1 2014-04-28 David Malcolm <dmalcolm@redhat.com>
3 * doc/gimple.texi: Replace the description of the now-defunct
4 union gimple_statement_d with a diagram showing the
5 gimple_statement_base class hierarchy and its relationships to
6 the GSS_ and GIMPLE_ enums.
8 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
10 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
11 * config/aarch64/aarch64.c
12 (aarch64_cannot_change_mode_class): Weaken conditions.
13 (aarch64_modes_tieable_p): New.
14 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
16 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
18 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
19 (loadsync_<mode>): Change mode.
20 (load_quadpti, store_quadpti): New.
21 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
22 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
24 2014-04-28 Martin Jambor <mjambor@suse.cz>
26 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
27 same alias type as the original statement.
28 (subreplacement_assignment_data): New type.
29 (handle_unscalarized_data_in_subtree): New type of parameter,
30 generate new memory accesses with same alias type as the original
32 (load_assign_lhs_subreplacements): Likewise.
33 (sra_modify_constructor_assign): Generate new memory accesses with
34 same alias type as the original statement.
36 2014-04-28 Richard Biener <rguenther@suse.de>
38 * tree-pass.h (TODO_verify_il): Define.
39 (TODO_verify_all): Complete properly.
40 * passes.c (execute_function_todo): Move existing loop-closed
41 SSA verification under TODO_verify_il.
42 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
43 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
44 Fix tree sharing issue.
46 2014-04-28 Richard Biener <rguenther@suse.de>
49 * builtins.def (DEF_C11_BUILTIN): Add.
50 (BUILT_IN_ALIGNED_ALLOC): Likewise.
51 * coretypes.h (enum function_class): Add function_c11_misc.
52 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
53 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
54 (call_may_clobber_ref_p_1): Likewise.
55 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
56 (mark_all_reaching_defs_necessary_1): Likewise.
57 (propagate_necessity): Likewise.
58 (eliminate_unnecessary_stmts): Likewise.
59 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
61 2014-04-28 Richard Biener <rguenther@suse.de>
63 * tree-vrp.c (vrp_var_may_overflow): Remove.
64 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
65 with overflow immediately bump to one before that value and
66 let iteration figure out overflow status.
68 2014-04-28 Richard Biener <rguenther@suse.de>
70 * configure.ac: Do valgrind header checks unconditionally.
71 Add --enable-valgrind-annotations.
72 * system.h: Guard valgrind header inclusion with
73 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
74 * alloc-pool.c (pool_alloc, pool_free): Use
75 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
76 to guard possibly dead code.
77 * config.in: Regenerated.
78 * configure: Likewise.
80 2014-04-28 Jeff Law <law@redhat.com>
82 PR tree-optimization/60902
83 * tree-ssa-threadedge.c
84 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
85 over real defs when invalidating outputs from statements that do not
86 produce useful outputs for threading.
88 2014-04-28 Richard Biener <rguenther@suse.de>
90 PR tree-optimization/60979
91 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
92 SCOPs that end in a block with a successor with abnormal
95 2014-04-28 Richard Biener <rguenther@suse.de>
97 * tree-pass.h (execute_pass_list): Adjust prototype.
98 * passes.c (pass_manager::execute_early_local_passes):
100 (do_per_function): Change callback signature, push all actual
101 work to the callbals.
102 (do_per_function_toporder): Likewise.
103 (execute_function_dump): Adjust.
104 (execute_function_todo): Likewise.
105 (clear_last_verified): Likewise.
106 (verify_curr_properties): Likewise.
107 (update_properties_after_pass): Likewise.
108 (execute_pass_list_1): Split out from ...
109 (execute_pass_list): ... here. Adjust.
110 (execute_ipa_pass_list): Likewise.
111 * cgraphunit.c (cgraph_add_new_function): Adjust.
112 (analyze_function): Likewise.
113 (expand_function): Likewise.
114 * cgraph.c (release_function_body): Free dominance info
115 here instead of asserting it was magically freed elsewhere.
117 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
119 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
120 * configure: Regenerate.
121 * config/sparc/sparc.opt (muser-mode): New option.
122 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
124 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
125 * doc/invoke.texi (SPARC options): Document -muser-mode.
127 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
129 * cselib.c (find_slot_memmode): Delete.
130 (cselib_hasher): Change compare_type to a struct.
131 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
133 (preserve_constants_and_equivs): Adjust for new compare_type.
134 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
135 (wrap_constant): Delete.
136 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
138 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
140 * doc/install.texi (Building with profile feedback): Remove
143 2014-04-26 Tom de Vries <tom@codesourcery.com>
145 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
148 2014-04-25 Cary Coutant <ccoutant@google.com>
151 * dwarf2out.c (should_move_die_to_comdat): A type definition
152 can contain a subprogram definition, but don't move it to a
154 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
155 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
157 (clone_tree_hash): Rename to...
158 (clone_tree_partial): ...this; change callers. Copy
159 DW_TAG_subprogram DIEs as declarations.
160 (copy_decls_walk): Don't copy children of a declaration into a
163 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
166 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
169 2014-04-25 Jiong Wang <jiong.wang@arm.com>
171 * config/arm/predicates.md (call_insn_operand): Add long_call check.
172 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
174 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
177 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
179 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
182 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
184 PR tree-optimization/60930
185 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
186 creating a multiply candidate by folding two constant
187 multiplicands when the result overflows.
189 2014-04-25 Jakub Jelinek <jakub@redhat.com>
191 PR tree-optimization/60960
192 * tree-vect-generic.c (expand_vector_operation): Only call
193 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
195 2014-04-25 Tom de Vries <tom@codesourcery.com>
197 * expr.c (clobber_reg_mode): New function.
198 * expr.h (clobber_reg): New function.
200 2014-04-25 Tom de Vries <tom@codesourcery.com>
202 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
205 2014-04-25 Radovan Obradovic <robradovic@mips.com>
206 Tom de Vries <tom@codesourcery.com>
208 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
210 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
211 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
212 new argument to find_all_hard_reg_sets call.
214 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
216 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
217 Use HOST_WIDE_INT_C for mask literal.
218 (aarch_rev16_shleft_mask_imm_p): Likewise.
220 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
223 * config/sparc/sparc.md (ashlsi3_extend): Delete.
225 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
227 PR preprocessor/56540
228 * config/i386/i386-c.c (ix86_target_macros): Define
229 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
231 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
233 * configure.ac (tga_func): Remove.
234 (LIB_TLS_SPEC): Remove.
235 * configure: Regenerate.
236 * config.in: Regenerate.
237 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
239 2014-04-25 Richard Biener <rguenther@suse.de>
242 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
243 call stmt use/clobber sets during stmt walk instead of
244 walking the possibly incomplete set of caller edges.
246 2014-04-25 Richard Biener <rguenther@suse.de>
249 * passes.c (apply_ipa_transforms): Inline into only caller ...
250 (execute_one_pass): ... here. Properly bring in function
251 bodies for nodes we want to apply IPA transforms to.
253 2014-04-24 Cong Hou <congh@google.com>
255 PR tree-optimization/60896
256 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
257 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
258 (vect_mark_pattern_stmts): Set the def type of all statements in
259 PATTERN_DEF_SEQ as vect_internal_def.
261 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
263 * doc/extend.texi (PowerPC Built-in Functions): Document new
264 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
266 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
268 * config/rs6000/predicates.md (const_0_to_3_operand): New
269 predicate to match 0..3 integer constants.
271 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
272 to support adding miscellaneous builtin functions.
273 (BU_DFP_MISC_2): Likewise.
274 (BU_P7_MISC_1): Likewise.
275 (BU_P7_MISC_2): Likewise.
276 (BU_P8V_MISC_3): Likewise.
277 (BU_MISC_1): Likewise.
278 (BU_MISC_2): Likewise.
279 (DIVWE): Add extended divide builtin functions.
287 (DXEX): Add decimal floating-point builtin functions.
299 (CDTBCD): Add new BCD builtin functions.
303 (BCDADD_LT): Likewise.
304 (BCDADD_EQ): Likewise.
305 (BCDADD_GT): Likewise.
306 (BCDADD_OV): Likewise.
308 (BCDSUB_LT): Likewise.
309 (BCDSUB_EQ): Likewise.
310 (BCDSUB_GT): Likewise.
311 (BCDSUB_OV): Likewise.
312 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
313 (UNPACK_TD): Likewise.
315 (UNPACK_TF): Likewise.
316 (UNPACK_TF_0): Likewise.
317 (UNPACK_TF_1): Likewise.
318 (PACK_V1TI): Likewise.
319 (UNPACK_V1TI): Likewise.
321 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
322 support for decimal floating point builtin functions.
323 (rs6000_expand_ternop_builtin): Add checks for the new builtin
324 functions that take constant arguments.
325 (rs6000_invalid_builtin): Add decimal floating point builtin
327 (rs6000_init_builtins): Setup long double, _Decimal64, and
328 _Decimal128 types for new builtin functions.
329 (builtin_function_type): Set the unsigned flags appropriately for
330 the new builtin functions.
331 (rs6000_opt_masks): Add support for decimal floating point builtin
334 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
335 floating point builtin functions.
336 (RS6000_BTM_COMMON): Likewise.
337 (RS6000_BTI_long_double): Likewise.
338 (RS6000_BTI_dfloat64): Likewise.
339 (RS6000_BTI_dfloat128): Likewise.
340 (long_double_type_internal_node): Likewise.
341 (dfloat64_type_internal_node): Likewise.
342 (dfloat128_type_internal_node): Likewise.
344 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
345 2.07 bcd arithmetic instructions.
346 (UNSPEC_BCDSUB): Likewise.
347 (UNSPEC_BCD_OVERFLOW): Likewise.
348 (UNSPEC_BCD_ADD_SUB): Likewise.
349 (bcd_add_sub): Likewise.
350 (BCD_TEST): Likewise.
351 (bcd<bcd_add_sub>): Likewise.
352 (bcd<bcd_add_sub>_test): Likewise.
353 (bcd<bcd_add_sub>_test2): Likewise.
354 (bcd<bcd_add_sub>_<code>): Likewise.
355 (peephole2 for combined bcd ops): Likewise.
357 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
358 decimal floating point builtin functions.
359 (UNSPEC_DENBCD): Likewise.
360 (UNSPEC_DXEX): Likewise.
361 (UNSPEC_DIEX): Likewise.
362 (UNSPEC_DSCLI): Likewise.
363 (UNSPEC_DSCRI): Likewise.
364 (D64_D128): Likewise.
365 (dfp_suffix): Likewise.
366 (dfp_ddedpd_<mode>): Likewise.
367 (dfp_denbcd_<mode>): Likewise.
368 (dfp_dxex_<mode>): Likewise.
369 (dfp_diex_<mode>): Likewise.
370 (dfp_dscli_<mode>): Likewise.
371 (dfp_dscri_<mode>): Likewise.
373 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
375 (UNSPEC_CDTBCD): Likewise.
376 (UNSPEC_CBCDTD): Likewise.
377 (UNSPEC_DIVE): Add support for new extended divide builtin
379 (UNSPEC_DIVEO): Likewise.
380 (UNSPEC_DIVEU): Likewise.
381 (UNSPEC_DIVEUO): Likewise.
382 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
383 pack/unpack 128-bit types.
384 (UNSPEC_PACK_128BIT): Likewise.
385 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
386 (udiv<mode>3): Use idiv_ldiv mode attribute.
387 (div<mode>3): Likewise.
388 (addg6s): Add new BCD builtin functions.
391 (UNSPEC_DIV_EXTEND): Add support for new extended divide
393 (div_extend): Likewise.
394 (div<div_extend>_<mode>"): Likewise.
395 (FP128_64): Add support for new builtin functions to pack/unpack
397 (unpack<mode>): Likewise.
398 (unpacktf_0): Likewise.
399 (unpacktf_1): Likewise.
400 (unpack<mode>_dm): Likewise.
401 (unpack<mode>_nodm): Likewise.
402 (pack<mode>): Likewise.
403 (unpackv1ti): Likewise.
404 (packv1ti): Likewise.
406 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
408 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
411 2014-04-24 Jakub Jelinek <jakub@redhat.com>
413 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
414 * gimplify.c (omp_is_private): Change last argument's type to int.
415 Only diagnose lastprivate if the simd argument is 1, only diagnose
416 linear if the simd argument is 2.
417 (gimplify_omp_for): Adjust omp_is_private callers. When adding
418 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
419 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
420 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
421 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
422 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
423 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
424 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
425 * tree-nested.c (convert_nonlocal_omp_clauses,
426 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
428 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
431 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
434 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
436 * flag-types.h (enum ivar_visibility): Add.
438 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
440 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
443 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
445 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
447 2014-04-24 Radovan Obradovic <robradovic@mips.com>
448 Tom de Vries <tom@codesourcery.com>
450 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
451 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
453 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
454 * emit-rtl.c (try_split): Same.
456 2014-04-24 Radovan Obradovic <robradovic@mips.com>
457 Tom de Vries <tom@codesourcery.com>
459 * common.opt (fuse-caller-save): New option.
461 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
463 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
466 2014-04-24 Richard Biener <rguenther@suse.de>
468 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
469 during TER and instead use the sepops interface for expanding
470 non-GIMPLE_SINGLE_RHS.
472 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
474 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
475 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
477 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
479 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
480 assembler 64-bit option.
481 * configure: Regenerate.
483 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
485 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
486 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
487 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
488 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
489 (TARGET_CRYPTO): Take TARGET_SIMD into account.
491 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
493 * config/aarch64/aarch64-builtins.c
494 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
495 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
496 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
497 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
499 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
500 (Vrevsuff): New mode attribute.
502 2014-04-24 Terry Guo <terry.guo@arm.com>
504 * config/arm/arm.h (machine_function): Define variable
505 after_arm_reorg here.
506 * config/arm/arm.c (after_arm_reorg): Remove the definition.
507 (arm_split_constant): Update the way to access variable
510 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
512 2014-04-23 Tom de Vries <tom@codesourcery.com>
514 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
516 2014-04-23 David Malcolm <dmalcolm@redhat.com>
518 * is-a.h: Update comments to reflect the following changes to the
519 "pointerness" of the API, making the template parameter match the
520 return type, allowing use of is-a.h with typedefs of pointers.
521 (is_a_helper::cast): Return a T rather then a pointer to a T, so
522 that the return type matches the parameter to the is_a_helper.
524 (dyn_cast): Likewise.
526 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
527 pointer from the is-a.h API.
529 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
530 (is_a_helper <cgraph_node *>::test): ...this, matching change to
532 (is_a_helper <varpool_node>::test): Likewise, convert to...
533 (is_a_helper <varpool_node *>::test): ...this.
535 (varpool_first_variable): Update for removal of implicit pointer
537 (varpool_next_variable): Likewise.
538 (varpool_first_static_initializer): Likewise.
539 (varpool_next_static_initializer): Likewise.
540 (varpool_first_defined_variable): Likewise.
541 (varpool_next_defined_variable): Likewise.
542 (cgraph_first_defined_function): Likewise.
543 (cgraph_next_defined_function): Likewise.
544 (cgraph_first_function): Likewise.
545 (cgraph_next_function): Likewise.
546 (cgraph_first_function_with_gimple_body): Likewise.
547 (cgraph_next_function_with_gimple_body): Likewise.
548 (cgraph_alias_target): Likewise.
549 (varpool_alias_target): Likewise.
550 (cgraph_function_or_thunk_node): Likewise.
551 (varpool_variable_node): Likewise.
552 (symtab_real_symbol_p): Likewise.
553 * cgraphunit.c (referred_to_p): Likewise.
554 (analyze_functions): Likewise.
555 (handle_alias_pairs): Likewise.
556 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
557 * gimple-ssa.h (gimple_vuse_op): Likewise.
558 (gimple_vdef_op): Likewise.
559 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
560 * gimple.c (gimple_build_asm_1): Likewise.
561 (gimple_build_try): Likewise.
562 (gimple_build_resx): Likewise.
563 (gimple_build_eh_dispatch): Likewise.
564 (gimple_build_omp_for): Likewise.
565 (gimple_omp_for_set_clauses): Likewise.
567 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
568 (is_a_helper <gimple_statement_asm *>::test): ...this.
569 (is_a_helper <gimple_statement_bind>::test): Convert to...
570 (is_a_helper <gimple_statement_bind *>::test): ...this.
571 (is_a_helper <gimple_statement_call>::test): Convert to...
572 (is_a_helper <gimple_statement_call *>::test): ...this.
573 (is_a_helper <gimple_statement_catch>::test): Convert to...
574 (is_a_helper <gimple_statement_catch *>::test): ...this.
575 (is_a_helper <gimple_statement_resx>::test): Convert to...
576 (is_a_helper <gimple_statement_resx *>::test): ...this.
577 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
578 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
579 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
580 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
581 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
582 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
583 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
584 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
585 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
586 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
587 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
588 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
589 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
590 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
591 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
592 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
593 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
594 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
595 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
596 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
597 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
598 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
599 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
600 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
601 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
602 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
603 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
604 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
605 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
606 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
607 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
608 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
609 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
610 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
611 (is_a_helper <gimple_statement_phi>::test): Convert to...
612 (is_a_helper <gimple_statement_phi *>::test): ...this.
613 (is_a_helper <gimple_statement_transaction>::test): Convert to...
614 (is_a_helper <gimple_statement_transaction *>::test): ...this.
615 (is_a_helper <gimple_statement_try>::test): Convert to...
616 (is_a_helper <gimple_statement_try *>::test): ...this.
617 (is_a_helper <gimple_statement_wce>::test): Convert to...
618 (is_a_helper <gimple_statement_wce *>::test): ...this.
619 (is_a_helper <const gimple_statement_asm>::test): Convert to...
620 (is_a_helper <const gimple_statement_asm *>::test): ...this.
621 (is_a_helper <const gimple_statement_bind>::test): Convert to...
622 (is_a_helper <const gimple_statement_bind *>::test): ...this.
623 (is_a_helper <const gimple_statement_call>::test): Convert to...
624 (is_a_helper <const gimple_statement_call *>::test): ...this.
625 (is_a_helper <const gimple_statement_catch>::test): Convert to...
626 (is_a_helper <const gimple_statement_catch *>::test): ...this.
627 (is_a_helper <const gimple_statement_resx>::test): Convert to...
628 (is_a_helper <const gimple_statement_resx *>::test): ...this.
629 (is_a_helper <const gimple_statement_eh_dispatch>::test):
631 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
633 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
635 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
636 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
638 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
640 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
642 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
644 (is_a_helper <const gimple_statement_omp_return>::test): Convert
646 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
647 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
649 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
650 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
652 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
653 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
654 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
655 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
657 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
658 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
660 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
661 (is_a_helper <const gimple_statement_omp_target>::test): Convert
663 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
664 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
666 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
667 (is_a_helper <const gimple_statement_omp_single>::test): Convert
669 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
670 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
672 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
673 (is_a_helper <const gimple_statement_omp_task>::test): Convert
675 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
676 (is_a_helper <const gimple_statement_phi>::test): Convert to...
677 (is_a_helper <const gimple_statement_phi *>::test): ...this.
678 (is_a_helper <const gimple_statement_transaction>::test): Convert
680 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
681 (is_a_helper <const gimple_statement_with_ops>::test): Convert
683 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
684 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
685 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
686 (is_a_helper <const gimple_statement_with_memory_ops>::test):
688 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
690 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
692 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
694 (gimple_use_ops): Update for removal of implicit pointer from the
696 (gimple_set_use_ops): Likewise.
697 (gimple_vuse): Likewise.
698 (gimple_vdef): Likewise.
699 (gimple_vuse_ptr): Likewise.
700 (gimple_vdef_ptr): Likewise.
701 (gimple_set_vuse): Likewise.
702 (gimple_set_vdef): Likewise.
703 (gimple_omp_return_set_lhs): Likewise.
704 (gimple_omp_return_lhs): Likewise.
705 (gimple_omp_return_lhs_ptr): Likewise.
706 (gimple_call_fntype): Likewise.
707 (gimple_call_set_fntype): Likewise.
708 (gimple_call_set_internal_fn): Likewise.
709 (gimple_call_use_set): Likewise.
710 (gimple_call_clobber_set): Likewise.
711 (gimple_bind_vars): Likewise.
712 (gimple_bind_set_vars): Likewise.
713 (gimple_bind_body_ptr): Likewise.
714 (gimple_bind_set_body): Likewise.
715 (gimple_bind_add_stmt): Likewise.
716 (gimple_bind_block): Likewise.
717 (gimple_bind_set_block): Likewise.
718 (gimple_asm_ninputs): Likewise.
719 (gimple_asm_noutputs): Likewise.
720 (gimple_asm_nclobbers): Likewise.
721 (gimple_asm_nlabels): Likewise.
722 (gimple_asm_input_op): Likewise.
723 (gimple_asm_input_op_ptr): Likewise.
724 (gimple_asm_output_op): Likewise.
725 (gimple_asm_output_op_ptr): Likewise.
726 (gimple_asm_set_output_op): Likewise.
727 (gimple_asm_clobber_op): Likewise.
728 (gimple_asm_set_clobber_op): Likewise.
729 (gimple_asm_label_op): Likewise.
730 (gimple_asm_set_label_op): Likewise.
731 (gimple_asm_string): Likewise.
732 (gimple_catch_types): Likewise.
733 (gimple_catch_types_ptr): Likewise.
734 (gimple_catch_handler_ptr): Likewise.
735 (gimple_catch_set_types): Likewise.
736 (gimple_catch_set_handler): Likewise.
737 (gimple_eh_filter_types): Likewise.
738 (gimple_eh_filter_types_ptr): Likewise.
739 (gimple_eh_filter_failure_ptr): Likewise.
740 (gimple_eh_filter_set_types): Likewise.
741 (gimple_eh_filter_set_failure): Likewise.
742 (gimple_eh_must_not_throw_fndecl): Likewise.
743 (gimple_eh_must_not_throw_set_fndecl): Likewise.
744 (gimple_eh_else_n_body_ptr): Likewise.
745 (gimple_eh_else_e_body_ptr): Likewise.
746 (gimple_eh_else_set_n_body): Likewise.
747 (gimple_eh_else_set_e_body): Likewise.
748 (gimple_try_eval_ptr): Likewise.
749 (gimple_try_cleanup_ptr): Likewise.
750 (gimple_try_set_eval): Likewise.
751 (gimple_try_set_cleanup): Likewise.
752 (gimple_wce_cleanup_ptr): Likewise.
753 (gimple_wce_set_cleanup): Likewise.
754 (gimple_phi_capacity): Likewise.
755 (gimple_phi_num_args): Likewise.
756 (gimple_phi_result): Likewise.
757 (gimple_phi_result_ptr): Likewise.
758 (gimple_phi_set_result): Likewise.
759 (gimple_phi_arg): Likewise.
760 (gimple_phi_set_arg): Likewise.
761 (gimple_resx_region): Likewise.
762 (gimple_resx_set_region): Likewise.
763 (gimple_eh_dispatch_region): Likewise.
764 (gimple_eh_dispatch_set_region): Likewise.
765 (gimple_omp_critical_name): Likewise.
766 (gimple_omp_critical_name_ptr): Likewise.
767 (gimple_omp_critical_set_name): Likewise.
768 (gimple_omp_for_clauses): Likewise.
769 (gimple_omp_for_clauses_ptr): Likewise.
770 (gimple_omp_for_set_clauses): Likewise.
771 (gimple_omp_for_collapse): Likewise.
772 (gimple_omp_for_index): Likewise.
773 (gimple_omp_for_index_ptr): Likewise.
774 (gimple_omp_for_set_index): Likewise.
775 (gimple_omp_for_initial): Likewise.
776 (gimple_omp_for_initial_ptr): Likewise.
777 (gimple_omp_for_set_initial): Likewise.
778 (gimple_omp_for_final): Likewise.
779 (gimple_omp_for_final_ptr): Likewise.
780 (gimple_omp_for_set_final): Likewise.
781 (gimple_omp_for_incr): Likewise.
782 (gimple_omp_for_incr_ptr): Likewise.
783 (gimple_omp_for_set_incr): Likewise.
784 (gimple_omp_for_pre_body_ptr): Likewise.
785 (gimple_omp_for_set_pre_body): Likewise.
786 (gimple_omp_parallel_clauses): Likewise.
787 (gimple_omp_parallel_clauses_ptr): Likewise.
788 (gimple_omp_parallel_set_clauses): Likewise.
789 (gimple_omp_parallel_child_fn): Likewise.
790 (gimple_omp_parallel_child_fn_ptr): Likewise.
791 (gimple_omp_parallel_set_child_fn): Likewise.
792 (gimple_omp_parallel_data_arg): Likewise.
793 (gimple_omp_parallel_data_arg_ptr): Likewise.
794 (gimple_omp_parallel_set_data_arg): Likewise.
795 (gimple_omp_task_clauses): Likewise.
796 (gimple_omp_task_clauses_ptr): Likewise.
797 (gimple_omp_task_set_clauses): Likewise.
798 (gimple_omp_task_child_fn): Likewise.
799 (gimple_omp_task_child_fn_ptr): Likewise.
800 (gimple_omp_task_set_child_fn): Likewise.
801 (gimple_omp_task_data_arg): Likewise.
802 (gimple_omp_task_data_arg_ptr): Likewise.
803 (gimple_omp_task_set_data_arg): Likewise.
804 (gimple_omp_taskreg_clauses): Likewise.
805 (gimple_omp_taskreg_clauses_ptr): Likewise.
806 (gimple_omp_taskreg_set_clauses): Likewise.
807 (gimple_omp_taskreg_child_fn): Likewise.
808 (gimple_omp_taskreg_child_fn_ptr): Likewise.
809 (gimple_omp_taskreg_set_child_fn): Likewise.
810 (gimple_omp_taskreg_data_arg): Likewise.
811 (gimple_omp_taskreg_data_arg_ptr): Likewise.
812 (gimple_omp_taskreg_set_data_arg): Likewise.
813 (gimple_omp_task_copy_fn): Likewise.
814 (gimple_omp_task_copy_fn_ptr): Likewise.
815 (gimple_omp_task_set_copy_fn): Likewise.
816 (gimple_omp_task_arg_size): Likewise.
817 (gimple_omp_task_arg_size_ptr): Likewise.
818 (gimple_omp_task_set_arg_size): Likewise.
819 (gimple_omp_task_arg_align): Likewise.
820 (gimple_omp_task_arg_align_ptr): Likewise.
821 (gimple_omp_task_set_arg_align): Likewise.
822 (gimple_omp_single_clauses): Likewise.
823 (gimple_omp_single_clauses_ptr): Likewise.
824 (gimple_omp_single_set_clauses): Likewise.
825 (gimple_omp_target_clauses): Likewise.
826 (gimple_omp_target_clauses_ptr): Likewise.
827 (gimple_omp_target_set_clauses): Likewise.
828 (gimple_omp_target_child_fn): Likewise.
829 (gimple_omp_target_child_fn_ptr): Likewise.
830 (gimple_omp_target_set_child_fn): Likewise.
831 (gimple_omp_target_data_arg): Likewise.
832 (gimple_omp_target_data_arg_ptr): Likewise.
833 (gimple_omp_target_set_data_arg): Likewise.
834 (gimple_omp_teams_clauses): Likewise.
835 (gimple_omp_teams_clauses_ptr): Likewise.
836 (gimple_omp_teams_set_clauses): Likewise.
837 (gimple_omp_sections_clauses): Likewise.
838 (gimple_omp_sections_clauses_ptr): Likewise.
839 (gimple_omp_sections_set_clauses): Likewise.
840 (gimple_omp_sections_control): Likewise.
841 (gimple_omp_sections_control_ptr): Likewise.
842 (gimple_omp_sections_set_control): Likewise.
843 (gimple_omp_for_set_cond): Likewise.
844 (gimple_omp_for_cond): Likewise.
845 (gimple_omp_atomic_store_set_val): Likewise.
846 (gimple_omp_atomic_store_val): Likewise.
847 (gimple_omp_atomic_store_val_ptr): Likewise.
848 (gimple_omp_atomic_load_set_lhs): Likewise.
849 (gimple_omp_atomic_load_lhs): Likewise.
850 (gimple_omp_atomic_load_lhs_ptr): Likewise.
851 (gimple_omp_atomic_load_set_rhs): Likewise.
852 (gimple_omp_atomic_load_rhs): Likewise.
853 (gimple_omp_atomic_load_rhs_ptr): Likewise.
854 (gimple_omp_continue_control_def): Likewise.
855 (gimple_omp_continue_control_def_ptr): Likewise.
856 (gimple_omp_continue_set_control_def): Likewise.
857 (gimple_omp_continue_control_use): Likewise.
858 (gimple_omp_continue_control_use_ptr): Likewise.
859 (gimple_omp_continue_set_control_use): Likewise.
860 (gimple_transaction_body_ptr): Likewise.
861 (gimple_transaction_label): Likewise.
862 (gimple_transaction_label_ptr): Likewise.
863 (gimple_transaction_set_body): Likewise.
864 (gimple_transaction_set_label): Likewise.
866 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
867 * ipa-inline-analysis.c (inline_write_summary): Likewise.
868 * ipa-ref.c (ipa_record_reference): Likewise.
869 * ipa-reference.c (analyze_function): Likewise.
870 (ipa_reference_write_optimization_summary): Likewise.
871 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
872 (address_taken_from_non_vtable_p): Likewise.
873 (comdat_can_be_unshared_p_1): Likewise.
874 * lto-cgraph.c (lto_output_ref): Likewise.
875 (add_references): Likewise.
876 (compute_ltrans_boundary): Likewise.
877 (output_symtab): Likewise.
878 (input_ref): Likewise.
879 (input_cgraph_1): Likewise.
880 (output_cgraph_opt_summary): Likewise.
881 * lto-streamer-out.c (lto_output): Likewise.
882 (output_symbol_p): Likewise.
883 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
884 (lsei_start_function_in_partition): Likewise.
885 (lsei_next_variable_in_partition): Likewise.
886 (lsei_start_variable_in_partition): Likewise.
887 * symtab.c (insert_to_assembler_name_hash): Likewise.
888 (unlink_from_assembler_name_hash): Likewise.
889 (symtab_unregister_node): Likewise.
890 (symtab_remove_node): Likewise.
891 (dump_symtab_node): Likewise.
892 (verify_symtab_base): Likewise.
893 (verify_symtab_node): Likewise.
894 (symtab_make_decl_local): Likewise.
895 (symtab_alias_ultimate_target): Likewise.
896 (symtab_resolve_alias): Likewise.
897 (symtab_get_symbol_partitioning_class): Likewise.
898 * tree-phinodes.c (allocate_phi_node): Likewise.
899 (reserve_phi_args_for_new_edge): Likewise.
900 (remove_phi_args): Likewise.
901 * varpool.c (varpool_node_for_asm): Likewise.
902 (varpool_remove_unreferenced_decls): Likewise.
904 2014-04-23 Jeff Law <law@redhat.com>
906 PR tree-optimization/60902
907 * tree-ssa-threadedge.c
908 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
909 invalidate outputs from statements that do not produce useful
910 outputs for threading.
912 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
914 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
915 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
916 machine descriptions for Stack Smashing Protector.
918 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
920 * aarch64.md (<optab>_rol<mode>3): New pattern.
921 (<optab>_rolsi3_uxtw): Likewise.
922 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
924 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
926 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
927 (arm_cortex_a12_tune): Likewise.
929 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
931 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
933 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
935 * config/arm/arm.md (arm_rev16si2): New pattern.
936 (arm_rev16si2_alt): Likewise.
937 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
939 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
941 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
942 (rev16<mode>2_alt): Likewise.
943 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
944 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
945 (aarch_rev16_shleft_mask_imm_p): Likewise.
946 (aarch_rev16_p_1): Likewise.
947 (aarch_rev16_p): Likewise.
948 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
949 (aarch_rev16_shright_mask_imm_p): Likewise.
950 (aarch_rev16_shleft_mask_imm_p): Likewise.
952 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
954 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
955 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
957 (cortex_a53_extra_costs): Likewise.
958 (cortex_a57_extra_costs): Likewise.
959 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
960 (cortexa7_extra_costs): Likewise.
961 (cortexa8_extra_costs): Likewise.
962 (cortexa12_extra_costs): Likewise.
963 (cortexa15_extra_costs): Likewise.
964 (v7m_extra_costs): Likewise.
965 (arm_new_rtx_costs): Handle BSWAP.
967 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
969 * config/arm/arm.c (cortexa8_extra_costs): New table.
970 (arm_cortex_a8_tune): New tuning struct.
971 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
973 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
975 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
977 2014-04-23 Richard Biener <rguenther@suse.de>
979 * Makefile.in (OBJS): Remove loop-unswitch.o.
980 * loop-unswitch.c: Delete.
981 * tree-pass.h (make_pass_rtl_unswitch): Remove.
982 * passes.def (pass_rtl_unswitch): Likewise.
983 * loop-init.c (gate_rtl_unswitch): Likewise.
984 (rtl_unswitch): Likewise.
985 (pass_data_rtl_unswitch): Likewise.
986 (pass_rtl_unswitch): Likewise.
987 (make_pass_rtl_unswitch): Likewise.
988 * rtl.h (reversed_condition): Likewise.
989 (compare_and_jump_seq): Likewise.
990 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
992 * loop-unroll.c (compare_and_jump_seq): Likewise.
994 2014-04-23 Richard Biener <rguenther@suse.de>
996 PR tree-optimization/60903
997 * tree-ssa-loop-im.c (analyze_memory_references): Remove
998 commented code block.
999 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
1000 loop flags to newly created BBs and edges.
1002 2014-04-23 Nick Clifton <nickc@redhat.com>
1004 * config/msp430/msp430.c (msp430_handle_option): Move function
1006 (msp430_option_override): Simplify mcu and mcpu option handling.
1007 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
1008 support for -mhwmult command line option.
1009 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
1010 -mhwmult command line option.
1011 (msp430_hwmult_enabled): Delete.
1012 (msp43o_output_labelref): Add support for -mhwmult command line
1014 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
1015 (umulsidi3): Likewise.
1016 * config/msp430/msp430.opt (mmcu): Add Report attribute.
1017 (mcpu, mlarge, msmall): Likewise.
1018 (mhwmult): New option.
1019 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
1021 (msp430_is_f5_mcu): Remove prototype.
1022 (msp430_use_f5_series_hwmult): Add prototype.
1023 * config/msp430/msp430-opts.h: New file.
1024 * common/config/msp430: New directory.
1025 * common/config/msp430/msp430-common.c: New file.
1026 * config.gcc (msp430): Remove target_has_targetm_common.
1027 * doc/invoke.texi: Document -mhwmult command line option.
1029 2014-04-23 Nick Clifton <nickc@redhat.com>
1031 * config/i386/cygwin.h (ENDFILE_SPEC): Include
1032 default-manifest.o if it can be found in the search path.
1033 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
1035 2014-04-23 Terry Guo <terry.guo@arm.com>
1037 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
1039 2014-04-23 Richard Biener <rguenther@suse.de>
1042 * tree-inline.c (declare_return_variable): Use mark_addressable.
1044 2014-04-23 Richard Biener <rguenther@suse.de>
1047 * loop-init.c (loop_optimizer_init): Make sure to apply
1048 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
1050 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1053 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
1055 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
1056 if flag_sanitize_undefined_trap_on_error.
1057 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
1058 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
1059 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
1060 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
1061 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
1062 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
1063 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
1064 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
1065 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
1066 * ubsan.c (ubsan_instrument_unreachable): Return
1067 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1068 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1069 if flag_sanitize_undefined_trap_on_error and
1070 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1071 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1072 instrument_bool_enum_load): Emit __builtin_trap () if
1073 flag_sanitize_undefined_trap_on_error and
1074 __builtin_handle_*_abort () if !flag_sanitize_recover.
1075 * doc/invoke.texi (-fsanitize-recover,
1076 -fsanitize-undefined-trap-on-error): Document.
1078 2014-04-22 Christian Bruel <christian.bruel@st.com>
1080 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1081 Force immediates to SImode.
1083 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1085 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1087 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1088 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1089 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1090 (nios2_fpu_insn): Add entry for round.
1091 (N2FPU_NO_ERRNO_P): Define.
1092 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1094 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1096 2014-04-22 Richard Henderson <rth@redhat.com>
1098 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1099 (add<GPI>3_compare0): Remove leading * from name.
1100 (add<GPI>3_carryin): Likewise.
1101 (sub<GPI>3_compare0): Likewise.
1102 (sub<GPI>3_carryin): Likewise.
1103 (<su_optab>mulditi3): New expander.
1104 (multi3): New expander.
1105 (madd<GPI>): Remove leading * from name.
1107 2014-04-22 Martin Jambor <mjambor@suse.cz>
1109 * cgraphclones.c (cgraph_function_versioning): Copy
1110 ipa_transforms_to_apply instead of asserting it is empty.
1112 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1115 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1116 on count_exp to get mode.
1118 2014-04-22 Andrew Pinski <apinski@cavium.com>
1120 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1121 Handle TLS for ILP32.
1122 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1123 (tlsie_small_<mode>): this and handle PTR.
1124 (tlsie_small_sidi): New pattern.
1125 (tlsle_small): Change to an expand to handle ILP32.
1126 (tlsle_small_<mode>): New pattern.
1127 (tlsdesc_small): Rename to ...
1128 (tlsdesc_small_<mode>): this and handle PTR.
1130 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1132 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1134 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1136 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1137 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1138 (aarch64_types_signed_poly_qualifiers): Likewise.
1139 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1140 (aarch64_types_poly_signed_qualifiers): Likewise.
1141 (TYPES_REINTERP_SS): Type macro added.
1142 (TYPES_REINTERP_SU): Likewise.
1143 (TYPES_REINTERP_SP): Likewise.
1144 (TYPES_REINTERP_US): Likewise.
1145 (TYPES_REINTERP_PS): Likewise.
1146 (aarch64_fold_builtin): New expression folding added.
1147 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1148 Declarations removed.
1149 (REINTERP_SS): Declarations added.
1150 (REINTERP_US): Likewise.
1151 (REINTERP_PS): Likewise.
1152 (REINTERP_SU): Likewise.
1153 (REINTERP_SP): Likewise.
1154 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1155 (vreinterpretq_p8_f64): Likewise.
1156 (vreinterpret_p16_f64): Likewise.
1157 (vreinterpretq_p16_f64): Likewise.
1158 (vreinterpret_f32_f64): Likewise.
1159 (vreinterpretq_f32_f64): Likewise.
1160 (vreinterpret_f64_f32): Likewise.
1161 (vreinterpret_f64_p8): Likewise.
1162 (vreinterpret_f64_p16): Likewise.
1163 (vreinterpret_f64_s8): Likewise.
1164 (vreinterpret_f64_s16): Likewise.
1165 (vreinterpret_f64_s32): Likewise.
1166 (vreinterpret_f64_s64): Likewise.
1167 (vreinterpret_f64_u8): Likewise.
1168 (vreinterpret_f64_u16): Likewise.
1169 (vreinterpret_f64_u32): Likewise.
1170 (vreinterpret_f64_u64): Likewise.
1171 (vreinterpretq_f64_f32): Likewise.
1172 (vreinterpretq_f64_p8): Likewise.
1173 (vreinterpretq_f64_p16): Likewise.
1174 (vreinterpretq_f64_s8): Likewise.
1175 (vreinterpretq_f64_s16): Likewise.
1176 (vreinterpretq_f64_s32): Likewise.
1177 (vreinterpretq_f64_s64): Likewise.
1178 (vreinterpretq_f64_u8): Likewise.
1179 (vreinterpretq_f64_u16): Likewise.
1180 (vreinterpretq_f64_u32): Likewise.
1181 (vreinterpretq_f64_u64): Likewise.
1182 (vreinterpret_s64_f64): Likewise.
1183 (vreinterpretq_s64_f64): Likewise.
1184 (vreinterpret_u64_f64): Likewise.
1185 (vreinterpretq_u64_f64): Likewise.
1186 (vreinterpret_s8_f64): Likewise.
1187 (vreinterpretq_s8_f64): Likewise.
1188 (vreinterpret_s16_f64): Likewise.
1189 (vreinterpretq_s16_f64): Likewise.
1190 (vreinterpret_s32_f64): Likewise.
1191 (vreinterpretq_s32_f64): Likewise.
1192 (vreinterpret_u8_f64): Likewise.
1193 (vreinterpretq_u8_f64): Likewise.
1194 (vreinterpret_u16_f64): Likewise.
1195 (vreinterpretq_u16_f64): Likewise.
1196 (vreinterpret_u32_f64): Likewise.
1197 (vreinterpretq_u32_f64): Likewise.
1199 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1201 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1202 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1203 (vreinterpret_p8_s8): Likewise.
1204 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1205 (vreinterpret_p8_s16): Likewise.
1206 (vreinterpret_p8_s32): Likewise.
1207 (vreinterpret_p8_s64): Likewise.
1208 (vreinterpret_p8_f32): Likewise.
1209 (vreinterpret_p8_u8): Likewise.
1210 (vreinterpret_p8_u16): Likewise.
1211 (vreinterpret_p8_u32): Likewise.
1212 (vreinterpret_p8_u64): Likewise.
1213 (vreinterpret_p8_p16): Likewise.
1214 (vreinterpretq_p8_s8): Likewise.
1215 (vreinterpretq_p8_s16): Likewise.
1216 (vreinterpretq_p8_s32): Likewise.
1217 (vreinterpretq_p8_s64): Likewise.
1218 (vreinterpretq_p8_f32): Likewise.
1219 (vreinterpretq_p8_u8): Likewise.
1220 (vreinterpretq_p8_u16): Likewise.
1221 (vreinterpretq_p8_u32): Likewise.
1222 (vreinterpretq_p8_u64): Likewise.
1223 (vreinterpretq_p8_p16): Likewise.
1224 (vreinterpret_p16_s8): Likewise.
1225 (vreinterpret_p16_s16): Likewise.
1226 (vreinterpret_p16_s32): Likewise.
1227 (vreinterpret_p16_s64): Likewise.
1228 (vreinterpret_p16_f32): Likewise.
1229 (vreinterpret_p16_u8): Likewise.
1230 (vreinterpret_p16_u16): Likewise.
1231 (vreinterpret_p16_u32): Likewise.
1232 (vreinterpret_p16_u64): Likewise.
1233 (vreinterpret_p16_p8): Likewise.
1234 (vreinterpretq_p16_s8): Likewise.
1235 (vreinterpretq_p16_s16): Likewise.
1236 (vreinterpretq_p16_s32): Likewise.
1237 (vreinterpretq_p16_s64): Likewise.
1238 (vreinterpretq_p16_f32): Likewise.
1239 (vreinterpretq_p16_u8): Likewise.
1240 (vreinterpretq_p16_u16): Likewise.
1241 (vreinterpretq_p16_u32): Likewise.
1242 (vreinterpretq_p16_u64): Likewise.
1243 (vreinterpretq_p16_p8): Likewise.
1244 (vreinterpret_f32_s8): Likewise.
1245 (vreinterpret_f32_s16): Likewise.
1246 (vreinterpret_f32_s32): Likewise.
1247 (vreinterpret_f32_s64): Likewise.
1248 (vreinterpret_f32_u8): Likewise.
1249 (vreinterpret_f32_u16): Likewise.
1250 (vreinterpret_f32_u32): Likewise.
1251 (vreinterpret_f32_u64): Likewise.
1252 (vreinterpret_f32_p8): Likewise.
1253 (vreinterpret_f32_p16): Likewise.
1254 (vreinterpretq_f32_s8): Likewise.
1255 (vreinterpretq_f32_s16): Likewise.
1256 (vreinterpretq_f32_s32): Likewise.
1257 (vreinterpretq_f32_s64): Likewise.
1258 (vreinterpretq_f32_u8): Likewise.
1259 (vreinterpretq_f32_u16): Likewise.
1260 (vreinterpretq_f32_u32): Likewise.
1261 (vreinterpretq_f32_u64): Likewise.
1262 (vreinterpretq_f32_p8): Likewise.
1263 (vreinterpretq_f32_p16): Likewise.
1264 (vreinterpret_s64_s8): Likewise.
1265 (vreinterpret_s64_s16): Likewise.
1266 (vreinterpret_s64_s32): Likewise.
1267 (vreinterpret_s64_f32): Likewise.
1268 (vreinterpret_s64_u8): Likewise.
1269 (vreinterpret_s64_u16): Likewise.
1270 (vreinterpret_s64_u32): Likewise.
1271 (vreinterpret_s64_u64): Likewise.
1272 (vreinterpret_s64_p8): Likewise.
1273 (vreinterpret_s64_p16): Likewise.
1274 (vreinterpretq_s64_s8): Likewise.
1275 (vreinterpretq_s64_s16): Likewise.
1276 (vreinterpretq_s64_s32): Likewise.
1277 (vreinterpretq_s64_f32): Likewise.
1278 (vreinterpretq_s64_u8): Likewise.
1279 (vreinterpretq_s64_u16): Likewise.
1280 (vreinterpretq_s64_u32): Likewise.
1281 (vreinterpretq_s64_u64): Likewise.
1282 (vreinterpretq_s64_p8): Likewise.
1283 (vreinterpretq_s64_p16): Likewise.
1284 (vreinterpret_u64_s8): Likewise.
1285 (vreinterpret_u64_s16): Likewise.
1286 (vreinterpret_u64_s32): Likewise.
1287 (vreinterpret_u64_s64): Likewise.
1288 (vreinterpret_u64_f32): Likewise.
1289 (vreinterpret_u64_u8): Likewise.
1290 (vreinterpret_u64_u16): Likewise.
1291 (vreinterpret_u64_u32): Likewise.
1292 (vreinterpret_u64_p8): Likewise.
1293 (vreinterpret_u64_p16): Likewise.
1294 (vreinterpretq_u64_s8): Likewise.
1295 (vreinterpretq_u64_s16): Likewise.
1296 (vreinterpretq_u64_s32): Likewise.
1297 (vreinterpretq_u64_s64): Likewise.
1298 (vreinterpretq_u64_f32): Likewise.
1299 (vreinterpretq_u64_u8): Likewise.
1300 (vreinterpretq_u64_u16): Likewise.
1301 (vreinterpretq_u64_u32): Likewise.
1302 (vreinterpretq_u64_p8): Likewise.
1303 (vreinterpretq_u64_p16): Likewise.
1304 (vreinterpret_s8_s16): Likewise.
1305 (vreinterpret_s8_s32): Likewise.
1306 (vreinterpret_s8_s64): Likewise.
1307 (vreinterpret_s8_f32): Likewise.
1308 (vreinterpret_s8_u8): Likewise.
1309 (vreinterpret_s8_u16): Likewise.
1310 (vreinterpret_s8_u32): Likewise.
1311 (vreinterpret_s8_u64): Likewise.
1312 (vreinterpret_s8_p8): Likewise.
1313 (vreinterpret_s8_p16): Likewise.
1314 (vreinterpretq_s8_s16): Likewise.
1315 (vreinterpretq_s8_s32): Likewise.
1316 (vreinterpretq_s8_s64): Likewise.
1317 (vreinterpretq_s8_f32): Likewise.
1318 (vreinterpretq_s8_u8): Likewise.
1319 (vreinterpretq_s8_u16): Likewise.
1320 (vreinterpretq_s8_u32): Likewise.
1321 (vreinterpretq_s8_u64): Likewise.
1322 (vreinterpretq_s8_p8): Likewise.
1323 (vreinterpretq_s8_p16): Likewise.
1324 (vreinterpret_s16_s8): Likewise.
1325 (vreinterpret_s16_s32): Likewise.
1326 (vreinterpret_s16_s64): Likewise.
1327 (vreinterpret_s16_f32): Likewise.
1328 (vreinterpret_s16_u8): Likewise.
1329 (vreinterpret_s16_u16): Likewise.
1330 (vreinterpret_s16_u32): Likewise.
1331 (vreinterpret_s16_u64): Likewise.
1332 (vreinterpret_s16_p8): Likewise.
1333 (vreinterpret_s16_p16): Likewise.
1334 (vreinterpretq_s16_s8): Likewise.
1335 (vreinterpretq_s16_s32): Likewise.
1336 (vreinterpretq_s16_s64): Likewise.
1337 (vreinterpretq_s16_f32): Likewise.
1338 (vreinterpretq_s16_u8): Likewise.
1339 (vreinterpretq_s16_u16): Likewise.
1340 (vreinterpretq_s16_u32): Likewise.
1341 (vreinterpretq_s16_u64): Likewise.
1342 (vreinterpretq_s16_p8): Likewise.
1343 (vreinterpretq_s16_p16): Likewise.
1344 (vreinterpret_s32_s8): Likewise.
1345 (vreinterpret_s32_s16): Likewise.
1346 (vreinterpret_s32_s64): Likewise.
1347 (vreinterpret_s32_f32): Likewise.
1348 (vreinterpret_s32_u8): Likewise.
1349 (vreinterpret_s32_u16): Likewise.
1350 (vreinterpret_s32_u32): Likewise.
1351 (vreinterpret_s32_u64): Likewise.
1352 (vreinterpret_s32_p8): Likewise.
1353 (vreinterpret_s32_p16): Likewise.
1354 (vreinterpretq_s32_s8): Likewise.
1355 (vreinterpretq_s32_s16): Likewise.
1356 (vreinterpretq_s32_s64): Likewise.
1357 (vreinterpretq_s32_f32): Likewise.
1358 (vreinterpretq_s32_u8): Likewise.
1359 (vreinterpretq_s32_u16): Likewise.
1360 (vreinterpretq_s32_u32): Likewise.
1361 (vreinterpretq_s32_u64): Likewise.
1362 (vreinterpretq_s32_p8): Likewise.
1363 (vreinterpretq_s32_p16): Likewise.
1364 (vreinterpret_u8_s8): Likewise.
1365 (vreinterpret_u8_s16): Likewise.
1366 (vreinterpret_u8_s32): Likewise.
1367 (vreinterpret_u8_s64): Likewise.
1368 (vreinterpret_u8_f32): Likewise.
1369 (vreinterpret_u8_u16): Likewise.
1370 (vreinterpret_u8_u32): Likewise.
1371 (vreinterpret_u8_u64): Likewise.
1372 (vreinterpret_u8_p8): Likewise.
1373 (vreinterpret_u8_p16): Likewise.
1374 (vreinterpretq_u8_s8): Likewise.
1375 (vreinterpretq_u8_s16): Likewise.
1376 (vreinterpretq_u8_s32): Likewise.
1377 (vreinterpretq_u8_s64): Likewise.
1378 (vreinterpretq_u8_f32): Likewise.
1379 (vreinterpretq_u8_u16): Likewise.
1380 (vreinterpretq_u8_u32): Likewise.
1381 (vreinterpretq_u8_u64): Likewise.
1382 (vreinterpretq_u8_p8): Likewise.
1383 (vreinterpretq_u8_p16): Likewise.
1384 (vreinterpret_u16_s8): Likewise.
1385 (vreinterpret_u16_s16): Likewise.
1386 (vreinterpret_u16_s32): Likewise.
1387 (vreinterpret_u16_s64): Likewise.
1388 (vreinterpret_u16_f32): Likewise.
1389 (vreinterpret_u16_u8): Likewise.
1390 (vreinterpret_u16_u32): Likewise.
1391 (vreinterpret_u16_u64): Likewise.
1392 (vreinterpret_u16_p8): Likewise.
1393 (vreinterpret_u16_p16): Likewise.
1394 (vreinterpretq_u16_s8): Likewise.
1395 (vreinterpretq_u16_s16): Likewise.
1396 (vreinterpretq_u16_s32): Likewise.
1397 (vreinterpretq_u16_s64): Likewise.
1398 (vreinterpretq_u16_f32): Likewise.
1399 (vreinterpretq_u16_u8): Likewise.
1400 (vreinterpretq_u16_u32): Likewise.
1401 (vreinterpretq_u16_u64): Likewise.
1402 (vreinterpretq_u16_p8): Likewise.
1403 (vreinterpretq_u16_p16): Likewise.
1404 (vreinterpret_u32_s8): Likewise.
1405 (vreinterpret_u32_s16): Likewise.
1406 (vreinterpret_u32_s32): Likewise.
1407 (vreinterpret_u32_s64): Likewise.
1408 (vreinterpret_u32_f32): Likewise.
1409 (vreinterpret_u32_u8): Likewise.
1410 (vreinterpret_u32_u16): Likewise.
1411 (vreinterpret_u32_u64): Likewise.
1412 (vreinterpret_u32_p8): Likewise.
1413 (vreinterpret_u32_p16): Likewise.
1414 (vreinterpretq_u32_s8): Likewise.
1415 (vreinterpretq_u32_s16): Likewise.
1416 (vreinterpretq_u32_s32): Likewise.
1417 (vreinterpretq_u32_s64): Likewise.
1418 (vreinterpretq_u32_f32): Likewise.
1419 (vreinterpretq_u32_u8): Likewise.
1420 (vreinterpretq_u32_u16): Likewise.
1421 (vreinterpretq_u32_u64): Likewise.
1422 (vreinterpretq_u32_p8): Likewise.
1423 (vreinterpretq_u32_p16): Likewise.
1425 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1427 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1429 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1432 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1433 (vqnegd_s64): Likewise.
1434 (vqabs_s64): Likewise.
1435 (vqabsd_s64): Likewise.
1437 2014-04-22 Richard Henderson <rth@redhat.com>
1439 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1440 computation to the top of the loop.
1442 2014-04-22 Renlin <renlin.li@arm.com>
1443 Jiong Wang <jiong.wang@arm.com>
1445 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1446 * config/aarch64/aarch64.c (aarch64_layout_frame)
1447 (aarch64_initial_elimination_offset): Likewise.
1449 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1451 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1454 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1456 * machmode.h (bitwise_mode_for_mode): Declare.
1457 * stor-layout.h (bitwise_type_for_mode): Likewise.
1458 * stor-layout.c (bitwise_mode_for_mode): New function.
1459 (bitwise_type_for_mode): Likewise.
1460 * builtins.c (fold_builtin_memory_op): Use it instead of
1461 int_mode_for_mode and build_nonstandard_integer_type.
1463 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1465 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1466 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1467 (*-*-solaris2*): Simplify.
1468 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1469 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1470 *-*-solaris2.9* handling.
1472 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1474 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1475 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1477 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1478 * configure: Regenerate.
1480 * config/i386/sol2-9.h: Remove.
1482 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1483 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1484 Remove Solaris 9 references.
1486 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1488 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1489 (floatuns<GPI:mode><GPF:mode>2): Remove.
1490 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1491 and floatuns conversions.
1492 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1493 and floatuns conversions.
1494 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1495 (w1,w2): New mode attributes for inequal width conversions.
1497 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1499 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1500 the output asm format.
1502 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1504 * config/aarch64/aarch64-simd.md
1505 (aarch64_cm<optab>di): Always split.
1506 (*aarch64_cm<optab>di): New.
1507 (aarch64_cmtstdi): Always split.
1508 (*aarch64_cmtstdi): New.
1510 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1512 PR tree-optimization/60823
1513 * omp-low.c (ipa_simd_modify_function_body): Go through
1514 all SSA_NAMEs and for those refering to vector arguments
1515 which are going to be replaced adjust SSA_NAME_VAR and,
1516 if it is a default definition, change it into a non-default
1517 definition assigned at the beginning of function from new_decl.
1518 (ipa_simd_modify_stmt_ops): Rewritten.
1519 * tree-dfa.c (set_ssa_default_def): When removing default def,
1520 check for NULL loc instead of NULL *loc.
1522 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1524 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1525 restrictions on core registers for DImode values in Thumb2.
1527 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1529 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1530 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1532 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1534 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1535 (*iordi_notzesidi_di): Likewise.
1536 (*iordi_notsesidi_di): Likewise.
1538 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1540 * config/arm/arm-protos.h (tune_params): New struct members.
1541 * config/arm/arm.c: Initialise tune_params per processor.
1542 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1543 for speed, based on new tune_params.
1545 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1547 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1549 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1551 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1553 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1554 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1555 (vrnda_f64): Likewise.
1556 (vrndi_f64): Likewise.
1557 (vrndm_f64): Likewise.
1558 (vrndn_f64): Likewise.
1559 (vrndp_f64): Likewise.
1560 (vrndx_f64): Likewise.
1562 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1564 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1565 GET_MODE_SIZE argument is enum machine_mode.
1567 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1570 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1571 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1573 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1576 * asan.c (asan_emit_stack_protection): Force the base to align to
1577 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1578 appropriate bits if STRICT_ALIGNMENT.
1579 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1581 (expand_used_vars): Leave a space in the stack frame for alignment
1582 if STRICT_ALIGNMENT.
1584 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1586 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1588 (gimple_store_p): Likewise.
1589 (gimple_assign_load_p): Likewise.
1590 (gimple_assign_cast_p): Likewise.
1591 (gimple_clobber_p): Likewise.
1593 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1594 rather than a gimple.
1595 (gimple_assign_cast_p): Likewise.
1597 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1600 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1601 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1603 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1604 more debug information for E500 if -mdebug=reg.
1606 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1609 * config/i386/i386.c (ix86_expand_builtin)
1610 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1611 register for target RTX.
1612 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1614 2014-04-18 Cong Hou <congh@google.com>
1616 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1617 the widen-mult pattern by handling two operands with different sizes,
1618 and operands whose size is smaller than half of the result type.
1620 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1622 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1623 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1624 (do_estimate_edge_time): Compute it.
1625 * ipa-inline.c (want_inline_small_function_p): Bypass
1626 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1628 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1630 * ipa-inline.c (spec_rem): New static variable.
1631 (dump_overall_stats): New function.
1632 (dump_inline_stats): New function.
1634 2014-04-18 Richard Henderson <rth@redhat.com>
1636 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1637 to GET_MODE_SIZE, not a reg_class_t.
1639 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1641 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1642 (vsx_xxmrglw_<mode>): Likewise.
1644 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1647 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1648 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1649 (rs6000_init_hard_regno_mode_ok): Likewise.
1651 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1653 * ipa-inline.c (inline_small_functions): Account only non-cold
1655 * doc/invoke.texi (inline-unit-growth): Update documentation.
1657 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1659 * config/rs6000/rs6000.md (addti3, subti3): New.
1661 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1664 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1665 comment. Check optimize_insn_for_size_p instead of
1666 optimize_insn_for_speed_p.
1668 2014-04-17 Martin Jambor <mjambor@suse.cz>
1670 * gimple-iterator.c (gsi_start_edge): New function.
1671 * gimple-iterator.h (gsi_start_edge): Declare.
1672 * tree-sra.c (single_non_eh_succ): New function.
1673 (disqualify_ops_if_throwing_stmt): Renamed to
1674 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1675 having one non-EH successor BB.
1676 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1677 generate loads into replacements.
1678 (sra_modify_assign): Likewise and and also use the simple path for
1680 (sra_modify_function_body): Commit statements on edges.
1682 2014-04-17 Richard Biener <rguenther@suse.de>
1685 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1686 comparison results and add clarifying comment.
1688 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1690 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1691 (blank_mode): Initialize it.
1692 (emit_mode_size_inline, emit_mode_nunits_inline,
1693 emit_mode_inner_inline): New functions.
1694 (emit_insn_modes_h): Call them and surround their output with
1695 #if GCC_VERSION >= 4001 ... #endif.
1696 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1697 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1698 mode_* arrays if the argument is __builtin_constant_p.
1699 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1700 is enum machine_mode.
1702 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1704 * passes.c (opt_pass::execute): Adjust.
1705 (pass_manager::execute_pass_mode_switching): Likewise.
1706 (early_local_passes::execute): Likewise.
1707 (execute_one_pass): Pass cfun to the pass's execute method.
1708 * tree-pass.h (opt_pass::execute): Add function * argument.
1709 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1710 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1711 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1712 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1713 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1714 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1715 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1716 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1717 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1718 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1719 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1720 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1721 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1722 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1723 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1724 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1725 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1726 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1727 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1728 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1729 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1730 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1731 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1732 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1733 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1734 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1735 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1736 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1737 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1740 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1742 * passes.c (opt_pass::gate): Take function * argument.
1743 (gate_all_early_local_passes): Merge into
1744 (early_local_passes::gate): this.
1745 (gate_all_early_optimizations): Merge into
1746 (all_early_optimizations::gate): this.
1747 (gate_all_optimizations): Mege into
1748 (all_optimizations::gate): this.
1749 (gate_all_optimizations_g): Merge into
1750 (all_optimizations_g::gate): this.
1751 (gate_rest_of_compilation): Mege into
1752 (rest_of_compilation::gate): this.
1753 (gate_postreload): Merge into
1754 (postreload::gate): this.
1755 (dump_one_pass): Pass cfun to the pass's gate method.
1756 (execute_ipa_summary_passes): Likewise.
1757 (execute_one_pass): Likewise.
1758 (ipa_write_summaries_2): Likewise.
1759 (ipa_write_optimization_summaries_1): Likewise.
1760 (ipa_read_summaries_1): Likewise.
1761 (ipa_read_optimization_summaries_1): Likewise.
1762 (execute_ipa_stmt_fixups): Likewise.
1763 * tree-pass.h (opt_pass::gate): Add function * argument.
1764 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
1765 combine-stack-adj.c, combine.c, compare-elim.c,
1766 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1767 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
1768 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
1769 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
1770 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1771 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1772 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1773 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
1774 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
1775 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
1776 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1777 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1778 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
1779 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1780 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1781 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1782 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1783 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1784 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1785 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1786 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1787 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1788 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1789 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
1790 var-tracking.c, vtable-verify.c, web.c: Adjust.
1792 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1794 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
1795 * configure: Regenerate.
1797 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1799 * passes.c (dump_one_pass): don't check pass->has_gate.
1800 (execute_ipa_summary_passes): Likewise.
1801 (execute_one_pass): Likewise.
1802 (ipa_write_summaries_2): Likewise.
1803 (ipa_write_optimization_summaries_1): Likewise.
1804 (ipa_read_optimization_summaries_1): Likewise.
1805 (execute_ipa_stmt_fixups): Likewise.
1806 * tree-pass.h (pass_data::has_gate): Remove.
1807 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1808 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1809 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1810 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1811 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1812 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
1813 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
1814 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
1815 gimple-low.c, gimple-ssa-isolate-paths.c,
1816 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
1817 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
1818 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1819 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
1820 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
1821 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
1822 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
1823 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1824 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
1825 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1826 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1827 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1828 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1829 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1830 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1831 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1832 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1833 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1834 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1835 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1836 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1837 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1840 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1842 * pass_manager.h (pass_manager::register_dump_files_1): Remove
1844 * passes.c (pass_manager::register_dump_files_1): Merge into
1845 (pass_manager::register_dump_files): this, and remove its handling of
1846 properties since the pass always has the properties anyway.
1847 (pass_manager::pass_manager): Adjust.
1849 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1851 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
1852 * passes.c (pass_manager::register_dump_files_1): Remove dead code
1853 dealing with properties.
1854 (pass_manager::register_dump_files): Adjust.
1856 2014-03-20 Mark Wielaard <mjw@redhat.com>
1858 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
1859 then represent the bound as normal constant value.
1861 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1864 Forward port from 4.8 branch
1865 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
1867 * config/i386/bmiintrin.h (_blsi_u32): New.
1871 (_blsmsk_u32): Ditto.
1872 (_blsmsk_u64): Ditto.
1873 (_tzcnt_u32): Ditto.
1874 (_tzcnt_u64): Ditto.
1876 2014-04-17 Kito Cheng <kito@0xlab.org>
1878 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
1880 2014-04-17 Richard Biener <rguenther@suse.de>
1883 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
1884 boolean results for comparisons.
1886 2014-04-17 Richard Biener <rguenther@suse.de>
1888 PR tree-optimization/60836
1889 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
1890 initial PHI args to be gimple values.
1892 2014-04-17 Richard Biener <rguenther@suse.de>
1894 PR tree-optimization/60841
1895 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
1896 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
1897 of stmts to SLP build.
1898 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
1899 (vect_analyze_slp): Likewise.
1900 (vect_analyze_slp_instance): Likewise.
1901 (vect_build_slp_tree): Limit overall SLP tree growth.
1902 * tree-vectorizer.h (vect_analyze_data_refs,
1903 vect_analyze_slp): Adjust prototypes.
1905 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1907 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
1910 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1912 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
1913 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
1914 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
1915 for TARGET_SLOW_PSHUFB
1917 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1919 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
1920 * config/i386/i386.c (intel_cost): Ditto.
1922 2014-04-17 Joey Ye <joey.ye@arm.com>
1924 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
1926 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1928 * opts.c (common_handle_option): Disable -fipa-reference coorectly
1931 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1933 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
1934 (type_all_derivations_known_p): New predicate.
1935 (type_all_ctors_visible_p): New predicate.
1936 (type_possibly_instantiated_p): New predicate.
1937 (get_odr_type): Compute all_derivations_known.
1938 (dump_odr_type): Dump the flag.
1939 (maybe_record_type): Cleanup.
1940 (record_target_from_binfo): Add bases_to_consider array;
1941 record bases for types w/o instances and skip CXX destructor.
1942 (possible_polymorphic_call_targets_1): Add bases_to_consider
1943 and consider_construction parameters; check if type may have instance.
1944 (get_polymorphic_call_info): Set maybe_in_construction to true
1945 when we know nothing.
1946 (record_targets_from_bases): Skip CXX destructors; they are
1947 never called for types in construction.
1948 (possible_polymorphic_call_targets): Do not record target when
1949 type may not have instance.
1951 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1954 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
1955 external aliases alive, too.
1957 2014-04-16 Andrew Pinski <apinski@cavium.com>
1959 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
1962 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1964 * final.c (compute_alignments): Do not apply loop alignment to a block
1965 falling through to the exit.
1967 2014-04-16 Catherine Moore <clm@codesourcery.com>
1969 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
1970 Adjust constraints for microMIPS store patterns.
1972 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
1974 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
1976 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1978 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
1979 (append_use): Run at -O0.
1980 (append_vdef): Likewise.
1981 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
1982 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
1984 2014-04-16 Jakub Jelinek <jakub@redhat.com>
1986 PR tree-optimization/60844
1987 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
1988 (propagate_op_to_single_use, remove_visited_stmt_chain,
1989 linearize_expr, repropagate_negates, reassociate_bb): Use it
1990 instead of gsi_remove.
1992 2014-04-16 Martin Jambor <mjambor@suse.cz>
1994 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
1995 ipa_transforms_to_apply.
1996 (cgraph_function_versioning): Assert that old_node has empty
1997 ipa_transforms_to_apply.
1998 * trans-mem.c (ipa_tm_create_version): Likewise.
1999 * tree-inline.c (tree_function_versioning): Do not duplicate
2000 ipa_transforms_to_apply.
2002 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2005 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
2007 Pass necessary as flags on 64-bit Solaris/x86.
2008 Use lowercase relocs for x86_64-*-*.
2009 * configure: Regenerate.
2011 2014-04-15 Jan Hubicka <jh@suse.cz>
2013 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
2014 (maybe_record_node, likely_target_p): Use it.
2016 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2019 Revert following patch
2021 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2024 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2025 software floating point or no floating point registers, do not
2026 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2027 in GPRs that occurs after we tested for GPRs that would never be
2030 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2031 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2032 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2033 specifically allow DDmode, since that does not use the SPE SIMD
2036 2014-03-21 Mark Wielaard <mjw@redhat.com>
2038 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
2039 as unsigned or int depending on type and value used.
2041 2014-04-15 Richard Biener <rguenther@suse.de>
2043 PR rtl-optimization/56965
2044 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
2045 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
2047 * alias.c (true_dependence_1): Do not call
2048 nonoverlapping_component_refs_p.
2049 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
2050 nonoverlapping_component_refs_p.
2051 (indirect_refs_may_alias_p): Likewise.
2053 2014-04-15 Teresa Johnson <tejohnson@google.com>
2055 * cfg.c (dump_bb_info): Fix flags check.
2056 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
2058 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2060 PR rtl-optimization/60663
2061 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
2064 2014-04-15 Richard Biener <rguenther@suse.de>
2066 * lto-streamer.h (LTO_major_version): Bump to 4.
2068 2014-04-15 Richard Biener <rguenther@suse.de>
2070 * common.opt (lto_partition_model): New enum.
2071 (flto-partition=): Merge separate options with a single with argument,
2072 add -flto-partition=one support.
2073 * flag-types.h (enum lto_partition_model): Declare.
2074 * opts.c (finish_options): Remove duplicate -flto-partition=
2076 * lto-wrapper.c (run_gcc): Adjust.
2078 2014-04-15 Richard Biener <rguenther@suse.de>
2080 * alias.c (ncr_compar): New function.
2081 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2083 2014-04-15 Richard Biener <rguenther@suse.de>
2085 * alias.c (record_component_aliases): Do not walk BINFOs.
2087 2014-04-15 Richard Biener <rguenther@suse.de>
2089 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2090 Add struct function argument and adjust.
2091 (find_func_aliases_for_call): Likewise.
2092 (find_func_aliases): Likewise.
2093 (find_func_clobbers): Likewise.
2094 (intra_create_variable_infos): Likewise.
2095 (compute_points_to_sets): Likewise.
2096 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2098 2014-04-15 Richard Biener <rguenther@suse.de>
2100 * tree.c (iterative_hash_expr): Use enum tree_code_class
2101 to store TREE_CODE_CLASS.
2102 (tree_block): Likewise.
2103 (tree_set_block): Likewise.
2104 * tree.h (fold_build_pointer_plus_loc): Use
2105 convert_to_ptrofftype_loc.
2107 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2110 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2113 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2115 * cfgloop.h (struct loop): Move force_vectorize down.
2116 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2117 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2118 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2119 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2120 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2121 * tree-core.h (enum annot_expr_kind): Add new kind values.
2122 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2123 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2125 * tree.def (ANNOTATE_EXPR): Tweak comment.
2127 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2129 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2132 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2134 * tree.h (TYPE_IDENTIFIER): Declare.
2135 * tree.c (subrange_type_for_debug_p): Use it.
2136 * godump.c (go_format_type): Likewise.
2137 * dwarf2out.c (is_cxx_auto, modified_type_die,
2138 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2139 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2141 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2144 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2146 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2148 * config/i386/i386.c (examine_argument): Return bool. Return true if
2149 parameter should be passed in memory.
2150 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2151 (construct_container): Update calls to examine_argument.
2152 (function_arg_advance_64): Ditto.
2153 (return_in_memory_32): Merge with ix86_return_in_memory.
2154 (return_in_memory_64): Ditto.
2155 (return_in_memory_ms_64): Ditto.
2157 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2159 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2160 * coverage.c (coverage_compute_profile_id): Handle externally visible
2163 2014-04-14 Martin Jambor <mjambor@suse.cz>
2165 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2166 DECL_DISREGARD_INLINE_LIMITS functions.
2168 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2171 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2173 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2176 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2177 optimize_insn_for_speed_p instead of
2178 optimize_function_for_speed_p.
2180 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2182 * doc/invoke.texi (free): Document AArch64.
2184 2014-04-14 Richard Biener <rguenther@suse.de>
2186 PR tree-optimization/60042
2187 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2188 (insert_into_preds_of_block): Do not prevent PHI insertion
2189 for REFERENCE exprs here ...
2190 (eliminate_dom_walker::before_dom_children): ... but prevent
2191 their use here under similar conditions when applied to the
2192 IL after PRE optimizations.
2194 2014-04-14 Richard Biener <rguenther@suse.de>
2196 * passes.def: Move early points-to after early SRA.
2198 2014-04-14 Richard Biener <rguenther@suse.de>
2200 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2201 check for which sign-changes we allow when forwarding
2202 a converted value into a switch.
2204 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2206 * stor-layout.c (place_field): Finalize non-constant offset for the
2209 2014-04-14 Richard Biener <rguenther@suse.de>
2211 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2213 (expand_switch_using_bit_tests_p): Likewise.
2214 (process_switch): Compute and pass on speed_p based on the
2216 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2217 optimize_bb_for_speed_p.
2219 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2221 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2222 * function.h (struct function): Rename has_force_vect_loops into
2223 has_force_vectorize_loops.
2224 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2225 (input_struct_function_base): Likewise.
2226 * lto-streamer-out.c (output_cfg): Likewise.
2227 (output_struct_function_base): Likewise.
2228 * omp-low.c (expand_omp_simd): Likewise.
2229 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2230 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2231 (version_loop_for_if_conversion): Likewise.
2232 (tree_if_conversion): Likewise.
2233 (main_tree_if_conversion): Likewise.
2234 (gate_tree_if_conversion): Likewise.
2235 * tree-inline.c (copy_loops): Likewise.
2236 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2237 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2238 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2239 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2240 * tree-vectorizer.c (vectorize_loops): Likewise.
2241 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2243 2014-04-14 Richard Biener <rguenther@suse.de>
2246 * lto-streamer-out.c (wrap_refs): New function.
2247 (lto_output): Wrap symbol references in global initializes in
2248 type-preserving MEM_REFs.
2250 2014-04-14 Christian Bruel <christian.bruel@st.com>
2252 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2254 2014-04-14 Christian Bruel <christian.bruel@st.com>
2256 * config/sh/sh.md (setmemqi): New expand pattern.
2257 * config/sh/sh.h (CLEAR_RATIO): Define.
2258 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2259 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2261 2014-04-14 Richard Biener <rguenther@suse.de>
2264 * fold-const.c (negate_expr_p): Don't negate directional rounding
2266 (fold_negate_expr): Likewise.
2268 2014-04-14 Richard Biener <rguenther@suse.de>
2270 PR tree-optimization/59817
2271 PR tree-optimization/60453
2272 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2273 recursion to catch all CHRECs in the scalar evolution and restrict
2274 the predicate for the remains appropriately.
2276 2014-04-12 Catherine Moore <clm@codesourcery.com>
2278 * config/mips/constraints.md: Add new register constraint "kb".
2279 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2280 (*movhi_internal): Likewise.
2281 (*movqi_internal): Likewise.
2282 * config/mips/mips.h (M16_STORE_REGS): New register class.
2283 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2284 (REG_CLASS_CONTENTS): Likewise.
2285 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2287 2014-04-11 Tobias Burnus <burnus@net-b.de>
2290 * doc/invoke.texi (-Wformat-signedness): Document it.
2291 (Wformat=2): Mention that this enables -Wformat-signedness.
2293 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2295 * common/config/epiphany/epiphany-common.c
2296 (epiphany_option_optimization_table): Enable section anchors by
2297 default at -O1 or higher.
2298 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2299 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2300 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2301 carries no extra cost.
2302 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2303 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2304 * config/epiphany/predicates.md (memclob_operand): New predicate.
2305 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2306 Use memclob_operand predicate and X constraint for operand 3.
2308 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2310 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2311 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2314 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2316 PR rtl-optimization/60651
2317 * mode-switching.c (optimize_mode_switching): Make sure to emit
2318 sets of a lower numbered entity before sets of a higher numbered
2319 entity to a mode of the same or lower priority.
2320 When creating a seginfo for a basic block that starts with a code
2321 label, move the insertion point past the code label.
2322 (new_seginfo): Document and enforce requirement that
2323 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2324 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2325 * doc/tm.texi: Regenerate.
2327 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2330 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2332 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2334 * BASE-VER: Set to 4.10.0.
2336 2014-04-11 Tobias Burnus <burnus@net-b.de>
2339 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2340 * doc/gcc.texi (Service): Update description in the @menu
2341 * doc/invoke.texi (Option Summary): Remove misplaced and
2344 2014-04-11 Steve Ellcey <sellcey@mips.com>
2345 Jakub Jelinek <jakub@redhat.com>
2348 * expr.c (convert_move): Use emit_store_flag_force instead of
2349 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2352 2014-04-11 Richard Biener <rguenther@suse.de>
2355 * varasm.c (assemble_alias): Avoid endless error reporting
2356 recursion by setting TREE_ASM_WRITTEN.
2358 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2360 * config/s390/s390.md: Add a splitter for NOT rtx.
2362 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2364 PR rtl-optimization/60663
2365 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2367 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2368 Jakub Jelinek <jakub@redhat.com>
2371 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2372 flag from decl_node to node.
2374 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2377 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2378 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2379 ameliorating the cases where it can be.
2381 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2384 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2386 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2387 (loadsync_<mode>): Change mode.
2388 (load_quadpti, store_quadpti): New.
2389 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2390 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2391 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2393 2014-04-09 Cong Hou <congh@google.com>
2396 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2399 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2401 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2402 instead of vnor to exploit possible fusion opportunity in the
2404 (altivec_expand_vec_perm_const_le): Likewise.
2406 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2408 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2409 (loadsync_<mode>): Change mode.
2410 (load_quadpti, store_quadpti): New.
2411 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2412 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2414 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2417 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2418 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2419 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2421 2014-04-08 Richard Biener <rguenther@suse.de>
2424 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2425 a 64bit widest int print double-int similar to on HWI64 hosts.
2427 2014-04-08 Richard Biener <rguenther@suse.de>
2429 PR tree-optimization/60785
2430 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2431 default defs properly.
2433 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2435 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2436 (Weffc++): Likewise.
2438 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2440 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2441 set completep to false rather than true.
2443 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2446 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2447 ARM_TARGET2_DWARF_FORMAT.
2449 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2452 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2453 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2456 2014-04-07 Richard Biener <rguenther@suse.de>
2458 PR tree-optimization/60766
2459 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2460 (may_eliminate_iv): Convert cand_value_at result to desired type.
2462 2014-04-07 Jason Merrill <jason@redhat.com>
2465 * common.opt (-fno-gnu-unique): Add.
2466 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2468 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2470 * haifa-sched.c: Fix outdated function reference and minor
2471 grammar errors in introductory comment.
2473 2014-04-07 Richard Biener <rguenther@suse.de>
2476 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2478 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2480 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2483 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2484 size accounting for thunks.
2485 (pa_asm_output_mi_thunk): Use final_start_function() and
2486 final_end_function() to output function start and end directives.
2488 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2490 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2491 device specific ISA/ feature information. Remove short_sp and
2492 errata_skip ds. Add avr_device_specific_features enum to have device
2494 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2495 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2496 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2497 updated device specific info.
2498 * config/avr/avr-mcus.def: Merge device specific details to
2499 dev_attribute field.
2500 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2502 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2503 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2504 assembler if RMW isa supported by current device.
2505 * config/avr/genmultilib.awk: Update as device info structure changed.
2506 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2508 2014-04-04 Cong Hou <congh@google.com>
2510 PR tree-optimization/60656
2511 * tree-vect-stmts.c (supportable_widening_operation):
2512 Fix a bug that elements in a vector with vect_used_by_reduction
2513 property are incorrectly reordered when the operation on it is not
2514 consistant with the one in reduction operation.
2516 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2518 PR rtl-optimization/60155
2519 * gcse.c (record_set_data): New function.
2520 (single_set_gcse): New function.
2521 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2522 (hoist_code): Likewise.
2523 (get_pressure_class_and_nregs): Likewise.
2525 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2527 * explow.c (probe_stack_range): Emit a final optimization blockage.
2529 2014-04-04 Anthony Green <green@moxielogic.com>
2531 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2534 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2537 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2538 flags are set only during streaming.
2539 * ipa.c (process_references, walk_polymorphic_call_targets,
2540 symtab_remove_unreachable_nodes): Drop bodies of always inline
2541 after early inlining.
2542 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2544 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2545 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2548 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2551 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2554 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2556 (cortex_a53_fdivd): Likewise.
2558 2014-04-04 Martin Jambor <mjambor@suse.cz>
2561 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2563 * cgraph.c (clone_of_p): Also return true if thunks match.
2564 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2565 cgraph_function_or_thunk_node and an obsolete comment.
2566 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2568 (build_function_decl_skip_args): Likewise.
2569 (set_new_clone_decl_and_node_flags): New function.
2570 (duplicate_thunk_for_node): Likewise.
2571 (redirect_edge_duplicating_thunks): Likewise.
2572 (cgraph_clone_node): New parameter args_to_skip, pass it to
2573 redirect_edge_duplicating_thunks which is called instead of
2574 cgraph_redirect_edge_callee.
2575 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2576 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2578 2014-04-04 Jeff Law <law@redhat.com>
2581 * config/arm/predicates.md (const_int_I_operand): New predicate.
2582 (const_int_M_operand): Similarly.
2583 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2585 (insv_t2, extv_reg, extzv_t2): Likewise.
2586 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2587 (pop_multiple_with_writeback_and_return): Likewise.
2588 (vfp_pop_multiple_with_writeback): Likewise
2590 2014-04-04 Richard Biener <rguenther@suse.de>
2593 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2594 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2596 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2597 * gimplify.c (gimple_add_tmp_var_fn): New function.
2598 * gimple-expr.h (create_tmp_reg_fn): Declare.
2599 * gimple-expr.c (create_tmp_reg_fn): New function.
2600 * gimple-low.c (record_vars_into): Don't change cfun.
2601 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2602 code generation without cfun.
2604 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2607 * Makefile.in (install-driver): Fix shell scripting.
2609 2014-04-03 Cong Hou <congh@google.com>
2611 PR tree-optimization/60505
2612 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2613 threshold of number of iterations below which no vectorization
2615 * tree-vect-loop.c (new_loop_vec_info):
2616 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2617 * tree-vect-loop.c (vect_analyze_loop_operations):
2618 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2619 * tree-vect-loop.c (vect_transform_loop):
2620 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2621 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2622 of iterations of the loop and see if we should build the epilogue.
2624 2014-04-03 Richard Biener <rguenther@suse.de>
2626 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2627 (streamer_tree_cache_create): Adjust.
2628 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2629 to allow optional nodes array.
2630 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2631 (streamer_tree_cache_append): Likewise.
2632 (streamer_tree_cache_create): Create nodes array optionally
2633 as specified by parameter.
2634 * lto-streamer-out.c (create_output_block): Avoid maintaining
2635 the node array in the writer cache.
2636 (DFS_write_tree): Remove assertion.
2637 (produce_asm_for_decls): Free the out decl state hash table early.
2638 * lto-streamer-in.c (lto_data_in_create): Adjust for
2639 streamer_tree_cache_create prototype change.
2641 2014-04-03 Richard Biener <rguenther@suse.de>
2643 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2644 set TREE_CHAIN to NULL_TREE.
2646 2014-04-03 Richard Biener <rguenther@suse.de>
2648 PR tree-optimization/60740
2649 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2650 over all GIMPLE_COND operands.
2652 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2654 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2655 (Weffc++): Remove Scott's numbering, merge lists and reference
2658 2014-04-03 Nick Clifton <nickc@redhat.com>
2660 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2663 2014-04-03 Martin Jambor <mjambor@suse.cz>
2665 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2666 mention gcc_unreachable before failing.
2667 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2670 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
2673 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2674 inconsistent code and instead mark the context inconsistent.
2675 (possible_polymorphic_call_targets): For inconsistent contexts
2676 return empty complete list.
2678 2014-04-02 Anthony Green <green@moxielogic.com>
2680 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2681 (extendqisi2, extendhisi2): Define.
2682 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2683 (WCHAR_TYPE): Change to unsigned int.
2685 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2687 PR tree-optimization/60733
2688 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2689 insertion point for PHI candidates to be the end of the feeding
2690 block for the PHI argument.
2692 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2694 PR rtl-optimization/60650
2695 * lra-constraints.c (process_alt_operands): Decrease reject for
2696 earlyclobber matching.
2698 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2700 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2702 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2704 * config/spu/spu.c (pad_bb): Do not crash when the last
2705 insn is CODE_FOR_blockage.
2707 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2709 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2710 lies outside the target mode.
2712 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2715 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2716 software floating point or no floating point registers, do not
2717 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2718 in GPRs that occurs after we tested for GPRs that would never be
2721 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2722 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2723 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2724 specifically allow DDmode, since that does not use the SPE SIMD
2727 2014-04-02 Richard Biener <rguenther@suse.de>
2730 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2731 MODE_INTs. Properly use negv_optab.
2732 (expand_abs): Likewise.
2734 2014-04-02 Richard Biener <rguenther@suse.de>
2737 * Makefile.in (install-driver): Guard extra installs with special
2740 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
2742 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2745 2014-04-01 Richard Henderson <rth@redhat.com>
2748 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
2749 alternative enabled before register allocation.
2751 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
2753 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
2754 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
2756 (nios2_large_got_address): Remove unneeded 'sym' parameter.
2757 (nios2_got_address): Update nios2_large_got_address call site.
2758 (nios2_delegitimize_address): New function.
2759 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
2760 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
2761 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
2763 2014-04-01 Martin Husemann <martin@duskware.de>
2765 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
2768 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
2770 PR rtl-optimization/60604
2771 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
2772 check from register_operand.
2773 (register_operand): Redefine in terms of general_operand.
2774 (nonmemory_operand): Use register_operand for the non-constant cases.
2776 2014-04-01 Richard Biener <rguenther@suse.de>
2778 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
2780 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
2782 * doc/invoke.texi (mapp-regs): Clarify.
2784 2014-03-31 Ulrich Drepper <drepper@gmail.com>
2786 * config/i386/avx512fintrin.h (__v32hi): Define type.
2787 (__v64qi): Likewise.
2788 (_mm512_set1_epi8): Define.
2789 (_mm512_set1_epi16): Define.
2790 (_mm512_set4_epi32): Define.
2791 (_mm512_set4_epi64): Define.
2792 (_mm512_set4_pd): Define.
2793 (_mm512_set4_ps): Define.
2794 (_mm512_setr4_epi64): Define.
2795 (_mm512_setr4_epi32): Define.
2796 (_mm512_setr4_pd): Define.
2797 (_mm512_setr4_ps): Define.
2798 (_mm512_setzero_epi32): Define.
2800 2014-03-31 Martin Jambor <mjambor@suse.cz>
2803 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
2804 callsite_arguments_match_p. Updated all callers. Also check types of
2805 corresponding formal parameters and actual arguments.
2806 (not_all_callers_have_enough_arguments_p) Renamed to
2807 some_callers_have_mismatched_arguments_p.
2809 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
2811 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
2813 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
2816 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
2819 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
2821 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
2822 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
2824 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
2825 Use FMAMODE_NOVF512 mode iterator.
2826 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
2827 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
2828 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
2829 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
2831 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
2832 Use VF_128_256 mode iterator.
2833 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
2836 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2838 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
2839 static chain if needed.
2841 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2844 * lra-constraints.c (index_part_to_reg): New.
2845 (process_address): Use it.
2847 2014-03-27 Jeff Law <law@redhat.com>
2848 Jakub Jelinek <jakub@redhat.com>
2851 * expr.c (do_tablejump): Use simplify_gen_binary rather than
2852 gen_rtx_{PLUS,MULT} to build up the address expression.
2854 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
2855 creating non-canonical RTL.
2857 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2860 * ipa-inline.c (want_inline_small_function_p): Short circuit large
2861 functions; reorganize to make cheap checks first.
2862 (inline_small_functions): Do not estimate growth when dumping;
2864 * ipa-inline.h (inline_summary): Add min_size.
2865 (growth_likely_positive): New function.
2866 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
2867 (set_cond_stmt_execution_predicate): Cleanup.
2868 (estimate_edge_size_and_time): Compute min_size.
2869 (estimate_calls_size_and_time): Likewise.
2870 (estimate_node_size_and_time): Likewise.
2871 (inline_update_overall_summary): Update min_size.
2872 (do_estimate_edge_time): Likewise.
2873 (do_estimate_edge_size): Update.
2874 (do_estimate_edge_hints): Update.
2875 (growth_likely_positive): New function.
2877 2014-03-28 Jakub Jelinek <jakub@redhat.com>
2880 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
2881 also if addr has VOIDmode.
2883 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2885 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
2886 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
2888 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
2889 instructions as well as AdvancedSIMD loads.
2891 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2893 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
2894 Use crypto_aese type.
2895 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
2896 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
2897 crypto_aese, crypto_aesmc. Move to types.md.
2898 * config/arm/types.md (crypto_aes): Split into crypto_aese,
2900 * config/arm/iterators.md (crypto_type): Likewise.
2902 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2904 * cgraph.c: Include expr.h and tree-dfa.h.
2905 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
2908 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2911 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
2912 regs from checking multi-reg pseudos.
2914 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2916 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
2918 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2920 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
2921 if it would clobber the stack pointer, even temporarily.
2923 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
2925 * mode-switching.c: Make small adjustments to the top comment.
2927 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
2929 * config/rs6000/constraints.md (wD constraint): New constraint to
2930 match the constant integer to get the top DImode/DFmode out of a
2931 vector in a VSX register.
2933 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
2934 match the constant integer to get the top DImode/DFmode out of a
2935 vector in a VSX register.
2937 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
2940 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
2943 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
2944 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
2946 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
2947 Optimize vec_extract of 64-bit values, where the value being
2948 extracted is in the top word, where we can use scalar
2949 instructions. Add direct move and store support. Combine the big
2950 endian/little endian vector select load support into a single insn.
2951 (vsx_extract_<mode>_internal1): Likewise.
2952 (vsx_extract_<mode>_internal2): Likewise.
2953 (vsx_extract_<mode>_load): Likewise.
2954 (vsx_extract_<mode>_store): Likewise.
2955 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
2956 combined into vsx_extract_<mode>_load.
2957 (vsx_extract_<mode>_one_le): Likewise.
2959 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
2960 define the top 64-bit vector element.
2962 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
2965 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2966 Document vec_vbpermq builtin.
2969 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
2970 enable use of xxsldwi and xxpermdi builtin functions.
2971 (vec_xxpermdi): Likewise.
2973 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2974 Document use of vec_xxsldwi and vec_xxpermdi builtins.
2976 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
2978 PR rtl-optimization/60650
2979 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
2981 (find_spills_for): New.
2982 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
2983 Spill all pseudos on the second iteration.
2985 2014-03-27 Marek Polacek <polacek@redhat.com>
2988 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
2991 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2993 * config/s390/s390.c (s390_can_use_return_insn): Check for
2994 call-saved FPRs on 31 bit.
2996 2014-03-27 Jakub Jelinek <jakub@redhat.com>
2999 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
3000 if they need regimplification, just drop them instead of
3001 calling gimple_regimplify_operands on them.
3003 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
3006 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
3007 (aarch64_frame_pointer_required): Adjust logic.
3008 (aarch64_can_eliminate): Adjust logic.
3009 (aarch64_override_options_after_change): Adjust logic.
3011 2014-03-27 Dehao Chen <dehao@google.com>
3013 * ipa-inline.c (early_inliner): Update node's inline info.
3015 2014-03-26 Dehao Chen <dehao@google.com>
3017 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
3018 compiler inserted conditional jumps for NAN float check.
3020 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3022 * ubsan.h (ubsan_create_data): Change second argument's type
3023 to const location_t *.
3024 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
3026 (ubsan_create_data): Change second argument to const location_t *PLOC.
3027 Create Loc field whenever PLOC is non-NULL.
3028 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
3029 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
3033 * real.c (real_to_integer2): Change type of low to UHWI.
3035 2014-03-26 Tobias Burnus <burnus@net-b.de>
3037 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
3038 (CILK_SELF_SPECS): New define.
3039 (driver_self_specs): Use it.
3041 2014-03-26 Richard Biener <rguenther@suse.de>
3043 * tree-pretty-print.c (percent_K_format): Implement special
3044 case for LTO and its stripped down BLOCK tree.
3046 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3049 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
3051 * tree-vrp.c (simplify_internal_call_using_ranges): If only
3052 one range is range_int_cst_p, but not both, at least optimize
3053 addition/subtraction of 0 and multiplication by 0 or 1.
3054 * gimple-fold.c (gimple_fold_call): Fold
3055 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
3056 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
3057 INTEGER_CSTs, try to fold at least x * 0 and y - y.
3059 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
3061 PR rtl-optimization/60452
3062 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
3063 <case REG>: Return 1 for invalid offsets from the frame pointer.
3065 2014-03-26 Marek Polacek <polacek@redhat.com>
3068 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3071 2014-03-26 Marek Polacek <polacek@redhat.com>
3074 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3077 2014-03-26 Marek Polacek <polacek@redhat.com>
3080 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3081 multiplication in unsigned type.
3083 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3085 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3087 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3089 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3091 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3094 * cif-code.def (UNREACHABLE) New code.
3095 * ipa-inline.c (inline_small_functions): Skip edges to
3096 __builtlin_unreachable.
3097 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3098 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3099 predicate to __bulitin_unreachable.
3100 (set_cond_stmt_execution_predicate): Fix issue when
3101 invert_tree_comparison returns ERROR_MARK.
3102 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3103 propagate to inline clones.
3104 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3106 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3107 * cgraphclones.c (cgraph_clone_node): If call destination is already
3108 ureachable, do not redirect it back.
3109 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3112 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3114 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3115 Do not modify inline clones.
3117 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3119 * config/i386/i386.md (general_sext_operand): New mode attr.
3120 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3121 don't generate (sign_extend (const_int)).
3122 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3123 operands[2]. Use We constraint instead of <i> and
3124 <general_sext_operand> predicate instead of <general_operand>.
3125 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3126 * config/i386/constraints.md (We): New constraint.
3127 * config/i386/predicates.md (x86_64_sext_operand,
3128 sext_operand): New predicates.
3130 2014-03-25 Martin Jambor <mjambor@suse.cz>
3133 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3134 inconsistent devirtualizations to __builtin_unreachable.
3136 2014-03-25 Marek Polacek <polacek@redhat.com>
3139 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3141 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3143 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3144 order of elements for big-endian.
3146 2014-03-25 Richard Biener <rguenther@suse.de>
3149 * gimplify-me.c (gimple_regimplify_operands): Update the
3152 2014-03-25 Martin Jambor <mjambor@suse.cz>
3155 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3156 (lto_output_varpool_node): Likewise.
3157 (input_overwrite_node): Likewise.
3158 (input_varpool_node): Likewise.
3160 2014-03-25 Richard Biener <rguenther@suse.de>
3162 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3163 (run_gcc): Likewise.
3165 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3167 * combine.c (simplify_compare_const): Add MODE argument.
3168 Handle mode_width 0 as very large mode_width.
3169 (try_combine, simplify_comparison): Adjust callers.
3171 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3172 type to avoid signed integer overflow.
3173 * explow.c (plus_constant): Likewise.
3175 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3177 * doc/generic.texi: Correct typos.
3179 2014-03-24 Tobias Burnus <burnus@net-b.de>
3181 * doc/invoke.texi (-flto): Expand section about
3182 using static libraries with LTO.
3184 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3186 PR rtl-optimization/60501
3187 * optabs.def (addptr3_optab): New optab.
3188 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3189 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3190 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3192 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3194 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3196 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3198 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3201 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3202 (_mm256_undefined_ps): Define.
3203 (_mm256_undefined_pd): Define.
3204 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3205 (_mm_undefined_pd): Define.
3206 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3207 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3208 (_mm512_undefined_ps): Define.
3209 (_mm512_undefined_pd): Define.
3210 Use _mm*_undefined_*.
3211 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3213 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3215 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3216 (lshr_simd): DI mode added.
3217 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3218 (aarch64_ushr_simddi): Likewise.
3219 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3220 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3221 (vshrd_n_u64): Likewise.
3223 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3225 * Makefile.in (s-macro_list): Depend on cc1.
3227 2014-03-23 Teresa Johnson <tejohnson@google.com>
3229 * ipa-utils.c (ipa_print_order): Use specified dump file.
3231 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3233 PR rtl-optimization/60601
3234 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3236 * gcc.c (eval_spec_function): Initialize save_growing_value.
3238 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3241 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3242 code == MINUS_EXPR, never swap op0 with op1.
3244 * toplev.c (init_local_tick): Avoid signed integer multiplication
3246 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3247 shift by first operand's bitsize.
3249 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3252 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3254 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3255 TARGET_ISA_64BIT_P(x).
3257 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3259 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3260 pattern for vector nor instead of subtract from splat(-1).
3261 (altivec_expand_vec_perm_const_le): Likewise.
3263 2014-03-21 Richard Henderson <rth@twiddle.net>
3266 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3267 related insns after epilogue_completed.
3269 2014-03-21 Martin Jambor <mjambor@suse.cz>
3272 * cgraph.h (symtab_node): New flag body_removed.
3273 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3274 when removing bodies.
3275 * symtab.c (dump_symtab_base): Dump body_removed flag.
3276 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3277 had their bodies removed.
3279 2014-03-21 Martin Jambor <mjambor@suse.cz>
3282 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3285 2014-03-21 Richard Biener <rguenther@suse.de>
3287 PR tree-optimization/60577
3288 * tree-core.h (struct tree_base): Document nothrow_flag use
3289 in VAR_DECL_NONALIASED.
3290 * tree.h (VAR_DECL_NONALIASED): New.
3291 (may_be_aliased): Adjust.
3292 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3294 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3296 * expr.c (expand_expr_real_1): Remove outdated comment.
3298 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3301 * ira.c (adjust_cleared_regs): Call copy_rtx on
3302 *reg_equiv[REGNO (loc)].src_p before passing it to
3303 simplify_replace_fn_rtx.
3306 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3307 into CONST, put pic register as first operand of PLUS. Use
3308 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3310 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3312 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3314 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3316 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3317 around for store forwarding issue in the FPU on the UT699.
3318 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3319 loads and operations if -mfix-ut699 is specified.
3320 (divtf3_hq): Tweak attribute.
3321 (sqrttf2_hq): Likewise.
3323 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3325 * calls.c (store_one_arg): Remove incorrect const qualification on the
3326 type of the temporary.
3327 * cfgexpand.c (expand_return): Likewise.
3328 * expr.c (expand_constructor): Likewise.
3329 (expand_expr_real_1): Likewise.
3331 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3333 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3336 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3339 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3341 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3343 * config/arm/aarch-common-protos.h
3344 (alu_cost_table): Fix spelling of "extend".
3345 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3347 2014-03-19 Richard Biener <rguenther@suse.de>
3350 * tree-core.h (tree_type_common): Re-order pointer members
3351 to reduce recursion depth during GC walks.
3353 2014-03-19 Marek Polacek <polacek@redhat.com>
3356 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3357 before accessing it.
3359 2014-03-19 Richard Biener <rguenther@suse.de>
3362 * lto-streamer-in.c (input_function): In WPA stage do not drop
3365 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3367 PR tree-optimization/60559
3368 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3369 with build_zero_cst assignment.
3371 2014-03-18 Kai Tietz <ktietz@redhat.com>
3373 PR rtl-optimization/56356
3374 * sdbout.c (sdbout_parms): Verify that parms'
3375 incoming argument is valid.
3376 (sdbout_reg_parms): Likewise.
3378 2014-03-18 Richard Henderson <rth@redhat.com>
3381 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3382 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3383 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3385 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3387 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3388 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3389 Italicize plugin event names in description. Explain that
3390 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3391 Remind that no GCC functions should be called after PLUGIN_FINISH.
3392 Explain what pragmas with expansion are.
3394 2014-03-18 Martin Liska <mliska@suse.cz>
3396 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3397 gimple call statement is update.
3398 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3399 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3401 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3404 * ubsan.c (ubsan_instrument_unreachable): Call
3405 initialize_sanitizer_builtins.
3406 (ubsan_pass): Likewise.
3409 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3410 varpool_finalize_decl instead of rest_of_decl_compilation.
3412 2014-03-18 Richard Biener <rguenther@suse.de>
3414 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3415 by using bitmap_and_compl instead of bitmap_and_compl_into.
3416 (df_rd_transfer_function): Likewise.
3418 2014-03-18 Richard Biener <rguenther@suse.de>
3420 * doc/lto.texi (fresolution): Fix typo.
3422 2014-03-18 Richard Biener <rguenther@suse.de>
3424 * doc/invoke.texi (flto): Update for changes in 4.9.
3426 2014-03-18 Richard Biener <rguenther@suse.de>
3428 * doc/loop.texi: Remove section on the removed lambda framework.
3429 Update loop docs with recent changes in preserving loop structure.
3431 2014-03-18 Richard Biener <rguenther@suse.de>
3433 * doc/lto.texi (-fresolution): Document.
3435 2014-03-18 Richard Biener <rguenther@suse.de>
3437 * doc/contrib.texi: Adjust my name.
3439 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3442 * internal-fn.c: Include diagnostic-core.h.
3443 (expand_BUILTIN_EXPECT): New function.
3444 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3445 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3446 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3447 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3449 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3450 Revert 3 argument __builtin_expect code.
3451 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3452 * gimple-fold.c (gimple_fold_call): Likewise.
3453 * tree.h (fold_builtin_expect): New prototype.
3454 * builtins.c (build_builtin_expect_predicate): Add predictor
3455 argument, if non-NULL, create 3 argument __builtin_expect.
3456 (fold_builtin_expect): No longer static. Add ARG2 argument,
3457 pass it through to build_builtin_expect_predicate.
3458 (fold_builtin_2): Adjust caller.
3459 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3460 * internal-fn.def (BUILTIN_EXPECT): New.
3462 2014-03-18 Tobias Burnus <burnus@net-b.de>
3465 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3466 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3467 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3469 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3472 * predict.c (combine_predictions_for_bb): Fix up formatting.
3473 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3474 fill what it points to if non-NULL.
3475 (tree_predict_by_opcode): Adjust caller, use the predictor.
3476 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3478 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3480 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3481 proper constant for the store mode.
3483 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3485 * symtab.c (change_decl_assembler_name): Fix transparent alias
3488 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3490 * config/aarch64/aarch64.c: Correct the comments about the
3491 aarch64 stack layout.
3493 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3495 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3496 check for GF_OMP_FOR_KIND_FOR.
3498 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3500 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3501 ymm and zmm register names.
3503 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3506 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3507 note creation for the 2010-08-31 changes.
3509 2014-03-17 Marek Polacek <polacek@redhat.com>
3512 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3513 as -fno-tree-loop-vectorize.
3514 (expand_omp_simd): Likewise.
3516 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3518 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3519 (eligible_for_call_delay): New prototype.
3520 * config/sparc/sparc.c (tls_call_delay): Rename into...
3521 (eligible_for_call_delay): ...this. Return false if the instruction
3522 cannot be put in the delay slot of a branch.
3523 (eligible_for_restore_insn): Simplify.
3524 (eligible_for_return_delay): Return false if the instruction cannot be
3525 put in the delay slot of a branch and simplify.
3526 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3527 put in the delay slot of a branch.
3528 * config/sparc/sparc.md (fix_ut699): New attribute.
3529 (tls_call_delay): Delete.
3530 (in_call_delay): Reimplement.
3531 (eligible_for_sibcall_delay): Rename into...
3532 (in_sibcall_delay): ...this.
3533 (eligible_for_return_delay): Rename into...
3534 (in_return_delay): ...this.
3535 (in_branch_delay): Reimplement.
3536 (in_uncond_branch_delay): Delete.
3537 (in_annul_branch_delay): Delete.
3539 2014-03-14 Richard Henderson <rth@redhat.com>
3542 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3543 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3544 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3545 (floathi splitters): Remove.
3546 (float<SWI48x>xf2): New pattern.
3547 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3548 code that tried to handle DImode for 32-bit, but which was excluded
3549 by the pattern's condition. Drop allocation of stack temporary.
3550 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3551 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3552 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3553 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3554 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3555 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3556 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3557 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3558 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3559 (*float<SWI48x><X87MODEF>2_i387): Remove.
3560 (all float _with_temp splitters): Remove.
3561 (*float<SWI48x><MODEF>2_i387): New pattern.
3562 (*float<SWI48><MODEF>2_sse): New pattern.
3563 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3564 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3566 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3567 Marek Polacek <polacek@redhat.com>
3570 * common.opt (dump_base_name_prefixed): New Variable.
3571 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3572 if x_dump_base_name_prefixed is already set, set it at the end.
3574 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3576 PR rtl-optimization/60508
3577 * lra-constraints.c (get_reload_reg): Add new parameter
3579 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3580 Pass the new parameter values.
3582 2014-03-14 Richard Biener <rguenther@suse.de>
3584 * common.opt: Revert unintented changes from r205065.
3587 2014-03-14 Richard Biener <rguenther@suse.de>
3590 * cfghooks.c (split_block): Properly adjust all loops the
3591 block was a latch of.
3593 2014-03-14 Martin Jambor <mjambor@suse.cz>
3596 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3599 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3602 * config/avr/avr.c (avr_set_current_function): Pass function name
3603 through default_strip_name_encoding before sanity checking instead
3604 of skipping the first char of the assembler name.
3606 2014-03-13 Richard Henderson <rth@redhat.com>
3609 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3610 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3611 * config/i386/i386-protos.h: Likewise.
3612 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3613 in the expander instead of a splitter.
3614 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3615 any possibility of requiring a memory.
3616 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3617 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3618 (fp branch splitters): Update for ix86_split_fp_branch.
3619 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3620 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3621 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3622 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3623 (*fop_<MODEF>_3_i387): Likewise.
3624 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3625 (splitters for the fop_* register patterns): Remove.
3626 (fscalexf4_i387): Rename from *fscalexf4_i387.
3627 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3629 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3631 PR tree-optimization/59779
3632 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3633 type for bitsize and maxsize instead of HOST_WIDE_INT.
3635 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3637 PR rtl-optimization/57320
3638 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3639 the CFG after thread_prologue_and_epilogue_insns.
3641 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3643 PR rtl-optimization/57189
3644 * lra-constraints.c (process_alt_operands): Disfavor spilling
3647 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3649 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3651 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3653 PR tree-optimization/59025
3655 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3656 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3658 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3661 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3662 calls of avr_out_plus_1.
3664 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3666 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3667 BB's single pred and update the father loop's latch info later.
3669 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3671 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3675 (VEC_base): Likewise.
3676 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3677 registers, we need to swap double words in little endian mode.
3679 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3680 to be a container mode for 128-bit integer operations added in ISA
3681 2.07. Unlike TImode and PTImode, the preferred register set is
3682 the Altivec/VMX registers for the 128-bit operations.
3684 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3686 (rs6000_split_128bit_ok_p): Likewise.
3688 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3689 macros for creating ISA 2.07 normal and overloaded builtin
3690 functions with 3 arguments.
3691 (BU_P8V_OVERLOAD_3): Likewise.
3692 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3693 for use as overloaded functions.
3694 (VPERM_1TI_UNS): Likewise.
3695 (VSEL_1TI): Likewise.
3696 (VSEL_1TI_UNS): Likewise.
3697 (ST_INTERNAL_1ti): Likewise.
3698 (LD_INTERNAL_1ti): Likewise.
3699 (XXSEL_1TI): Likewise.
3700 (XXSEL_1TI_UNS): Likewise.
3701 (VPERM_1TI): Likewise.
3702 (VPERM_1TI_UNS): Likewise.
3703 (XXPERMDI_1TI): Likewise.
3704 (SET_1TI): Likewise.
3705 (LXVD2X_V1TI): Likewise.
3706 (STXVD2X_V1TI): Likewise.
3707 (VEC_INIT_V1TI): Likewise.
3708 (VEC_SET_V1TI): Likewise.
3709 (VEC_EXT_V1TI): Likewise.
3710 (EQV_V1TI): Likewise.
3711 (NAND_V1TI): Likewise.
3712 (ORC_V1TI): Likewise.
3713 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3714 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3716 (VADDUQM): Likewise.
3717 (VSUBCUQ): Likewise.
3718 (VADDEUQM): Likewise.
3719 (VADDECUQ): Likewise.
3720 (VSUBEUQM): Likewise.
3721 (VSUBECUQ): Likewise.
3723 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3724 __int128_t and __uint128_t types.
3725 (__uint128_type): Likewise.
3726 (altivec_categorize_keyword): Add support for vector __int128_t,
3727 vector __uint128_t, vector __int128, and vector unsigned __int128
3728 as a container type for TImode operations that need to be done in
3729 VSX/Altivec registers.
3730 (rs6000_macro_to_expand): Likewise.
3731 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3732 to support 128-bit integer instructions vaddcuq, vadduqm,
3733 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3734 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
3736 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
3737 for V1TImode, and set up preferences to use VSX/Altivec registers.
3738 Setup VSX reload handlers.
3739 (rs6000_debug_reg_global): Likewise.
3740 (rs6000_init_hard_regno_mode_ok): Likewise.
3741 (rs6000_preferred_simd_mode): Likewise.
3742 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
3743 (easy_altivec_constant): Likewise.
3744 (output_vec_const_move): Likewise.
3745 (rs6000_expand_vector_set): Convert V1TImode set and extract to
3747 (rs6000_expand_vector_extract): Likewise.
3748 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
3750 (rs6000_const_vec): Add support for V1TImode.
3751 (rs6000_emit_le_vsx_load): Swap double words when loading or
3752 storing TImode/V1TImode.
3753 (rs6000_emit_le_vsx_store): Likewise.
3754 (rs6000_emit_le_vsx_move): Likewise.
3755 (rs6000_emit_move): Add support for V1TImode.
3756 (altivec_expand_ld_builtin): Likewise.
3757 (altivec_expand_st_builtin): Likewise.
3758 (altivec_expand_vec_init_builtin): Likewise.
3759 (altivec_expand_builtin): Likewise.
3760 (rs6000_init_builtins): Add support for V1TImode type. Add
3761 support for ISA 2.07 128-bit integer builtins. Define type names
3762 for the VSX/Altivec vector types.
3763 (altivec_init_builtins): Add support for overloaded vector
3764 functions with V1TImode type.
3765 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
3766 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
3768 (rs6000_split_128bit_ok_p): Likewise.
3769 (rs6000_handle_altivec_attribute): Create V1TImode from vector
3770 __int128_t and vector __uint128_t.
3772 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
3773 and mode attributes.
3780 (VS_scalar): Likewise.
3781 (VS_double): Likewise.
3782 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
3784 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
3785 we support the ISA 2.07 128-bit integer arithmetic instructions.
3786 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
3787 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
3788 and TImode types for use with the builtin functions.
3789 (V1TI_type_node): Likewise.
3790 (unsigned_V1TI_type_node): Likewise.
3791 (intTI_type_internal_node): Likewise.
3792 (uintTI_type_internal_node): Likewise.
3794 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
3795 128-bit builtin functions.
3796 (UNSPEC_VADDEUQM): Likewise.
3797 (UNSPEC_VADDECUQ): Likewise.
3798 (UNSPEC_VSUBCUQ): Likewise.
3799 (UNSPEC_VSUBEUQM): Likewise.
3800 (UNSPEC_VSUBECUQ): Likewise.
3801 (VM): Add V1TImode to vector mode iterators.
3803 (VI_unit): Likewise.
3804 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
3805 (altivec_vaddcuq): Likewise.
3806 (altivec_vsubuqm): Likewise.
3807 (altivec_vsubcuq): Likewise.
3808 (altivec_vaddeuqm): Likewise.
3809 (altivec_vaddecuq): Likewise.
3810 (altivec_vsubeuqm): Likewise.
3811 (altivec_vsubecuq): Likewise.
3813 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
3815 (BOOL_128): Likewise.
3816 (BOOL_REGS_OUTPUT): Likewise.
3817 (BOOL_REGS_OP1): Likewise.
3818 (BOOL_REGS_OP2): Likewise.
3819 (BOOL_REGS_UNARY): Likewise.
3820 (BOOL_REGS_AND_CR0): Likewise.
3822 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
3823 128-bit integer builtin support.
3824 (vec_vadduqm): Likewise.
3825 (vec_vaddecuq): Likewise.
3826 (vec_vaddeuqm): Likewise.
3827 (vec_vsubecuq): Likewise.
3828 (vec_vsubeuqm): Likewise.
3829 (vec_vsubcuq): Likewise.
3830 (vec_vsubuqm): Likewise.
3832 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3833 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
3834 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
3835 128-bit integer add/subtract to ISA 2.07.
3837 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
3839 * config/arc/arc.c (arc_predicate_delay_insns):
3840 Fix third argument passed to conditionalize_nonjump.
3842 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
3844 * config/aarch64/aarch64-builtins.c
3845 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
3846 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
3847 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
3848 instead of __builtin_lfloor.
3849 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
3851 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3853 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
3854 (tree_ssa_ifcombine_bb_1): New function.
3855 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
3856 is an empty forwarder block to then_bb or vice versa and then_bb
3857 and else_bb are effectively swapped.
3859 2014-03-12 Christian Bruel <christian.bruel@st.com>
3862 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
3863 REG_CFA_DEF_CFA note.
3864 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
3865 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
3867 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
3869 PR tree-optimization/60454
3870 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
3872 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3874 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
3875 Do not define target_cpu_default2 to generic.
3876 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
3877 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
3878 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
3880 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3881 Marc Glisse <marc.glisse@inria.fr>
3883 PR tree-optimization/60502
3884 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
3885 instead of build_low_bits_mask.
3887 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3890 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
3891 if there are multiple uses, but op doesn't live on E edge.
3892 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
3893 clobber stmts before __builtin_unreachable.
3895 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
3897 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
3898 hard_frame_pointer_rtx.
3899 * cse.c (cse_insn): Remove volatile check.
3900 * cselib.c (cselib_process_insn): Likewise.
3901 * dse.c (scan_insn): Likewise.
3903 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3905 * config/arc/arc.c (conditionalize_nonjump): New function,
3907 (arc_ifcvt): ... this.
3908 (arc_predicate_delay_insns): Use it.
3910 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3912 * config/arc/predicates.md (extend_operand): During/after reload,
3913 allow const_int_operand.
3914 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
3915 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
3916 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
3918 (umulsi3_highpart_i): Likewise.
3920 2014-03-11 Richard Biener <rguenther@suse.de>
3922 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
3923 Add asserts to guard possible wrong-code bugs.
3925 2014-03-11 Richard Biener <rguenther@suse.de>
3927 PR tree-optimization/60429
3928 PR tree-optimization/60485
3929 * tree-ssa-structalias.c (set_union_with_increment): Properly
3930 take into account all fields that overlap the shifted vars.
3931 (do_sd_constraint): Likewise.
3932 (do_ds_constraint): Likewise.
3933 (get_constraint_for_ptr_offset): Likewise.
3935 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
3937 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
3938 (nios2_compute_frame_layout):
3939 Add calculation of cfun->machine->fp_save_offset.
3940 (nios2_expand_prologue): Correct setting of frame pointer register
3942 (nios2_expand_epilogue): Update recovery of stack pointer from
3943 frame pointer accordingly.
3944 (nios2_initial_elimination_offset): Update calculation of offset
3945 for eliminating to HARD_FRAME_POINTER_REGNUM.
3947 2014-03-10 Jakub Jelinek <jakub@redhat.com>
3950 * ipa.c (symtab_remove_unreachable_nodes): Don't call
3951 cgraph_get_create_node on VAR_DECLs.
3953 2014-03-10 Richard Biener <rguenther@suse.de>
3956 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
3958 2014-03-08 Douglas B Rupp <rupp@gnat.com>
3960 * config/vms/vms.opt (vms_float_format): New variable.
3962 2014-03-08 Tobias Burnus <burnus@net-b.de>
3964 * doc/invoke.texi (-fcilkplus): Update implementation status.
3966 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
3967 Richard Biener <rguenther@suse.de>
3969 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
3970 consistently accross all TUs.
3971 (run_gcc): Enable -fshort-double automatically at link at link-time
3972 and disallow override.
3974 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
3977 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
3978 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
3979 if they can't be used.
3981 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3983 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
3984 for Solaris 11/x86 ld.
3985 * configure: Regenerate.
3987 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3989 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
3990 (LIB_TLS_SPEC): Save as ld_tls_libs.
3991 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
3992 (HAVE_AS_IX86_TLSLDM): New test.
3993 * configure, config.in: Regenerate.
3994 * config/i386/i386.c (legitimize_tls_address): Fall back to
3995 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
3996 cannot support TLS_MODEL_LOCAL_DYNAMIC.
3997 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
3998 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
4000 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
4002 * common.opt (fira-loop-pressure): Mark as optimization.
4004 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
4006 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
4007 an OpenMP mappable type.
4009 2014-03-06 Matthias Klose <doko@ubuntu.com>
4011 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
4012 MULTILIB_OSDIRNAMES is not defined.
4014 2014-03-06 Jakub Jelinek <jakub@redhat.com>
4015 Meador Inge <meadori@codesourcery.com>
4018 * config/arm/arm.c (arm_tls_symbol_p): Remove.
4019 (arm_legitimize_address): Call legitimize_tls_address for any
4020 arm_tls_referenced_p expression, handle constant addend. Call it
4021 before testing for !TARGET_ARM.
4022 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
4024 2014-03-06 Richard Biener <rguenther@suse.de>
4030 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4032 * tree-streamer.c (record_common_node): Assert we don't record
4033 nodes with type double.
4034 (preload_common_node): Skip type double, complex double and double
4035 pointer since it is now frontend dependent due to fshort-double option.
4037 2014-03-06 Richard Biener <rguenther@suse.de>
4039 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
4040 or -fno-lto is specified and the linker has full plugin support.
4041 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
4042 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
4043 * lto-wrapper.c (merge_and_complain): Merge compile-time
4044 optimization levels.
4045 (run_gcc): And pass it through to the link options.
4047 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
4051 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4053 * cselib.c (remove_useless_values): Skip to avoid quadratic
4054 behavior if the condition moved from...
4055 (cselib_process_insn): ... here holds.
4057 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4060 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
4061 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
4064 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
4065 (TM_H): Add x86-tune.def.
4067 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4069 * config/aarch64/aarch64.c (generic_tunings):
4070 Use cortexa57_extra_costs.
4072 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4075 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4076 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4077 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4080 2014-03-04 Heiher <r@hev.cc>
4082 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4083 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4085 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4087 * config/i386/predicates.md (const2356_operand): Change to ...
4088 (const2367_operand): ... this.
4089 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4091 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4092 (*avx512pf_scatterpf<mode>sf): Ditto.
4093 (avx512pf_scatterpf<mode>df): Ditto.
4094 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4095 (*avx512pf_scatterpf<mode>df): Ditto.
4096 * config/i386/i386.c (ix86_expand_builtin): Update
4097 incorrect hint operand error message.
4099 2014-03-04 Richard Biener <rguenther@suse.de>
4101 * lto-section-in.c (lto_get_section_data): Fix const cast.
4103 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4105 * tree-streamer.c (record_common_node): Assert we don't record
4106 nodes with type double.
4107 (preload_common_node): Skip type double, complex double and double
4108 pointer since it is now frontend dependent due to fshort-double option.
4110 2014-03-04 Richard Biener <rguenther@suse.de>
4113 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4114 (lto_input_toplevel_asms): Likewise.
4115 * lto-section-in.c (lto_get_section_data): Instead do it here
4118 2014-03-04 Richard Biener <rguenther@suse.de>
4120 PR tree-optimization/60382
4121 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4122 dead PHIs a reduction.
4124 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4126 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4128 (_mm_prefetch): Move out of GCC target("sse") pragma.
4129 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4130 GCC target("prfchw") pragma.
4131 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4133 * config/i386/i386.c (ix86_option_override_internal): Enable
4134 -mprfchw with -mprefetchwt1.
4136 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4138 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4141 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4143 * opts.h (CL_PCH_IGNORE): Define.
4144 * targhooks.c (option_affects_pch_p):
4145 Return false for options that have CL_PCH_IGNORE set.
4146 * opt-functions.awk: Process PchIgnore.
4147 * doc/options.texi: Document PchIgnore.
4149 * config/arc/arc.opt (misize): Add PchIgnore property.
4151 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4153 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4154 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4155 constraint on constants to permit them being loaded into
4156 GENERAL_REGS or BASE_REGS.
4158 2014-03-03 Nick Clifton <nickc@redhat.com>
4160 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4161 anti-cacnonical alternatives.
4162 (negandhi3_real): New pattern.
4163 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4165 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4167 * config/avr/avr-mcus.def: Remove atxmega16x1.
4168 * config/avr/avr-tables.opt: Regenerate.
4169 * config/avr/t-multilib: Regenerate.
4170 * doc/avr-mmcu.texi: Regenerate.
4172 2014-03-03 Tobias Grosser <tobias@grosser.es>
4173 Mircea Namolaru <mircea.namolaru@inria.fr>
4175 PR tree-optimization/58028
4176 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4179 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4181 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4182 not handled by recognizers.
4184 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4187 * function.c (expand_function_end): Don't emit
4188 clobber_return_register sequence if clobber_after is a BARRIER.
4189 * cfgexpand.c (construct_exit_block): Append instructions before
4190 return_label to prev_bb.
4192 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4194 * config/rs6000/constraints.md: Document reserved use of "wc".
4196 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4199 * ipa.c (function_and_variable_visibility): When dissolving comdat
4200 group, also set all symbols to local.
4202 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4207 2013-12-14 Jan Hubicka <jh@suse.cz>
4209 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4211 2014-03-02 Jon Beniston <jon@beniston.com>
4217 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4218 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4219 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4220 (simple_return, *simple_return): New patterns
4221 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4222 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4224 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4226 * dwarf2out.c (gen_subprogram_die): Tidy.
4228 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4231 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4232 (*mov_t_msb_neg_negc): ... this new insn.
4234 2014-02-28 Jason Merrill <jason@redhat.com>
4237 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4240 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4243 * dwarf2out.c (decltype_auto_die): New static.
4244 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4245 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4246 (is_cxx_auto): Likewise.
4248 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4250 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4251 we are not using general regs only.
4253 2014-02-28 Richard Biener <rguenther@suse.de>
4256 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4257 previous fix and only allow to remove trivial pre-headers
4258 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4259 (remove_forwarder_block): Properly update the latch of a loop.
4261 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4264 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4265 (cselib_preserved_hash_table): New.
4266 (preserve_constants_and_equivs): Move preserved vals to it.
4267 (cselib_find_slot): Look it up first.
4268 (cselib_init): Initialize it.
4269 (cselib_finish): Release it.
4270 (dump_cselib_table): Dump it.
4272 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4275 * cselib.c (remove_useless_values): Skip to avoid quadratic
4276 behavior if the condition moved from...
4277 (cselib_process_insn): ... here holds.
4279 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4282 * var-tracking.c (vt_initialize): Apply the same condition to
4283 preserve the CFA base value.
4285 2014-02-28 Joey Ye <joey.ye@arm.com>
4288 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4289 if reload in progress or completed.
4291 2014-02-28 Tobias Burnus <burnus@net-b.de>
4294 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4297 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4299 * doc/tm.texi.in (Condition Code Status): Update documention for
4300 relative locations of cc0-setter and cc0-user.
4302 2014-02-27 Jeff Law <law@redhat.com>
4304 PR rtl-optimization/52714
4305 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4306 into two independent simple sets, if I3 is a jump, ensure the
4307 pattern we place into I3 is a (set (pc) ...).
4309 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4310 Jeff Law <law@redhat.com>
4312 PR rtl-optimization/49847
4313 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4314 are in different blocks.
4315 * doc/tm.texi (Condition Code Status): Update documention for
4316 relative locations of cc0-setter and cc0-user.
4318 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4321 * lra.c (lra_emit_add): Check SUBREG too.
4323 2014-02-27 Andreas Schwab <schwab@suse.de>
4325 * config/m68k/m68k.c (m68k_option_override): Disable
4326 -flive-range-shrinkage for classic m68k.
4327 (m68k_override_options_after_change): Likewise.
4329 2014-02-27 Marek Polacek <polacek@redhat.com>
4332 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4333 -Wmaybe-uninitialized.
4335 2014-02-27 Alan Modra <amodra@gmail.com>
4338 * reload1.c (emit_input_reload_insns): When reload_override_in,
4339 set old to rl->in_reg when rl->in_reg is a subreg.
4341 2014-02-26 Richard Biener <rguenther@suse.de>
4344 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4346 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4348 * common/config/i386/predicates.md (const1256_operand): Remove.
4349 (const2356_operand): New.
4350 (const_1_to_2_operand): Remove.
4351 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4352 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4353 (*avx512pf_gatherpf<mode>sf): Ditto.
4354 (avx512pf_gatherpf<mode>df): Ditto.
4355 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4356 (*avx512pf_gatherpf<mode>df): Ditto.
4357 (avx512pf_scatterpf<mode>sf): Ditto.
4358 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4359 (*avx512pf_scatterpf<mode>sf): Ditto.
4360 (avx512pf_scatterpf<mode>df): Ditto.
4361 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4362 (*avx512pf_scatterpf<mode>df): Ditto.
4363 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4365 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4367 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4368 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4369 (_mm512_mask_testn_epi64_mask): Move to ...
4370 * config/i386/avx512cdintrin.h: Here.
4371 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4372 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4373 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4374 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4375 TARGET_AVX512F from TARGET_AVX512CD.
4377 2014-02-26 Richard Biener <rguenther@suse.de>
4380 * ipa.c (walk_polymorphic_call_targets): Properly guard
4381 call to inline_update_overall_summary.
4383 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4386 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4387 and latches only if requested. Fix latch if it is removed.
4388 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4389 LOOPS_HAVE_PREHEADERS.
4391 2014-02-25 Andrew Pinski <apinski@cavium.com>
4393 * builtins.c (expand_builtin_thread_pointer): Create a new target
4394 when the target is NULL.
4396 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4398 PR rtl-optimization/60317
4399 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4400 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4401 * lra-assigns.c: Include params.h.
4402 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4403 other reload pseudos considerations.
4405 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4407 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4408 to use canonical form for nor<mode>3.
4410 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4413 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4416 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4418 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4419 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4420 (ix86_handle_option): Handle OPT_mprefetchwt1.
4421 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4422 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4424 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4425 OPTION_MASK_ISA_PREFETCHWT1.
4426 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4427 (PTA_PREFETCHWT1): New.
4428 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4429 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4430 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4431 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4432 (*prefetch_avx512pf_<mode>_: Change into ...
4433 (*prefetch_prefetchwt1_<mode>: This.
4434 * config/i386/i386.opt (mprefetchwt1): New.
4435 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4436 (_mm_prefetch): Handle intent to write.
4437 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4439 2014-02-25 Richard Biener <rguenther@suse.de>
4442 * emit-rtl.c (mem_attrs_htab): Remove.
4443 (mem_attrs_htab_hash): Likewise.
4444 (mem_attrs_htab_eq): Likewise.
4445 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4446 (init_emit_once): Do not allocate mem_attrs_htab.
4448 2014-02-25 Richard Biener <rguenther@suse.de>
4451 * lto-opts.c (lto_write_options): Output non-explicit conservative
4452 -fwrapv, -fno-trapv and -fno-strict-overflow.
4453 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4454 (run_gcc): And pass them through.
4456 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4458 * sel-sched.c (calculate_new_fences): New parameter ptime.
4459 Calculate it as a maximum over all fence cycles.
4460 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4461 Print the final schedule timing when sched_verbose.
4463 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4465 PR rtl-optimization/60292
4466 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4467 bit fot the fence instruction.
4469 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4471 * calls.h: Fix typo in comment.
4473 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4475 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4476 adjusting offsetable addresses.
4478 2014-02-24 Guozhi Wei <carrot@google.com>
4480 * sparseset.h (sparseset_pop): Fix the wrong index.
4482 2014-02-24 Walter Lee <walt@tilera.com>
4484 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4485 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4487 * common/config/tilegx/tilegx-common.c
4488 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4489 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4491 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4492 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4493 (tilegx_gimplify_va_arg_expr): Handle big endian.
4494 (tilegx_expand_unaligned_load): Ditto.
4495 (tilegx_expand_unaligned_store): Ditto.
4496 (TARGET_RETURN_IN_MSB): New.
4497 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4498 (TARGET_ENDIAN_DEFAULT): New.
4499 (TARGET_BIG_ENDIAN): Handle big endian.
4500 (BYTES_BIG_ENDIAN): Ditto.
4501 (WORDS_BIG_ENDIAN): Ditto.
4502 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4505 * config/tilegx/tilegx.md (extv): Handle big endian.
4507 (insn_st<n>): Ditto.
4508 (insn_st<n>_add<bitsuffix>): Ditto.
4509 (insn_stnt<n>): Ditto.
4510 (insn_stnt<n>_add<bitsuffix>):Ditto.
4511 (vec_interleave_highv8qi): Handle big endian.
4512 (vec_interleave_highv8qi_be): New.
4513 (vec_interleave_highv8qi_le): New.
4514 (insn_v1int_h): Handle big endian.
4515 (vec_interleave_lowv8qi): Handle big endian.
4516 (vec_interleave_lowv8qi_be): New.
4517 (vec_interleave_lowv8qi_le): New.
4518 (insn_v1int_l): Handle big endian.
4519 (vec_interleave_highv4hi): Handle big endian.
4520 (vec_interleave_highv4hi_be): New.
4521 (vec_interleave_highv4hi_le): New.
4522 (insn_v2int_h): Handle big endian.
4523 (vec_interleave_lowv4hi): Handle big endian.
4524 (vec_interleave_lowv4hi_be): New.
4525 (vec_interleave_lowv4hi_le): New.
4526 (insn_v2int_l): Handle big endian.
4527 (vec_interleave_highv2si): Handle big endian.
4528 (vec_interleave_highv2si_be): New.
4529 (vec_interleave_highv2si_le): New.
4530 (insn_v4int_h): Handle big endian.
4531 (vec_interleave_lowv2si): Handle big endian.
4532 (vec_interleave_lowv2si_be): New.
4533 (vec_interleave_lowv2si_le): New.
4534 (insn_v4int_l): Handle big endian.
4535 * config/tilegx/tilegx.opt (mbig-endian): New option.
4536 (mlittle-endian): New option.
4537 * doc/install.texi: Document tilegxbe-linux.
4538 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4540 2014-02-24 Martin Jambor <mjambor@suse.cz>
4543 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4544 there are no parameter descriptors.
4546 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4548 PR rtl-optimization/60268
4549 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4550 initialization to ...
4551 (sched_rgn_init): ... here.
4552 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4554 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4556 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4559 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4561 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4564 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4566 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4567 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4569 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4571 * config/microblaze/predicates.md: Add cmp_op predicate.
4572 * config/microblaze/microblaze.md: Add branch_compare instruction
4573 which uses cmp_op predicate and emits cmp insn before branch.
4574 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4575 to microblaze_expand_conditional_branch and consolidate logic.
4576 (microblaze_expand_conditional_branch): emit branch_compare
4577 insn instead of handling cmp op separate from branch insn.
4579 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4581 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4584 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4586 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4587 define_insn with define_expand and new define_insn
4588 *altivec_lve<VI_char>x_internal.
4589 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4590 and new define_insn *altivec_stve<VI_char>x_internal.
4591 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4593 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4595 (altivec_expand_stvex_be): New function.
4597 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4599 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4600 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4601 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4602 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4604 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4607 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4608 instead of emit_move_insn.
4610 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4612 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4614 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4615 gen_altivec_vsumsws.
4617 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4619 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4620 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4621 (altivec_lvxl_<mode>): New define_expand incorporating
4622 -maltivec=be semantics where needed.
4623 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4624 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4625 semantics where needed.
4626 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4627 (altivec_stvx_<mode>): New define_expand incorporating
4628 -maltivec=be semantics where needed.
4629 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4630 VM2 iterator instead of V4SI.
4631 (altivec_stvxl_<mode>): New define_expand incorporating
4632 -maltivec=be semantics where needed.
4633 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4634 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4635 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4636 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4637 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4638 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4639 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4640 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4641 ALTIVEC_BUILTIN_STVXL.
4642 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4643 (altivec_expand_stvx_be): Likewise.
4644 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4645 (altivec_expand_lvx_be): Likewise.
4646 (altivec_expand_stvx_be): Likewise.
4647 (altivec_expand_builtin): Add cases for
4648 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4649 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4650 (altivec_init_builtins): Add definitions for
4651 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4652 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4654 2014-02-21 Catherine Moore <clm@codesourcery.com>
4656 * doc/invoke.texi (mvirt, mno-virt): Document.
4657 * config/mips/mips.opt (mvirt): New option.
4658 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4660 2014-02-21 Richard Biener <rguenther@suse.de>
4662 PR tree-optimization/60276
4663 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4664 (STMT_VINFO_MIN_NEG_DIST): New macro.
4665 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4666 STMT_VINFO_MIN_NEG_DIST.
4667 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4668 made for negative dependence distances still hold.
4670 2014-02-21 Richard Biener <rguenther@suse.de>
4673 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4674 DECL_INITIAL for globals not in the current function context.
4676 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4678 PR tree-optimization/56490
4679 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4680 * tree-ssa-uninit.c: Include params.h.
4681 (compute_control_dep_chain): Add num_calls argument, return false
4682 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4683 num_calls to recursive call.
4684 (find_predicates): Change dep_chain into normal array,
4685 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4686 variable and adjust compute_control_dep_chain caller.
4687 (find_def_preds): Likewise.
4689 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4691 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4692 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4694 2014-02-21 Nick Clifton <nickc@redhat.com>
4696 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4697 (pushhi1): Likewise.
4698 (popqi1): Add mode to pre_dec.
4701 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4703 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4704 mode for mask of V8SFmode permutation.
4706 2014-02-20 Richard Henderson <rth@redhat.com>
4709 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4710 a new pseudo for OLDVAL.
4712 2014-02-20 Jakub Jelinek <jakub@redhat.com>
4715 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4716 gen_reg_rtx if d->testing_p.
4717 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4718 if d->testing_p and we will certainly return true.
4719 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4722 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
4724 * emit-rtl.c (gen_reg_rtx): Assert that
4725 crtl->emit.regno_pointer_align_length is non-zero.
4727 2014-02-20 Richard Henderson <rth@redhat.com>
4730 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4731 on failure the store back into EXPECT.
4733 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4734 Sandra Loosemore <sandra@codesourcery.com>
4736 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
4737 * config/nios2/nios2.c (nios2_function_profiler): Add
4738 -fPIC (flag_pic == 2) support.
4739 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
4740 (nios2_large_offset_p): New function.
4741 (nios2_unspec_reloc_p): Move up position, update to use
4742 nios2_large_offset_p.
4743 (nios2_unspec_address): Remove function.
4744 (nios2_unspec_offset): New function.
4745 (nios2_large_got_address): New function.
4746 (nios2_got_address): Add large offset support.
4747 (nios2_legitimize_tls_address): Update usage of removed and new
4749 (nios2_symbol_binds_local_p): New function.
4750 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
4751 (nios2_legitimize_address): Update to use nios2_large_offset_p.
4752 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
4753 (nios2_print_operand): Merge H/L processing, add hiadj/lo
4754 processing for (const (unspec ...)).
4755 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
4757 2014-02-20 Richard Biener <rguenther@suse.de>
4759 * tree-cfg.c (replace_uses_by): Mark altered BBs before
4760 doing the substitution.
4761 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
4763 2014-02-20 Martin Jambor <mjambor@suse.cz>
4766 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
4767 info when checking whether lattices are bottom.
4769 2014-02-20 Richard Biener <rguenther@suse.de>
4772 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
4775 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
4778 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
4779 parameter specifying the scaling.
4780 (inline_call): Update.
4781 (want_inline_recursively): Guard division by zero.
4782 (recursive_inlining): Update.
4783 * ipa-inline.h (clone_inlined_nodes): Update.
4785 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4788 * config/i386/i386.c (classify_argument): Pass structures of size
4789 64 bytes or less in register.
4791 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4792 Kirill Yukhin <kirill.yukhin@intel.com>
4794 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
4795 (_mm_rcp28_round_ss): Ditto.
4796 (_mm_rsqrt28_round_sd): Ditto.
4797 (_mm_rsqrt28_round_ss): Ditto.
4798 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
4799 (_mm_rcp14_round_ss): Ditto.
4800 (_mm_rsqrt14_round_sd): Ditto.
4801 (_mm_rsqrt14_round_ss): Ditto.
4802 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
4803 the first input operand, get rid of match_dup.
4804 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
4806 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
4808 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
4809 operand as the first input operand, set type attribute.
4810 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
4812 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
4813 operand as the first input operand, set type attribute.
4815 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4817 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
4820 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
4823 * config/i386/i386.c (construct_container): Remove TFmode check
4824 for X86_64_INTEGER_CLASS.
4826 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
4829 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
4830 only when -Wpsabi is enabled.
4832 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
4835 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
4836 passing arrays in registers are the same as for structs, so remove the
4837 special case for them.
4839 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
4841 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
4842 destination type, extract only the valid bits if the source type is not
4843 integral and has a different mode.
4845 2014-02-19 Richard Biener <rguenther@suse.de>
4848 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
4851 2014-02-19 Richard Biener <rguenther@suse.de>
4854 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
4855 (ipa_modify_call_arguments): Emit an argument load explicitely and
4856 preserve virtual SSA form there and for the replacement call.
4857 Do not update SSA form nor free dominance info.
4859 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4861 * ipa.c (function_and_variable_visibility): Also clear WEAK
4862 flag when disolving COMDAT_GROUP.
4864 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4866 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
4867 * ipa-prop.c (ipa_set_jf_known_type): Return early when
4869 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
4870 do more sanity checks.
4871 (detect_type_change): Return true when giving up early.
4872 (compute_complex_assign_jump_func): Fix type parameter of
4873 ipa_set_ancestor_jf.
4874 (compute_complex_ancestor_jump_func): Likewise.
4875 (update_jump_functions_after_inlining): Fix updating of
4877 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
4879 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4881 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
4882 inline clones when edge disappears.
4884 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4887 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
4888 Split 64-bit moves into 2 patterns. Do not allow the use of
4889 direct move for TDmode in little endian, since the decimal value
4890 has little endian bytes within a word, but the 64-bit pieces are
4891 ordered in a big endian fashion, and normal subreg's of TDmode are
4893 (mov<mode>_64bit_dm): Likewise.
4894 (movtd_64bit_nodm): Likewise.
4896 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4898 PR tree-optimization/60174
4899 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
4900 statement of an SSA_NAME that occurs in an abnormal PHI node.
4902 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4905 * final.c (SEEN_BB): Remove.
4906 (SEEN_NOTE, SEEN_EMITTED): Renumber.
4907 (final_scan_insn): Don't force_source_line on second
4908 NOTE_INSN_BASIC_BLOCK.
4910 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
4913 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
4914 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
4915 (type_natural_mode): Warn ABI change when %zmm register is not
4916 available for AVX512F vector value passing.
4918 2014-02-18 Kai Tietz <ktietz@redhat.com>
4921 * config/i386/i386.c (ix86_expand_prologue): Use value in
4922 rax register as displacement when restoring %r10 or %rax.
4923 Fix wrong offset when restoring both registers.
4925 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4927 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
4928 assertion with conditional return.
4930 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4931 Uros Bizjak <ubizjak@gmail.com>
4934 * config/i386/driver-i386.c (host_detect_local_cpu): If
4935 YMM state is not saved by the OS, also clear has_f16c. Move
4936 CPUID 0x80000001 handling before YMM state saving checking.
4938 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
4940 PR rtl-optimization/58960
4941 * haifa-sched.c (alloc_global_sched_pressure_data): New,
4942 factored out from ...
4943 (sched_init): ... here.
4944 (free_global_sched_pressure_data): New, factored out from ...
4945 (sched_finish): ... here.
4946 * sched-int.h (free_global_sched_pressure_data): Declare.
4947 * sched-rgn.c (nr_regions_initial): New static global.
4948 (haifa_find_rgns): Initialize it.
4949 (schedule_region): Disable sched-pressure for the newly
4952 2014-02-17 Richard Biener <rguenther@suse.de>
4954 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
4955 release SSA defs of pattern stmts.
4957 2014-02-17 Richard Biener <rguenther@suse.de>
4959 * tree-inline.c (expand_call_inline): Release the virtual
4960 operand defined by the call we are about to inline.
4962 2014-02-17 Richard Biener <rguenther@suse.de>
4964 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
4966 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
4967 Ilya Tocar <ilya.tocar@intel.com>
4969 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
4970 arguments order in builtin.
4971 (_mm512_permutexvar_epi64): Ditto.
4972 (_mm512_mask_permutexvar_epi64): Ditto
4973 (_mm512_maskz_permutexvar_epi32): Ditto
4974 (_mm512_permutexvar_epi32): Ditto
4975 (_mm512_mask_permutexvar_epi32): Ditto
4977 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4979 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
4980 (p8_vmrgow): Likewise.
4982 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4984 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
4987 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
4990 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
4991 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
4992 into 64-bit and 32-bit moves. On 64-bit moves, add support for
4993 using direct move instructions on ISA 2.07. Also adjust
4994 instruction length for 64-bit.
4995 (mov<mode>_64bit, TFmode/TDmode): Likewise.
4996 (mov<mode>_32bit, TFmode/TDmode): Likewise.
4998 2014-02-15 Alan Modra <amodra@gmail.com>
5002 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
5003 find_replacement on parts of insn rtl that might be reloaded.
5005 2014-02-15 Richard Biener <rguenther@suse.de>
5007 PR tree-optimization/60183
5008 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
5009 (tree_ssa_phiprop): Calculate and free post-dominators.
5011 2014-02-14 Jeff Law <law@redhat.com>
5013 PR rtl-optimization/60131
5014 * ree.c (get_extended_src_reg): New function.
5015 (combine_reaching_defs): Use it rather than assuming location of REG.
5016 (find_and_remove_re): Verify first operand of extension is
5017 a REG before adding the insns to the copy list.
5019 2014-02-14 Roland McGrath <mcgrathr@google.com>
5021 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
5022 * configure: Regenerated.
5023 * config.in: Regenerated.
5024 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
5025 instead of ASM_SHORT.
5027 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
5028 Richard Earnshaw <rearnsha@arm.com>
5030 PR rtl-optimization/59535
5031 * lra-constraints.c (process_alt_operands): Encourage alternative
5032 when unassigned pseudo class is superset of the alternative class.
5033 (inherit_reload_reg): Don't inherit when optimizing for code size.
5034 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
5035 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
5036 modes not less than 4 for Thumb1.
5038 2014-02-14 Kyle McMartin <kyle@redhat.com>
5041 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
5043 2014-02-14 Richard Biener <rguenther@suse.de>
5045 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
5046 (get_frame_arg): Drop the assert with langhook types_compatible_p.
5047 Do not strip INDIRECT_REFs.
5049 2014-02-14 Richard Biener <rguenther@suse.de>
5052 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
5053 DECL_FUNCTION_SPECIFIC_TARGET.
5054 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
5055 * tree-streamer-out.c (pack_ts_target_option): Remove.
5056 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
5057 (write_ts_function_decl_tree_pointers): Do not stream
5058 DECL_FUNCTION_SPECIFIC_TARGET.
5059 * tree-streamer-in.c (unpack_ts_target_option): Remove.
5060 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
5061 (lto_input_ts_function_decl_tree_pointers): Do not stream
5062 DECL_FUNCTION_SPECIFIC_TARGET.
5064 2014-02-14 Jakub Jelinek <jakub@redhat.com>
5066 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
5067 (get_initial_def_for_induction, vectorizable_induction): Ignore
5068 debug stmts when looking for exit_phi.
5069 (vectorizable_live_operation): Fix up condition.
5071 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5073 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5074 nreverse() because it changes the content of original tree list.
5076 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5078 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5079 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5081 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5083 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5084 GNU coding standards.
5086 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5089 * dwarf2out.c (gen_subprogram_die): Don't call
5090 add_calling_convention_attribute if subr_die is old_die.
5092 2014-02-13 Sharad Singhai <singhai@google.com>
5094 * doc/optinfo.texi: Fix order of nodes.
5096 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5098 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5099 operands[2], not operands[3].
5101 2014-02-13 Richard Biener <rguenther@suse.de>
5104 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5105 mention the possibility of an in-tree build.
5106 (CLooG): Update recommended version to 0.18.1, mention the
5107 possibility of an in-tree build and clarify that the ISL
5108 bundled with CLooG does not work.
5110 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5113 * expr.c (compress_float_constant): If x is a hard register,
5114 extend into a pseudo and then move to x.
5116 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5118 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5119 caused by bad second argument to warning_at() with -mhotpatch and
5120 nested functions (e.g. with gfortran).
5122 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5124 * opts.c (option_name): Remove "enabled by default" rider.
5126 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5128 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5130 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5131 Uros Bizjak <ubizjak@gmail.com>
5134 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5135 * configure: Regenerated.
5137 2014-02-12 Richard Biener <rguenther@suse.de>
5139 * vec.c (vec_prefix::calculate_allocation): Move as
5140 inline variant to vec.h.
5141 (vec_prefix::calculate_allocation_1): New out-of-line version.
5142 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5143 (vec_prefix::m_has_auto_buf): Rename to ...
5144 (vec_prefix::m_using_auto_storage): ... this.
5145 (vec_prefix::calculate_allocation): Inline the easy cases
5146 and dispatch to calculate_allocation_1 which doesn't need the
5148 (va_heap::reserve): Use gcc_checking_assert.
5149 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5150 m_using_auto_storage.
5151 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5153 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5154 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5155 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5157 2014-02-12 Richard Biener <rguenther@suse.de>
5159 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5160 when we found a dependence.
5162 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5164 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5166 (maybe_fold_stmt): ... into this new function.
5167 * omp-low.c (lower_omp): Update comment.
5169 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5172 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5175 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5177 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5178 identifiers in comments.
5179 (cortexa53_extra_costs): Likewise.
5180 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5181 (cortexa7_extra_costs): Likewise.
5182 (cortexa12_extra_costs): Likewise.
5183 (cortexa15_extra_costs): Likewise.
5184 (v7m_extra_costs): Likewise.
5186 2014-02-12 Richard Biener <rguenther@suse.de>
5189 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5190 of posix_memalign being successful.
5191 (lower_stmt): Restrict lowering of posix_memalign to when
5192 -ftree-bit-ccp is enabled.
5194 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5196 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5198 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5200 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5202 PR rtl-optimization/60116
5203 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5204 other_insn once the combination has been validated.
5206 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5209 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5211 * ipa-devirt.c: Include demangle.h
5212 (odr_violation_reported): New static variable.
5213 (add_type_duplicate): Update odr_violations.
5214 (maybe_record_node): Add completep parameter; update it.
5215 (record_target_from_binfo): Add COMPLETEP parameter;
5216 update it as needed.
5217 (possible_polymorphic_call_targets_1): Likewise.
5218 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5219 rename FINAL to COMPLETE.
5220 (record_targets_from_bases): Sanity check we found the binfo;
5221 fix COMPLETEP updating.
5222 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5223 parameter, fix computing of COMPLETEP.
5224 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5225 at LTO time do demangling.
5226 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5227 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5229 (gimple_get_virt_method_for_binfo): Likewise.
5230 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5231 gimple_get_virt_method_for_vtable): Update prototypes.
5233 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5236 * genautomata.c (add_presence_absence): Fix typo with
5237 {final_}presence_list.
5239 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5242 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5243 for VSX/Altivec vectors that land in GPR registers.
5245 2014-02-11 Richard Henderson <rth@redhat.com>
5246 Jakub Jelinek <jakub@redhat.com>
5249 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5250 around drhs if type conversion to lacc->type is not useless.
5252 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5254 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5256 (cortex-a57.cortex-a53): Likewise.
5257 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5259 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5261 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5264 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5266 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5267 add_options_for_arm_vfp3.
5269 2014-02-11 Jeff Law <law@redhat.com>
5272 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5273 object with an undesirable mode.
5275 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5278 * config/i386/sol2-9.h: New file.
5279 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5280 *-*-solaris2.9*): Use it.
5282 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5284 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5285 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5287 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5289 * config/microblaze/microblaze.c: Extend mcpu version format
5291 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5293 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5295 2014-02-10 Richard Henderson <rth@redhat.com>
5298 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5299 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5300 ms-abi vs -mno-accumulate-outgoing-args.
5301 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5302 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5305 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5308 * cfgexpand.c (expand_asm_operands): Attach source location to
5309 ASM_INPUT rtx objects.
5310 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5312 2014-02-10 Nick Clifton <nickc@redhat.com>
5314 * config/mn10300/mn10300.c (popcount): New function.
5315 (mn10300_expand_prologue): Include saved registers in stack usage
5318 2014-02-10 Jeff Law <law@redhat.com>
5321 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5322 when changing the SET_DEST of a prior insn to avoid an input reload.
5324 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5326 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5327 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5328 -mcall-openbsd, or -mcall-linux.
5329 (CC1_ENDIAN_BIG_SPEC): Remove.
5330 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5331 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5332 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5333 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5334 and %cc1_endian_default.
5335 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5337 2014-02-10 Richard Biener <rguenther@suse.de>
5339 PR tree-optimization/60115
5340 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5341 MEM_REF handling. Properly verify that the accesses are not
5342 out of the objects bound.
5344 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5346 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5349 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5351 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5352 proper constants and fix formatting.
5353 (possible_polymorphic_call_targets): Fix formatting.
5355 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5356 Ilya Tocar <ilya.tocar@intel.com>
5358 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5359 (_mm512_loadu_epi32): Renamed into...
5360 (_mm512_loadu_si512): This.
5361 (_mm512_storeu_epi32): Renamed into...
5362 (_mm512_storeu_si512): This.
5363 (_mm512_maskz_ceil_ps): Removed.
5364 (_mm512_maskz_ceil_pd): Ditto.
5365 (_mm512_maskz_floor_ps): Ditto.
5366 (_mm512_maskz_floor_pd): Ditto.
5367 (_mm512_floor_round_ps): Ditto.
5368 (_mm512_floor_round_pd): Ditto.
5369 (_mm512_ceil_round_ps): Ditto.
5370 (_mm512_ceil_round_pd): Ditto.
5371 (_mm512_mask_floor_round_ps): Ditto.
5372 (_mm512_mask_floor_round_pd): Ditto.
5373 (_mm512_mask_ceil_round_ps): Ditto.
5374 (_mm512_mask_ceil_round_pd): Ditto.
5375 (_mm512_maskz_floor_round_ps): Ditto.
5376 (_mm512_maskz_floor_round_pd): Ditto.
5377 (_mm512_maskz_ceil_round_ps): Ditto.
5378 (_mm512_maskz_ceil_round_pd): Ditto.
5379 (_mm512_expand_pd): Ditto.
5380 (_mm512_expand_ps): Ditto.
5381 * config/i386/i386.c (ix86_builtins): Remove
5382 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5383 (bdesc_args): Ditto.
5384 * config/i386/predicates.md (const1256_operand): New.
5385 (const_1_to_2_operand): Ditto.
5386 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5387 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5388 (*avx512pf_gatherpf<mode>sf): Ditto.
5389 (avx512pf_gatherpf<mode>df): Ditto.
5390 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5391 (*avx512pf_gatherpf<mode>df): Ditto.
5392 (avx512pf_scatterpf<mode>sf): Ditto.
5393 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5394 (*avx512pf_scatterpf<mode>sf): Ditto.
5395 (avx512pf_scatterpf<mode>df): Ditto.
5396 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5397 (*avx512pf_scatterpf<mode>df): Ditto.
5398 (avx512f_expand<mode>): Removed.
5399 (<shift_insn><mode>3<mask_name>): Change predicate type.
5401 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5403 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5404 not at the end of datarefs vector use ordered_remove to avoid
5405 reordering datarefs vector.
5408 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5409 mark local addressable non-static vars as GOVD_PRIVATE
5410 instead of GOVD_LOCAL.
5411 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5412 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5416 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5417 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5418 assume_aligned or alloc_align attributes.
5419 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5420 arguments. Handle also assume_aligned and alloc_align attributes.
5421 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5422 calls to functions with assume_aligned or alloc_align attributes.
5423 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5425 2014-02-08 Terry Guo <terry.guo@arm.com>
5427 * doc/invoke.texi: Document ARM -march=armv7e-m.
5429 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5431 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5432 flag on __cilkrts_rethrow builtin.
5435 * ipa-cp.c (determine_versionability): Fail at -O0
5436 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5437 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5440 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5443 * tree-inline.c (copy_forbidden): Fail for
5444 __attribute__((optimize (0))) functions.
5446 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5448 * varpool.c: Include pointer-set.h.
5449 (varpool_remove_unreferenced_decls): Variables in other partitions
5450 will not be output; be however careful to not lose information
5453 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5455 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5456 lookup in the vtable constructor.
5458 2014-02-07 Jeff Law <law@redhat.com>
5461 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5462 define_insn_and_split.
5464 * ipa-inline.c (inline_small_functions): Fix typos.
5466 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5468 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5469 (s390_can_use_return_insn): Declare.
5470 * config/s390/s390.h (EPILOGUE_USES): Define.
5471 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5473 (s390_chunkify_start): Handle return JUMP_LABELs.
5474 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5475 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5476 (s390_can_use_return_insn): New functions.
5477 (s390_fix_long_loop_prediction): Handle conditional returns.
5478 (TARGET_SET_UP_BY_PROLOGUE): Define.
5479 * config/s390/s390.md (ANY_RETURN): New code iterator.
5480 (*creturn, *csimple_return, return, simple_return): New patterns.
5482 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5484 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5485 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5486 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5487 REG_CFA_RESTORE list when deciding not to restore a register.
5489 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5491 * config/s390/s390.c: Include tree-pass.h and context.h.
5492 (s390_early_mach): New function, split out from...
5493 (s390_emit_prologue): ...here.
5494 (pass_data_s390_early_mach): New pass structure.
5495 (pass_s390_early_mach): New class.
5496 (s390_option_override): Create and register early_mach pass.
5497 Move to end of file.
5499 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5501 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5502 to match for the exit block.
5504 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5506 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5507 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5508 Reject misaligned operands.
5510 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5512 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5514 2014-02-07 Richard Biener <rguenther@suse.de>
5517 * gimple-low.c (lower_builtin_posix_memalign): New function.
5518 (lower_stmt): Call it to lower posix_memalign in a way
5519 to make alignment info accessible.
5521 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5524 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5525 __builtin_setjmp_receiver.
5527 2014-02-07 Richard Biener <rguenther@suse.de>
5530 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5531 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5532 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5533 Handle BUILT_IN_POSIX_MEMALIGN.
5534 (find_func_clobbers): Likewise.
5535 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5536 (call_may_clobber_ref_p_1): Likewise.
5538 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5541 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5544 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5547 * lto-cgraph.c (lto_output_node): Use
5548 symtab_get_symbol_partitioning_class.
5549 (lto_output_varpool_node): likewise.
5550 (symtab_get_symbol_partitioning_class): Move here from
5552 * cgraph.h (symbol_partitioning_class): Likewise.
5553 (symtab_get_symbol_partitioning_class): Declare.
5555 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5557 * ggc.h (ggc_internal_cleared_alloc): New macro.
5558 * vec.h (vec_safe_copy): Handle memory stats.
5559 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5560 * target-globals.c (save_target_globals): Likewise.
5562 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5565 * expr.c (emit_move_resolve_push): Export; be bit more selective
5566 on when to clear alias set.
5567 * expr.h (emit_move_resolve_push): Declare.
5568 * function.h (struct function): Add tail_call_marked.
5569 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5570 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5571 * config/i386/i386.md (TImode move expander): De not call
5573 (FP push expanders): Preserve memory attributes.
5574 * config/i386/sse.md (push<mode>1): Remove.
5575 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5576 (ix86_expand_push): Remove.
5577 * config/i386/mmx.md (push<mode>1): Remove.
5579 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5581 PR rtl-optimization/60030
5582 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5583 lopart with paradoxical subreg before shifting it up by hprec.
5585 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5587 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5588 Remove extra newline at end of file.
5589 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5590 (arm_issue_rate): Handle cortexa57.
5591 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5592 (cortex-a57.cortex-a53): Likewise.
5594 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5597 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5598 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5600 (arm_expand_prologue): Adjust all callers.
5601 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5602 info, registers also at the lowest numbered registers side. Use
5603 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5607 * var-tracking.c (adjust_mems): Before adding a SET to
5608 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5610 2014-02-06 Alan Modra <amodra@gmail.com>
5613 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5614 change SDmode to DDmode when lra_in_progress.
5616 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5619 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5620 free_data_ref on the dr first, and before goto again also set dr
5621 to the next dr. For simd_lane_access, free old datarefs[i] before
5622 overwriting it. For get_vectype_for_scalar_type failure, don't
5623 free_data_ref if simd_lane_access.
5625 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5628 * tree.h (opts_for_fn): New inline function.
5629 (opt_for_fn): Define.
5630 * config/i386/i386.c (ix86_function_regparm): Use
5631 opt_for_fn (decl, optimize) instead of optimize.
5633 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5635 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5636 for SYMBOL_REF in large memory model.
5638 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5640 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5642 (cortex-a57): Likewise.
5643 (cortex-a57.cortex-a53): Likewise.
5645 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5646 Kugan Vivekanandarajah <kuganv@linaro.org>
5648 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5650 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5652 2014-02-06 Richard Biener <rguenther@suse.de>
5654 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5655 set_loop_copy and initialize_original_copy_tables.
5657 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5659 * config/aarch64/aarch64-simd.md
5660 (aarch64_ashr_simddi): Change QI to SI.
5662 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5663 Jakub Jelinek <jakub@redhat.com>
5666 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5669 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5671 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5672 CODE_FOR_altivec_vpku[hw]um to
5673 CODE_FOR_altivec_vpku[hw]um_direct.
5674 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5675 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5676 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5677 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5679 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5681 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5682 generation for -maltivec=be.
5683 (altivec_vsumsws): Simplify redundant test.
5685 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5687 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5688 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5689 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5690 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5691 gen_altivec_vpkuwum.
5692 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5694 (altivec_vpks<VI_char>ss): Likewise.
5695 (altivec_vpks<VI_char>us): Likewise.
5696 (altivec_vpku<VI_char>us): Likewise.
5697 (altivec_vpku<VI_char>um): Likewise.
5698 (altivec_vpku<VI_char>um_direct): New (copy of
5699 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5701 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5702 target is little endian and -maltivec=be is not specified.
5703 (*altivec_vupkhs<VU_char>_direct): New (copy of
5704 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
5705 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5706 target is little endian and -maltivec=be is not specified.
5707 (*altivec_vupkls<VU_char>_direct): New (copy of
5708 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
5709 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5710 little endian and -maltivec=be is not specified.
5711 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5712 little endian and -maltivec=be is not specified.
5714 2014-02-05 Richard Henderson <rth@redhat.com>
5717 * combine-stack-adj.c: Revert r206943.
5718 * sched-int.h (struct deps_desc): Add last_args_size.
5719 * sched-deps.c (init_deps): Initialize it.
5720 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5721 contain REG_ARGS_SIZE notes.
5723 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5725 * lto-cgraph.c (asm_nodes_output): Make global.
5726 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
5727 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
5728 (driver_handle_option): Handle OPT_fwpa.
5730 2014-02-05 Jakub Jelinek <jakub@redhat.com>
5733 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5734 a comment typo and formatting issue. If odr_hash hasn't been
5735 created, return vNULL and set *completep to false.
5738 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
5739 bb with no successors.
5741 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
5744 * doc/invoke.texi (-march): Clarify documentation for ARM.
5748 2014-02-05 Richard Biener <rguenther@suse.de>
5750 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
5751 when not vectorizing because of too many alias checks.
5752 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
5753 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
5755 2014-02-05 Nick Clifton <nickc@redhat.com>
5757 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
5758 accept extended registers in any mode when compiling for the MN10300.
5760 2014-02-05 Yury Gribov <y.gribov@samsung.com>
5762 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
5763 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
5764 sanitization attributes.
5765 (can_inline_edge_p): Likewise.
5766 (sanitize_attrs_match_for_inline_p): New function.
5768 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5770 * ipa-prop.c (detect_type_change): Shor circuit testing of
5771 type changes on THIS pointer.
5773 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
5776 * config/pa/pa.c (legitimize_tls_address): Return original address
5777 if not passed a SYMBOL_REF rtx.
5778 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
5780 (pa_emit_move_sequence): Simplify TLS source operands.
5781 (pa_legitimate_constant_p): Reject all TLS constants.
5782 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
5783 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
5785 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5787 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
5788 groups when we know they are controlled by LTO.
5789 * varasm.c (default_binds_local_p_1): If object is in other partition,
5790 it will be resolved locally.
5792 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5794 * config/host-linux.c (linux_gt_pch_use_address): Don't
5795 use SSIZE_MAX because it is not always defined.
5797 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
5800 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
5801 threshold for pseudo splitting.
5802 (update_ebb_live_info): Process call argument hard registers and
5803 hard registers from insn definition too.
5804 (max_small_class_regs_num): New constant.
5805 (inherit_in_ebb): Update live hard regs through EBBs. Update
5806 reloads_num only for small register classes. Don't split for
5809 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
5812 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
5815 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5817 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
5820 2014-02-04 Marek Polacek <polacek@redhat.com>
5822 * gdbinit.in (pel): Define.
5824 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5826 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
5829 2014-02-04 Richard Biener <rguenther@suse.de>
5832 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
5833 in function context local.
5834 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
5835 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
5836 similar to LTO_imported_decl_ref.
5838 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5840 PR tree-optimization/60002
5841 * cgraphclones.c (build_function_decl_skip_args): Clear
5844 PR tree-optimization/60023
5845 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
5846 false to gsi_replace.
5847 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
5848 has been in some EH region and vec_stmt could throw, add
5849 vec_stmt into the same EH region.
5850 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
5851 has no lhs, ignore it.
5852 * internal-fn.c (expand_MASK_LOAD): Likewise.
5855 * tree-inline.c (copy_forbidden): Fail for
5856 __attribute__((optimize (0))) functions.
5859 * omp-low.c (simd_clone_struct_copy): If from->inbranch
5860 is set, copy one less argument.
5861 (expand_simd_clones): Don't subtract clone_info->inbranch
5862 from simd_clone_struct_alloc argument.
5864 PR rtl-optimization/57915
5865 * recog.c (simplify_while_replacing): If all unary/binary/relational
5866 operation arguments are constant, attempt to simplify those.
5869 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
5870 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
5872 2014-02-04 Richard Biener <rguenther@suse.de>
5874 PR tree-optimization/60012
5875 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
5876 TBAA disambiguation to all DDRs.
5878 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5881 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
5882 (LINK_SPEC): Use it for -shared, -shared-libgcc.
5884 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5887 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
5889 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5891 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
5892 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
5894 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5897 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
5898 to figure out targets of polymorphic calls with known decl.
5899 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5900 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
5901 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
5902 (get_polymorphic_call_info): ... here.
5903 (get_polymorphic_call_info_from_invariant): New function.
5905 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5907 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
5908 lookup via vtable pointer; check for type consistency
5909 and turn inconsitent facts into UNREACHABLE.
5910 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5911 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
5912 type inconsistent querries; return UNREACHABLE instead.
5914 2014-02-03 Richard Henderson <rth@twiddle.net>
5917 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
5918 already processed this node.
5919 (normalize_one_pred_1): Pass along mark_set.
5920 (normalize_one_pred): Create and destroy a pointer_set_t.
5921 (normalize_one_pred_chain): Likewise.
5923 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
5925 PR gcov-profile/58602
5926 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
5928 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5931 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
5932 -fno-devirtualize; try to devirtualize by the knowledge of
5933 virtual table pointer given by aggregate propagation.
5934 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5935 (ipa_print_node_jump_functions): Dump also offset that
5936 is relevant for polymorphic calls.
5937 (determine_known_aggregate_parts): Add arg_type parameter; use it
5938 instead of determining the type from pointer type.
5939 (ipa_compute_jump_functions_for_edge): Update call of
5940 determine_known_aggregate_parts.
5941 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
5942 (gimple_get_virt_method_for_binfo): ... here; simplify using
5943 vtable_pointer_value_to_vtable.
5944 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
5945 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
5946 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
5947 (vtable_pointer_value_to_vtable): Break out from ...; handle also
5949 (vtable_pointer_value_to_binfo): ... here.
5950 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
5952 2014-02-03 Teresa Johnson <tejohnson@google.com>
5954 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
5955 redef of outer loop index variable.
5957 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
5961 * doc/extend.texi (Function Attributes): Typo.
5963 2014-02-03 Cong Hou <congh@google.com>
5965 PR tree-optimization/60000
5966 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
5967 if the vectorized statement is a store. A store statement can only
5968 appear at the end of pattern statements.
5970 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5972 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
5973 (ix86_option_override_internal): Default long double to 64-bit for
5974 32-bit Bionic and to 128-bit for 64-bit Bionic.
5976 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
5977 TARGET_LONG_DOUBLE_128 is true.
5978 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
5980 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
5981 (mlong-double-64): Negate -mlong-double-128.
5982 (mlong-double-128): New option.
5984 * config/i386/i386-c.c (ix86_target_macros): Define
5985 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
5987 * doc/invoke.texi: Document -mlong-double-128.
5989 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5991 PR rtl-optimization/60024
5992 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
5994 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
5996 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
5998 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6000 PR rtl-optimization/57662
6001 * sel-sched.c (code_motion_path_driver): Do not mark already not
6002 existing blocks in the visiting bitmap.
6004 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6006 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
6007 on the insn being emitted.
6009 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
6010 Will Deacon <will.deacon@arm.com>
6012 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
6014 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6016 * config/arm/arm-tables.opt: Regenerate.
6018 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6020 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
6021 for vector types other than V16QImode.
6022 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
6023 define_expand, and call altivec_expand_vec_perm_le when producing
6024 code with little endian element order.
6025 (*altivec_vperm_<mode>_internal): New insn having previous
6026 behavior of altivec_vperm_<mode>.
6027 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
6028 altivec_expand_vec_perm_le when producing code with little endian
6030 (*altivec_vperm_<mode>_uns_internal): New insn having previous
6031 behavior of altivec_vperm_<mode>_uns.
6033 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6035 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
6036 (altivec_vsumsws): Add handling for -maltivec=be with a little
6038 (altivec_vsumsws_direct): New.
6039 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
6040 gen_altivec_vsumsws.
6042 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6044 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
6045 vtable_pointer_value_to_binfo): New functions.
6046 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
6047 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
6049 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
6051 * config/nios2/nios2.md (load_got_register): Initialize GOT
6052 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
6053 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
6055 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6057 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
6058 preserverd by passthrough, do not propagate the type.
6060 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6062 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
6063 (mips_atomic_assign_expand_fenv): New function.
6064 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
6066 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6068 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6069 (__builtin_mips_set_fcsr): Likewise.
6070 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6071 MIPS_USI_FTYPE_VOID.
6072 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6073 (mips16_expand_set_fcsr): Likewise.
6074 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6075 (mips16_set_fcsr_stub): Likewise.
6076 (mips16_get_fcsr_one_only_stub): New class.
6077 (mips16_set_fcsr_one_only_stub): Likewise.
6078 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6079 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6080 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6081 (hard_float): New availability predicate.
6082 (mips_builtins): Add get_fcsr and set_fcsr.
6083 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6084 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6085 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6086 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6087 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6090 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6092 * config/mips/mips.c (mips_one_only_stub): New class.
6093 (mips_need_mips16_rdhwr_p): Replace with...
6094 (mips16_rdhwr_stub): ...this new variable.
6095 (mips16_stub_call_address): New function.
6096 (mips16_rdhwr_one_only_stub): New class.
6097 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6098 (mips_output_mips16_rdhwr): Delete.
6099 (mips_finish_stub): New function.
6100 (mips_code_end): Use it to handle rdhwr stubs.
6102 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6105 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6106 when calculating size of integer atomic types.
6108 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6110 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6112 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6114 PR tree-optimization/60003
6115 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6116 * profile.c (branch_prob): Use gimple_call_builtin_p
6117 to check for BUILT_IN_SETJMP_RECEIVER.
6118 * tree-inline.c (copy_bb): Call notice_special_calls.
6120 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6123 * lra-constraints.c (process_alt_operands): Update reload_sum only
6126 2014-01-31 Richard Henderson <rth@redhat.com>
6129 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6130 until after else_eh is processed.
6132 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6134 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6135 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6136 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6137 in smmintrin.h, remove them.
6138 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6139 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6140 * config/i386/i386.md (ROUND_SAE): Fix value.
6141 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6142 (const48_operand): New.
6143 * config/i386/subst.md (round), (round_expand): Use
6144 const_4_or_8_to_11_operand.
6145 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6147 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6149 * config/i386/constraints.md (Yk): Swap meaning with k.
6150 * config/i386/i386.md (movhi_internal): Change Yk to k.
6151 (movqi_internal): Ditto.
6152 (*k<logic><mode>): Ditto.
6155 (kandn<mode>): Ditto.
6156 (*<code>hi_1): Ditto.
6157 (*<code>qi_1): Ditto.
6158 (kxnor<mode>): Ditto.
6159 (kortestzhi): Ditto.
6160 (kortestchi): Ditto.
6162 (*one_cmplhi2_1): Ditto.
6163 (*one_cmplqi2_1): Ditto.
6164 * config/i386/sse.md (): Change k to Yk.
6165 (avx512f_load<mode>_mask): Ditto.
6166 (avx512f_blendm<mode>): Ditto.
6167 (avx512f_store<mode>_mask): Ditto.
6168 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6169 (avx512f_storedqu<mode>_mask): Ditto.
6170 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6172 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6173 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6174 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6175 (avx512f_maskcmp<mode>3): Ditto.
6176 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6177 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6178 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6179 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6180 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6181 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6182 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6183 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6184 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6185 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6186 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6187 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6188 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6189 (vec_extract_lo_<mode>_maskm): Ditto.
6190 (vec_extract_hi_<mode>_maskm): Ditto.
6191 (avx512f_vternlog<mode>_mask): Ditto.
6192 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6193 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6194 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6195 (avx512f_<code>v8div16qi2_mask): Ditto.
6196 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6197 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6198 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6199 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6200 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6201 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6202 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6203 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6204 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6205 (avx512cd_maskb_vec_dupv8di): Ditto.
6206 (avx512cd_maskw_vec_dupv16si): Ditto.
6207 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6208 (avx512f_vpermi2var<mode>3_mask): Ditto.
6209 (avx512f_vpermi2var<mode>3_mask): Ditto.
6210 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6211 (*avx512f_gathersi<mode>): Ditto.
6212 (*avx512f_gathersi<mode>_2): Ditto.
6213 (*avx512f_gatherdi<mode>): Ditto.
6214 (*avx512f_gatherdi<mode>_2): Ditto.
6215 (*avx512f_scattersi<mode>): Ditto.
6216 (*avx512f_scatterdi<mode>): Ditto.
6217 (avx512f_compress<mode>_mask): Ditto.
6218 (avx512f_compressstore<mode>_mask): Ditto.
6219 (avx512f_expand<mode>_mask): Ditto.
6220 * config/i386/subst.md (mask): Change k to Yk.
6221 (mask_scalar_merge): Ditto.
6224 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6226 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6228 2014-01-31 Richard Biener <rguenther@suse.de>
6231 * builtins.c (fold_builtin_memory_op): Make sure to not
6232 use a floating-point mode or a boolean or enumeral type for
6235 2014-01-30 DJ Delorie <dj@redhat.com>
6237 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6238 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6239 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6240 whenever main() has an epilogue.
6242 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6244 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6245 unused variable "field".
6246 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6247 (vsx_mergeh_<mode>): Likewise.
6248 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6249 (altivec_vmrghh): Likewise.
6250 (altivec_vmrghw): Likewise.
6251 (altivec_vmrglb): Likewise.
6252 (altivec_vmrglh): Likewise.
6253 (altivec_vmrglw): Likewise.
6254 (altivec_vspltb): Add missing uses.
6255 (altivec_vsplth): Likewise.
6256 (altivec_vspltw): Likewise.
6257 (altivec_vspltsf): Likewise.
6259 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6262 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6263 frame related instructions.
6265 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6267 PR rtl-optimization/59959
6268 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6269 any reload of register whose subreg is invalid.
6271 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6273 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6274 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6275 Add missing return type - void.
6277 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6279 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6280 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6281 remove element index adjustment for endian (now handled in vsx.md
6283 (altivec_expand_vec_perm_const): Use
6284 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6285 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6286 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6287 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6288 define_expand and a new define_insn *altivec_vspltb_internal;
6289 adjust for -maltivec=be on a little endian target.
6290 (altivec_vspltb_direct): New.
6291 (altivec_vsplth): Divide into a define_expand and a new
6292 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6293 little endian target.
6294 (altivec_vsplth_direct): New.
6295 (altivec_vspltw): Divide into a define_expand and a new
6296 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6297 little endian target.
6298 (altivec_vspltw_direct): New.
6299 (altivec_vspltsf): Divide into a define_expand and a new
6300 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6301 a little endian target.
6303 2014-01-30 Richard Biener <rguenther@suse.de>
6305 PR tree-optimization/59993
6306 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6307 can propagate form the earlier stmt and avoid the transform
6308 when the intermediate result is needed.
6310 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6312 * README.Portability: Fix typo.
6314 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6316 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6317 comparison_operator with ordered_comparison_operator.
6319 2014-01-30 Nick Clifton <nickc@redhat.com>
6321 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6322 Rename to mn10300_store_multiple_regs.
6323 * config/mn10300/mn10300.c: Likewise.
6324 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6325 store_multiple_regs.
6326 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6327 Call mn10300_store_multiple_regs.
6329 2014-01-30 Nick Clifton <nickc@redhat.com>
6330 DJ Delorie <dj@redhat.com>
6332 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6333 %fp 2 to keep registers after it properly word-aligned.
6334 (rl78_alloc_physical_registers_umul): Handle the case where both
6335 input operands are the same.
6337 2014-01-30 Richard Biener <rguenther@suse.de>
6339 PR tree-optimization/59903
6340 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6343 2014-01-30 Jason Merrill <jason@redhat.com>
6346 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6349 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6351 2014-01-30 Richard Biener <rguenther@suse.de>
6353 PR tree-optimization/59951
6354 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6356 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6359 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6360 SFmode to DFmode case.
6362 2014-01-29 DJ Delorie <dj@redhat.com>
6364 * config/msp430/msp430.opt (-minrt): New.
6365 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6367 (ENDFILE_SPEC): Likewise.
6369 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6371 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6372 (estimate_function_body_sizes): Use it.
6374 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6377 * dwarf2out.c (is_cxx_auto): New.
6378 (is_base_type): Use it.
6379 (gen_type_die_with_usage): Likewise.
6381 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6383 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6384 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6385 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6386 -maltivec=be with LE targets.
6387 (vsx_mergeh_<mode>): Likewise.
6388 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6389 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6390 (altivec_vmrghb): Replace with define_expand and new
6391 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6392 (altivec_vmrghb_direct): New define_insn.
6393 (altivec_vmrghh): Replace with define_expand and new
6394 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6395 (altivec_vmrghh_direct): New define_insn.
6396 (altivec_vmrghw): Replace with define_expand and new
6397 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6398 (altivec_vmrghw_direct): New define_insn.
6399 (*altivec_vmrghsf): Adjust for endianness.
6400 (altivec_vmrglb): Replace with define_expand and new
6401 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6402 (altivec_vmrglb_direct): New define_insn.
6403 (altivec_vmrglh): Replace with define_expand and new
6404 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6405 (altivec_vmrglh_direct): New define_insn.
6406 (altivec_vmrglw): Replace with define_expand and new
6407 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6408 (altivec_vmrglw_direct): New define_insn.
6409 (*altivec_vmrglsf): Adjust for endianness.
6410 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6411 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6412 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6413 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6414 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6415 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6416 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6417 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6419 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6421 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6422 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6425 2014-01-29 Richard Biener <rguenther@suse.de>
6427 PR tree-optimization/58742
6428 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6429 associate_pointerplus_align.
6430 (associate_pointerplus_diff): New function.
6431 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6432 and associate_pointerplus_diff.
6434 2014-01-29 Richard Biener <rguenther@suse.de>
6436 * lto-streamer.h (LTO_major_version): Bump to 3.
6437 (LTO_minor_version): Reset to 0.
6439 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6441 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6442 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6443 (arm_file_start): Generate correct asm header for armv7ve.
6444 * config/arm/bpabi.h: Add multilib support for armv7ve.
6445 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6446 and cortex-a15 to armv7ve.
6447 * config/arm/t-aprofile: Add multilib support for armv7ve.
6448 * doc/invoke.texi: Document -march=armv7ve.
6450 2014-01-29 Richard Biener <rguenther@suse.de>
6452 PR tree-optimization/58742
6453 * tree-ssa-forwprop.c (associate_plusminus): Return true
6454 if we changed sth, defer EH cleanup to ...
6455 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6456 (simplify_mult): New function.
6458 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6461 PR tree-optimization/59920
6462 * tree.c (build_common_builtin_nodes): Remove
6463 __builtin_setjmp_dispatcher initialization.
6464 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6465 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6466 instead of gsi_after_labels + manually skipping debug stmts.
6467 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6468 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6469 * tree-inline.c (copy_edges_for_bb): Remove
6470 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6471 argument. Ignore computed_goto_p stmts. Don't call
6472 make_abnormal_goto_edges. If a call might need abnormal edges
6473 for non-local gotos, see if it already has an edge to
6474 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6475 with true argument, don't do anything then, otherwise add
6476 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6477 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6479 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6480 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6481 (lower_stmt): Don't set data->calls_builtin_setjmp.
6482 (lower_builtin_setjmp): Adjust comment.
6483 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6484 * tree-cfg.c (found_computed_goto): Remove.
6485 (factor_computed_gotos): Remove.
6486 (make_goto_expr_edges): Return bool, true for computed gotos.
6487 Don't call make_abnormal_goto_edges.
6488 (build_gimple_cfg): Don't set found_computed_goto, don't call
6489 factor_computed_gotos.
6490 (computed_goto_p): No longer static.
6491 (make_blocks): Don't set found_computed_goto.
6492 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6493 (make_edges): If make_goto_expr_edges returns true, push bb
6494 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6495 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6496 vector. Record mapping between bbs and OpenMP regions if there
6497 are any, adjust make_gimple_omp_edges caller. Call
6498 handle_abnormal_edges.
6499 (make_abnormal_goto_edges): Remove.
6500 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6501 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6502 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6503 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6504 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6505 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6506 filling *region also set *region_idx to (*region)->entry->index.
6509 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6510 For REGs set ORIGINAL_REGNO.
6512 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6514 * doc/md.texi: Mention that a target shouldn't implement
6515 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6518 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6520 PR tree-optimization/59594
6521 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6522 a copy of the datarefs vector rather than the vector itself.
6524 2014-01-28 Jason Merrill <jason@redhat.com>
6527 * dwarf2out.c (auto_die): New static.
6528 (gen_type_die_with_usage): Handle C++1y 'auto'.
6529 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6532 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6535 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6536 (SPEC_X32): Likewise.
6537 (SPEC_64): Likewise.
6538 * config/i386/i386.c (ix86_option_override_internal): Turn off
6539 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6541 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6542 * config/i386/i386.h (TARGET_16BIT): New macro.
6543 (TARGET_16BIT_P): Likewise.
6544 * config/i386/i386.opt: Add m16.
6545 * doc/invoke.texi: Document -m16.
6547 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6549 PR preprocessor/59935
6550 * input.c (location_get_source_line): Bail out on when line number
6551 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6553 2014-01-28 Richard Biener <rguenther@suse.de>
6555 PR tree-optimization/58742
6556 * tree-ssa-forwprop.c (associate_plusminus): Handle
6557 pointer subtraction of the form (T)(P + A) - (T)P.
6559 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6561 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6564 2014-01-28 Richard Biener <rguenther@suse.de>
6567 2014-01-28 Richard Biener <rguenther@suse.de>
6569 PR rtl-optimization/45364
6570 PR rtl-optimization/59890
6571 * var-tracking.c (local_get_addr_clear_given_value): Handle
6572 already cleared slot.
6573 (val_reset): Handle not allocated local_get_addr_cache.
6574 (vt_find_locations): Use post-order on the inverted CFG.
6576 2014-01-28 Richard Biener <rguenther@suse.de>
6578 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6580 2014-01-28 Richard Biener <rguenther@suse.de>
6582 PR rtl-optimization/45364
6583 PR rtl-optimization/59890
6584 * var-tracking.c (local_get_addr_clear_given_value): Handle
6585 already cleared slot.
6586 (val_reset): Handle not allocated local_get_addr_cache.
6587 (vt_find_locations): Use post-order on the inverted CFG.
6589 2014-01-28 Alan Modra <amodra@gmail.com>
6591 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6592 * configure.ac <recursive call for build != host>: Define
6593 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6594 and LD_FOR_BUILD too.
6595 * configure: Regenerate.
6597 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6599 * config/i386/i386.c (get_builtin_code_for_version): Separate
6600 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6601 Broadwell from Haswell.
6603 2014-01-27 Steve Ellcey <sellcey@mips.com>
6605 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6606 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6607 * config/mips/mips.c (mips_option_override): Change setting
6609 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6610 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6611 Change from Mask to Var.
6613 2014-01-27 Jeff Law <law@redhat.com>
6615 * ipa-inline.c (inline_small_functions): Fix typo.
6617 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6619 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6620 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6621 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6622 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6623 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6624 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6625 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6626 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6627 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6628 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6629 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6630 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6631 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6632 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6633 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6634 (_mm512_storeu_epi64): Ditto.
6635 (_mm512_cmpge_epi32_mask): Ditto.
6636 (_mm512_cmpge_epu32_mask): Ditto.
6637 (_mm512_cmpge_epi64_mask): Ditto.
6638 (_mm512_cmpge_epu64_mask): Ditto.
6639 (_mm512_cmple_epi32_mask): Ditto.
6640 (_mm512_cmple_epu32_mask): Ditto.
6641 (_mm512_cmple_epi64_mask): Ditto.
6642 (_mm512_cmple_epu64_mask): Ditto.
6643 (_mm512_cmplt_epi32_mask): Ditto.
6644 (_mm512_cmplt_epu32_mask): Ditto.
6645 (_mm512_cmplt_epi64_mask): Ditto.
6646 (_mm512_cmplt_epu64_mask): Ditto.
6647 (_mm512_cmpneq_epi32_mask): Ditto.
6648 (_mm512_cmpneq_epu32_mask): Ditto.
6649 (_mm512_cmpneq_epi64_mask): Ditto.
6650 (_mm512_cmpneq_epu64_mask): Ditto.
6651 (_mm512_expand_pd): Ditto.
6652 (_mm512_expand_ps): Ditto.
6653 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6654 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6655 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6656 * config/i386/i386.c (ix86_builtins): Add
6657 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6658 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6659 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6660 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6661 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6662 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6663 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6664 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6665 IX86_BUILTIN_PMOVUSQW512_MEM.
6666 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6667 __builtin_ia32_pmovsqd512mem_mask,
6668 __builtin_ia32_pmovqd512mem_mask,
6669 __builtin_ia32_pmovusqw512mem_mask,
6670 __builtin_ia32_pmovsqw512mem_mask,
6671 __builtin_ia32_pmovqw512mem_mask,
6672 __builtin_ia32_pmovusdw512mem_mask,
6673 __builtin_ia32_pmovsdw512mem_mask,
6674 __builtin_ia32_pmovdw512mem_mask,
6675 __builtin_ia32_pmovqb512mem_mask,
6676 __builtin_ia32_pmovusqb512mem_mask,
6677 __builtin_ia32_pmovsqb512mem_mask,
6678 __builtin_ia32_pmovusdb512mem_mask,
6679 __builtin_ia32_pmovsdb512mem_mask,
6680 __builtin_ia32_pmovdb512mem_mask.
6681 (bdesc_args): Add __builtin_ia32_expanddf512,
6682 __builtin_ia32_expandsf512.
6683 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6684 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6685 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6686 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6687 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6688 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6689 (avx512f_<code>v8div16qi2_mask_store): This.
6690 (avx512f_expand<mode>): New.
6692 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6694 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
6696 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6697 (_mm512_prefetch_i32scatter_pd): Ditto.
6698 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6699 (_mm512_prefetch_i64scatter_pd): Ditto.
6700 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6701 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6702 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6703 (_mm512_prefetch_i32scatter_ps): Ditto.
6704 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6705 (_mm512_prefetch_i64scatter_ps): Ditto.
6706 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6707 * config/i386/i386-builtin-types.def: Define
6708 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6709 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
6710 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6711 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6712 IX86_BUILTIN_SCATTERPFQPD.
6713 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6714 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6715 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6716 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6717 __builtin_ia32_scatterpfqps.
6718 (ix86_expand_builtin): Expand new built-ins.
6719 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6720 fix memory access data type.
6721 (*avx512pf_gatherpf<mode>_mask): Ditto.
6722 (*avx512pf_gatherpf<mode>): Ditto.
6723 (avx512pf_scatterpf<mode>): Ditto.
6724 (*avx512pf_scatterpf<mode>_mask): Ditto.
6725 (*avx512pf_scatterpf<mode>): Ditto.
6726 (GATHER_SCATTER_SF_MEM_MODE): New.
6727 (avx512pf_gatherpf<mode>df): Ditto.
6728 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6729 (*avx512pf_scatterpf<mode>df): Ditto.
6731 2014-01-27 Jakub Jelinek <jakub@redhat.com>
6734 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
6735 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
6738 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6740 * common/config/arm/arm-common.c
6741 (arm_rewrite_mcpu): Handle multiple names.
6743 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6745 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6747 * gimple-builder.h (create_gimple_tmp): Delete.
6749 2014-01-27 Christian Bruel <christian.bruel@st.com>
6751 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
6754 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
6756 * config/pa/pa.md (call): Generate indirect long calls to non-local
6757 functions when outputing 32-bit code.
6758 (call_value): Likewise except for special call to buggy powf function.
6760 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
6761 portable runtime and PIC indirect calls.
6762 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
6763 and PIC call sequences. Use ldo instead of blr to set return register
6764 in PIC call sequence.
6766 2014-01-25 Walter Lee <walt@tilera.com>
6768 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
6769 avoid clobbering a live register.
6771 2014-01-25 Walter Lee <walt@tilera.com>
6773 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
6774 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
6775 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
6776 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
6778 2014-01-25 Walter Lee <walt@tilera.com>
6780 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
6781 arguments on even registers.
6782 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
6784 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
6785 (BIGGEST_ALIGNMENT): Ditto.
6786 (BIGGEST_FIELD_ALIGNMENT): Ditto.
6788 2014-01-25 Walter Lee <walt@tilera.com>
6790 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
6791 insns before bundling.
6792 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
6794 2014-01-25 Walter Lee <walt@tilera.com>
6796 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
6797 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
6798 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
6800 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6802 * config/mips/constraints.md (kl): Delete.
6803 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
6804 define expands, using...
6805 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
6806 instructions for MIPS16.
6807 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
6808 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
6810 2014-01-25 Walter Lee <walt@tilera.com>
6812 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
6816 2014-01-25 Walter Lee <walt@tilera.com>
6818 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
6819 (TARGET_EXPAND_TO_RTL_HOOK): Define.
6821 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6823 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
6826 2014-01-25 Jakub Jelinek <jakub@redhat.com>
6828 * print-rtl.c (in_call_function_usage): New var.
6829 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
6830 EXPR_LIST mode as mode and not as reg note name.
6833 * cfgloopmanip.c (copy_loop_info): If
6834 loop->warned_aggressive_loop_optimizations, make sure
6835 the flag is set in target loop too.
6837 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
6839 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
6841 * builtins.def: Likewise.
6842 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
6843 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
6844 * ira.c (ira_setup_eliminable_regset): Likewise.
6845 * omp-low.c (gate_expand_omp): Likewise.
6846 (execute_lower_omp): Likewise.
6847 (diagnose_sb_0): Likewise.
6848 (gate_diagnose_omp_blocks): Likewise.
6849 (simd_clone_clauses_extract): Likewise.
6852 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6854 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
6855 correction for little endian...
6856 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
6859 2014-01-24 Jeff Law <law@redhat.com>
6861 PR tree-optimization/59919
6862 * tree-vrp.c (find_assert_locations_1): Do not register asserts
6863 for non-returning calls.
6865 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
6867 * common/config/aarch64/aarch64-common.c
6868 (aarch64_rewrite_mcpu): Handle multiple names.
6869 * config/aarch64/aarch64.h
6870 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6872 2014-01-24 Dodji Seketeli <dodji@redhat.com>
6874 * input.c (add_file_to_cache_tab): Handle the case where fopen
6877 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
6880 * config/i386/i386.md (pushsf splitter): Get stack adjustment
6881 from push operand if code of push isn't PRE_DEC.
6883 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6886 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
6887 -mquad-memory-atomic. Update -mquad-memory documentation to say
6888 it is only used for non-atomic loads/stores.
6890 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
6891 -mquad-memory or -mquad-memory-atomic switches.
6893 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
6894 -mquad-memory-atomic to ISA 2.07 support.
6896 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
6897 to separate support of normal quad word memory operations (ldq, stq)
6898 from the atomic quad word memory operations.
6900 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
6901 support to separate non-atomic quad word operations from atomic
6902 quad word operations. Disable non-atomic quad word operations in
6903 little endian mode so that we don't have to swap words after the
6904 load and before the store.
6905 (quad_load_store_p): Add comment about atomic quad word support.
6906 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
6907 options printed with -mdebug=reg.
6909 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
6910 -mquad-memory-atomic as the test for whether we have quad word
6911 atomic instructions.
6912 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
6913 or -mp8-vector are used, allow byte/half-word atomic operations.
6915 * config/rs6000/sync.md (load_lockedti): Insure that the address
6916 is a proper indexed or indirect address for the lqarx instruction.
6917 On little endian systems, swap the hi/lo registers after the lqarx
6919 (load_lockedpti): Use indexed_or_indirect_operand predicate to
6920 insure the address is valid for the lqarx instruction.
6921 (store_conditionalti): Insure that the address is a proper indexed
6922 or indirect address for the stqcrx. instruction. On little endian
6923 systems, swap the hi/lo registers before doing the stqcrx.
6925 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
6926 insure the address is valid for the stqcrx. instruction.
6928 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
6929 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
6930 type of quad memory support is available.
6932 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
6935 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
6936 there is a danger of looping.
6938 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
6940 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
6941 force flag_ira_loop_pressure if set via command line.
6943 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
6945 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
6946 (ashr_simd): New builtin handling DI mode.
6947 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
6948 (aarch64_sshr_simddi): New match pattern.
6949 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
6950 (vshrd_n_s64): Likewise.
6951 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
6953 2014-01-23 Nick Clifton <nickc@redhat.com>
6955 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
6956 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
6957 favour of mcu specific scripts.
6958 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
6961 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
6962 Alex Velenko <Alex.Velenko@arm.com>
6964 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
6965 (vaddv_s16): Likewise.
6966 (vaddv_s32): Likewise.
6967 (vaddv_u8): Likewise.
6968 (vaddv_u16): Likewise.
6969 (vaddv_u32): Likewise.
6970 (vaddvq_s8): Likewise.
6971 (vaddvq_s16): Likewise.
6972 (vaddvq_s32): Likewise.
6973 (vaddvq_s64): Likewise.
6974 (vaddvq_u8): Likewise.
6975 (vaddvq_u16): Likewise.
6976 (vaddvq_u32): Likewise.
6977 (vaddvq_u64): Likewise.
6978 (vaddv_f32): Likewise.
6979 (vaddvq_f32): Likewise.
6980 (vaddvq_f64): Likewise.
6981 (vmaxv_f32): Likewise.
6982 (vmaxv_s8): Likewise.
6983 (vmaxv_s16): Likewise.
6984 (vmaxv_s32): Likewise.
6985 (vmaxv_u8): Likewise.
6986 (vmaxv_u16): Likewise.
6987 (vmaxv_u32): Likewise.
6988 (vmaxvq_f32): Likewise.
6989 (vmaxvq_f64): Likewise.
6990 (vmaxvq_s8): Likewise.
6991 (vmaxvq_s16): Likewise.
6992 (vmaxvq_s32): Likewise.
6993 (vmaxvq_u8): Likewise.
6994 (vmaxvq_u16): Likewise.
6995 (vmaxvq_u32): Likewise.
6996 (vmaxnmv_f32): Likewise.
6997 (vmaxnmvq_f32): Likewise.
6998 (vmaxnmvq_f64): Likewise.
6999 (vminv_f32): Likewise.
7000 (vminv_s8): Likewise.
7001 (vminv_s16): Likewise.
7002 (vminv_s32): Likewise.
7003 (vminv_u8): Likewise.
7004 (vminv_u16): Likewise.
7005 (vminv_u32): Likewise.
7006 (vminvq_f32): Likewise.
7007 (vminvq_f64): Likewise.
7008 (vminvq_s8): Likewise.
7009 (vminvq_s16): Likewise.
7010 (vminvq_s32): Likewise.
7011 (vminvq_u8): Likewise.
7012 (vminvq_u16): Likewise.
7013 (vminvq_u32): Likewise.
7014 (vminnmv_f32): Likewise.
7015 (vminnmvq_f32): Likewise.
7016 (vminnmvq_f64): Likewise.
7018 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7020 * config/aarch64/aarch64-simd.md
7021 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
7022 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
7023 (*aarch64_mul3_elt<mode>): Likewise.
7024 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
7025 (*aarch64_mul3_elt_to_64v2df): Likewise.
7026 (*aarch64_mla_elt<mode>): Likewise.
7027 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7028 (*aarch64_mls_elt<mode>): Likewise.
7029 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7030 (*aarch64_fma4_elt<mode>): Likewise.
7031 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7032 (*aarch64_fma4_elt_to_64v2df): Likewise.
7033 (*aarch64_fnma4_elt<mode>): Likewise.
7034 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7035 (*aarch64_fnma4_elt_to_64v2df): Likewise.
7036 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7037 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
7038 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7039 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7040 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
7041 (aarch64_sqdmull_lane<mode>_internal): Likewise.
7042 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
7044 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
7046 * config/aarch64/aarch64-simd.md
7047 (aarch64_be_checked_get_lane<mode>): New define_expand.
7048 * config/aarch64/aarch64-simd-builtins.def
7049 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
7050 New builtin definition.
7051 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
7052 Use new safe be builtin.
7054 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7056 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
7058 (aarch64_be_st1<mode>): Likewise.
7059 (aarch_ld1<VALL:mode>): Define_expand modified.
7060 (aarch_st1<VALL:mode>): Likewise.
7061 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
7062 (UNSPEC_ST1): Likewise.
7064 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
7066 * config/microblaze/microblaze.md: Add trap insn and attribute
7068 2014-01-23 Dodji Seketeli <dodji@redhat.com>
7070 PR preprocessor/58580
7071 * input.h (location_get_source_line): Take an additional line_size
7073 (void diagnostics_file_cache_fini): Declare new function.
7074 * input.c (struct fcache): New type.
7075 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7076 New static constants.
7077 (diagnostic_file_cache_init, total_lines_num)
7078 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7079 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7080 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7081 (get_next_line, read_next_line, goto_next_line, read_line_num):
7082 New static function definitions.
7083 (diagnostic_file_cache_fini): New function.
7084 (location_get_source_line): Take an additional output line_len
7085 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7087 * diagnostic.c (diagnostic_finish): Call
7088 diagnostic_file_cache_fini.
7089 (adjust_line): Take an additional input parameter for the length
7090 of the line, rather than calculating it with strlen.
7091 (diagnostic_show_locus): Adjust the use of
7092 location_get_source_line and adjust_line with respect to their new
7093 signature. While displaying a line now, do not stop at the first
7094 null byte. Rather, display the zero byte as a space and keep
7095 going until we reach the size of the line.
7096 * Makefile.in: Add vec.o to OBJS-libcommon
7098 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7099 Ilya Tocar <ilya.tocar@intel.com>
7101 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7102 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7103 (__builtin_ia32_kmov16): Ditto.
7104 * config/i386/i386.md (UNSPEC_KMOV): New.
7107 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7109 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7110 (_mm512_storeu_si512): Ditto.
7112 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7115 * rtl.h (get_referenced_operands): Declare.
7116 * recog.c (get_referenced_operands): New function.
7117 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7118 operands have been referenced when recording LO_SUM references.
7120 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7122 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7124 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7126 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7127 Enable for generic and recent AMD targets.
7129 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7131 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7132 ARG_SIZE note when adjustment was eliminated.
7134 2014-01-22 Jeff Law <law@redhat.com>
7136 PR tree-optimization/59597
7137 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7138 in file. Accept new argument REGISTERING and use it to modify
7139 dump output appropriately.
7140 (register_jump_thread): Corresponding changes.
7141 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7142 thread paths involving joiner blocks. Add code to dump cancelled
7143 jump threading paths.
7145 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7147 PR rtl-optimization/59477
7148 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7149 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7151 2014-01-22 Tom Tromey <tromey@redhat.com>
7153 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7155 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7157 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7159 PR rtl-optimization/59896
7160 * lra-constraints.c (process_alt_operands): Check unused note for
7161 matched operands of insn with no output reloads.
7163 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7165 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7166 (mips_move_from_gpr_cost): Likewise.
7168 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7170 PR rtl-optimization/59858
7171 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7172 ira_class_hard_regs_num.
7173 (process_alt_operands): Increase reject for dying matched operand.
7175 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7178 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7179 smaller than size, perform several stores or loads and stores
7180 at dst + count - size to store or copy all of size bytes, rather
7181 than just last modesize bytes.
7183 2014-01-20 DJ Delorie <dj@redhat.com>
7185 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7186 that CLOBBERs are REGs before propogating their values.
7188 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7191 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7192 (cgraph_inline_failed_type): New function.
7193 * cgraph.h (DEFCIFCODE): Add type.
7194 (cgraph_inline_failed_type_t): New enum.
7195 (cgraph_inline_failed_type): New prototype.
7196 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7197 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7198 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7199 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7200 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7201 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7202 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7203 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7204 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7205 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7206 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7207 OPTIMIZATION_MISMATCH.
7208 * tree-inline.c (expand_call_inline): Emit errors during
7209 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7211 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7214 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7215 mode attribute in insn output.
7217 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7219 * output.h (output_constant): Delete.
7220 * varasm.c (output_constant): Make private.
7222 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7224 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7226 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7229 * tree.h (fold_builtin_strcat): New prototype.
7230 * builtins.c (fold_builtin_strcat): No longer static. Add len
7231 argument, if non-NULL, don't call c_strlen. Optimize
7232 directly into __builtin_memcpy instead of __builtin_strcpy.
7233 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7234 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7236 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7238 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7239 for SImode_address_operand operands, having only a REG argument.
7241 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7243 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7244 loader name using mbig-endian.
7245 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7247 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7249 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7253 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7255 * config/aarch64/aarch64-protos.h
7256 (aarch64_cannot_change_mode_class_ptr): Declare.
7257 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7258 aarch64_cannot_change_mode_class_ptr): New.
7259 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7260 backend hook aarch64_cannot_change_mode_class.
7262 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7264 * common/config/aarch64/aarch64-common.c
7265 (aarch64_handle_option): Don't handle any option order logic here.
7266 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7267 selected_cpu, warn on architecture version mismatch.
7268 (aarch64_override_options): Fix parsing order for option strings.
7270 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7271 Iain Sandoe <iain@codesourcery.com>
7274 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7275 warning. Amend comment to reflect current functionality.
7277 2014-01-20 Richard Biener <rguenther@suse.de>
7280 * builtins.c (fold_builtin_strcat): Remove case better handled
7281 by tree-ssa-strlen.c.
7283 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7285 * config/aarch64/aarch64.opt
7286 (mcpu, march, mtune): Make case-insensitive.
7288 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7291 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7292 if operands[1] is a REG or ZERO_EXTEND of a REG.
7294 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7296 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7298 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7300 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7301 long non-pic millicode calls.
7303 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7305 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7307 2014-01-19 Kito Cheng <kito@0xlab.org>
7309 * builtins.c (expand_movstr): Check movstr expand done or fail.
7311 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7312 H.J. Lu <hongjiu.lu@intel.com>
7315 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7316 to DImode for zero-extended addresses.
7318 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7320 PR rtl-optimization/57763
7321 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7322 on the new indirect jump_insn and increment LABEL_NUSES (label).
7324 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7328 * config.gcc (x86_archs): New variable.
7329 (x86_64_archs): Likewise.
7330 (x86_cpus): Likewise.
7331 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7332 --with-arch/--with-cpu= options.
7333 Support --with-arch=/--with-cpu={nehalem,westmere,
7334 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7336 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7338 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7339 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7341 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7343 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7345 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7348 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7349 clear cpp_get_options (parse_in)->warn_unused_macros for
7350 ix86_target_macros_internal with cpp_define.
7352 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7354 * jump.c (delete_related_insns): Keep (use (insn))s.
7355 * reorg.c (redundant_insn): Check for barriers too.
7357 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7359 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7361 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7363 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7364 call to $$dyncall when TARGET_LONG_CALLS is true.
7366 2014-01-17 Jeff Law <law@redhat.com>
7368 * ree.c (combine_set_extension): Temporarily disable test for
7369 changing number of hard registers.
7371 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7374 * ipa-inline-analysis.c (inline_free_summary):
7375 Do not free summary of aliases.
7377 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7380 * gimplify.c (gimplify_expr): Use create_tmp_var
7381 instead of create_tmp_var_raw. If cond doesn't have
7382 integral type, don't add the IFN_ANNOTATE builtin at all.
7384 2014-01-17 Martin Jambor <mjambor@suse.cz>
7387 * ipa-cp.c (prev_edge_clone): New variable.
7388 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7389 Also resize prev_edge_clone vector.
7390 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7391 (ipcp_edge_removal_hook): New function.
7392 (ipcp_driver): Register ipcp_edge_removal_hook.
7394 2014-01-17 Andrew Pinski <apinski@cavium.com>
7395 Steve Ellcey <sellcey@mips.com>
7398 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7401 2014-01-17 Jeff Law <law@redhat.com>
7404 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7405 so that pass_ccp runs first.
7407 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7409 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7410 (ix86_adjust_cost): Use !TARGET_XXX.
7411 (do_reorder_for_imul): Likewise.
7412 (swap_top_of_ready_list): Likewise.
7413 (ix86_sched_reorder): Likewise.
7415 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7417 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7418 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7419 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7420 (intel_memset): New. Duplicate slm_memset.
7421 (intel_cost): New. Duplicate slm_cost.
7422 (m_INTEL): New macro.
7423 (processor_target_table): Add "intel".
7424 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7425 with PROCESSOR_INTEL for "intel".
7426 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7427 PROCESSOR_SILVERMONT.
7428 (ix86_issue_rate): Likewise.
7429 (ix86_adjust_cost): Likewise.
7430 (ia32_multipass_dfa_lookahead): Likewise.
7431 (swap_top_of_ready_list): Likewise.
7432 (ix86_sched_reorder): Likewise.
7433 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7434 instead of TARGET_OPT_AGU.
7435 * config/i386/i386.h (TARGET_INTEL): New.
7436 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7437 (processor_type): Add PROCESSOR_INTEL.
7438 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7439 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7441 2014-01-17 Marek Polacek <polacek@redhat.com>
7444 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7447 2014-01-17 Richard Biener <rguenther@suse.de>
7449 PR tree-optimization/46590
7450 * opts.c (default_options_table): Add entries for
7451 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7452 all enabled at -O1 but not for -Og.
7453 * common.opt (fbranch-count-reg): Remove Init(1).
7454 (fmove-loop-invariants): Likewise.
7455 (ftree-pta): Likewise.
7457 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7459 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7460 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7461 decls to at least the GCC 4.8 used alignments.
7464 * tree-nested.c (convert_nonlocal_reference_stmt,
7465 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7466 of GIMPLE_BIND stmts, adjust associated decls.
7468 2014-01-17 Richard Biener <rguenther@suse.de>
7470 PR tree-optimization/46590
7471 * vec.h (vec<>::bseach): New member function implementing
7472 binary search according to C89 bsearch.
7473 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7474 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7475 bitmap pointer again. Make accesses_in_loop a flat array.
7476 (mem_ref_obstack): New global.
7477 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7478 (mark_ref_stored): Likewise.
7479 (ref_indep_loop_p_2): Likewise.
7480 (set_ref_stored_in_loop): New helper function.
7481 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7482 (memref_free): Adjust.
7483 (record_mem_ref_loc): Simplify.
7484 (gather_mem_refs_stmt): Adjust.
7485 (sort_locs_in_loop_postorder_cmp): New function.
7486 (analyze_memory_references): Sort accesses_in_loop after
7487 loop postorder number.
7488 (find_ref_loc_in_loop_cmp): New function.
7489 (for_all_locs_in_loop): Find relevant cluster of locs in
7490 accesses_in_loop and iterate without recursion.
7491 (execute_sm): Avoid uninit warning.
7492 (struct ref_always_accessed): Simplify.
7493 (ref_always_accessed::operator ()): Likewise.
7494 (ref_always_accessed_p): Likewise.
7495 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7496 loop postorder numbers here.
7497 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7500 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7503 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7504 on decls for which assemble_alias has been called.
7506 2014-01-17 Nick Clifton <nickc@redhat.com>
7508 * config/msp430/msp430.opt: (mcpu): New option.
7509 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7510 (msp430_option_override): Parse target_cpu. If the MCU name
7511 matches a generic string, clear target_mcu.
7512 (msp430_attr): Allow numeric interrupt values up to 63.
7513 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7514 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7516 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7518 * config/msp430/msp430.md (popm): Use %J rather than %I.
7519 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7520 (addhi_cy_i): Use immediate_operand for operand 2.
7521 * doc/invoke.texi: Document -mcpu option.
7523 2014-01-17 Richard Biener <rguenther@suse.de>
7525 PR rtl-optimization/38518
7526 * df.h (df_analyze_loop): Declare.
7527 * df-core.c: Include cfgloop.h.
7528 (df_analyze_1): Split out main part of df_analyze.
7529 (df_analyze): Adjust.
7530 (loop_inverted_post_order_compute): New function.
7531 (loop_post_order_compute): Likewise.
7532 (df_analyze_loop): New function avoiding whole-function
7534 * loop-invariant.c (find_defs): Use df_analyze_loop.
7535 (find_invariants): Adjust.
7536 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7538 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7540 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7541 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7543 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7545 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7546 traversal when removing references.
7548 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7551 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7553 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7556 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7557 pushing a reload for an autoinc when we had previously reloaded an
7558 inner part of the address.
7560 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7562 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7564 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7565 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7566 when not giving up or versioning for alias only because of
7568 (vect_analyze_data_ref_dependences): Set to true.
7569 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7571 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7572 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7576 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7579 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7580 operand 0 predicate for gathers, use a new pseudo as subtarget.
7582 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7585 * lra-constraints.c (process_alt_operands): Add printing debug info.
7586 Check absence of input/output reloads for matched operands too.
7588 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7590 PR rtl-optimization/59835
7591 * ira.c (ira_init_register_move_cost): Increase cost for
7594 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7596 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7598 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7601 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7602 non-register objects. Use gen_(high/low)part more consistently.
7605 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7608 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7609 endian support, remove tests for WORDS_BIG_ENDIAN.
7610 (p8_mfvsrd_3_<mode>): Likewise.
7611 (reload_gpr_from_vsx<mode>): Likewise.
7612 (reload_gpr_from_vsxsf): Likewise.
7613 (p8_mfvsrd_4_disf): Likewise.
7615 2014-01-16 Richard Biener <rguenther@suse.de>
7617 PR rtl-optimization/46590
7618 * lcm.c (compute_antinout_edge): Use postorder iteration.
7619 (compute_laterin): Use inverted postorder iteration.
7621 2014-01-16 Nick Clifton <nickc@redhat.com>
7624 * varasm.c (output_constant): Return the number of bytes actually
7626 (output_constructor_array_range): Update the field size with the
7627 number of bytes emitted by output_constant.
7628 (output_constructor_regular_field): Likewise. Also do not
7629 complain if the total number of bytes emitted is now greater
7630 than the expected fieldpos.
7631 * output.h (output_constant): Update prototype and descriptive comment.
7633 2014-01-16 Marek Polacek <polacek@redhat.com>
7636 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7637 it is error_mark_node.
7639 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7641 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7642 VALID_AVX256_REG_OR_OI_MODE.
7644 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7646 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7647 current procedure should be profiled.
7649 2014-01-15 Andrew Pinski <apinski@cavium.com>
7651 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7652 of moving from/to the STACK_REG register class.
7654 2014-01-15 Richard Henderson <rth@redhat.com>
7657 * reginfo.c (global_regs_decl): Globalize.
7658 * rtl.h (global_regs_decl): Declare.
7659 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7660 reserved via global_regs.
7662 2014-01-15 Teresa Johnson <tejohnson@google.com>
7664 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7666 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7668 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7669 and vmulosh rather than call gen_vec_widen_smult_*.
7670 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7671 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7672 (vec_widen_smult_even_v16qi): Likewise.
7673 (vec_widen_umult_even_v8hi): Likewise.
7674 (vec_widen_smult_even_v8hi): Likewise.
7675 (vec_widen_umult_odd_v16qi): Likewise.
7676 (vec_widen_smult_odd_v16qi): Likewise.
7677 (vec_widen_umult_odd_v8hi): Likewise.
7678 (vec_widen_smult_odd_v8hi): Likewise.
7679 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7680 vmuloub rather than call gen_vec_widen_umult_*.
7681 (vec_widen_umult_lo_v16qi): Likewise.
7682 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7683 vmulosb rather than call gen_vec_widen_smult_*.
7684 (vec_widen_smult_lo_v16qi): Likewise.
7685 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7686 rather than call gen_vec_widen_umult_*.
7687 (vec_widen_umult_lo_v8hi): Likewise.
7688 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7689 rather than call gen_vec_widen_smult_*.
7690 (vec_widen_smult_lo_v8hi): Likewise.
7692 2014-01-15 Jeff Law <law@redhat.com>
7694 PR tree-optimization/59747
7695 * ree.c (find_and_remove_re): Properly handle case where a second
7696 eliminated extension requires widening a copy created for elimination
7697 of a prior extension.
7698 (combine_set_extension): Ensure that the number of hard regs needed
7699 for a destination register does not change when we widen it.
7701 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7703 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7704 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7705 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7706 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7707 (avr-*-rtems*): Likewise.
7708 (bfin*-rtems*): Likewise.
7709 (moxie-*-rtems*): Likewise.
7710 (h8300-*-rtems*): Likewise.
7711 (i[34567]86-*-rtems*): Likewise.
7712 (lm32-*-rtems*): Likewise.
7713 (m32r-*-rtems*): Likewise.
7714 (m68k-*-rtems*): Likewise.
7715 (microblaze*-*-rtems*): Likewise.
7716 (mips*-*-rtems*): Likewise.
7717 (powerpc-*-rtems*): Likewise.
7718 (sh-*-rtems*): Likewise.
7719 (sparc-*-rtems*): Likewise.
7720 (sparc64-*-rtems*): Likewise.
7721 (v850-*-rtems*): Likewise.
7722 (m32c-*-rtems*): Likewise.
7724 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7726 PR rtl-optimization/59511
7727 * ira.c (ira_init_register_move_cost): Use memory costs for some
7728 cases of register move cost calculations.
7729 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7730 instead of BB frequency.
7731 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7732 * lra-assigns.c (find_hard_regno_for): Ditto.
7734 2014-01-15 Richard Biener <rguenther@suse.de>
7736 PR tree-optimization/59822
7737 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
7738 (vectorizable_load): Use it to hoist defs of uses of invariant
7739 loads out of the loop.
7741 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7742 Kugan Vivekanandarajah <kuganv@linaro.org>
7745 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
7748 2014-01-15 Richard Biener <rguenther@suse.de>
7750 PR rtl-optimization/59802
7751 * lcm.c (compute_available): Use inverted postorder to seed
7752 the initial worklist.
7754 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7757 * config/s390/s390.c (s390_preferred_reload_class): Don't return
7758 ADDR_REGS for invalid symrefs in non-PIC code.
7760 2014-01-15 Jakub Jelinek <jakub@redhat.com>
7763 * builtins.c (determine_block_size): Initialize *probable_max_size
7764 even if len_rtx is CONST_INT.
7766 2014-01-14 Andrew Pinski <apinski@cavium.com>
7768 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
7769 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
7770 (cortexa53_tunings): Likewise.
7771 (aarch64_sched_issue_rate): New function.
7772 (TARGET_SCHED_ISSUE_RATE): Define.
7774 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7776 * ira-costs.c (find_costs_and_classes): Add missed
7777 ira_init_register_move_cost_if_necessary.
7779 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7782 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
7784 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
7787 * config/i386/i386.c (type_natural_mode): Add a bool parameter
7788 to indicate if type is used for function return value. Warn ABI
7789 change if the vector mode isn't available for function return value.
7790 (ix86_function_arg_advance): Pass false to type_natural_mode.
7791 (ix86_function_arg): Likewise.
7792 (ix86_gimplify_va_arg): Likewise.
7793 (function_arg_32): Don't warn ABI change.
7794 (ix86_function_value): Pass true to type_natural_mode.
7795 (ix86_return_in_memory): Likewise.
7796 (ix86_struct_value_rtx): Removed.
7797 (TARGET_STRUCT_VALUE_RTX): Likewise.
7799 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7801 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
7802 converting a conditional jump into a conditional return.
7804 2014-01-14 Richard Biener <rguenther@suse.de>
7806 PR tree-optimization/58921
7807 PR tree-optimization/59006
7808 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
7809 hoisting invariant stmts.
7810 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
7811 invariant loads on the preheader edge if possible.
7813 2014-01-14 Joey Ye <joey.ye@arm.com>
7815 * doc/plugin.texi (Building GCC plugins): Update to C++.
7817 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
7819 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
7820 (_mm_rcp28_round_ss): Ditto.
7821 (_mm_rsqrt28_round_sd): Ditto.
7822 (_mm_rsqrt28_round_ss): Ditto.
7823 (_mm_rcp28_sd): Ditto.
7824 (_mm_rcp28_ss): Ditto.
7825 (_mm_rsqrt28_sd): Ditto.
7826 (_mm_rsqrt28_ss): Ditto.
7827 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
7828 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
7829 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
7830 (IX86_BUILTIN_RCP28SD): Ditto.
7831 (IX86_BUILTIN_RCP28SS): Ditto.
7832 (IX86_BUILTIN_RSQRT28SD): Ditto.
7833 (IX86_BUILTIN_RSQRT28SS): Ditto.
7834 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
7835 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
7836 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
7837 (ix86_expand_special_args_builtin): Expand new FTYPE.
7838 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
7839 (srcp14<mode>): Make insn unary.
7840 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
7841 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
7842 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
7843 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
7844 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
7845 Fix rounding: make it SAE only.
7846 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7848 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7850 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
7851 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
7852 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
7853 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
7854 (round_saeonly_mask_scalar_operand4): Ditto.
7855 (round_saeonly_mask_scalar_op3): Ditto.
7856 (round_saeonly_mask_scalar_op4): Ditto.
7858 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7860 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7861 Implement -maltivec=be for vec_insert and vec_extract.
7863 2014-01-10 DJ Delorie <dj@redhat.com>
7865 * config/msp430/msp430.md (call_internal): Don't allow memory
7866 references with SP as the base register.
7867 (call_value_internal): Likewise.
7868 * config/msp430/constraints.md (Yc): New. For memory references
7869 that don't use SP as a base register.
7871 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
7872 "an integer without a # prefix"
7873 * config/msp430/msp430.md (epilogue_helper): Use it.
7875 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7878 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
7879 AVX512F gather builtins.
7880 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
7881 on gather decls with INTEGER_TYPE masktype.
7882 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
7883 directly into the builtin rather than hoisting it before loop.
7885 PR tree-optimization/59387
7886 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
7887 (scev_const_prop): If folded_casts and type has undefined overflow,
7888 use force_gimple_operand instead of force_gimple_operand_gsi and
7889 for each added stmt if it is assign with
7890 arith_code_with_undefined_signed_overflow, call
7891 rewrite_to_defined_overflow.
7892 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
7893 gimple-fold.h instead.
7894 (arith_code_with_undefined_signed_overflow,
7895 rewrite_to_defined_overflow): Moved to ...
7896 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
7897 rewrite_to_defined_overflow): ... here. No longer static.
7898 Include gimplify-me.h.
7899 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
7900 rewrite_to_defined_overflow): New prototypes.
7902 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7904 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
7906 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
7908 * builtins.c (get_object_alignment_2): Minor tweak.
7909 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
7911 2014-01-13 Christian Bruel <christian.bruel@st.com>
7913 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
7914 optimized non constant lengths.
7916 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7919 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
7920 load as __atomic_load_N if possible.
7922 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7924 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
7926 (rs6000_expand_builtin): Adjust call.
7928 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7931 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
7932 * config/rs6000/rs6000.c: Include target-globals.h.
7933 (rs6000_set_current_function): Instead of doing target_reinit
7934 unconditionally, use save_target_globals_default_opts and
7935 restore_target_globals.
7937 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
7939 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
7940 (rs6000_expand_builtin): Handle mffs and mtfsf.
7941 (rs6000_init_builtins): Define mffs and mtfsf.
7942 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
7943 (rs6000_mffs): New pattern.
7944 (rs6000_mtfsf): New pattern.
7946 2014-01-11 Bin Cheng <bin.cheng@arm.com>
7948 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
7949 Start narrowing with START. Apply candidate-use pair
7950 and check overall cost in narrowing.
7951 (iv_ca_prune): Pass new argument.
7953 2014-01-10 Jeff Law <law@redhat.com>
7956 * ree.c (combine_reaching_defs): Ensure the defining statement
7957 occurs before the extension when optimizing extensions with
7958 different source and destination hard registers.
7960 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
7963 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
7964 vtables into the type inheritance graph.
7966 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7968 PR rtl-optimization/59754
7969 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
7970 modes in the REGNO != REGNO case.
7972 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7974 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
7976 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7978 PR tree-optimization/59745
7979 * tree-predcom.c (tree_predictive_commoning_loop): Call
7980 free_affine_expand_cache if giving up because components is NULL.
7982 * target-globals.c (save_target_globals): Allocate < 4KB structs using
7983 GC in payload of target_globals struct instead of allocating them on
7984 the heap and the larger structs separately using GC.
7985 * target-globals.h (struct target_globals): Make regs, hard_regs,
7986 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
7987 of GTY((skip)) and change type to void *.
7988 (reset_target_globals): Cast loads from those fields to corresponding
7991 2014-01-10 Steve Ellcey <sellcey@mips.com>
7994 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
7995 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
7996 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
7998 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
8001 * aarch64-modes.def (CC_Zmode): New flags mode.
8002 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
8003 represents an equality.
8004 (aarch64_get_condition_code): Handle CC_Zmode.
8005 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
8007 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8009 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
8010 extraction in good case.
8012 2014-01-10 Richard Biener <rguenther@suse.de>
8014 PR tree-optimization/59374
8015 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
8016 checking after SLP discovery. Mark stmts not participating
8017 in any SLP instance properly.
8019 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8021 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
8022 when handling a SET rtx.
8024 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8026 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
8027 (cortex-a57): Likewise.
8028 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
8030 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8032 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
8033 non-iwmmxt builtins.
8035 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8039 * ipa-devirt.c record_target_from_binfo): Take as argument
8040 stack of binfos and lookup matching one for virtual inheritance.
8041 (possible_polymorphic_call_targets_1): Update.
8043 2014-01-10 Huacai Chen <chenhc@lemote.com>
8045 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
8046 kernel strings for Loongson-2E/2F/3A.
8048 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8051 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
8052 is_gimple_call before calling gimple_call_internal_p.
8054 2014-01-09 Steve Ellcey <sellcey@mips.com>
8056 * Makefile.in (TREE_FLOW_H): Remove.
8057 (TREE_SSA_H): Add file names from tree-flow.h.
8058 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
8059 * tree.h: Remove tree-flow.h reference.
8060 * hash-table.h: Remove tree-flow.h reference.
8061 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
8062 reference with tree-ssa-loop.h.
8064 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8066 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8067 default element-order behavior for -maltivec.
8068 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8069 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8070 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8071 when targeting big endian, at least for now.
8072 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8074 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8077 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8078 var satisfies use_register_for_decl, just take into account type
8079 alignment, rather than decl alignment.
8081 PR tree-optimization/59622
8082 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8083 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8084 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8085 Don't devirtualize for inplace at all. For targets.length () == 1,
8086 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8088 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8090 * config/i386/i386.md (cpu): Remove the unused btver1.
8092 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8094 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8096 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8099 * tree-core.h (struct target_globals): New forward declaration.
8100 (struct tree_target_option): Add globals field.
8101 * tree.h (TREE_TARGET_GLOBALS): Define.
8102 (prepare_target_option_nodes_for_pch): New prototype.
8103 * target-globals.h (struct target_globals): Define even if
8105 * tree.c (prepare_target_option_node_for_pch,
8106 prepare_target_option_nodes_for_pch): New functions.
8107 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8108 * config/i386/i386.c: Include target-globals.h.
8109 (ix86_set_current_function): Instead of doing target_reinit
8110 unconditionally, use save_target_globals_default_opts and
8111 restore_target_globals.
8113 2014-01-09 Richard Biener <rguenther@suse.de>
8115 PR tree-optimization/59715
8116 * tree-cfg.h (split_critical_edges): Declare.
8117 * tree-cfg.c (split_critical_edges): Export.
8118 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8120 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8122 * cfgexpand.c (expand_stack_vars): Optionally disable
8123 asan stack protection.
8124 (expand_used_vars): Likewise.
8125 (partition_stack_vars): Likewise.
8126 * asan.c (asan_emit_stack_protection): Optionally disable
8127 after return stack usage.
8128 (instrument_derefs): Optionally disable memory access instrumentation.
8129 (instrument_builtin_call): Likewise.
8130 (instrument_strlen_call): Likewise.
8131 (asan_protect_global): Optionally disable global variables protection.
8132 * doc/invoke.texi: Added doc for new options.
8133 * params.def: Added new options.
8134 * params.h: Likewise.
8136 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8138 PR rtl-optimization/59724
8139 * ifcvt.c (cond_exec_process_if_block): Don't call
8140 flow_find_head_matching_sequence with 0 longest_match.
8141 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8142 non-active insns if !stop_after.
8143 (try_head_merge_bb): Revert 2014-01-07 changes.
8145 2014-01-08 Jeff Law <law@redhat.com>
8147 * ree.c (get_sub_rtx): New function, extracted from...
8148 (merge_def_and_ext): Here.
8149 (combine_reaching_defs): Use get_sub_rtx.
8151 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8153 * cgraph.h (varpool_variable_node): Do not choke on null node.
8155 2014-01-08 Catherine Moore <clm@codesourcery.com>
8157 * config/mips/mips.md (simple_return): Attempt to use JRC
8159 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8161 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8163 PR rtl-optimization/59137
8164 * reorg.c (steal_delay_list_from_target): Call update_block for
8166 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8168 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8170 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8171 two duplicate entries.
8173 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8176 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8178 * config/mips/mips.c (mips_truncated_op_cost): New function.
8179 (mips_rtx_costs): Adjust test for BADDU.
8180 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8182 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8184 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8185 (*baddu_si): ...this new pattern.
8187 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8190 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8192 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8195 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8197 (expand_expr, expand_normal): Adjust.
8198 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8199 inner_reference_p. Use inner_reference_p to expand inner references.
8200 (store_expr): Adjust.
8201 * cfgexpand.c (expand_call_stmt): Adjust.
8203 2014-01-08 Rong Xu <xur@google.com>
8205 * gcov-io.c (gcov_var): Move from gcov-io.h.
8206 (gcov_position): Ditto.
8207 (gcov_is_error): Ditto.
8208 (gcov_rewrite): Ditto.
8209 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8210 only part to libgcc/libgcov.h.
8212 2014-01-08 Marek Polacek <polacek@redhat.com>
8215 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8217 2014-01-08 Marek Polacek <polacek@redhat.com>
8220 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8222 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8224 PR rtl-optimization/59649
8225 * stor-layout.c (get_mode_bounds): For BImode return
8226 0 and STORE_FLAG_VALUE.
8228 2014-01-08 Richard Biener <rguenther@suse.de>
8231 * gimple.h (is_gimple_builtin_call): Remove.
8232 (gimple_builtin_call_types_compatible_p): New.
8233 (gimple_call_builtin_p): New overload.
8234 * gimple.c (is_gimple_builtin_call): Remove.
8235 (validate_call): Rename to ...
8236 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8238 (validate_type): New static function.
8239 (gimple_call_builtin_p): New overload and adjust.
8240 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8241 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8242 (gimple_fold_stmt_to_constant_1): Likewise.
8243 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8245 2014-01-08 Richard Biener <rguenther@suse.de>
8248 * gimplify.c (gimplify_expr): Gimplify register-register type
8249 VIEW_CONVERT_EXPRs to separate stmts.
8251 2014-01-07 Jeff Law <law@redhat.com>
8254 * ree.c (combine_set_extension): Handle case where source
8255 and destination registers in an extension insn are different.
8256 (combine_reaching_defs): Allow source and destination registers
8257 in extension to be different under limited circumstances.
8258 (add_removable_extension): Remove restriction that the
8259 source and destination registers in the extension are the same.
8260 (find_and_remove_re): Emit a copy from the extension's
8261 destination to its source after the defining insn if
8262 the source and destination registers are different.
8265 * ifcvt.c (merge_if_block): If we are merging a block with more than
8266 one successor with a block with no successors, remove any BARRIER
8267 after the second block.
8269 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8271 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8273 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8276 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8277 for 14-bit register offsets when INT14_OK_STRICT is false.
8279 2014-01-07 Roland Stigge <stigge@antcom.de>
8280 Michael Meissner <meissner@linux.vnet.ibm.com>
8283 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8284 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8286 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8288 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8291 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8293 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8294 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8295 rtx is const0_rtx or not.
8297 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8300 * target-globals.c (save_target_globals): Remove this_fn_optab
8302 * toplev.c: Include optabs.h.
8303 (target_reinit): Temporarily restore the global options if another
8304 set of options are in force.
8306 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8308 PR rtl-optimization/58668
8309 * cfgcleanup.c (flow_find_cross_jump): Don't count
8310 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8311 to determine what is counted.
8312 (flow_find_head_matching_sequence): Use active_insn_p to determine
8314 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8316 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8317 determine what is counted.
8319 PR tree-optimization/59643
8320 * tree-predcom.c (split_data_refs_to_components): If one dr is
8321 read and one write, determine_offset fails and the write isn't
8322 in the bad component, just put the read into the bad component.
8324 2014-01-07 Mike Stump <mikestump@comcast.net>
8325 Jakub Jelinek <jakub@redhat.com>
8328 * tree-core.h (struct tree_optimization_option): Change optabs
8329 type from unsigned char * to void *.
8330 * optabs.c (init_tree_optimization_optabs): Adjust
8331 TREE_OPTIMIZATION_OPTABS initialization.
8333 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8336 * config/i386/i386.h (struct machine_function): Add
8337 no_drap_save_restore field.
8338 * config/i386/i386.c (ix86_save_reg): Use
8339 !cfun->machine->no_drap_save_restore instead of
8340 crtl->stack_realign_needed.
8341 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8342 this function clears frame_pointer_needed. Set
8343 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8344 and DRAP reg is needed.
8346 2014-01-06 Marek Polacek <polacek@redhat.com>
8349 * doc/implement-c.texi: Mention that other integer types are
8350 permitted as bit-field types in strictly conforming mode.
8352 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8354 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8357 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8359 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8361 2014-01-06 Martin Jambor <mjambor@suse.cz>
8364 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8366 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8368 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8372 * var-tracking.c (add_stores): Preserve the value of the source even if
8373 we don't record the store.
8375 2014-01-06 Terry Guo <terry.guo@arm.com>
8377 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8379 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8382 * config/darwin.c (darwin_function_section): Adjust return values to
8383 correspond to optimisation changes made in r206070.
8385 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8387 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8388 from prefetch_block tune setting.
8389 (nocona_cost): Correct size of prefetch block to 64.
8391 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8393 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8394 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8395 used to save the static chain register in the computation of the offset
8396 from which the FP registers need to be restored.
8398 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8400 PR tree-optimization/59519
8401 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8402 ICE if get_current_def (current_new_name) is already non-NULL, as long
8403 as it is a phi result of some other phi in *new_exit_bb that has
8406 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8407 or vmovdqu* for misaligned_operand.
8408 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8409 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8410 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8411 aligned_mem for AVX512F masked aligned load and store builtins and for
8414 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8416 PR tree-optimization/59651
8417 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8418 Address range for negative step should be added by TYPE_SIZE_UNIT.
8420 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8422 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8423 overlapping registers also for registers other than the stack pointer.
8425 2014-01-03 Marek Polacek <polacek@redhat.com>
8428 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8431 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8434 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8437 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8438 (push splitter): Use <P:MODE_SIZE> instead of
8439 GET_MODE_SIZE (<P:MODE>mode).
8440 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8441 (mov -1, reg peephole2): Likewise.
8442 * config/i386/sse.md (*mov<mode>_internal,
8443 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8444 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8445 *<code><mode>3, *andnot<mode>3<mask_name>,
8446 <mask_codefor><code><mode>3<mask_name>): Likewise.
8447 * config/i386/subst.md (mask_mode512bit_condition,
8448 sd_mask_mode512bit_condition): Likewise.
8450 2014-01-02 Xinliang David Li <davidxl@google.com>
8452 PR tree-optimization/59303
8453 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8454 (dump_predicates): Better output format.
8455 (pred_equal_p): New function.
8456 (is_neq_relop_p): Ditto.
8457 (is_neq_zero_form_p): Ditto.
8458 (pred_expr_equal_p): Ditto.
8459 (pred_neg_p): Ditto.
8460 (simplify_pred): Ditto.
8461 (simplify_preds_2): Ditto.
8462 (simplify_preds_3): Ditto.
8463 (simplify_preds_4): Ditto.
8464 (simplify_preds): Ditto.
8466 (push_to_worklist): Ditto.
8467 (get_pred_info_from_cmp): Ditto.
8468 (is_degenerated_phi): Ditto.
8469 (normalize_one_pred_1): Ditto.
8470 (normalize_one_pred): Ditto.
8471 (normalize_one_pred_chain): Ditto.
8472 (normalize_preds): Ditto.
8473 (normalize_cond_1): Remove function.
8474 (normalize_cond): Ditto.
8475 (is_gcond_subset_of): Ditto.
8476 (is_subset_of_any): Ditto.
8477 (is_or_set_subset_of): Ditto.
8478 (is_and_set_subset_of): Ditto.
8479 (is_norm_cond_subset_of): Ditto.
8480 (pred_chain_length_cmp): Ditto.
8481 (convert_control_dep_chain_into_preds): Type change.
8482 (find_predicates): Ditto.
8483 (find_def_preds): Ditto.
8484 (destroy_predicates_vecs): Ditto.
8485 (find_matching_predicates_in_rest_chains): Ditto.
8486 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8487 (is_pred_expr_subset): Ditto.
8488 (is_pred_chain_subset_of): Ditto.
8489 (is_included_in): Ditto.
8490 (is_superset_of): Ditto.
8492 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8494 Update copyright years.
8496 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8498 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8499 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8500 config/arc/arc.md, config/arc/arc.opt,
8501 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8502 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8503 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8504 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8505 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8506 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8507 vtable-verify.c, vtable-verify.h: Use the standard form for the
8510 2014-01-02 Tobias Burnus <burnus@net-b.de>
8512 * gcc.c (process_command): Update copyright notice dates.
8513 * gcov-dump.c: Ditto.
8515 * doc/cpp.texi: Bump @copying's copyright year.
8516 * doc/cppinternals.texi: Ditto.
8517 * doc/gcc.texi: Ditto.
8518 * doc/gccint.texi: Ditto.
8519 * doc/gcov.texi: Ditto.
8520 * doc/install.texi: Ditto.
8521 * doc/invoke.texi: Ditto.
8523 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8525 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8527 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8529 * config/i386/sse.md (*mov<mode>_internal): Guard
8530 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8532 PR rtl-optimization/59647
8533 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8534 new_rtx into UNSIGNED_FLOAT rtxes.
8536 Copyright (C) 2014 Free Software Foundation, Inc.
8538 Copying and distribution of this file, with or without modification,
8539 are permitted in any medium without royalty provided the copyright
8540 notice and this notice are preserved.