1 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
3 * config/aarch64/aarch64-builtins.c
4 (aarch64_types_storestruct_lane_qualifiers): New.
5 (TYPES_STORESTRUCT_LANE): Likewise.
6 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
9 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
10 (vec_store_lanesci_lane<mode>): Likewise.
11 (vec_store_lanesxi_lane<mode>): Likewise.
12 (aarch64_st2_lane<VQ:mode>): Likewise.
13 (aarch64_st3_lane<VQ:mode>): Likewise.
14 (aarch64_st4_lane<VQ:mode>): Likewise.
15 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
16 * config/aarch64/arm_neon.h
17 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
18 use new macro arguments.
19 (__ST3_LANE_FUNC): Likewise.
20 (__ST4_LANE_FUNC): Likewise.
21 * config/aarch64/iterators.md (V_TWO_ELEM): New.
22 (V_THREE_ELEM): Likewise.
23 (V_FOUR_ELEM): Likewise.
25 2014-04-28 David Malcolm <dmalcolm@redhat.com>
27 * doc/gimple.texi: Replace the description of the now-defunct
28 union gimple_statement_d with a diagram showing the
29 gimple_statement_base class hierarchy and its relationships to
30 the GSS_ and GIMPLE_ enums.
32 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
34 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
35 * config/aarch64/aarch64.c
36 (aarch64_cannot_change_mode_class): Weaken conditions.
37 (aarch64_modes_tieable_p): New.
38 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
40 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
42 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
43 (loadsync_<mode>): Change mode.
44 (load_quadpti, store_quadpti): New.
45 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
46 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
48 2014-04-28 Martin Jambor <mjambor@suse.cz>
50 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
51 same alias type as the original statement.
52 (subreplacement_assignment_data): New type.
53 (handle_unscalarized_data_in_subtree): New type of parameter,
54 generate new memory accesses with same alias type as the original
56 (load_assign_lhs_subreplacements): Likewise.
57 (sra_modify_constructor_assign): Generate new memory accesses with
58 same alias type as the original statement.
60 2014-04-28 Richard Biener <rguenther@suse.de>
62 * tree-pass.h (TODO_verify_il): Define.
63 (TODO_verify_all): Complete properly.
64 * passes.c (execute_function_todo): Move existing loop-closed
65 SSA verification under TODO_verify_il.
66 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
67 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
68 Fix tree sharing issue.
70 2014-04-28 Richard Biener <rguenther@suse.de>
73 * builtins.def (DEF_C11_BUILTIN): Add.
74 (BUILT_IN_ALIGNED_ALLOC): Likewise.
75 * coretypes.h (enum function_class): Add function_c11_misc.
76 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
77 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
78 (call_may_clobber_ref_p_1): Likewise.
79 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
80 (mark_all_reaching_defs_necessary_1): Likewise.
81 (propagate_necessity): Likewise.
82 (eliminate_unnecessary_stmts): Likewise.
83 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
85 2014-04-28 Richard Biener <rguenther@suse.de>
87 * tree-vrp.c (vrp_var_may_overflow): Remove.
88 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
89 with overflow immediately bump to one before that value and
90 let iteration figure out overflow status.
92 2014-04-28 Richard Biener <rguenther@suse.de>
94 * configure.ac: Do valgrind header checks unconditionally.
95 Add --enable-valgrind-annotations.
96 * system.h: Guard valgrind header inclusion with
97 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
98 * alloc-pool.c (pool_alloc, pool_free): Use
99 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
100 to guard possibly dead code.
101 * config.in: Regenerated.
102 * configure: Likewise.
104 2014-04-28 Jeff Law <law@redhat.com>
106 PR tree-optimization/60902
107 * tree-ssa-threadedge.c
108 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
109 over real defs when invalidating outputs from statements that do not
110 produce useful outputs for threading.
112 2014-04-28 Richard Biener <rguenther@suse.de>
114 PR tree-optimization/60979
115 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
116 SCOPs that end in a block with a successor with abnormal
119 2014-04-28 Richard Biener <rguenther@suse.de>
121 * tree-pass.h (execute_pass_list): Adjust prototype.
122 * passes.c (pass_manager::execute_early_local_passes):
124 (do_per_function): Change callback signature, push all actual
125 work to the callbals.
126 (do_per_function_toporder): Likewise.
127 (execute_function_dump): Adjust.
128 (execute_function_todo): Likewise.
129 (clear_last_verified): Likewise.
130 (verify_curr_properties): Likewise.
131 (update_properties_after_pass): Likewise.
132 (execute_pass_list_1): Split out from ...
133 (execute_pass_list): ... here. Adjust.
134 (execute_ipa_pass_list): Likewise.
135 * cgraphunit.c (cgraph_add_new_function): Adjust.
136 (analyze_function): Likewise.
137 (expand_function): Likewise.
138 * cgraph.c (release_function_body): Free dominance info
139 here instead of asserting it was magically freed elsewhere.
141 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
143 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
144 * configure: Regenerate.
145 * config/sparc/sparc.opt (muser-mode): New option.
146 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
148 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
149 * doc/invoke.texi (SPARC options): Document -muser-mode.
151 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
153 * cselib.c (find_slot_memmode): Delete.
154 (cselib_hasher): Change compare_type to a struct.
155 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
157 (preserve_constants_and_equivs): Adjust for new compare_type.
158 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
159 (wrap_constant): Delete.
160 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
162 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
164 * doc/install.texi (Building with profile feedback): Remove
167 2014-04-26 Tom de Vries <tom@codesourcery.com>
169 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
172 2014-04-25 Cary Coutant <ccoutant@google.com>
175 * dwarf2out.c (should_move_die_to_comdat): A type definition
176 can contain a subprogram definition, but don't move it to a
178 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
179 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
181 (clone_tree_hash): Rename to...
182 (clone_tree_partial): ...this; change callers. Copy
183 DW_TAG_subprogram DIEs as declarations.
184 (copy_decls_walk): Don't copy children of a declaration into a
187 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
190 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
193 2014-04-25 Jiong Wang <jiong.wang@arm.com>
195 * config/arm/predicates.md (call_insn_operand): Add long_call check.
196 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
198 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
201 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
203 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
206 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
208 PR tree-optimization/60930
209 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
210 creating a multiply candidate by folding two constant
211 multiplicands when the result overflows.
213 2014-04-25 Jakub Jelinek <jakub@redhat.com>
215 PR tree-optimization/60960
216 * tree-vect-generic.c (expand_vector_operation): Only call
217 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
219 2014-04-25 Tom de Vries <tom@codesourcery.com>
221 * expr.c (clobber_reg_mode): New function.
222 * expr.h (clobber_reg): New function.
224 2014-04-25 Tom de Vries <tom@codesourcery.com>
226 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
229 2014-04-25 Radovan Obradovic <robradovic@mips.com>
230 Tom de Vries <tom@codesourcery.com>
232 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
234 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
235 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
236 new argument to find_all_hard_reg_sets call.
238 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
240 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
241 Use HOST_WIDE_INT_C for mask literal.
242 (aarch_rev16_shleft_mask_imm_p): Likewise.
244 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
247 * config/sparc/sparc.md (ashlsi3_extend): Delete.
249 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
251 PR preprocessor/56540
252 * config/i386/i386-c.c (ix86_target_macros): Define
253 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
255 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
257 * configure.ac (tga_func): Remove.
258 (LIB_TLS_SPEC): Remove.
259 * configure: Regenerate.
260 * config.in: Regenerate.
261 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
263 2014-04-25 Richard Biener <rguenther@suse.de>
266 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
267 call stmt use/clobber sets during stmt walk instead of
268 walking the possibly incomplete set of caller edges.
270 2014-04-25 Richard Biener <rguenther@suse.de>
273 * passes.c (apply_ipa_transforms): Inline into only caller ...
274 (execute_one_pass): ... here. Properly bring in function
275 bodies for nodes we want to apply IPA transforms to.
277 2014-04-24 Cong Hou <congh@google.com>
279 PR tree-optimization/60896
280 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
281 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
282 (vect_mark_pattern_stmts): Set the def type of all statements in
283 PATTERN_DEF_SEQ as vect_internal_def.
285 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
287 * doc/extend.texi (PowerPC Built-in Functions): Document new
288 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
290 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
292 * config/rs6000/predicates.md (const_0_to_3_operand): New
293 predicate to match 0..3 integer constants.
295 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
296 to support adding miscellaneous builtin functions.
297 (BU_DFP_MISC_2): Likewise.
298 (BU_P7_MISC_1): Likewise.
299 (BU_P7_MISC_2): Likewise.
300 (BU_P8V_MISC_3): Likewise.
301 (BU_MISC_1): Likewise.
302 (BU_MISC_2): Likewise.
303 (DIVWE): Add extended divide builtin functions.
311 (DXEX): Add decimal floating-point builtin functions.
323 (CDTBCD): Add new BCD builtin functions.
327 (BCDADD_LT): Likewise.
328 (BCDADD_EQ): Likewise.
329 (BCDADD_GT): Likewise.
330 (BCDADD_OV): Likewise.
332 (BCDSUB_LT): Likewise.
333 (BCDSUB_EQ): Likewise.
334 (BCDSUB_GT): Likewise.
335 (BCDSUB_OV): Likewise.
336 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
337 (UNPACK_TD): Likewise.
339 (UNPACK_TF): Likewise.
340 (UNPACK_TF_0): Likewise.
341 (UNPACK_TF_1): Likewise.
342 (PACK_V1TI): Likewise.
343 (UNPACK_V1TI): Likewise.
345 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
346 support for decimal floating point builtin functions.
347 (rs6000_expand_ternop_builtin): Add checks for the new builtin
348 functions that take constant arguments.
349 (rs6000_invalid_builtin): Add decimal floating point builtin
351 (rs6000_init_builtins): Setup long double, _Decimal64, and
352 _Decimal128 types for new builtin functions.
353 (builtin_function_type): Set the unsigned flags appropriately for
354 the new builtin functions.
355 (rs6000_opt_masks): Add support for decimal floating point builtin
358 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
359 floating point builtin functions.
360 (RS6000_BTM_COMMON): Likewise.
361 (RS6000_BTI_long_double): Likewise.
362 (RS6000_BTI_dfloat64): Likewise.
363 (RS6000_BTI_dfloat128): Likewise.
364 (long_double_type_internal_node): Likewise.
365 (dfloat64_type_internal_node): Likewise.
366 (dfloat128_type_internal_node): Likewise.
368 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
369 2.07 bcd arithmetic instructions.
370 (UNSPEC_BCDSUB): Likewise.
371 (UNSPEC_BCD_OVERFLOW): Likewise.
372 (UNSPEC_BCD_ADD_SUB): Likewise.
373 (bcd_add_sub): Likewise.
374 (BCD_TEST): Likewise.
375 (bcd<bcd_add_sub>): Likewise.
376 (bcd<bcd_add_sub>_test): Likewise.
377 (bcd<bcd_add_sub>_test2): Likewise.
378 (bcd<bcd_add_sub>_<code>): Likewise.
379 (peephole2 for combined bcd ops): Likewise.
381 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
382 decimal floating point builtin functions.
383 (UNSPEC_DENBCD): Likewise.
384 (UNSPEC_DXEX): Likewise.
385 (UNSPEC_DIEX): Likewise.
386 (UNSPEC_DSCLI): Likewise.
387 (UNSPEC_DSCRI): Likewise.
388 (D64_D128): Likewise.
389 (dfp_suffix): Likewise.
390 (dfp_ddedpd_<mode>): Likewise.
391 (dfp_denbcd_<mode>): Likewise.
392 (dfp_dxex_<mode>): Likewise.
393 (dfp_diex_<mode>): Likewise.
394 (dfp_dscli_<mode>): Likewise.
395 (dfp_dscri_<mode>): Likewise.
397 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
399 (UNSPEC_CDTBCD): Likewise.
400 (UNSPEC_CBCDTD): Likewise.
401 (UNSPEC_DIVE): Add support for new extended divide builtin
403 (UNSPEC_DIVEO): Likewise.
404 (UNSPEC_DIVEU): Likewise.
405 (UNSPEC_DIVEUO): Likewise.
406 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
407 pack/unpack 128-bit types.
408 (UNSPEC_PACK_128BIT): Likewise.
409 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
410 (udiv<mode>3): Use idiv_ldiv mode attribute.
411 (div<mode>3): Likewise.
412 (addg6s): Add new BCD builtin functions.
415 (UNSPEC_DIV_EXTEND): Add support for new extended divide
417 (div_extend): Likewise.
418 (div<div_extend>_<mode>"): Likewise.
419 (FP128_64): Add support for new builtin functions to pack/unpack
421 (unpack<mode>): Likewise.
422 (unpacktf_0): Likewise.
423 (unpacktf_1): Likewise.
424 (unpack<mode>_dm): Likewise.
425 (unpack<mode>_nodm): Likewise.
426 (pack<mode>): Likewise.
427 (unpackv1ti): Likewise.
428 (packv1ti): Likewise.
430 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
432 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
435 2014-04-24 Jakub Jelinek <jakub@redhat.com>
437 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
438 * gimplify.c (omp_is_private): Change last argument's type to int.
439 Only diagnose lastprivate if the simd argument is 1, only diagnose
440 linear if the simd argument is 2.
441 (gimplify_omp_for): Adjust omp_is_private callers. When adding
442 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
443 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
444 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
445 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
446 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
447 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
448 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
449 * tree-nested.c (convert_nonlocal_omp_clauses,
450 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
452 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
455 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
458 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
460 * flag-types.h (enum ivar_visibility): Add.
462 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
464 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
467 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
469 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
471 2014-04-24 Radovan Obradovic <robradovic@mips.com>
472 Tom de Vries <tom@codesourcery.com>
474 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
475 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
477 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
478 * emit-rtl.c (try_split): Same.
480 2014-04-24 Radovan Obradovic <robradovic@mips.com>
481 Tom de Vries <tom@codesourcery.com>
483 * common.opt (fuse-caller-save): New option.
485 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
487 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
490 2014-04-24 Richard Biener <rguenther@suse.de>
492 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
493 during TER and instead use the sepops interface for expanding
494 non-GIMPLE_SINGLE_RHS.
496 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
498 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
499 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
501 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
503 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
504 assembler 64-bit option.
505 * configure: Regenerate.
507 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
509 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
510 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
511 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
512 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
513 (TARGET_CRYPTO): Take TARGET_SIMD into account.
515 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
517 * config/aarch64/aarch64-builtins.c
518 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
519 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
520 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
521 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
523 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
524 (Vrevsuff): New mode attribute.
526 2014-04-24 Terry Guo <terry.guo@arm.com>
528 * config/arm/arm.h (machine_function): Define variable
529 after_arm_reorg here.
530 * config/arm/arm.c (after_arm_reorg): Remove the definition.
531 (arm_split_constant): Update the way to access variable
534 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
536 2014-04-23 Tom de Vries <tom@codesourcery.com>
538 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
540 2014-04-23 David Malcolm <dmalcolm@redhat.com>
542 * is-a.h: Update comments to reflect the following changes to the
543 "pointerness" of the API, making the template parameter match the
544 return type, allowing use of is-a.h with typedefs of pointers.
545 (is_a_helper::cast): Return a T rather then a pointer to a T, so
546 that the return type matches the parameter to the is_a_helper.
548 (dyn_cast): Likewise.
550 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
551 pointer from the is-a.h API.
553 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
554 (is_a_helper <cgraph_node *>::test): ...this, matching change to
556 (is_a_helper <varpool_node>::test): Likewise, convert to...
557 (is_a_helper <varpool_node *>::test): ...this.
559 (varpool_first_variable): Update for removal of implicit pointer
561 (varpool_next_variable): Likewise.
562 (varpool_first_static_initializer): Likewise.
563 (varpool_next_static_initializer): Likewise.
564 (varpool_first_defined_variable): Likewise.
565 (varpool_next_defined_variable): Likewise.
566 (cgraph_first_defined_function): Likewise.
567 (cgraph_next_defined_function): Likewise.
568 (cgraph_first_function): Likewise.
569 (cgraph_next_function): Likewise.
570 (cgraph_first_function_with_gimple_body): Likewise.
571 (cgraph_next_function_with_gimple_body): Likewise.
572 (cgraph_alias_target): Likewise.
573 (varpool_alias_target): Likewise.
574 (cgraph_function_or_thunk_node): Likewise.
575 (varpool_variable_node): Likewise.
576 (symtab_real_symbol_p): Likewise.
577 * cgraphunit.c (referred_to_p): Likewise.
578 (analyze_functions): Likewise.
579 (handle_alias_pairs): Likewise.
580 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
581 * gimple-ssa.h (gimple_vuse_op): Likewise.
582 (gimple_vdef_op): Likewise.
583 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
584 * gimple.c (gimple_build_asm_1): Likewise.
585 (gimple_build_try): Likewise.
586 (gimple_build_resx): Likewise.
587 (gimple_build_eh_dispatch): Likewise.
588 (gimple_build_omp_for): Likewise.
589 (gimple_omp_for_set_clauses): Likewise.
591 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
592 (is_a_helper <gimple_statement_asm *>::test): ...this.
593 (is_a_helper <gimple_statement_bind>::test): Convert to...
594 (is_a_helper <gimple_statement_bind *>::test): ...this.
595 (is_a_helper <gimple_statement_call>::test): Convert to...
596 (is_a_helper <gimple_statement_call *>::test): ...this.
597 (is_a_helper <gimple_statement_catch>::test): Convert to...
598 (is_a_helper <gimple_statement_catch *>::test): ...this.
599 (is_a_helper <gimple_statement_resx>::test): Convert to...
600 (is_a_helper <gimple_statement_resx *>::test): ...this.
601 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
602 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
603 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
604 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
605 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
606 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
607 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
608 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
609 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
610 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
611 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
612 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
613 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
614 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
615 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
616 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
617 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
618 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
619 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
620 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
621 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
622 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
623 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
624 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
625 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
626 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
627 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
628 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
629 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
630 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
631 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
632 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
633 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
634 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
635 (is_a_helper <gimple_statement_phi>::test): Convert to...
636 (is_a_helper <gimple_statement_phi *>::test): ...this.
637 (is_a_helper <gimple_statement_transaction>::test): Convert to...
638 (is_a_helper <gimple_statement_transaction *>::test): ...this.
639 (is_a_helper <gimple_statement_try>::test): Convert to...
640 (is_a_helper <gimple_statement_try *>::test): ...this.
641 (is_a_helper <gimple_statement_wce>::test): Convert to...
642 (is_a_helper <gimple_statement_wce *>::test): ...this.
643 (is_a_helper <const gimple_statement_asm>::test): Convert to...
644 (is_a_helper <const gimple_statement_asm *>::test): ...this.
645 (is_a_helper <const gimple_statement_bind>::test): Convert to...
646 (is_a_helper <const gimple_statement_bind *>::test): ...this.
647 (is_a_helper <const gimple_statement_call>::test): Convert to...
648 (is_a_helper <const gimple_statement_call *>::test): ...this.
649 (is_a_helper <const gimple_statement_catch>::test): Convert to...
650 (is_a_helper <const gimple_statement_catch *>::test): ...this.
651 (is_a_helper <const gimple_statement_resx>::test): Convert to...
652 (is_a_helper <const gimple_statement_resx *>::test): ...this.
653 (is_a_helper <const gimple_statement_eh_dispatch>::test):
655 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
657 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
659 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
660 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
662 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
664 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
666 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
668 (is_a_helper <const gimple_statement_omp_return>::test): Convert
670 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
671 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
673 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
674 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
676 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
677 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
678 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
679 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
681 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
682 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
684 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
685 (is_a_helper <const gimple_statement_omp_target>::test): Convert
687 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
688 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
690 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
691 (is_a_helper <const gimple_statement_omp_single>::test): Convert
693 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
694 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
696 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
697 (is_a_helper <const gimple_statement_omp_task>::test): Convert
699 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
700 (is_a_helper <const gimple_statement_phi>::test): Convert to...
701 (is_a_helper <const gimple_statement_phi *>::test): ...this.
702 (is_a_helper <const gimple_statement_transaction>::test): Convert
704 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
705 (is_a_helper <const gimple_statement_with_ops>::test): Convert
707 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
708 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
709 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
710 (is_a_helper <const gimple_statement_with_memory_ops>::test):
712 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
714 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
716 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
718 (gimple_use_ops): Update for removal of implicit pointer from the
720 (gimple_set_use_ops): Likewise.
721 (gimple_vuse): Likewise.
722 (gimple_vdef): Likewise.
723 (gimple_vuse_ptr): Likewise.
724 (gimple_vdef_ptr): Likewise.
725 (gimple_set_vuse): Likewise.
726 (gimple_set_vdef): Likewise.
727 (gimple_omp_return_set_lhs): Likewise.
728 (gimple_omp_return_lhs): Likewise.
729 (gimple_omp_return_lhs_ptr): Likewise.
730 (gimple_call_fntype): Likewise.
731 (gimple_call_set_fntype): Likewise.
732 (gimple_call_set_internal_fn): Likewise.
733 (gimple_call_use_set): Likewise.
734 (gimple_call_clobber_set): Likewise.
735 (gimple_bind_vars): Likewise.
736 (gimple_bind_set_vars): Likewise.
737 (gimple_bind_body_ptr): Likewise.
738 (gimple_bind_set_body): Likewise.
739 (gimple_bind_add_stmt): Likewise.
740 (gimple_bind_block): Likewise.
741 (gimple_bind_set_block): Likewise.
742 (gimple_asm_ninputs): Likewise.
743 (gimple_asm_noutputs): Likewise.
744 (gimple_asm_nclobbers): Likewise.
745 (gimple_asm_nlabels): Likewise.
746 (gimple_asm_input_op): Likewise.
747 (gimple_asm_input_op_ptr): Likewise.
748 (gimple_asm_output_op): Likewise.
749 (gimple_asm_output_op_ptr): Likewise.
750 (gimple_asm_set_output_op): Likewise.
751 (gimple_asm_clobber_op): Likewise.
752 (gimple_asm_set_clobber_op): Likewise.
753 (gimple_asm_label_op): Likewise.
754 (gimple_asm_set_label_op): Likewise.
755 (gimple_asm_string): Likewise.
756 (gimple_catch_types): Likewise.
757 (gimple_catch_types_ptr): Likewise.
758 (gimple_catch_handler_ptr): Likewise.
759 (gimple_catch_set_types): Likewise.
760 (gimple_catch_set_handler): Likewise.
761 (gimple_eh_filter_types): Likewise.
762 (gimple_eh_filter_types_ptr): Likewise.
763 (gimple_eh_filter_failure_ptr): Likewise.
764 (gimple_eh_filter_set_types): Likewise.
765 (gimple_eh_filter_set_failure): Likewise.
766 (gimple_eh_must_not_throw_fndecl): Likewise.
767 (gimple_eh_must_not_throw_set_fndecl): Likewise.
768 (gimple_eh_else_n_body_ptr): Likewise.
769 (gimple_eh_else_e_body_ptr): Likewise.
770 (gimple_eh_else_set_n_body): Likewise.
771 (gimple_eh_else_set_e_body): Likewise.
772 (gimple_try_eval_ptr): Likewise.
773 (gimple_try_cleanup_ptr): Likewise.
774 (gimple_try_set_eval): Likewise.
775 (gimple_try_set_cleanup): Likewise.
776 (gimple_wce_cleanup_ptr): Likewise.
777 (gimple_wce_set_cleanup): Likewise.
778 (gimple_phi_capacity): Likewise.
779 (gimple_phi_num_args): Likewise.
780 (gimple_phi_result): Likewise.
781 (gimple_phi_result_ptr): Likewise.
782 (gimple_phi_set_result): Likewise.
783 (gimple_phi_arg): Likewise.
784 (gimple_phi_set_arg): Likewise.
785 (gimple_resx_region): Likewise.
786 (gimple_resx_set_region): Likewise.
787 (gimple_eh_dispatch_region): Likewise.
788 (gimple_eh_dispatch_set_region): Likewise.
789 (gimple_omp_critical_name): Likewise.
790 (gimple_omp_critical_name_ptr): Likewise.
791 (gimple_omp_critical_set_name): Likewise.
792 (gimple_omp_for_clauses): Likewise.
793 (gimple_omp_for_clauses_ptr): Likewise.
794 (gimple_omp_for_set_clauses): Likewise.
795 (gimple_omp_for_collapse): Likewise.
796 (gimple_omp_for_index): Likewise.
797 (gimple_omp_for_index_ptr): Likewise.
798 (gimple_omp_for_set_index): Likewise.
799 (gimple_omp_for_initial): Likewise.
800 (gimple_omp_for_initial_ptr): Likewise.
801 (gimple_omp_for_set_initial): Likewise.
802 (gimple_omp_for_final): Likewise.
803 (gimple_omp_for_final_ptr): Likewise.
804 (gimple_omp_for_set_final): Likewise.
805 (gimple_omp_for_incr): Likewise.
806 (gimple_omp_for_incr_ptr): Likewise.
807 (gimple_omp_for_set_incr): Likewise.
808 (gimple_omp_for_pre_body_ptr): Likewise.
809 (gimple_omp_for_set_pre_body): Likewise.
810 (gimple_omp_parallel_clauses): Likewise.
811 (gimple_omp_parallel_clauses_ptr): Likewise.
812 (gimple_omp_parallel_set_clauses): Likewise.
813 (gimple_omp_parallel_child_fn): Likewise.
814 (gimple_omp_parallel_child_fn_ptr): Likewise.
815 (gimple_omp_parallel_set_child_fn): Likewise.
816 (gimple_omp_parallel_data_arg): Likewise.
817 (gimple_omp_parallel_data_arg_ptr): Likewise.
818 (gimple_omp_parallel_set_data_arg): Likewise.
819 (gimple_omp_task_clauses): Likewise.
820 (gimple_omp_task_clauses_ptr): Likewise.
821 (gimple_omp_task_set_clauses): Likewise.
822 (gimple_omp_task_child_fn): Likewise.
823 (gimple_omp_task_child_fn_ptr): Likewise.
824 (gimple_omp_task_set_child_fn): Likewise.
825 (gimple_omp_task_data_arg): Likewise.
826 (gimple_omp_task_data_arg_ptr): Likewise.
827 (gimple_omp_task_set_data_arg): Likewise.
828 (gimple_omp_taskreg_clauses): Likewise.
829 (gimple_omp_taskreg_clauses_ptr): Likewise.
830 (gimple_omp_taskreg_set_clauses): Likewise.
831 (gimple_omp_taskreg_child_fn): Likewise.
832 (gimple_omp_taskreg_child_fn_ptr): Likewise.
833 (gimple_omp_taskreg_set_child_fn): Likewise.
834 (gimple_omp_taskreg_data_arg): Likewise.
835 (gimple_omp_taskreg_data_arg_ptr): Likewise.
836 (gimple_omp_taskreg_set_data_arg): Likewise.
837 (gimple_omp_task_copy_fn): Likewise.
838 (gimple_omp_task_copy_fn_ptr): Likewise.
839 (gimple_omp_task_set_copy_fn): Likewise.
840 (gimple_omp_task_arg_size): Likewise.
841 (gimple_omp_task_arg_size_ptr): Likewise.
842 (gimple_omp_task_set_arg_size): Likewise.
843 (gimple_omp_task_arg_align): Likewise.
844 (gimple_omp_task_arg_align_ptr): Likewise.
845 (gimple_omp_task_set_arg_align): Likewise.
846 (gimple_omp_single_clauses): Likewise.
847 (gimple_omp_single_clauses_ptr): Likewise.
848 (gimple_omp_single_set_clauses): Likewise.
849 (gimple_omp_target_clauses): Likewise.
850 (gimple_omp_target_clauses_ptr): Likewise.
851 (gimple_omp_target_set_clauses): Likewise.
852 (gimple_omp_target_child_fn): Likewise.
853 (gimple_omp_target_child_fn_ptr): Likewise.
854 (gimple_omp_target_set_child_fn): Likewise.
855 (gimple_omp_target_data_arg): Likewise.
856 (gimple_omp_target_data_arg_ptr): Likewise.
857 (gimple_omp_target_set_data_arg): Likewise.
858 (gimple_omp_teams_clauses): Likewise.
859 (gimple_omp_teams_clauses_ptr): Likewise.
860 (gimple_omp_teams_set_clauses): Likewise.
861 (gimple_omp_sections_clauses): Likewise.
862 (gimple_omp_sections_clauses_ptr): Likewise.
863 (gimple_omp_sections_set_clauses): Likewise.
864 (gimple_omp_sections_control): Likewise.
865 (gimple_omp_sections_control_ptr): Likewise.
866 (gimple_omp_sections_set_control): Likewise.
867 (gimple_omp_for_set_cond): Likewise.
868 (gimple_omp_for_cond): Likewise.
869 (gimple_omp_atomic_store_set_val): Likewise.
870 (gimple_omp_atomic_store_val): Likewise.
871 (gimple_omp_atomic_store_val_ptr): Likewise.
872 (gimple_omp_atomic_load_set_lhs): Likewise.
873 (gimple_omp_atomic_load_lhs): Likewise.
874 (gimple_omp_atomic_load_lhs_ptr): Likewise.
875 (gimple_omp_atomic_load_set_rhs): Likewise.
876 (gimple_omp_atomic_load_rhs): Likewise.
877 (gimple_omp_atomic_load_rhs_ptr): Likewise.
878 (gimple_omp_continue_control_def): Likewise.
879 (gimple_omp_continue_control_def_ptr): Likewise.
880 (gimple_omp_continue_set_control_def): Likewise.
881 (gimple_omp_continue_control_use): Likewise.
882 (gimple_omp_continue_control_use_ptr): Likewise.
883 (gimple_omp_continue_set_control_use): Likewise.
884 (gimple_transaction_body_ptr): Likewise.
885 (gimple_transaction_label): Likewise.
886 (gimple_transaction_label_ptr): Likewise.
887 (gimple_transaction_set_body): Likewise.
888 (gimple_transaction_set_label): Likewise.
890 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
891 * ipa-inline-analysis.c (inline_write_summary): Likewise.
892 * ipa-ref.c (ipa_record_reference): Likewise.
893 * ipa-reference.c (analyze_function): Likewise.
894 (ipa_reference_write_optimization_summary): Likewise.
895 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
896 (address_taken_from_non_vtable_p): Likewise.
897 (comdat_can_be_unshared_p_1): Likewise.
898 * lto-cgraph.c (lto_output_ref): Likewise.
899 (add_references): Likewise.
900 (compute_ltrans_boundary): Likewise.
901 (output_symtab): Likewise.
902 (input_ref): Likewise.
903 (input_cgraph_1): Likewise.
904 (output_cgraph_opt_summary): Likewise.
905 * lto-streamer-out.c (lto_output): Likewise.
906 (output_symbol_p): Likewise.
907 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
908 (lsei_start_function_in_partition): Likewise.
909 (lsei_next_variable_in_partition): Likewise.
910 (lsei_start_variable_in_partition): Likewise.
911 * symtab.c (insert_to_assembler_name_hash): Likewise.
912 (unlink_from_assembler_name_hash): Likewise.
913 (symtab_unregister_node): Likewise.
914 (symtab_remove_node): Likewise.
915 (dump_symtab_node): Likewise.
916 (verify_symtab_base): Likewise.
917 (verify_symtab_node): Likewise.
918 (symtab_make_decl_local): Likewise.
919 (symtab_alias_ultimate_target): Likewise.
920 (symtab_resolve_alias): Likewise.
921 (symtab_get_symbol_partitioning_class): Likewise.
922 * tree-phinodes.c (allocate_phi_node): Likewise.
923 (reserve_phi_args_for_new_edge): Likewise.
924 (remove_phi_args): Likewise.
925 * varpool.c (varpool_node_for_asm): Likewise.
926 (varpool_remove_unreferenced_decls): Likewise.
928 2014-04-23 Jeff Law <law@redhat.com>
930 PR tree-optimization/60902
931 * tree-ssa-threadedge.c
932 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
933 invalidate outputs from statements that do not produce useful
934 outputs for threading.
936 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
938 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
939 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
940 machine descriptions for Stack Smashing Protector.
942 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
944 * aarch64.md (<optab>_rol<mode>3): New pattern.
945 (<optab>_rolsi3_uxtw): Likewise.
946 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
948 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
950 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
951 (arm_cortex_a12_tune): Likewise.
953 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
955 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
957 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
959 * config/arm/arm.md (arm_rev16si2): New pattern.
960 (arm_rev16si2_alt): Likewise.
961 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
963 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
965 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
966 (rev16<mode>2_alt): Likewise.
967 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
968 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
969 (aarch_rev16_shleft_mask_imm_p): Likewise.
970 (aarch_rev16_p_1): Likewise.
971 (aarch_rev16_p): Likewise.
972 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
973 (aarch_rev16_shright_mask_imm_p): Likewise.
974 (aarch_rev16_shleft_mask_imm_p): Likewise.
976 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
978 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
979 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
981 (cortex_a53_extra_costs): Likewise.
982 (cortex_a57_extra_costs): Likewise.
983 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
984 (cortexa7_extra_costs): Likewise.
985 (cortexa8_extra_costs): Likewise.
986 (cortexa12_extra_costs): Likewise.
987 (cortexa15_extra_costs): Likewise.
988 (v7m_extra_costs): Likewise.
989 (arm_new_rtx_costs): Handle BSWAP.
991 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
993 * config/arm/arm.c (cortexa8_extra_costs): New table.
994 (arm_cortex_a8_tune): New tuning struct.
995 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
997 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
999 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
1001 2014-04-23 Richard Biener <rguenther@suse.de>
1003 * Makefile.in (OBJS): Remove loop-unswitch.o.
1004 * loop-unswitch.c: Delete.
1005 * tree-pass.h (make_pass_rtl_unswitch): Remove.
1006 * passes.def (pass_rtl_unswitch): Likewise.
1007 * loop-init.c (gate_rtl_unswitch): Likewise.
1008 (rtl_unswitch): Likewise.
1009 (pass_data_rtl_unswitch): Likewise.
1010 (pass_rtl_unswitch): Likewise.
1011 (make_pass_rtl_unswitch): Likewise.
1012 * rtl.h (reversed_condition): Likewise.
1013 (compare_and_jump_seq): Likewise.
1014 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
1016 * loop-unroll.c (compare_and_jump_seq): Likewise.
1018 2014-04-23 Richard Biener <rguenther@suse.de>
1020 PR tree-optimization/60903
1021 * tree-ssa-loop-im.c (analyze_memory_references): Remove
1022 commented code block.
1023 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
1024 loop flags to newly created BBs and edges.
1026 2014-04-23 Nick Clifton <nickc@redhat.com>
1028 * config/msp430/msp430.c (msp430_handle_option): Move function
1030 (msp430_option_override): Simplify mcu and mcpu option handling.
1031 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
1032 support for -mhwmult command line option.
1033 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
1034 -mhwmult command line option.
1035 (msp430_hwmult_enabled): Delete.
1036 (msp43o_output_labelref): Add support for -mhwmult command line
1038 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
1039 (umulsidi3): Likewise.
1040 * config/msp430/msp430.opt (mmcu): Add Report attribute.
1041 (mcpu, mlarge, msmall): Likewise.
1042 (mhwmult): New option.
1043 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
1045 (msp430_is_f5_mcu): Remove prototype.
1046 (msp430_use_f5_series_hwmult): Add prototype.
1047 * config/msp430/msp430-opts.h: New file.
1048 * common/config/msp430: New directory.
1049 * common/config/msp430/msp430-common.c: New file.
1050 * config.gcc (msp430): Remove target_has_targetm_common.
1051 * doc/invoke.texi: Document -mhwmult command line option.
1053 2014-04-23 Nick Clifton <nickc@redhat.com>
1055 * config/i386/cygwin.h (ENDFILE_SPEC): Include
1056 default-manifest.o if it can be found in the search path.
1057 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
1059 2014-04-23 Terry Guo <terry.guo@arm.com>
1061 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
1063 2014-04-23 Richard Biener <rguenther@suse.de>
1066 * tree-inline.c (declare_return_variable): Use mark_addressable.
1068 2014-04-23 Richard Biener <rguenther@suse.de>
1071 * loop-init.c (loop_optimizer_init): Make sure to apply
1072 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
1074 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1077 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
1079 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
1080 if flag_sanitize_undefined_trap_on_error.
1081 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
1082 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
1083 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
1084 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
1085 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
1086 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
1087 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
1088 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
1089 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
1090 * ubsan.c (ubsan_instrument_unreachable): Return
1091 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1092 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1093 if flag_sanitize_undefined_trap_on_error and
1094 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1095 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1096 instrument_bool_enum_load): Emit __builtin_trap () if
1097 flag_sanitize_undefined_trap_on_error and
1098 __builtin_handle_*_abort () if !flag_sanitize_recover.
1099 * doc/invoke.texi (-fsanitize-recover,
1100 -fsanitize-undefined-trap-on-error): Document.
1102 2014-04-22 Christian Bruel <christian.bruel@st.com>
1104 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1105 Force immediates to SImode.
1107 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1109 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1111 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1112 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1113 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1114 (nios2_fpu_insn): Add entry for round.
1115 (N2FPU_NO_ERRNO_P): Define.
1116 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1118 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1120 2014-04-22 Richard Henderson <rth@redhat.com>
1122 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1123 (add<GPI>3_compare0): Remove leading * from name.
1124 (add<GPI>3_carryin): Likewise.
1125 (sub<GPI>3_compare0): Likewise.
1126 (sub<GPI>3_carryin): Likewise.
1127 (<su_optab>mulditi3): New expander.
1128 (multi3): New expander.
1129 (madd<GPI>): Remove leading * from name.
1131 2014-04-22 Martin Jambor <mjambor@suse.cz>
1133 * cgraphclones.c (cgraph_function_versioning): Copy
1134 ipa_transforms_to_apply instead of asserting it is empty.
1136 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1139 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1140 on count_exp to get mode.
1142 2014-04-22 Andrew Pinski <apinski@cavium.com>
1144 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1145 Handle TLS for ILP32.
1146 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1147 (tlsie_small_<mode>): this and handle PTR.
1148 (tlsie_small_sidi): New pattern.
1149 (tlsle_small): Change to an expand to handle ILP32.
1150 (tlsle_small_<mode>): New pattern.
1151 (tlsdesc_small): Rename to ...
1152 (tlsdesc_small_<mode>): this and handle PTR.
1154 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1156 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1158 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1160 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1161 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1162 (aarch64_types_signed_poly_qualifiers): Likewise.
1163 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1164 (aarch64_types_poly_signed_qualifiers): Likewise.
1165 (TYPES_REINTERP_SS): Type macro added.
1166 (TYPES_REINTERP_SU): Likewise.
1167 (TYPES_REINTERP_SP): Likewise.
1168 (TYPES_REINTERP_US): Likewise.
1169 (TYPES_REINTERP_PS): Likewise.
1170 (aarch64_fold_builtin): New expression folding added.
1171 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1172 Declarations removed.
1173 (REINTERP_SS): Declarations added.
1174 (REINTERP_US): Likewise.
1175 (REINTERP_PS): Likewise.
1176 (REINTERP_SU): Likewise.
1177 (REINTERP_SP): Likewise.
1178 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1179 (vreinterpretq_p8_f64): Likewise.
1180 (vreinterpret_p16_f64): Likewise.
1181 (vreinterpretq_p16_f64): Likewise.
1182 (vreinterpret_f32_f64): Likewise.
1183 (vreinterpretq_f32_f64): Likewise.
1184 (vreinterpret_f64_f32): Likewise.
1185 (vreinterpret_f64_p8): Likewise.
1186 (vreinterpret_f64_p16): Likewise.
1187 (vreinterpret_f64_s8): Likewise.
1188 (vreinterpret_f64_s16): Likewise.
1189 (vreinterpret_f64_s32): Likewise.
1190 (vreinterpret_f64_s64): Likewise.
1191 (vreinterpret_f64_u8): Likewise.
1192 (vreinterpret_f64_u16): Likewise.
1193 (vreinterpret_f64_u32): Likewise.
1194 (vreinterpret_f64_u64): Likewise.
1195 (vreinterpretq_f64_f32): Likewise.
1196 (vreinterpretq_f64_p8): Likewise.
1197 (vreinterpretq_f64_p16): Likewise.
1198 (vreinterpretq_f64_s8): Likewise.
1199 (vreinterpretq_f64_s16): Likewise.
1200 (vreinterpretq_f64_s32): Likewise.
1201 (vreinterpretq_f64_s64): Likewise.
1202 (vreinterpretq_f64_u8): Likewise.
1203 (vreinterpretq_f64_u16): Likewise.
1204 (vreinterpretq_f64_u32): Likewise.
1205 (vreinterpretq_f64_u64): Likewise.
1206 (vreinterpret_s64_f64): Likewise.
1207 (vreinterpretq_s64_f64): Likewise.
1208 (vreinterpret_u64_f64): Likewise.
1209 (vreinterpretq_u64_f64): Likewise.
1210 (vreinterpret_s8_f64): Likewise.
1211 (vreinterpretq_s8_f64): Likewise.
1212 (vreinterpret_s16_f64): Likewise.
1213 (vreinterpretq_s16_f64): Likewise.
1214 (vreinterpret_s32_f64): Likewise.
1215 (vreinterpretq_s32_f64): Likewise.
1216 (vreinterpret_u8_f64): Likewise.
1217 (vreinterpretq_u8_f64): Likewise.
1218 (vreinterpret_u16_f64): Likewise.
1219 (vreinterpretq_u16_f64): Likewise.
1220 (vreinterpret_u32_f64): Likewise.
1221 (vreinterpretq_u32_f64): Likewise.
1223 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1225 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1226 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1227 (vreinterpret_p8_s8): Likewise.
1228 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1229 (vreinterpret_p8_s16): Likewise.
1230 (vreinterpret_p8_s32): Likewise.
1231 (vreinterpret_p8_s64): Likewise.
1232 (vreinterpret_p8_f32): Likewise.
1233 (vreinterpret_p8_u8): Likewise.
1234 (vreinterpret_p8_u16): Likewise.
1235 (vreinterpret_p8_u32): Likewise.
1236 (vreinterpret_p8_u64): Likewise.
1237 (vreinterpret_p8_p16): Likewise.
1238 (vreinterpretq_p8_s8): Likewise.
1239 (vreinterpretq_p8_s16): Likewise.
1240 (vreinterpretq_p8_s32): Likewise.
1241 (vreinterpretq_p8_s64): Likewise.
1242 (vreinterpretq_p8_f32): Likewise.
1243 (vreinterpretq_p8_u8): Likewise.
1244 (vreinterpretq_p8_u16): Likewise.
1245 (vreinterpretq_p8_u32): Likewise.
1246 (vreinterpretq_p8_u64): Likewise.
1247 (vreinterpretq_p8_p16): Likewise.
1248 (vreinterpret_p16_s8): Likewise.
1249 (vreinterpret_p16_s16): Likewise.
1250 (vreinterpret_p16_s32): Likewise.
1251 (vreinterpret_p16_s64): Likewise.
1252 (vreinterpret_p16_f32): Likewise.
1253 (vreinterpret_p16_u8): Likewise.
1254 (vreinterpret_p16_u16): Likewise.
1255 (vreinterpret_p16_u32): Likewise.
1256 (vreinterpret_p16_u64): Likewise.
1257 (vreinterpret_p16_p8): Likewise.
1258 (vreinterpretq_p16_s8): Likewise.
1259 (vreinterpretq_p16_s16): Likewise.
1260 (vreinterpretq_p16_s32): Likewise.
1261 (vreinterpretq_p16_s64): Likewise.
1262 (vreinterpretq_p16_f32): Likewise.
1263 (vreinterpretq_p16_u8): Likewise.
1264 (vreinterpretq_p16_u16): Likewise.
1265 (vreinterpretq_p16_u32): Likewise.
1266 (vreinterpretq_p16_u64): Likewise.
1267 (vreinterpretq_p16_p8): Likewise.
1268 (vreinterpret_f32_s8): Likewise.
1269 (vreinterpret_f32_s16): Likewise.
1270 (vreinterpret_f32_s32): Likewise.
1271 (vreinterpret_f32_s64): Likewise.
1272 (vreinterpret_f32_u8): Likewise.
1273 (vreinterpret_f32_u16): Likewise.
1274 (vreinterpret_f32_u32): Likewise.
1275 (vreinterpret_f32_u64): Likewise.
1276 (vreinterpret_f32_p8): Likewise.
1277 (vreinterpret_f32_p16): Likewise.
1278 (vreinterpretq_f32_s8): Likewise.
1279 (vreinterpretq_f32_s16): Likewise.
1280 (vreinterpretq_f32_s32): Likewise.
1281 (vreinterpretq_f32_s64): Likewise.
1282 (vreinterpretq_f32_u8): Likewise.
1283 (vreinterpretq_f32_u16): Likewise.
1284 (vreinterpretq_f32_u32): Likewise.
1285 (vreinterpretq_f32_u64): Likewise.
1286 (vreinterpretq_f32_p8): Likewise.
1287 (vreinterpretq_f32_p16): Likewise.
1288 (vreinterpret_s64_s8): Likewise.
1289 (vreinterpret_s64_s16): Likewise.
1290 (vreinterpret_s64_s32): Likewise.
1291 (vreinterpret_s64_f32): Likewise.
1292 (vreinterpret_s64_u8): Likewise.
1293 (vreinterpret_s64_u16): Likewise.
1294 (vreinterpret_s64_u32): Likewise.
1295 (vreinterpret_s64_u64): Likewise.
1296 (vreinterpret_s64_p8): Likewise.
1297 (vreinterpret_s64_p16): Likewise.
1298 (vreinterpretq_s64_s8): Likewise.
1299 (vreinterpretq_s64_s16): Likewise.
1300 (vreinterpretq_s64_s32): Likewise.
1301 (vreinterpretq_s64_f32): Likewise.
1302 (vreinterpretq_s64_u8): Likewise.
1303 (vreinterpretq_s64_u16): Likewise.
1304 (vreinterpretq_s64_u32): Likewise.
1305 (vreinterpretq_s64_u64): Likewise.
1306 (vreinterpretq_s64_p8): Likewise.
1307 (vreinterpretq_s64_p16): Likewise.
1308 (vreinterpret_u64_s8): Likewise.
1309 (vreinterpret_u64_s16): Likewise.
1310 (vreinterpret_u64_s32): Likewise.
1311 (vreinterpret_u64_s64): Likewise.
1312 (vreinterpret_u64_f32): Likewise.
1313 (vreinterpret_u64_u8): Likewise.
1314 (vreinterpret_u64_u16): Likewise.
1315 (vreinterpret_u64_u32): Likewise.
1316 (vreinterpret_u64_p8): Likewise.
1317 (vreinterpret_u64_p16): Likewise.
1318 (vreinterpretq_u64_s8): Likewise.
1319 (vreinterpretq_u64_s16): Likewise.
1320 (vreinterpretq_u64_s32): Likewise.
1321 (vreinterpretq_u64_s64): Likewise.
1322 (vreinterpretq_u64_f32): Likewise.
1323 (vreinterpretq_u64_u8): Likewise.
1324 (vreinterpretq_u64_u16): Likewise.
1325 (vreinterpretq_u64_u32): Likewise.
1326 (vreinterpretq_u64_p8): Likewise.
1327 (vreinterpretq_u64_p16): Likewise.
1328 (vreinterpret_s8_s16): Likewise.
1329 (vreinterpret_s8_s32): Likewise.
1330 (vreinterpret_s8_s64): Likewise.
1331 (vreinterpret_s8_f32): Likewise.
1332 (vreinterpret_s8_u8): Likewise.
1333 (vreinterpret_s8_u16): Likewise.
1334 (vreinterpret_s8_u32): Likewise.
1335 (vreinterpret_s8_u64): Likewise.
1336 (vreinterpret_s8_p8): Likewise.
1337 (vreinterpret_s8_p16): Likewise.
1338 (vreinterpretq_s8_s16): Likewise.
1339 (vreinterpretq_s8_s32): Likewise.
1340 (vreinterpretq_s8_s64): Likewise.
1341 (vreinterpretq_s8_f32): Likewise.
1342 (vreinterpretq_s8_u8): Likewise.
1343 (vreinterpretq_s8_u16): Likewise.
1344 (vreinterpretq_s8_u32): Likewise.
1345 (vreinterpretq_s8_u64): Likewise.
1346 (vreinterpretq_s8_p8): Likewise.
1347 (vreinterpretq_s8_p16): Likewise.
1348 (vreinterpret_s16_s8): Likewise.
1349 (vreinterpret_s16_s32): Likewise.
1350 (vreinterpret_s16_s64): Likewise.
1351 (vreinterpret_s16_f32): Likewise.
1352 (vreinterpret_s16_u8): Likewise.
1353 (vreinterpret_s16_u16): Likewise.
1354 (vreinterpret_s16_u32): Likewise.
1355 (vreinterpret_s16_u64): Likewise.
1356 (vreinterpret_s16_p8): Likewise.
1357 (vreinterpret_s16_p16): Likewise.
1358 (vreinterpretq_s16_s8): Likewise.
1359 (vreinterpretq_s16_s32): Likewise.
1360 (vreinterpretq_s16_s64): Likewise.
1361 (vreinterpretq_s16_f32): Likewise.
1362 (vreinterpretq_s16_u8): Likewise.
1363 (vreinterpretq_s16_u16): Likewise.
1364 (vreinterpretq_s16_u32): Likewise.
1365 (vreinterpretq_s16_u64): Likewise.
1366 (vreinterpretq_s16_p8): Likewise.
1367 (vreinterpretq_s16_p16): Likewise.
1368 (vreinterpret_s32_s8): Likewise.
1369 (vreinterpret_s32_s16): Likewise.
1370 (vreinterpret_s32_s64): Likewise.
1371 (vreinterpret_s32_f32): Likewise.
1372 (vreinterpret_s32_u8): Likewise.
1373 (vreinterpret_s32_u16): Likewise.
1374 (vreinterpret_s32_u32): Likewise.
1375 (vreinterpret_s32_u64): Likewise.
1376 (vreinterpret_s32_p8): Likewise.
1377 (vreinterpret_s32_p16): Likewise.
1378 (vreinterpretq_s32_s8): Likewise.
1379 (vreinterpretq_s32_s16): Likewise.
1380 (vreinterpretq_s32_s64): Likewise.
1381 (vreinterpretq_s32_f32): Likewise.
1382 (vreinterpretq_s32_u8): Likewise.
1383 (vreinterpretq_s32_u16): Likewise.
1384 (vreinterpretq_s32_u32): Likewise.
1385 (vreinterpretq_s32_u64): Likewise.
1386 (vreinterpretq_s32_p8): Likewise.
1387 (vreinterpretq_s32_p16): Likewise.
1388 (vreinterpret_u8_s8): Likewise.
1389 (vreinterpret_u8_s16): Likewise.
1390 (vreinterpret_u8_s32): Likewise.
1391 (vreinterpret_u8_s64): Likewise.
1392 (vreinterpret_u8_f32): Likewise.
1393 (vreinterpret_u8_u16): Likewise.
1394 (vreinterpret_u8_u32): Likewise.
1395 (vreinterpret_u8_u64): Likewise.
1396 (vreinterpret_u8_p8): Likewise.
1397 (vreinterpret_u8_p16): Likewise.
1398 (vreinterpretq_u8_s8): Likewise.
1399 (vreinterpretq_u8_s16): Likewise.
1400 (vreinterpretq_u8_s32): Likewise.
1401 (vreinterpretq_u8_s64): Likewise.
1402 (vreinterpretq_u8_f32): Likewise.
1403 (vreinterpretq_u8_u16): Likewise.
1404 (vreinterpretq_u8_u32): Likewise.
1405 (vreinterpretq_u8_u64): Likewise.
1406 (vreinterpretq_u8_p8): Likewise.
1407 (vreinterpretq_u8_p16): Likewise.
1408 (vreinterpret_u16_s8): Likewise.
1409 (vreinterpret_u16_s16): Likewise.
1410 (vreinterpret_u16_s32): Likewise.
1411 (vreinterpret_u16_s64): Likewise.
1412 (vreinterpret_u16_f32): Likewise.
1413 (vreinterpret_u16_u8): Likewise.
1414 (vreinterpret_u16_u32): Likewise.
1415 (vreinterpret_u16_u64): Likewise.
1416 (vreinterpret_u16_p8): Likewise.
1417 (vreinterpret_u16_p16): Likewise.
1418 (vreinterpretq_u16_s8): Likewise.
1419 (vreinterpretq_u16_s16): Likewise.
1420 (vreinterpretq_u16_s32): Likewise.
1421 (vreinterpretq_u16_s64): Likewise.
1422 (vreinterpretq_u16_f32): Likewise.
1423 (vreinterpretq_u16_u8): Likewise.
1424 (vreinterpretq_u16_u32): Likewise.
1425 (vreinterpretq_u16_u64): Likewise.
1426 (vreinterpretq_u16_p8): Likewise.
1427 (vreinterpretq_u16_p16): Likewise.
1428 (vreinterpret_u32_s8): Likewise.
1429 (vreinterpret_u32_s16): Likewise.
1430 (vreinterpret_u32_s32): Likewise.
1431 (vreinterpret_u32_s64): Likewise.
1432 (vreinterpret_u32_f32): Likewise.
1433 (vreinterpret_u32_u8): Likewise.
1434 (vreinterpret_u32_u16): Likewise.
1435 (vreinterpret_u32_u64): Likewise.
1436 (vreinterpret_u32_p8): Likewise.
1437 (vreinterpret_u32_p16): Likewise.
1438 (vreinterpretq_u32_s8): Likewise.
1439 (vreinterpretq_u32_s16): Likewise.
1440 (vreinterpretq_u32_s32): Likewise.
1441 (vreinterpretq_u32_s64): Likewise.
1442 (vreinterpretq_u32_f32): Likewise.
1443 (vreinterpretq_u32_u8): Likewise.
1444 (vreinterpretq_u32_u16): Likewise.
1445 (vreinterpretq_u32_u64): Likewise.
1446 (vreinterpretq_u32_p8): Likewise.
1447 (vreinterpretq_u32_p16): Likewise.
1449 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1451 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1453 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1456 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1457 (vqnegd_s64): Likewise.
1458 (vqabs_s64): Likewise.
1459 (vqabsd_s64): Likewise.
1461 2014-04-22 Richard Henderson <rth@redhat.com>
1463 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1464 computation to the top of the loop.
1466 2014-04-22 Renlin <renlin.li@arm.com>
1467 Jiong Wang <jiong.wang@arm.com>
1469 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1470 * config/aarch64/aarch64.c (aarch64_layout_frame)
1471 (aarch64_initial_elimination_offset): Likewise.
1473 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1475 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1478 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1480 * machmode.h (bitwise_mode_for_mode): Declare.
1481 * stor-layout.h (bitwise_type_for_mode): Likewise.
1482 * stor-layout.c (bitwise_mode_for_mode): New function.
1483 (bitwise_type_for_mode): Likewise.
1484 * builtins.c (fold_builtin_memory_op): Use it instead of
1485 int_mode_for_mode and build_nonstandard_integer_type.
1487 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1489 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1490 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1491 (*-*-solaris2*): Simplify.
1492 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1493 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1494 *-*-solaris2.9* handling.
1496 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1498 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1499 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1501 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1502 * configure: Regenerate.
1504 * config/i386/sol2-9.h: Remove.
1506 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1507 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1508 Remove Solaris 9 references.
1510 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1512 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1513 (floatuns<GPI:mode><GPF:mode>2): Remove.
1514 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1515 and floatuns conversions.
1516 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1517 and floatuns conversions.
1518 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1519 (w1,w2): New mode attributes for inequal width conversions.
1521 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1523 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1524 the output asm format.
1526 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1528 * config/aarch64/aarch64-simd.md
1529 (aarch64_cm<optab>di): Always split.
1530 (*aarch64_cm<optab>di): New.
1531 (aarch64_cmtstdi): Always split.
1532 (*aarch64_cmtstdi): New.
1534 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1536 PR tree-optimization/60823
1537 * omp-low.c (ipa_simd_modify_function_body): Go through
1538 all SSA_NAMEs and for those refering to vector arguments
1539 which are going to be replaced adjust SSA_NAME_VAR and,
1540 if it is a default definition, change it into a non-default
1541 definition assigned at the beginning of function from new_decl.
1542 (ipa_simd_modify_stmt_ops): Rewritten.
1543 * tree-dfa.c (set_ssa_default_def): When removing default def,
1544 check for NULL loc instead of NULL *loc.
1546 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1548 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1549 restrictions on core registers for DImode values in Thumb2.
1551 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1553 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1554 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1556 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1558 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1559 (*iordi_notzesidi_di): Likewise.
1560 (*iordi_notsesidi_di): Likewise.
1562 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1564 * config/arm/arm-protos.h (tune_params): New struct members.
1565 * config/arm/arm.c: Initialise tune_params per processor.
1566 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1567 for speed, based on new tune_params.
1569 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1571 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1573 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1575 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1577 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1578 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1579 (vrnda_f64): Likewise.
1580 (vrndi_f64): Likewise.
1581 (vrndm_f64): Likewise.
1582 (vrndn_f64): Likewise.
1583 (vrndp_f64): Likewise.
1584 (vrndx_f64): Likewise.
1586 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1588 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1589 GET_MODE_SIZE argument is enum machine_mode.
1591 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1594 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1595 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1597 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1600 * asan.c (asan_emit_stack_protection): Force the base to align to
1601 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1602 appropriate bits if STRICT_ALIGNMENT.
1603 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1605 (expand_used_vars): Leave a space in the stack frame for alignment
1606 if STRICT_ALIGNMENT.
1608 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1610 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1612 (gimple_store_p): Likewise.
1613 (gimple_assign_load_p): Likewise.
1614 (gimple_assign_cast_p): Likewise.
1615 (gimple_clobber_p): Likewise.
1617 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1618 rather than a gimple.
1619 (gimple_assign_cast_p): Likewise.
1621 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1624 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1625 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1627 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1628 more debug information for E500 if -mdebug=reg.
1630 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1633 * config/i386/i386.c (ix86_expand_builtin)
1634 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1635 register for target RTX.
1636 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1638 2014-04-18 Cong Hou <congh@google.com>
1640 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1641 the widen-mult pattern by handling two operands with different sizes,
1642 and operands whose size is smaller than half of the result type.
1644 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1646 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1647 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1648 (do_estimate_edge_time): Compute it.
1649 * ipa-inline.c (want_inline_small_function_p): Bypass
1650 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1652 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1654 * ipa-inline.c (spec_rem): New static variable.
1655 (dump_overall_stats): New function.
1656 (dump_inline_stats): New function.
1658 2014-04-18 Richard Henderson <rth@redhat.com>
1660 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1661 to GET_MODE_SIZE, not a reg_class_t.
1663 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1665 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1666 (vsx_xxmrglw_<mode>): Likewise.
1668 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1671 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1672 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1673 (rs6000_init_hard_regno_mode_ok): Likewise.
1675 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1677 * ipa-inline.c (inline_small_functions): Account only non-cold
1679 * doc/invoke.texi (inline-unit-growth): Update documentation.
1681 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1683 * config/rs6000/rs6000.md (addti3, subti3): New.
1685 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1688 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1689 comment. Check optimize_insn_for_size_p instead of
1690 optimize_insn_for_speed_p.
1692 2014-04-17 Martin Jambor <mjambor@suse.cz>
1694 * gimple-iterator.c (gsi_start_edge): New function.
1695 * gimple-iterator.h (gsi_start_edge): Declare.
1696 * tree-sra.c (single_non_eh_succ): New function.
1697 (disqualify_ops_if_throwing_stmt): Renamed to
1698 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1699 having one non-EH successor BB.
1700 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1701 generate loads into replacements.
1702 (sra_modify_assign): Likewise and and also use the simple path for
1704 (sra_modify_function_body): Commit statements on edges.
1706 2014-04-17 Richard Biener <rguenther@suse.de>
1709 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1710 comparison results and add clarifying comment.
1712 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1714 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1715 (blank_mode): Initialize it.
1716 (emit_mode_size_inline, emit_mode_nunits_inline,
1717 emit_mode_inner_inline): New functions.
1718 (emit_insn_modes_h): Call them and surround their output with
1719 #if GCC_VERSION >= 4001 ... #endif.
1720 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1721 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1722 mode_* arrays if the argument is __builtin_constant_p.
1723 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1724 is enum machine_mode.
1726 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1728 * passes.c (opt_pass::execute): Adjust.
1729 (pass_manager::execute_pass_mode_switching): Likewise.
1730 (early_local_passes::execute): Likewise.
1731 (execute_one_pass): Pass cfun to the pass's execute method.
1732 * tree-pass.h (opt_pass::execute): Add function * argument.
1733 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1734 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1735 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1736 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1737 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1738 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1739 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1740 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1741 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1742 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1743 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1744 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1745 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1746 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1747 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1748 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1749 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1750 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1751 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1752 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1753 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1754 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1755 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1756 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1757 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1758 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1759 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1760 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1761 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1764 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1766 * passes.c (opt_pass::gate): Take function * argument.
1767 (gate_all_early_local_passes): Merge into
1768 (early_local_passes::gate): this.
1769 (gate_all_early_optimizations): Merge into
1770 (all_early_optimizations::gate): this.
1771 (gate_all_optimizations): Mege into
1772 (all_optimizations::gate): this.
1773 (gate_all_optimizations_g): Merge into
1774 (all_optimizations_g::gate): this.
1775 (gate_rest_of_compilation): Mege into
1776 (rest_of_compilation::gate): this.
1777 (gate_postreload): Merge into
1778 (postreload::gate): this.
1779 (dump_one_pass): Pass cfun to the pass's gate method.
1780 (execute_ipa_summary_passes): Likewise.
1781 (execute_one_pass): Likewise.
1782 (ipa_write_summaries_2): Likewise.
1783 (ipa_write_optimization_summaries_1): Likewise.
1784 (ipa_read_summaries_1): Likewise.
1785 (ipa_read_optimization_summaries_1): Likewise.
1786 (execute_ipa_stmt_fixups): Likewise.
1787 * tree-pass.h (opt_pass::gate): Add function * argument.
1788 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
1789 combine-stack-adj.c, combine.c, compare-elim.c,
1790 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1791 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
1792 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
1793 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
1794 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1795 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1796 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1797 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
1798 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
1799 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
1800 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1801 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1802 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
1803 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1804 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1805 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1806 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1807 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1808 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1809 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1810 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1811 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1812 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1813 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
1814 var-tracking.c, vtable-verify.c, web.c: Adjust.
1816 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1818 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
1819 * configure: Regenerate.
1821 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1823 * passes.c (dump_one_pass): don't check pass->has_gate.
1824 (execute_ipa_summary_passes): Likewise.
1825 (execute_one_pass): Likewise.
1826 (ipa_write_summaries_2): Likewise.
1827 (ipa_write_optimization_summaries_1): Likewise.
1828 (ipa_read_optimization_summaries_1): Likewise.
1829 (execute_ipa_stmt_fixups): Likewise.
1830 * tree-pass.h (pass_data::has_gate): Remove.
1831 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1832 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1833 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1834 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1835 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1836 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
1837 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
1838 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
1839 gimple-low.c, gimple-ssa-isolate-paths.c,
1840 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
1841 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
1842 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1843 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
1844 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
1845 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
1846 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
1847 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1848 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
1849 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1850 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1851 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1852 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1853 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1854 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1855 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1856 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1857 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1858 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1859 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1860 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1861 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1864 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1866 * pass_manager.h (pass_manager::register_dump_files_1): Remove
1868 * passes.c (pass_manager::register_dump_files_1): Merge into
1869 (pass_manager::register_dump_files): this, and remove its handling of
1870 properties since the pass always has the properties anyway.
1871 (pass_manager::pass_manager): Adjust.
1873 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1875 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
1876 * passes.c (pass_manager::register_dump_files_1): Remove dead code
1877 dealing with properties.
1878 (pass_manager::register_dump_files): Adjust.
1880 2014-03-20 Mark Wielaard <mjw@redhat.com>
1882 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
1883 then represent the bound as normal constant value.
1885 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1888 Forward port from 4.8 branch
1889 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
1891 * config/i386/bmiintrin.h (_blsi_u32): New.
1895 (_blsmsk_u32): Ditto.
1896 (_blsmsk_u64): Ditto.
1897 (_tzcnt_u32): Ditto.
1898 (_tzcnt_u64): Ditto.
1900 2014-04-17 Kito Cheng <kito@0xlab.org>
1902 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
1904 2014-04-17 Richard Biener <rguenther@suse.de>
1907 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
1908 boolean results for comparisons.
1910 2014-04-17 Richard Biener <rguenther@suse.de>
1912 PR tree-optimization/60836
1913 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
1914 initial PHI args to be gimple values.
1916 2014-04-17 Richard Biener <rguenther@suse.de>
1918 PR tree-optimization/60841
1919 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
1920 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
1921 of stmts to SLP build.
1922 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
1923 (vect_analyze_slp): Likewise.
1924 (vect_analyze_slp_instance): Likewise.
1925 (vect_build_slp_tree): Limit overall SLP tree growth.
1926 * tree-vectorizer.h (vect_analyze_data_refs,
1927 vect_analyze_slp): Adjust prototypes.
1929 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1931 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
1934 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1936 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
1937 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
1938 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
1939 for TARGET_SLOW_PSHUFB
1941 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1943 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
1944 * config/i386/i386.c (intel_cost): Ditto.
1946 2014-04-17 Joey Ye <joey.ye@arm.com>
1948 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
1950 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1952 * opts.c (common_handle_option): Disable -fipa-reference coorectly
1955 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1957 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
1958 (type_all_derivations_known_p): New predicate.
1959 (type_all_ctors_visible_p): New predicate.
1960 (type_possibly_instantiated_p): New predicate.
1961 (get_odr_type): Compute all_derivations_known.
1962 (dump_odr_type): Dump the flag.
1963 (maybe_record_type): Cleanup.
1964 (record_target_from_binfo): Add bases_to_consider array;
1965 record bases for types w/o instances and skip CXX destructor.
1966 (possible_polymorphic_call_targets_1): Add bases_to_consider
1967 and consider_construction parameters; check if type may have instance.
1968 (get_polymorphic_call_info): Set maybe_in_construction to true
1969 when we know nothing.
1970 (record_targets_from_bases): Skip CXX destructors; they are
1971 never called for types in construction.
1972 (possible_polymorphic_call_targets): Do not record target when
1973 type may not have instance.
1975 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1978 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
1979 external aliases alive, too.
1981 2014-04-16 Andrew Pinski <apinski@cavium.com>
1983 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
1986 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1988 * final.c (compute_alignments): Do not apply loop alignment to a block
1989 falling through to the exit.
1991 2014-04-16 Catherine Moore <clm@codesourcery.com>
1993 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
1994 Adjust constraints for microMIPS store patterns.
1996 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
1998 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
2000 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2002 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
2003 (append_use): Run at -O0.
2004 (append_vdef): Likewise.
2005 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
2006 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
2008 2014-04-16 Jakub Jelinek <jakub@redhat.com>
2010 PR tree-optimization/60844
2011 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
2012 (propagate_op_to_single_use, remove_visited_stmt_chain,
2013 linearize_expr, repropagate_negates, reassociate_bb): Use it
2014 instead of gsi_remove.
2016 2014-04-16 Martin Jambor <mjambor@suse.cz>
2018 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
2019 ipa_transforms_to_apply.
2020 (cgraph_function_versioning): Assert that old_node has empty
2021 ipa_transforms_to_apply.
2022 * trans-mem.c (ipa_tm_create_version): Likewise.
2023 * tree-inline.c (tree_function_versioning): Do not duplicate
2024 ipa_transforms_to_apply.
2026 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2029 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
2031 Pass necessary as flags on 64-bit Solaris/x86.
2032 Use lowercase relocs for x86_64-*-*.
2033 * configure: Regenerate.
2035 2014-04-15 Jan Hubicka <jh@suse.cz>
2037 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
2038 (maybe_record_node, likely_target_p): Use it.
2040 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2043 Revert following patch
2045 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2048 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2049 software floating point or no floating point registers, do not
2050 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2051 in GPRs that occurs after we tested for GPRs that would never be
2054 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2055 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2056 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2057 specifically allow DDmode, since that does not use the SPE SIMD
2060 2014-03-21 Mark Wielaard <mjw@redhat.com>
2062 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
2063 as unsigned or int depending on type and value used.
2065 2014-04-15 Richard Biener <rguenther@suse.de>
2067 PR rtl-optimization/56965
2068 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
2069 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
2071 * alias.c (true_dependence_1): Do not call
2072 nonoverlapping_component_refs_p.
2073 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
2074 nonoverlapping_component_refs_p.
2075 (indirect_refs_may_alias_p): Likewise.
2077 2014-04-15 Teresa Johnson <tejohnson@google.com>
2079 * cfg.c (dump_bb_info): Fix flags check.
2080 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
2082 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2084 PR rtl-optimization/60663
2085 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
2088 2014-04-15 Richard Biener <rguenther@suse.de>
2090 * lto-streamer.h (LTO_major_version): Bump to 4.
2092 2014-04-15 Richard Biener <rguenther@suse.de>
2094 * common.opt (lto_partition_model): New enum.
2095 (flto-partition=): Merge separate options with a single with argument,
2096 add -flto-partition=one support.
2097 * flag-types.h (enum lto_partition_model): Declare.
2098 * opts.c (finish_options): Remove duplicate -flto-partition=
2100 * lto-wrapper.c (run_gcc): Adjust.
2102 2014-04-15 Richard Biener <rguenther@suse.de>
2104 * alias.c (ncr_compar): New function.
2105 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2107 2014-04-15 Richard Biener <rguenther@suse.de>
2109 * alias.c (record_component_aliases): Do not walk BINFOs.
2111 2014-04-15 Richard Biener <rguenther@suse.de>
2113 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2114 Add struct function argument and adjust.
2115 (find_func_aliases_for_call): Likewise.
2116 (find_func_aliases): Likewise.
2117 (find_func_clobbers): Likewise.
2118 (intra_create_variable_infos): Likewise.
2119 (compute_points_to_sets): Likewise.
2120 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2122 2014-04-15 Richard Biener <rguenther@suse.de>
2124 * tree.c (iterative_hash_expr): Use enum tree_code_class
2125 to store TREE_CODE_CLASS.
2126 (tree_block): Likewise.
2127 (tree_set_block): Likewise.
2128 * tree.h (fold_build_pointer_plus_loc): Use
2129 convert_to_ptrofftype_loc.
2131 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2134 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2137 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2139 * cfgloop.h (struct loop): Move force_vectorize down.
2140 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2141 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2142 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2143 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2144 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2145 * tree-core.h (enum annot_expr_kind): Add new kind values.
2146 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2147 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2149 * tree.def (ANNOTATE_EXPR): Tweak comment.
2151 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2153 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2156 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2158 * tree.h (TYPE_IDENTIFIER): Declare.
2159 * tree.c (subrange_type_for_debug_p): Use it.
2160 * godump.c (go_format_type): Likewise.
2161 * dwarf2out.c (is_cxx_auto, modified_type_die,
2162 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2163 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2165 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2168 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2170 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2172 * config/i386/i386.c (examine_argument): Return bool. Return true if
2173 parameter should be passed in memory.
2174 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2175 (construct_container): Update calls to examine_argument.
2176 (function_arg_advance_64): Ditto.
2177 (return_in_memory_32): Merge with ix86_return_in_memory.
2178 (return_in_memory_64): Ditto.
2179 (return_in_memory_ms_64): Ditto.
2181 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2183 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2184 * coverage.c (coverage_compute_profile_id): Handle externally visible
2187 2014-04-14 Martin Jambor <mjambor@suse.cz>
2189 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2190 DECL_DISREGARD_INLINE_LIMITS functions.
2192 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2195 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2197 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2200 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2201 optimize_insn_for_speed_p instead of
2202 optimize_function_for_speed_p.
2204 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2206 * doc/invoke.texi (free): Document AArch64.
2208 2014-04-14 Richard Biener <rguenther@suse.de>
2210 PR tree-optimization/60042
2211 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2212 (insert_into_preds_of_block): Do not prevent PHI insertion
2213 for REFERENCE exprs here ...
2214 (eliminate_dom_walker::before_dom_children): ... but prevent
2215 their use here under similar conditions when applied to the
2216 IL after PRE optimizations.
2218 2014-04-14 Richard Biener <rguenther@suse.de>
2220 * passes.def: Move early points-to after early SRA.
2222 2014-04-14 Richard Biener <rguenther@suse.de>
2224 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2225 check for which sign-changes we allow when forwarding
2226 a converted value into a switch.
2228 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2230 * stor-layout.c (place_field): Finalize non-constant offset for the
2233 2014-04-14 Richard Biener <rguenther@suse.de>
2235 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2237 (expand_switch_using_bit_tests_p): Likewise.
2238 (process_switch): Compute and pass on speed_p based on the
2240 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2241 optimize_bb_for_speed_p.
2243 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2245 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2246 * function.h (struct function): Rename has_force_vect_loops into
2247 has_force_vectorize_loops.
2248 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2249 (input_struct_function_base): Likewise.
2250 * lto-streamer-out.c (output_cfg): Likewise.
2251 (output_struct_function_base): Likewise.
2252 * omp-low.c (expand_omp_simd): Likewise.
2253 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2254 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2255 (version_loop_for_if_conversion): Likewise.
2256 (tree_if_conversion): Likewise.
2257 (main_tree_if_conversion): Likewise.
2258 (gate_tree_if_conversion): Likewise.
2259 * tree-inline.c (copy_loops): Likewise.
2260 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2261 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2262 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2263 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2264 * tree-vectorizer.c (vectorize_loops): Likewise.
2265 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2267 2014-04-14 Richard Biener <rguenther@suse.de>
2270 * lto-streamer-out.c (wrap_refs): New function.
2271 (lto_output): Wrap symbol references in global initializes in
2272 type-preserving MEM_REFs.
2274 2014-04-14 Christian Bruel <christian.bruel@st.com>
2276 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2278 2014-04-14 Christian Bruel <christian.bruel@st.com>
2280 * config/sh/sh.md (setmemqi): New expand pattern.
2281 * config/sh/sh.h (CLEAR_RATIO): Define.
2282 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2283 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2285 2014-04-14 Richard Biener <rguenther@suse.de>
2288 * fold-const.c (negate_expr_p): Don't negate directional rounding
2290 (fold_negate_expr): Likewise.
2292 2014-04-14 Richard Biener <rguenther@suse.de>
2294 PR tree-optimization/59817
2295 PR tree-optimization/60453
2296 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2297 recursion to catch all CHRECs in the scalar evolution and restrict
2298 the predicate for the remains appropriately.
2300 2014-04-12 Catherine Moore <clm@codesourcery.com>
2302 * config/mips/constraints.md: Add new register constraint "kb".
2303 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2304 (*movhi_internal): Likewise.
2305 (*movqi_internal): Likewise.
2306 * config/mips/mips.h (M16_STORE_REGS): New register class.
2307 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2308 (REG_CLASS_CONTENTS): Likewise.
2309 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2311 2014-04-11 Tobias Burnus <burnus@net-b.de>
2314 * doc/invoke.texi (-Wformat-signedness): Document it.
2315 (Wformat=2): Mention that this enables -Wformat-signedness.
2317 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2319 * common/config/epiphany/epiphany-common.c
2320 (epiphany_option_optimization_table): Enable section anchors by
2321 default at -O1 or higher.
2322 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2323 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2324 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2325 carries no extra cost.
2326 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2327 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2328 * config/epiphany/predicates.md (memclob_operand): New predicate.
2329 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2330 Use memclob_operand predicate and X constraint for operand 3.
2332 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2334 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2335 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2338 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2340 PR rtl-optimization/60651
2341 * mode-switching.c (optimize_mode_switching): Make sure to emit
2342 sets of a lower numbered entity before sets of a higher numbered
2343 entity to a mode of the same or lower priority.
2344 When creating a seginfo for a basic block that starts with a code
2345 label, move the insertion point past the code label.
2346 (new_seginfo): Document and enforce requirement that
2347 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2348 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2349 * doc/tm.texi: Regenerate.
2351 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2354 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2356 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2358 * BASE-VER: Set to 4.10.0.
2360 2014-04-11 Tobias Burnus <burnus@net-b.de>
2363 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2364 * doc/gcc.texi (Service): Update description in the @menu
2365 * doc/invoke.texi (Option Summary): Remove misplaced and
2368 2014-04-11 Steve Ellcey <sellcey@mips.com>
2369 Jakub Jelinek <jakub@redhat.com>
2372 * expr.c (convert_move): Use emit_store_flag_force instead of
2373 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2376 2014-04-11 Richard Biener <rguenther@suse.de>
2379 * varasm.c (assemble_alias): Avoid endless error reporting
2380 recursion by setting TREE_ASM_WRITTEN.
2382 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2384 * config/s390/s390.md: Add a splitter for NOT rtx.
2386 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2388 PR rtl-optimization/60663
2389 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2391 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2392 Jakub Jelinek <jakub@redhat.com>
2395 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2396 flag from decl_node to node.
2398 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2401 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2402 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2403 ameliorating the cases where it can be.
2405 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2408 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2410 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2411 (loadsync_<mode>): Change mode.
2412 (load_quadpti, store_quadpti): New.
2413 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2414 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2415 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2417 2014-04-09 Cong Hou <congh@google.com>
2420 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2423 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2425 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2426 instead of vnor to exploit possible fusion opportunity in the
2428 (altivec_expand_vec_perm_const_le): Likewise.
2430 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2432 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2433 (loadsync_<mode>): Change mode.
2434 (load_quadpti, store_quadpti): New.
2435 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2436 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2438 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2441 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2442 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2443 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2445 2014-04-08 Richard Biener <rguenther@suse.de>
2448 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2449 a 64bit widest int print double-int similar to on HWI64 hosts.
2451 2014-04-08 Richard Biener <rguenther@suse.de>
2453 PR tree-optimization/60785
2454 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2455 default defs properly.
2457 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2459 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2460 (Weffc++): Likewise.
2462 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2464 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2465 set completep to false rather than true.
2467 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2470 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2471 ARM_TARGET2_DWARF_FORMAT.
2473 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2476 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2477 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2480 2014-04-07 Richard Biener <rguenther@suse.de>
2482 PR tree-optimization/60766
2483 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2484 (may_eliminate_iv): Convert cand_value_at result to desired type.
2486 2014-04-07 Jason Merrill <jason@redhat.com>
2489 * common.opt (-fno-gnu-unique): Add.
2490 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2492 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2494 * haifa-sched.c: Fix outdated function reference and minor
2495 grammar errors in introductory comment.
2497 2014-04-07 Richard Biener <rguenther@suse.de>
2500 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2502 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2504 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2507 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2508 size accounting for thunks.
2509 (pa_asm_output_mi_thunk): Use final_start_function() and
2510 final_end_function() to output function start and end directives.
2512 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2514 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2515 device specific ISA/ feature information. Remove short_sp and
2516 errata_skip ds. Add avr_device_specific_features enum to have device
2518 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2519 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2520 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2521 updated device specific info.
2522 * config/avr/avr-mcus.def: Merge device specific details to
2523 dev_attribute field.
2524 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2526 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2527 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2528 assembler if RMW isa supported by current device.
2529 * config/avr/genmultilib.awk: Update as device info structure changed.
2530 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2532 2014-04-04 Cong Hou <congh@google.com>
2534 PR tree-optimization/60656
2535 * tree-vect-stmts.c (supportable_widening_operation):
2536 Fix a bug that elements in a vector with vect_used_by_reduction
2537 property are incorrectly reordered when the operation on it is not
2538 consistant with the one in reduction operation.
2540 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2542 PR rtl-optimization/60155
2543 * gcse.c (record_set_data): New function.
2544 (single_set_gcse): New function.
2545 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2546 (hoist_code): Likewise.
2547 (get_pressure_class_and_nregs): Likewise.
2549 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2551 * explow.c (probe_stack_range): Emit a final optimization blockage.
2553 2014-04-04 Anthony Green <green@moxielogic.com>
2555 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2558 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2561 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2562 flags are set only during streaming.
2563 * ipa.c (process_references, walk_polymorphic_call_targets,
2564 symtab_remove_unreachable_nodes): Drop bodies of always inline
2565 after early inlining.
2566 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2568 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2569 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2572 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2575 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2578 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2580 (cortex_a53_fdivd): Likewise.
2582 2014-04-04 Martin Jambor <mjambor@suse.cz>
2585 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2587 * cgraph.c (clone_of_p): Also return true if thunks match.
2588 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2589 cgraph_function_or_thunk_node and an obsolete comment.
2590 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2592 (build_function_decl_skip_args): Likewise.
2593 (set_new_clone_decl_and_node_flags): New function.
2594 (duplicate_thunk_for_node): Likewise.
2595 (redirect_edge_duplicating_thunks): Likewise.
2596 (cgraph_clone_node): New parameter args_to_skip, pass it to
2597 redirect_edge_duplicating_thunks which is called instead of
2598 cgraph_redirect_edge_callee.
2599 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2600 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2602 2014-04-04 Jeff Law <law@redhat.com>
2605 * config/arm/predicates.md (const_int_I_operand): New predicate.
2606 (const_int_M_operand): Similarly.
2607 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2609 (insv_t2, extv_reg, extzv_t2): Likewise.
2610 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2611 (pop_multiple_with_writeback_and_return): Likewise.
2612 (vfp_pop_multiple_with_writeback): Likewise
2614 2014-04-04 Richard Biener <rguenther@suse.de>
2617 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2618 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2620 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2621 * gimplify.c (gimple_add_tmp_var_fn): New function.
2622 * gimple-expr.h (create_tmp_reg_fn): Declare.
2623 * gimple-expr.c (create_tmp_reg_fn): New function.
2624 * gimple-low.c (record_vars_into): Don't change cfun.
2625 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2626 code generation without cfun.
2628 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2631 * Makefile.in (install-driver): Fix shell scripting.
2633 2014-04-03 Cong Hou <congh@google.com>
2635 PR tree-optimization/60505
2636 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2637 threshold of number of iterations below which no vectorization
2639 * tree-vect-loop.c (new_loop_vec_info):
2640 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2641 * tree-vect-loop.c (vect_analyze_loop_operations):
2642 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2643 * tree-vect-loop.c (vect_transform_loop):
2644 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2645 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2646 of iterations of the loop and see if we should build the epilogue.
2648 2014-04-03 Richard Biener <rguenther@suse.de>
2650 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2651 (streamer_tree_cache_create): Adjust.
2652 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2653 to allow optional nodes array.
2654 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2655 (streamer_tree_cache_append): Likewise.
2656 (streamer_tree_cache_create): Create nodes array optionally
2657 as specified by parameter.
2658 * lto-streamer-out.c (create_output_block): Avoid maintaining
2659 the node array in the writer cache.
2660 (DFS_write_tree): Remove assertion.
2661 (produce_asm_for_decls): Free the out decl state hash table early.
2662 * lto-streamer-in.c (lto_data_in_create): Adjust for
2663 streamer_tree_cache_create prototype change.
2665 2014-04-03 Richard Biener <rguenther@suse.de>
2667 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2668 set TREE_CHAIN to NULL_TREE.
2670 2014-04-03 Richard Biener <rguenther@suse.de>
2672 PR tree-optimization/60740
2673 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2674 over all GIMPLE_COND operands.
2676 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2678 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2679 (Weffc++): Remove Scott's numbering, merge lists and reference
2682 2014-04-03 Nick Clifton <nickc@redhat.com>
2684 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2687 2014-04-03 Martin Jambor <mjambor@suse.cz>
2689 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2690 mention gcc_unreachable before failing.
2691 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2694 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
2697 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2698 inconsistent code and instead mark the context inconsistent.
2699 (possible_polymorphic_call_targets): For inconsistent contexts
2700 return empty complete list.
2702 2014-04-02 Anthony Green <green@moxielogic.com>
2704 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2705 (extendqisi2, extendhisi2): Define.
2706 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2707 (WCHAR_TYPE): Change to unsigned int.
2709 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2711 PR tree-optimization/60733
2712 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2713 insertion point for PHI candidates to be the end of the feeding
2714 block for the PHI argument.
2716 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2718 PR rtl-optimization/60650
2719 * lra-constraints.c (process_alt_operands): Decrease reject for
2720 earlyclobber matching.
2722 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2724 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2726 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2728 * config/spu/spu.c (pad_bb): Do not crash when the last
2729 insn is CODE_FOR_blockage.
2731 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2733 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2734 lies outside the target mode.
2736 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2739 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2740 software floating point or no floating point registers, do not
2741 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2742 in GPRs that occurs after we tested for GPRs that would never be
2745 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2746 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2747 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2748 specifically allow DDmode, since that does not use the SPE SIMD
2751 2014-04-02 Richard Biener <rguenther@suse.de>
2754 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2755 MODE_INTs. Properly use negv_optab.
2756 (expand_abs): Likewise.
2758 2014-04-02 Richard Biener <rguenther@suse.de>
2761 * Makefile.in (install-driver): Guard extra installs with special
2764 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
2766 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2769 2014-04-01 Richard Henderson <rth@redhat.com>
2772 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
2773 alternative enabled before register allocation.
2775 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
2777 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
2778 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
2780 (nios2_large_got_address): Remove unneeded 'sym' parameter.
2781 (nios2_got_address): Update nios2_large_got_address call site.
2782 (nios2_delegitimize_address): New function.
2783 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
2784 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
2785 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
2787 2014-04-01 Martin Husemann <martin@duskware.de>
2789 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
2792 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
2794 PR rtl-optimization/60604
2795 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
2796 check from register_operand.
2797 (register_operand): Redefine in terms of general_operand.
2798 (nonmemory_operand): Use register_operand for the non-constant cases.
2800 2014-04-01 Richard Biener <rguenther@suse.de>
2802 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
2804 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
2806 * doc/invoke.texi (mapp-regs): Clarify.
2808 2014-03-31 Ulrich Drepper <drepper@gmail.com>
2810 * config/i386/avx512fintrin.h (__v32hi): Define type.
2811 (__v64qi): Likewise.
2812 (_mm512_set1_epi8): Define.
2813 (_mm512_set1_epi16): Define.
2814 (_mm512_set4_epi32): Define.
2815 (_mm512_set4_epi64): Define.
2816 (_mm512_set4_pd): Define.
2817 (_mm512_set4_ps): Define.
2818 (_mm512_setr4_epi64): Define.
2819 (_mm512_setr4_epi32): Define.
2820 (_mm512_setr4_pd): Define.
2821 (_mm512_setr4_ps): Define.
2822 (_mm512_setzero_epi32): Define.
2824 2014-03-31 Martin Jambor <mjambor@suse.cz>
2827 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
2828 callsite_arguments_match_p. Updated all callers. Also check types of
2829 corresponding formal parameters and actual arguments.
2830 (not_all_callers_have_enough_arguments_p) Renamed to
2831 some_callers_have_mismatched_arguments_p.
2833 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
2835 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
2837 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
2840 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
2843 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
2845 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
2846 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
2848 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
2849 Use FMAMODE_NOVF512 mode iterator.
2850 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
2851 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
2852 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
2853 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
2855 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
2856 Use VF_128_256 mode iterator.
2857 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
2860 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2862 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
2863 static chain if needed.
2865 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2868 * lra-constraints.c (index_part_to_reg): New.
2869 (process_address): Use it.
2871 2014-03-27 Jeff Law <law@redhat.com>
2872 Jakub Jelinek <jakub@redhat.com>
2875 * expr.c (do_tablejump): Use simplify_gen_binary rather than
2876 gen_rtx_{PLUS,MULT} to build up the address expression.
2878 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
2879 creating non-canonical RTL.
2881 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2884 * ipa-inline.c (want_inline_small_function_p): Short circuit large
2885 functions; reorganize to make cheap checks first.
2886 (inline_small_functions): Do not estimate growth when dumping;
2888 * ipa-inline.h (inline_summary): Add min_size.
2889 (growth_likely_positive): New function.
2890 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
2891 (set_cond_stmt_execution_predicate): Cleanup.
2892 (estimate_edge_size_and_time): Compute min_size.
2893 (estimate_calls_size_and_time): Likewise.
2894 (estimate_node_size_and_time): Likewise.
2895 (inline_update_overall_summary): Update min_size.
2896 (do_estimate_edge_time): Likewise.
2897 (do_estimate_edge_size): Update.
2898 (do_estimate_edge_hints): Update.
2899 (growth_likely_positive): New function.
2901 2014-03-28 Jakub Jelinek <jakub@redhat.com>
2904 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
2905 also if addr has VOIDmode.
2907 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2909 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
2910 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
2912 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
2913 instructions as well as AdvancedSIMD loads.
2915 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2917 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
2918 Use crypto_aese type.
2919 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
2920 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
2921 crypto_aese, crypto_aesmc. Move to types.md.
2922 * config/arm/types.md (crypto_aes): Split into crypto_aese,
2924 * config/arm/iterators.md (crypto_type): Likewise.
2926 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2928 * cgraph.c: Include expr.h and tree-dfa.h.
2929 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
2932 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2935 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
2936 regs from checking multi-reg pseudos.
2938 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2940 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
2942 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2944 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
2945 if it would clobber the stack pointer, even temporarily.
2947 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
2949 * mode-switching.c: Make small adjustments to the top comment.
2951 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
2953 * config/rs6000/constraints.md (wD constraint): New constraint to
2954 match the constant integer to get the top DImode/DFmode out of a
2955 vector in a VSX register.
2957 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
2958 match the constant integer to get the top DImode/DFmode out of a
2959 vector in a VSX register.
2961 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
2964 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
2967 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
2968 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
2970 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
2971 Optimize vec_extract of 64-bit values, where the value being
2972 extracted is in the top word, where we can use scalar
2973 instructions. Add direct move and store support. Combine the big
2974 endian/little endian vector select load support into a single insn.
2975 (vsx_extract_<mode>_internal1): Likewise.
2976 (vsx_extract_<mode>_internal2): Likewise.
2977 (vsx_extract_<mode>_load): Likewise.
2978 (vsx_extract_<mode>_store): Likewise.
2979 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
2980 combined into vsx_extract_<mode>_load.
2981 (vsx_extract_<mode>_one_le): Likewise.
2983 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
2984 define the top 64-bit vector element.
2986 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
2989 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2990 Document vec_vbpermq builtin.
2993 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
2994 enable use of xxsldwi and xxpermdi builtin functions.
2995 (vec_xxpermdi): Likewise.
2997 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2998 Document use of vec_xxsldwi and vec_xxpermdi builtins.
3000 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
3002 PR rtl-optimization/60650
3003 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
3005 (find_spills_for): New.
3006 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
3007 Spill all pseudos on the second iteration.
3009 2014-03-27 Marek Polacek <polacek@redhat.com>
3012 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
3015 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3017 * config/s390/s390.c (s390_can_use_return_insn): Check for
3018 call-saved FPRs on 31 bit.
3020 2014-03-27 Jakub Jelinek <jakub@redhat.com>
3023 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
3024 if they need regimplification, just drop them instead of
3025 calling gimple_regimplify_operands on them.
3027 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
3030 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
3031 (aarch64_frame_pointer_required): Adjust logic.
3032 (aarch64_can_eliminate): Adjust logic.
3033 (aarch64_override_options_after_change): Adjust logic.
3035 2014-03-27 Dehao Chen <dehao@google.com>
3037 * ipa-inline.c (early_inliner): Update node's inline info.
3039 2014-03-26 Dehao Chen <dehao@google.com>
3041 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
3042 compiler inserted conditional jumps for NAN float check.
3044 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3046 * ubsan.h (ubsan_create_data): Change second argument's type
3047 to const location_t *.
3048 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
3050 (ubsan_create_data): Change second argument to const location_t *PLOC.
3051 Create Loc field whenever PLOC is non-NULL.
3052 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
3053 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
3057 * real.c (real_to_integer2): Change type of low to UHWI.
3059 2014-03-26 Tobias Burnus <burnus@net-b.de>
3061 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
3062 (CILK_SELF_SPECS): New define.
3063 (driver_self_specs): Use it.
3065 2014-03-26 Richard Biener <rguenther@suse.de>
3067 * tree-pretty-print.c (percent_K_format): Implement special
3068 case for LTO and its stripped down BLOCK tree.
3070 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3073 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
3075 * tree-vrp.c (simplify_internal_call_using_ranges): If only
3076 one range is range_int_cst_p, but not both, at least optimize
3077 addition/subtraction of 0 and multiplication by 0 or 1.
3078 * gimple-fold.c (gimple_fold_call): Fold
3079 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
3080 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
3081 INTEGER_CSTs, try to fold at least x * 0 and y - y.
3083 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
3085 PR rtl-optimization/60452
3086 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
3087 <case REG>: Return 1 for invalid offsets from the frame pointer.
3089 2014-03-26 Marek Polacek <polacek@redhat.com>
3092 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3095 2014-03-26 Marek Polacek <polacek@redhat.com>
3098 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3101 2014-03-26 Marek Polacek <polacek@redhat.com>
3104 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3105 multiplication in unsigned type.
3107 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3109 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3111 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3113 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3115 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3118 * cif-code.def (UNREACHABLE) New code.
3119 * ipa-inline.c (inline_small_functions): Skip edges to
3120 __builtlin_unreachable.
3121 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3122 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3123 predicate to __bulitin_unreachable.
3124 (set_cond_stmt_execution_predicate): Fix issue when
3125 invert_tree_comparison returns ERROR_MARK.
3126 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3127 propagate to inline clones.
3128 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3130 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3131 * cgraphclones.c (cgraph_clone_node): If call destination is already
3132 ureachable, do not redirect it back.
3133 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3136 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3138 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3139 Do not modify inline clones.
3141 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3143 * config/i386/i386.md (general_sext_operand): New mode attr.
3144 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3145 don't generate (sign_extend (const_int)).
3146 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3147 operands[2]. Use We constraint instead of <i> and
3148 <general_sext_operand> predicate instead of <general_operand>.
3149 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3150 * config/i386/constraints.md (We): New constraint.
3151 * config/i386/predicates.md (x86_64_sext_operand,
3152 sext_operand): New predicates.
3154 2014-03-25 Martin Jambor <mjambor@suse.cz>
3157 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3158 inconsistent devirtualizations to __builtin_unreachable.
3160 2014-03-25 Marek Polacek <polacek@redhat.com>
3163 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3165 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3167 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3168 order of elements for big-endian.
3170 2014-03-25 Richard Biener <rguenther@suse.de>
3173 * gimplify-me.c (gimple_regimplify_operands): Update the
3176 2014-03-25 Martin Jambor <mjambor@suse.cz>
3179 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3180 (lto_output_varpool_node): Likewise.
3181 (input_overwrite_node): Likewise.
3182 (input_varpool_node): Likewise.
3184 2014-03-25 Richard Biener <rguenther@suse.de>
3186 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3187 (run_gcc): Likewise.
3189 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3191 * combine.c (simplify_compare_const): Add MODE argument.
3192 Handle mode_width 0 as very large mode_width.
3193 (try_combine, simplify_comparison): Adjust callers.
3195 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3196 type to avoid signed integer overflow.
3197 * explow.c (plus_constant): Likewise.
3199 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3201 * doc/generic.texi: Correct typos.
3203 2014-03-24 Tobias Burnus <burnus@net-b.de>
3205 * doc/invoke.texi (-flto): Expand section about
3206 using static libraries with LTO.
3208 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3210 PR rtl-optimization/60501
3211 * optabs.def (addptr3_optab): New optab.
3212 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3213 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3214 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3216 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3218 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3220 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3222 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3225 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3226 (_mm256_undefined_ps): Define.
3227 (_mm256_undefined_pd): Define.
3228 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3229 (_mm_undefined_pd): Define.
3230 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3231 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3232 (_mm512_undefined_ps): Define.
3233 (_mm512_undefined_pd): Define.
3234 Use _mm*_undefined_*.
3235 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3237 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3239 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3240 (lshr_simd): DI mode added.
3241 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3242 (aarch64_ushr_simddi): Likewise.
3243 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3244 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3245 (vshrd_n_u64): Likewise.
3247 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3249 * Makefile.in (s-macro_list): Depend on cc1.
3251 2014-03-23 Teresa Johnson <tejohnson@google.com>
3253 * ipa-utils.c (ipa_print_order): Use specified dump file.
3255 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3257 PR rtl-optimization/60601
3258 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3260 * gcc.c (eval_spec_function): Initialize save_growing_value.
3262 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3265 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3266 code == MINUS_EXPR, never swap op0 with op1.
3268 * toplev.c (init_local_tick): Avoid signed integer multiplication
3270 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3271 shift by first operand's bitsize.
3273 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3276 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3278 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3279 TARGET_ISA_64BIT_P(x).
3281 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3283 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3284 pattern for vector nor instead of subtract from splat(-1).
3285 (altivec_expand_vec_perm_const_le): Likewise.
3287 2014-03-21 Richard Henderson <rth@twiddle.net>
3290 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3291 related insns after epilogue_completed.
3293 2014-03-21 Martin Jambor <mjambor@suse.cz>
3296 * cgraph.h (symtab_node): New flag body_removed.
3297 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3298 when removing bodies.
3299 * symtab.c (dump_symtab_base): Dump body_removed flag.
3300 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3301 had their bodies removed.
3303 2014-03-21 Martin Jambor <mjambor@suse.cz>
3306 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3309 2014-03-21 Richard Biener <rguenther@suse.de>
3311 PR tree-optimization/60577
3312 * tree-core.h (struct tree_base): Document nothrow_flag use
3313 in VAR_DECL_NONALIASED.
3314 * tree.h (VAR_DECL_NONALIASED): New.
3315 (may_be_aliased): Adjust.
3316 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3318 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3320 * expr.c (expand_expr_real_1): Remove outdated comment.
3322 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3325 * ira.c (adjust_cleared_regs): Call copy_rtx on
3326 *reg_equiv[REGNO (loc)].src_p before passing it to
3327 simplify_replace_fn_rtx.
3330 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3331 into CONST, put pic register as first operand of PLUS. Use
3332 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3334 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3336 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3338 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3340 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3341 around for store forwarding issue in the FPU on the UT699.
3342 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3343 loads and operations if -mfix-ut699 is specified.
3344 (divtf3_hq): Tweak attribute.
3345 (sqrttf2_hq): Likewise.
3347 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3349 * calls.c (store_one_arg): Remove incorrect const qualification on the
3350 type of the temporary.
3351 * cfgexpand.c (expand_return): Likewise.
3352 * expr.c (expand_constructor): Likewise.
3353 (expand_expr_real_1): Likewise.
3355 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3357 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3360 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3363 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3365 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3367 * config/arm/aarch-common-protos.h
3368 (alu_cost_table): Fix spelling of "extend".
3369 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3371 2014-03-19 Richard Biener <rguenther@suse.de>
3374 * tree-core.h (tree_type_common): Re-order pointer members
3375 to reduce recursion depth during GC walks.
3377 2014-03-19 Marek Polacek <polacek@redhat.com>
3380 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3381 before accessing it.
3383 2014-03-19 Richard Biener <rguenther@suse.de>
3386 * lto-streamer-in.c (input_function): In WPA stage do not drop
3389 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3391 PR tree-optimization/60559
3392 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3393 with build_zero_cst assignment.
3395 2014-03-18 Kai Tietz <ktietz@redhat.com>
3397 PR rtl-optimization/56356
3398 * sdbout.c (sdbout_parms): Verify that parms'
3399 incoming argument is valid.
3400 (sdbout_reg_parms): Likewise.
3402 2014-03-18 Richard Henderson <rth@redhat.com>
3405 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3406 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3407 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3409 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3411 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3412 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3413 Italicize plugin event names in description. Explain that
3414 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3415 Remind that no GCC functions should be called after PLUGIN_FINISH.
3416 Explain what pragmas with expansion are.
3418 2014-03-18 Martin Liska <mliska@suse.cz>
3420 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3421 gimple call statement is update.
3422 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3423 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3425 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3428 * ubsan.c (ubsan_instrument_unreachable): Call
3429 initialize_sanitizer_builtins.
3430 (ubsan_pass): Likewise.
3433 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3434 varpool_finalize_decl instead of rest_of_decl_compilation.
3436 2014-03-18 Richard Biener <rguenther@suse.de>
3438 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3439 by using bitmap_and_compl instead of bitmap_and_compl_into.
3440 (df_rd_transfer_function): Likewise.
3442 2014-03-18 Richard Biener <rguenther@suse.de>
3444 * doc/lto.texi (fresolution): Fix typo.
3446 2014-03-18 Richard Biener <rguenther@suse.de>
3448 * doc/invoke.texi (flto): Update for changes in 4.9.
3450 2014-03-18 Richard Biener <rguenther@suse.de>
3452 * doc/loop.texi: Remove section on the removed lambda framework.
3453 Update loop docs with recent changes in preserving loop structure.
3455 2014-03-18 Richard Biener <rguenther@suse.de>
3457 * doc/lto.texi (-fresolution): Document.
3459 2014-03-18 Richard Biener <rguenther@suse.de>
3461 * doc/contrib.texi: Adjust my name.
3463 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3466 * internal-fn.c: Include diagnostic-core.h.
3467 (expand_BUILTIN_EXPECT): New function.
3468 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3469 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3470 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3471 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3473 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3474 Revert 3 argument __builtin_expect code.
3475 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3476 * gimple-fold.c (gimple_fold_call): Likewise.
3477 * tree.h (fold_builtin_expect): New prototype.
3478 * builtins.c (build_builtin_expect_predicate): Add predictor
3479 argument, if non-NULL, create 3 argument __builtin_expect.
3480 (fold_builtin_expect): No longer static. Add ARG2 argument,
3481 pass it through to build_builtin_expect_predicate.
3482 (fold_builtin_2): Adjust caller.
3483 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3484 * internal-fn.def (BUILTIN_EXPECT): New.
3486 2014-03-18 Tobias Burnus <burnus@net-b.de>
3489 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3490 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3491 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3493 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3496 * predict.c (combine_predictions_for_bb): Fix up formatting.
3497 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3498 fill what it points to if non-NULL.
3499 (tree_predict_by_opcode): Adjust caller, use the predictor.
3500 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3502 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3504 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3505 proper constant for the store mode.
3507 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3509 * symtab.c (change_decl_assembler_name): Fix transparent alias
3512 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3514 * config/aarch64/aarch64.c: Correct the comments about the
3515 aarch64 stack layout.
3517 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3519 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3520 check for GF_OMP_FOR_KIND_FOR.
3522 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3524 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3525 ymm and zmm register names.
3527 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3530 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3531 note creation for the 2010-08-31 changes.
3533 2014-03-17 Marek Polacek <polacek@redhat.com>
3536 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3537 as -fno-tree-loop-vectorize.
3538 (expand_omp_simd): Likewise.
3540 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3542 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3543 (eligible_for_call_delay): New prototype.
3544 * config/sparc/sparc.c (tls_call_delay): Rename into...
3545 (eligible_for_call_delay): ...this. Return false if the instruction
3546 cannot be put in the delay slot of a branch.
3547 (eligible_for_restore_insn): Simplify.
3548 (eligible_for_return_delay): Return false if the instruction cannot be
3549 put in the delay slot of a branch and simplify.
3550 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3551 put in the delay slot of a branch.
3552 * config/sparc/sparc.md (fix_ut699): New attribute.
3553 (tls_call_delay): Delete.
3554 (in_call_delay): Reimplement.
3555 (eligible_for_sibcall_delay): Rename into...
3556 (in_sibcall_delay): ...this.
3557 (eligible_for_return_delay): Rename into...
3558 (in_return_delay): ...this.
3559 (in_branch_delay): Reimplement.
3560 (in_uncond_branch_delay): Delete.
3561 (in_annul_branch_delay): Delete.
3563 2014-03-14 Richard Henderson <rth@redhat.com>
3566 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3567 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3568 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3569 (floathi splitters): Remove.
3570 (float<SWI48x>xf2): New pattern.
3571 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3572 code that tried to handle DImode for 32-bit, but which was excluded
3573 by the pattern's condition. Drop allocation of stack temporary.
3574 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3575 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3576 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3577 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3578 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3579 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3580 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3581 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3582 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3583 (*float<SWI48x><X87MODEF>2_i387): Remove.
3584 (all float _with_temp splitters): Remove.
3585 (*float<SWI48x><MODEF>2_i387): New pattern.
3586 (*float<SWI48><MODEF>2_sse): New pattern.
3587 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3588 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3590 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3591 Marek Polacek <polacek@redhat.com>
3594 * common.opt (dump_base_name_prefixed): New Variable.
3595 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3596 if x_dump_base_name_prefixed is already set, set it at the end.
3598 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3600 PR rtl-optimization/60508
3601 * lra-constraints.c (get_reload_reg): Add new parameter
3603 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3604 Pass the new parameter values.
3606 2014-03-14 Richard Biener <rguenther@suse.de>
3608 * common.opt: Revert unintented changes from r205065.
3611 2014-03-14 Richard Biener <rguenther@suse.de>
3614 * cfghooks.c (split_block): Properly adjust all loops the
3615 block was a latch of.
3617 2014-03-14 Martin Jambor <mjambor@suse.cz>
3620 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3623 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3626 * config/avr/avr.c (avr_set_current_function): Pass function name
3627 through default_strip_name_encoding before sanity checking instead
3628 of skipping the first char of the assembler name.
3630 2014-03-13 Richard Henderson <rth@redhat.com>
3633 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3634 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3635 * config/i386/i386-protos.h: Likewise.
3636 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3637 in the expander instead of a splitter.
3638 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3639 any possibility of requiring a memory.
3640 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3641 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3642 (fp branch splitters): Update for ix86_split_fp_branch.
3643 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3644 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3645 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3646 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3647 (*fop_<MODEF>_3_i387): Likewise.
3648 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3649 (splitters for the fop_* register patterns): Remove.
3650 (fscalexf4_i387): Rename from *fscalexf4_i387.
3651 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3653 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3655 PR tree-optimization/59779
3656 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3657 type for bitsize and maxsize instead of HOST_WIDE_INT.
3659 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3661 PR rtl-optimization/57320
3662 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3663 the CFG after thread_prologue_and_epilogue_insns.
3665 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3667 PR rtl-optimization/57189
3668 * lra-constraints.c (process_alt_operands): Disfavor spilling
3671 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3673 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3675 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3677 PR tree-optimization/59025
3679 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3680 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3682 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3685 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3686 calls of avr_out_plus_1.
3688 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3690 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3691 BB's single pred and update the father loop's latch info later.
3693 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3695 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3699 (VEC_base): Likewise.
3700 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3701 registers, we need to swap double words in little endian mode.
3703 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3704 to be a container mode for 128-bit integer operations added in ISA
3705 2.07. Unlike TImode and PTImode, the preferred register set is
3706 the Altivec/VMX registers for the 128-bit operations.
3708 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3710 (rs6000_split_128bit_ok_p): Likewise.
3712 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3713 macros for creating ISA 2.07 normal and overloaded builtin
3714 functions with 3 arguments.
3715 (BU_P8V_OVERLOAD_3): Likewise.
3716 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3717 for use as overloaded functions.
3718 (VPERM_1TI_UNS): Likewise.
3719 (VSEL_1TI): Likewise.
3720 (VSEL_1TI_UNS): Likewise.
3721 (ST_INTERNAL_1ti): Likewise.
3722 (LD_INTERNAL_1ti): Likewise.
3723 (XXSEL_1TI): Likewise.
3724 (XXSEL_1TI_UNS): Likewise.
3725 (VPERM_1TI): Likewise.
3726 (VPERM_1TI_UNS): Likewise.
3727 (XXPERMDI_1TI): Likewise.
3728 (SET_1TI): Likewise.
3729 (LXVD2X_V1TI): Likewise.
3730 (STXVD2X_V1TI): Likewise.
3731 (VEC_INIT_V1TI): Likewise.
3732 (VEC_SET_V1TI): Likewise.
3733 (VEC_EXT_V1TI): Likewise.
3734 (EQV_V1TI): Likewise.
3735 (NAND_V1TI): Likewise.
3736 (ORC_V1TI): Likewise.
3737 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3738 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3740 (VADDUQM): Likewise.
3741 (VSUBCUQ): Likewise.
3742 (VADDEUQM): Likewise.
3743 (VADDECUQ): Likewise.
3744 (VSUBEUQM): Likewise.
3745 (VSUBECUQ): Likewise.
3747 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3748 __int128_t and __uint128_t types.
3749 (__uint128_type): Likewise.
3750 (altivec_categorize_keyword): Add support for vector __int128_t,
3751 vector __uint128_t, vector __int128, and vector unsigned __int128
3752 as a container type for TImode operations that need to be done in
3753 VSX/Altivec registers.
3754 (rs6000_macro_to_expand): Likewise.
3755 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3756 to support 128-bit integer instructions vaddcuq, vadduqm,
3757 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3758 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
3760 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
3761 for V1TImode, and set up preferences to use VSX/Altivec registers.
3762 Setup VSX reload handlers.
3763 (rs6000_debug_reg_global): Likewise.
3764 (rs6000_init_hard_regno_mode_ok): Likewise.
3765 (rs6000_preferred_simd_mode): Likewise.
3766 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
3767 (easy_altivec_constant): Likewise.
3768 (output_vec_const_move): Likewise.
3769 (rs6000_expand_vector_set): Convert V1TImode set and extract to
3771 (rs6000_expand_vector_extract): Likewise.
3772 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
3774 (rs6000_const_vec): Add support for V1TImode.
3775 (rs6000_emit_le_vsx_load): Swap double words when loading or
3776 storing TImode/V1TImode.
3777 (rs6000_emit_le_vsx_store): Likewise.
3778 (rs6000_emit_le_vsx_move): Likewise.
3779 (rs6000_emit_move): Add support for V1TImode.
3780 (altivec_expand_ld_builtin): Likewise.
3781 (altivec_expand_st_builtin): Likewise.
3782 (altivec_expand_vec_init_builtin): Likewise.
3783 (altivec_expand_builtin): Likewise.
3784 (rs6000_init_builtins): Add support for V1TImode type. Add
3785 support for ISA 2.07 128-bit integer builtins. Define type names
3786 for the VSX/Altivec vector types.
3787 (altivec_init_builtins): Add support for overloaded vector
3788 functions with V1TImode type.
3789 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
3790 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
3792 (rs6000_split_128bit_ok_p): Likewise.
3793 (rs6000_handle_altivec_attribute): Create V1TImode from vector
3794 __int128_t and vector __uint128_t.
3796 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
3797 and mode attributes.
3804 (VS_scalar): Likewise.
3805 (VS_double): Likewise.
3806 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
3808 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
3809 we support the ISA 2.07 128-bit integer arithmetic instructions.
3810 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
3811 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
3812 and TImode types for use with the builtin functions.
3813 (V1TI_type_node): Likewise.
3814 (unsigned_V1TI_type_node): Likewise.
3815 (intTI_type_internal_node): Likewise.
3816 (uintTI_type_internal_node): Likewise.
3818 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
3819 128-bit builtin functions.
3820 (UNSPEC_VADDEUQM): Likewise.
3821 (UNSPEC_VADDECUQ): Likewise.
3822 (UNSPEC_VSUBCUQ): Likewise.
3823 (UNSPEC_VSUBEUQM): Likewise.
3824 (UNSPEC_VSUBECUQ): Likewise.
3825 (VM): Add V1TImode to vector mode iterators.
3827 (VI_unit): Likewise.
3828 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
3829 (altivec_vaddcuq): Likewise.
3830 (altivec_vsubuqm): Likewise.
3831 (altivec_vsubcuq): Likewise.
3832 (altivec_vaddeuqm): Likewise.
3833 (altivec_vaddecuq): Likewise.
3834 (altivec_vsubeuqm): Likewise.
3835 (altivec_vsubecuq): Likewise.
3837 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
3839 (BOOL_128): Likewise.
3840 (BOOL_REGS_OUTPUT): Likewise.
3841 (BOOL_REGS_OP1): Likewise.
3842 (BOOL_REGS_OP2): Likewise.
3843 (BOOL_REGS_UNARY): Likewise.
3844 (BOOL_REGS_AND_CR0): Likewise.
3846 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
3847 128-bit integer builtin support.
3848 (vec_vadduqm): Likewise.
3849 (vec_vaddecuq): Likewise.
3850 (vec_vaddeuqm): Likewise.
3851 (vec_vsubecuq): Likewise.
3852 (vec_vsubeuqm): Likewise.
3853 (vec_vsubcuq): Likewise.
3854 (vec_vsubuqm): Likewise.
3856 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3857 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
3858 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
3859 128-bit integer add/subtract to ISA 2.07.
3861 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
3863 * config/arc/arc.c (arc_predicate_delay_insns):
3864 Fix third argument passed to conditionalize_nonjump.
3866 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
3868 * config/aarch64/aarch64-builtins.c
3869 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
3870 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
3871 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
3872 instead of __builtin_lfloor.
3873 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
3875 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3877 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
3878 (tree_ssa_ifcombine_bb_1): New function.
3879 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
3880 is an empty forwarder block to then_bb or vice versa and then_bb
3881 and else_bb are effectively swapped.
3883 2014-03-12 Christian Bruel <christian.bruel@st.com>
3886 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
3887 REG_CFA_DEF_CFA note.
3888 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
3889 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
3891 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
3893 PR tree-optimization/60454
3894 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
3896 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3898 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
3899 Do not define target_cpu_default2 to generic.
3900 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
3901 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
3902 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
3904 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3905 Marc Glisse <marc.glisse@inria.fr>
3907 PR tree-optimization/60502
3908 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
3909 instead of build_low_bits_mask.
3911 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3914 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
3915 if there are multiple uses, but op doesn't live on E edge.
3916 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
3917 clobber stmts before __builtin_unreachable.
3919 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
3921 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
3922 hard_frame_pointer_rtx.
3923 * cse.c (cse_insn): Remove volatile check.
3924 * cselib.c (cselib_process_insn): Likewise.
3925 * dse.c (scan_insn): Likewise.
3927 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3929 * config/arc/arc.c (conditionalize_nonjump): New function,
3931 (arc_ifcvt): ... this.
3932 (arc_predicate_delay_insns): Use it.
3934 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3936 * config/arc/predicates.md (extend_operand): During/after reload,
3937 allow const_int_operand.
3938 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
3939 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
3940 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
3942 (umulsi3_highpart_i): Likewise.
3944 2014-03-11 Richard Biener <rguenther@suse.de>
3946 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
3947 Add asserts to guard possible wrong-code bugs.
3949 2014-03-11 Richard Biener <rguenther@suse.de>
3951 PR tree-optimization/60429
3952 PR tree-optimization/60485
3953 * tree-ssa-structalias.c (set_union_with_increment): Properly
3954 take into account all fields that overlap the shifted vars.
3955 (do_sd_constraint): Likewise.
3956 (do_ds_constraint): Likewise.
3957 (get_constraint_for_ptr_offset): Likewise.
3959 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
3961 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
3962 (nios2_compute_frame_layout):
3963 Add calculation of cfun->machine->fp_save_offset.
3964 (nios2_expand_prologue): Correct setting of frame pointer register
3966 (nios2_expand_epilogue): Update recovery of stack pointer from
3967 frame pointer accordingly.
3968 (nios2_initial_elimination_offset): Update calculation of offset
3969 for eliminating to HARD_FRAME_POINTER_REGNUM.
3971 2014-03-10 Jakub Jelinek <jakub@redhat.com>
3974 * ipa.c (symtab_remove_unreachable_nodes): Don't call
3975 cgraph_get_create_node on VAR_DECLs.
3977 2014-03-10 Richard Biener <rguenther@suse.de>
3980 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
3982 2014-03-08 Douglas B Rupp <rupp@gnat.com>
3984 * config/vms/vms.opt (vms_float_format): New variable.
3986 2014-03-08 Tobias Burnus <burnus@net-b.de>
3988 * doc/invoke.texi (-fcilkplus): Update implementation status.
3990 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
3991 Richard Biener <rguenther@suse.de>
3993 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
3994 consistently accross all TUs.
3995 (run_gcc): Enable -fshort-double automatically at link at link-time
3996 and disallow override.
3998 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
4001 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
4002 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
4003 if they can't be used.
4005 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4007 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
4008 for Solaris 11/x86 ld.
4009 * configure: Regenerate.
4011 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4013 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
4014 (LIB_TLS_SPEC): Save as ld_tls_libs.
4015 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
4016 (HAVE_AS_IX86_TLSLDM): New test.
4017 * configure, config.in: Regenerate.
4018 * config/i386/i386.c (legitimize_tls_address): Fall back to
4019 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
4020 cannot support TLS_MODEL_LOCAL_DYNAMIC.
4021 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
4022 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
4024 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
4026 * common.opt (fira-loop-pressure): Mark as optimization.
4028 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
4030 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
4031 an OpenMP mappable type.
4033 2014-03-06 Matthias Klose <doko@ubuntu.com>
4035 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
4036 MULTILIB_OSDIRNAMES is not defined.
4038 2014-03-06 Jakub Jelinek <jakub@redhat.com>
4039 Meador Inge <meadori@codesourcery.com>
4042 * config/arm/arm.c (arm_tls_symbol_p): Remove.
4043 (arm_legitimize_address): Call legitimize_tls_address for any
4044 arm_tls_referenced_p expression, handle constant addend. Call it
4045 before testing for !TARGET_ARM.
4046 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
4048 2014-03-06 Richard Biener <rguenther@suse.de>
4054 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4056 * tree-streamer.c (record_common_node): Assert we don't record
4057 nodes with type double.
4058 (preload_common_node): Skip type double, complex double and double
4059 pointer since it is now frontend dependent due to fshort-double option.
4061 2014-03-06 Richard Biener <rguenther@suse.de>
4063 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
4064 or -fno-lto is specified and the linker has full plugin support.
4065 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
4066 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
4067 * lto-wrapper.c (merge_and_complain): Merge compile-time
4068 optimization levels.
4069 (run_gcc): And pass it through to the link options.
4071 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
4075 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4077 * cselib.c (remove_useless_values): Skip to avoid quadratic
4078 behavior if the condition moved from...
4079 (cselib_process_insn): ... here holds.
4081 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4084 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
4085 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
4088 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
4089 (TM_H): Add x86-tune.def.
4091 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4093 * config/aarch64/aarch64.c (generic_tunings):
4094 Use cortexa57_extra_costs.
4096 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4099 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4100 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4101 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4104 2014-03-04 Heiher <r@hev.cc>
4106 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4107 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4109 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4111 * config/i386/predicates.md (const2356_operand): Change to ...
4112 (const2367_operand): ... this.
4113 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4115 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4116 (*avx512pf_scatterpf<mode>sf): Ditto.
4117 (avx512pf_scatterpf<mode>df): Ditto.
4118 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4119 (*avx512pf_scatterpf<mode>df): Ditto.
4120 * config/i386/i386.c (ix86_expand_builtin): Update
4121 incorrect hint operand error message.
4123 2014-03-04 Richard Biener <rguenther@suse.de>
4125 * lto-section-in.c (lto_get_section_data): Fix const cast.
4127 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4129 * tree-streamer.c (record_common_node): Assert we don't record
4130 nodes with type double.
4131 (preload_common_node): Skip type double, complex double and double
4132 pointer since it is now frontend dependent due to fshort-double option.
4134 2014-03-04 Richard Biener <rguenther@suse.de>
4137 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4138 (lto_input_toplevel_asms): Likewise.
4139 * lto-section-in.c (lto_get_section_data): Instead do it here
4142 2014-03-04 Richard Biener <rguenther@suse.de>
4144 PR tree-optimization/60382
4145 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4146 dead PHIs a reduction.
4148 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4150 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4152 (_mm_prefetch): Move out of GCC target("sse") pragma.
4153 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4154 GCC target("prfchw") pragma.
4155 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4157 * config/i386/i386.c (ix86_option_override_internal): Enable
4158 -mprfchw with -mprefetchwt1.
4160 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4162 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4165 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4167 * opts.h (CL_PCH_IGNORE): Define.
4168 * targhooks.c (option_affects_pch_p):
4169 Return false for options that have CL_PCH_IGNORE set.
4170 * opt-functions.awk: Process PchIgnore.
4171 * doc/options.texi: Document PchIgnore.
4173 * config/arc/arc.opt (misize): Add PchIgnore property.
4175 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4177 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4178 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4179 constraint on constants to permit them being loaded into
4180 GENERAL_REGS or BASE_REGS.
4182 2014-03-03 Nick Clifton <nickc@redhat.com>
4184 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4185 anti-cacnonical alternatives.
4186 (negandhi3_real): New pattern.
4187 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4189 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4191 * config/avr/avr-mcus.def: Remove atxmega16x1.
4192 * config/avr/avr-tables.opt: Regenerate.
4193 * config/avr/t-multilib: Regenerate.
4194 * doc/avr-mmcu.texi: Regenerate.
4196 2014-03-03 Tobias Grosser <tobias@grosser.es>
4197 Mircea Namolaru <mircea.namolaru@inria.fr>
4199 PR tree-optimization/58028
4200 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4203 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4205 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4206 not handled by recognizers.
4208 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4211 * function.c (expand_function_end): Don't emit
4212 clobber_return_register sequence if clobber_after is a BARRIER.
4213 * cfgexpand.c (construct_exit_block): Append instructions before
4214 return_label to prev_bb.
4216 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4218 * config/rs6000/constraints.md: Document reserved use of "wc".
4220 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4223 * ipa.c (function_and_variable_visibility): When dissolving comdat
4224 group, also set all symbols to local.
4226 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4231 2013-12-14 Jan Hubicka <jh@suse.cz>
4233 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4235 2014-03-02 Jon Beniston <jon@beniston.com>
4241 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4242 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4243 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4244 (simple_return, *simple_return): New patterns
4245 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4246 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4248 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4250 * dwarf2out.c (gen_subprogram_die): Tidy.
4252 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4255 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4256 (*mov_t_msb_neg_negc): ... this new insn.
4258 2014-02-28 Jason Merrill <jason@redhat.com>
4261 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4264 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4267 * dwarf2out.c (decltype_auto_die): New static.
4268 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4269 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4270 (is_cxx_auto): Likewise.
4272 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4274 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4275 we are not using general regs only.
4277 2014-02-28 Richard Biener <rguenther@suse.de>
4280 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4281 previous fix and only allow to remove trivial pre-headers
4282 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4283 (remove_forwarder_block): Properly update the latch of a loop.
4285 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4288 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4289 (cselib_preserved_hash_table): New.
4290 (preserve_constants_and_equivs): Move preserved vals to it.
4291 (cselib_find_slot): Look it up first.
4292 (cselib_init): Initialize it.
4293 (cselib_finish): Release it.
4294 (dump_cselib_table): Dump it.
4296 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4299 * cselib.c (remove_useless_values): Skip to avoid quadratic
4300 behavior if the condition moved from...
4301 (cselib_process_insn): ... here holds.
4303 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4306 * var-tracking.c (vt_initialize): Apply the same condition to
4307 preserve the CFA base value.
4309 2014-02-28 Joey Ye <joey.ye@arm.com>
4312 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4313 if reload in progress or completed.
4315 2014-02-28 Tobias Burnus <burnus@net-b.de>
4318 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4321 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4323 * doc/tm.texi.in (Condition Code Status): Update documention for
4324 relative locations of cc0-setter and cc0-user.
4326 2014-02-27 Jeff Law <law@redhat.com>
4328 PR rtl-optimization/52714
4329 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4330 into two independent simple sets, if I3 is a jump, ensure the
4331 pattern we place into I3 is a (set (pc) ...).
4333 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4334 Jeff Law <law@redhat.com>
4336 PR rtl-optimization/49847
4337 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4338 are in different blocks.
4339 * doc/tm.texi (Condition Code Status): Update documention for
4340 relative locations of cc0-setter and cc0-user.
4342 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4345 * lra.c (lra_emit_add): Check SUBREG too.
4347 2014-02-27 Andreas Schwab <schwab@suse.de>
4349 * config/m68k/m68k.c (m68k_option_override): Disable
4350 -flive-range-shrinkage for classic m68k.
4351 (m68k_override_options_after_change): Likewise.
4353 2014-02-27 Marek Polacek <polacek@redhat.com>
4356 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4357 -Wmaybe-uninitialized.
4359 2014-02-27 Alan Modra <amodra@gmail.com>
4362 * reload1.c (emit_input_reload_insns): When reload_override_in,
4363 set old to rl->in_reg when rl->in_reg is a subreg.
4365 2014-02-26 Richard Biener <rguenther@suse.de>
4368 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4370 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4372 * common/config/i386/predicates.md (const1256_operand): Remove.
4373 (const2356_operand): New.
4374 (const_1_to_2_operand): Remove.
4375 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4376 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4377 (*avx512pf_gatherpf<mode>sf): Ditto.
4378 (avx512pf_gatherpf<mode>df): Ditto.
4379 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4380 (*avx512pf_gatherpf<mode>df): Ditto.
4381 (avx512pf_scatterpf<mode>sf): Ditto.
4382 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4383 (*avx512pf_scatterpf<mode>sf): Ditto.
4384 (avx512pf_scatterpf<mode>df): Ditto.
4385 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4386 (*avx512pf_scatterpf<mode>df): Ditto.
4387 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4389 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4391 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4392 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4393 (_mm512_mask_testn_epi64_mask): Move to ...
4394 * config/i386/avx512cdintrin.h: Here.
4395 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4396 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4397 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4398 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4399 TARGET_AVX512F from TARGET_AVX512CD.
4401 2014-02-26 Richard Biener <rguenther@suse.de>
4404 * ipa.c (walk_polymorphic_call_targets): Properly guard
4405 call to inline_update_overall_summary.
4407 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4410 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4411 and latches only if requested. Fix latch if it is removed.
4412 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4413 LOOPS_HAVE_PREHEADERS.
4415 2014-02-25 Andrew Pinski <apinski@cavium.com>
4417 * builtins.c (expand_builtin_thread_pointer): Create a new target
4418 when the target is NULL.
4420 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4422 PR rtl-optimization/60317
4423 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4424 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4425 * lra-assigns.c: Include params.h.
4426 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4427 other reload pseudos considerations.
4429 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4431 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4432 to use canonical form for nor<mode>3.
4434 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4437 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4440 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4442 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4443 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4444 (ix86_handle_option): Handle OPT_mprefetchwt1.
4445 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4446 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4448 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4449 OPTION_MASK_ISA_PREFETCHWT1.
4450 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4451 (PTA_PREFETCHWT1): New.
4452 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4453 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4454 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4455 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4456 (*prefetch_avx512pf_<mode>_: Change into ...
4457 (*prefetch_prefetchwt1_<mode>: This.
4458 * config/i386/i386.opt (mprefetchwt1): New.
4459 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4460 (_mm_prefetch): Handle intent to write.
4461 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4463 2014-02-25 Richard Biener <rguenther@suse.de>
4466 * emit-rtl.c (mem_attrs_htab): Remove.
4467 (mem_attrs_htab_hash): Likewise.
4468 (mem_attrs_htab_eq): Likewise.
4469 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4470 (init_emit_once): Do not allocate mem_attrs_htab.
4472 2014-02-25 Richard Biener <rguenther@suse.de>
4475 * lto-opts.c (lto_write_options): Output non-explicit conservative
4476 -fwrapv, -fno-trapv and -fno-strict-overflow.
4477 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4478 (run_gcc): And pass them through.
4480 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4482 * sel-sched.c (calculate_new_fences): New parameter ptime.
4483 Calculate it as a maximum over all fence cycles.
4484 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4485 Print the final schedule timing when sched_verbose.
4487 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4489 PR rtl-optimization/60292
4490 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4491 bit fot the fence instruction.
4493 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4495 * calls.h: Fix typo in comment.
4497 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4499 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4500 adjusting offsetable addresses.
4502 2014-02-24 Guozhi Wei <carrot@google.com>
4504 * sparseset.h (sparseset_pop): Fix the wrong index.
4506 2014-02-24 Walter Lee <walt@tilera.com>
4508 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4509 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4511 * common/config/tilegx/tilegx-common.c
4512 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4513 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4515 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4516 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4517 (tilegx_gimplify_va_arg_expr): Handle big endian.
4518 (tilegx_expand_unaligned_load): Ditto.
4519 (tilegx_expand_unaligned_store): Ditto.
4520 (TARGET_RETURN_IN_MSB): New.
4521 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4522 (TARGET_ENDIAN_DEFAULT): New.
4523 (TARGET_BIG_ENDIAN): Handle big endian.
4524 (BYTES_BIG_ENDIAN): Ditto.
4525 (WORDS_BIG_ENDIAN): Ditto.
4526 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4529 * config/tilegx/tilegx.md (extv): Handle big endian.
4531 (insn_st<n>): Ditto.
4532 (insn_st<n>_add<bitsuffix>): Ditto.
4533 (insn_stnt<n>): Ditto.
4534 (insn_stnt<n>_add<bitsuffix>):Ditto.
4535 (vec_interleave_highv8qi): Handle big endian.
4536 (vec_interleave_highv8qi_be): New.
4537 (vec_interleave_highv8qi_le): New.
4538 (insn_v1int_h): Handle big endian.
4539 (vec_interleave_lowv8qi): Handle big endian.
4540 (vec_interleave_lowv8qi_be): New.
4541 (vec_interleave_lowv8qi_le): New.
4542 (insn_v1int_l): Handle big endian.
4543 (vec_interleave_highv4hi): Handle big endian.
4544 (vec_interleave_highv4hi_be): New.
4545 (vec_interleave_highv4hi_le): New.
4546 (insn_v2int_h): Handle big endian.
4547 (vec_interleave_lowv4hi): Handle big endian.
4548 (vec_interleave_lowv4hi_be): New.
4549 (vec_interleave_lowv4hi_le): New.
4550 (insn_v2int_l): Handle big endian.
4551 (vec_interleave_highv2si): Handle big endian.
4552 (vec_interleave_highv2si_be): New.
4553 (vec_interleave_highv2si_le): New.
4554 (insn_v4int_h): Handle big endian.
4555 (vec_interleave_lowv2si): Handle big endian.
4556 (vec_interleave_lowv2si_be): New.
4557 (vec_interleave_lowv2si_le): New.
4558 (insn_v4int_l): Handle big endian.
4559 * config/tilegx/tilegx.opt (mbig-endian): New option.
4560 (mlittle-endian): New option.
4561 * doc/install.texi: Document tilegxbe-linux.
4562 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4564 2014-02-24 Martin Jambor <mjambor@suse.cz>
4567 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4568 there are no parameter descriptors.
4570 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4572 PR rtl-optimization/60268
4573 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4574 initialization to ...
4575 (sched_rgn_init): ... here.
4576 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4578 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4580 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4583 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4585 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4588 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4590 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4591 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4593 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4595 * config/microblaze/predicates.md: Add cmp_op predicate.
4596 * config/microblaze/microblaze.md: Add branch_compare instruction
4597 which uses cmp_op predicate and emits cmp insn before branch.
4598 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4599 to microblaze_expand_conditional_branch and consolidate logic.
4600 (microblaze_expand_conditional_branch): emit branch_compare
4601 insn instead of handling cmp op separate from branch insn.
4603 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4605 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4608 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4610 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4611 define_insn with define_expand and new define_insn
4612 *altivec_lve<VI_char>x_internal.
4613 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4614 and new define_insn *altivec_stve<VI_char>x_internal.
4615 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4617 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4619 (altivec_expand_stvex_be): New function.
4621 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4623 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4624 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4625 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4626 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4628 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4631 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4632 instead of emit_move_insn.
4634 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4636 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4638 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4639 gen_altivec_vsumsws.
4641 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4643 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4644 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4645 (altivec_lvxl_<mode>): New define_expand incorporating
4646 -maltivec=be semantics where needed.
4647 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4648 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4649 semantics where needed.
4650 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4651 (altivec_stvx_<mode>): New define_expand incorporating
4652 -maltivec=be semantics where needed.
4653 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4654 VM2 iterator instead of V4SI.
4655 (altivec_stvxl_<mode>): New define_expand incorporating
4656 -maltivec=be semantics where needed.
4657 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4658 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4659 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4660 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4661 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4662 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4663 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4664 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4665 ALTIVEC_BUILTIN_STVXL.
4666 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4667 (altivec_expand_stvx_be): Likewise.
4668 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4669 (altivec_expand_lvx_be): Likewise.
4670 (altivec_expand_stvx_be): Likewise.
4671 (altivec_expand_builtin): Add cases for
4672 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4673 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4674 (altivec_init_builtins): Add definitions for
4675 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4676 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4678 2014-02-21 Catherine Moore <clm@codesourcery.com>
4680 * doc/invoke.texi (mvirt, mno-virt): Document.
4681 * config/mips/mips.opt (mvirt): New option.
4682 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4684 2014-02-21 Richard Biener <rguenther@suse.de>
4686 PR tree-optimization/60276
4687 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4688 (STMT_VINFO_MIN_NEG_DIST): New macro.
4689 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4690 STMT_VINFO_MIN_NEG_DIST.
4691 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4692 made for negative dependence distances still hold.
4694 2014-02-21 Richard Biener <rguenther@suse.de>
4697 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4698 DECL_INITIAL for globals not in the current function context.
4700 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4702 PR tree-optimization/56490
4703 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4704 * tree-ssa-uninit.c: Include params.h.
4705 (compute_control_dep_chain): Add num_calls argument, return false
4706 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4707 num_calls to recursive call.
4708 (find_predicates): Change dep_chain into normal array,
4709 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4710 variable and adjust compute_control_dep_chain caller.
4711 (find_def_preds): Likewise.
4713 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4715 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4716 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4718 2014-02-21 Nick Clifton <nickc@redhat.com>
4720 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4721 (pushhi1): Likewise.
4722 (popqi1): Add mode to pre_dec.
4725 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4727 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4728 mode for mask of V8SFmode permutation.
4730 2014-02-20 Richard Henderson <rth@redhat.com>
4733 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4734 a new pseudo for OLDVAL.
4736 2014-02-20 Jakub Jelinek <jakub@redhat.com>
4739 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4740 gen_reg_rtx if d->testing_p.
4741 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4742 if d->testing_p and we will certainly return true.
4743 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4746 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
4748 * emit-rtl.c (gen_reg_rtx): Assert that
4749 crtl->emit.regno_pointer_align_length is non-zero.
4751 2014-02-20 Richard Henderson <rth@redhat.com>
4754 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4755 on failure the store back into EXPECT.
4757 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4758 Sandra Loosemore <sandra@codesourcery.com>
4760 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
4761 * config/nios2/nios2.c (nios2_function_profiler): Add
4762 -fPIC (flag_pic == 2) support.
4763 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
4764 (nios2_large_offset_p): New function.
4765 (nios2_unspec_reloc_p): Move up position, update to use
4766 nios2_large_offset_p.
4767 (nios2_unspec_address): Remove function.
4768 (nios2_unspec_offset): New function.
4769 (nios2_large_got_address): New function.
4770 (nios2_got_address): Add large offset support.
4771 (nios2_legitimize_tls_address): Update usage of removed and new
4773 (nios2_symbol_binds_local_p): New function.
4774 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
4775 (nios2_legitimize_address): Update to use nios2_large_offset_p.
4776 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
4777 (nios2_print_operand): Merge H/L processing, add hiadj/lo
4778 processing for (const (unspec ...)).
4779 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
4781 2014-02-20 Richard Biener <rguenther@suse.de>
4783 * tree-cfg.c (replace_uses_by): Mark altered BBs before
4784 doing the substitution.
4785 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
4787 2014-02-20 Martin Jambor <mjambor@suse.cz>
4790 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
4791 info when checking whether lattices are bottom.
4793 2014-02-20 Richard Biener <rguenther@suse.de>
4796 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
4799 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
4802 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
4803 parameter specifying the scaling.
4804 (inline_call): Update.
4805 (want_inline_recursively): Guard division by zero.
4806 (recursive_inlining): Update.
4807 * ipa-inline.h (clone_inlined_nodes): Update.
4809 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4812 * config/i386/i386.c (classify_argument): Pass structures of size
4813 64 bytes or less in register.
4815 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4816 Kirill Yukhin <kirill.yukhin@intel.com>
4818 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
4819 (_mm_rcp28_round_ss): Ditto.
4820 (_mm_rsqrt28_round_sd): Ditto.
4821 (_mm_rsqrt28_round_ss): Ditto.
4822 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
4823 (_mm_rcp14_round_ss): Ditto.
4824 (_mm_rsqrt14_round_sd): Ditto.
4825 (_mm_rsqrt14_round_ss): Ditto.
4826 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
4827 the first input operand, get rid of match_dup.
4828 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
4830 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
4832 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
4833 operand as the first input operand, set type attribute.
4834 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
4836 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
4837 operand as the first input operand, set type attribute.
4839 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4841 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
4844 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
4847 * config/i386/i386.c (construct_container): Remove TFmode check
4848 for X86_64_INTEGER_CLASS.
4850 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
4853 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
4854 only when -Wpsabi is enabled.
4856 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
4859 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
4860 passing arrays in registers are the same as for structs, so remove the
4861 special case for them.
4863 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
4865 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
4866 destination type, extract only the valid bits if the source type is not
4867 integral and has a different mode.
4869 2014-02-19 Richard Biener <rguenther@suse.de>
4872 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
4875 2014-02-19 Richard Biener <rguenther@suse.de>
4878 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
4879 (ipa_modify_call_arguments): Emit an argument load explicitely and
4880 preserve virtual SSA form there and for the replacement call.
4881 Do not update SSA form nor free dominance info.
4883 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4885 * ipa.c (function_and_variable_visibility): Also clear WEAK
4886 flag when disolving COMDAT_GROUP.
4888 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4890 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
4891 * ipa-prop.c (ipa_set_jf_known_type): Return early when
4893 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
4894 do more sanity checks.
4895 (detect_type_change): Return true when giving up early.
4896 (compute_complex_assign_jump_func): Fix type parameter of
4897 ipa_set_ancestor_jf.
4898 (compute_complex_ancestor_jump_func): Likewise.
4899 (update_jump_functions_after_inlining): Fix updating of
4901 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
4903 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4905 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
4906 inline clones when edge disappears.
4908 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4911 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
4912 Split 64-bit moves into 2 patterns. Do not allow the use of
4913 direct move for TDmode in little endian, since the decimal value
4914 has little endian bytes within a word, but the 64-bit pieces are
4915 ordered in a big endian fashion, and normal subreg's of TDmode are
4917 (mov<mode>_64bit_dm): Likewise.
4918 (movtd_64bit_nodm): Likewise.
4920 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4922 PR tree-optimization/60174
4923 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
4924 statement of an SSA_NAME that occurs in an abnormal PHI node.
4926 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4929 * final.c (SEEN_BB): Remove.
4930 (SEEN_NOTE, SEEN_EMITTED): Renumber.
4931 (final_scan_insn): Don't force_source_line on second
4932 NOTE_INSN_BASIC_BLOCK.
4934 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
4937 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
4938 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
4939 (type_natural_mode): Warn ABI change when %zmm register is not
4940 available for AVX512F vector value passing.
4942 2014-02-18 Kai Tietz <ktietz@redhat.com>
4945 * config/i386/i386.c (ix86_expand_prologue): Use value in
4946 rax register as displacement when restoring %r10 or %rax.
4947 Fix wrong offset when restoring both registers.
4949 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4951 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
4952 assertion with conditional return.
4954 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4955 Uros Bizjak <ubizjak@gmail.com>
4958 * config/i386/driver-i386.c (host_detect_local_cpu): If
4959 YMM state is not saved by the OS, also clear has_f16c. Move
4960 CPUID 0x80000001 handling before YMM state saving checking.
4962 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
4964 PR rtl-optimization/58960
4965 * haifa-sched.c (alloc_global_sched_pressure_data): New,
4966 factored out from ...
4967 (sched_init): ... here.
4968 (free_global_sched_pressure_data): New, factored out from ...
4969 (sched_finish): ... here.
4970 * sched-int.h (free_global_sched_pressure_data): Declare.
4971 * sched-rgn.c (nr_regions_initial): New static global.
4972 (haifa_find_rgns): Initialize it.
4973 (schedule_region): Disable sched-pressure for the newly
4976 2014-02-17 Richard Biener <rguenther@suse.de>
4978 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
4979 release SSA defs of pattern stmts.
4981 2014-02-17 Richard Biener <rguenther@suse.de>
4983 * tree-inline.c (expand_call_inline): Release the virtual
4984 operand defined by the call we are about to inline.
4986 2014-02-17 Richard Biener <rguenther@suse.de>
4988 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
4990 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
4991 Ilya Tocar <ilya.tocar@intel.com>
4993 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
4994 arguments order in builtin.
4995 (_mm512_permutexvar_epi64): Ditto.
4996 (_mm512_mask_permutexvar_epi64): Ditto
4997 (_mm512_maskz_permutexvar_epi32): Ditto
4998 (_mm512_permutexvar_epi32): Ditto
4999 (_mm512_mask_permutexvar_epi32): Ditto
5001 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5003 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
5004 (p8_vmrgow): Likewise.
5006 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5008 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
5011 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
5014 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
5015 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
5016 into 64-bit and 32-bit moves. On 64-bit moves, add support for
5017 using direct move instructions on ISA 2.07. Also adjust
5018 instruction length for 64-bit.
5019 (mov<mode>_64bit, TFmode/TDmode): Likewise.
5020 (mov<mode>_32bit, TFmode/TDmode): Likewise.
5022 2014-02-15 Alan Modra <amodra@gmail.com>
5026 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
5027 find_replacement on parts of insn rtl that might be reloaded.
5029 2014-02-15 Richard Biener <rguenther@suse.de>
5031 PR tree-optimization/60183
5032 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
5033 (tree_ssa_phiprop): Calculate and free post-dominators.
5035 2014-02-14 Jeff Law <law@redhat.com>
5037 PR rtl-optimization/60131
5038 * ree.c (get_extended_src_reg): New function.
5039 (combine_reaching_defs): Use it rather than assuming location of REG.
5040 (find_and_remove_re): Verify first operand of extension is
5041 a REG before adding the insns to the copy list.
5043 2014-02-14 Roland McGrath <mcgrathr@google.com>
5045 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
5046 * configure: Regenerated.
5047 * config.in: Regenerated.
5048 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
5049 instead of ASM_SHORT.
5051 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
5052 Richard Earnshaw <rearnsha@arm.com>
5054 PR rtl-optimization/59535
5055 * lra-constraints.c (process_alt_operands): Encourage alternative
5056 when unassigned pseudo class is superset of the alternative class.
5057 (inherit_reload_reg): Don't inherit when optimizing for code size.
5058 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
5059 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
5060 modes not less than 4 for Thumb1.
5062 2014-02-14 Kyle McMartin <kyle@redhat.com>
5065 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
5067 2014-02-14 Richard Biener <rguenther@suse.de>
5069 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
5070 (get_frame_arg): Drop the assert with langhook types_compatible_p.
5071 Do not strip INDIRECT_REFs.
5073 2014-02-14 Richard Biener <rguenther@suse.de>
5076 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
5077 DECL_FUNCTION_SPECIFIC_TARGET.
5078 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
5079 * tree-streamer-out.c (pack_ts_target_option): Remove.
5080 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
5081 (write_ts_function_decl_tree_pointers): Do not stream
5082 DECL_FUNCTION_SPECIFIC_TARGET.
5083 * tree-streamer-in.c (unpack_ts_target_option): Remove.
5084 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
5085 (lto_input_ts_function_decl_tree_pointers): Do not stream
5086 DECL_FUNCTION_SPECIFIC_TARGET.
5088 2014-02-14 Jakub Jelinek <jakub@redhat.com>
5090 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
5091 (get_initial_def_for_induction, vectorizable_induction): Ignore
5092 debug stmts when looking for exit_phi.
5093 (vectorizable_live_operation): Fix up condition.
5095 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5097 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5098 nreverse() because it changes the content of original tree list.
5100 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5102 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5103 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5105 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5107 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5108 GNU coding standards.
5110 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5113 * dwarf2out.c (gen_subprogram_die): Don't call
5114 add_calling_convention_attribute if subr_die is old_die.
5116 2014-02-13 Sharad Singhai <singhai@google.com>
5118 * doc/optinfo.texi: Fix order of nodes.
5120 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5122 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5123 operands[2], not operands[3].
5125 2014-02-13 Richard Biener <rguenther@suse.de>
5128 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5129 mention the possibility of an in-tree build.
5130 (CLooG): Update recommended version to 0.18.1, mention the
5131 possibility of an in-tree build and clarify that the ISL
5132 bundled with CLooG does not work.
5134 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5137 * expr.c (compress_float_constant): If x is a hard register,
5138 extend into a pseudo and then move to x.
5140 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5142 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5143 caused by bad second argument to warning_at() with -mhotpatch and
5144 nested functions (e.g. with gfortran).
5146 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5148 * opts.c (option_name): Remove "enabled by default" rider.
5150 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5152 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5154 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5155 Uros Bizjak <ubizjak@gmail.com>
5158 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5159 * configure: Regenerated.
5161 2014-02-12 Richard Biener <rguenther@suse.de>
5163 * vec.c (vec_prefix::calculate_allocation): Move as
5164 inline variant to vec.h.
5165 (vec_prefix::calculate_allocation_1): New out-of-line version.
5166 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5167 (vec_prefix::m_has_auto_buf): Rename to ...
5168 (vec_prefix::m_using_auto_storage): ... this.
5169 (vec_prefix::calculate_allocation): Inline the easy cases
5170 and dispatch to calculate_allocation_1 which doesn't need the
5172 (va_heap::reserve): Use gcc_checking_assert.
5173 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5174 m_using_auto_storage.
5175 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5177 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5178 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5179 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5181 2014-02-12 Richard Biener <rguenther@suse.de>
5183 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5184 when we found a dependence.
5186 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5188 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5190 (maybe_fold_stmt): ... into this new function.
5191 * omp-low.c (lower_omp): Update comment.
5193 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5196 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5199 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5201 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5202 identifiers in comments.
5203 (cortexa53_extra_costs): Likewise.
5204 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5205 (cortexa7_extra_costs): Likewise.
5206 (cortexa12_extra_costs): Likewise.
5207 (cortexa15_extra_costs): Likewise.
5208 (v7m_extra_costs): Likewise.
5210 2014-02-12 Richard Biener <rguenther@suse.de>
5213 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5214 of posix_memalign being successful.
5215 (lower_stmt): Restrict lowering of posix_memalign to when
5216 -ftree-bit-ccp is enabled.
5218 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5220 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5222 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5224 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5226 PR rtl-optimization/60116
5227 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5228 other_insn once the combination has been validated.
5230 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5233 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5235 * ipa-devirt.c: Include demangle.h
5236 (odr_violation_reported): New static variable.
5237 (add_type_duplicate): Update odr_violations.
5238 (maybe_record_node): Add completep parameter; update it.
5239 (record_target_from_binfo): Add COMPLETEP parameter;
5240 update it as needed.
5241 (possible_polymorphic_call_targets_1): Likewise.
5242 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5243 rename FINAL to COMPLETE.
5244 (record_targets_from_bases): Sanity check we found the binfo;
5245 fix COMPLETEP updating.
5246 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5247 parameter, fix computing of COMPLETEP.
5248 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5249 at LTO time do demangling.
5250 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5251 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5253 (gimple_get_virt_method_for_binfo): Likewise.
5254 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5255 gimple_get_virt_method_for_vtable): Update prototypes.
5257 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5260 * genautomata.c (add_presence_absence): Fix typo with
5261 {final_}presence_list.
5263 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5266 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5267 for VSX/Altivec vectors that land in GPR registers.
5269 2014-02-11 Richard Henderson <rth@redhat.com>
5270 Jakub Jelinek <jakub@redhat.com>
5273 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5274 around drhs if type conversion to lacc->type is not useless.
5276 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5278 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5280 (cortex-a57.cortex-a53): Likewise.
5281 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5283 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5285 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5288 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5290 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5291 add_options_for_arm_vfp3.
5293 2014-02-11 Jeff Law <law@redhat.com>
5296 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5297 object with an undesirable mode.
5299 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5302 * config/i386/sol2-9.h: New file.
5303 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5304 *-*-solaris2.9*): Use it.
5306 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5308 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5309 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5311 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5313 * config/microblaze/microblaze.c: Extend mcpu version format
5315 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5317 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5319 2014-02-10 Richard Henderson <rth@redhat.com>
5322 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5323 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5324 ms-abi vs -mno-accumulate-outgoing-args.
5325 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5326 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5329 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5332 * cfgexpand.c (expand_asm_operands): Attach source location to
5333 ASM_INPUT rtx objects.
5334 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5336 2014-02-10 Nick Clifton <nickc@redhat.com>
5338 * config/mn10300/mn10300.c (popcount): New function.
5339 (mn10300_expand_prologue): Include saved registers in stack usage
5342 2014-02-10 Jeff Law <law@redhat.com>
5345 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5346 when changing the SET_DEST of a prior insn to avoid an input reload.
5348 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5350 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5351 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5352 -mcall-openbsd, or -mcall-linux.
5353 (CC1_ENDIAN_BIG_SPEC): Remove.
5354 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5355 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5356 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5357 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5358 and %cc1_endian_default.
5359 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5361 2014-02-10 Richard Biener <rguenther@suse.de>
5363 PR tree-optimization/60115
5364 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5365 MEM_REF handling. Properly verify that the accesses are not
5366 out of the objects bound.
5368 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5370 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5373 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5375 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5376 proper constants and fix formatting.
5377 (possible_polymorphic_call_targets): Fix formatting.
5379 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5380 Ilya Tocar <ilya.tocar@intel.com>
5382 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5383 (_mm512_loadu_epi32): Renamed into...
5384 (_mm512_loadu_si512): This.
5385 (_mm512_storeu_epi32): Renamed into...
5386 (_mm512_storeu_si512): This.
5387 (_mm512_maskz_ceil_ps): Removed.
5388 (_mm512_maskz_ceil_pd): Ditto.
5389 (_mm512_maskz_floor_ps): Ditto.
5390 (_mm512_maskz_floor_pd): Ditto.
5391 (_mm512_floor_round_ps): Ditto.
5392 (_mm512_floor_round_pd): Ditto.
5393 (_mm512_ceil_round_ps): Ditto.
5394 (_mm512_ceil_round_pd): Ditto.
5395 (_mm512_mask_floor_round_ps): Ditto.
5396 (_mm512_mask_floor_round_pd): Ditto.
5397 (_mm512_mask_ceil_round_ps): Ditto.
5398 (_mm512_mask_ceil_round_pd): Ditto.
5399 (_mm512_maskz_floor_round_ps): Ditto.
5400 (_mm512_maskz_floor_round_pd): Ditto.
5401 (_mm512_maskz_ceil_round_ps): Ditto.
5402 (_mm512_maskz_ceil_round_pd): Ditto.
5403 (_mm512_expand_pd): Ditto.
5404 (_mm512_expand_ps): Ditto.
5405 * config/i386/i386.c (ix86_builtins): Remove
5406 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5407 (bdesc_args): Ditto.
5408 * config/i386/predicates.md (const1256_operand): New.
5409 (const_1_to_2_operand): Ditto.
5410 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5411 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5412 (*avx512pf_gatherpf<mode>sf): Ditto.
5413 (avx512pf_gatherpf<mode>df): Ditto.
5414 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5415 (*avx512pf_gatherpf<mode>df): Ditto.
5416 (avx512pf_scatterpf<mode>sf): Ditto.
5417 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5418 (*avx512pf_scatterpf<mode>sf): Ditto.
5419 (avx512pf_scatterpf<mode>df): Ditto.
5420 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5421 (*avx512pf_scatterpf<mode>df): Ditto.
5422 (avx512f_expand<mode>): Removed.
5423 (<shift_insn><mode>3<mask_name>): Change predicate type.
5425 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5427 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5428 not at the end of datarefs vector use ordered_remove to avoid
5429 reordering datarefs vector.
5432 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5433 mark local addressable non-static vars as GOVD_PRIVATE
5434 instead of GOVD_LOCAL.
5435 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5436 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5440 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5441 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5442 assume_aligned or alloc_align attributes.
5443 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5444 arguments. Handle also assume_aligned and alloc_align attributes.
5445 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5446 calls to functions with assume_aligned or alloc_align attributes.
5447 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5449 2014-02-08 Terry Guo <terry.guo@arm.com>
5451 * doc/invoke.texi: Document ARM -march=armv7e-m.
5453 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5455 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5456 flag on __cilkrts_rethrow builtin.
5459 * ipa-cp.c (determine_versionability): Fail at -O0
5460 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5461 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5464 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5467 * tree-inline.c (copy_forbidden): Fail for
5468 __attribute__((optimize (0))) functions.
5470 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5472 * varpool.c: Include pointer-set.h.
5473 (varpool_remove_unreferenced_decls): Variables in other partitions
5474 will not be output; be however careful to not lose information
5477 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5479 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5480 lookup in the vtable constructor.
5482 2014-02-07 Jeff Law <law@redhat.com>
5485 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5486 define_insn_and_split.
5488 * ipa-inline.c (inline_small_functions): Fix typos.
5490 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5492 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5493 (s390_can_use_return_insn): Declare.
5494 * config/s390/s390.h (EPILOGUE_USES): Define.
5495 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5497 (s390_chunkify_start): Handle return JUMP_LABELs.
5498 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5499 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5500 (s390_can_use_return_insn): New functions.
5501 (s390_fix_long_loop_prediction): Handle conditional returns.
5502 (TARGET_SET_UP_BY_PROLOGUE): Define.
5503 * config/s390/s390.md (ANY_RETURN): New code iterator.
5504 (*creturn, *csimple_return, return, simple_return): New patterns.
5506 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5508 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5509 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5510 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5511 REG_CFA_RESTORE list when deciding not to restore a register.
5513 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5515 * config/s390/s390.c: Include tree-pass.h and context.h.
5516 (s390_early_mach): New function, split out from...
5517 (s390_emit_prologue): ...here.
5518 (pass_data_s390_early_mach): New pass structure.
5519 (pass_s390_early_mach): New class.
5520 (s390_option_override): Create and register early_mach pass.
5521 Move to end of file.
5523 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5525 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5526 to match for the exit block.
5528 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5530 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5531 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5532 Reject misaligned operands.
5534 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5536 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5538 2014-02-07 Richard Biener <rguenther@suse.de>
5541 * gimple-low.c (lower_builtin_posix_memalign): New function.
5542 (lower_stmt): Call it to lower posix_memalign in a way
5543 to make alignment info accessible.
5545 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5548 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5549 __builtin_setjmp_receiver.
5551 2014-02-07 Richard Biener <rguenther@suse.de>
5554 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5555 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5556 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5557 Handle BUILT_IN_POSIX_MEMALIGN.
5558 (find_func_clobbers): Likewise.
5559 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5560 (call_may_clobber_ref_p_1): Likewise.
5562 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5565 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5568 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5571 * lto-cgraph.c (lto_output_node): Use
5572 symtab_get_symbol_partitioning_class.
5573 (lto_output_varpool_node): likewise.
5574 (symtab_get_symbol_partitioning_class): Move here from
5576 * cgraph.h (symbol_partitioning_class): Likewise.
5577 (symtab_get_symbol_partitioning_class): Declare.
5579 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5581 * ggc.h (ggc_internal_cleared_alloc): New macro.
5582 * vec.h (vec_safe_copy): Handle memory stats.
5583 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5584 * target-globals.c (save_target_globals): Likewise.
5586 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5589 * expr.c (emit_move_resolve_push): Export; be bit more selective
5590 on when to clear alias set.
5591 * expr.h (emit_move_resolve_push): Declare.
5592 * function.h (struct function): Add tail_call_marked.
5593 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5594 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5595 * config/i386/i386.md (TImode move expander): De not call
5597 (FP push expanders): Preserve memory attributes.
5598 * config/i386/sse.md (push<mode>1): Remove.
5599 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5600 (ix86_expand_push): Remove.
5601 * config/i386/mmx.md (push<mode>1): Remove.
5603 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5605 PR rtl-optimization/60030
5606 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5607 lopart with paradoxical subreg before shifting it up by hprec.
5609 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5611 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5612 Remove extra newline at end of file.
5613 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5614 (arm_issue_rate): Handle cortexa57.
5615 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5616 (cortex-a57.cortex-a53): Likewise.
5618 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5621 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5622 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5624 (arm_expand_prologue): Adjust all callers.
5625 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5626 info, registers also at the lowest numbered registers side. Use
5627 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5631 * var-tracking.c (adjust_mems): Before adding a SET to
5632 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5634 2014-02-06 Alan Modra <amodra@gmail.com>
5637 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5638 change SDmode to DDmode when lra_in_progress.
5640 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5643 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5644 free_data_ref on the dr first, and before goto again also set dr
5645 to the next dr. For simd_lane_access, free old datarefs[i] before
5646 overwriting it. For get_vectype_for_scalar_type failure, don't
5647 free_data_ref if simd_lane_access.
5649 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5652 * tree.h (opts_for_fn): New inline function.
5653 (opt_for_fn): Define.
5654 * config/i386/i386.c (ix86_function_regparm): Use
5655 opt_for_fn (decl, optimize) instead of optimize.
5657 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5659 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5660 for SYMBOL_REF in large memory model.
5662 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5664 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5666 (cortex-a57): Likewise.
5667 (cortex-a57.cortex-a53): Likewise.
5669 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5670 Kugan Vivekanandarajah <kuganv@linaro.org>
5672 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5674 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5676 2014-02-06 Richard Biener <rguenther@suse.de>
5678 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5679 set_loop_copy and initialize_original_copy_tables.
5681 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5683 * config/aarch64/aarch64-simd.md
5684 (aarch64_ashr_simddi): Change QI to SI.
5686 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5687 Jakub Jelinek <jakub@redhat.com>
5690 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5693 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5695 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5696 CODE_FOR_altivec_vpku[hw]um to
5697 CODE_FOR_altivec_vpku[hw]um_direct.
5698 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5699 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5700 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5701 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5703 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5705 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5706 generation for -maltivec=be.
5707 (altivec_vsumsws): Simplify redundant test.
5709 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5711 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5712 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5713 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5714 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5715 gen_altivec_vpkuwum.
5716 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5718 (altivec_vpks<VI_char>ss): Likewise.
5719 (altivec_vpks<VI_char>us): Likewise.
5720 (altivec_vpku<VI_char>us): Likewise.
5721 (altivec_vpku<VI_char>um): Likewise.
5722 (altivec_vpku<VI_char>um_direct): New (copy of
5723 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5725 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5726 target is little endian and -maltivec=be is not specified.
5727 (*altivec_vupkhs<VU_char>_direct): New (copy of
5728 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
5729 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5730 target is little endian and -maltivec=be is not specified.
5731 (*altivec_vupkls<VU_char>_direct): New (copy of
5732 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
5733 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5734 little endian and -maltivec=be is not specified.
5735 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5736 little endian and -maltivec=be is not specified.
5738 2014-02-05 Richard Henderson <rth@redhat.com>
5741 * combine-stack-adj.c: Revert r206943.
5742 * sched-int.h (struct deps_desc): Add last_args_size.
5743 * sched-deps.c (init_deps): Initialize it.
5744 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5745 contain REG_ARGS_SIZE notes.
5747 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5749 * lto-cgraph.c (asm_nodes_output): Make global.
5750 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
5751 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
5752 (driver_handle_option): Handle OPT_fwpa.
5754 2014-02-05 Jakub Jelinek <jakub@redhat.com>
5757 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5758 a comment typo and formatting issue. If odr_hash hasn't been
5759 created, return vNULL and set *completep to false.
5762 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
5763 bb with no successors.
5765 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
5768 * doc/invoke.texi (-march): Clarify documentation for ARM.
5772 2014-02-05 Richard Biener <rguenther@suse.de>
5774 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
5775 when not vectorizing because of too many alias checks.
5776 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
5777 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
5779 2014-02-05 Nick Clifton <nickc@redhat.com>
5781 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
5782 accept extended registers in any mode when compiling for the MN10300.
5784 2014-02-05 Yury Gribov <y.gribov@samsung.com>
5786 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
5787 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
5788 sanitization attributes.
5789 (can_inline_edge_p): Likewise.
5790 (sanitize_attrs_match_for_inline_p): New function.
5792 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5794 * ipa-prop.c (detect_type_change): Shor circuit testing of
5795 type changes on THIS pointer.
5797 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
5800 * config/pa/pa.c (legitimize_tls_address): Return original address
5801 if not passed a SYMBOL_REF rtx.
5802 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
5804 (pa_emit_move_sequence): Simplify TLS source operands.
5805 (pa_legitimate_constant_p): Reject all TLS constants.
5806 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
5807 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
5809 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5811 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
5812 groups when we know they are controlled by LTO.
5813 * varasm.c (default_binds_local_p_1): If object is in other partition,
5814 it will be resolved locally.
5816 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5818 * config/host-linux.c (linux_gt_pch_use_address): Don't
5819 use SSIZE_MAX because it is not always defined.
5821 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
5824 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
5825 threshold for pseudo splitting.
5826 (update_ebb_live_info): Process call argument hard registers and
5827 hard registers from insn definition too.
5828 (max_small_class_regs_num): New constant.
5829 (inherit_in_ebb): Update live hard regs through EBBs. Update
5830 reloads_num only for small register classes. Don't split for
5833 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
5836 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
5839 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5841 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
5844 2014-02-04 Marek Polacek <polacek@redhat.com>
5846 * gdbinit.in (pel): Define.
5848 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5850 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
5853 2014-02-04 Richard Biener <rguenther@suse.de>
5856 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
5857 in function context local.
5858 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
5859 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
5860 similar to LTO_imported_decl_ref.
5862 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5864 PR tree-optimization/60002
5865 * cgraphclones.c (build_function_decl_skip_args): Clear
5868 PR tree-optimization/60023
5869 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
5870 false to gsi_replace.
5871 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
5872 has been in some EH region and vec_stmt could throw, add
5873 vec_stmt into the same EH region.
5874 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
5875 has no lhs, ignore it.
5876 * internal-fn.c (expand_MASK_LOAD): Likewise.
5879 * tree-inline.c (copy_forbidden): Fail for
5880 __attribute__((optimize (0))) functions.
5883 * omp-low.c (simd_clone_struct_copy): If from->inbranch
5884 is set, copy one less argument.
5885 (expand_simd_clones): Don't subtract clone_info->inbranch
5886 from simd_clone_struct_alloc argument.
5888 PR rtl-optimization/57915
5889 * recog.c (simplify_while_replacing): If all unary/binary/relational
5890 operation arguments are constant, attempt to simplify those.
5893 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
5894 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
5896 2014-02-04 Richard Biener <rguenther@suse.de>
5898 PR tree-optimization/60012
5899 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
5900 TBAA disambiguation to all DDRs.
5902 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5905 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
5906 (LINK_SPEC): Use it for -shared, -shared-libgcc.
5908 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5911 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
5913 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5915 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
5916 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
5918 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5921 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
5922 to figure out targets of polymorphic calls with known decl.
5923 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5924 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
5925 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
5926 (get_polymorphic_call_info): ... here.
5927 (get_polymorphic_call_info_from_invariant): New function.
5929 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5931 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
5932 lookup via vtable pointer; check for type consistency
5933 and turn inconsitent facts into UNREACHABLE.
5934 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5935 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
5936 type inconsistent querries; return UNREACHABLE instead.
5938 2014-02-03 Richard Henderson <rth@twiddle.net>
5941 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
5942 already processed this node.
5943 (normalize_one_pred_1): Pass along mark_set.
5944 (normalize_one_pred): Create and destroy a pointer_set_t.
5945 (normalize_one_pred_chain): Likewise.
5947 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
5949 PR gcov-profile/58602
5950 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
5952 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5955 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
5956 -fno-devirtualize; try to devirtualize by the knowledge of
5957 virtual table pointer given by aggregate propagation.
5958 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5959 (ipa_print_node_jump_functions): Dump also offset that
5960 is relevant for polymorphic calls.
5961 (determine_known_aggregate_parts): Add arg_type parameter; use it
5962 instead of determining the type from pointer type.
5963 (ipa_compute_jump_functions_for_edge): Update call of
5964 determine_known_aggregate_parts.
5965 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
5966 (gimple_get_virt_method_for_binfo): ... here; simplify using
5967 vtable_pointer_value_to_vtable.
5968 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
5969 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
5970 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
5971 (vtable_pointer_value_to_vtable): Break out from ...; handle also
5973 (vtable_pointer_value_to_binfo): ... here.
5974 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
5976 2014-02-03 Teresa Johnson <tejohnson@google.com>
5978 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
5979 redef of outer loop index variable.
5981 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
5985 * doc/extend.texi (Function Attributes): Typo.
5987 2014-02-03 Cong Hou <congh@google.com>
5989 PR tree-optimization/60000
5990 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
5991 if the vectorized statement is a store. A store statement can only
5992 appear at the end of pattern statements.
5994 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5996 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
5997 (ix86_option_override_internal): Default long double to 64-bit for
5998 32-bit Bionic and to 128-bit for 64-bit Bionic.
6000 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
6001 TARGET_LONG_DOUBLE_128 is true.
6002 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
6004 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
6005 (mlong-double-64): Negate -mlong-double-128.
6006 (mlong-double-128): New option.
6008 * config/i386/i386-c.c (ix86_target_macros): Define
6009 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
6011 * doc/invoke.texi: Document -mlong-double-128.
6013 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6015 PR rtl-optimization/60024
6016 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
6018 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
6020 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
6022 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6024 PR rtl-optimization/57662
6025 * sel-sched.c (code_motion_path_driver): Do not mark already not
6026 existing blocks in the visiting bitmap.
6028 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6030 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
6031 on the insn being emitted.
6033 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
6034 Will Deacon <will.deacon@arm.com>
6036 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
6038 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6040 * config/arm/arm-tables.opt: Regenerate.
6042 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6044 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
6045 for vector types other than V16QImode.
6046 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
6047 define_expand, and call altivec_expand_vec_perm_le when producing
6048 code with little endian element order.
6049 (*altivec_vperm_<mode>_internal): New insn having previous
6050 behavior of altivec_vperm_<mode>.
6051 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
6052 altivec_expand_vec_perm_le when producing code with little endian
6054 (*altivec_vperm_<mode>_uns_internal): New insn having previous
6055 behavior of altivec_vperm_<mode>_uns.
6057 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6059 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
6060 (altivec_vsumsws): Add handling for -maltivec=be with a little
6062 (altivec_vsumsws_direct): New.
6063 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
6064 gen_altivec_vsumsws.
6066 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6068 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
6069 vtable_pointer_value_to_binfo): New functions.
6070 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
6071 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
6073 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
6075 * config/nios2/nios2.md (load_got_register): Initialize GOT
6076 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
6077 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
6079 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6081 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
6082 preserverd by passthrough, do not propagate the type.
6084 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6086 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
6087 (mips_atomic_assign_expand_fenv): New function.
6088 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
6090 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6092 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6093 (__builtin_mips_set_fcsr): Likewise.
6094 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6095 MIPS_USI_FTYPE_VOID.
6096 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6097 (mips16_expand_set_fcsr): Likewise.
6098 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6099 (mips16_set_fcsr_stub): Likewise.
6100 (mips16_get_fcsr_one_only_stub): New class.
6101 (mips16_set_fcsr_one_only_stub): Likewise.
6102 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6103 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6104 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6105 (hard_float): New availability predicate.
6106 (mips_builtins): Add get_fcsr and set_fcsr.
6107 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6108 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6109 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6110 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6111 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6114 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6116 * config/mips/mips.c (mips_one_only_stub): New class.
6117 (mips_need_mips16_rdhwr_p): Replace with...
6118 (mips16_rdhwr_stub): ...this new variable.
6119 (mips16_stub_call_address): New function.
6120 (mips16_rdhwr_one_only_stub): New class.
6121 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6122 (mips_output_mips16_rdhwr): Delete.
6123 (mips_finish_stub): New function.
6124 (mips_code_end): Use it to handle rdhwr stubs.
6126 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6129 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6130 when calculating size of integer atomic types.
6132 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6134 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6136 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6138 PR tree-optimization/60003
6139 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6140 * profile.c (branch_prob): Use gimple_call_builtin_p
6141 to check for BUILT_IN_SETJMP_RECEIVER.
6142 * tree-inline.c (copy_bb): Call notice_special_calls.
6144 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6147 * lra-constraints.c (process_alt_operands): Update reload_sum only
6150 2014-01-31 Richard Henderson <rth@redhat.com>
6153 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6154 until after else_eh is processed.
6156 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6158 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6159 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6160 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6161 in smmintrin.h, remove them.
6162 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6163 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6164 * config/i386/i386.md (ROUND_SAE): Fix value.
6165 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6166 (const48_operand): New.
6167 * config/i386/subst.md (round), (round_expand): Use
6168 const_4_or_8_to_11_operand.
6169 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6171 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6173 * config/i386/constraints.md (Yk): Swap meaning with k.
6174 * config/i386/i386.md (movhi_internal): Change Yk to k.
6175 (movqi_internal): Ditto.
6176 (*k<logic><mode>): Ditto.
6179 (kandn<mode>): Ditto.
6180 (*<code>hi_1): Ditto.
6181 (*<code>qi_1): Ditto.
6182 (kxnor<mode>): Ditto.
6183 (kortestzhi): Ditto.
6184 (kortestchi): Ditto.
6186 (*one_cmplhi2_1): Ditto.
6187 (*one_cmplqi2_1): Ditto.
6188 * config/i386/sse.md (): Change k to Yk.
6189 (avx512f_load<mode>_mask): Ditto.
6190 (avx512f_blendm<mode>): Ditto.
6191 (avx512f_store<mode>_mask): Ditto.
6192 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6193 (avx512f_storedqu<mode>_mask): Ditto.
6194 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6196 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6197 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6198 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6199 (avx512f_maskcmp<mode>3): Ditto.
6200 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6201 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6202 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6203 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6204 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6205 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6206 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6207 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6208 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6209 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6210 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6211 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6212 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6213 (vec_extract_lo_<mode>_maskm): Ditto.
6214 (vec_extract_hi_<mode>_maskm): Ditto.
6215 (avx512f_vternlog<mode>_mask): Ditto.
6216 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6217 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6218 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6219 (avx512f_<code>v8div16qi2_mask): Ditto.
6220 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6221 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6222 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6223 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6224 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6225 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6226 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6227 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6228 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6229 (avx512cd_maskb_vec_dupv8di): Ditto.
6230 (avx512cd_maskw_vec_dupv16si): Ditto.
6231 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6232 (avx512f_vpermi2var<mode>3_mask): Ditto.
6233 (avx512f_vpermi2var<mode>3_mask): Ditto.
6234 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6235 (*avx512f_gathersi<mode>): Ditto.
6236 (*avx512f_gathersi<mode>_2): Ditto.
6237 (*avx512f_gatherdi<mode>): Ditto.
6238 (*avx512f_gatherdi<mode>_2): Ditto.
6239 (*avx512f_scattersi<mode>): Ditto.
6240 (*avx512f_scatterdi<mode>): Ditto.
6241 (avx512f_compress<mode>_mask): Ditto.
6242 (avx512f_compressstore<mode>_mask): Ditto.
6243 (avx512f_expand<mode>_mask): Ditto.
6244 * config/i386/subst.md (mask): Change k to Yk.
6245 (mask_scalar_merge): Ditto.
6248 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6250 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6252 2014-01-31 Richard Biener <rguenther@suse.de>
6255 * builtins.c (fold_builtin_memory_op): Make sure to not
6256 use a floating-point mode or a boolean or enumeral type for
6259 2014-01-30 DJ Delorie <dj@redhat.com>
6261 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6262 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6263 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6264 whenever main() has an epilogue.
6266 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6268 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6269 unused variable "field".
6270 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6271 (vsx_mergeh_<mode>): Likewise.
6272 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6273 (altivec_vmrghh): Likewise.
6274 (altivec_vmrghw): Likewise.
6275 (altivec_vmrglb): Likewise.
6276 (altivec_vmrglh): Likewise.
6277 (altivec_vmrglw): Likewise.
6278 (altivec_vspltb): Add missing uses.
6279 (altivec_vsplth): Likewise.
6280 (altivec_vspltw): Likewise.
6281 (altivec_vspltsf): Likewise.
6283 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6286 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6287 frame related instructions.
6289 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6291 PR rtl-optimization/59959
6292 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6293 any reload of register whose subreg is invalid.
6295 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6297 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6298 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6299 Add missing return type - void.
6301 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6303 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6304 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6305 remove element index adjustment for endian (now handled in vsx.md
6307 (altivec_expand_vec_perm_const): Use
6308 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6309 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6310 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6311 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6312 define_expand and a new define_insn *altivec_vspltb_internal;
6313 adjust for -maltivec=be on a little endian target.
6314 (altivec_vspltb_direct): New.
6315 (altivec_vsplth): Divide into a define_expand and a new
6316 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6317 little endian target.
6318 (altivec_vsplth_direct): New.
6319 (altivec_vspltw): Divide into a define_expand and a new
6320 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6321 little endian target.
6322 (altivec_vspltw_direct): New.
6323 (altivec_vspltsf): Divide into a define_expand and a new
6324 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6325 a little endian target.
6327 2014-01-30 Richard Biener <rguenther@suse.de>
6329 PR tree-optimization/59993
6330 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6331 can propagate form the earlier stmt and avoid the transform
6332 when the intermediate result is needed.
6334 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6336 * README.Portability: Fix typo.
6338 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6340 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6341 comparison_operator with ordered_comparison_operator.
6343 2014-01-30 Nick Clifton <nickc@redhat.com>
6345 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6346 Rename to mn10300_store_multiple_regs.
6347 * config/mn10300/mn10300.c: Likewise.
6348 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6349 store_multiple_regs.
6350 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6351 Call mn10300_store_multiple_regs.
6353 2014-01-30 Nick Clifton <nickc@redhat.com>
6354 DJ Delorie <dj@redhat.com>
6356 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6357 %fp 2 to keep registers after it properly word-aligned.
6358 (rl78_alloc_physical_registers_umul): Handle the case where both
6359 input operands are the same.
6361 2014-01-30 Richard Biener <rguenther@suse.de>
6363 PR tree-optimization/59903
6364 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6367 2014-01-30 Jason Merrill <jason@redhat.com>
6370 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6373 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6375 2014-01-30 Richard Biener <rguenther@suse.de>
6377 PR tree-optimization/59951
6378 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6380 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6383 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6384 SFmode to DFmode case.
6386 2014-01-29 DJ Delorie <dj@redhat.com>
6388 * config/msp430/msp430.opt (-minrt): New.
6389 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6391 (ENDFILE_SPEC): Likewise.
6393 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6395 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6396 (estimate_function_body_sizes): Use it.
6398 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6401 * dwarf2out.c (is_cxx_auto): New.
6402 (is_base_type): Use it.
6403 (gen_type_die_with_usage): Likewise.
6405 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6407 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6408 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6409 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6410 -maltivec=be with LE targets.
6411 (vsx_mergeh_<mode>): Likewise.
6412 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6413 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6414 (altivec_vmrghb): Replace with define_expand and new
6415 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6416 (altivec_vmrghb_direct): New define_insn.
6417 (altivec_vmrghh): Replace with define_expand and new
6418 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6419 (altivec_vmrghh_direct): New define_insn.
6420 (altivec_vmrghw): Replace with define_expand and new
6421 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6422 (altivec_vmrghw_direct): New define_insn.
6423 (*altivec_vmrghsf): Adjust for endianness.
6424 (altivec_vmrglb): Replace with define_expand and new
6425 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6426 (altivec_vmrglb_direct): New define_insn.
6427 (altivec_vmrglh): Replace with define_expand and new
6428 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6429 (altivec_vmrglh_direct): New define_insn.
6430 (altivec_vmrglw): Replace with define_expand and new
6431 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6432 (altivec_vmrglw_direct): New define_insn.
6433 (*altivec_vmrglsf): Adjust for endianness.
6434 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6435 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6436 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6437 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6438 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6439 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6440 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6441 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6443 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6445 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6446 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6449 2014-01-29 Richard Biener <rguenther@suse.de>
6451 PR tree-optimization/58742
6452 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6453 associate_pointerplus_align.
6454 (associate_pointerplus_diff): New function.
6455 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6456 and associate_pointerplus_diff.
6458 2014-01-29 Richard Biener <rguenther@suse.de>
6460 * lto-streamer.h (LTO_major_version): Bump to 3.
6461 (LTO_minor_version): Reset to 0.
6463 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6465 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6466 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6467 (arm_file_start): Generate correct asm header for armv7ve.
6468 * config/arm/bpabi.h: Add multilib support for armv7ve.
6469 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6470 and cortex-a15 to armv7ve.
6471 * config/arm/t-aprofile: Add multilib support for armv7ve.
6472 * doc/invoke.texi: Document -march=armv7ve.
6474 2014-01-29 Richard Biener <rguenther@suse.de>
6476 PR tree-optimization/58742
6477 * tree-ssa-forwprop.c (associate_plusminus): Return true
6478 if we changed sth, defer EH cleanup to ...
6479 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6480 (simplify_mult): New function.
6482 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6485 PR tree-optimization/59920
6486 * tree.c (build_common_builtin_nodes): Remove
6487 __builtin_setjmp_dispatcher initialization.
6488 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6489 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6490 instead of gsi_after_labels + manually skipping debug stmts.
6491 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6492 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6493 * tree-inline.c (copy_edges_for_bb): Remove
6494 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6495 argument. Ignore computed_goto_p stmts. Don't call
6496 make_abnormal_goto_edges. If a call might need abnormal edges
6497 for non-local gotos, see if it already has an edge to
6498 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6499 with true argument, don't do anything then, otherwise add
6500 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6501 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6503 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6504 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6505 (lower_stmt): Don't set data->calls_builtin_setjmp.
6506 (lower_builtin_setjmp): Adjust comment.
6507 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6508 * tree-cfg.c (found_computed_goto): Remove.
6509 (factor_computed_gotos): Remove.
6510 (make_goto_expr_edges): Return bool, true for computed gotos.
6511 Don't call make_abnormal_goto_edges.
6512 (build_gimple_cfg): Don't set found_computed_goto, don't call
6513 factor_computed_gotos.
6514 (computed_goto_p): No longer static.
6515 (make_blocks): Don't set found_computed_goto.
6516 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6517 (make_edges): If make_goto_expr_edges returns true, push bb
6518 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6519 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6520 vector. Record mapping between bbs and OpenMP regions if there
6521 are any, adjust make_gimple_omp_edges caller. Call
6522 handle_abnormal_edges.
6523 (make_abnormal_goto_edges): Remove.
6524 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6525 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6526 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6527 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6528 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6529 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6530 filling *region also set *region_idx to (*region)->entry->index.
6533 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6534 For REGs set ORIGINAL_REGNO.
6536 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6538 * doc/md.texi: Mention that a target shouldn't implement
6539 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6542 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6544 PR tree-optimization/59594
6545 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6546 a copy of the datarefs vector rather than the vector itself.
6548 2014-01-28 Jason Merrill <jason@redhat.com>
6551 * dwarf2out.c (auto_die): New static.
6552 (gen_type_die_with_usage): Handle C++1y 'auto'.
6553 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6556 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6559 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6560 (SPEC_X32): Likewise.
6561 (SPEC_64): Likewise.
6562 * config/i386/i386.c (ix86_option_override_internal): Turn off
6563 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6565 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6566 * config/i386/i386.h (TARGET_16BIT): New macro.
6567 (TARGET_16BIT_P): Likewise.
6568 * config/i386/i386.opt: Add m16.
6569 * doc/invoke.texi: Document -m16.
6571 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6573 PR preprocessor/59935
6574 * input.c (location_get_source_line): Bail out on when line number
6575 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6577 2014-01-28 Richard Biener <rguenther@suse.de>
6579 PR tree-optimization/58742
6580 * tree-ssa-forwprop.c (associate_plusminus): Handle
6581 pointer subtraction of the form (T)(P + A) - (T)P.
6583 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6585 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6588 2014-01-28 Richard Biener <rguenther@suse.de>
6591 2014-01-28 Richard Biener <rguenther@suse.de>
6593 PR rtl-optimization/45364
6594 PR rtl-optimization/59890
6595 * var-tracking.c (local_get_addr_clear_given_value): Handle
6596 already cleared slot.
6597 (val_reset): Handle not allocated local_get_addr_cache.
6598 (vt_find_locations): Use post-order on the inverted CFG.
6600 2014-01-28 Richard Biener <rguenther@suse.de>
6602 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6604 2014-01-28 Richard Biener <rguenther@suse.de>
6606 PR rtl-optimization/45364
6607 PR rtl-optimization/59890
6608 * var-tracking.c (local_get_addr_clear_given_value): Handle
6609 already cleared slot.
6610 (val_reset): Handle not allocated local_get_addr_cache.
6611 (vt_find_locations): Use post-order on the inverted CFG.
6613 2014-01-28 Alan Modra <amodra@gmail.com>
6615 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6616 * configure.ac <recursive call for build != host>: Define
6617 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6618 and LD_FOR_BUILD too.
6619 * configure: Regenerate.
6621 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6623 * config/i386/i386.c (get_builtin_code_for_version): Separate
6624 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6625 Broadwell from Haswell.
6627 2014-01-27 Steve Ellcey <sellcey@mips.com>
6629 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6630 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6631 * config/mips/mips.c (mips_option_override): Change setting
6633 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6634 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6635 Change from Mask to Var.
6637 2014-01-27 Jeff Law <law@redhat.com>
6639 * ipa-inline.c (inline_small_functions): Fix typo.
6641 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6643 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6644 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6645 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6646 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6647 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6648 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6649 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6650 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6651 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6652 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6653 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6654 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6655 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6656 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6657 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6658 (_mm512_storeu_epi64): Ditto.
6659 (_mm512_cmpge_epi32_mask): Ditto.
6660 (_mm512_cmpge_epu32_mask): Ditto.
6661 (_mm512_cmpge_epi64_mask): Ditto.
6662 (_mm512_cmpge_epu64_mask): Ditto.
6663 (_mm512_cmple_epi32_mask): Ditto.
6664 (_mm512_cmple_epu32_mask): Ditto.
6665 (_mm512_cmple_epi64_mask): Ditto.
6666 (_mm512_cmple_epu64_mask): Ditto.
6667 (_mm512_cmplt_epi32_mask): Ditto.
6668 (_mm512_cmplt_epu32_mask): Ditto.
6669 (_mm512_cmplt_epi64_mask): Ditto.
6670 (_mm512_cmplt_epu64_mask): Ditto.
6671 (_mm512_cmpneq_epi32_mask): Ditto.
6672 (_mm512_cmpneq_epu32_mask): Ditto.
6673 (_mm512_cmpneq_epi64_mask): Ditto.
6674 (_mm512_cmpneq_epu64_mask): Ditto.
6675 (_mm512_expand_pd): Ditto.
6676 (_mm512_expand_ps): Ditto.
6677 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6678 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6679 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6680 * config/i386/i386.c (ix86_builtins): Add
6681 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6682 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6683 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6684 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6685 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6686 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6687 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6688 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6689 IX86_BUILTIN_PMOVUSQW512_MEM.
6690 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6691 __builtin_ia32_pmovsqd512mem_mask,
6692 __builtin_ia32_pmovqd512mem_mask,
6693 __builtin_ia32_pmovusqw512mem_mask,
6694 __builtin_ia32_pmovsqw512mem_mask,
6695 __builtin_ia32_pmovqw512mem_mask,
6696 __builtin_ia32_pmovusdw512mem_mask,
6697 __builtin_ia32_pmovsdw512mem_mask,
6698 __builtin_ia32_pmovdw512mem_mask,
6699 __builtin_ia32_pmovqb512mem_mask,
6700 __builtin_ia32_pmovusqb512mem_mask,
6701 __builtin_ia32_pmovsqb512mem_mask,
6702 __builtin_ia32_pmovusdb512mem_mask,
6703 __builtin_ia32_pmovsdb512mem_mask,
6704 __builtin_ia32_pmovdb512mem_mask.
6705 (bdesc_args): Add __builtin_ia32_expanddf512,
6706 __builtin_ia32_expandsf512.
6707 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6708 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6709 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6710 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6711 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6712 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6713 (avx512f_<code>v8div16qi2_mask_store): This.
6714 (avx512f_expand<mode>): New.
6716 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6718 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
6720 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6721 (_mm512_prefetch_i32scatter_pd): Ditto.
6722 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6723 (_mm512_prefetch_i64scatter_pd): Ditto.
6724 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6725 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6726 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6727 (_mm512_prefetch_i32scatter_ps): Ditto.
6728 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6729 (_mm512_prefetch_i64scatter_ps): Ditto.
6730 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6731 * config/i386/i386-builtin-types.def: Define
6732 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6733 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
6734 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6735 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6736 IX86_BUILTIN_SCATTERPFQPD.
6737 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6738 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6739 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6740 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6741 __builtin_ia32_scatterpfqps.
6742 (ix86_expand_builtin): Expand new built-ins.
6743 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6744 fix memory access data type.
6745 (*avx512pf_gatherpf<mode>_mask): Ditto.
6746 (*avx512pf_gatherpf<mode>): Ditto.
6747 (avx512pf_scatterpf<mode>): Ditto.
6748 (*avx512pf_scatterpf<mode>_mask): Ditto.
6749 (*avx512pf_scatterpf<mode>): Ditto.
6750 (GATHER_SCATTER_SF_MEM_MODE): New.
6751 (avx512pf_gatherpf<mode>df): Ditto.
6752 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6753 (*avx512pf_scatterpf<mode>df): Ditto.
6755 2014-01-27 Jakub Jelinek <jakub@redhat.com>
6758 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
6759 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
6762 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6764 * common/config/arm/arm-common.c
6765 (arm_rewrite_mcpu): Handle multiple names.
6767 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6769 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6771 * gimple-builder.h (create_gimple_tmp): Delete.
6773 2014-01-27 Christian Bruel <christian.bruel@st.com>
6775 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
6778 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
6780 * config/pa/pa.md (call): Generate indirect long calls to non-local
6781 functions when outputing 32-bit code.
6782 (call_value): Likewise except for special call to buggy powf function.
6784 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
6785 portable runtime and PIC indirect calls.
6786 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
6787 and PIC call sequences. Use ldo instead of blr to set return register
6788 in PIC call sequence.
6790 2014-01-25 Walter Lee <walt@tilera.com>
6792 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
6793 avoid clobbering a live register.
6795 2014-01-25 Walter Lee <walt@tilera.com>
6797 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
6798 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
6799 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
6800 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
6802 2014-01-25 Walter Lee <walt@tilera.com>
6804 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
6805 arguments on even registers.
6806 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
6808 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
6809 (BIGGEST_ALIGNMENT): Ditto.
6810 (BIGGEST_FIELD_ALIGNMENT): Ditto.
6812 2014-01-25 Walter Lee <walt@tilera.com>
6814 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
6815 insns before bundling.
6816 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
6818 2014-01-25 Walter Lee <walt@tilera.com>
6820 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
6821 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
6822 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
6824 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6826 * config/mips/constraints.md (kl): Delete.
6827 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
6828 define expands, using...
6829 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
6830 instructions for MIPS16.
6831 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
6832 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
6834 2014-01-25 Walter Lee <walt@tilera.com>
6836 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
6840 2014-01-25 Walter Lee <walt@tilera.com>
6842 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
6843 (TARGET_EXPAND_TO_RTL_HOOK): Define.
6845 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6847 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
6850 2014-01-25 Jakub Jelinek <jakub@redhat.com>
6852 * print-rtl.c (in_call_function_usage): New var.
6853 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
6854 EXPR_LIST mode as mode and not as reg note name.
6857 * cfgloopmanip.c (copy_loop_info): If
6858 loop->warned_aggressive_loop_optimizations, make sure
6859 the flag is set in target loop too.
6861 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
6863 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
6865 * builtins.def: Likewise.
6866 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
6867 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
6868 * ira.c (ira_setup_eliminable_regset): Likewise.
6869 * omp-low.c (gate_expand_omp): Likewise.
6870 (execute_lower_omp): Likewise.
6871 (diagnose_sb_0): Likewise.
6872 (gate_diagnose_omp_blocks): Likewise.
6873 (simd_clone_clauses_extract): Likewise.
6876 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6878 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
6879 correction for little endian...
6880 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
6883 2014-01-24 Jeff Law <law@redhat.com>
6885 PR tree-optimization/59919
6886 * tree-vrp.c (find_assert_locations_1): Do not register asserts
6887 for non-returning calls.
6889 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
6891 * common/config/aarch64/aarch64-common.c
6892 (aarch64_rewrite_mcpu): Handle multiple names.
6893 * config/aarch64/aarch64.h
6894 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6896 2014-01-24 Dodji Seketeli <dodji@redhat.com>
6898 * input.c (add_file_to_cache_tab): Handle the case where fopen
6901 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
6904 * config/i386/i386.md (pushsf splitter): Get stack adjustment
6905 from push operand if code of push isn't PRE_DEC.
6907 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6910 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
6911 -mquad-memory-atomic. Update -mquad-memory documentation to say
6912 it is only used for non-atomic loads/stores.
6914 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
6915 -mquad-memory or -mquad-memory-atomic switches.
6917 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
6918 -mquad-memory-atomic to ISA 2.07 support.
6920 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
6921 to separate support of normal quad word memory operations (ldq, stq)
6922 from the atomic quad word memory operations.
6924 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
6925 support to separate non-atomic quad word operations from atomic
6926 quad word operations. Disable non-atomic quad word operations in
6927 little endian mode so that we don't have to swap words after the
6928 load and before the store.
6929 (quad_load_store_p): Add comment about atomic quad word support.
6930 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
6931 options printed with -mdebug=reg.
6933 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
6934 -mquad-memory-atomic as the test for whether we have quad word
6935 atomic instructions.
6936 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
6937 or -mp8-vector are used, allow byte/half-word atomic operations.
6939 * config/rs6000/sync.md (load_lockedti): Insure that the address
6940 is a proper indexed or indirect address for the lqarx instruction.
6941 On little endian systems, swap the hi/lo registers after the lqarx
6943 (load_lockedpti): Use indexed_or_indirect_operand predicate to
6944 insure the address is valid for the lqarx instruction.
6945 (store_conditionalti): Insure that the address is a proper indexed
6946 or indirect address for the stqcrx. instruction. On little endian
6947 systems, swap the hi/lo registers before doing the stqcrx.
6949 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
6950 insure the address is valid for the stqcrx. instruction.
6952 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
6953 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
6954 type of quad memory support is available.
6956 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
6959 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
6960 there is a danger of looping.
6962 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
6964 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
6965 force flag_ira_loop_pressure if set via command line.
6967 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
6969 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
6970 (ashr_simd): New builtin handling DI mode.
6971 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
6972 (aarch64_sshr_simddi): New match pattern.
6973 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
6974 (vshrd_n_s64): Likewise.
6975 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
6977 2014-01-23 Nick Clifton <nickc@redhat.com>
6979 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
6980 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
6981 favour of mcu specific scripts.
6982 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
6985 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
6986 Alex Velenko <Alex.Velenko@arm.com>
6988 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
6989 (vaddv_s16): Likewise.
6990 (vaddv_s32): Likewise.
6991 (vaddv_u8): Likewise.
6992 (vaddv_u16): Likewise.
6993 (vaddv_u32): Likewise.
6994 (vaddvq_s8): Likewise.
6995 (vaddvq_s16): Likewise.
6996 (vaddvq_s32): Likewise.
6997 (vaddvq_s64): Likewise.
6998 (vaddvq_u8): Likewise.
6999 (vaddvq_u16): Likewise.
7000 (vaddvq_u32): Likewise.
7001 (vaddvq_u64): Likewise.
7002 (vaddv_f32): Likewise.
7003 (vaddvq_f32): Likewise.
7004 (vaddvq_f64): Likewise.
7005 (vmaxv_f32): Likewise.
7006 (vmaxv_s8): Likewise.
7007 (vmaxv_s16): Likewise.
7008 (vmaxv_s32): Likewise.
7009 (vmaxv_u8): Likewise.
7010 (vmaxv_u16): Likewise.
7011 (vmaxv_u32): Likewise.
7012 (vmaxvq_f32): Likewise.
7013 (vmaxvq_f64): Likewise.
7014 (vmaxvq_s8): Likewise.
7015 (vmaxvq_s16): Likewise.
7016 (vmaxvq_s32): Likewise.
7017 (vmaxvq_u8): Likewise.
7018 (vmaxvq_u16): Likewise.
7019 (vmaxvq_u32): Likewise.
7020 (vmaxnmv_f32): Likewise.
7021 (vmaxnmvq_f32): Likewise.
7022 (vmaxnmvq_f64): Likewise.
7023 (vminv_f32): Likewise.
7024 (vminv_s8): Likewise.
7025 (vminv_s16): Likewise.
7026 (vminv_s32): Likewise.
7027 (vminv_u8): Likewise.
7028 (vminv_u16): Likewise.
7029 (vminv_u32): Likewise.
7030 (vminvq_f32): Likewise.
7031 (vminvq_f64): Likewise.
7032 (vminvq_s8): Likewise.
7033 (vminvq_s16): Likewise.
7034 (vminvq_s32): Likewise.
7035 (vminvq_u8): Likewise.
7036 (vminvq_u16): Likewise.
7037 (vminvq_u32): Likewise.
7038 (vminnmv_f32): Likewise.
7039 (vminnmvq_f32): Likewise.
7040 (vminnmvq_f64): Likewise.
7042 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7044 * config/aarch64/aarch64-simd.md
7045 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
7046 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
7047 (*aarch64_mul3_elt<mode>): Likewise.
7048 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
7049 (*aarch64_mul3_elt_to_64v2df): Likewise.
7050 (*aarch64_mla_elt<mode>): Likewise.
7051 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7052 (*aarch64_mls_elt<mode>): Likewise.
7053 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7054 (*aarch64_fma4_elt<mode>): Likewise.
7055 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7056 (*aarch64_fma4_elt_to_64v2df): Likewise.
7057 (*aarch64_fnma4_elt<mode>): Likewise.
7058 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7059 (*aarch64_fnma4_elt_to_64v2df): Likewise.
7060 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7061 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
7062 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7063 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7064 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
7065 (aarch64_sqdmull_lane<mode>_internal): Likewise.
7066 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
7068 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
7070 * config/aarch64/aarch64-simd.md
7071 (aarch64_be_checked_get_lane<mode>): New define_expand.
7072 * config/aarch64/aarch64-simd-builtins.def
7073 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
7074 New builtin definition.
7075 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
7076 Use new safe be builtin.
7078 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7080 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
7082 (aarch64_be_st1<mode>): Likewise.
7083 (aarch_ld1<VALL:mode>): Define_expand modified.
7084 (aarch_st1<VALL:mode>): Likewise.
7085 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
7086 (UNSPEC_ST1): Likewise.
7088 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
7090 * config/microblaze/microblaze.md: Add trap insn and attribute
7092 2014-01-23 Dodji Seketeli <dodji@redhat.com>
7094 PR preprocessor/58580
7095 * input.h (location_get_source_line): Take an additional line_size
7097 (void diagnostics_file_cache_fini): Declare new function.
7098 * input.c (struct fcache): New type.
7099 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7100 New static constants.
7101 (diagnostic_file_cache_init, total_lines_num)
7102 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7103 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7104 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7105 (get_next_line, read_next_line, goto_next_line, read_line_num):
7106 New static function definitions.
7107 (diagnostic_file_cache_fini): New function.
7108 (location_get_source_line): Take an additional output line_len
7109 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7111 * diagnostic.c (diagnostic_finish): Call
7112 diagnostic_file_cache_fini.
7113 (adjust_line): Take an additional input parameter for the length
7114 of the line, rather than calculating it with strlen.
7115 (diagnostic_show_locus): Adjust the use of
7116 location_get_source_line and adjust_line with respect to their new
7117 signature. While displaying a line now, do not stop at the first
7118 null byte. Rather, display the zero byte as a space and keep
7119 going until we reach the size of the line.
7120 * Makefile.in: Add vec.o to OBJS-libcommon
7122 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7123 Ilya Tocar <ilya.tocar@intel.com>
7125 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7126 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7127 (__builtin_ia32_kmov16): Ditto.
7128 * config/i386/i386.md (UNSPEC_KMOV): New.
7131 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7133 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7134 (_mm512_storeu_si512): Ditto.
7136 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7139 * rtl.h (get_referenced_operands): Declare.
7140 * recog.c (get_referenced_operands): New function.
7141 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7142 operands have been referenced when recording LO_SUM references.
7144 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7146 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7148 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7150 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7151 Enable for generic and recent AMD targets.
7153 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7155 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7156 ARG_SIZE note when adjustment was eliminated.
7158 2014-01-22 Jeff Law <law@redhat.com>
7160 PR tree-optimization/59597
7161 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7162 in file. Accept new argument REGISTERING and use it to modify
7163 dump output appropriately.
7164 (register_jump_thread): Corresponding changes.
7165 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7166 thread paths involving joiner blocks. Add code to dump cancelled
7167 jump threading paths.
7169 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7171 PR rtl-optimization/59477
7172 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7173 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7175 2014-01-22 Tom Tromey <tromey@redhat.com>
7177 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7179 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7181 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7183 PR rtl-optimization/59896
7184 * lra-constraints.c (process_alt_operands): Check unused note for
7185 matched operands of insn with no output reloads.
7187 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7189 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7190 (mips_move_from_gpr_cost): Likewise.
7192 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7194 PR rtl-optimization/59858
7195 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7196 ira_class_hard_regs_num.
7197 (process_alt_operands): Increase reject for dying matched operand.
7199 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7202 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7203 smaller than size, perform several stores or loads and stores
7204 at dst + count - size to store or copy all of size bytes, rather
7205 than just last modesize bytes.
7207 2014-01-20 DJ Delorie <dj@redhat.com>
7209 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7210 that CLOBBERs are REGs before propogating their values.
7212 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7215 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7216 (cgraph_inline_failed_type): New function.
7217 * cgraph.h (DEFCIFCODE): Add type.
7218 (cgraph_inline_failed_type_t): New enum.
7219 (cgraph_inline_failed_type): New prototype.
7220 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7221 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7222 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7223 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7224 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7225 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7226 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7227 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7228 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7229 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7230 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7231 OPTIMIZATION_MISMATCH.
7232 * tree-inline.c (expand_call_inline): Emit errors during
7233 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7235 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7238 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7239 mode attribute in insn output.
7241 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7243 * output.h (output_constant): Delete.
7244 * varasm.c (output_constant): Make private.
7246 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7248 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7250 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7253 * tree.h (fold_builtin_strcat): New prototype.
7254 * builtins.c (fold_builtin_strcat): No longer static. Add len
7255 argument, if non-NULL, don't call c_strlen. Optimize
7256 directly into __builtin_memcpy instead of __builtin_strcpy.
7257 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7258 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7260 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7262 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7263 for SImode_address_operand operands, having only a REG argument.
7265 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7267 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7268 loader name using mbig-endian.
7269 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7271 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7273 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7277 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7279 * config/aarch64/aarch64-protos.h
7280 (aarch64_cannot_change_mode_class_ptr): Declare.
7281 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7282 aarch64_cannot_change_mode_class_ptr): New.
7283 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7284 backend hook aarch64_cannot_change_mode_class.
7286 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7288 * common/config/aarch64/aarch64-common.c
7289 (aarch64_handle_option): Don't handle any option order logic here.
7290 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7291 selected_cpu, warn on architecture version mismatch.
7292 (aarch64_override_options): Fix parsing order for option strings.
7294 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7295 Iain Sandoe <iain@codesourcery.com>
7298 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7299 warning. Amend comment to reflect current functionality.
7301 2014-01-20 Richard Biener <rguenther@suse.de>
7304 * builtins.c (fold_builtin_strcat): Remove case better handled
7305 by tree-ssa-strlen.c.
7307 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7309 * config/aarch64/aarch64.opt
7310 (mcpu, march, mtune): Make case-insensitive.
7312 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7315 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7316 if operands[1] is a REG or ZERO_EXTEND of a REG.
7318 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7320 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7322 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7324 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7325 long non-pic millicode calls.
7327 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7329 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7331 2014-01-19 Kito Cheng <kito@0xlab.org>
7333 * builtins.c (expand_movstr): Check movstr expand done or fail.
7335 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7336 H.J. Lu <hongjiu.lu@intel.com>
7339 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7340 to DImode for zero-extended addresses.
7342 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7344 PR rtl-optimization/57763
7345 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7346 on the new indirect jump_insn and increment LABEL_NUSES (label).
7348 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7352 * config.gcc (x86_archs): New variable.
7353 (x86_64_archs): Likewise.
7354 (x86_cpus): Likewise.
7355 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7356 --with-arch/--with-cpu= options.
7357 Support --with-arch=/--with-cpu={nehalem,westmere,
7358 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7360 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7362 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7363 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7365 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7367 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7369 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7372 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7373 clear cpp_get_options (parse_in)->warn_unused_macros for
7374 ix86_target_macros_internal with cpp_define.
7376 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7378 * jump.c (delete_related_insns): Keep (use (insn))s.
7379 * reorg.c (redundant_insn): Check for barriers too.
7381 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7383 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7385 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7387 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7388 call to $$dyncall when TARGET_LONG_CALLS is true.
7390 2014-01-17 Jeff Law <law@redhat.com>
7392 * ree.c (combine_set_extension): Temporarily disable test for
7393 changing number of hard registers.
7395 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7398 * ipa-inline-analysis.c (inline_free_summary):
7399 Do not free summary of aliases.
7401 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7404 * gimplify.c (gimplify_expr): Use create_tmp_var
7405 instead of create_tmp_var_raw. If cond doesn't have
7406 integral type, don't add the IFN_ANNOTATE builtin at all.
7408 2014-01-17 Martin Jambor <mjambor@suse.cz>
7411 * ipa-cp.c (prev_edge_clone): New variable.
7412 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7413 Also resize prev_edge_clone vector.
7414 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7415 (ipcp_edge_removal_hook): New function.
7416 (ipcp_driver): Register ipcp_edge_removal_hook.
7418 2014-01-17 Andrew Pinski <apinski@cavium.com>
7419 Steve Ellcey <sellcey@mips.com>
7422 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7425 2014-01-17 Jeff Law <law@redhat.com>
7428 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7429 so that pass_ccp runs first.
7431 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7433 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7434 (ix86_adjust_cost): Use !TARGET_XXX.
7435 (do_reorder_for_imul): Likewise.
7436 (swap_top_of_ready_list): Likewise.
7437 (ix86_sched_reorder): Likewise.
7439 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7441 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7442 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7443 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7444 (intel_memset): New. Duplicate slm_memset.
7445 (intel_cost): New. Duplicate slm_cost.
7446 (m_INTEL): New macro.
7447 (processor_target_table): Add "intel".
7448 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7449 with PROCESSOR_INTEL for "intel".
7450 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7451 PROCESSOR_SILVERMONT.
7452 (ix86_issue_rate): Likewise.
7453 (ix86_adjust_cost): Likewise.
7454 (ia32_multipass_dfa_lookahead): Likewise.
7455 (swap_top_of_ready_list): Likewise.
7456 (ix86_sched_reorder): Likewise.
7457 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7458 instead of TARGET_OPT_AGU.
7459 * config/i386/i386.h (TARGET_INTEL): New.
7460 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7461 (processor_type): Add PROCESSOR_INTEL.
7462 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7463 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7465 2014-01-17 Marek Polacek <polacek@redhat.com>
7468 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7471 2014-01-17 Richard Biener <rguenther@suse.de>
7473 PR tree-optimization/46590
7474 * opts.c (default_options_table): Add entries for
7475 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7476 all enabled at -O1 but not for -Og.
7477 * common.opt (fbranch-count-reg): Remove Init(1).
7478 (fmove-loop-invariants): Likewise.
7479 (ftree-pta): Likewise.
7481 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7483 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7484 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7485 decls to at least the GCC 4.8 used alignments.
7488 * tree-nested.c (convert_nonlocal_reference_stmt,
7489 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7490 of GIMPLE_BIND stmts, adjust associated decls.
7492 2014-01-17 Richard Biener <rguenther@suse.de>
7494 PR tree-optimization/46590
7495 * vec.h (vec<>::bseach): New member function implementing
7496 binary search according to C89 bsearch.
7497 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7498 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7499 bitmap pointer again. Make accesses_in_loop a flat array.
7500 (mem_ref_obstack): New global.
7501 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7502 (mark_ref_stored): Likewise.
7503 (ref_indep_loop_p_2): Likewise.
7504 (set_ref_stored_in_loop): New helper function.
7505 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7506 (memref_free): Adjust.
7507 (record_mem_ref_loc): Simplify.
7508 (gather_mem_refs_stmt): Adjust.
7509 (sort_locs_in_loop_postorder_cmp): New function.
7510 (analyze_memory_references): Sort accesses_in_loop after
7511 loop postorder number.
7512 (find_ref_loc_in_loop_cmp): New function.
7513 (for_all_locs_in_loop): Find relevant cluster of locs in
7514 accesses_in_loop and iterate without recursion.
7515 (execute_sm): Avoid uninit warning.
7516 (struct ref_always_accessed): Simplify.
7517 (ref_always_accessed::operator ()): Likewise.
7518 (ref_always_accessed_p): Likewise.
7519 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7520 loop postorder numbers here.
7521 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7524 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7527 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7528 on decls for which assemble_alias has been called.
7530 2014-01-17 Nick Clifton <nickc@redhat.com>
7532 * config/msp430/msp430.opt: (mcpu): New option.
7533 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7534 (msp430_option_override): Parse target_cpu. If the MCU name
7535 matches a generic string, clear target_mcu.
7536 (msp430_attr): Allow numeric interrupt values up to 63.
7537 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7538 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7540 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7542 * config/msp430/msp430.md (popm): Use %J rather than %I.
7543 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7544 (addhi_cy_i): Use immediate_operand for operand 2.
7545 * doc/invoke.texi: Document -mcpu option.
7547 2014-01-17 Richard Biener <rguenther@suse.de>
7549 PR rtl-optimization/38518
7550 * df.h (df_analyze_loop): Declare.
7551 * df-core.c: Include cfgloop.h.
7552 (df_analyze_1): Split out main part of df_analyze.
7553 (df_analyze): Adjust.
7554 (loop_inverted_post_order_compute): New function.
7555 (loop_post_order_compute): Likewise.
7556 (df_analyze_loop): New function avoiding whole-function
7558 * loop-invariant.c (find_defs): Use df_analyze_loop.
7559 (find_invariants): Adjust.
7560 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7562 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7564 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7565 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7567 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7569 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7570 traversal when removing references.
7572 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7575 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7577 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7580 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7581 pushing a reload for an autoinc when we had previously reloaded an
7582 inner part of the address.
7584 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7586 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7588 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7589 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7590 when not giving up or versioning for alias only because of
7592 (vect_analyze_data_ref_dependences): Set to true.
7593 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7595 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7596 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7600 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7603 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7604 operand 0 predicate for gathers, use a new pseudo as subtarget.
7606 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7609 * lra-constraints.c (process_alt_operands): Add printing debug info.
7610 Check absence of input/output reloads for matched operands too.
7612 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7614 PR rtl-optimization/59835
7615 * ira.c (ira_init_register_move_cost): Increase cost for
7618 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7620 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7622 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7625 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7626 non-register objects. Use gen_(high/low)part more consistently.
7629 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7632 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7633 endian support, remove tests for WORDS_BIG_ENDIAN.
7634 (p8_mfvsrd_3_<mode>): Likewise.
7635 (reload_gpr_from_vsx<mode>): Likewise.
7636 (reload_gpr_from_vsxsf): Likewise.
7637 (p8_mfvsrd_4_disf): Likewise.
7639 2014-01-16 Richard Biener <rguenther@suse.de>
7641 PR rtl-optimization/46590
7642 * lcm.c (compute_antinout_edge): Use postorder iteration.
7643 (compute_laterin): Use inverted postorder iteration.
7645 2014-01-16 Nick Clifton <nickc@redhat.com>
7648 * varasm.c (output_constant): Return the number of bytes actually
7650 (output_constructor_array_range): Update the field size with the
7651 number of bytes emitted by output_constant.
7652 (output_constructor_regular_field): Likewise. Also do not
7653 complain if the total number of bytes emitted is now greater
7654 than the expected fieldpos.
7655 * output.h (output_constant): Update prototype and descriptive comment.
7657 2014-01-16 Marek Polacek <polacek@redhat.com>
7660 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7661 it is error_mark_node.
7663 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7665 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7666 VALID_AVX256_REG_OR_OI_MODE.
7668 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7670 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7671 current procedure should be profiled.
7673 2014-01-15 Andrew Pinski <apinski@cavium.com>
7675 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7676 of moving from/to the STACK_REG register class.
7678 2014-01-15 Richard Henderson <rth@redhat.com>
7681 * reginfo.c (global_regs_decl): Globalize.
7682 * rtl.h (global_regs_decl): Declare.
7683 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7684 reserved via global_regs.
7686 2014-01-15 Teresa Johnson <tejohnson@google.com>
7688 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7690 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7692 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7693 and vmulosh rather than call gen_vec_widen_smult_*.
7694 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7695 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7696 (vec_widen_smult_even_v16qi): Likewise.
7697 (vec_widen_umult_even_v8hi): Likewise.
7698 (vec_widen_smult_even_v8hi): Likewise.
7699 (vec_widen_umult_odd_v16qi): Likewise.
7700 (vec_widen_smult_odd_v16qi): Likewise.
7701 (vec_widen_umult_odd_v8hi): Likewise.
7702 (vec_widen_smult_odd_v8hi): Likewise.
7703 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7704 vmuloub rather than call gen_vec_widen_umult_*.
7705 (vec_widen_umult_lo_v16qi): Likewise.
7706 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7707 vmulosb rather than call gen_vec_widen_smult_*.
7708 (vec_widen_smult_lo_v16qi): Likewise.
7709 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7710 rather than call gen_vec_widen_umult_*.
7711 (vec_widen_umult_lo_v8hi): Likewise.
7712 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7713 rather than call gen_vec_widen_smult_*.
7714 (vec_widen_smult_lo_v8hi): Likewise.
7716 2014-01-15 Jeff Law <law@redhat.com>
7718 PR tree-optimization/59747
7719 * ree.c (find_and_remove_re): Properly handle case where a second
7720 eliminated extension requires widening a copy created for elimination
7721 of a prior extension.
7722 (combine_set_extension): Ensure that the number of hard regs needed
7723 for a destination register does not change when we widen it.
7725 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7727 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7728 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7729 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7730 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7731 (avr-*-rtems*): Likewise.
7732 (bfin*-rtems*): Likewise.
7733 (moxie-*-rtems*): Likewise.
7734 (h8300-*-rtems*): Likewise.
7735 (i[34567]86-*-rtems*): Likewise.
7736 (lm32-*-rtems*): Likewise.
7737 (m32r-*-rtems*): Likewise.
7738 (m68k-*-rtems*): Likewise.
7739 (microblaze*-*-rtems*): Likewise.
7740 (mips*-*-rtems*): Likewise.
7741 (powerpc-*-rtems*): Likewise.
7742 (sh-*-rtems*): Likewise.
7743 (sparc-*-rtems*): Likewise.
7744 (sparc64-*-rtems*): Likewise.
7745 (v850-*-rtems*): Likewise.
7746 (m32c-*-rtems*): Likewise.
7748 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7750 PR rtl-optimization/59511
7751 * ira.c (ira_init_register_move_cost): Use memory costs for some
7752 cases of register move cost calculations.
7753 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7754 instead of BB frequency.
7755 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7756 * lra-assigns.c (find_hard_regno_for): Ditto.
7758 2014-01-15 Richard Biener <rguenther@suse.de>
7760 PR tree-optimization/59822
7761 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
7762 (vectorizable_load): Use it to hoist defs of uses of invariant
7763 loads out of the loop.
7765 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7766 Kugan Vivekanandarajah <kuganv@linaro.org>
7769 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
7772 2014-01-15 Richard Biener <rguenther@suse.de>
7774 PR rtl-optimization/59802
7775 * lcm.c (compute_available): Use inverted postorder to seed
7776 the initial worklist.
7778 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7781 * config/s390/s390.c (s390_preferred_reload_class): Don't return
7782 ADDR_REGS for invalid symrefs in non-PIC code.
7784 2014-01-15 Jakub Jelinek <jakub@redhat.com>
7787 * builtins.c (determine_block_size): Initialize *probable_max_size
7788 even if len_rtx is CONST_INT.
7790 2014-01-14 Andrew Pinski <apinski@cavium.com>
7792 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
7793 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
7794 (cortexa53_tunings): Likewise.
7795 (aarch64_sched_issue_rate): New function.
7796 (TARGET_SCHED_ISSUE_RATE): Define.
7798 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7800 * ira-costs.c (find_costs_and_classes): Add missed
7801 ira_init_register_move_cost_if_necessary.
7803 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7806 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
7808 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
7811 * config/i386/i386.c (type_natural_mode): Add a bool parameter
7812 to indicate if type is used for function return value. Warn ABI
7813 change if the vector mode isn't available for function return value.
7814 (ix86_function_arg_advance): Pass false to type_natural_mode.
7815 (ix86_function_arg): Likewise.
7816 (ix86_gimplify_va_arg): Likewise.
7817 (function_arg_32): Don't warn ABI change.
7818 (ix86_function_value): Pass true to type_natural_mode.
7819 (ix86_return_in_memory): Likewise.
7820 (ix86_struct_value_rtx): Removed.
7821 (TARGET_STRUCT_VALUE_RTX): Likewise.
7823 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7825 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
7826 converting a conditional jump into a conditional return.
7828 2014-01-14 Richard Biener <rguenther@suse.de>
7830 PR tree-optimization/58921
7831 PR tree-optimization/59006
7832 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
7833 hoisting invariant stmts.
7834 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
7835 invariant loads on the preheader edge if possible.
7837 2014-01-14 Joey Ye <joey.ye@arm.com>
7839 * doc/plugin.texi (Building GCC plugins): Update to C++.
7841 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
7843 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
7844 (_mm_rcp28_round_ss): Ditto.
7845 (_mm_rsqrt28_round_sd): Ditto.
7846 (_mm_rsqrt28_round_ss): Ditto.
7847 (_mm_rcp28_sd): Ditto.
7848 (_mm_rcp28_ss): Ditto.
7849 (_mm_rsqrt28_sd): Ditto.
7850 (_mm_rsqrt28_ss): Ditto.
7851 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
7852 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
7853 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
7854 (IX86_BUILTIN_RCP28SD): Ditto.
7855 (IX86_BUILTIN_RCP28SS): Ditto.
7856 (IX86_BUILTIN_RSQRT28SD): Ditto.
7857 (IX86_BUILTIN_RSQRT28SS): Ditto.
7858 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
7859 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
7860 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
7861 (ix86_expand_special_args_builtin): Expand new FTYPE.
7862 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
7863 (srcp14<mode>): Make insn unary.
7864 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
7865 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
7866 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
7867 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
7868 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
7869 Fix rounding: make it SAE only.
7870 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7872 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7874 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
7875 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
7876 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
7877 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
7878 (round_saeonly_mask_scalar_operand4): Ditto.
7879 (round_saeonly_mask_scalar_op3): Ditto.
7880 (round_saeonly_mask_scalar_op4): Ditto.
7882 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7884 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7885 Implement -maltivec=be for vec_insert and vec_extract.
7887 2014-01-10 DJ Delorie <dj@redhat.com>
7889 * config/msp430/msp430.md (call_internal): Don't allow memory
7890 references with SP as the base register.
7891 (call_value_internal): Likewise.
7892 * config/msp430/constraints.md (Yc): New. For memory references
7893 that don't use SP as a base register.
7895 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
7896 "an integer without a # prefix"
7897 * config/msp430/msp430.md (epilogue_helper): Use it.
7899 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7902 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
7903 AVX512F gather builtins.
7904 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
7905 on gather decls with INTEGER_TYPE masktype.
7906 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
7907 directly into the builtin rather than hoisting it before loop.
7909 PR tree-optimization/59387
7910 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
7911 (scev_const_prop): If folded_casts and type has undefined overflow,
7912 use force_gimple_operand instead of force_gimple_operand_gsi and
7913 for each added stmt if it is assign with
7914 arith_code_with_undefined_signed_overflow, call
7915 rewrite_to_defined_overflow.
7916 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
7917 gimple-fold.h instead.
7918 (arith_code_with_undefined_signed_overflow,
7919 rewrite_to_defined_overflow): Moved to ...
7920 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
7921 rewrite_to_defined_overflow): ... here. No longer static.
7922 Include gimplify-me.h.
7923 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
7924 rewrite_to_defined_overflow): New prototypes.
7926 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7928 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
7930 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
7932 * builtins.c (get_object_alignment_2): Minor tweak.
7933 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
7935 2014-01-13 Christian Bruel <christian.bruel@st.com>
7937 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
7938 optimized non constant lengths.
7940 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7943 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
7944 load as __atomic_load_N if possible.
7946 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7948 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
7950 (rs6000_expand_builtin): Adjust call.
7952 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7955 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
7956 * config/rs6000/rs6000.c: Include target-globals.h.
7957 (rs6000_set_current_function): Instead of doing target_reinit
7958 unconditionally, use save_target_globals_default_opts and
7959 restore_target_globals.
7961 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
7963 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
7964 (rs6000_expand_builtin): Handle mffs and mtfsf.
7965 (rs6000_init_builtins): Define mffs and mtfsf.
7966 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
7967 (rs6000_mffs): New pattern.
7968 (rs6000_mtfsf): New pattern.
7970 2014-01-11 Bin Cheng <bin.cheng@arm.com>
7972 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
7973 Start narrowing with START. Apply candidate-use pair
7974 and check overall cost in narrowing.
7975 (iv_ca_prune): Pass new argument.
7977 2014-01-10 Jeff Law <law@redhat.com>
7980 * ree.c (combine_reaching_defs): Ensure the defining statement
7981 occurs before the extension when optimizing extensions with
7982 different source and destination hard registers.
7984 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
7987 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
7988 vtables into the type inheritance graph.
7990 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7992 PR rtl-optimization/59754
7993 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
7994 modes in the REGNO != REGNO case.
7996 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7998 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
8000 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8002 PR tree-optimization/59745
8003 * tree-predcom.c (tree_predictive_commoning_loop): Call
8004 free_affine_expand_cache if giving up because components is NULL.
8006 * target-globals.c (save_target_globals): Allocate < 4KB structs using
8007 GC in payload of target_globals struct instead of allocating them on
8008 the heap and the larger structs separately using GC.
8009 * target-globals.h (struct target_globals): Make regs, hard_regs,
8010 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
8011 of GTY((skip)) and change type to void *.
8012 (reset_target_globals): Cast loads from those fields to corresponding
8015 2014-01-10 Steve Ellcey <sellcey@mips.com>
8018 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
8019 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
8020 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
8022 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
8025 * aarch64-modes.def (CC_Zmode): New flags mode.
8026 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
8027 represents an equality.
8028 (aarch64_get_condition_code): Handle CC_Zmode.
8029 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
8031 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8033 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
8034 extraction in good case.
8036 2014-01-10 Richard Biener <rguenther@suse.de>
8038 PR tree-optimization/59374
8039 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
8040 checking after SLP discovery. Mark stmts not participating
8041 in any SLP instance properly.
8043 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8045 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
8046 when handling a SET rtx.
8048 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8050 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
8051 (cortex-a57): Likewise.
8052 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
8054 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8056 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
8057 non-iwmmxt builtins.
8059 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8063 * ipa-devirt.c record_target_from_binfo): Take as argument
8064 stack of binfos and lookup matching one for virtual inheritance.
8065 (possible_polymorphic_call_targets_1): Update.
8067 2014-01-10 Huacai Chen <chenhc@lemote.com>
8069 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
8070 kernel strings for Loongson-2E/2F/3A.
8072 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8075 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
8076 is_gimple_call before calling gimple_call_internal_p.
8078 2014-01-09 Steve Ellcey <sellcey@mips.com>
8080 * Makefile.in (TREE_FLOW_H): Remove.
8081 (TREE_SSA_H): Add file names from tree-flow.h.
8082 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
8083 * tree.h: Remove tree-flow.h reference.
8084 * hash-table.h: Remove tree-flow.h reference.
8085 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
8086 reference with tree-ssa-loop.h.
8088 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8090 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8091 default element-order behavior for -maltivec.
8092 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8093 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8094 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8095 when targeting big endian, at least for now.
8096 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8098 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8101 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8102 var satisfies use_register_for_decl, just take into account type
8103 alignment, rather than decl alignment.
8105 PR tree-optimization/59622
8106 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8107 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8108 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8109 Don't devirtualize for inplace at all. For targets.length () == 1,
8110 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8112 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8114 * config/i386/i386.md (cpu): Remove the unused btver1.
8116 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8118 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8120 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8123 * tree-core.h (struct target_globals): New forward declaration.
8124 (struct tree_target_option): Add globals field.
8125 * tree.h (TREE_TARGET_GLOBALS): Define.
8126 (prepare_target_option_nodes_for_pch): New prototype.
8127 * target-globals.h (struct target_globals): Define even if
8129 * tree.c (prepare_target_option_node_for_pch,
8130 prepare_target_option_nodes_for_pch): New functions.
8131 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8132 * config/i386/i386.c: Include target-globals.h.
8133 (ix86_set_current_function): Instead of doing target_reinit
8134 unconditionally, use save_target_globals_default_opts and
8135 restore_target_globals.
8137 2014-01-09 Richard Biener <rguenther@suse.de>
8139 PR tree-optimization/59715
8140 * tree-cfg.h (split_critical_edges): Declare.
8141 * tree-cfg.c (split_critical_edges): Export.
8142 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8144 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8146 * cfgexpand.c (expand_stack_vars): Optionally disable
8147 asan stack protection.
8148 (expand_used_vars): Likewise.
8149 (partition_stack_vars): Likewise.
8150 * asan.c (asan_emit_stack_protection): Optionally disable
8151 after return stack usage.
8152 (instrument_derefs): Optionally disable memory access instrumentation.
8153 (instrument_builtin_call): Likewise.
8154 (instrument_strlen_call): Likewise.
8155 (asan_protect_global): Optionally disable global variables protection.
8156 * doc/invoke.texi: Added doc for new options.
8157 * params.def: Added new options.
8158 * params.h: Likewise.
8160 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8162 PR rtl-optimization/59724
8163 * ifcvt.c (cond_exec_process_if_block): Don't call
8164 flow_find_head_matching_sequence with 0 longest_match.
8165 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8166 non-active insns if !stop_after.
8167 (try_head_merge_bb): Revert 2014-01-07 changes.
8169 2014-01-08 Jeff Law <law@redhat.com>
8171 * ree.c (get_sub_rtx): New function, extracted from...
8172 (merge_def_and_ext): Here.
8173 (combine_reaching_defs): Use get_sub_rtx.
8175 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8177 * cgraph.h (varpool_variable_node): Do not choke on null node.
8179 2014-01-08 Catherine Moore <clm@codesourcery.com>
8181 * config/mips/mips.md (simple_return): Attempt to use JRC
8183 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8185 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8187 PR rtl-optimization/59137
8188 * reorg.c (steal_delay_list_from_target): Call update_block for
8190 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8192 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8194 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8195 two duplicate entries.
8197 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8200 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8202 * config/mips/mips.c (mips_truncated_op_cost): New function.
8203 (mips_rtx_costs): Adjust test for BADDU.
8204 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8206 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8208 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8209 (*baddu_si): ...this new pattern.
8211 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8214 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8216 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8219 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8221 (expand_expr, expand_normal): Adjust.
8222 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8223 inner_reference_p. Use inner_reference_p to expand inner references.
8224 (store_expr): Adjust.
8225 * cfgexpand.c (expand_call_stmt): Adjust.
8227 2014-01-08 Rong Xu <xur@google.com>
8229 * gcov-io.c (gcov_var): Move from gcov-io.h.
8230 (gcov_position): Ditto.
8231 (gcov_is_error): Ditto.
8232 (gcov_rewrite): Ditto.
8233 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8234 only part to libgcc/libgcov.h.
8236 2014-01-08 Marek Polacek <polacek@redhat.com>
8239 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8241 2014-01-08 Marek Polacek <polacek@redhat.com>
8244 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8246 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8248 PR rtl-optimization/59649
8249 * stor-layout.c (get_mode_bounds): For BImode return
8250 0 and STORE_FLAG_VALUE.
8252 2014-01-08 Richard Biener <rguenther@suse.de>
8255 * gimple.h (is_gimple_builtin_call): Remove.
8256 (gimple_builtin_call_types_compatible_p): New.
8257 (gimple_call_builtin_p): New overload.
8258 * gimple.c (is_gimple_builtin_call): Remove.
8259 (validate_call): Rename to ...
8260 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8262 (validate_type): New static function.
8263 (gimple_call_builtin_p): New overload and adjust.
8264 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8265 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8266 (gimple_fold_stmt_to_constant_1): Likewise.
8267 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8269 2014-01-08 Richard Biener <rguenther@suse.de>
8272 * gimplify.c (gimplify_expr): Gimplify register-register type
8273 VIEW_CONVERT_EXPRs to separate stmts.
8275 2014-01-07 Jeff Law <law@redhat.com>
8278 * ree.c (combine_set_extension): Handle case where source
8279 and destination registers in an extension insn are different.
8280 (combine_reaching_defs): Allow source and destination registers
8281 in extension to be different under limited circumstances.
8282 (add_removable_extension): Remove restriction that the
8283 source and destination registers in the extension are the same.
8284 (find_and_remove_re): Emit a copy from the extension's
8285 destination to its source after the defining insn if
8286 the source and destination registers are different.
8289 * ifcvt.c (merge_if_block): If we are merging a block with more than
8290 one successor with a block with no successors, remove any BARRIER
8291 after the second block.
8293 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8295 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8297 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8300 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8301 for 14-bit register offsets when INT14_OK_STRICT is false.
8303 2014-01-07 Roland Stigge <stigge@antcom.de>
8304 Michael Meissner <meissner@linux.vnet.ibm.com>
8307 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8308 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8310 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8312 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8315 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8317 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8318 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8319 rtx is const0_rtx or not.
8321 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8324 * target-globals.c (save_target_globals): Remove this_fn_optab
8326 * toplev.c: Include optabs.h.
8327 (target_reinit): Temporarily restore the global options if another
8328 set of options are in force.
8330 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8332 PR rtl-optimization/58668
8333 * cfgcleanup.c (flow_find_cross_jump): Don't count
8334 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8335 to determine what is counted.
8336 (flow_find_head_matching_sequence): Use active_insn_p to determine
8338 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8340 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8341 determine what is counted.
8343 PR tree-optimization/59643
8344 * tree-predcom.c (split_data_refs_to_components): If one dr is
8345 read and one write, determine_offset fails and the write isn't
8346 in the bad component, just put the read into the bad component.
8348 2014-01-07 Mike Stump <mikestump@comcast.net>
8349 Jakub Jelinek <jakub@redhat.com>
8352 * tree-core.h (struct tree_optimization_option): Change optabs
8353 type from unsigned char * to void *.
8354 * optabs.c (init_tree_optimization_optabs): Adjust
8355 TREE_OPTIMIZATION_OPTABS initialization.
8357 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8360 * config/i386/i386.h (struct machine_function): Add
8361 no_drap_save_restore field.
8362 * config/i386/i386.c (ix86_save_reg): Use
8363 !cfun->machine->no_drap_save_restore instead of
8364 crtl->stack_realign_needed.
8365 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8366 this function clears frame_pointer_needed. Set
8367 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8368 and DRAP reg is needed.
8370 2014-01-06 Marek Polacek <polacek@redhat.com>
8373 * doc/implement-c.texi: Mention that other integer types are
8374 permitted as bit-field types in strictly conforming mode.
8376 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8378 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8381 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8383 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8385 2014-01-06 Martin Jambor <mjambor@suse.cz>
8388 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8390 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8392 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8396 * var-tracking.c (add_stores): Preserve the value of the source even if
8397 we don't record the store.
8399 2014-01-06 Terry Guo <terry.guo@arm.com>
8401 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8403 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8406 * config/darwin.c (darwin_function_section): Adjust return values to
8407 correspond to optimisation changes made in r206070.
8409 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8411 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8412 from prefetch_block tune setting.
8413 (nocona_cost): Correct size of prefetch block to 64.
8415 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8417 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8418 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8419 used to save the static chain register in the computation of the offset
8420 from which the FP registers need to be restored.
8422 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8424 PR tree-optimization/59519
8425 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8426 ICE if get_current_def (current_new_name) is already non-NULL, as long
8427 as it is a phi result of some other phi in *new_exit_bb that has
8430 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8431 or vmovdqu* for misaligned_operand.
8432 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8433 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8434 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8435 aligned_mem for AVX512F masked aligned load and store builtins and for
8438 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8440 PR tree-optimization/59651
8441 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8442 Address range for negative step should be added by TYPE_SIZE_UNIT.
8444 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8446 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8447 overlapping registers also for registers other than the stack pointer.
8449 2014-01-03 Marek Polacek <polacek@redhat.com>
8452 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8455 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8458 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8461 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8462 (push splitter): Use <P:MODE_SIZE> instead of
8463 GET_MODE_SIZE (<P:MODE>mode).
8464 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8465 (mov -1, reg peephole2): Likewise.
8466 * config/i386/sse.md (*mov<mode>_internal,
8467 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8468 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8469 *<code><mode>3, *andnot<mode>3<mask_name>,
8470 <mask_codefor><code><mode>3<mask_name>): Likewise.
8471 * config/i386/subst.md (mask_mode512bit_condition,
8472 sd_mask_mode512bit_condition): Likewise.
8474 2014-01-02 Xinliang David Li <davidxl@google.com>
8476 PR tree-optimization/59303
8477 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8478 (dump_predicates): Better output format.
8479 (pred_equal_p): New function.
8480 (is_neq_relop_p): Ditto.
8481 (is_neq_zero_form_p): Ditto.
8482 (pred_expr_equal_p): Ditto.
8483 (pred_neg_p): Ditto.
8484 (simplify_pred): Ditto.
8485 (simplify_preds_2): Ditto.
8486 (simplify_preds_3): Ditto.
8487 (simplify_preds_4): Ditto.
8488 (simplify_preds): Ditto.
8490 (push_to_worklist): Ditto.
8491 (get_pred_info_from_cmp): Ditto.
8492 (is_degenerated_phi): Ditto.
8493 (normalize_one_pred_1): Ditto.
8494 (normalize_one_pred): Ditto.
8495 (normalize_one_pred_chain): Ditto.
8496 (normalize_preds): Ditto.
8497 (normalize_cond_1): Remove function.
8498 (normalize_cond): Ditto.
8499 (is_gcond_subset_of): Ditto.
8500 (is_subset_of_any): Ditto.
8501 (is_or_set_subset_of): Ditto.
8502 (is_and_set_subset_of): Ditto.
8503 (is_norm_cond_subset_of): Ditto.
8504 (pred_chain_length_cmp): Ditto.
8505 (convert_control_dep_chain_into_preds): Type change.
8506 (find_predicates): Ditto.
8507 (find_def_preds): Ditto.
8508 (destroy_predicates_vecs): Ditto.
8509 (find_matching_predicates_in_rest_chains): Ditto.
8510 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8511 (is_pred_expr_subset): Ditto.
8512 (is_pred_chain_subset_of): Ditto.
8513 (is_included_in): Ditto.
8514 (is_superset_of): Ditto.
8516 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8518 Update copyright years.
8520 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8522 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8523 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8524 config/arc/arc.md, config/arc/arc.opt,
8525 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8526 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8527 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8528 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8529 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8530 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8531 vtable-verify.c, vtable-verify.h: Use the standard form for the
8534 2014-01-02 Tobias Burnus <burnus@net-b.de>
8536 * gcc.c (process_command): Update copyright notice dates.
8537 * gcov-dump.c: Ditto.
8539 * doc/cpp.texi: Bump @copying's copyright year.
8540 * doc/cppinternals.texi: Ditto.
8541 * doc/gcc.texi: Ditto.
8542 * doc/gccint.texi: Ditto.
8543 * doc/gcov.texi: Ditto.
8544 * doc/install.texi: Ditto.
8545 * doc/invoke.texi: Ditto.
8547 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8549 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8551 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8553 * config/i386/sse.md (*mov<mode>_internal): Guard
8554 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8556 PR rtl-optimization/59647
8557 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8558 new_rtx into UNSIGNED_FLOAT rtxes.
8560 Copyright (C) 2014 Free Software Foundation, Inc.
8562 Copying and distribution of this file, with or without modification,
8563 are permitted in any medium without royalty provided the copyright
8564 notice and this notice are preserved.