No longer static.
7772 Include gimplify-me.h.
7773 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
7774 rewrite_to_defined_overflow): New prototypes.
7776 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7778 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
7780 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
7782 * builtins.c (get_object_alignment_2): Minor tweak.
7783 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
7785 2014-01-13 Christian Bruel <christian.bruel@st.com>
7787 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
7788 optimized non constant lengths.
7790 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7793 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
7794 load as __atomic_load_N if possible.
7796 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7798 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
7800 (rs6000_expand_builtin): Adjust call.
7802 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7805 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
7806 * config/rs6000/rs6000.c: Include target-globals.h.
7807 (rs6000_set_current_function): Instead of doing target_reinit
7808 unconditionally, use save_target_globals_default_opts and
7809 restore_target_globals.
7811 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
7813 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
7814 (rs6000_expand_builtin): Handle mffs and mtfsf.
7815 (rs6000_init_builtins): Define mffs and mtfsf.
7816 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
7817 (rs6000_mffs): New pattern.
7818 (rs6000_mtfsf): New pattern.
7820 2014-01-11 Bin Cheng <bin.cheng@arm.com>
7822 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
7823 Start narrowing with START. Apply candidate-use pair
7824 and check overall cost in narrowing.
7825 (iv_ca_prune): Pass new argument.
7827 2014-01-10 Jeff Law <law@redhat.com>
7830 * ree.c (combine_reaching_defs): Ensure the defining statement
7831 occurs before the extension when optimizing extensions with
7832 different source and destination hard registers.
7834 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
7837 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
7838 vtables into the type inheritance graph.
7840 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7842 PR rtl-optimization/59754
7843 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
7844 modes in the REGNO != REGNO case.
7846 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7848 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
7850 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7852 PR tree-optimization/59745
7853 * tree-predcom.c (tree_predictive_commoning_loop): Call
7854 free_affine_expand_cache if giving up because components is NULL.
7856 * target-globals.c (save_target_globals): Allocate < 4KB structs using
7857 GC in payload of target_globals struct instead of allocating them on
7858 the heap and the larger structs separately using GC.
7859 * target-globals.h (struct target_globals): Make regs, hard_regs,
7860 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
7861 of GTY((skip)) and change type to void *.
7862 (reset_target_globals): Cast loads from those fields to corresponding
7865 2014-01-10 Steve Ellcey <sellcey@mips.com>
7868 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
7869 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
7870 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
7872 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
7875 * aarch64-modes.def (CC_Zmode): New flags mode.
7876 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
7877 represents an equality.
7878 (aarch64_get_condition_code): Handle CC_Zmode.
7879 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
7881 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7883 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
7884 extraction in good case.
7886 2014-01-10 Richard Biener <rguenther@suse.de>
7888 PR tree-optimization/59374
7889 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
7890 checking after SLP discovery. Mark stmts not participating
7891 in any SLP instance properly.
7893 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7895 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
7896 when handling a SET rtx.
7898 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7900 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
7901 (cortex-a57): Likewise.
7902 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
7904 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7906 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
7907 non-iwmmxt builtins.
7909 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
7913 * ipa-devirt.c record_target_from_binfo): Take as argument
7914 stack of binfos and lookup matching one for virtual inheritance.
7915 (possible_polymorphic_call_targets_1): Update.
7917 2014-01-10 Huacai Chen <chenhc@lemote.com>
7919 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
7920 kernel strings for Loongson-2E/2F/3A.
7922 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7925 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
7926 is_gimple_call before calling gimple_call_internal_p.
7928 2014-01-09 Steve Ellcey <sellcey@mips.com>
7930 * Makefile.in (TREE_FLOW_H): Remove.
7931 (TREE_SSA_H): Add file names from tree-flow.h.
7932 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
7933 * tree.h: Remove tree-flow.h reference.
7934 * hash-table.h: Remove tree-flow.h reference.
7935 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
7936 reference with tree-ssa-loop.h.
7938 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7940 * doc/invoke.texi: Add -maltivec={be,le} options, and document
7941 default element-order behavior for -maltivec.
7942 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
7943 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
7944 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
7945 when targeting big endian, at least for now.
7946 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
7948 2014-01-09 Jakub Jelinek <jakub@redhat.com>
7951 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
7952 var satisfies use_register_for_decl, just take into account type
7953 alignment, rather than decl alignment.
7955 PR tree-optimization/59622
7956 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
7957 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
7958 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
7959 Don't devirtualize for inplace at all. For targets.length () == 1,
7960 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
7962 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
7964 * config/i386/i386.md (cpu): Remove the unused btver1.
7966 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
7968 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
7970 2014-01-09 Jakub Jelinek <jakub@redhat.com>
7973 * tree-core.h (struct target_globals): New forward declaration.
7974 (struct tree_target_option): Add globals field.
7975 * tree.h (TREE_TARGET_GLOBALS): Define.
7976 (prepare_target_option_nodes_for_pch): New prototype.
7977 * target-globals.h (struct target_globals): Define even if
7979 * tree.c (prepare_target_option_node_for_pch,
7980 prepare_target_option_nodes_for_pch): New functions.
7981 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
7982 * config/i386/i386.c: Include target-globals.h.
7983 (ix86_set_current_function): Instead of doing target_reinit
7984 unconditionally, use save_target_globals_default_opts and
7985 restore_target_globals.
7987 2014-01-09 Richard Biener <rguenther@suse.de>
7989 PR tree-optimization/59715
7990 * tree-cfg.h (split_critical_edges): Declare.
7991 * tree-cfg.c (split_critical_edges): Export.
7992 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
7994 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
7996 * cfgexpand.c (expand_stack_vars): Optionally disable
7997 asan stack protection.
7998 (expand_used_vars): Likewise.
7999 (partition_stack_vars): Likewise.
8000 * asan.c (asan_emit_stack_protection): Optionally disable
8001 after return stack usage.
8002 (instrument_derefs): Optionally disable memory access instrumentation.
8003 (instrument_builtin_call): Likewise.
8004 (instrument_strlen_call): Likewise.
8005 (asan_protect_global): Optionally disable global variables protection.
8006 * doc/invoke.texi: Added doc for new options.
8007 * params.def: Added new options.
8008 * params.h: Likewise.
8010 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8012 PR rtl-optimization/59724
8013 * ifcvt.c (cond_exec_process_if_block): Don't call
8014 flow_find_head_matching_sequence with 0 longest_match.
8015 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8016 non-active insns if !stop_after.
8017 (try_head_merge_bb): Revert 2014-01-07 changes.
8019 2014-01-08 Jeff Law <law@redhat.com>
8021 * ree.c (get_sub_rtx): New function, extracted from...
8022 (merge_def_and_ext): Here.
8023 (combine_reaching_defs): Use get_sub_rtx.
8025 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8027 * cgraph.h (varpool_variable_node): Do not choke on null node.
8029 2014-01-08 Catherine Moore <clm@codesourcery.com>
8031 * config/mips/mips.md (simple_return): Attempt to use JRC
8033 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8035 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8037 PR rtl-optimization/59137
8038 * reorg.c (steal_delay_list_from_target): Call update_block for
8040 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8042 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8044 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8045 two duplicate entries.
8047 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8050 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8052 * config/mips/mips.c (mips_truncated_op_cost): New function.
8053 (mips_rtx_costs): Adjust test for BADDU.
8054 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8056 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8058 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8059 (*baddu_si): ...this new pattern.
8061 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8064 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8066 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8069 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8071 (expand_expr, expand_normal): Adjust.
8072 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8073 inner_reference_p. Use inner_reference_p to expand inner references.
8074 (store_expr): Adjust.
8075 * cfgexpand.c (expand_call_stmt): Adjust.
8077 2014-01-08 Rong Xu <xur@google.com>
8079 * gcov-io.c (gcov_var): Move from gcov-io.h.
8080 (gcov_position): Ditto.
8081 (gcov_is_error): Ditto.
8082 (gcov_rewrite): Ditto.
8083 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8084 only part to libgcc/libgcov.h.
8086 2014-01-08 Marek Polacek <polacek@redhat.com>
8089 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8091 2014-01-08 Marek Polacek <polacek@redhat.com>
8094 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8096 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8098 PR rtl-optimization/59649
8099 * stor-layout.c (get_mode_bounds): For BImode return
8100 0 and STORE_FLAG_VALUE.
8102 2014-01-08 Richard Biener <rguenther@suse.de>
8105 * gimple.h (is_gimple_builtin_call): Remove.
8106 (gimple_builtin_call_types_compatible_p): New.
8107 (gimple_call_builtin_p): New overload.
8108 * gimple.c (is_gimple_builtin_call): Remove.
8109 (validate_call): Rename to ...
8110 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8112 (validate_type): New static function.
8113 (gimple_call_builtin_p): New overload and adjust.
8114 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8115 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8116 (gimple_fold_stmt_to_constant_1): Likewise.
8117 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8119 2014-01-08 Richard Biener <rguenther@suse.de>
8122 * gimplify.c (gimplify_expr): Gimplify register-register type
8123 VIEW_CONVERT_EXPRs to separate stmts.
8125 2014-01-07 Jeff Law <law@redhat.com>
8128 * ree.c (combine_set_extension): Handle case where source
8129 and destination registers in an extension insn are different.
8130 (combine_reaching_defs): Allow source and destination registers
8131 in extension to be different under limited circumstances.
8132 (add_removable_extension): Remove restriction that the
8133 source and destination registers in the extension are the same.
8134 (find_and_remove_re): Emit a copy from the extension's
8135 destination to its source after the defining insn if
8136 the source and destination registers are different.
8139 * ifcvt.c (merge_if_block): If we are merging a block with more than
8140 one successor with a block with no successors, remove any BARRIER
8141 after the second block.
8143 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8145 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8147 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8150 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8151 for 14-bit register offsets when INT14_OK_STRICT is false.
8153 2014-01-07 Roland Stigge <stigge@antcom.de>
8154 Michael Meissner <meissner@linux.vnet.ibm.com>
8157 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8158 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8160 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8162 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8165 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8167 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8168 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8169 rtx is const0_rtx or not.
8171 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8174 * target-globals.c (save_target_globals): Remove this_fn_optab
8176 * toplev.c: Include optabs.h.
8177 (target_reinit): Temporarily restore the global options if another
8178 set of options are in force.
8180 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8182 PR rtl-optimization/58668
8183 * cfgcleanup.c (flow_find_cross_jump): Don't count
8184 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8185 to determine what is counted.
8186 (flow_find_head_matching_sequence): Use active_insn_p to determine
8188 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8190 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8191 determine what is counted.
8193 PR tree-optimization/59643
8194 * tree-predcom.c (split_data_refs_to_components): If one dr is
8195 read and one write, determine_offset fails and the write isn't
8196 in the bad component, just put the read into the bad component.
8198 2014-01-07 Mike Stump <mikestump@comcast.net>
8199 Jakub Jelinek <jakub@redhat.com>
8202 * tree-core.h (struct tree_optimization_option): Change optabs
8203 type from unsigned char * to void *.
8204 * optabs.c (init_tree_optimization_optabs): Adjust
8205 TREE_OPTIMIZATION_OPTABS initialization.
8207 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8210 * config/i386/i386.h (struct machine_function): Add
8211 no_drap_save_restore field.
8212 * config/i386/i386.c (ix86_save_reg): Use
8213 !cfun->machine->no_drap_save_restore instead of
8214 crtl->stack_realign_needed.
8215 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8216 this function clears frame_pointer_needed. Set
8217 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8218 and DRAP reg is needed.
8220 2014-01-06 Marek Polacek <polacek@redhat.com>
8223 * doc/implement-c.texi: Mention that other integer types are
8224 permitted as bit-field types in strictly conforming mode.
8226 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8228 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8231 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8233 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8235 2014-01-06 Martin Jambor <mjambor@suse.cz>
8238 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8240 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8242 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8246 * var-tracking.c (add_stores): Preserve the value of the source even if
8247 we don't record the store.
8249 2014-01-06 Terry Guo <terry.guo@arm.com>
8251 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8253 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8256 * config/darwin.c (darwin_function_section): Adjust return values to
8257 correspond to optimisation changes made in r206070.
8259 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8261 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8262 from prefetch_block tune setting.
8263 (nocona_cost): Correct size of prefetch block to 64.
8265 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8267 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8268 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8269 used to save the static chain register in the computation of the offset
8270 from which the FP registers need to be restored.
8272 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8274 PR tree-optimization/59519
8275 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8276 ICE if get_current_def (current_new_name) is already non-NULL, as long
8277 as it is a phi result of some other phi in *new_exit_bb that has
8280 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8281 or vmovdqu* for misaligned_operand.
8282 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8283 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8284 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8285 aligned_mem for AVX512F masked aligned load and store builtins and for
8288 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8290 PR tree-optimization/59651
8291 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8292 Address range for negative step should be added by TYPE_SIZE_UNIT.
8294 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8296 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8297 overlapping registers also for registers other than the stack pointer.
8299 2014-01-03 Marek Polacek <polacek@redhat.com>
8302 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8305 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8308 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8311 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8312 (push splitter): Use <P:MODE_SIZE> instead of
8313 GET_MODE_SIZE (<P:MODE>mode).
8314 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8315 (mov -1, reg peephole2): Likewise.
8316 * config/i386/sse.md (*mov<mode>_internal,
8317 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8318 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8319 *<code><mode>3, *andnot<mode>3<mask_name>,
8320 <mask_codefor><code><mode>3<mask_name>): Likewise.
8321 * config/i386/subst.md (mask_mode512bit_condition,
8322 sd_mask_mode512bit_condition): Likewise.
8324 2014-01-02 Xinliang David Li <davidxl@google.com>
8326 PR tree-optimization/59303
8327 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8328 (dump_predicates): Better output format.
8329 (pred_equal_p): New function.
8330 (is_neq_relop_p): Ditto.
8331 (is_neq_zero_form_p): Ditto.
8332 (pred_expr_equal_p): Ditto.
8333 (pred_neg_p): Ditto.
8334 (simplify_pred): Ditto.
8335 (simplify_preds_2): Ditto.
8336 (simplify_preds_3): Ditto.
8337 (simplify_preds_4): Ditto.
8338 (simplify_preds): Ditto.
8340 (push_to_worklist): Ditto.
8341 (get_pred_info_from_cmp): Ditto.
8342 (is_degenerated_phi): Ditto.
8343 (normalize_one_pred_1): Ditto.
8344 (normalize_one_pred): Ditto.
8345 (normalize_one_pred_chain): Ditto.
8346 (normalize_preds): Ditto.
8347 (normalize_cond_1): Remove function.
8348 (normalize_cond): Ditto.
8349 (is_gcond_subset_of): Ditto.
8350 (is_subset_of_any): Ditto.
8351 (is_or_set_subset_of): Ditto.
8352 (is_and_set_subset_of): Ditto.
8353 (is_norm_cond_subset_of): Ditto.
8354 (pred_chain_length_cmp): Ditto.
8355 (convert_control_dep_chain_into_preds): Type change.
8356 (find_predicates): Ditto.
8357 (find_def_preds): Ditto.
8358 (destroy_predicates_vecs): Ditto.
8359 (find_matching_predicates_in_rest_chains): Ditto.
8360 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8361 (is_pred_expr_subset): Ditto.
8362 (is_pred_chain_subset_of): Ditto.
8363 (is_included_in): Ditto.
8364 (is_superset_of): Ditto.
8366 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8368 Update copyright years.
8370 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8372 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8373 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8374 config/arc/arc.md, config/arc/arc.opt,
8375 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8376 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8377 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8378 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8379 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8380 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8381 vtable-verify.c, vtable-verify.h: Use the standard form for the
8384 2014-01-02 Tobias Burnus <burnus@net-b.de>
8386 * gcc.c (process_command): Update copyright notice dates.
8387 * gcov-dump.c: Ditto.
8389 * doc/cpp.texi: Bump @copying's copyright year.
8390 * doc/cppinternals.texi: Ditto.
8391 * doc/gcc.texi: Ditto.
8392 * doc/gccint.texi: Ditto.
8393 * doc/gcov.texi: Ditto.
8394 * doc/install.texi: Ditto.
8395 * doc/invoke.texi: Ditto.
8397 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8399 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8401 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8403 * config/i386/sse.md (*mov<mode>_internal): Guard
8404 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8406 PR rtl-optimization/59647
8407 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8408 new_rtx into UNSIGNED_FLOAT rtxes.
8410 Copyright (C) 2014 Free Software Foundation, Inc.
8412 Copying and distribution of this file, with or without modification,
8413 are permitted in any medium without royalty provided the copyright
8414 notice and this notice are preserved.