1 2014-04-29 David Malcolm <dmalcolm@redhat.com>
3 * tree-cfg.c (dump_function_to_file): Dump the return type of
4 functions, in a line to itself before the function body, mimicking
5 the layout of a C function.
7 2014-04-29 Jakub Jelinek <jakub@redhat.com>
9 PR tree-optimization/60971
10 * tree-tailcall.c (process_assignment): Reject conversions which
13 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15 * calls.c (initialize_argument_information): Always treat
16 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
17 (expand_call): Likewise.
18 (emit_library_call_calue_1): Likewise.
19 * expr.c (PUSH_ARGS_REVERSED): Do not define.
20 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
23 2014-04-29 Nick Clifton <nickc@redhat.com>
25 * config/msp430/msp430.md (umulsidi): Fix typo.
26 (mulhisi3): Enable even inside interrupt handlers.
27 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
28 bigger return address pushed in large mode.
30 2014-04-29 Nick Clifton <nickc@redhat.com>
32 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
33 (arc_init_reg_tables): Use a machine_mode enum to iterate over
35 * config/m32r/m32r.c (init_reg_tables): Likewise.
36 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
37 enum to hold the modes.
39 2014-04-29 Richard Biener <rguenther@suse.de>
41 * dominance.c (free_dominance_info): Add overload with
43 (dom_info_state): Likewise.
44 (dom_info_available_p): Likewise.
45 * basic-block.h (free_dominance_info, dom_info_state,
46 dom_info_available_p): Declare overloads.
47 * passes.c (execute_function_todo): Verify that verifiers
48 don't change dominator info state. Drop dominator info
49 for IPA pass invocations.
50 * cgraph.c (release_function_body): Restore asserts that
51 dominator information is released.
53 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
55 * doc/invoke.texi: Fix typo.
56 * tree-vrp.c: Fix typos.
57 * gimple.c (infer_nonnull_range): Reorder operands of an &&
60 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
62 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
64 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
66 * config/aarch64/aarch64-builtins.c
67 (aarch64_types_storestruct_lane_qualifiers): New.
68 (TYPES_STORESTRUCT_LANE): Likewise.
69 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
72 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
73 (vec_store_lanesci_lane<mode>): Likewise.
74 (vec_store_lanesxi_lane<mode>): Likewise.
75 (aarch64_st2_lane<VQ:mode>): Likewise.
76 (aarch64_st3_lane<VQ:mode>): Likewise.
77 (aarch64_st4_lane<VQ:mode>): Likewise.
78 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
79 * config/aarch64/arm_neon.h
80 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
81 use new macro arguments.
82 (__ST3_LANE_FUNC): Likewise.
83 (__ST4_LANE_FUNC): Likewise.
84 * config/aarch64/iterators.md (V_TWO_ELEM): New.
85 (V_THREE_ELEM): Likewise.
86 (V_FOUR_ELEM): Likewise.
88 2014-04-28 David Malcolm <dmalcolm@redhat.com>
90 * doc/gimple.texi: Replace the description of the now-defunct
91 union gimple_statement_d with a diagram showing the
92 gimple_statement_base class hierarchy and its relationships to
93 the GSS_ and GIMPLE_ enums.
95 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
97 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
98 * config/aarch64/aarch64.c
99 (aarch64_cannot_change_mode_class): Weaken conditions.
100 (aarch64_modes_tieable_p): New.
101 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
103 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
105 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
106 (loadsync_<mode>): Change mode.
107 (load_quadpti, store_quadpti): New.
108 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
109 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
111 2014-04-28 Martin Jambor <mjambor@suse.cz>
113 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
114 same alias type as the original statement.
115 (subreplacement_assignment_data): New type.
116 (handle_unscalarized_data_in_subtree): New type of parameter,
117 generate new memory accesses with same alias type as the original
119 (load_assign_lhs_subreplacements): Likewise.
120 (sra_modify_constructor_assign): Generate new memory accesses with
121 same alias type as the original statement.
123 2014-04-28 Richard Biener <rguenther@suse.de>
125 * tree-pass.h (TODO_verify_il): Define.
126 (TODO_verify_all): Complete properly.
127 * passes.c (execute_function_todo): Move existing loop-closed
128 SSA verification under TODO_verify_il.
129 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
130 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
131 Fix tree sharing issue.
133 2014-04-28 Richard Biener <rguenther@suse.de>
136 * builtins.def (DEF_C11_BUILTIN): Add.
137 (BUILT_IN_ALIGNED_ALLOC): Likewise.
138 * coretypes.h (enum function_class): Add function_c11_misc.
139 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
140 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
141 (call_may_clobber_ref_p_1): Likewise.
142 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
143 (mark_all_reaching_defs_necessary_1): Likewise.
144 (propagate_necessity): Likewise.
145 (eliminate_unnecessary_stmts): Likewise.
146 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
148 2014-04-28 Richard Biener <rguenther@suse.de>
150 * tree-vrp.c (vrp_var_may_overflow): Remove.
151 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
152 with overflow immediately bump to one before that value and
153 let iteration figure out overflow status.
155 2014-04-28 Richard Biener <rguenther@suse.de>
157 * configure.ac: Do valgrind header checks unconditionally.
158 Add --enable-valgrind-annotations.
159 * system.h: Guard valgrind header inclusion with
160 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
161 * alloc-pool.c (pool_alloc, pool_free): Use
162 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
163 to guard possibly dead code.
164 * config.in: Regenerated.
165 * configure: Likewise.
167 2014-04-28 Jeff Law <law@redhat.com>
169 PR tree-optimization/60902
170 * tree-ssa-threadedge.c
171 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
172 over real defs when invalidating outputs from statements that do not
173 produce useful outputs for threading.
175 2014-04-28 Richard Biener <rguenther@suse.de>
177 PR tree-optimization/60979
178 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
179 SCOPs that end in a block with a successor with abnormal
182 2014-04-28 Richard Biener <rguenther@suse.de>
184 * tree-pass.h (execute_pass_list): Adjust prototype.
185 * passes.c (pass_manager::execute_early_local_passes):
187 (do_per_function): Change callback signature, push all actual
188 work to the callbals.
189 (do_per_function_toporder): Likewise.
190 (execute_function_dump): Adjust.
191 (execute_function_todo): Likewise.
192 (clear_last_verified): Likewise.
193 (verify_curr_properties): Likewise.
194 (update_properties_after_pass): Likewise.
195 (execute_pass_list_1): Split out from ...
196 (execute_pass_list): ... here. Adjust.
197 (execute_ipa_pass_list): Likewise.
198 * cgraphunit.c (cgraph_add_new_function): Adjust.
199 (analyze_function): Likewise.
200 (expand_function): Likewise.
201 * cgraph.c (release_function_body): Free dominance info
202 here instead of asserting it was magically freed elsewhere.
204 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
206 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
207 * configure: Regenerate.
208 * config/sparc/sparc.opt (muser-mode): New option.
209 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
211 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
212 * doc/invoke.texi (SPARC options): Document -muser-mode.
214 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
216 * cselib.c (find_slot_memmode): Delete.
217 (cselib_hasher): Change compare_type to a struct.
218 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
220 (preserve_constants_and_equivs): Adjust for new compare_type.
221 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
222 (wrap_constant): Delete.
223 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
225 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
227 * doc/install.texi (Building with profile feedback): Remove
230 2014-04-26 Tom de Vries <tom@codesourcery.com>
232 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
235 2014-04-25 Cary Coutant <ccoutant@google.com>
238 * dwarf2out.c (should_move_die_to_comdat): A type definition
239 can contain a subprogram definition, but don't move it to a
241 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
242 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
244 (clone_tree_hash): Rename to...
245 (clone_tree_partial): ...this; change callers. Copy
246 DW_TAG_subprogram DIEs as declarations.
247 (copy_decls_walk): Don't copy children of a declaration into a
250 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
253 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
256 2014-04-25 Jiong Wang <jiong.wang@arm.com>
258 * config/arm/predicates.md (call_insn_operand): Add long_call check.
259 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
261 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
264 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
266 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
269 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
271 PR tree-optimization/60930
272 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
273 creating a multiply candidate by folding two constant
274 multiplicands when the result overflows.
276 2014-04-25 Jakub Jelinek <jakub@redhat.com>
278 PR tree-optimization/60960
279 * tree-vect-generic.c (expand_vector_operation): Only call
280 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
282 2014-04-25 Tom de Vries <tom@codesourcery.com>
284 * expr.c (clobber_reg_mode): New function.
285 * expr.h (clobber_reg): New function.
287 2014-04-25 Tom de Vries <tom@codesourcery.com>
289 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
292 2014-04-25 Radovan Obradovic <robradovic@mips.com>
293 Tom de Vries <tom@codesourcery.com>
295 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
297 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
298 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
299 new argument to find_all_hard_reg_sets call.
301 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
303 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
304 Use HOST_WIDE_INT_C for mask literal.
305 (aarch_rev16_shleft_mask_imm_p): Likewise.
307 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
310 * config/sparc/sparc.md (ashlsi3_extend): Delete.
312 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
314 PR preprocessor/56540
315 * config/i386/i386-c.c (ix86_target_macros): Define
316 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
318 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
320 * configure.ac (tga_func): Remove.
321 (LIB_TLS_SPEC): Remove.
322 * configure: Regenerate.
323 * config.in: Regenerate.
324 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
326 2014-04-25 Richard Biener <rguenther@suse.de>
329 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
330 call stmt use/clobber sets during stmt walk instead of
331 walking the possibly incomplete set of caller edges.
333 2014-04-25 Richard Biener <rguenther@suse.de>
336 * passes.c (apply_ipa_transforms): Inline into only caller ...
337 (execute_one_pass): ... here. Properly bring in function
338 bodies for nodes we want to apply IPA transforms to.
340 2014-04-24 Cong Hou <congh@google.com>
342 PR tree-optimization/60896
343 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
344 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
345 (vect_mark_pattern_stmts): Set the def type of all statements in
346 PATTERN_DEF_SEQ as vect_internal_def.
348 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
350 * doc/extend.texi (PowerPC Built-in Functions): Document new
351 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
353 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
355 * config/rs6000/predicates.md (const_0_to_3_operand): New
356 predicate to match 0..3 integer constants.
358 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
359 to support adding miscellaneous builtin functions.
360 (BU_DFP_MISC_2): Likewise.
361 (BU_P7_MISC_1): Likewise.
362 (BU_P7_MISC_2): Likewise.
363 (BU_P8V_MISC_3): Likewise.
364 (BU_MISC_1): Likewise.
365 (BU_MISC_2): Likewise.
366 (DIVWE): Add extended divide builtin functions.
374 (DXEX): Add decimal floating-point builtin functions.
386 (CDTBCD): Add new BCD builtin functions.
390 (BCDADD_LT): Likewise.
391 (BCDADD_EQ): Likewise.
392 (BCDADD_GT): Likewise.
393 (BCDADD_OV): Likewise.
395 (BCDSUB_LT): Likewise.
396 (BCDSUB_EQ): Likewise.
397 (BCDSUB_GT): Likewise.
398 (BCDSUB_OV): Likewise.
399 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
400 (UNPACK_TD): Likewise.
402 (UNPACK_TF): Likewise.
403 (UNPACK_TF_0): Likewise.
404 (UNPACK_TF_1): Likewise.
405 (PACK_V1TI): Likewise.
406 (UNPACK_V1TI): Likewise.
408 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
409 support for decimal floating point builtin functions.
410 (rs6000_expand_ternop_builtin): Add checks for the new builtin
411 functions that take constant arguments.
412 (rs6000_invalid_builtin): Add decimal floating point builtin
414 (rs6000_init_builtins): Setup long double, _Decimal64, and
415 _Decimal128 types for new builtin functions.
416 (builtin_function_type): Set the unsigned flags appropriately for
417 the new builtin functions.
418 (rs6000_opt_masks): Add support for decimal floating point builtin
421 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
422 floating point builtin functions.
423 (RS6000_BTM_COMMON): Likewise.
424 (RS6000_BTI_long_double): Likewise.
425 (RS6000_BTI_dfloat64): Likewise.
426 (RS6000_BTI_dfloat128): Likewise.
427 (long_double_type_internal_node): Likewise.
428 (dfloat64_type_internal_node): Likewise.
429 (dfloat128_type_internal_node): Likewise.
431 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
432 2.07 bcd arithmetic instructions.
433 (UNSPEC_BCDSUB): Likewise.
434 (UNSPEC_BCD_OVERFLOW): Likewise.
435 (UNSPEC_BCD_ADD_SUB): Likewise.
436 (bcd_add_sub): Likewise.
437 (BCD_TEST): Likewise.
438 (bcd<bcd_add_sub>): Likewise.
439 (bcd<bcd_add_sub>_test): Likewise.
440 (bcd<bcd_add_sub>_test2): Likewise.
441 (bcd<bcd_add_sub>_<code>): Likewise.
442 (peephole2 for combined bcd ops): Likewise.
444 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
445 decimal floating point builtin functions.
446 (UNSPEC_DENBCD): Likewise.
447 (UNSPEC_DXEX): Likewise.
448 (UNSPEC_DIEX): Likewise.
449 (UNSPEC_DSCLI): Likewise.
450 (UNSPEC_DSCRI): Likewise.
451 (D64_D128): Likewise.
452 (dfp_suffix): Likewise.
453 (dfp_ddedpd_<mode>): Likewise.
454 (dfp_denbcd_<mode>): Likewise.
455 (dfp_dxex_<mode>): Likewise.
456 (dfp_diex_<mode>): Likewise.
457 (dfp_dscli_<mode>): Likewise.
458 (dfp_dscri_<mode>): Likewise.
460 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
462 (UNSPEC_CDTBCD): Likewise.
463 (UNSPEC_CBCDTD): Likewise.
464 (UNSPEC_DIVE): Add support for new extended divide builtin
466 (UNSPEC_DIVEO): Likewise.
467 (UNSPEC_DIVEU): Likewise.
468 (UNSPEC_DIVEUO): Likewise.
469 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
470 pack/unpack 128-bit types.
471 (UNSPEC_PACK_128BIT): Likewise.
472 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
473 (udiv<mode>3): Use idiv_ldiv mode attribute.
474 (div<mode>3): Likewise.
475 (addg6s): Add new BCD builtin functions.
478 (UNSPEC_DIV_EXTEND): Add support for new extended divide
480 (div_extend): Likewise.
481 (div<div_extend>_<mode>"): Likewise.
482 (FP128_64): Add support for new builtin functions to pack/unpack
484 (unpack<mode>): Likewise.
485 (unpacktf_0): Likewise.
486 (unpacktf_1): Likewise.
487 (unpack<mode>_dm): Likewise.
488 (unpack<mode>_nodm): Likewise.
489 (pack<mode>): Likewise.
490 (unpackv1ti): Likewise.
491 (packv1ti): Likewise.
493 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
495 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
498 2014-04-24 Jakub Jelinek <jakub@redhat.com>
500 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
501 * gimplify.c (omp_is_private): Change last argument's type to int.
502 Only diagnose lastprivate if the simd argument is 1, only diagnose
503 linear if the simd argument is 2.
504 (gimplify_omp_for): Adjust omp_is_private callers. When adding
505 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
506 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
507 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
508 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
509 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
510 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
511 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
512 * tree-nested.c (convert_nonlocal_omp_clauses,
513 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
515 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
518 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
521 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
523 * flag-types.h (enum ivar_visibility): Add.
525 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
527 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
530 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
532 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
534 2014-04-24 Radovan Obradovic <robradovic@mips.com>
535 Tom de Vries <tom@codesourcery.com>
537 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
538 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
540 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
541 * emit-rtl.c (try_split): Same.
543 2014-04-24 Radovan Obradovic <robradovic@mips.com>
544 Tom de Vries <tom@codesourcery.com>
546 * common.opt (fuse-caller-save): New option.
548 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
550 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
553 2014-04-24 Richard Biener <rguenther@suse.de>
555 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
556 during TER and instead use the sepops interface for expanding
557 non-GIMPLE_SINGLE_RHS.
559 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
561 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
562 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
564 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
566 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
567 assembler 64-bit option.
568 * configure: Regenerate.
570 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
572 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
573 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
574 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
575 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
576 (TARGET_CRYPTO): Take TARGET_SIMD into account.
578 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
580 * config/aarch64/aarch64-builtins.c
581 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
582 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
583 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
584 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
586 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
587 (Vrevsuff): New mode attribute.
589 2014-04-24 Terry Guo <terry.guo@arm.com>
591 * config/arm/arm.h (machine_function): Define variable
592 after_arm_reorg here.
593 * config/arm/arm.c (after_arm_reorg): Remove the definition.
594 (arm_split_constant): Update the way to access variable
597 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
599 2014-04-23 Tom de Vries <tom@codesourcery.com>
601 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
603 2014-04-23 David Malcolm <dmalcolm@redhat.com>
605 * is-a.h: Update comments to reflect the following changes to the
606 "pointerness" of the API, making the template parameter match the
607 return type, allowing use of is-a.h with typedefs of pointers.
608 (is_a_helper::cast): Return a T rather then a pointer to a T, so
609 that the return type matches the parameter to the is_a_helper.
611 (dyn_cast): Likewise.
613 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
614 pointer from the is-a.h API.
616 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
617 (is_a_helper <cgraph_node *>::test): ...this, matching change to
619 (is_a_helper <varpool_node>::test): Likewise, convert to...
620 (is_a_helper <varpool_node *>::test): ...this.
622 (varpool_first_variable): Update for removal of implicit pointer
624 (varpool_next_variable): Likewise.
625 (varpool_first_static_initializer): Likewise.
626 (varpool_next_static_initializer): Likewise.
627 (varpool_first_defined_variable): Likewise.
628 (varpool_next_defined_variable): Likewise.
629 (cgraph_first_defined_function): Likewise.
630 (cgraph_next_defined_function): Likewise.
631 (cgraph_first_function): Likewise.
632 (cgraph_next_function): Likewise.
633 (cgraph_first_function_with_gimple_body): Likewise.
634 (cgraph_next_function_with_gimple_body): Likewise.
635 (cgraph_alias_target): Likewise.
636 (varpool_alias_target): Likewise.
637 (cgraph_function_or_thunk_node): Likewise.
638 (varpool_variable_node): Likewise.
639 (symtab_real_symbol_p): Likewise.
640 * cgraphunit.c (referred_to_p): Likewise.
641 (analyze_functions): Likewise.
642 (handle_alias_pairs): Likewise.
643 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
644 * gimple-ssa.h (gimple_vuse_op): Likewise.
645 (gimple_vdef_op): Likewise.
646 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
647 * gimple.c (gimple_build_asm_1): Likewise.
648 (gimple_build_try): Likewise.
649 (gimple_build_resx): Likewise.
650 (gimple_build_eh_dispatch): Likewise.
651 (gimple_build_omp_for): Likewise.
652 (gimple_omp_for_set_clauses): Likewise.
654 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
655 (is_a_helper <gimple_statement_asm *>::test): ...this.
656 (is_a_helper <gimple_statement_bind>::test): Convert to...
657 (is_a_helper <gimple_statement_bind *>::test): ...this.
658 (is_a_helper <gimple_statement_call>::test): Convert to...
659 (is_a_helper <gimple_statement_call *>::test): ...this.
660 (is_a_helper <gimple_statement_catch>::test): Convert to...
661 (is_a_helper <gimple_statement_catch *>::test): ...this.
662 (is_a_helper <gimple_statement_resx>::test): Convert to...
663 (is_a_helper <gimple_statement_resx *>::test): ...this.
664 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
665 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
666 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
667 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
668 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
669 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
670 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
671 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
672 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
673 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
674 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
675 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
676 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
677 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
678 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
679 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
680 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
681 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
682 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
683 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
684 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
685 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
686 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
687 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
688 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
689 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
690 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
691 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
692 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
693 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
694 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
695 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
696 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
697 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
698 (is_a_helper <gimple_statement_phi>::test): Convert to...
699 (is_a_helper <gimple_statement_phi *>::test): ...this.
700 (is_a_helper <gimple_statement_transaction>::test): Convert to...
701 (is_a_helper <gimple_statement_transaction *>::test): ...this.
702 (is_a_helper <gimple_statement_try>::test): Convert to...
703 (is_a_helper <gimple_statement_try *>::test): ...this.
704 (is_a_helper <gimple_statement_wce>::test): Convert to...
705 (is_a_helper <gimple_statement_wce *>::test): ...this.
706 (is_a_helper <const gimple_statement_asm>::test): Convert to...
707 (is_a_helper <const gimple_statement_asm *>::test): ...this.
708 (is_a_helper <const gimple_statement_bind>::test): Convert to...
709 (is_a_helper <const gimple_statement_bind *>::test): ...this.
710 (is_a_helper <const gimple_statement_call>::test): Convert to...
711 (is_a_helper <const gimple_statement_call *>::test): ...this.
712 (is_a_helper <const gimple_statement_catch>::test): Convert to...
713 (is_a_helper <const gimple_statement_catch *>::test): ...this.
714 (is_a_helper <const gimple_statement_resx>::test): Convert to...
715 (is_a_helper <const gimple_statement_resx *>::test): ...this.
716 (is_a_helper <const gimple_statement_eh_dispatch>::test):
718 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
720 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
722 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
723 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
725 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
727 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
729 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
731 (is_a_helper <const gimple_statement_omp_return>::test): Convert
733 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
734 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
736 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
737 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
739 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
740 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
741 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
742 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
744 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
745 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
747 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
748 (is_a_helper <const gimple_statement_omp_target>::test): Convert
750 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
751 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
753 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
754 (is_a_helper <const gimple_statement_omp_single>::test): Convert
756 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
757 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
759 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
760 (is_a_helper <const gimple_statement_omp_task>::test): Convert
762 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
763 (is_a_helper <const gimple_statement_phi>::test): Convert to...
764 (is_a_helper <const gimple_statement_phi *>::test): ...this.
765 (is_a_helper <const gimple_statement_transaction>::test): Convert
767 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
768 (is_a_helper <const gimple_statement_with_ops>::test): Convert
770 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
771 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
772 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
773 (is_a_helper <const gimple_statement_with_memory_ops>::test):
775 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
777 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
779 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
781 (gimple_use_ops): Update for removal of implicit pointer from the
783 (gimple_set_use_ops): Likewise.
784 (gimple_vuse): Likewise.
785 (gimple_vdef): Likewise.
786 (gimple_vuse_ptr): Likewise.
787 (gimple_vdef_ptr): Likewise.
788 (gimple_set_vuse): Likewise.
789 (gimple_set_vdef): Likewise.
790 (gimple_omp_return_set_lhs): Likewise.
791 (gimple_omp_return_lhs): Likewise.
792 (gimple_omp_return_lhs_ptr): Likewise.
793 (gimple_call_fntype): Likewise.
794 (gimple_call_set_fntype): Likewise.
795 (gimple_call_set_internal_fn): Likewise.
796 (gimple_call_use_set): Likewise.
797 (gimple_call_clobber_set): Likewise.
798 (gimple_bind_vars): Likewise.
799 (gimple_bind_set_vars): Likewise.
800 (gimple_bind_body_ptr): Likewise.
801 (gimple_bind_set_body): Likewise.
802 (gimple_bind_add_stmt): Likewise.
803 (gimple_bind_block): Likewise.
804 (gimple_bind_set_block): Likewise.
805 (gimple_asm_ninputs): Likewise.
806 (gimple_asm_noutputs): Likewise.
807 (gimple_asm_nclobbers): Likewise.
808 (gimple_asm_nlabels): Likewise.
809 (gimple_asm_input_op): Likewise.
810 (gimple_asm_input_op_ptr): Likewise.
811 (gimple_asm_output_op): Likewise.
812 (gimple_asm_output_op_ptr): Likewise.
813 (gimple_asm_set_output_op): Likewise.
814 (gimple_asm_clobber_op): Likewise.
815 (gimple_asm_set_clobber_op): Likewise.
816 (gimple_asm_label_op): Likewise.
817 (gimple_asm_set_label_op): Likewise.
818 (gimple_asm_string): Likewise.
819 (gimple_catch_types): Likewise.
820 (gimple_catch_types_ptr): Likewise.
821 (gimple_catch_handler_ptr): Likewise.
822 (gimple_catch_set_types): Likewise.
823 (gimple_catch_set_handler): Likewise.
824 (gimple_eh_filter_types): Likewise.
825 (gimple_eh_filter_types_ptr): Likewise.
826 (gimple_eh_filter_failure_ptr): Likewise.
827 (gimple_eh_filter_set_types): Likewise.
828 (gimple_eh_filter_set_failure): Likewise.
829 (gimple_eh_must_not_throw_fndecl): Likewise.
830 (gimple_eh_must_not_throw_set_fndecl): Likewise.
831 (gimple_eh_else_n_body_ptr): Likewise.
832 (gimple_eh_else_e_body_ptr): Likewise.
833 (gimple_eh_else_set_n_body): Likewise.
834 (gimple_eh_else_set_e_body): Likewise.
835 (gimple_try_eval_ptr): Likewise.
836 (gimple_try_cleanup_ptr): Likewise.
837 (gimple_try_set_eval): Likewise.
838 (gimple_try_set_cleanup): Likewise.
839 (gimple_wce_cleanup_ptr): Likewise.
840 (gimple_wce_set_cleanup): Likewise.
841 (gimple_phi_capacity): Likewise.
842 (gimple_phi_num_args): Likewise.
843 (gimple_phi_result): Likewise.
844 (gimple_phi_result_ptr): Likewise.
845 (gimple_phi_set_result): Likewise.
846 (gimple_phi_arg): Likewise.
847 (gimple_phi_set_arg): Likewise.
848 (gimple_resx_region): Likewise.
849 (gimple_resx_set_region): Likewise.
850 (gimple_eh_dispatch_region): Likewise.
851 (gimple_eh_dispatch_set_region): Likewise.
852 (gimple_omp_critical_name): Likewise.
853 (gimple_omp_critical_name_ptr): Likewise.
854 (gimple_omp_critical_set_name): Likewise.
855 (gimple_omp_for_clauses): Likewise.
856 (gimple_omp_for_clauses_ptr): Likewise.
857 (gimple_omp_for_set_clauses): Likewise.
858 (gimple_omp_for_collapse): Likewise.
859 (gimple_omp_for_index): Likewise.
860 (gimple_omp_for_index_ptr): Likewise.
861 (gimple_omp_for_set_index): Likewise.
862 (gimple_omp_for_initial): Likewise.
863 (gimple_omp_for_initial_ptr): Likewise.
864 (gimple_omp_for_set_initial): Likewise.
865 (gimple_omp_for_final): Likewise.
866 (gimple_omp_for_final_ptr): Likewise.
867 (gimple_omp_for_set_final): Likewise.
868 (gimple_omp_for_incr): Likewise.
869 (gimple_omp_for_incr_ptr): Likewise.
870 (gimple_omp_for_set_incr): Likewise.
871 (gimple_omp_for_pre_body_ptr): Likewise.
872 (gimple_omp_for_set_pre_body): Likewise.
873 (gimple_omp_parallel_clauses): Likewise.
874 (gimple_omp_parallel_clauses_ptr): Likewise.
875 (gimple_omp_parallel_set_clauses): Likewise.
876 (gimple_omp_parallel_child_fn): Likewise.
877 (gimple_omp_parallel_child_fn_ptr): Likewise.
878 (gimple_omp_parallel_set_child_fn): Likewise.
879 (gimple_omp_parallel_data_arg): Likewise.
880 (gimple_omp_parallel_data_arg_ptr): Likewise.
881 (gimple_omp_parallel_set_data_arg): Likewise.
882 (gimple_omp_task_clauses): Likewise.
883 (gimple_omp_task_clauses_ptr): Likewise.
884 (gimple_omp_task_set_clauses): Likewise.
885 (gimple_omp_task_child_fn): Likewise.
886 (gimple_omp_task_child_fn_ptr): Likewise.
887 (gimple_omp_task_set_child_fn): Likewise.
888 (gimple_omp_task_data_arg): Likewise.
889 (gimple_omp_task_data_arg_ptr): Likewise.
890 (gimple_omp_task_set_data_arg): Likewise.
891 (gimple_omp_taskreg_clauses): Likewise.
892 (gimple_omp_taskreg_clauses_ptr): Likewise.
893 (gimple_omp_taskreg_set_clauses): Likewise.
894 (gimple_omp_taskreg_child_fn): Likewise.
895 (gimple_omp_taskreg_child_fn_ptr): Likewise.
896 (gimple_omp_taskreg_set_child_fn): Likewise.
897 (gimple_omp_taskreg_data_arg): Likewise.
898 (gimple_omp_taskreg_data_arg_ptr): Likewise.
899 (gimple_omp_taskreg_set_data_arg): Likewise.
900 (gimple_omp_task_copy_fn): Likewise.
901 (gimple_omp_task_copy_fn_ptr): Likewise.
902 (gimple_omp_task_set_copy_fn): Likewise.
903 (gimple_omp_task_arg_size): Likewise.
904 (gimple_omp_task_arg_size_ptr): Likewise.
905 (gimple_omp_task_set_arg_size): Likewise.
906 (gimple_omp_task_arg_align): Likewise.
907 (gimple_omp_task_arg_align_ptr): Likewise.
908 (gimple_omp_task_set_arg_align): Likewise.
909 (gimple_omp_single_clauses): Likewise.
910 (gimple_omp_single_clauses_ptr): Likewise.
911 (gimple_omp_single_set_clauses): Likewise.
912 (gimple_omp_target_clauses): Likewise.
913 (gimple_omp_target_clauses_ptr): Likewise.
914 (gimple_omp_target_set_clauses): Likewise.
915 (gimple_omp_target_child_fn): Likewise.
916 (gimple_omp_target_child_fn_ptr): Likewise.
917 (gimple_omp_target_set_child_fn): Likewise.
918 (gimple_omp_target_data_arg): Likewise.
919 (gimple_omp_target_data_arg_ptr): Likewise.
920 (gimple_omp_target_set_data_arg): Likewise.
921 (gimple_omp_teams_clauses): Likewise.
922 (gimple_omp_teams_clauses_ptr): Likewise.
923 (gimple_omp_teams_set_clauses): Likewise.
924 (gimple_omp_sections_clauses): Likewise.
925 (gimple_omp_sections_clauses_ptr): Likewise.
926 (gimple_omp_sections_set_clauses): Likewise.
927 (gimple_omp_sections_control): Likewise.
928 (gimple_omp_sections_control_ptr): Likewise.
929 (gimple_omp_sections_set_control): Likewise.
930 (gimple_omp_for_set_cond): Likewise.
931 (gimple_omp_for_cond): Likewise.
932 (gimple_omp_atomic_store_set_val): Likewise.
933 (gimple_omp_atomic_store_val): Likewise.
934 (gimple_omp_atomic_store_val_ptr): Likewise.
935 (gimple_omp_atomic_load_set_lhs): Likewise.
936 (gimple_omp_atomic_load_lhs): Likewise.
937 (gimple_omp_atomic_load_lhs_ptr): Likewise.
938 (gimple_omp_atomic_load_set_rhs): Likewise.
939 (gimple_omp_atomic_load_rhs): Likewise.
940 (gimple_omp_atomic_load_rhs_ptr): Likewise.
941 (gimple_omp_continue_control_def): Likewise.
942 (gimple_omp_continue_control_def_ptr): Likewise.
943 (gimple_omp_continue_set_control_def): Likewise.
944 (gimple_omp_continue_control_use): Likewise.
945 (gimple_omp_continue_control_use_ptr): Likewise.
946 (gimple_omp_continue_set_control_use): Likewise.
947 (gimple_transaction_body_ptr): Likewise.
948 (gimple_transaction_label): Likewise.
949 (gimple_transaction_label_ptr): Likewise.
950 (gimple_transaction_set_body): Likewise.
951 (gimple_transaction_set_label): Likewise.
953 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
954 * ipa-inline-analysis.c (inline_write_summary): Likewise.
955 * ipa-ref.c (ipa_record_reference): Likewise.
956 * ipa-reference.c (analyze_function): Likewise.
957 (ipa_reference_write_optimization_summary): Likewise.
958 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
959 (address_taken_from_non_vtable_p): Likewise.
960 (comdat_can_be_unshared_p_1): Likewise.
961 * lto-cgraph.c (lto_output_ref): Likewise.
962 (add_references): Likewise.
963 (compute_ltrans_boundary): Likewise.
964 (output_symtab): Likewise.
965 (input_ref): Likewise.
966 (input_cgraph_1): Likewise.
967 (output_cgraph_opt_summary): Likewise.
968 * lto-streamer-out.c (lto_output): Likewise.
969 (output_symbol_p): Likewise.
970 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
971 (lsei_start_function_in_partition): Likewise.
972 (lsei_next_variable_in_partition): Likewise.
973 (lsei_start_variable_in_partition): Likewise.
974 * symtab.c (insert_to_assembler_name_hash): Likewise.
975 (unlink_from_assembler_name_hash): Likewise.
976 (symtab_unregister_node): Likewise.
977 (symtab_remove_node): Likewise.
978 (dump_symtab_node): Likewise.
979 (verify_symtab_base): Likewise.
980 (verify_symtab_node): Likewise.
981 (symtab_make_decl_local): Likewise.
982 (symtab_alias_ultimate_target): Likewise.
983 (symtab_resolve_alias): Likewise.
984 (symtab_get_symbol_partitioning_class): Likewise.
985 * tree-phinodes.c (allocate_phi_node): Likewise.
986 (reserve_phi_args_for_new_edge): Likewise.
987 (remove_phi_args): Likewise.
988 * varpool.c (varpool_node_for_asm): Likewise.
989 (varpool_remove_unreferenced_decls): Likewise.
991 2014-04-23 Jeff Law <law@redhat.com>
993 PR tree-optimization/60902
994 * tree-ssa-threadedge.c
995 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
996 invalidate outputs from statements that do not produce useful
997 outputs for threading.
999 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
1001 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
1002 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
1003 machine descriptions for Stack Smashing Protector.
1005 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
1007 * aarch64.md (<optab>_rol<mode>3): New pattern.
1008 (<optab>_rolsi3_uxtw): Likewise.
1009 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
1011 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
1013 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
1014 (arm_cortex_a12_tune): Likewise.
1016 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1018 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
1020 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1022 * config/arm/arm.md (arm_rev16si2): New pattern.
1023 (arm_rev16si2_alt): Likewise.
1024 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
1026 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1028 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
1029 (rev16<mode>2_alt): Likewise.
1030 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
1031 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
1032 (aarch_rev16_shleft_mask_imm_p): Likewise.
1033 (aarch_rev16_p_1): Likewise.
1034 (aarch_rev16_p): Likewise.
1035 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
1036 (aarch_rev16_shright_mask_imm_p): Likewise.
1037 (aarch_rev16_shleft_mask_imm_p): Likewise.
1039 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1041 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
1042 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
1044 (cortex_a53_extra_costs): Likewise.
1045 (cortex_a57_extra_costs): Likewise.
1046 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
1047 (cortexa7_extra_costs): Likewise.
1048 (cortexa8_extra_costs): Likewise.
1049 (cortexa12_extra_costs): Likewise.
1050 (cortexa15_extra_costs): Likewise.
1051 (v7m_extra_costs): Likewise.
1052 (arm_new_rtx_costs): Handle BSWAP.
1054 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1056 * config/arm/arm.c (cortexa8_extra_costs): New table.
1057 (arm_cortex_a8_tune): New tuning struct.
1058 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
1060 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1062 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
1064 2014-04-23 Richard Biener <rguenther@suse.de>
1066 * Makefile.in (OBJS): Remove loop-unswitch.o.
1067 * loop-unswitch.c: Delete.
1068 * tree-pass.h (make_pass_rtl_unswitch): Remove.
1069 * passes.def (pass_rtl_unswitch): Likewise.
1070 * loop-init.c (gate_rtl_unswitch): Likewise.
1071 (rtl_unswitch): Likewise.
1072 (pass_data_rtl_unswitch): Likewise.
1073 (pass_rtl_unswitch): Likewise.
1074 (make_pass_rtl_unswitch): Likewise.
1075 * rtl.h (reversed_condition): Likewise.
1076 (compare_and_jump_seq): Likewise.
1077 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
1079 * loop-unroll.c (compare_and_jump_seq): Likewise.
1081 2014-04-23 Richard Biener <rguenther@suse.de>
1083 PR tree-optimization/60903
1084 * tree-ssa-loop-im.c (analyze_memory_references): Remove
1085 commented code block.
1086 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
1087 loop flags to newly created BBs and edges.
1089 2014-04-23 Nick Clifton <nickc@redhat.com>
1091 * config/msp430/msp430.c (msp430_handle_option): Move function
1093 (msp430_option_override): Simplify mcu and mcpu option handling.
1094 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
1095 support for -mhwmult command line option.
1096 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
1097 -mhwmult command line option.
1098 (msp430_hwmult_enabled): Delete.
1099 (msp43o_output_labelref): Add support for -mhwmult command line
1101 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
1102 (umulsidi3): Likewise.
1103 * config/msp430/msp430.opt (mmcu): Add Report attribute.
1104 (mcpu, mlarge, msmall): Likewise.
1105 (mhwmult): New option.
1106 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
1108 (msp430_is_f5_mcu): Remove prototype.
1109 (msp430_use_f5_series_hwmult): Add prototype.
1110 * config/msp430/msp430-opts.h: New file.
1111 * common/config/msp430: New directory.
1112 * common/config/msp430/msp430-common.c: New file.
1113 * config.gcc (msp430): Remove target_has_targetm_common.
1114 * doc/invoke.texi: Document -mhwmult command line option.
1116 2014-04-23 Nick Clifton <nickc@redhat.com>
1118 * config/i386/cygwin.h (ENDFILE_SPEC): Include
1119 default-manifest.o if it can be found in the search path.
1120 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
1122 2014-04-23 Terry Guo <terry.guo@arm.com>
1124 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
1126 2014-04-23 Richard Biener <rguenther@suse.de>
1129 * tree-inline.c (declare_return_variable): Use mark_addressable.
1131 2014-04-23 Richard Biener <rguenther@suse.de>
1134 * loop-init.c (loop_optimizer_init): Make sure to apply
1135 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
1137 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1140 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
1142 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
1143 if flag_sanitize_undefined_trap_on_error.
1144 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
1145 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
1146 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
1147 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
1148 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
1149 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
1150 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
1151 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
1152 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
1153 * ubsan.c (ubsan_instrument_unreachable): Return
1154 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1155 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1156 if flag_sanitize_undefined_trap_on_error and
1157 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1158 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1159 instrument_bool_enum_load): Emit __builtin_trap () if
1160 flag_sanitize_undefined_trap_on_error and
1161 __builtin_handle_*_abort () if !flag_sanitize_recover.
1162 * doc/invoke.texi (-fsanitize-recover,
1163 -fsanitize-undefined-trap-on-error): Document.
1165 2014-04-22 Christian Bruel <christian.bruel@st.com>
1167 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1168 Force immediates to SImode.
1170 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1172 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1174 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1175 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1176 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1177 (nios2_fpu_insn): Add entry for round.
1178 (N2FPU_NO_ERRNO_P): Define.
1179 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1181 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1183 2014-04-22 Richard Henderson <rth@redhat.com>
1185 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1186 (add<GPI>3_compare0): Remove leading * from name.
1187 (add<GPI>3_carryin): Likewise.
1188 (sub<GPI>3_compare0): Likewise.
1189 (sub<GPI>3_carryin): Likewise.
1190 (<su_optab>mulditi3): New expander.
1191 (multi3): New expander.
1192 (madd<GPI>): Remove leading * from name.
1194 2014-04-22 Martin Jambor <mjambor@suse.cz>
1196 * cgraphclones.c (cgraph_function_versioning): Copy
1197 ipa_transforms_to_apply instead of asserting it is empty.
1199 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1202 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1203 on count_exp to get mode.
1205 2014-04-22 Andrew Pinski <apinski@cavium.com>
1207 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1208 Handle TLS for ILP32.
1209 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1210 (tlsie_small_<mode>): this and handle PTR.
1211 (tlsie_small_sidi): New pattern.
1212 (tlsle_small): Change to an expand to handle ILP32.
1213 (tlsle_small_<mode>): New pattern.
1214 (tlsdesc_small): Rename to ...
1215 (tlsdesc_small_<mode>): this and handle PTR.
1217 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1219 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1221 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1223 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1224 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1225 (aarch64_types_signed_poly_qualifiers): Likewise.
1226 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1227 (aarch64_types_poly_signed_qualifiers): Likewise.
1228 (TYPES_REINTERP_SS): Type macro added.
1229 (TYPES_REINTERP_SU): Likewise.
1230 (TYPES_REINTERP_SP): Likewise.
1231 (TYPES_REINTERP_US): Likewise.
1232 (TYPES_REINTERP_PS): Likewise.
1233 (aarch64_fold_builtin): New expression folding added.
1234 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1235 Declarations removed.
1236 (REINTERP_SS): Declarations added.
1237 (REINTERP_US): Likewise.
1238 (REINTERP_PS): Likewise.
1239 (REINTERP_SU): Likewise.
1240 (REINTERP_SP): Likewise.
1241 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1242 (vreinterpretq_p8_f64): Likewise.
1243 (vreinterpret_p16_f64): Likewise.
1244 (vreinterpretq_p16_f64): Likewise.
1245 (vreinterpret_f32_f64): Likewise.
1246 (vreinterpretq_f32_f64): Likewise.
1247 (vreinterpret_f64_f32): Likewise.
1248 (vreinterpret_f64_p8): Likewise.
1249 (vreinterpret_f64_p16): Likewise.
1250 (vreinterpret_f64_s8): Likewise.
1251 (vreinterpret_f64_s16): Likewise.
1252 (vreinterpret_f64_s32): Likewise.
1253 (vreinterpret_f64_s64): Likewise.
1254 (vreinterpret_f64_u8): Likewise.
1255 (vreinterpret_f64_u16): Likewise.
1256 (vreinterpret_f64_u32): Likewise.
1257 (vreinterpret_f64_u64): Likewise.
1258 (vreinterpretq_f64_f32): Likewise.
1259 (vreinterpretq_f64_p8): Likewise.
1260 (vreinterpretq_f64_p16): Likewise.
1261 (vreinterpretq_f64_s8): Likewise.
1262 (vreinterpretq_f64_s16): Likewise.
1263 (vreinterpretq_f64_s32): Likewise.
1264 (vreinterpretq_f64_s64): Likewise.
1265 (vreinterpretq_f64_u8): Likewise.
1266 (vreinterpretq_f64_u16): Likewise.
1267 (vreinterpretq_f64_u32): Likewise.
1268 (vreinterpretq_f64_u64): Likewise.
1269 (vreinterpret_s64_f64): Likewise.
1270 (vreinterpretq_s64_f64): Likewise.
1271 (vreinterpret_u64_f64): Likewise.
1272 (vreinterpretq_u64_f64): Likewise.
1273 (vreinterpret_s8_f64): Likewise.
1274 (vreinterpretq_s8_f64): Likewise.
1275 (vreinterpret_s16_f64): Likewise.
1276 (vreinterpretq_s16_f64): Likewise.
1277 (vreinterpret_s32_f64): Likewise.
1278 (vreinterpretq_s32_f64): Likewise.
1279 (vreinterpret_u8_f64): Likewise.
1280 (vreinterpretq_u8_f64): Likewise.
1281 (vreinterpret_u16_f64): Likewise.
1282 (vreinterpretq_u16_f64): Likewise.
1283 (vreinterpret_u32_f64): Likewise.
1284 (vreinterpretq_u32_f64): Likewise.
1286 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1288 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1289 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1290 (vreinterpret_p8_s8): Likewise.
1291 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1292 (vreinterpret_p8_s16): Likewise.
1293 (vreinterpret_p8_s32): Likewise.
1294 (vreinterpret_p8_s64): Likewise.
1295 (vreinterpret_p8_f32): Likewise.
1296 (vreinterpret_p8_u8): Likewise.
1297 (vreinterpret_p8_u16): Likewise.
1298 (vreinterpret_p8_u32): Likewise.
1299 (vreinterpret_p8_u64): Likewise.
1300 (vreinterpret_p8_p16): Likewise.
1301 (vreinterpretq_p8_s8): Likewise.
1302 (vreinterpretq_p8_s16): Likewise.
1303 (vreinterpretq_p8_s32): Likewise.
1304 (vreinterpretq_p8_s64): Likewise.
1305 (vreinterpretq_p8_f32): Likewise.
1306 (vreinterpretq_p8_u8): Likewise.
1307 (vreinterpretq_p8_u16): Likewise.
1308 (vreinterpretq_p8_u32): Likewise.
1309 (vreinterpretq_p8_u64): Likewise.
1310 (vreinterpretq_p8_p16): Likewise.
1311 (vreinterpret_p16_s8): Likewise.
1312 (vreinterpret_p16_s16): Likewise.
1313 (vreinterpret_p16_s32): Likewise.
1314 (vreinterpret_p16_s64): Likewise.
1315 (vreinterpret_p16_f32): Likewise.
1316 (vreinterpret_p16_u8): Likewise.
1317 (vreinterpret_p16_u16): Likewise.
1318 (vreinterpret_p16_u32): Likewise.
1319 (vreinterpret_p16_u64): Likewise.
1320 (vreinterpret_p16_p8): Likewise.
1321 (vreinterpretq_p16_s8): Likewise.
1322 (vreinterpretq_p16_s16): Likewise.
1323 (vreinterpretq_p16_s32): Likewise.
1324 (vreinterpretq_p16_s64): Likewise.
1325 (vreinterpretq_p16_f32): Likewise.
1326 (vreinterpretq_p16_u8): Likewise.
1327 (vreinterpretq_p16_u16): Likewise.
1328 (vreinterpretq_p16_u32): Likewise.
1329 (vreinterpretq_p16_u64): Likewise.
1330 (vreinterpretq_p16_p8): Likewise.
1331 (vreinterpret_f32_s8): Likewise.
1332 (vreinterpret_f32_s16): Likewise.
1333 (vreinterpret_f32_s32): Likewise.
1334 (vreinterpret_f32_s64): Likewise.
1335 (vreinterpret_f32_u8): Likewise.
1336 (vreinterpret_f32_u16): Likewise.
1337 (vreinterpret_f32_u32): Likewise.
1338 (vreinterpret_f32_u64): Likewise.
1339 (vreinterpret_f32_p8): Likewise.
1340 (vreinterpret_f32_p16): Likewise.
1341 (vreinterpretq_f32_s8): Likewise.
1342 (vreinterpretq_f32_s16): Likewise.
1343 (vreinterpretq_f32_s32): Likewise.
1344 (vreinterpretq_f32_s64): Likewise.
1345 (vreinterpretq_f32_u8): Likewise.
1346 (vreinterpretq_f32_u16): Likewise.
1347 (vreinterpretq_f32_u32): Likewise.
1348 (vreinterpretq_f32_u64): Likewise.
1349 (vreinterpretq_f32_p8): Likewise.
1350 (vreinterpretq_f32_p16): Likewise.
1351 (vreinterpret_s64_s8): Likewise.
1352 (vreinterpret_s64_s16): Likewise.
1353 (vreinterpret_s64_s32): Likewise.
1354 (vreinterpret_s64_f32): Likewise.
1355 (vreinterpret_s64_u8): Likewise.
1356 (vreinterpret_s64_u16): Likewise.
1357 (vreinterpret_s64_u32): Likewise.
1358 (vreinterpret_s64_u64): Likewise.
1359 (vreinterpret_s64_p8): Likewise.
1360 (vreinterpret_s64_p16): Likewise.
1361 (vreinterpretq_s64_s8): Likewise.
1362 (vreinterpretq_s64_s16): Likewise.
1363 (vreinterpretq_s64_s32): Likewise.
1364 (vreinterpretq_s64_f32): Likewise.
1365 (vreinterpretq_s64_u8): Likewise.
1366 (vreinterpretq_s64_u16): Likewise.
1367 (vreinterpretq_s64_u32): Likewise.
1368 (vreinterpretq_s64_u64): Likewise.
1369 (vreinterpretq_s64_p8): Likewise.
1370 (vreinterpretq_s64_p16): Likewise.
1371 (vreinterpret_u64_s8): Likewise.
1372 (vreinterpret_u64_s16): Likewise.
1373 (vreinterpret_u64_s32): Likewise.
1374 (vreinterpret_u64_s64): Likewise.
1375 (vreinterpret_u64_f32): Likewise.
1376 (vreinterpret_u64_u8): Likewise.
1377 (vreinterpret_u64_u16): Likewise.
1378 (vreinterpret_u64_u32): Likewise.
1379 (vreinterpret_u64_p8): Likewise.
1380 (vreinterpret_u64_p16): Likewise.
1381 (vreinterpretq_u64_s8): Likewise.
1382 (vreinterpretq_u64_s16): Likewise.
1383 (vreinterpretq_u64_s32): Likewise.
1384 (vreinterpretq_u64_s64): Likewise.
1385 (vreinterpretq_u64_f32): Likewise.
1386 (vreinterpretq_u64_u8): Likewise.
1387 (vreinterpretq_u64_u16): Likewise.
1388 (vreinterpretq_u64_u32): Likewise.
1389 (vreinterpretq_u64_p8): Likewise.
1390 (vreinterpretq_u64_p16): Likewise.
1391 (vreinterpret_s8_s16): Likewise.
1392 (vreinterpret_s8_s32): Likewise.
1393 (vreinterpret_s8_s64): Likewise.
1394 (vreinterpret_s8_f32): Likewise.
1395 (vreinterpret_s8_u8): Likewise.
1396 (vreinterpret_s8_u16): Likewise.
1397 (vreinterpret_s8_u32): Likewise.
1398 (vreinterpret_s8_u64): Likewise.
1399 (vreinterpret_s8_p8): Likewise.
1400 (vreinterpret_s8_p16): Likewise.
1401 (vreinterpretq_s8_s16): Likewise.
1402 (vreinterpretq_s8_s32): Likewise.
1403 (vreinterpretq_s8_s64): Likewise.
1404 (vreinterpretq_s8_f32): Likewise.
1405 (vreinterpretq_s8_u8): Likewise.
1406 (vreinterpretq_s8_u16): Likewise.
1407 (vreinterpretq_s8_u32): Likewise.
1408 (vreinterpretq_s8_u64): Likewise.
1409 (vreinterpretq_s8_p8): Likewise.
1410 (vreinterpretq_s8_p16): Likewise.
1411 (vreinterpret_s16_s8): Likewise.
1412 (vreinterpret_s16_s32): Likewise.
1413 (vreinterpret_s16_s64): Likewise.
1414 (vreinterpret_s16_f32): Likewise.
1415 (vreinterpret_s16_u8): Likewise.
1416 (vreinterpret_s16_u16): Likewise.
1417 (vreinterpret_s16_u32): Likewise.
1418 (vreinterpret_s16_u64): Likewise.
1419 (vreinterpret_s16_p8): Likewise.
1420 (vreinterpret_s16_p16): Likewise.
1421 (vreinterpretq_s16_s8): Likewise.
1422 (vreinterpretq_s16_s32): Likewise.
1423 (vreinterpretq_s16_s64): Likewise.
1424 (vreinterpretq_s16_f32): Likewise.
1425 (vreinterpretq_s16_u8): Likewise.
1426 (vreinterpretq_s16_u16): Likewise.
1427 (vreinterpretq_s16_u32): Likewise.
1428 (vreinterpretq_s16_u64): Likewise.
1429 (vreinterpretq_s16_p8): Likewise.
1430 (vreinterpretq_s16_p16): Likewise.
1431 (vreinterpret_s32_s8): Likewise.
1432 (vreinterpret_s32_s16): Likewise.
1433 (vreinterpret_s32_s64): Likewise.
1434 (vreinterpret_s32_f32): Likewise.
1435 (vreinterpret_s32_u8): Likewise.
1436 (vreinterpret_s32_u16): Likewise.
1437 (vreinterpret_s32_u32): Likewise.
1438 (vreinterpret_s32_u64): Likewise.
1439 (vreinterpret_s32_p8): Likewise.
1440 (vreinterpret_s32_p16): Likewise.
1441 (vreinterpretq_s32_s8): Likewise.
1442 (vreinterpretq_s32_s16): Likewise.
1443 (vreinterpretq_s32_s64): Likewise.
1444 (vreinterpretq_s32_f32): Likewise.
1445 (vreinterpretq_s32_u8): Likewise.
1446 (vreinterpretq_s32_u16): Likewise.
1447 (vreinterpretq_s32_u32): Likewise.
1448 (vreinterpretq_s32_u64): Likewise.
1449 (vreinterpretq_s32_p8): Likewise.
1450 (vreinterpretq_s32_p16): Likewise.
1451 (vreinterpret_u8_s8): Likewise.
1452 (vreinterpret_u8_s16): Likewise.
1453 (vreinterpret_u8_s32): Likewise.
1454 (vreinterpret_u8_s64): Likewise.
1455 (vreinterpret_u8_f32): Likewise.
1456 (vreinterpret_u8_u16): Likewise.
1457 (vreinterpret_u8_u32): Likewise.
1458 (vreinterpret_u8_u64): Likewise.
1459 (vreinterpret_u8_p8): Likewise.
1460 (vreinterpret_u8_p16): Likewise.
1461 (vreinterpretq_u8_s8): Likewise.
1462 (vreinterpretq_u8_s16): Likewise.
1463 (vreinterpretq_u8_s32): Likewise.
1464 (vreinterpretq_u8_s64): Likewise.
1465 (vreinterpretq_u8_f32): Likewise.
1466 (vreinterpretq_u8_u16): Likewise.
1467 (vreinterpretq_u8_u32): Likewise.
1468 (vreinterpretq_u8_u64): Likewise.
1469 (vreinterpretq_u8_p8): Likewise.
1470 (vreinterpretq_u8_p16): Likewise.
1471 (vreinterpret_u16_s8): Likewise.
1472 (vreinterpret_u16_s16): Likewise.
1473 (vreinterpret_u16_s32): Likewise.
1474 (vreinterpret_u16_s64): Likewise.
1475 (vreinterpret_u16_f32): Likewise.
1476 (vreinterpret_u16_u8): Likewise.
1477 (vreinterpret_u16_u32): Likewise.
1478 (vreinterpret_u16_u64): Likewise.
1479 (vreinterpret_u16_p8): Likewise.
1480 (vreinterpret_u16_p16): Likewise.
1481 (vreinterpretq_u16_s8): Likewise.
1482 (vreinterpretq_u16_s16): Likewise.
1483 (vreinterpretq_u16_s32): Likewise.
1484 (vreinterpretq_u16_s64): Likewise.
1485 (vreinterpretq_u16_f32): Likewise.
1486 (vreinterpretq_u16_u8): Likewise.
1487 (vreinterpretq_u16_u32): Likewise.
1488 (vreinterpretq_u16_u64): Likewise.
1489 (vreinterpretq_u16_p8): Likewise.
1490 (vreinterpretq_u16_p16): Likewise.
1491 (vreinterpret_u32_s8): Likewise.
1492 (vreinterpret_u32_s16): Likewise.
1493 (vreinterpret_u32_s32): Likewise.
1494 (vreinterpret_u32_s64): Likewise.
1495 (vreinterpret_u32_f32): Likewise.
1496 (vreinterpret_u32_u8): Likewise.
1497 (vreinterpret_u32_u16): Likewise.
1498 (vreinterpret_u32_u64): Likewise.
1499 (vreinterpret_u32_p8): Likewise.
1500 (vreinterpret_u32_p16): Likewise.
1501 (vreinterpretq_u32_s8): Likewise.
1502 (vreinterpretq_u32_s16): Likewise.
1503 (vreinterpretq_u32_s32): Likewise.
1504 (vreinterpretq_u32_s64): Likewise.
1505 (vreinterpretq_u32_f32): Likewise.
1506 (vreinterpretq_u32_u8): Likewise.
1507 (vreinterpretq_u32_u16): Likewise.
1508 (vreinterpretq_u32_u64): Likewise.
1509 (vreinterpretq_u32_p8): Likewise.
1510 (vreinterpretq_u32_p16): Likewise.
1512 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1514 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1516 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1519 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1520 (vqnegd_s64): Likewise.
1521 (vqabs_s64): Likewise.
1522 (vqabsd_s64): Likewise.
1524 2014-04-22 Richard Henderson <rth@redhat.com>
1526 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1527 computation to the top of the loop.
1529 2014-04-22 Renlin <renlin.li@arm.com>
1530 Jiong Wang <jiong.wang@arm.com>
1532 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1533 * config/aarch64/aarch64.c (aarch64_layout_frame)
1534 (aarch64_initial_elimination_offset): Likewise.
1536 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1538 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1541 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1543 * machmode.h (bitwise_mode_for_mode): Declare.
1544 * stor-layout.h (bitwise_type_for_mode): Likewise.
1545 * stor-layout.c (bitwise_mode_for_mode): New function.
1546 (bitwise_type_for_mode): Likewise.
1547 * builtins.c (fold_builtin_memory_op): Use it instead of
1548 int_mode_for_mode and build_nonstandard_integer_type.
1550 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1552 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1553 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1554 (*-*-solaris2*): Simplify.
1555 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1556 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1557 *-*-solaris2.9* handling.
1559 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1561 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1562 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1564 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1565 * configure: Regenerate.
1567 * config/i386/sol2-9.h: Remove.
1569 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1570 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1571 Remove Solaris 9 references.
1573 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1575 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1576 (floatuns<GPI:mode><GPF:mode>2): Remove.
1577 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1578 and floatuns conversions.
1579 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1580 and floatuns conversions.
1581 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1582 (w1,w2): New mode attributes for inequal width conversions.
1584 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1586 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1587 the output asm format.
1589 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1591 * config/aarch64/aarch64-simd.md
1592 (aarch64_cm<optab>di): Always split.
1593 (*aarch64_cm<optab>di): New.
1594 (aarch64_cmtstdi): Always split.
1595 (*aarch64_cmtstdi): New.
1597 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1599 PR tree-optimization/60823
1600 * omp-low.c (ipa_simd_modify_function_body): Go through
1601 all SSA_NAMEs and for those refering to vector arguments
1602 which are going to be replaced adjust SSA_NAME_VAR and,
1603 if it is a default definition, change it into a non-default
1604 definition assigned at the beginning of function from new_decl.
1605 (ipa_simd_modify_stmt_ops): Rewritten.
1606 * tree-dfa.c (set_ssa_default_def): When removing default def,
1607 check for NULL loc instead of NULL *loc.
1609 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1611 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1612 restrictions on core registers for DImode values in Thumb2.
1614 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1616 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1617 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1619 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1621 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1622 (*iordi_notzesidi_di): Likewise.
1623 (*iordi_notsesidi_di): Likewise.
1625 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1627 * config/arm/arm-protos.h (tune_params): New struct members.
1628 * config/arm/arm.c: Initialise tune_params per processor.
1629 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1630 for speed, based on new tune_params.
1632 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1634 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1636 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1638 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1640 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1641 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1642 (vrnda_f64): Likewise.
1643 (vrndi_f64): Likewise.
1644 (vrndm_f64): Likewise.
1645 (vrndn_f64): Likewise.
1646 (vrndp_f64): Likewise.
1647 (vrndx_f64): Likewise.
1649 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1651 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1652 GET_MODE_SIZE argument is enum machine_mode.
1654 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1657 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1658 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1660 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1663 * asan.c (asan_emit_stack_protection): Force the base to align to
1664 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1665 appropriate bits if STRICT_ALIGNMENT.
1666 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1668 (expand_used_vars): Leave a space in the stack frame for alignment
1669 if STRICT_ALIGNMENT.
1671 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1673 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1675 (gimple_store_p): Likewise.
1676 (gimple_assign_load_p): Likewise.
1677 (gimple_assign_cast_p): Likewise.
1678 (gimple_clobber_p): Likewise.
1680 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1681 rather than a gimple.
1682 (gimple_assign_cast_p): Likewise.
1684 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1687 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1688 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1690 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1691 more debug information for E500 if -mdebug=reg.
1693 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1696 * config/i386/i386.c (ix86_expand_builtin)
1697 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1698 register for target RTX.
1699 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1701 2014-04-18 Cong Hou <congh@google.com>
1703 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1704 the widen-mult pattern by handling two operands with different sizes,
1705 and operands whose size is smaller than half of the result type.
1707 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1709 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1710 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1711 (do_estimate_edge_time): Compute it.
1712 * ipa-inline.c (want_inline_small_function_p): Bypass
1713 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1715 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1717 * ipa-inline.c (spec_rem): New static variable.
1718 (dump_overall_stats): New function.
1719 (dump_inline_stats): New function.
1721 2014-04-18 Richard Henderson <rth@redhat.com>
1723 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1724 to GET_MODE_SIZE, not a reg_class_t.
1726 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1728 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1729 (vsx_xxmrglw_<mode>): Likewise.
1731 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1734 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1735 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1736 (rs6000_init_hard_regno_mode_ok): Likewise.
1738 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1740 * ipa-inline.c (inline_small_functions): Account only non-cold
1742 * doc/invoke.texi (inline-unit-growth): Update documentation.
1744 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1746 * config/rs6000/rs6000.md (addti3, subti3): New.
1748 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1751 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1752 comment. Check optimize_insn_for_size_p instead of
1753 optimize_insn_for_speed_p.
1755 2014-04-17 Martin Jambor <mjambor@suse.cz>
1757 * gimple-iterator.c (gsi_start_edge): New function.
1758 * gimple-iterator.h (gsi_start_edge): Declare.
1759 * tree-sra.c (single_non_eh_succ): New function.
1760 (disqualify_ops_if_throwing_stmt): Renamed to
1761 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1762 having one non-EH successor BB.
1763 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1764 generate loads into replacements.
1765 (sra_modify_assign): Likewise and and also use the simple path for
1767 (sra_modify_function_body): Commit statements on edges.
1769 2014-04-17 Richard Biener <rguenther@suse.de>
1772 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1773 comparison results and add clarifying comment.
1775 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1777 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1778 (blank_mode): Initialize it.
1779 (emit_mode_size_inline, emit_mode_nunits_inline,
1780 emit_mode_inner_inline): New functions.
1781 (emit_insn_modes_h): Call them and surround their output with
1782 #if GCC_VERSION >= 4001 ... #endif.
1783 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1784 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1785 mode_* arrays if the argument is __builtin_constant_p.
1786 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1787 is enum machine_mode.
1789 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1791 * passes.c (opt_pass::execute): Adjust.
1792 (pass_manager::execute_pass_mode_switching): Likewise.
1793 (early_local_passes::execute): Likewise.
1794 (execute_one_pass): Pass cfun to the pass's execute method.
1795 * tree-pass.h (opt_pass::execute): Add function * argument.
1796 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1797 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1798 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1799 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1800 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1801 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1802 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1803 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1804 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1805 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1806 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1807 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1808 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1809 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1810 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1811 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1812 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1813 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1814 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1815 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1816 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1817 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1818 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1819 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1820 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1821 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1822 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1823 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1824 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1827 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1829 * passes.c (opt_pass::gate): Take function * argument.
1830 (gate_all_early_local_passes): Merge into
1831 (early_local_passes::gate): this.
1832 (gate_all_early_optimizations): Merge into
1833 (all_early_optimizations::gate): this.
1834 (gate_all_optimizations): Mege into
1835 (all_optimizations::gate): this.
1836 (gate_all_optimizations_g): Merge into
1837 (all_optimizations_g::gate): this.
1838 (gate_rest_of_compilation): Mege into
1839 (rest_of_compilation::gate): this.
1840 (gate_postreload): Merge into
1841 (postreload::gate): this.
1842 (dump_one_pass): Pass cfun to the pass's gate method.
1843 (execute_ipa_summary_passes): Likewise.
1844 (execute_one_pass): Likewise.
1845 (ipa_write_summaries_2): Likewise.
1846 (ipa_write_optimization_summaries_1): Likewise.
1847 (ipa_read_summaries_1): Likewise.
1848 (ipa_read_optimization_summaries_1): Likewise.
1849 (execute_ipa_stmt_fixups): Likewise.
1850 * tree-pass.h (opt_pass::gate): Add function * argument.
1851 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
1852 combine-stack-adj.c, combine.c, compare-elim.c,
1853 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1854 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
1855 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
1856 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
1857 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1858 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1859 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1860 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
1861 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
1862 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
1863 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1864 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1865 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
1866 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1867 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1868 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1869 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1870 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1871 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1872 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1873 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1874 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1875 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1876 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
1877 var-tracking.c, vtable-verify.c, web.c: Adjust.
1879 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1881 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
1882 * configure: Regenerate.
1884 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1886 * passes.c (dump_one_pass): don't check pass->has_gate.
1887 (execute_ipa_summary_passes): Likewise.
1888 (execute_one_pass): Likewise.
1889 (ipa_write_summaries_2): Likewise.
1890 (ipa_write_optimization_summaries_1): Likewise.
1891 (ipa_read_optimization_summaries_1): Likewise.
1892 (execute_ipa_stmt_fixups): Likewise.
1893 * tree-pass.h (pass_data::has_gate): Remove.
1894 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1895 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1896 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1897 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1898 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1899 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
1900 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
1901 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
1902 gimple-low.c, gimple-ssa-isolate-paths.c,
1903 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
1904 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
1905 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1906 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
1907 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
1908 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
1909 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
1910 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1911 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
1912 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1913 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1914 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1915 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1916 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1917 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1918 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1919 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1920 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1921 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1922 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1923 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1924 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1927 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1929 * pass_manager.h (pass_manager::register_dump_files_1): Remove
1931 * passes.c (pass_manager::register_dump_files_1): Merge into
1932 (pass_manager::register_dump_files): this, and remove its handling of
1933 properties since the pass always has the properties anyway.
1934 (pass_manager::pass_manager): Adjust.
1936 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1938 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
1939 * passes.c (pass_manager::register_dump_files_1): Remove dead code
1940 dealing with properties.
1941 (pass_manager::register_dump_files): Adjust.
1943 2014-03-20 Mark Wielaard <mjw@redhat.com>
1945 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
1946 then represent the bound as normal constant value.
1948 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1951 Forward port from 4.8 branch
1952 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
1954 * config/i386/bmiintrin.h (_blsi_u32): New.
1958 (_blsmsk_u32): Ditto.
1959 (_blsmsk_u64): Ditto.
1960 (_tzcnt_u32): Ditto.
1961 (_tzcnt_u64): Ditto.
1963 2014-04-17 Kito Cheng <kito@0xlab.org>
1965 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
1967 2014-04-17 Richard Biener <rguenther@suse.de>
1970 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
1971 boolean results for comparisons.
1973 2014-04-17 Richard Biener <rguenther@suse.de>
1975 PR tree-optimization/60836
1976 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
1977 initial PHI args to be gimple values.
1979 2014-04-17 Richard Biener <rguenther@suse.de>
1981 PR tree-optimization/60841
1982 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
1983 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
1984 of stmts to SLP build.
1985 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
1986 (vect_analyze_slp): Likewise.
1987 (vect_analyze_slp_instance): Likewise.
1988 (vect_build_slp_tree): Limit overall SLP tree growth.
1989 * tree-vectorizer.h (vect_analyze_data_refs,
1990 vect_analyze_slp): Adjust prototypes.
1992 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1994 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
1997 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1999 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
2000 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
2001 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
2002 for TARGET_SLOW_PSHUFB
2004 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2006 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
2007 * config/i386/i386.c (intel_cost): Ditto.
2009 2014-04-17 Joey Ye <joey.ye@arm.com>
2011 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
2013 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2015 * opts.c (common_handle_option): Disable -fipa-reference coorectly
2018 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2020 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
2021 (type_all_derivations_known_p): New predicate.
2022 (type_all_ctors_visible_p): New predicate.
2023 (type_possibly_instantiated_p): New predicate.
2024 (get_odr_type): Compute all_derivations_known.
2025 (dump_odr_type): Dump the flag.
2026 (maybe_record_type): Cleanup.
2027 (record_target_from_binfo): Add bases_to_consider array;
2028 record bases for types w/o instances and skip CXX destructor.
2029 (possible_polymorphic_call_targets_1): Add bases_to_consider
2030 and consider_construction parameters; check if type may have instance.
2031 (get_polymorphic_call_info): Set maybe_in_construction to true
2032 when we know nothing.
2033 (record_targets_from_bases): Skip CXX destructors; they are
2034 never called for types in construction.
2035 (possible_polymorphic_call_targets): Do not record target when
2036 type may not have instance.
2038 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2041 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
2042 external aliases alive, too.
2044 2014-04-16 Andrew Pinski <apinski@cavium.com>
2046 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
2049 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2051 * final.c (compute_alignments): Do not apply loop alignment to a block
2052 falling through to the exit.
2054 2014-04-16 Catherine Moore <clm@codesourcery.com>
2056 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
2057 Adjust constraints for microMIPS store patterns.
2059 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2061 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
2063 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2065 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
2066 (append_use): Run at -O0.
2067 (append_vdef): Likewise.
2068 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
2069 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
2071 2014-04-16 Jakub Jelinek <jakub@redhat.com>
2073 PR tree-optimization/60844
2074 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
2075 (propagate_op_to_single_use, remove_visited_stmt_chain,
2076 linearize_expr, repropagate_negates, reassociate_bb): Use it
2077 instead of gsi_remove.
2079 2014-04-16 Martin Jambor <mjambor@suse.cz>
2081 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
2082 ipa_transforms_to_apply.
2083 (cgraph_function_versioning): Assert that old_node has empty
2084 ipa_transforms_to_apply.
2085 * trans-mem.c (ipa_tm_create_version): Likewise.
2086 * tree-inline.c (tree_function_versioning): Do not duplicate
2087 ipa_transforms_to_apply.
2089 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2092 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
2094 Pass necessary as flags on 64-bit Solaris/x86.
2095 Use lowercase relocs for x86_64-*-*.
2096 * configure: Regenerate.
2098 2014-04-15 Jan Hubicka <jh@suse.cz>
2100 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
2101 (maybe_record_node, likely_target_p): Use it.
2103 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2106 Revert following patch
2108 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2111 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2112 software floating point or no floating point registers, do not
2113 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2114 in GPRs that occurs after we tested for GPRs that would never be
2117 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2118 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2119 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2120 specifically allow DDmode, since that does not use the SPE SIMD
2123 2014-03-21 Mark Wielaard <mjw@redhat.com>
2125 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
2126 as unsigned or int depending on type and value used.
2128 2014-04-15 Richard Biener <rguenther@suse.de>
2130 PR rtl-optimization/56965
2131 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
2132 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
2134 * alias.c (true_dependence_1): Do not call
2135 nonoverlapping_component_refs_p.
2136 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
2137 nonoverlapping_component_refs_p.
2138 (indirect_refs_may_alias_p): Likewise.
2140 2014-04-15 Teresa Johnson <tejohnson@google.com>
2142 * cfg.c (dump_bb_info): Fix flags check.
2143 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
2145 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2147 PR rtl-optimization/60663
2148 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
2151 2014-04-15 Richard Biener <rguenther@suse.de>
2153 * lto-streamer.h (LTO_major_version): Bump to 4.
2155 2014-04-15 Richard Biener <rguenther@suse.de>
2157 * common.opt (lto_partition_model): New enum.
2158 (flto-partition=): Merge separate options with a single with argument,
2159 add -flto-partition=one support.
2160 * flag-types.h (enum lto_partition_model): Declare.
2161 * opts.c (finish_options): Remove duplicate -flto-partition=
2163 * lto-wrapper.c (run_gcc): Adjust.
2165 2014-04-15 Richard Biener <rguenther@suse.de>
2167 * alias.c (ncr_compar): New function.
2168 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2170 2014-04-15 Richard Biener <rguenther@suse.de>
2172 * alias.c (record_component_aliases): Do not walk BINFOs.
2174 2014-04-15 Richard Biener <rguenther@suse.de>
2176 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2177 Add struct function argument and adjust.
2178 (find_func_aliases_for_call): Likewise.
2179 (find_func_aliases): Likewise.
2180 (find_func_clobbers): Likewise.
2181 (intra_create_variable_infos): Likewise.
2182 (compute_points_to_sets): Likewise.
2183 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2185 2014-04-15 Richard Biener <rguenther@suse.de>
2187 * tree.c (iterative_hash_expr): Use enum tree_code_class
2188 to store TREE_CODE_CLASS.
2189 (tree_block): Likewise.
2190 (tree_set_block): Likewise.
2191 * tree.h (fold_build_pointer_plus_loc): Use
2192 convert_to_ptrofftype_loc.
2194 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2197 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2200 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2202 * cfgloop.h (struct loop): Move force_vectorize down.
2203 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2204 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2205 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2206 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2207 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2208 * tree-core.h (enum annot_expr_kind): Add new kind values.
2209 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2210 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2212 * tree.def (ANNOTATE_EXPR): Tweak comment.
2214 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2216 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2219 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2221 * tree.h (TYPE_IDENTIFIER): Declare.
2222 * tree.c (subrange_type_for_debug_p): Use it.
2223 * godump.c (go_format_type): Likewise.
2224 * dwarf2out.c (is_cxx_auto, modified_type_die,
2225 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2226 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2228 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2231 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2233 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2235 * config/i386/i386.c (examine_argument): Return bool. Return true if
2236 parameter should be passed in memory.
2237 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2238 (construct_container): Update calls to examine_argument.
2239 (function_arg_advance_64): Ditto.
2240 (return_in_memory_32): Merge with ix86_return_in_memory.
2241 (return_in_memory_64): Ditto.
2242 (return_in_memory_ms_64): Ditto.
2244 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2246 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2247 * coverage.c (coverage_compute_profile_id): Handle externally visible
2250 2014-04-14 Martin Jambor <mjambor@suse.cz>
2252 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2253 DECL_DISREGARD_INLINE_LIMITS functions.
2255 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2258 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2260 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2263 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2264 optimize_insn_for_speed_p instead of
2265 optimize_function_for_speed_p.
2267 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2269 * doc/invoke.texi (free): Document AArch64.
2271 2014-04-14 Richard Biener <rguenther@suse.de>
2273 PR tree-optimization/60042
2274 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2275 (insert_into_preds_of_block): Do not prevent PHI insertion
2276 for REFERENCE exprs here ...
2277 (eliminate_dom_walker::before_dom_children): ... but prevent
2278 their use here under similar conditions when applied to the
2279 IL after PRE optimizations.
2281 2014-04-14 Richard Biener <rguenther@suse.de>
2283 * passes.def: Move early points-to after early SRA.
2285 2014-04-14 Richard Biener <rguenther@suse.de>
2287 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2288 check for which sign-changes we allow when forwarding
2289 a converted value into a switch.
2291 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2293 * stor-layout.c (place_field): Finalize non-constant offset for the
2296 2014-04-14 Richard Biener <rguenther@suse.de>
2298 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2300 (expand_switch_using_bit_tests_p): Likewise.
2301 (process_switch): Compute and pass on speed_p based on the
2303 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2304 optimize_bb_for_speed_p.
2306 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2308 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2309 * function.h (struct function): Rename has_force_vect_loops into
2310 has_force_vectorize_loops.
2311 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2312 (input_struct_function_base): Likewise.
2313 * lto-streamer-out.c (output_cfg): Likewise.
2314 (output_struct_function_base): Likewise.
2315 * omp-low.c (expand_omp_simd): Likewise.
2316 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2317 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2318 (version_loop_for_if_conversion): Likewise.
2319 (tree_if_conversion): Likewise.
2320 (main_tree_if_conversion): Likewise.
2321 (gate_tree_if_conversion): Likewise.
2322 * tree-inline.c (copy_loops): Likewise.
2323 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2324 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2325 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2326 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2327 * tree-vectorizer.c (vectorize_loops): Likewise.
2328 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2330 2014-04-14 Richard Biener <rguenther@suse.de>
2333 * lto-streamer-out.c (wrap_refs): New function.
2334 (lto_output): Wrap symbol references in global initializes in
2335 type-preserving MEM_REFs.
2337 2014-04-14 Christian Bruel <christian.bruel@st.com>
2339 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2341 2014-04-14 Christian Bruel <christian.bruel@st.com>
2343 * config/sh/sh.md (setmemqi): New expand pattern.
2344 * config/sh/sh.h (CLEAR_RATIO): Define.
2345 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2346 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2348 2014-04-14 Richard Biener <rguenther@suse.de>
2351 * fold-const.c (negate_expr_p): Don't negate directional rounding
2353 (fold_negate_expr): Likewise.
2355 2014-04-14 Richard Biener <rguenther@suse.de>
2357 PR tree-optimization/59817
2358 PR tree-optimization/60453
2359 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2360 recursion to catch all CHRECs in the scalar evolution and restrict
2361 the predicate for the remains appropriately.
2363 2014-04-12 Catherine Moore <clm@codesourcery.com>
2365 * config/mips/constraints.md: Add new register constraint "kb".
2366 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2367 (*movhi_internal): Likewise.
2368 (*movqi_internal): Likewise.
2369 * config/mips/mips.h (M16_STORE_REGS): New register class.
2370 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2371 (REG_CLASS_CONTENTS): Likewise.
2372 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2374 2014-04-11 Tobias Burnus <burnus@net-b.de>
2377 * doc/invoke.texi (-Wformat-signedness): Document it.
2378 (Wformat=2): Mention that this enables -Wformat-signedness.
2380 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2382 * common/config/epiphany/epiphany-common.c
2383 (epiphany_option_optimization_table): Enable section anchors by
2384 default at -O1 or higher.
2385 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2386 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2387 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2388 carries no extra cost.
2389 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2390 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2391 * config/epiphany/predicates.md (memclob_operand): New predicate.
2392 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2393 Use memclob_operand predicate and X constraint for operand 3.
2395 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2397 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2398 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2401 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2403 PR rtl-optimization/60651
2404 * mode-switching.c (optimize_mode_switching): Make sure to emit
2405 sets of a lower numbered entity before sets of a higher numbered
2406 entity to a mode of the same or lower priority.
2407 When creating a seginfo for a basic block that starts with a code
2408 label, move the insertion point past the code label.
2409 (new_seginfo): Document and enforce requirement that
2410 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2411 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2412 * doc/tm.texi: Regenerate.
2414 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2417 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2419 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2421 * BASE-VER: Set to 4.10.0.
2423 2014-04-11 Tobias Burnus <burnus@net-b.de>
2426 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2427 * doc/gcc.texi (Service): Update description in the @menu
2428 * doc/invoke.texi (Option Summary): Remove misplaced and
2431 2014-04-11 Steve Ellcey <sellcey@mips.com>
2432 Jakub Jelinek <jakub@redhat.com>
2435 * expr.c (convert_move): Use emit_store_flag_force instead of
2436 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2439 2014-04-11 Richard Biener <rguenther@suse.de>
2442 * varasm.c (assemble_alias): Avoid endless error reporting
2443 recursion by setting TREE_ASM_WRITTEN.
2445 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2447 * config/s390/s390.md: Add a splitter for NOT rtx.
2449 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2451 PR rtl-optimization/60663
2452 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2454 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2455 Jakub Jelinek <jakub@redhat.com>
2458 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2459 flag from decl_node to node.
2461 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2464 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2465 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2466 ameliorating the cases where it can be.
2468 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2471 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2473 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2474 (loadsync_<mode>): Change mode.
2475 (load_quadpti, store_quadpti): New.
2476 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2477 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2478 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2480 2014-04-09 Cong Hou <congh@google.com>
2483 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2486 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2488 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2489 instead of vnor to exploit possible fusion opportunity in the
2491 (altivec_expand_vec_perm_const_le): Likewise.
2493 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2495 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2496 (loadsync_<mode>): Change mode.
2497 (load_quadpti, store_quadpti): New.
2498 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2499 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2501 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2504 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2505 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2506 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2508 2014-04-08 Richard Biener <rguenther@suse.de>
2511 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2512 a 64bit widest int print double-int similar to on HWI64 hosts.
2514 2014-04-08 Richard Biener <rguenther@suse.de>
2516 PR tree-optimization/60785
2517 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2518 default defs properly.
2520 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2522 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2523 (Weffc++): Likewise.
2525 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2527 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2528 set completep to false rather than true.
2530 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2533 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2534 ARM_TARGET2_DWARF_FORMAT.
2536 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2539 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2540 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2543 2014-04-07 Richard Biener <rguenther@suse.de>
2545 PR tree-optimization/60766
2546 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2547 (may_eliminate_iv): Convert cand_value_at result to desired type.
2549 2014-04-07 Jason Merrill <jason@redhat.com>
2552 * common.opt (-fno-gnu-unique): Add.
2553 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2555 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2557 * haifa-sched.c: Fix outdated function reference and minor
2558 grammar errors in introductory comment.
2560 2014-04-07 Richard Biener <rguenther@suse.de>
2563 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2565 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2567 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2570 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2571 size accounting for thunks.
2572 (pa_asm_output_mi_thunk): Use final_start_function() and
2573 final_end_function() to output function start and end directives.
2575 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2577 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2578 device specific ISA/ feature information. Remove short_sp and
2579 errata_skip ds. Add avr_device_specific_features enum to have device
2581 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2582 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2583 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2584 updated device specific info.
2585 * config/avr/avr-mcus.def: Merge device specific details to
2586 dev_attribute field.
2587 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2589 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2590 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2591 assembler if RMW isa supported by current device.
2592 * config/avr/genmultilib.awk: Update as device info structure changed.
2593 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2595 2014-04-04 Cong Hou <congh@google.com>
2597 PR tree-optimization/60656
2598 * tree-vect-stmts.c (supportable_widening_operation):
2599 Fix a bug that elements in a vector with vect_used_by_reduction
2600 property are incorrectly reordered when the operation on it is not
2601 consistant with the one in reduction operation.
2603 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2605 PR rtl-optimization/60155
2606 * gcse.c (record_set_data): New function.
2607 (single_set_gcse): New function.
2608 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2609 (hoist_code): Likewise.
2610 (get_pressure_class_and_nregs): Likewise.
2612 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2614 * explow.c (probe_stack_range): Emit a final optimization blockage.
2616 2014-04-04 Anthony Green <green@moxielogic.com>
2618 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2621 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2624 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2625 flags are set only during streaming.
2626 * ipa.c (process_references, walk_polymorphic_call_targets,
2627 symtab_remove_unreachable_nodes): Drop bodies of always inline
2628 after early inlining.
2629 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2631 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2632 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2635 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2638 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2641 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2643 (cortex_a53_fdivd): Likewise.
2645 2014-04-04 Martin Jambor <mjambor@suse.cz>
2648 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2650 * cgraph.c (clone_of_p): Also return true if thunks match.
2651 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2652 cgraph_function_or_thunk_node and an obsolete comment.
2653 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2655 (build_function_decl_skip_args): Likewise.
2656 (set_new_clone_decl_and_node_flags): New function.
2657 (duplicate_thunk_for_node): Likewise.
2658 (redirect_edge_duplicating_thunks): Likewise.
2659 (cgraph_clone_node): New parameter args_to_skip, pass it to
2660 redirect_edge_duplicating_thunks which is called instead of
2661 cgraph_redirect_edge_callee.
2662 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2663 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2665 2014-04-04 Jeff Law <law@redhat.com>
2668 * config/arm/predicates.md (const_int_I_operand): New predicate.
2669 (const_int_M_operand): Similarly.
2670 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2672 (insv_t2, extv_reg, extzv_t2): Likewise.
2673 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2674 (pop_multiple_with_writeback_and_return): Likewise.
2675 (vfp_pop_multiple_with_writeback): Likewise
2677 2014-04-04 Richard Biener <rguenther@suse.de>
2680 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2681 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2683 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2684 * gimplify.c (gimple_add_tmp_var_fn): New function.
2685 * gimple-expr.h (create_tmp_reg_fn): Declare.
2686 * gimple-expr.c (create_tmp_reg_fn): New function.
2687 * gimple-low.c (record_vars_into): Don't change cfun.
2688 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2689 code generation without cfun.
2691 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2694 * Makefile.in (install-driver): Fix shell scripting.
2696 2014-04-03 Cong Hou <congh@google.com>
2698 PR tree-optimization/60505
2699 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2700 threshold of number of iterations below which no vectorization
2702 * tree-vect-loop.c (new_loop_vec_info):
2703 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2704 * tree-vect-loop.c (vect_analyze_loop_operations):
2705 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2706 * tree-vect-loop.c (vect_transform_loop):
2707 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2708 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2709 of iterations of the loop and see if we should build the epilogue.
2711 2014-04-03 Richard Biener <rguenther@suse.de>
2713 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2714 (streamer_tree_cache_create): Adjust.
2715 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2716 to allow optional nodes array.
2717 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2718 (streamer_tree_cache_append): Likewise.
2719 (streamer_tree_cache_create): Create nodes array optionally
2720 as specified by parameter.
2721 * lto-streamer-out.c (create_output_block): Avoid maintaining
2722 the node array in the writer cache.
2723 (DFS_write_tree): Remove assertion.
2724 (produce_asm_for_decls): Free the out decl state hash table early.
2725 * lto-streamer-in.c (lto_data_in_create): Adjust for
2726 streamer_tree_cache_create prototype change.
2728 2014-04-03 Richard Biener <rguenther@suse.de>
2730 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2731 set TREE_CHAIN to NULL_TREE.
2733 2014-04-03 Richard Biener <rguenther@suse.de>
2735 PR tree-optimization/60740
2736 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2737 over all GIMPLE_COND operands.
2739 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2741 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2742 (Weffc++): Remove Scott's numbering, merge lists and reference
2745 2014-04-03 Nick Clifton <nickc@redhat.com>
2747 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2750 2014-04-03 Martin Jambor <mjambor@suse.cz>
2752 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2753 mention gcc_unreachable before failing.
2754 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2757 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
2760 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2761 inconsistent code and instead mark the context inconsistent.
2762 (possible_polymorphic_call_targets): For inconsistent contexts
2763 return empty complete list.
2765 2014-04-02 Anthony Green <green@moxielogic.com>
2767 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2768 (extendqisi2, extendhisi2): Define.
2769 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2770 (WCHAR_TYPE): Change to unsigned int.
2772 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2774 PR tree-optimization/60733
2775 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2776 insertion point for PHI candidates to be the end of the feeding
2777 block for the PHI argument.
2779 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2781 PR rtl-optimization/60650
2782 * lra-constraints.c (process_alt_operands): Decrease reject for
2783 earlyclobber matching.
2785 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2787 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2789 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2791 * config/spu/spu.c (pad_bb): Do not crash when the last
2792 insn is CODE_FOR_blockage.
2794 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2796 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2797 lies outside the target mode.
2799 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2802 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2803 software floating point or no floating point registers, do not
2804 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2805 in GPRs that occurs after we tested for GPRs that would never be
2808 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2809 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2810 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2811 specifically allow DDmode, since that does not use the SPE SIMD
2814 2014-04-02 Richard Biener <rguenther@suse.de>
2817 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2818 MODE_INTs. Properly use negv_optab.
2819 (expand_abs): Likewise.
2821 2014-04-02 Richard Biener <rguenther@suse.de>
2824 * Makefile.in (install-driver): Guard extra installs with special
2827 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
2829 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2832 2014-04-01 Richard Henderson <rth@redhat.com>
2835 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
2836 alternative enabled before register allocation.
2838 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
2840 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
2841 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
2843 (nios2_large_got_address): Remove unneeded 'sym' parameter.
2844 (nios2_got_address): Update nios2_large_got_address call site.
2845 (nios2_delegitimize_address): New function.
2846 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
2847 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
2848 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
2850 2014-04-01 Martin Husemann <martin@duskware.de>
2852 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
2855 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
2857 PR rtl-optimization/60604
2858 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
2859 check from register_operand.
2860 (register_operand): Redefine in terms of general_operand.
2861 (nonmemory_operand): Use register_operand for the non-constant cases.
2863 2014-04-01 Richard Biener <rguenther@suse.de>
2865 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
2867 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
2869 * doc/invoke.texi (mapp-regs): Clarify.
2871 2014-03-31 Ulrich Drepper <drepper@gmail.com>
2873 * config/i386/avx512fintrin.h (__v32hi): Define type.
2874 (__v64qi): Likewise.
2875 (_mm512_set1_epi8): Define.
2876 (_mm512_set1_epi16): Define.
2877 (_mm512_set4_epi32): Define.
2878 (_mm512_set4_epi64): Define.
2879 (_mm512_set4_pd): Define.
2880 (_mm512_set4_ps): Define.
2881 (_mm512_setr4_epi64): Define.
2882 (_mm512_setr4_epi32): Define.
2883 (_mm512_setr4_pd): Define.
2884 (_mm512_setr4_ps): Define.
2885 (_mm512_setzero_epi32): Define.
2887 2014-03-31 Martin Jambor <mjambor@suse.cz>
2890 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
2891 callsite_arguments_match_p. Updated all callers. Also check types of
2892 corresponding formal parameters and actual arguments.
2893 (not_all_callers_have_enough_arguments_p) Renamed to
2894 some_callers_have_mismatched_arguments_p.
2896 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
2898 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
2900 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
2903 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
2906 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
2908 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
2909 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
2911 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
2912 Use FMAMODE_NOVF512 mode iterator.
2913 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
2914 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
2915 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
2916 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
2918 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
2919 Use VF_128_256 mode iterator.
2920 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
2923 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2925 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
2926 static chain if needed.
2928 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2931 * lra-constraints.c (index_part_to_reg): New.
2932 (process_address): Use it.
2934 2014-03-27 Jeff Law <law@redhat.com>
2935 Jakub Jelinek <jakub@redhat.com>
2938 * expr.c (do_tablejump): Use simplify_gen_binary rather than
2939 gen_rtx_{PLUS,MULT} to build up the address expression.
2941 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
2942 creating non-canonical RTL.
2944 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2947 * ipa-inline.c (want_inline_small_function_p): Short circuit large
2948 functions; reorganize to make cheap checks first.
2949 (inline_small_functions): Do not estimate growth when dumping;
2951 * ipa-inline.h (inline_summary): Add min_size.
2952 (growth_likely_positive): New function.
2953 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
2954 (set_cond_stmt_execution_predicate): Cleanup.
2955 (estimate_edge_size_and_time): Compute min_size.
2956 (estimate_calls_size_and_time): Likewise.
2957 (estimate_node_size_and_time): Likewise.
2958 (inline_update_overall_summary): Update min_size.
2959 (do_estimate_edge_time): Likewise.
2960 (do_estimate_edge_size): Update.
2961 (do_estimate_edge_hints): Update.
2962 (growth_likely_positive): New function.
2964 2014-03-28 Jakub Jelinek <jakub@redhat.com>
2967 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
2968 also if addr has VOIDmode.
2970 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2972 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
2973 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
2975 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
2976 instructions as well as AdvancedSIMD loads.
2978 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2980 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
2981 Use crypto_aese type.
2982 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
2983 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
2984 crypto_aese, crypto_aesmc. Move to types.md.
2985 * config/arm/types.md (crypto_aes): Split into crypto_aese,
2987 * config/arm/iterators.md (crypto_type): Likewise.
2989 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2991 * cgraph.c: Include expr.h and tree-dfa.h.
2992 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
2995 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2998 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
2999 regs from checking multi-reg pseudos.
3001 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3003 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
3005 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3007 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
3008 if it would clobber the stack pointer, even temporarily.
3010 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
3012 * mode-switching.c: Make small adjustments to the top comment.
3014 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3016 * config/rs6000/constraints.md (wD constraint): New constraint to
3017 match the constant integer to get the top DImode/DFmode out of a
3018 vector in a VSX register.
3020 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
3021 match the constant integer to get the top DImode/DFmode out of a
3022 vector in a VSX register.
3024 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
3027 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
3030 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
3031 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
3033 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
3034 Optimize vec_extract of 64-bit values, where the value being
3035 extracted is in the top word, where we can use scalar
3036 instructions. Add direct move and store support. Combine the big
3037 endian/little endian vector select load support into a single insn.
3038 (vsx_extract_<mode>_internal1): Likewise.
3039 (vsx_extract_<mode>_internal2): Likewise.
3040 (vsx_extract_<mode>_load): Likewise.
3041 (vsx_extract_<mode>_store): Likewise.
3042 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
3043 combined into vsx_extract_<mode>_load.
3044 (vsx_extract_<mode>_one_le): Likewise.
3046 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
3047 define the top 64-bit vector element.
3049 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
3052 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3053 Document vec_vbpermq builtin.
3056 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
3057 enable use of xxsldwi and xxpermdi builtin functions.
3058 (vec_xxpermdi): Likewise.
3060 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3061 Document use of vec_xxsldwi and vec_xxpermdi builtins.
3063 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
3065 PR rtl-optimization/60650
3066 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
3068 (find_spills_for): New.
3069 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
3070 Spill all pseudos on the second iteration.
3072 2014-03-27 Marek Polacek <polacek@redhat.com>
3075 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
3078 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3080 * config/s390/s390.c (s390_can_use_return_insn): Check for
3081 call-saved FPRs on 31 bit.
3083 2014-03-27 Jakub Jelinek <jakub@redhat.com>
3086 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
3087 if they need regimplification, just drop them instead of
3088 calling gimple_regimplify_operands on them.
3090 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
3093 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
3094 (aarch64_frame_pointer_required): Adjust logic.
3095 (aarch64_can_eliminate): Adjust logic.
3096 (aarch64_override_options_after_change): Adjust logic.
3098 2014-03-27 Dehao Chen <dehao@google.com>
3100 * ipa-inline.c (early_inliner): Update node's inline info.
3102 2014-03-26 Dehao Chen <dehao@google.com>
3104 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
3105 compiler inserted conditional jumps for NAN float check.
3107 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3109 * ubsan.h (ubsan_create_data): Change second argument's type
3110 to const location_t *.
3111 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
3113 (ubsan_create_data): Change second argument to const location_t *PLOC.
3114 Create Loc field whenever PLOC is non-NULL.
3115 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
3116 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
3120 * real.c (real_to_integer2): Change type of low to UHWI.
3122 2014-03-26 Tobias Burnus <burnus@net-b.de>
3124 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
3125 (CILK_SELF_SPECS): New define.
3126 (driver_self_specs): Use it.
3128 2014-03-26 Richard Biener <rguenther@suse.de>
3130 * tree-pretty-print.c (percent_K_format): Implement special
3131 case for LTO and its stripped down BLOCK tree.
3133 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3136 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
3138 * tree-vrp.c (simplify_internal_call_using_ranges): If only
3139 one range is range_int_cst_p, but not both, at least optimize
3140 addition/subtraction of 0 and multiplication by 0 or 1.
3141 * gimple-fold.c (gimple_fold_call): Fold
3142 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
3143 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
3144 INTEGER_CSTs, try to fold at least x * 0 and y - y.
3146 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
3148 PR rtl-optimization/60452
3149 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
3150 <case REG>: Return 1 for invalid offsets from the frame pointer.
3152 2014-03-26 Marek Polacek <polacek@redhat.com>
3155 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3158 2014-03-26 Marek Polacek <polacek@redhat.com>
3161 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3164 2014-03-26 Marek Polacek <polacek@redhat.com>
3167 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3168 multiplication in unsigned type.
3170 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3172 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3174 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3176 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3178 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3181 * cif-code.def (UNREACHABLE) New code.
3182 * ipa-inline.c (inline_small_functions): Skip edges to
3183 __builtlin_unreachable.
3184 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3185 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3186 predicate to __bulitin_unreachable.
3187 (set_cond_stmt_execution_predicate): Fix issue when
3188 invert_tree_comparison returns ERROR_MARK.
3189 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3190 propagate to inline clones.
3191 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3193 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3194 * cgraphclones.c (cgraph_clone_node): If call destination is already
3195 ureachable, do not redirect it back.
3196 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3199 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3201 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3202 Do not modify inline clones.
3204 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3206 * config/i386/i386.md (general_sext_operand): New mode attr.
3207 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3208 don't generate (sign_extend (const_int)).
3209 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3210 operands[2]. Use We constraint instead of <i> and
3211 <general_sext_operand> predicate instead of <general_operand>.
3212 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3213 * config/i386/constraints.md (We): New constraint.
3214 * config/i386/predicates.md (x86_64_sext_operand,
3215 sext_operand): New predicates.
3217 2014-03-25 Martin Jambor <mjambor@suse.cz>
3220 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3221 inconsistent devirtualizations to __builtin_unreachable.
3223 2014-03-25 Marek Polacek <polacek@redhat.com>
3226 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3228 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3230 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3231 order of elements for big-endian.
3233 2014-03-25 Richard Biener <rguenther@suse.de>
3236 * gimplify-me.c (gimple_regimplify_operands): Update the
3239 2014-03-25 Martin Jambor <mjambor@suse.cz>
3242 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3243 (lto_output_varpool_node): Likewise.
3244 (input_overwrite_node): Likewise.
3245 (input_varpool_node): Likewise.
3247 2014-03-25 Richard Biener <rguenther@suse.de>
3249 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3250 (run_gcc): Likewise.
3252 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3254 * combine.c (simplify_compare_const): Add MODE argument.
3255 Handle mode_width 0 as very large mode_width.
3256 (try_combine, simplify_comparison): Adjust callers.
3258 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3259 type to avoid signed integer overflow.
3260 * explow.c (plus_constant): Likewise.
3262 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3264 * doc/generic.texi: Correct typos.
3266 2014-03-24 Tobias Burnus <burnus@net-b.de>
3268 * doc/invoke.texi (-flto): Expand section about
3269 using static libraries with LTO.
3271 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3273 PR rtl-optimization/60501
3274 * optabs.def (addptr3_optab): New optab.
3275 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3276 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3277 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3279 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3281 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3283 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3285 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3288 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3289 (_mm256_undefined_ps): Define.
3290 (_mm256_undefined_pd): Define.
3291 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3292 (_mm_undefined_pd): Define.
3293 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3294 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3295 (_mm512_undefined_ps): Define.
3296 (_mm512_undefined_pd): Define.
3297 Use _mm*_undefined_*.
3298 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3300 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3302 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3303 (lshr_simd): DI mode added.
3304 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3305 (aarch64_ushr_simddi): Likewise.
3306 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3307 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3308 (vshrd_n_u64): Likewise.
3310 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3312 * Makefile.in (s-macro_list): Depend on cc1.
3314 2014-03-23 Teresa Johnson <tejohnson@google.com>
3316 * ipa-utils.c (ipa_print_order): Use specified dump file.
3318 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3320 PR rtl-optimization/60601
3321 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3323 * gcc.c (eval_spec_function): Initialize save_growing_value.
3325 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3328 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3329 code == MINUS_EXPR, never swap op0 with op1.
3331 * toplev.c (init_local_tick): Avoid signed integer multiplication
3333 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3334 shift by first operand's bitsize.
3336 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3339 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3341 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3342 TARGET_ISA_64BIT_P(x).
3344 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3346 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3347 pattern for vector nor instead of subtract from splat(-1).
3348 (altivec_expand_vec_perm_const_le): Likewise.
3350 2014-03-21 Richard Henderson <rth@twiddle.net>
3353 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3354 related insns after epilogue_completed.
3356 2014-03-21 Martin Jambor <mjambor@suse.cz>
3359 * cgraph.h (symtab_node): New flag body_removed.
3360 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3361 when removing bodies.
3362 * symtab.c (dump_symtab_base): Dump body_removed flag.
3363 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3364 had their bodies removed.
3366 2014-03-21 Martin Jambor <mjambor@suse.cz>
3369 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3372 2014-03-21 Richard Biener <rguenther@suse.de>
3374 PR tree-optimization/60577
3375 * tree-core.h (struct tree_base): Document nothrow_flag use
3376 in VAR_DECL_NONALIASED.
3377 * tree.h (VAR_DECL_NONALIASED): New.
3378 (may_be_aliased): Adjust.
3379 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3381 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3383 * expr.c (expand_expr_real_1): Remove outdated comment.
3385 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3388 * ira.c (adjust_cleared_regs): Call copy_rtx on
3389 *reg_equiv[REGNO (loc)].src_p before passing it to
3390 simplify_replace_fn_rtx.
3393 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3394 into CONST, put pic register as first operand of PLUS. Use
3395 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3397 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3399 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3401 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3403 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3404 around for store forwarding issue in the FPU on the UT699.
3405 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3406 loads and operations if -mfix-ut699 is specified.
3407 (divtf3_hq): Tweak attribute.
3408 (sqrttf2_hq): Likewise.
3410 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3412 * calls.c (store_one_arg): Remove incorrect const qualification on the
3413 type of the temporary.
3414 * cfgexpand.c (expand_return): Likewise.
3415 * expr.c (expand_constructor): Likewise.
3416 (expand_expr_real_1): Likewise.
3418 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3420 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3423 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3426 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3428 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3430 * config/arm/aarch-common-protos.h
3431 (alu_cost_table): Fix spelling of "extend".
3432 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3434 2014-03-19 Richard Biener <rguenther@suse.de>
3437 * tree-core.h (tree_type_common): Re-order pointer members
3438 to reduce recursion depth during GC walks.
3440 2014-03-19 Marek Polacek <polacek@redhat.com>
3443 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3444 before accessing it.
3446 2014-03-19 Richard Biener <rguenther@suse.de>
3449 * lto-streamer-in.c (input_function): In WPA stage do not drop
3452 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3454 PR tree-optimization/60559
3455 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3456 with build_zero_cst assignment.
3458 2014-03-18 Kai Tietz <ktietz@redhat.com>
3460 PR rtl-optimization/56356
3461 * sdbout.c (sdbout_parms): Verify that parms'
3462 incoming argument is valid.
3463 (sdbout_reg_parms): Likewise.
3465 2014-03-18 Richard Henderson <rth@redhat.com>
3468 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3469 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3470 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3472 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3474 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3475 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3476 Italicize plugin event names in description. Explain that
3477 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3478 Remind that no GCC functions should be called after PLUGIN_FINISH.
3479 Explain what pragmas with expansion are.
3481 2014-03-18 Martin Liska <mliska@suse.cz>
3483 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3484 gimple call statement is update.
3485 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3486 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3488 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3491 * ubsan.c (ubsan_instrument_unreachable): Call
3492 initialize_sanitizer_builtins.
3493 (ubsan_pass): Likewise.
3496 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3497 varpool_finalize_decl instead of rest_of_decl_compilation.
3499 2014-03-18 Richard Biener <rguenther@suse.de>
3501 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3502 by using bitmap_and_compl instead of bitmap_and_compl_into.
3503 (df_rd_transfer_function): Likewise.
3505 2014-03-18 Richard Biener <rguenther@suse.de>
3507 * doc/lto.texi (fresolution): Fix typo.
3509 2014-03-18 Richard Biener <rguenther@suse.de>
3511 * doc/invoke.texi (flto): Update for changes in 4.9.
3513 2014-03-18 Richard Biener <rguenther@suse.de>
3515 * doc/loop.texi: Remove section on the removed lambda framework.
3516 Update loop docs with recent changes in preserving loop structure.
3518 2014-03-18 Richard Biener <rguenther@suse.de>
3520 * doc/lto.texi (-fresolution): Document.
3522 2014-03-18 Richard Biener <rguenther@suse.de>
3524 * doc/contrib.texi: Adjust my name.
3526 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3529 * internal-fn.c: Include diagnostic-core.h.
3530 (expand_BUILTIN_EXPECT): New function.
3531 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3532 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3533 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3534 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3536 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3537 Revert 3 argument __builtin_expect code.
3538 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3539 * gimple-fold.c (gimple_fold_call): Likewise.
3540 * tree.h (fold_builtin_expect): New prototype.
3541 * builtins.c (build_builtin_expect_predicate): Add predictor
3542 argument, if non-NULL, create 3 argument __builtin_expect.
3543 (fold_builtin_expect): No longer static. Add ARG2 argument,
3544 pass it through to build_builtin_expect_predicate.
3545 (fold_builtin_2): Adjust caller.
3546 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3547 * internal-fn.def (BUILTIN_EXPECT): New.
3549 2014-03-18 Tobias Burnus <burnus@net-b.de>
3552 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3553 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3554 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3556 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3559 * predict.c (combine_predictions_for_bb): Fix up formatting.
3560 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3561 fill what it points to if non-NULL.
3562 (tree_predict_by_opcode): Adjust caller, use the predictor.
3563 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3565 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3567 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3568 proper constant for the store mode.
3570 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3572 * symtab.c (change_decl_assembler_name): Fix transparent alias
3575 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3577 * config/aarch64/aarch64.c: Correct the comments about the
3578 aarch64 stack layout.
3580 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3582 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3583 check for GF_OMP_FOR_KIND_FOR.
3585 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3587 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3588 ymm and zmm register names.
3590 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3593 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3594 note creation for the 2010-08-31 changes.
3596 2014-03-17 Marek Polacek <polacek@redhat.com>
3599 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3600 as -fno-tree-loop-vectorize.
3601 (expand_omp_simd): Likewise.
3603 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3605 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3606 (eligible_for_call_delay): New prototype.
3607 * config/sparc/sparc.c (tls_call_delay): Rename into...
3608 (eligible_for_call_delay): ...this. Return false if the instruction
3609 cannot be put in the delay slot of a branch.
3610 (eligible_for_restore_insn): Simplify.
3611 (eligible_for_return_delay): Return false if the instruction cannot be
3612 put in the delay slot of a branch and simplify.
3613 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3614 put in the delay slot of a branch.
3615 * config/sparc/sparc.md (fix_ut699): New attribute.
3616 (tls_call_delay): Delete.
3617 (in_call_delay): Reimplement.
3618 (eligible_for_sibcall_delay): Rename into...
3619 (in_sibcall_delay): ...this.
3620 (eligible_for_return_delay): Rename into...
3621 (in_return_delay): ...this.
3622 (in_branch_delay): Reimplement.
3623 (in_uncond_branch_delay): Delete.
3624 (in_annul_branch_delay): Delete.
3626 2014-03-14 Richard Henderson <rth@redhat.com>
3629 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3630 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3631 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3632 (floathi splitters): Remove.
3633 (float<SWI48x>xf2): New pattern.
3634 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3635 code that tried to handle DImode for 32-bit, but which was excluded
3636 by the pattern's condition. Drop allocation of stack temporary.
3637 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3638 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3639 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3640 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3641 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3642 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3643 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3644 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3645 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3646 (*float<SWI48x><X87MODEF>2_i387): Remove.
3647 (all float _with_temp splitters): Remove.
3648 (*float<SWI48x><MODEF>2_i387): New pattern.
3649 (*float<SWI48><MODEF>2_sse): New pattern.
3650 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3651 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3653 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3654 Marek Polacek <polacek@redhat.com>
3657 * common.opt (dump_base_name_prefixed): New Variable.
3658 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3659 if x_dump_base_name_prefixed is already set, set it at the end.
3661 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3663 PR rtl-optimization/60508
3664 * lra-constraints.c (get_reload_reg): Add new parameter
3666 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3667 Pass the new parameter values.
3669 2014-03-14 Richard Biener <rguenther@suse.de>
3671 * common.opt: Revert unintented changes from r205065.
3674 2014-03-14 Richard Biener <rguenther@suse.de>
3677 * cfghooks.c (split_block): Properly adjust all loops the
3678 block was a latch of.
3680 2014-03-14 Martin Jambor <mjambor@suse.cz>
3683 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3686 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3689 * config/avr/avr.c (avr_set_current_function): Pass function name
3690 through default_strip_name_encoding before sanity checking instead
3691 of skipping the first char of the assembler name.
3693 2014-03-13 Richard Henderson <rth@redhat.com>
3696 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3697 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3698 * config/i386/i386-protos.h: Likewise.
3699 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3700 in the expander instead of a splitter.
3701 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3702 any possibility of requiring a memory.
3703 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3704 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3705 (fp branch splitters): Update for ix86_split_fp_branch.
3706 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3707 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3708 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3709 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3710 (*fop_<MODEF>_3_i387): Likewise.
3711 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3712 (splitters for the fop_* register patterns): Remove.
3713 (fscalexf4_i387): Rename from *fscalexf4_i387.
3714 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3716 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3718 PR tree-optimization/59779
3719 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3720 type for bitsize and maxsize instead of HOST_WIDE_INT.
3722 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3724 PR rtl-optimization/57320
3725 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3726 the CFG after thread_prologue_and_epilogue_insns.
3728 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3730 PR rtl-optimization/57189
3731 * lra-constraints.c (process_alt_operands): Disfavor spilling
3734 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3736 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3738 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3740 PR tree-optimization/59025
3742 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3743 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3745 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3748 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3749 calls of avr_out_plus_1.
3751 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3753 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3754 BB's single pred and update the father loop's latch info later.
3756 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3758 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3762 (VEC_base): Likewise.
3763 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3764 registers, we need to swap double words in little endian mode.
3766 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3767 to be a container mode for 128-bit integer operations added in ISA
3768 2.07. Unlike TImode and PTImode, the preferred register set is
3769 the Altivec/VMX registers for the 128-bit operations.
3771 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3773 (rs6000_split_128bit_ok_p): Likewise.
3775 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3776 macros for creating ISA 2.07 normal and overloaded builtin
3777 functions with 3 arguments.
3778 (BU_P8V_OVERLOAD_3): Likewise.
3779 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3780 for use as overloaded functions.
3781 (VPERM_1TI_UNS): Likewise.
3782 (VSEL_1TI): Likewise.
3783 (VSEL_1TI_UNS): Likewise.
3784 (ST_INTERNAL_1ti): Likewise.
3785 (LD_INTERNAL_1ti): Likewise.
3786 (XXSEL_1TI): Likewise.
3787 (XXSEL_1TI_UNS): Likewise.
3788 (VPERM_1TI): Likewise.
3789 (VPERM_1TI_UNS): Likewise.
3790 (XXPERMDI_1TI): Likewise.
3791 (SET_1TI): Likewise.
3792 (LXVD2X_V1TI): Likewise.
3793 (STXVD2X_V1TI): Likewise.
3794 (VEC_INIT_V1TI): Likewise.
3795 (VEC_SET_V1TI): Likewise.
3796 (VEC_EXT_V1TI): Likewise.
3797 (EQV_V1TI): Likewise.
3798 (NAND_V1TI): Likewise.
3799 (ORC_V1TI): Likewise.
3800 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3801 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3803 (VADDUQM): Likewise.
3804 (VSUBCUQ): Likewise.
3805 (VADDEUQM): Likewise.
3806 (VADDECUQ): Likewise.
3807 (VSUBEUQM): Likewise.
3808 (VSUBECUQ): Likewise.
3810 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3811 __int128_t and __uint128_t types.
3812 (__uint128_type): Likewise.
3813 (altivec_categorize_keyword): Add support for vector __int128_t,
3814 vector __uint128_t, vector __int128, and vector unsigned __int128
3815 as a container type for TImode operations that need to be done in
3816 VSX/Altivec registers.
3817 (rs6000_macro_to_expand): Likewise.
3818 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3819 to support 128-bit integer instructions vaddcuq, vadduqm,
3820 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3821 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
3823 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
3824 for V1TImode, and set up preferences to use VSX/Altivec registers.
3825 Setup VSX reload handlers.
3826 (rs6000_debug_reg_global): Likewise.
3827 (rs6000_init_hard_regno_mode_ok): Likewise.
3828 (rs6000_preferred_simd_mode): Likewise.
3829 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
3830 (easy_altivec_constant): Likewise.
3831 (output_vec_const_move): Likewise.
3832 (rs6000_expand_vector_set): Convert V1TImode set and extract to
3834 (rs6000_expand_vector_extract): Likewise.
3835 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
3837 (rs6000_const_vec): Add support for V1TImode.
3838 (rs6000_emit_le_vsx_load): Swap double words when loading or
3839 storing TImode/V1TImode.
3840 (rs6000_emit_le_vsx_store): Likewise.
3841 (rs6000_emit_le_vsx_move): Likewise.
3842 (rs6000_emit_move): Add support for V1TImode.
3843 (altivec_expand_ld_builtin): Likewise.
3844 (altivec_expand_st_builtin): Likewise.
3845 (altivec_expand_vec_init_builtin): Likewise.
3846 (altivec_expand_builtin): Likewise.
3847 (rs6000_init_builtins): Add support for V1TImode type. Add
3848 support for ISA 2.07 128-bit integer builtins. Define type names
3849 for the VSX/Altivec vector types.
3850 (altivec_init_builtins): Add support for overloaded vector
3851 functions with V1TImode type.
3852 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
3853 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
3855 (rs6000_split_128bit_ok_p): Likewise.
3856 (rs6000_handle_altivec_attribute): Create V1TImode from vector
3857 __int128_t and vector __uint128_t.
3859 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
3860 and mode attributes.
3867 (VS_scalar): Likewise.
3868 (VS_double): Likewise.
3869 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
3871 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
3872 we support the ISA 2.07 128-bit integer arithmetic instructions.
3873 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
3874 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
3875 and TImode types for use with the builtin functions.
3876 (V1TI_type_node): Likewise.
3877 (unsigned_V1TI_type_node): Likewise.
3878 (intTI_type_internal_node): Likewise.
3879 (uintTI_type_internal_node): Likewise.
3881 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
3882 128-bit builtin functions.
3883 (UNSPEC_VADDEUQM): Likewise.
3884 (UNSPEC_VADDECUQ): Likewise.
3885 (UNSPEC_VSUBCUQ): Likewise.
3886 (UNSPEC_VSUBEUQM): Likewise.
3887 (UNSPEC_VSUBECUQ): Likewise.
3888 (VM): Add V1TImode to vector mode iterators.
3890 (VI_unit): Likewise.
3891 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
3892 (altivec_vaddcuq): Likewise.
3893 (altivec_vsubuqm): Likewise.
3894 (altivec_vsubcuq): Likewise.
3895 (altivec_vaddeuqm): Likewise.
3896 (altivec_vaddecuq): Likewise.
3897 (altivec_vsubeuqm): Likewise.
3898 (altivec_vsubecuq): Likewise.
3900 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
3902 (BOOL_128): Likewise.
3903 (BOOL_REGS_OUTPUT): Likewise.
3904 (BOOL_REGS_OP1): Likewise.
3905 (BOOL_REGS_OP2): Likewise.
3906 (BOOL_REGS_UNARY): Likewise.
3907 (BOOL_REGS_AND_CR0): Likewise.
3909 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
3910 128-bit integer builtin support.
3911 (vec_vadduqm): Likewise.
3912 (vec_vaddecuq): Likewise.
3913 (vec_vaddeuqm): Likewise.
3914 (vec_vsubecuq): Likewise.
3915 (vec_vsubeuqm): Likewise.
3916 (vec_vsubcuq): Likewise.
3917 (vec_vsubuqm): Likewise.
3919 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3920 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
3921 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
3922 128-bit integer add/subtract to ISA 2.07.
3924 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
3926 * config/arc/arc.c (arc_predicate_delay_insns):
3927 Fix third argument passed to conditionalize_nonjump.
3929 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
3931 * config/aarch64/aarch64-builtins.c
3932 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
3933 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
3934 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
3935 instead of __builtin_lfloor.
3936 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
3938 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3940 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
3941 (tree_ssa_ifcombine_bb_1): New function.
3942 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
3943 is an empty forwarder block to then_bb or vice versa and then_bb
3944 and else_bb are effectively swapped.
3946 2014-03-12 Christian Bruel <christian.bruel@st.com>
3949 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
3950 REG_CFA_DEF_CFA note.
3951 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
3952 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
3954 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
3956 PR tree-optimization/60454
3957 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
3959 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3961 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
3962 Do not define target_cpu_default2 to generic.
3963 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
3964 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
3965 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
3967 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3968 Marc Glisse <marc.glisse@inria.fr>
3970 PR tree-optimization/60502
3971 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
3972 instead of build_low_bits_mask.
3974 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3977 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
3978 if there are multiple uses, but op doesn't live on E edge.
3979 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
3980 clobber stmts before __builtin_unreachable.
3982 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
3984 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
3985 hard_frame_pointer_rtx.
3986 * cse.c (cse_insn): Remove volatile check.
3987 * cselib.c (cselib_process_insn): Likewise.
3988 * dse.c (scan_insn): Likewise.
3990 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3992 * config/arc/arc.c (conditionalize_nonjump): New function,
3994 (arc_ifcvt): ... this.
3995 (arc_predicate_delay_insns): Use it.
3997 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3999 * config/arc/predicates.md (extend_operand): During/after reload,
4000 allow const_int_operand.
4001 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
4002 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
4003 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
4005 (umulsi3_highpart_i): Likewise.
4007 2014-03-11 Richard Biener <rguenther@suse.de>
4009 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
4010 Add asserts to guard possible wrong-code bugs.
4012 2014-03-11 Richard Biener <rguenther@suse.de>
4014 PR tree-optimization/60429
4015 PR tree-optimization/60485
4016 * tree-ssa-structalias.c (set_union_with_increment): Properly
4017 take into account all fields that overlap the shifted vars.
4018 (do_sd_constraint): Likewise.
4019 (do_ds_constraint): Likewise.
4020 (get_constraint_for_ptr_offset): Likewise.
4022 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
4024 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
4025 (nios2_compute_frame_layout):
4026 Add calculation of cfun->machine->fp_save_offset.
4027 (nios2_expand_prologue): Correct setting of frame pointer register
4029 (nios2_expand_epilogue): Update recovery of stack pointer from
4030 frame pointer accordingly.
4031 (nios2_initial_elimination_offset): Update calculation of offset
4032 for eliminating to HARD_FRAME_POINTER_REGNUM.
4034 2014-03-10 Jakub Jelinek <jakub@redhat.com>
4037 * ipa.c (symtab_remove_unreachable_nodes): Don't call
4038 cgraph_get_create_node on VAR_DECLs.
4040 2014-03-10 Richard Biener <rguenther@suse.de>
4043 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
4045 2014-03-08 Douglas B Rupp <rupp@gnat.com>
4047 * config/vms/vms.opt (vms_float_format): New variable.
4049 2014-03-08 Tobias Burnus <burnus@net-b.de>
4051 * doc/invoke.texi (-fcilkplus): Update implementation status.
4053 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
4054 Richard Biener <rguenther@suse.de>
4056 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
4057 consistently accross all TUs.
4058 (run_gcc): Enable -fshort-double automatically at link at link-time
4059 and disallow override.
4061 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
4064 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
4065 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
4066 if they can't be used.
4068 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4070 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
4071 for Solaris 11/x86 ld.
4072 * configure: Regenerate.
4074 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4076 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
4077 (LIB_TLS_SPEC): Save as ld_tls_libs.
4078 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
4079 (HAVE_AS_IX86_TLSLDM): New test.
4080 * configure, config.in: Regenerate.
4081 * config/i386/i386.c (legitimize_tls_address): Fall back to
4082 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
4083 cannot support TLS_MODEL_LOCAL_DYNAMIC.
4084 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
4085 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
4087 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
4089 * common.opt (fira-loop-pressure): Mark as optimization.
4091 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
4093 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
4094 an OpenMP mappable type.
4096 2014-03-06 Matthias Klose <doko@ubuntu.com>
4098 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
4099 MULTILIB_OSDIRNAMES is not defined.
4101 2014-03-06 Jakub Jelinek <jakub@redhat.com>
4102 Meador Inge <meadori@codesourcery.com>
4105 * config/arm/arm.c (arm_tls_symbol_p): Remove.
4106 (arm_legitimize_address): Call legitimize_tls_address for any
4107 arm_tls_referenced_p expression, handle constant addend. Call it
4108 before testing for !TARGET_ARM.
4109 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
4111 2014-03-06 Richard Biener <rguenther@suse.de>
4117 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4119 * tree-streamer.c (record_common_node): Assert we don't record
4120 nodes with type double.
4121 (preload_common_node): Skip type double, complex double and double
4122 pointer since it is now frontend dependent due to fshort-double option.
4124 2014-03-06 Richard Biener <rguenther@suse.de>
4126 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
4127 or -fno-lto is specified and the linker has full plugin support.
4128 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
4129 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
4130 * lto-wrapper.c (merge_and_complain): Merge compile-time
4131 optimization levels.
4132 (run_gcc): And pass it through to the link options.
4134 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
4138 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4140 * cselib.c (remove_useless_values): Skip to avoid quadratic
4141 behavior if the condition moved from...
4142 (cselib_process_insn): ... here holds.
4144 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4147 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
4148 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
4151 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
4152 (TM_H): Add x86-tune.def.
4154 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4156 * config/aarch64/aarch64.c (generic_tunings):
4157 Use cortexa57_extra_costs.
4159 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4162 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4163 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4164 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4167 2014-03-04 Heiher <r@hev.cc>
4169 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4170 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4172 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4174 * config/i386/predicates.md (const2356_operand): Change to ...
4175 (const2367_operand): ... this.
4176 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4178 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4179 (*avx512pf_scatterpf<mode>sf): Ditto.
4180 (avx512pf_scatterpf<mode>df): Ditto.
4181 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4182 (*avx512pf_scatterpf<mode>df): Ditto.
4183 * config/i386/i386.c (ix86_expand_builtin): Update
4184 incorrect hint operand error message.
4186 2014-03-04 Richard Biener <rguenther@suse.de>
4188 * lto-section-in.c (lto_get_section_data): Fix const cast.
4190 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4192 * tree-streamer.c (record_common_node): Assert we don't record
4193 nodes with type double.
4194 (preload_common_node): Skip type double, complex double and double
4195 pointer since it is now frontend dependent due to fshort-double option.
4197 2014-03-04 Richard Biener <rguenther@suse.de>
4200 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4201 (lto_input_toplevel_asms): Likewise.
4202 * lto-section-in.c (lto_get_section_data): Instead do it here
4205 2014-03-04 Richard Biener <rguenther@suse.de>
4207 PR tree-optimization/60382
4208 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4209 dead PHIs a reduction.
4211 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4213 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4215 (_mm_prefetch): Move out of GCC target("sse") pragma.
4216 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4217 GCC target("prfchw") pragma.
4218 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4220 * config/i386/i386.c (ix86_option_override_internal): Enable
4221 -mprfchw with -mprefetchwt1.
4223 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4225 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4228 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4230 * opts.h (CL_PCH_IGNORE): Define.
4231 * targhooks.c (option_affects_pch_p):
4232 Return false for options that have CL_PCH_IGNORE set.
4233 * opt-functions.awk: Process PchIgnore.
4234 * doc/options.texi: Document PchIgnore.
4236 * config/arc/arc.opt (misize): Add PchIgnore property.
4238 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4240 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4241 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4242 constraint on constants to permit them being loaded into
4243 GENERAL_REGS or BASE_REGS.
4245 2014-03-03 Nick Clifton <nickc@redhat.com>
4247 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4248 anti-cacnonical alternatives.
4249 (negandhi3_real): New pattern.
4250 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4252 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4254 * config/avr/avr-mcus.def: Remove atxmega16x1.
4255 * config/avr/avr-tables.opt: Regenerate.
4256 * config/avr/t-multilib: Regenerate.
4257 * doc/avr-mmcu.texi: Regenerate.
4259 2014-03-03 Tobias Grosser <tobias@grosser.es>
4260 Mircea Namolaru <mircea.namolaru@inria.fr>
4262 PR tree-optimization/58028
4263 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4266 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4268 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4269 not handled by recognizers.
4271 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4274 * function.c (expand_function_end): Don't emit
4275 clobber_return_register sequence if clobber_after is a BARRIER.
4276 * cfgexpand.c (construct_exit_block): Append instructions before
4277 return_label to prev_bb.
4279 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4281 * config/rs6000/constraints.md: Document reserved use of "wc".
4283 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4286 * ipa.c (function_and_variable_visibility): When dissolving comdat
4287 group, also set all symbols to local.
4289 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4294 2013-12-14 Jan Hubicka <jh@suse.cz>
4296 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4298 2014-03-02 Jon Beniston <jon@beniston.com>
4304 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4305 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4306 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4307 (simple_return, *simple_return): New patterns
4308 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4309 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4311 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4313 * dwarf2out.c (gen_subprogram_die): Tidy.
4315 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4318 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4319 (*mov_t_msb_neg_negc): ... this new insn.
4321 2014-02-28 Jason Merrill <jason@redhat.com>
4324 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4327 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4330 * dwarf2out.c (decltype_auto_die): New static.
4331 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4332 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4333 (is_cxx_auto): Likewise.
4335 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4337 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4338 we are not using general regs only.
4340 2014-02-28 Richard Biener <rguenther@suse.de>
4343 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4344 previous fix and only allow to remove trivial pre-headers
4345 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4346 (remove_forwarder_block): Properly update the latch of a loop.
4348 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4351 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4352 (cselib_preserved_hash_table): New.
4353 (preserve_constants_and_equivs): Move preserved vals to it.
4354 (cselib_find_slot): Look it up first.
4355 (cselib_init): Initialize it.
4356 (cselib_finish): Release it.
4357 (dump_cselib_table): Dump it.
4359 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4362 * cselib.c (remove_useless_values): Skip to avoid quadratic
4363 behavior if the condition moved from...
4364 (cselib_process_insn): ... here holds.
4366 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4369 * var-tracking.c (vt_initialize): Apply the same condition to
4370 preserve the CFA base value.
4372 2014-02-28 Joey Ye <joey.ye@arm.com>
4375 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4376 if reload in progress or completed.
4378 2014-02-28 Tobias Burnus <burnus@net-b.de>
4381 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4384 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4386 * doc/tm.texi.in (Condition Code Status): Update documention for
4387 relative locations of cc0-setter and cc0-user.
4389 2014-02-27 Jeff Law <law@redhat.com>
4391 PR rtl-optimization/52714
4392 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4393 into two independent simple sets, if I3 is a jump, ensure the
4394 pattern we place into I3 is a (set (pc) ...).
4396 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4397 Jeff Law <law@redhat.com>
4399 PR rtl-optimization/49847
4400 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4401 are in different blocks.
4402 * doc/tm.texi (Condition Code Status): Update documention for
4403 relative locations of cc0-setter and cc0-user.
4405 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4408 * lra.c (lra_emit_add): Check SUBREG too.
4410 2014-02-27 Andreas Schwab <schwab@suse.de>
4412 * config/m68k/m68k.c (m68k_option_override): Disable
4413 -flive-range-shrinkage for classic m68k.
4414 (m68k_override_options_after_change): Likewise.
4416 2014-02-27 Marek Polacek <polacek@redhat.com>
4419 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4420 -Wmaybe-uninitialized.
4422 2014-02-27 Alan Modra <amodra@gmail.com>
4425 * reload1.c (emit_input_reload_insns): When reload_override_in,
4426 set old to rl->in_reg when rl->in_reg is a subreg.
4428 2014-02-26 Richard Biener <rguenther@suse.de>
4431 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4433 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4435 * common/config/i386/predicates.md (const1256_operand): Remove.
4436 (const2356_operand): New.
4437 (const_1_to_2_operand): Remove.
4438 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4439 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4440 (*avx512pf_gatherpf<mode>sf): Ditto.
4441 (avx512pf_gatherpf<mode>df): Ditto.
4442 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4443 (*avx512pf_gatherpf<mode>df): Ditto.
4444 (avx512pf_scatterpf<mode>sf): Ditto.
4445 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4446 (*avx512pf_scatterpf<mode>sf): Ditto.
4447 (avx512pf_scatterpf<mode>df): Ditto.
4448 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4449 (*avx512pf_scatterpf<mode>df): Ditto.
4450 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4452 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4454 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4455 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4456 (_mm512_mask_testn_epi64_mask): Move to ...
4457 * config/i386/avx512cdintrin.h: Here.
4458 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4459 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4460 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4461 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4462 TARGET_AVX512F from TARGET_AVX512CD.
4464 2014-02-26 Richard Biener <rguenther@suse.de>
4467 * ipa.c (walk_polymorphic_call_targets): Properly guard
4468 call to inline_update_overall_summary.
4470 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4473 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4474 and latches only if requested. Fix latch if it is removed.
4475 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4476 LOOPS_HAVE_PREHEADERS.
4478 2014-02-25 Andrew Pinski <apinski@cavium.com>
4480 * builtins.c (expand_builtin_thread_pointer): Create a new target
4481 when the target is NULL.
4483 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4485 PR rtl-optimization/60317
4486 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4487 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4488 * lra-assigns.c: Include params.h.
4489 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4490 other reload pseudos considerations.
4492 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4494 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4495 to use canonical form for nor<mode>3.
4497 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4500 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4503 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4505 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4506 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4507 (ix86_handle_option): Handle OPT_mprefetchwt1.
4508 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4509 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4511 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4512 OPTION_MASK_ISA_PREFETCHWT1.
4513 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4514 (PTA_PREFETCHWT1): New.
4515 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4516 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4517 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4518 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4519 (*prefetch_avx512pf_<mode>_: Change into ...
4520 (*prefetch_prefetchwt1_<mode>: This.
4521 * config/i386/i386.opt (mprefetchwt1): New.
4522 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4523 (_mm_prefetch): Handle intent to write.
4524 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4526 2014-02-25 Richard Biener <rguenther@suse.de>
4529 * emit-rtl.c (mem_attrs_htab): Remove.
4530 (mem_attrs_htab_hash): Likewise.
4531 (mem_attrs_htab_eq): Likewise.
4532 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4533 (init_emit_once): Do not allocate mem_attrs_htab.
4535 2014-02-25 Richard Biener <rguenther@suse.de>
4538 * lto-opts.c (lto_write_options): Output non-explicit conservative
4539 -fwrapv, -fno-trapv and -fno-strict-overflow.
4540 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4541 (run_gcc): And pass them through.
4543 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4545 * sel-sched.c (calculate_new_fences): New parameter ptime.
4546 Calculate it as a maximum over all fence cycles.
4547 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4548 Print the final schedule timing when sched_verbose.
4550 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4552 PR rtl-optimization/60292
4553 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4554 bit fot the fence instruction.
4556 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4558 * calls.h: Fix typo in comment.
4560 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4562 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4563 adjusting offsetable addresses.
4565 2014-02-24 Guozhi Wei <carrot@google.com>
4567 * sparseset.h (sparseset_pop): Fix the wrong index.
4569 2014-02-24 Walter Lee <walt@tilera.com>
4571 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4572 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4574 * common/config/tilegx/tilegx-common.c
4575 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4576 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4578 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4579 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4580 (tilegx_gimplify_va_arg_expr): Handle big endian.
4581 (tilegx_expand_unaligned_load): Ditto.
4582 (tilegx_expand_unaligned_store): Ditto.
4583 (TARGET_RETURN_IN_MSB): New.
4584 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4585 (TARGET_ENDIAN_DEFAULT): New.
4586 (TARGET_BIG_ENDIAN): Handle big endian.
4587 (BYTES_BIG_ENDIAN): Ditto.
4588 (WORDS_BIG_ENDIAN): Ditto.
4589 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4592 * config/tilegx/tilegx.md (extv): Handle big endian.
4594 (insn_st<n>): Ditto.
4595 (insn_st<n>_add<bitsuffix>): Ditto.
4596 (insn_stnt<n>): Ditto.
4597 (insn_stnt<n>_add<bitsuffix>):Ditto.
4598 (vec_interleave_highv8qi): Handle big endian.
4599 (vec_interleave_highv8qi_be): New.
4600 (vec_interleave_highv8qi_le): New.
4601 (insn_v1int_h): Handle big endian.
4602 (vec_interleave_lowv8qi): Handle big endian.
4603 (vec_interleave_lowv8qi_be): New.
4604 (vec_interleave_lowv8qi_le): New.
4605 (insn_v1int_l): Handle big endian.
4606 (vec_interleave_highv4hi): Handle big endian.
4607 (vec_interleave_highv4hi_be): New.
4608 (vec_interleave_highv4hi_le): New.
4609 (insn_v2int_h): Handle big endian.
4610 (vec_interleave_lowv4hi): Handle big endian.
4611 (vec_interleave_lowv4hi_be): New.
4612 (vec_interleave_lowv4hi_le): New.
4613 (insn_v2int_l): Handle big endian.
4614 (vec_interleave_highv2si): Handle big endian.
4615 (vec_interleave_highv2si_be): New.
4616 (vec_interleave_highv2si_le): New.
4617 (insn_v4int_h): Handle big endian.
4618 (vec_interleave_lowv2si): Handle big endian.
4619 (vec_interleave_lowv2si_be): New.
4620 (vec_interleave_lowv2si_le): New.
4621 (insn_v4int_l): Handle big endian.
4622 * config/tilegx/tilegx.opt (mbig-endian): New option.
4623 (mlittle-endian): New option.
4624 * doc/install.texi: Document tilegxbe-linux.
4625 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4627 2014-02-24 Martin Jambor <mjambor@suse.cz>
4630 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4631 there are no parameter descriptors.
4633 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4635 PR rtl-optimization/60268
4636 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4637 initialization to ...
4638 (sched_rgn_init): ... here.
4639 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4641 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4643 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4646 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4648 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4651 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4653 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4654 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4656 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4658 * config/microblaze/predicates.md: Add cmp_op predicate.
4659 * config/microblaze/microblaze.md: Add branch_compare instruction
4660 which uses cmp_op predicate and emits cmp insn before branch.
4661 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4662 to microblaze_expand_conditional_branch and consolidate logic.
4663 (microblaze_expand_conditional_branch): emit branch_compare
4664 insn instead of handling cmp op separate from branch insn.
4666 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4668 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4671 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4673 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4674 define_insn with define_expand and new define_insn
4675 *altivec_lve<VI_char>x_internal.
4676 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4677 and new define_insn *altivec_stve<VI_char>x_internal.
4678 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4680 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4682 (altivec_expand_stvex_be): New function.
4684 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4686 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4687 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4688 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4689 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4691 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4694 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4695 instead of emit_move_insn.
4697 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4699 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4701 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4702 gen_altivec_vsumsws.
4704 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4706 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4707 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4708 (altivec_lvxl_<mode>): New define_expand incorporating
4709 -maltivec=be semantics where needed.
4710 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4711 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4712 semantics where needed.
4713 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4714 (altivec_stvx_<mode>): New define_expand incorporating
4715 -maltivec=be semantics where needed.
4716 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4717 VM2 iterator instead of V4SI.
4718 (altivec_stvxl_<mode>): New define_expand incorporating
4719 -maltivec=be semantics where needed.
4720 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4721 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4722 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4723 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4724 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4725 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4726 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4727 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4728 ALTIVEC_BUILTIN_STVXL.
4729 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4730 (altivec_expand_stvx_be): Likewise.
4731 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4732 (altivec_expand_lvx_be): Likewise.
4733 (altivec_expand_stvx_be): Likewise.
4734 (altivec_expand_builtin): Add cases for
4735 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4736 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4737 (altivec_init_builtins): Add definitions for
4738 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4739 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4741 2014-02-21 Catherine Moore <clm@codesourcery.com>
4743 * doc/invoke.texi (mvirt, mno-virt): Document.
4744 * config/mips/mips.opt (mvirt): New option.
4745 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4747 2014-02-21 Richard Biener <rguenther@suse.de>
4749 PR tree-optimization/60276
4750 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4751 (STMT_VINFO_MIN_NEG_DIST): New macro.
4752 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4753 STMT_VINFO_MIN_NEG_DIST.
4754 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4755 made for negative dependence distances still hold.
4757 2014-02-21 Richard Biener <rguenther@suse.de>
4760 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4761 DECL_INITIAL for globals not in the current function context.
4763 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4765 PR tree-optimization/56490
4766 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4767 * tree-ssa-uninit.c: Include params.h.
4768 (compute_control_dep_chain): Add num_calls argument, return false
4769 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4770 num_calls to recursive call.
4771 (find_predicates): Change dep_chain into normal array,
4772 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4773 variable and adjust compute_control_dep_chain caller.
4774 (find_def_preds): Likewise.
4776 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4778 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4779 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4781 2014-02-21 Nick Clifton <nickc@redhat.com>
4783 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4784 (pushhi1): Likewise.
4785 (popqi1): Add mode to pre_dec.
4788 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4790 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4791 mode for mask of V8SFmode permutation.
4793 2014-02-20 Richard Henderson <rth@redhat.com>
4796 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4797 a new pseudo for OLDVAL.
4799 2014-02-20 Jakub Jelinek <jakub@redhat.com>
4802 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4803 gen_reg_rtx if d->testing_p.
4804 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4805 if d->testing_p and we will certainly return true.
4806 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4809 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
4811 * emit-rtl.c (gen_reg_rtx): Assert that
4812 crtl->emit.regno_pointer_align_length is non-zero.
4814 2014-02-20 Richard Henderson <rth@redhat.com>
4817 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4818 on failure the store back into EXPECT.
4820 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4821 Sandra Loosemore <sandra@codesourcery.com>
4823 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
4824 * config/nios2/nios2.c (nios2_function_profiler): Add
4825 -fPIC (flag_pic == 2) support.
4826 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
4827 (nios2_large_offset_p): New function.
4828 (nios2_unspec_reloc_p): Move up position, update to use
4829 nios2_large_offset_p.
4830 (nios2_unspec_address): Remove function.
4831 (nios2_unspec_offset): New function.
4832 (nios2_large_got_address): New function.
4833 (nios2_got_address): Add large offset support.
4834 (nios2_legitimize_tls_address): Update usage of removed and new
4836 (nios2_symbol_binds_local_p): New function.
4837 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
4838 (nios2_legitimize_address): Update to use nios2_large_offset_p.
4839 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
4840 (nios2_print_operand): Merge H/L processing, add hiadj/lo
4841 processing for (const (unspec ...)).
4842 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
4844 2014-02-20 Richard Biener <rguenther@suse.de>
4846 * tree-cfg.c (replace_uses_by): Mark altered BBs before
4847 doing the substitution.
4848 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
4850 2014-02-20 Martin Jambor <mjambor@suse.cz>
4853 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
4854 info when checking whether lattices are bottom.
4856 2014-02-20 Richard Biener <rguenther@suse.de>
4859 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
4862 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
4865 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
4866 parameter specifying the scaling.
4867 (inline_call): Update.
4868 (want_inline_recursively): Guard division by zero.
4869 (recursive_inlining): Update.
4870 * ipa-inline.h (clone_inlined_nodes): Update.
4872 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4875 * config/i386/i386.c (classify_argument): Pass structures of size
4876 64 bytes or less in register.
4878 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4879 Kirill Yukhin <kirill.yukhin@intel.com>
4881 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
4882 (_mm_rcp28_round_ss): Ditto.
4883 (_mm_rsqrt28_round_sd): Ditto.
4884 (_mm_rsqrt28_round_ss): Ditto.
4885 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
4886 (_mm_rcp14_round_ss): Ditto.
4887 (_mm_rsqrt14_round_sd): Ditto.
4888 (_mm_rsqrt14_round_ss): Ditto.
4889 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
4890 the first input operand, get rid of match_dup.
4891 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
4893 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
4895 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
4896 operand as the first input operand, set type attribute.
4897 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
4899 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
4900 operand as the first input operand, set type attribute.
4902 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4904 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
4907 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
4910 * config/i386/i386.c (construct_container): Remove TFmode check
4911 for X86_64_INTEGER_CLASS.
4913 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
4916 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
4917 only when -Wpsabi is enabled.
4919 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
4922 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
4923 passing arrays in registers are the same as for structs, so remove the
4924 special case for them.
4926 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
4928 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
4929 destination type, extract only the valid bits if the source type is not
4930 integral and has a different mode.
4932 2014-02-19 Richard Biener <rguenther@suse.de>
4935 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
4938 2014-02-19 Richard Biener <rguenther@suse.de>
4941 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
4942 (ipa_modify_call_arguments): Emit an argument load explicitely and
4943 preserve virtual SSA form there and for the replacement call.
4944 Do not update SSA form nor free dominance info.
4946 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4948 * ipa.c (function_and_variable_visibility): Also clear WEAK
4949 flag when disolving COMDAT_GROUP.
4951 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4953 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
4954 * ipa-prop.c (ipa_set_jf_known_type): Return early when
4956 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
4957 do more sanity checks.
4958 (detect_type_change): Return true when giving up early.
4959 (compute_complex_assign_jump_func): Fix type parameter of
4960 ipa_set_ancestor_jf.
4961 (compute_complex_ancestor_jump_func): Likewise.
4962 (update_jump_functions_after_inlining): Fix updating of
4964 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
4966 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4968 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
4969 inline clones when edge disappears.
4971 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4974 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
4975 Split 64-bit moves into 2 patterns. Do not allow the use of
4976 direct move for TDmode in little endian, since the decimal value
4977 has little endian bytes within a word, but the 64-bit pieces are
4978 ordered in a big endian fashion, and normal subreg's of TDmode are
4980 (mov<mode>_64bit_dm): Likewise.
4981 (movtd_64bit_nodm): Likewise.
4983 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4985 PR tree-optimization/60174
4986 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
4987 statement of an SSA_NAME that occurs in an abnormal PHI node.
4989 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4992 * final.c (SEEN_BB): Remove.
4993 (SEEN_NOTE, SEEN_EMITTED): Renumber.
4994 (final_scan_insn): Don't force_source_line on second
4995 NOTE_INSN_BASIC_BLOCK.
4997 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
5000 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
5001 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
5002 (type_natural_mode): Warn ABI change when %zmm register is not
5003 available for AVX512F vector value passing.
5005 2014-02-18 Kai Tietz <ktietz@redhat.com>
5008 * config/i386/i386.c (ix86_expand_prologue): Use value in
5009 rax register as displacement when restoring %r10 or %rax.
5010 Fix wrong offset when restoring both registers.
5012 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5014 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
5015 assertion with conditional return.
5017 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5018 Uros Bizjak <ubizjak@gmail.com>
5021 * config/i386/driver-i386.c (host_detect_local_cpu): If
5022 YMM state is not saved by the OS, also clear has_f16c. Move
5023 CPUID 0x80000001 handling before YMM state saving checking.
5025 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
5027 PR rtl-optimization/58960
5028 * haifa-sched.c (alloc_global_sched_pressure_data): New,
5029 factored out from ...
5030 (sched_init): ... here.
5031 (free_global_sched_pressure_data): New, factored out from ...
5032 (sched_finish): ... here.
5033 * sched-int.h (free_global_sched_pressure_data): Declare.
5034 * sched-rgn.c (nr_regions_initial): New static global.
5035 (haifa_find_rgns): Initialize it.
5036 (schedule_region): Disable sched-pressure for the newly
5039 2014-02-17 Richard Biener <rguenther@suse.de>
5041 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
5042 release SSA defs of pattern stmts.
5044 2014-02-17 Richard Biener <rguenther@suse.de>
5046 * tree-inline.c (expand_call_inline): Release the virtual
5047 operand defined by the call we are about to inline.
5049 2014-02-17 Richard Biener <rguenther@suse.de>
5051 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
5053 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
5054 Ilya Tocar <ilya.tocar@intel.com>
5056 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
5057 arguments order in builtin.
5058 (_mm512_permutexvar_epi64): Ditto.
5059 (_mm512_mask_permutexvar_epi64): Ditto
5060 (_mm512_maskz_permutexvar_epi32): Ditto
5061 (_mm512_permutexvar_epi32): Ditto
5062 (_mm512_mask_permutexvar_epi32): Ditto
5064 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5066 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
5067 (p8_vmrgow): Likewise.
5069 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5071 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
5074 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
5077 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
5078 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
5079 into 64-bit and 32-bit moves. On 64-bit moves, add support for
5080 using direct move instructions on ISA 2.07. Also adjust
5081 instruction length for 64-bit.
5082 (mov<mode>_64bit, TFmode/TDmode): Likewise.
5083 (mov<mode>_32bit, TFmode/TDmode): Likewise.
5085 2014-02-15 Alan Modra <amodra@gmail.com>
5089 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
5090 find_replacement on parts of insn rtl that might be reloaded.
5092 2014-02-15 Richard Biener <rguenther@suse.de>
5094 PR tree-optimization/60183
5095 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
5096 (tree_ssa_phiprop): Calculate and free post-dominators.
5098 2014-02-14 Jeff Law <law@redhat.com>
5100 PR rtl-optimization/60131
5101 * ree.c (get_extended_src_reg): New function.
5102 (combine_reaching_defs): Use it rather than assuming location of REG.
5103 (find_and_remove_re): Verify first operand of extension is
5104 a REG before adding the insns to the copy list.
5106 2014-02-14 Roland McGrath <mcgrathr@google.com>
5108 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
5109 * configure: Regenerated.
5110 * config.in: Regenerated.
5111 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
5112 instead of ASM_SHORT.
5114 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
5115 Richard Earnshaw <rearnsha@arm.com>
5117 PR rtl-optimization/59535
5118 * lra-constraints.c (process_alt_operands): Encourage alternative
5119 when unassigned pseudo class is superset of the alternative class.
5120 (inherit_reload_reg): Don't inherit when optimizing for code size.
5121 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
5122 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
5123 modes not less than 4 for Thumb1.
5125 2014-02-14 Kyle McMartin <kyle@redhat.com>
5128 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
5130 2014-02-14 Richard Biener <rguenther@suse.de>
5132 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
5133 (get_frame_arg): Drop the assert with langhook types_compatible_p.
5134 Do not strip INDIRECT_REFs.
5136 2014-02-14 Richard Biener <rguenther@suse.de>
5139 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
5140 DECL_FUNCTION_SPECIFIC_TARGET.
5141 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
5142 * tree-streamer-out.c (pack_ts_target_option): Remove.
5143 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
5144 (write_ts_function_decl_tree_pointers): Do not stream
5145 DECL_FUNCTION_SPECIFIC_TARGET.
5146 * tree-streamer-in.c (unpack_ts_target_option): Remove.
5147 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
5148 (lto_input_ts_function_decl_tree_pointers): Do not stream
5149 DECL_FUNCTION_SPECIFIC_TARGET.
5151 2014-02-14 Jakub Jelinek <jakub@redhat.com>
5153 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
5154 (get_initial_def_for_induction, vectorizable_induction): Ignore
5155 debug stmts when looking for exit_phi.
5156 (vectorizable_live_operation): Fix up condition.
5158 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5160 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5161 nreverse() because it changes the content of original tree list.
5163 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5165 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5166 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5168 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5170 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5171 GNU coding standards.
5173 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5176 * dwarf2out.c (gen_subprogram_die): Don't call
5177 add_calling_convention_attribute if subr_die is old_die.
5179 2014-02-13 Sharad Singhai <singhai@google.com>
5181 * doc/optinfo.texi: Fix order of nodes.
5183 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5185 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5186 operands[2], not operands[3].
5188 2014-02-13 Richard Biener <rguenther@suse.de>
5191 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5192 mention the possibility of an in-tree build.
5193 (CLooG): Update recommended version to 0.18.1, mention the
5194 possibility of an in-tree build and clarify that the ISL
5195 bundled with CLooG does not work.
5197 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5200 * expr.c (compress_float_constant): If x is a hard register,
5201 extend into a pseudo and then move to x.
5203 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5205 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5206 caused by bad second argument to warning_at() with -mhotpatch and
5207 nested functions (e.g. with gfortran).
5209 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5211 * opts.c (option_name): Remove "enabled by default" rider.
5213 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5215 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5217 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5218 Uros Bizjak <ubizjak@gmail.com>
5221 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5222 * configure: Regenerated.
5224 2014-02-12 Richard Biener <rguenther@suse.de>
5226 * vec.c (vec_prefix::calculate_allocation): Move as
5227 inline variant to vec.h.
5228 (vec_prefix::calculate_allocation_1): New out-of-line version.
5229 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5230 (vec_prefix::m_has_auto_buf): Rename to ...
5231 (vec_prefix::m_using_auto_storage): ... this.
5232 (vec_prefix::calculate_allocation): Inline the easy cases
5233 and dispatch to calculate_allocation_1 which doesn't need the
5235 (va_heap::reserve): Use gcc_checking_assert.
5236 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5237 m_using_auto_storage.
5238 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5240 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5241 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5242 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5244 2014-02-12 Richard Biener <rguenther@suse.de>
5246 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5247 when we found a dependence.
5249 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5251 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5253 (maybe_fold_stmt): ... into this new function.
5254 * omp-low.c (lower_omp): Update comment.
5256 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5259 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5262 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5264 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5265 identifiers in comments.
5266 (cortexa53_extra_costs): Likewise.
5267 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5268 (cortexa7_extra_costs): Likewise.
5269 (cortexa12_extra_costs): Likewise.
5270 (cortexa15_extra_costs): Likewise.
5271 (v7m_extra_costs): Likewise.
5273 2014-02-12 Richard Biener <rguenther@suse.de>
5276 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5277 of posix_memalign being successful.
5278 (lower_stmt): Restrict lowering of posix_memalign to when
5279 -ftree-bit-ccp is enabled.
5281 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5283 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5285 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5287 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5289 PR rtl-optimization/60116
5290 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5291 other_insn once the combination has been validated.
5293 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5296 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5298 * ipa-devirt.c: Include demangle.h
5299 (odr_violation_reported): New static variable.
5300 (add_type_duplicate): Update odr_violations.
5301 (maybe_record_node): Add completep parameter; update it.
5302 (record_target_from_binfo): Add COMPLETEP parameter;
5303 update it as needed.
5304 (possible_polymorphic_call_targets_1): Likewise.
5305 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5306 rename FINAL to COMPLETE.
5307 (record_targets_from_bases): Sanity check we found the binfo;
5308 fix COMPLETEP updating.
5309 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5310 parameter, fix computing of COMPLETEP.
5311 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5312 at LTO time do demangling.
5313 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5314 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5316 (gimple_get_virt_method_for_binfo): Likewise.
5317 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5318 gimple_get_virt_method_for_vtable): Update prototypes.
5320 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5323 * genautomata.c (add_presence_absence): Fix typo with
5324 {final_}presence_list.
5326 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5329 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5330 for VSX/Altivec vectors that land in GPR registers.
5332 2014-02-11 Richard Henderson <rth@redhat.com>
5333 Jakub Jelinek <jakub@redhat.com>
5336 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5337 around drhs if type conversion to lacc->type is not useless.
5339 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5341 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5343 (cortex-a57.cortex-a53): Likewise.
5344 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5346 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5348 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5351 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5353 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5354 add_options_for_arm_vfp3.
5356 2014-02-11 Jeff Law <law@redhat.com>
5359 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5360 object with an undesirable mode.
5362 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5365 * config/i386/sol2-9.h: New file.
5366 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5367 *-*-solaris2.9*): Use it.
5369 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5371 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5372 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5374 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5376 * config/microblaze/microblaze.c: Extend mcpu version format
5378 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5380 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5382 2014-02-10 Richard Henderson <rth@redhat.com>
5385 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5386 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5387 ms-abi vs -mno-accumulate-outgoing-args.
5388 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5389 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5392 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5395 * cfgexpand.c (expand_asm_operands): Attach source location to
5396 ASM_INPUT rtx objects.
5397 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5399 2014-02-10 Nick Clifton <nickc@redhat.com>
5401 * config/mn10300/mn10300.c (popcount): New function.
5402 (mn10300_expand_prologue): Include saved registers in stack usage
5405 2014-02-10 Jeff Law <law@redhat.com>
5408 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5409 when changing the SET_DEST of a prior insn to avoid an input reload.
5411 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5413 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5414 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5415 -mcall-openbsd, or -mcall-linux.
5416 (CC1_ENDIAN_BIG_SPEC): Remove.
5417 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5418 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5419 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5420 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5421 and %cc1_endian_default.
5422 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5424 2014-02-10 Richard Biener <rguenther@suse.de>
5426 PR tree-optimization/60115
5427 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5428 MEM_REF handling. Properly verify that the accesses are not
5429 out of the objects bound.
5431 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5433 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5436 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5438 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5439 proper constants and fix formatting.
5440 (possible_polymorphic_call_targets): Fix formatting.
5442 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5443 Ilya Tocar <ilya.tocar@intel.com>
5445 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5446 (_mm512_loadu_epi32): Renamed into...
5447 (_mm512_loadu_si512): This.
5448 (_mm512_storeu_epi32): Renamed into...
5449 (_mm512_storeu_si512): This.
5450 (_mm512_maskz_ceil_ps): Removed.
5451 (_mm512_maskz_ceil_pd): Ditto.
5452 (_mm512_maskz_floor_ps): Ditto.
5453 (_mm512_maskz_floor_pd): Ditto.
5454 (_mm512_floor_round_ps): Ditto.
5455 (_mm512_floor_round_pd): Ditto.
5456 (_mm512_ceil_round_ps): Ditto.
5457 (_mm512_ceil_round_pd): Ditto.
5458 (_mm512_mask_floor_round_ps): Ditto.
5459 (_mm512_mask_floor_round_pd): Ditto.
5460 (_mm512_mask_ceil_round_ps): Ditto.
5461 (_mm512_mask_ceil_round_pd): Ditto.
5462 (_mm512_maskz_floor_round_ps): Ditto.
5463 (_mm512_maskz_floor_round_pd): Ditto.
5464 (_mm512_maskz_ceil_round_ps): Ditto.
5465 (_mm512_maskz_ceil_round_pd): Ditto.
5466 (_mm512_expand_pd): Ditto.
5467 (_mm512_expand_ps): Ditto.
5468 * config/i386/i386.c (ix86_builtins): Remove
5469 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5470 (bdesc_args): Ditto.
5471 * config/i386/predicates.md (const1256_operand): New.
5472 (const_1_to_2_operand): Ditto.
5473 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5474 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5475 (*avx512pf_gatherpf<mode>sf): Ditto.
5476 (avx512pf_gatherpf<mode>df): Ditto.
5477 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5478 (*avx512pf_gatherpf<mode>df): Ditto.
5479 (avx512pf_scatterpf<mode>sf): Ditto.
5480 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5481 (*avx512pf_scatterpf<mode>sf): Ditto.
5482 (avx512pf_scatterpf<mode>df): Ditto.
5483 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5484 (*avx512pf_scatterpf<mode>df): Ditto.
5485 (avx512f_expand<mode>): Removed.
5486 (<shift_insn><mode>3<mask_name>): Change predicate type.
5488 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5490 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5491 not at the end of datarefs vector use ordered_remove to avoid
5492 reordering datarefs vector.
5495 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5496 mark local addressable non-static vars as GOVD_PRIVATE
5497 instead of GOVD_LOCAL.
5498 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5499 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5503 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5504 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5505 assume_aligned or alloc_align attributes.
5506 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5507 arguments. Handle also assume_aligned and alloc_align attributes.
5508 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5509 calls to functions with assume_aligned or alloc_align attributes.
5510 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5512 2014-02-08 Terry Guo <terry.guo@arm.com>
5514 * doc/invoke.texi: Document ARM -march=armv7e-m.
5516 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5518 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5519 flag on __cilkrts_rethrow builtin.
5522 * ipa-cp.c (determine_versionability): Fail at -O0
5523 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5524 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5527 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5530 * tree-inline.c (copy_forbidden): Fail for
5531 __attribute__((optimize (0))) functions.
5533 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5535 * varpool.c: Include pointer-set.h.
5536 (varpool_remove_unreferenced_decls): Variables in other partitions
5537 will not be output; be however careful to not lose information
5540 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5542 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5543 lookup in the vtable constructor.
5545 2014-02-07 Jeff Law <law@redhat.com>
5548 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5549 define_insn_and_split.
5551 * ipa-inline.c (inline_small_functions): Fix typos.
5553 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5555 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5556 (s390_can_use_return_insn): Declare.
5557 * config/s390/s390.h (EPILOGUE_USES): Define.
5558 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5560 (s390_chunkify_start): Handle return JUMP_LABELs.
5561 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5562 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5563 (s390_can_use_return_insn): New functions.
5564 (s390_fix_long_loop_prediction): Handle conditional returns.
5565 (TARGET_SET_UP_BY_PROLOGUE): Define.
5566 * config/s390/s390.md (ANY_RETURN): New code iterator.
5567 (*creturn, *csimple_return, return, simple_return): New patterns.
5569 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5571 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5572 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5573 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5574 REG_CFA_RESTORE list when deciding not to restore a register.
5576 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5578 * config/s390/s390.c: Include tree-pass.h and context.h.
5579 (s390_early_mach): New function, split out from...
5580 (s390_emit_prologue): ...here.
5581 (pass_data_s390_early_mach): New pass structure.
5582 (pass_s390_early_mach): New class.
5583 (s390_option_override): Create and register early_mach pass.
5584 Move to end of file.
5586 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5588 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5589 to match for the exit block.
5591 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5593 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5594 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5595 Reject misaligned operands.
5597 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5599 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5601 2014-02-07 Richard Biener <rguenther@suse.de>
5604 * gimple-low.c (lower_builtin_posix_memalign): New function.
5605 (lower_stmt): Call it to lower posix_memalign in a way
5606 to make alignment info accessible.
5608 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5611 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5612 __builtin_setjmp_receiver.
5614 2014-02-07 Richard Biener <rguenther@suse.de>
5617 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5618 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5619 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5620 Handle BUILT_IN_POSIX_MEMALIGN.
5621 (find_func_clobbers): Likewise.
5622 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5623 (call_may_clobber_ref_p_1): Likewise.
5625 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5628 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5631 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5634 * lto-cgraph.c (lto_output_node): Use
5635 symtab_get_symbol_partitioning_class.
5636 (lto_output_varpool_node): likewise.
5637 (symtab_get_symbol_partitioning_class): Move here from
5639 * cgraph.h (symbol_partitioning_class): Likewise.
5640 (symtab_get_symbol_partitioning_class): Declare.
5642 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5644 * ggc.h (ggc_internal_cleared_alloc): New macro.
5645 * vec.h (vec_safe_copy): Handle memory stats.
5646 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5647 * target-globals.c (save_target_globals): Likewise.
5649 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5652 * expr.c (emit_move_resolve_push): Export; be bit more selective
5653 on when to clear alias set.
5654 * expr.h (emit_move_resolve_push): Declare.
5655 * function.h (struct function): Add tail_call_marked.
5656 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5657 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5658 * config/i386/i386.md (TImode move expander): De not call
5660 (FP push expanders): Preserve memory attributes.
5661 * config/i386/sse.md (push<mode>1): Remove.
5662 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5663 (ix86_expand_push): Remove.
5664 * config/i386/mmx.md (push<mode>1): Remove.
5666 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5668 PR rtl-optimization/60030
5669 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5670 lopart with paradoxical subreg before shifting it up by hprec.
5672 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5674 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5675 Remove extra newline at end of file.
5676 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5677 (arm_issue_rate): Handle cortexa57.
5678 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5679 (cortex-a57.cortex-a53): Likewise.
5681 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5684 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5685 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5687 (arm_expand_prologue): Adjust all callers.
5688 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5689 info, registers also at the lowest numbered registers side. Use
5690 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5694 * var-tracking.c (adjust_mems): Before adding a SET to
5695 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5697 2014-02-06 Alan Modra <amodra@gmail.com>
5700 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5701 change SDmode to DDmode when lra_in_progress.
5703 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5706 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5707 free_data_ref on the dr first, and before goto again also set dr
5708 to the next dr. For simd_lane_access, free old datarefs[i] before
5709 overwriting it. For get_vectype_for_scalar_type failure, don't
5710 free_data_ref if simd_lane_access.
5712 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5715 * tree.h (opts_for_fn): New inline function.
5716 (opt_for_fn): Define.
5717 * config/i386/i386.c (ix86_function_regparm): Use
5718 opt_for_fn (decl, optimize) instead of optimize.
5720 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5722 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5723 for SYMBOL_REF in large memory model.
5725 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5727 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5729 (cortex-a57): Likewise.
5730 (cortex-a57.cortex-a53): Likewise.
5732 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5733 Kugan Vivekanandarajah <kuganv@linaro.org>
5735 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5737 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5739 2014-02-06 Richard Biener <rguenther@suse.de>
5741 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5742 set_loop_copy and initialize_original_copy_tables.
5744 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5746 * config/aarch64/aarch64-simd.md
5747 (aarch64_ashr_simddi): Change QI to SI.
5749 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5750 Jakub Jelinek <jakub@redhat.com>
5753 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5756 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5758 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5759 CODE_FOR_altivec_vpku[hw]um to
5760 CODE_FOR_altivec_vpku[hw]um_direct.
5761 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5762 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5763 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5764 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5766 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5768 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5769 generation for -maltivec=be.
5770 (altivec_vsumsws): Simplify redundant test.
5772 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5774 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5775 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5776 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5777 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5778 gen_altivec_vpkuwum.
5779 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5781 (altivec_vpks<VI_char>ss): Likewise.
5782 (altivec_vpks<VI_char>us): Likewise.
5783 (altivec_vpku<VI_char>us): Likewise.
5784 (altivec_vpku<VI_char>um): Likewise.
5785 (altivec_vpku<VI_char>um_direct): New (copy of
5786 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5788 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5789 target is little endian and -maltivec=be is not specified.
5790 (*altivec_vupkhs<VU_char>_direct): New (copy of
5791 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
5792 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5793 target is little endian and -maltivec=be is not specified.
5794 (*altivec_vupkls<VU_char>_direct): New (copy of
5795 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
5796 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5797 little endian and -maltivec=be is not specified.
5798 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5799 little endian and -maltivec=be is not specified.
5801 2014-02-05 Richard Henderson <rth@redhat.com>
5804 * combine-stack-adj.c: Revert r206943.
5805 * sched-int.h (struct deps_desc): Add last_args_size.
5806 * sched-deps.c (init_deps): Initialize it.
5807 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5808 contain REG_ARGS_SIZE notes.
5810 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5812 * lto-cgraph.c (asm_nodes_output): Make global.
5813 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
5814 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
5815 (driver_handle_option): Handle OPT_fwpa.
5817 2014-02-05 Jakub Jelinek <jakub@redhat.com>
5820 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5821 a comment typo and formatting issue. If odr_hash hasn't been
5822 created, return vNULL and set *completep to false.
5825 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
5826 bb with no successors.
5828 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
5831 * doc/invoke.texi (-march): Clarify documentation for ARM.
5835 2014-02-05 Richard Biener <rguenther@suse.de>
5837 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
5838 when not vectorizing because of too many alias checks.
5839 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
5840 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
5842 2014-02-05 Nick Clifton <nickc@redhat.com>
5844 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
5845 accept extended registers in any mode when compiling for the MN10300.
5847 2014-02-05 Yury Gribov <y.gribov@samsung.com>
5849 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
5850 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
5851 sanitization attributes.
5852 (can_inline_edge_p): Likewise.
5853 (sanitize_attrs_match_for_inline_p): New function.
5855 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5857 * ipa-prop.c (detect_type_change): Shor circuit testing of
5858 type changes on THIS pointer.
5860 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
5863 * config/pa/pa.c (legitimize_tls_address): Return original address
5864 if not passed a SYMBOL_REF rtx.
5865 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
5867 (pa_emit_move_sequence): Simplify TLS source operands.
5868 (pa_legitimate_constant_p): Reject all TLS constants.
5869 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
5870 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
5872 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5874 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
5875 groups when we know they are controlled by LTO.
5876 * varasm.c (default_binds_local_p_1): If object is in other partition,
5877 it will be resolved locally.
5879 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5881 * config/host-linux.c (linux_gt_pch_use_address): Don't
5882 use SSIZE_MAX because it is not always defined.
5884 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
5887 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
5888 threshold for pseudo splitting.
5889 (update_ebb_live_info): Process call argument hard registers and
5890 hard registers from insn definition too.
5891 (max_small_class_regs_num): New constant.
5892 (inherit_in_ebb): Update live hard regs through EBBs. Update
5893 reloads_num only for small register classes. Don't split for
5896 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
5899 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
5902 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5904 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
5907 2014-02-04 Marek Polacek <polacek@redhat.com>
5909 * gdbinit.in (pel): Define.
5911 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5913 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
5916 2014-02-04 Richard Biener <rguenther@suse.de>
5919 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
5920 in function context local.
5921 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
5922 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
5923 similar to LTO_imported_decl_ref.
5925 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5927 PR tree-optimization/60002
5928 * cgraphclones.c (build_function_decl_skip_args): Clear
5931 PR tree-optimization/60023
5932 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
5933 false to gsi_replace.
5934 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
5935 has been in some EH region and vec_stmt could throw, add
5936 vec_stmt into the same EH region.
5937 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
5938 has no lhs, ignore it.
5939 * internal-fn.c (expand_MASK_LOAD): Likewise.
5942 * tree-inline.c (copy_forbidden): Fail for
5943 __attribute__((optimize (0))) functions.
5946 * omp-low.c (simd_clone_struct_copy): If from->inbranch
5947 is set, copy one less argument.
5948 (expand_simd_clones): Don't subtract clone_info->inbranch
5949 from simd_clone_struct_alloc argument.
5951 PR rtl-optimization/57915
5952 * recog.c (simplify_while_replacing): If all unary/binary/relational
5953 operation arguments are constant, attempt to simplify those.
5956 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
5957 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
5959 2014-02-04 Richard Biener <rguenther@suse.de>
5961 PR tree-optimization/60012
5962 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
5963 TBAA disambiguation to all DDRs.
5965 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5968 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
5969 (LINK_SPEC): Use it for -shared, -shared-libgcc.
5971 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5974 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
5976 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5978 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
5979 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
5981 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5984 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
5985 to figure out targets of polymorphic calls with known decl.
5986 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5987 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
5988 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
5989 (get_polymorphic_call_info): ... here.
5990 (get_polymorphic_call_info_from_invariant): New function.
5992 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5994 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
5995 lookup via vtable pointer; check for type consistency
5996 and turn inconsitent facts into UNREACHABLE.
5997 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5998 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
5999 type inconsistent querries; return UNREACHABLE instead.
6001 2014-02-03 Richard Henderson <rth@twiddle.net>
6004 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
6005 already processed this node.
6006 (normalize_one_pred_1): Pass along mark_set.
6007 (normalize_one_pred): Create and destroy a pointer_set_t.
6008 (normalize_one_pred_chain): Likewise.
6010 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
6012 PR gcov-profile/58602
6013 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
6015 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6018 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
6019 -fno-devirtualize; try to devirtualize by the knowledge of
6020 virtual table pointer given by aggregate propagation.
6021 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6022 (ipa_print_node_jump_functions): Dump also offset that
6023 is relevant for polymorphic calls.
6024 (determine_known_aggregate_parts): Add arg_type parameter; use it
6025 instead of determining the type from pointer type.
6026 (ipa_compute_jump_functions_for_edge): Update call of
6027 determine_known_aggregate_parts.
6028 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
6029 (gimple_get_virt_method_for_binfo): ... here; simplify using
6030 vtable_pointer_value_to_vtable.
6031 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
6032 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
6033 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
6034 (vtable_pointer_value_to_vtable): Break out from ...; handle also
6036 (vtable_pointer_value_to_binfo): ... here.
6037 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
6039 2014-02-03 Teresa Johnson <tejohnson@google.com>
6041 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
6042 redef of outer loop index variable.
6044 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
6048 * doc/extend.texi (Function Attributes): Typo.
6050 2014-02-03 Cong Hou <congh@google.com>
6052 PR tree-optimization/60000
6053 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
6054 if the vectorized statement is a store. A store statement can only
6055 appear at the end of pattern statements.
6057 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6059 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
6060 (ix86_option_override_internal): Default long double to 64-bit for
6061 32-bit Bionic and to 128-bit for 64-bit Bionic.
6063 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
6064 TARGET_LONG_DOUBLE_128 is true.
6065 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
6067 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
6068 (mlong-double-64): Negate -mlong-double-128.
6069 (mlong-double-128): New option.
6071 * config/i386/i386-c.c (ix86_target_macros): Define
6072 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
6074 * doc/invoke.texi: Document -mlong-double-128.
6076 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6078 PR rtl-optimization/60024
6079 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
6081 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
6083 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
6085 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6087 PR rtl-optimization/57662
6088 * sel-sched.c (code_motion_path_driver): Do not mark already not
6089 existing blocks in the visiting bitmap.
6091 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6093 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
6094 on the insn being emitted.
6096 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
6097 Will Deacon <will.deacon@arm.com>
6099 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
6101 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6103 * config/arm/arm-tables.opt: Regenerate.
6105 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6107 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
6108 for vector types other than V16QImode.
6109 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
6110 define_expand, and call altivec_expand_vec_perm_le when producing
6111 code with little endian element order.
6112 (*altivec_vperm_<mode>_internal): New insn having previous
6113 behavior of altivec_vperm_<mode>.
6114 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
6115 altivec_expand_vec_perm_le when producing code with little endian
6117 (*altivec_vperm_<mode>_uns_internal): New insn having previous
6118 behavior of altivec_vperm_<mode>_uns.
6120 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6122 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
6123 (altivec_vsumsws): Add handling for -maltivec=be with a little
6125 (altivec_vsumsws_direct): New.
6126 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
6127 gen_altivec_vsumsws.
6129 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6131 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
6132 vtable_pointer_value_to_binfo): New functions.
6133 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
6134 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
6136 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
6138 * config/nios2/nios2.md (load_got_register): Initialize GOT
6139 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
6140 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
6142 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6144 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
6145 preserverd by passthrough, do not propagate the type.
6147 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6149 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
6150 (mips_atomic_assign_expand_fenv): New function.
6151 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
6153 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6155 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6156 (__builtin_mips_set_fcsr): Likewise.
6157 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6158 MIPS_USI_FTYPE_VOID.
6159 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6160 (mips16_expand_set_fcsr): Likewise.
6161 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6162 (mips16_set_fcsr_stub): Likewise.
6163 (mips16_get_fcsr_one_only_stub): New class.
6164 (mips16_set_fcsr_one_only_stub): Likewise.
6165 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6166 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6167 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6168 (hard_float): New availability predicate.
6169 (mips_builtins): Add get_fcsr and set_fcsr.
6170 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6171 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6172 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6173 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6174 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6177 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6179 * config/mips/mips.c (mips_one_only_stub): New class.
6180 (mips_need_mips16_rdhwr_p): Replace with...
6181 (mips16_rdhwr_stub): ...this new variable.
6182 (mips16_stub_call_address): New function.
6183 (mips16_rdhwr_one_only_stub): New class.
6184 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6185 (mips_output_mips16_rdhwr): Delete.
6186 (mips_finish_stub): New function.
6187 (mips_code_end): Use it to handle rdhwr stubs.
6189 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6192 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6193 when calculating size of integer atomic types.
6195 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6197 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6199 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6201 PR tree-optimization/60003
6202 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6203 * profile.c (branch_prob): Use gimple_call_builtin_p
6204 to check for BUILT_IN_SETJMP_RECEIVER.
6205 * tree-inline.c (copy_bb): Call notice_special_calls.
6207 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6210 * lra-constraints.c (process_alt_operands): Update reload_sum only
6213 2014-01-31 Richard Henderson <rth@redhat.com>
6216 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6217 until after else_eh is processed.
6219 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6221 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6222 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6223 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6224 in smmintrin.h, remove them.
6225 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6226 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6227 * config/i386/i386.md (ROUND_SAE): Fix value.
6228 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6229 (const48_operand): New.
6230 * config/i386/subst.md (round), (round_expand): Use
6231 const_4_or_8_to_11_operand.
6232 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6234 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6236 * config/i386/constraints.md (Yk): Swap meaning with k.
6237 * config/i386/i386.md (movhi_internal): Change Yk to k.
6238 (movqi_internal): Ditto.
6239 (*k<logic><mode>): Ditto.
6242 (kandn<mode>): Ditto.
6243 (*<code>hi_1): Ditto.
6244 (*<code>qi_1): Ditto.
6245 (kxnor<mode>): Ditto.
6246 (kortestzhi): Ditto.
6247 (kortestchi): Ditto.
6249 (*one_cmplhi2_1): Ditto.
6250 (*one_cmplqi2_1): Ditto.
6251 * config/i386/sse.md (): Change k to Yk.
6252 (avx512f_load<mode>_mask): Ditto.
6253 (avx512f_blendm<mode>): Ditto.
6254 (avx512f_store<mode>_mask): Ditto.
6255 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6256 (avx512f_storedqu<mode>_mask): Ditto.
6257 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6259 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6260 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6261 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6262 (avx512f_maskcmp<mode>3): Ditto.
6263 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6264 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6265 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6266 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6267 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6268 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6269 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6270 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6271 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6272 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6273 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6274 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6275 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6276 (vec_extract_lo_<mode>_maskm): Ditto.
6277 (vec_extract_hi_<mode>_maskm): Ditto.
6278 (avx512f_vternlog<mode>_mask): Ditto.
6279 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6280 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6281 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6282 (avx512f_<code>v8div16qi2_mask): Ditto.
6283 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6284 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6285 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6286 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6287 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6288 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6289 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6290 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6291 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6292 (avx512cd_maskb_vec_dupv8di): Ditto.
6293 (avx512cd_maskw_vec_dupv16si): Ditto.
6294 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6295 (avx512f_vpermi2var<mode>3_mask): Ditto.
6296 (avx512f_vpermi2var<mode>3_mask): Ditto.
6297 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6298 (*avx512f_gathersi<mode>): Ditto.
6299 (*avx512f_gathersi<mode>_2): Ditto.
6300 (*avx512f_gatherdi<mode>): Ditto.
6301 (*avx512f_gatherdi<mode>_2): Ditto.
6302 (*avx512f_scattersi<mode>): Ditto.
6303 (*avx512f_scatterdi<mode>): Ditto.
6304 (avx512f_compress<mode>_mask): Ditto.
6305 (avx512f_compressstore<mode>_mask): Ditto.
6306 (avx512f_expand<mode>_mask): Ditto.
6307 * config/i386/subst.md (mask): Change k to Yk.
6308 (mask_scalar_merge): Ditto.
6311 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6313 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6315 2014-01-31 Richard Biener <rguenther@suse.de>
6318 * builtins.c (fold_builtin_memory_op): Make sure to not
6319 use a floating-point mode or a boolean or enumeral type for
6322 2014-01-30 DJ Delorie <dj@redhat.com>
6324 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6325 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6326 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6327 whenever main() has an epilogue.
6329 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6331 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6332 unused variable "field".
6333 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6334 (vsx_mergeh_<mode>): Likewise.
6335 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6336 (altivec_vmrghh): Likewise.
6337 (altivec_vmrghw): Likewise.
6338 (altivec_vmrglb): Likewise.
6339 (altivec_vmrglh): Likewise.
6340 (altivec_vmrglw): Likewise.
6341 (altivec_vspltb): Add missing uses.
6342 (altivec_vsplth): Likewise.
6343 (altivec_vspltw): Likewise.
6344 (altivec_vspltsf): Likewise.
6346 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6349 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6350 frame related instructions.
6352 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6354 PR rtl-optimization/59959
6355 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6356 any reload of register whose subreg is invalid.
6358 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6360 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6361 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6362 Add missing return type - void.
6364 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6366 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6367 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6368 remove element index adjustment for endian (now handled in vsx.md
6370 (altivec_expand_vec_perm_const): Use
6371 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6372 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6373 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6374 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6375 define_expand and a new define_insn *altivec_vspltb_internal;
6376 adjust for -maltivec=be on a little endian target.
6377 (altivec_vspltb_direct): New.
6378 (altivec_vsplth): Divide into a define_expand and a new
6379 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6380 little endian target.
6381 (altivec_vsplth_direct): New.
6382 (altivec_vspltw): Divide into a define_expand and a new
6383 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6384 little endian target.
6385 (altivec_vspltw_direct): New.
6386 (altivec_vspltsf): Divide into a define_expand and a new
6387 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6388 a little endian target.
6390 2014-01-30 Richard Biener <rguenther@suse.de>
6392 PR tree-optimization/59993
6393 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6394 can propagate form the earlier stmt and avoid the transform
6395 when the intermediate result is needed.
6397 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6399 * README.Portability: Fix typo.
6401 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6403 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6404 comparison_operator with ordered_comparison_operator.
6406 2014-01-30 Nick Clifton <nickc@redhat.com>
6408 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6409 Rename to mn10300_store_multiple_regs.
6410 * config/mn10300/mn10300.c: Likewise.
6411 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6412 store_multiple_regs.
6413 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6414 Call mn10300_store_multiple_regs.
6416 2014-01-30 Nick Clifton <nickc@redhat.com>
6417 DJ Delorie <dj@redhat.com>
6419 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6420 %fp 2 to keep registers after it properly word-aligned.
6421 (rl78_alloc_physical_registers_umul): Handle the case where both
6422 input operands are the same.
6424 2014-01-30 Richard Biener <rguenther@suse.de>
6426 PR tree-optimization/59903
6427 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6430 2014-01-30 Jason Merrill <jason@redhat.com>
6433 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6436 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6438 2014-01-30 Richard Biener <rguenther@suse.de>
6440 PR tree-optimization/59951
6441 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6443 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6446 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6447 SFmode to DFmode case.
6449 2014-01-29 DJ Delorie <dj@redhat.com>
6451 * config/msp430/msp430.opt (-minrt): New.
6452 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6454 (ENDFILE_SPEC): Likewise.
6456 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6458 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6459 (estimate_function_body_sizes): Use it.
6461 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6464 * dwarf2out.c (is_cxx_auto): New.
6465 (is_base_type): Use it.
6466 (gen_type_die_with_usage): Likewise.
6468 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6470 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6471 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6472 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6473 -maltivec=be with LE targets.
6474 (vsx_mergeh_<mode>): Likewise.
6475 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6476 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6477 (altivec_vmrghb): Replace with define_expand and new
6478 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6479 (altivec_vmrghb_direct): New define_insn.
6480 (altivec_vmrghh): Replace with define_expand and new
6481 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6482 (altivec_vmrghh_direct): New define_insn.
6483 (altivec_vmrghw): Replace with define_expand and new
6484 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6485 (altivec_vmrghw_direct): New define_insn.
6486 (*altivec_vmrghsf): Adjust for endianness.
6487 (altivec_vmrglb): Replace with define_expand and new
6488 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6489 (altivec_vmrglb_direct): New define_insn.
6490 (altivec_vmrglh): Replace with define_expand and new
6491 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6492 (altivec_vmrglh_direct): New define_insn.
6493 (altivec_vmrglw): Replace with define_expand and new
6494 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6495 (altivec_vmrglw_direct): New define_insn.
6496 (*altivec_vmrglsf): Adjust for endianness.
6497 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6498 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6499 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6500 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6501 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6502 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6503 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6504 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6506 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6508 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6509 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6512 2014-01-29 Richard Biener <rguenther@suse.de>
6514 PR tree-optimization/58742
6515 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6516 associate_pointerplus_align.
6517 (associate_pointerplus_diff): New function.
6518 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6519 and associate_pointerplus_diff.
6521 2014-01-29 Richard Biener <rguenther@suse.de>
6523 * lto-streamer.h (LTO_major_version): Bump to 3.
6524 (LTO_minor_version): Reset to 0.
6526 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6528 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6529 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6530 (arm_file_start): Generate correct asm header for armv7ve.
6531 * config/arm/bpabi.h: Add multilib support for armv7ve.
6532 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6533 and cortex-a15 to armv7ve.
6534 * config/arm/t-aprofile: Add multilib support for armv7ve.
6535 * doc/invoke.texi: Document -march=armv7ve.
6537 2014-01-29 Richard Biener <rguenther@suse.de>
6539 PR tree-optimization/58742
6540 * tree-ssa-forwprop.c (associate_plusminus): Return true
6541 if we changed sth, defer EH cleanup to ...
6542 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6543 (simplify_mult): New function.
6545 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6548 PR tree-optimization/59920
6549 * tree.c (build_common_builtin_nodes): Remove
6550 __builtin_setjmp_dispatcher initialization.
6551 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6552 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6553 instead of gsi_after_labels + manually skipping debug stmts.
6554 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6555 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6556 * tree-inline.c (copy_edges_for_bb): Remove
6557 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6558 argument. Ignore computed_goto_p stmts. Don't call
6559 make_abnormal_goto_edges. If a call might need abnormal edges
6560 for non-local gotos, see if it already has an edge to
6561 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6562 with true argument, don't do anything then, otherwise add
6563 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6564 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6566 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6567 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6568 (lower_stmt): Don't set data->calls_builtin_setjmp.
6569 (lower_builtin_setjmp): Adjust comment.
6570 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6571 * tree-cfg.c (found_computed_goto): Remove.
6572 (factor_computed_gotos): Remove.
6573 (make_goto_expr_edges): Return bool, true for computed gotos.
6574 Don't call make_abnormal_goto_edges.
6575 (build_gimple_cfg): Don't set found_computed_goto, don't call
6576 factor_computed_gotos.
6577 (computed_goto_p): No longer static.
6578 (make_blocks): Don't set found_computed_goto.
6579 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6580 (make_edges): If make_goto_expr_edges returns true, push bb
6581 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6582 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6583 vector. Record mapping between bbs and OpenMP regions if there
6584 are any, adjust make_gimple_omp_edges caller. Call
6585 handle_abnormal_edges.
6586 (make_abnormal_goto_edges): Remove.
6587 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6588 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6589 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6590 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6591 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6592 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6593 filling *region also set *region_idx to (*region)->entry->index.
6596 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6597 For REGs set ORIGINAL_REGNO.
6599 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6601 * doc/md.texi: Mention that a target shouldn't implement
6602 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6605 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6607 PR tree-optimization/59594
6608 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6609 a copy of the datarefs vector rather than the vector itself.
6611 2014-01-28 Jason Merrill <jason@redhat.com>
6614 * dwarf2out.c (auto_die): New static.
6615 (gen_type_die_with_usage): Handle C++1y 'auto'.
6616 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6619 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6622 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6623 (SPEC_X32): Likewise.
6624 (SPEC_64): Likewise.
6625 * config/i386/i386.c (ix86_option_override_internal): Turn off
6626 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6628 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6629 * config/i386/i386.h (TARGET_16BIT): New macro.
6630 (TARGET_16BIT_P): Likewise.
6631 * config/i386/i386.opt: Add m16.
6632 * doc/invoke.texi: Document -m16.
6634 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6636 PR preprocessor/59935
6637 * input.c (location_get_source_line): Bail out on when line number
6638 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6640 2014-01-28 Richard Biener <rguenther@suse.de>
6642 PR tree-optimization/58742
6643 * tree-ssa-forwprop.c (associate_plusminus): Handle
6644 pointer subtraction of the form (T)(P + A) - (T)P.
6646 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6648 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6651 2014-01-28 Richard Biener <rguenther@suse.de>
6654 2014-01-28 Richard Biener <rguenther@suse.de>
6656 PR rtl-optimization/45364
6657 PR rtl-optimization/59890
6658 * var-tracking.c (local_get_addr_clear_given_value): Handle
6659 already cleared slot.
6660 (val_reset): Handle not allocated local_get_addr_cache.
6661 (vt_find_locations): Use post-order on the inverted CFG.
6663 2014-01-28 Richard Biener <rguenther@suse.de>
6665 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6667 2014-01-28 Richard Biener <rguenther@suse.de>
6669 PR rtl-optimization/45364
6670 PR rtl-optimization/59890
6671 * var-tracking.c (local_get_addr_clear_given_value): Handle
6672 already cleared slot.
6673 (val_reset): Handle not allocated local_get_addr_cache.
6674 (vt_find_locations): Use post-order on the inverted CFG.
6676 2014-01-28 Alan Modra <amodra@gmail.com>
6678 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6679 * configure.ac <recursive call for build != host>: Define
6680 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6681 and LD_FOR_BUILD too.
6682 * configure: Regenerate.
6684 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6686 * config/i386/i386.c (get_builtin_code_for_version): Separate
6687 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6688 Broadwell from Haswell.
6690 2014-01-27 Steve Ellcey <sellcey@mips.com>
6692 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6693 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6694 * config/mips/mips.c (mips_option_override): Change setting
6696 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6697 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6698 Change from Mask to Var.
6700 2014-01-27 Jeff Law <law@redhat.com>
6702 * ipa-inline.c (inline_small_functions): Fix typo.
6704 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6706 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6707 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6708 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6709 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6710 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6711 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6712 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6713 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6714 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6715 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6716 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6717 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6718 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6719 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6720 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6721 (_mm512_storeu_epi64): Ditto.
6722 (_mm512_cmpge_epi32_mask): Ditto.
6723 (_mm512_cmpge_epu32_mask): Ditto.
6724 (_mm512_cmpge_epi64_mask): Ditto.
6725 (_mm512_cmpge_epu64_mask): Ditto.
6726 (_mm512_cmple_epi32_mask): Ditto.
6727 (_mm512_cmple_epu32_mask): Ditto.
6728 (_mm512_cmple_epi64_mask): Ditto.
6729 (_mm512_cmple_epu64_mask): Ditto.
6730 (_mm512_cmplt_epi32_mask): Ditto.
6731 (_mm512_cmplt_epu32_mask): Ditto.
6732 (_mm512_cmplt_epi64_mask): Ditto.
6733 (_mm512_cmplt_epu64_mask): Ditto.
6734 (_mm512_cmpneq_epi32_mask): Ditto.
6735 (_mm512_cmpneq_epu32_mask): Ditto.
6736 (_mm512_cmpneq_epi64_mask): Ditto.
6737 (_mm512_cmpneq_epu64_mask): Ditto.
6738 (_mm512_expand_pd): Ditto.
6739 (_mm512_expand_ps): Ditto.
6740 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6741 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6742 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6743 * config/i386/i386.c (ix86_builtins): Add
6744 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6745 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6746 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6747 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6748 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6749 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6750 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6751 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6752 IX86_BUILTIN_PMOVUSQW512_MEM.
6753 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6754 __builtin_ia32_pmovsqd512mem_mask,
6755 __builtin_ia32_pmovqd512mem_mask,
6756 __builtin_ia32_pmovusqw512mem_mask,
6757 __builtin_ia32_pmovsqw512mem_mask,
6758 __builtin_ia32_pmovqw512mem_mask,
6759 __builtin_ia32_pmovusdw512mem_mask,
6760 __builtin_ia32_pmovsdw512mem_mask,
6761 __builtin_ia32_pmovdw512mem_mask,
6762 __builtin_ia32_pmovqb512mem_mask,
6763 __builtin_ia32_pmovusqb512mem_mask,
6764 __builtin_ia32_pmovsqb512mem_mask,
6765 __builtin_ia32_pmovusdb512mem_mask,
6766 __builtin_ia32_pmovsdb512mem_mask,
6767 __builtin_ia32_pmovdb512mem_mask.
6768 (bdesc_args): Add __builtin_ia32_expanddf512,
6769 __builtin_ia32_expandsf512.
6770 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6771 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6772 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6773 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6774 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6775 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6776 (avx512f_<code>v8div16qi2_mask_store): This.
6777 (avx512f_expand<mode>): New.
6779 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6781 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
6783 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6784 (_mm512_prefetch_i32scatter_pd): Ditto.
6785 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6786 (_mm512_prefetch_i64scatter_pd): Ditto.
6787 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6788 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6789 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6790 (_mm512_prefetch_i32scatter_ps): Ditto.
6791 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6792 (_mm512_prefetch_i64scatter_ps): Ditto.
6793 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6794 * config/i386/i386-builtin-types.def: Define
6795 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6796 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
6797 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6798 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6799 IX86_BUILTIN_SCATTERPFQPD.
6800 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6801 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6802 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6803 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6804 __builtin_ia32_scatterpfqps.
6805 (ix86_expand_builtin): Expand new built-ins.
6806 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6807 fix memory access data type.
6808 (*avx512pf_gatherpf<mode>_mask): Ditto.
6809 (*avx512pf_gatherpf<mode>): Ditto.
6810 (avx512pf_scatterpf<mode>): Ditto.
6811 (*avx512pf_scatterpf<mode>_mask): Ditto.
6812 (*avx512pf_scatterpf<mode>): Ditto.
6813 (GATHER_SCATTER_SF_MEM_MODE): New.
6814 (avx512pf_gatherpf<mode>df): Ditto.
6815 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6816 (*avx512pf_scatterpf<mode>df): Ditto.
6818 2014-01-27 Jakub Jelinek <jakub@redhat.com>
6821 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
6822 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
6825 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6827 * common/config/arm/arm-common.c
6828 (arm_rewrite_mcpu): Handle multiple names.
6830 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6832 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6834 * gimple-builder.h (create_gimple_tmp): Delete.
6836 2014-01-27 Christian Bruel <christian.bruel@st.com>
6838 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
6841 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
6843 * config/pa/pa.md (call): Generate indirect long calls to non-local
6844 functions when outputing 32-bit code.
6845 (call_value): Likewise except for special call to buggy powf function.
6847 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
6848 portable runtime and PIC indirect calls.
6849 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
6850 and PIC call sequences. Use ldo instead of blr to set return register
6851 in PIC call sequence.
6853 2014-01-25 Walter Lee <walt@tilera.com>
6855 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
6856 avoid clobbering a live register.
6858 2014-01-25 Walter Lee <walt@tilera.com>
6860 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
6861 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
6862 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
6863 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
6865 2014-01-25 Walter Lee <walt@tilera.com>
6867 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
6868 arguments on even registers.
6869 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
6871 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
6872 (BIGGEST_ALIGNMENT): Ditto.
6873 (BIGGEST_FIELD_ALIGNMENT): Ditto.
6875 2014-01-25 Walter Lee <walt@tilera.com>
6877 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
6878 insns before bundling.
6879 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
6881 2014-01-25 Walter Lee <walt@tilera.com>
6883 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
6884 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
6885 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
6887 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6889 * config/mips/constraints.md (kl): Delete.
6890 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
6891 define expands, using...
6892 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
6893 instructions for MIPS16.
6894 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
6895 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
6897 2014-01-25 Walter Lee <walt@tilera.com>
6899 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
6903 2014-01-25 Walter Lee <walt@tilera.com>
6905 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
6906 (TARGET_EXPAND_TO_RTL_HOOK): Define.
6908 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6910 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
6913 2014-01-25 Jakub Jelinek <jakub@redhat.com>
6915 * print-rtl.c (in_call_function_usage): New var.
6916 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
6917 EXPR_LIST mode as mode and not as reg note name.
6920 * cfgloopmanip.c (copy_loop_info): If
6921 loop->warned_aggressive_loop_optimizations, make sure
6922 the flag is set in target loop too.
6924 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
6926 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
6928 * builtins.def: Likewise.
6929 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
6930 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
6931 * ira.c (ira_setup_eliminable_regset): Likewise.
6932 * omp-low.c (gate_expand_omp): Likewise.
6933 (execute_lower_omp): Likewise.
6934 (diagnose_sb_0): Likewise.
6935 (gate_diagnose_omp_blocks): Likewise.
6936 (simd_clone_clauses_extract): Likewise.
6939 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6941 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
6942 correction for little endian...
6943 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
6946 2014-01-24 Jeff Law <law@redhat.com>
6948 PR tree-optimization/59919
6949 * tree-vrp.c (find_assert_locations_1): Do not register asserts
6950 for non-returning calls.
6952 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
6954 * common/config/aarch64/aarch64-common.c
6955 (aarch64_rewrite_mcpu): Handle multiple names.
6956 * config/aarch64/aarch64.h
6957 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6959 2014-01-24 Dodji Seketeli <dodji@redhat.com>
6961 * input.c (add_file_to_cache_tab): Handle the case where fopen
6964 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
6967 * config/i386/i386.md (pushsf splitter): Get stack adjustment
6968 from push operand if code of push isn't PRE_DEC.
6970 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6973 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
6974 -mquad-memory-atomic. Update -mquad-memory documentation to say
6975 it is only used for non-atomic loads/stores.
6977 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
6978 -mquad-memory or -mquad-memory-atomic switches.
6980 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
6981 -mquad-memory-atomic to ISA 2.07 support.
6983 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
6984 to separate support of normal quad word memory operations (ldq, stq)
6985 from the atomic quad word memory operations.
6987 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
6988 support to separate non-atomic quad word operations from atomic
6989 quad word operations. Disable non-atomic quad word operations in
6990 little endian mode so that we don't have to swap words after the
6991 load and before the store.
6992 (quad_load_store_p): Add comment about atomic quad word support.
6993 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
6994 options printed with -mdebug=reg.
6996 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
6997 -mquad-memory-atomic as the test for whether we have quad word
6998 atomic instructions.
6999 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
7000 or -mp8-vector are used, allow byte/half-word atomic operations.
7002 * config/rs6000/sync.md (load_lockedti): Insure that the address
7003 is a proper indexed or indirect address for the lqarx instruction.
7004 On little endian systems, swap the hi/lo registers after the lqarx
7006 (load_lockedpti): Use indexed_or_indirect_operand predicate to
7007 insure the address is valid for the lqarx instruction.
7008 (store_conditionalti): Insure that the address is a proper indexed
7009 or indirect address for the stqcrx. instruction. On little endian
7010 systems, swap the hi/lo registers before doing the stqcrx.
7012 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
7013 insure the address is valid for the stqcrx. instruction.
7015 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
7016 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
7017 type of quad memory support is available.
7019 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
7022 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
7023 there is a danger of looping.
7025 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
7027 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
7028 force flag_ira_loop_pressure if set via command line.
7030 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7032 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
7033 (ashr_simd): New builtin handling DI mode.
7034 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
7035 (aarch64_sshr_simddi): New match pattern.
7036 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
7037 (vshrd_n_s64): Likewise.
7038 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
7040 2014-01-23 Nick Clifton <nickc@redhat.com>
7042 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
7043 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
7044 favour of mcu specific scripts.
7045 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
7048 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7049 Alex Velenko <Alex.Velenko@arm.com>
7051 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
7052 (vaddv_s16): Likewise.
7053 (vaddv_s32): Likewise.
7054 (vaddv_u8): Likewise.
7055 (vaddv_u16): Likewise.
7056 (vaddv_u32): Likewise.
7057 (vaddvq_s8): Likewise.
7058 (vaddvq_s16): Likewise.
7059 (vaddvq_s32): Likewise.
7060 (vaddvq_s64): Likewise.
7061 (vaddvq_u8): Likewise.
7062 (vaddvq_u16): Likewise.
7063 (vaddvq_u32): Likewise.
7064 (vaddvq_u64): Likewise.
7065 (vaddv_f32): Likewise.
7066 (vaddvq_f32): Likewise.
7067 (vaddvq_f64): Likewise.
7068 (vmaxv_f32): Likewise.
7069 (vmaxv_s8): Likewise.
7070 (vmaxv_s16): Likewise.
7071 (vmaxv_s32): Likewise.
7072 (vmaxv_u8): Likewise.
7073 (vmaxv_u16): Likewise.
7074 (vmaxv_u32): Likewise.
7075 (vmaxvq_f32): Likewise.
7076 (vmaxvq_f64): Likewise.
7077 (vmaxvq_s8): Likewise.
7078 (vmaxvq_s16): Likewise.
7079 (vmaxvq_s32): Likewise.
7080 (vmaxvq_u8): Likewise.
7081 (vmaxvq_u16): Likewise.
7082 (vmaxvq_u32): Likewise.
7083 (vmaxnmv_f32): Likewise.
7084 (vmaxnmvq_f32): Likewise.
7085 (vmaxnmvq_f64): Likewise.
7086 (vminv_f32): Likewise.
7087 (vminv_s8): Likewise.
7088 (vminv_s16): Likewise.
7089 (vminv_s32): Likewise.
7090 (vminv_u8): Likewise.
7091 (vminv_u16): Likewise.
7092 (vminv_u32): Likewise.
7093 (vminvq_f32): Likewise.
7094 (vminvq_f64): Likewise.
7095 (vminvq_s8): Likewise.
7096 (vminvq_s16): Likewise.
7097 (vminvq_s32): Likewise.
7098 (vminvq_u8): Likewise.
7099 (vminvq_u16): Likewise.
7100 (vminvq_u32): Likewise.
7101 (vminnmv_f32): Likewise.
7102 (vminnmvq_f32): Likewise.
7103 (vminnmvq_f64): Likewise.
7105 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7107 * config/aarch64/aarch64-simd.md
7108 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
7109 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
7110 (*aarch64_mul3_elt<mode>): Likewise.
7111 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
7112 (*aarch64_mul3_elt_to_64v2df): Likewise.
7113 (*aarch64_mla_elt<mode>): Likewise.
7114 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7115 (*aarch64_mls_elt<mode>): Likewise.
7116 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7117 (*aarch64_fma4_elt<mode>): Likewise.
7118 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7119 (*aarch64_fma4_elt_to_64v2df): Likewise.
7120 (*aarch64_fnma4_elt<mode>): Likewise.
7121 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7122 (*aarch64_fnma4_elt_to_64v2df): Likewise.
7123 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7124 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
7125 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7126 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7127 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
7128 (aarch64_sqdmull_lane<mode>_internal): Likewise.
7129 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
7131 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
7133 * config/aarch64/aarch64-simd.md
7134 (aarch64_be_checked_get_lane<mode>): New define_expand.
7135 * config/aarch64/aarch64-simd-builtins.def
7136 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
7137 New builtin definition.
7138 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
7139 Use new safe be builtin.
7141 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7143 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
7145 (aarch64_be_st1<mode>): Likewise.
7146 (aarch_ld1<VALL:mode>): Define_expand modified.
7147 (aarch_st1<VALL:mode>): Likewise.
7148 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
7149 (UNSPEC_ST1): Likewise.
7151 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
7153 * config/microblaze/microblaze.md: Add trap insn and attribute
7155 2014-01-23 Dodji Seketeli <dodji@redhat.com>
7157 PR preprocessor/58580
7158 * input.h (location_get_source_line): Take an additional line_size
7160 (void diagnostics_file_cache_fini): Declare new function.
7161 * input.c (struct fcache): New type.
7162 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7163 New static constants.
7164 (diagnostic_file_cache_init, total_lines_num)
7165 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7166 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7167 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7168 (get_next_line, read_next_line, goto_next_line, read_line_num):
7169 New static function definitions.
7170 (diagnostic_file_cache_fini): New function.
7171 (location_get_source_line): Take an additional output line_len
7172 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7174 * diagnostic.c (diagnostic_finish): Call
7175 diagnostic_file_cache_fini.
7176 (adjust_line): Take an additional input parameter for the length
7177 of the line, rather than calculating it with strlen.
7178 (diagnostic_show_locus): Adjust the use of
7179 location_get_source_line and adjust_line with respect to their new
7180 signature. While displaying a line now, do not stop at the first
7181 null byte. Rather, display the zero byte as a space and keep
7182 going until we reach the size of the line.
7183 * Makefile.in: Add vec.o to OBJS-libcommon
7185 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7186 Ilya Tocar <ilya.tocar@intel.com>
7188 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7189 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7190 (__builtin_ia32_kmov16): Ditto.
7191 * config/i386/i386.md (UNSPEC_KMOV): New.
7194 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7196 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7197 (_mm512_storeu_si512): Ditto.
7199 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7202 * rtl.h (get_referenced_operands): Declare.
7203 * recog.c (get_referenced_operands): New function.
7204 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7205 operands have been referenced when recording LO_SUM references.
7207 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7209 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7211 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7213 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7214 Enable for generic and recent AMD targets.
7216 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7218 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7219 ARG_SIZE note when adjustment was eliminated.
7221 2014-01-22 Jeff Law <law@redhat.com>
7223 PR tree-optimization/59597
7224 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7225 in file. Accept new argument REGISTERING and use it to modify
7226 dump output appropriately.
7227 (register_jump_thread): Corresponding changes.
7228 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7229 thread paths involving joiner blocks. Add code to dump cancelled
7230 jump threading paths.
7232 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7234 PR rtl-optimization/59477
7235 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7236 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7238 2014-01-22 Tom Tromey <tromey@redhat.com>
7240 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7242 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7244 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7246 PR rtl-optimization/59896
7247 * lra-constraints.c (process_alt_operands): Check unused note for
7248 matched operands of insn with no output reloads.
7250 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7252 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7253 (mips_move_from_gpr_cost): Likewise.
7255 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7257 PR rtl-optimization/59858
7258 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7259 ira_class_hard_regs_num.
7260 (process_alt_operands): Increase reject for dying matched operand.
7262 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7265 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7266 smaller than size, perform several stores or loads and stores
7267 at dst + count - size to store or copy all of size bytes, rather
7268 than just last modesize bytes.
7270 2014-01-20 DJ Delorie <dj@redhat.com>
7272 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7273 that CLOBBERs are REGs before propogating their values.
7275 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7278 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7279 (cgraph_inline_failed_type): New function.
7280 * cgraph.h (DEFCIFCODE): Add type.
7281 (cgraph_inline_failed_type_t): New enum.
7282 (cgraph_inline_failed_type): New prototype.
7283 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7284 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7285 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7286 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7287 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7288 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7289 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7290 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7291 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7292 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7293 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7294 OPTIMIZATION_MISMATCH.
7295 * tree-inline.c (expand_call_inline): Emit errors during
7296 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7298 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7301 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7302 mode attribute in insn output.
7304 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7306 * output.h (output_constant): Delete.
7307 * varasm.c (output_constant): Make private.
7309 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7311 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7313 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7316 * tree.h (fold_builtin_strcat): New prototype.
7317 * builtins.c (fold_builtin_strcat): No longer static. Add len
7318 argument, if non-NULL, don't call c_strlen. Optimize
7319 directly into __builtin_memcpy instead of __builtin_strcpy.
7320 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7321 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7323 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7325 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7326 for SImode_address_operand operands, having only a REG argument.
7328 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7330 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7331 loader name using mbig-endian.
7332 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7334 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7336 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7340 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7342 * config/aarch64/aarch64-protos.h
7343 (aarch64_cannot_change_mode_class_ptr): Declare.
7344 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7345 aarch64_cannot_change_mode_class_ptr): New.
7346 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7347 backend hook aarch64_cannot_change_mode_class.
7349 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7351 * common/config/aarch64/aarch64-common.c
7352 (aarch64_handle_option): Don't handle any option order logic here.
7353 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7354 selected_cpu, warn on architecture version mismatch.
7355 (aarch64_override_options): Fix parsing order for option strings.
7357 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7358 Iain Sandoe <iain@codesourcery.com>
7361 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7362 warning. Amend comment to reflect current functionality.
7364 2014-01-20 Richard Biener <rguenther@suse.de>
7367 * builtins.c (fold_builtin_strcat): Remove case better handled
7368 by tree-ssa-strlen.c.
7370 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7372 * config/aarch64/aarch64.opt
7373 (mcpu, march, mtune): Make case-insensitive.
7375 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7378 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7379 if operands[1] is a REG or ZERO_EXTEND of a REG.
7381 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7383 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7385 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7387 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7388 long non-pic millicode calls.
7390 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7392 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7394 2014-01-19 Kito Cheng <kito@0xlab.org>
7396 * builtins.c (expand_movstr): Check movstr expand done or fail.
7398 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7399 H.J. Lu <hongjiu.lu@intel.com>
7402 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7403 to DImode for zero-extended addresses.
7405 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7407 PR rtl-optimization/57763
7408 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7409 on the new indirect jump_insn and increment LABEL_NUSES (label).
7411 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7415 * config.gcc (x86_archs): New variable.
7416 (x86_64_archs): Likewise.
7417 (x86_cpus): Likewise.
7418 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7419 --with-arch/--with-cpu= options.
7420 Support --with-arch=/--with-cpu={nehalem,westmere,
7421 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7423 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7425 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7426 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7428 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7430 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7432 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7435 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7436 clear cpp_get_options (parse_in)->warn_unused_macros for
7437 ix86_target_macros_internal with cpp_define.
7439 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7441 * jump.c (delete_related_insns): Keep (use (insn))s.
7442 * reorg.c (redundant_insn): Check for barriers too.
7444 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7446 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7448 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7450 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7451 call to $$dyncall when TARGET_LONG_CALLS is true.
7453 2014-01-17 Jeff Law <law@redhat.com>
7455 * ree.c (combine_set_extension): Temporarily disable test for
7456 changing number of hard registers.
7458 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7461 * ipa-inline-analysis.c (inline_free_summary):
7462 Do not free summary of aliases.
7464 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7467 * gimplify.c (gimplify_expr): Use create_tmp_var
7468 instead of create_tmp_var_raw. If cond doesn't have
7469 integral type, don't add the IFN_ANNOTATE builtin at all.
7471 2014-01-17 Martin Jambor <mjambor@suse.cz>
7474 * ipa-cp.c (prev_edge_clone): New variable.
7475 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7476 Also resize prev_edge_clone vector.
7477 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7478 (ipcp_edge_removal_hook): New function.
7479 (ipcp_driver): Register ipcp_edge_removal_hook.
7481 2014-01-17 Andrew Pinski <apinski@cavium.com>
7482 Steve Ellcey <sellcey@mips.com>
7485 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7488 2014-01-17 Jeff Law <law@redhat.com>
7491 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7492 so that pass_ccp runs first.
7494 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7496 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7497 (ix86_adjust_cost): Use !TARGET_XXX.
7498 (do_reorder_for_imul): Likewise.
7499 (swap_top_of_ready_list): Likewise.
7500 (ix86_sched_reorder): Likewise.
7502 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7504 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7505 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7506 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7507 (intel_memset): New. Duplicate slm_memset.
7508 (intel_cost): New. Duplicate slm_cost.
7509 (m_INTEL): New macro.
7510 (processor_target_table): Add "intel".
7511 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7512 with PROCESSOR_INTEL for "intel".
7513 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7514 PROCESSOR_SILVERMONT.
7515 (ix86_issue_rate): Likewise.
7516 (ix86_adjust_cost): Likewise.
7517 (ia32_multipass_dfa_lookahead): Likewise.
7518 (swap_top_of_ready_list): Likewise.
7519 (ix86_sched_reorder): Likewise.
7520 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7521 instead of TARGET_OPT_AGU.
7522 * config/i386/i386.h (TARGET_INTEL): New.
7523 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7524 (processor_type): Add PROCESSOR_INTEL.
7525 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7526 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7528 2014-01-17 Marek Polacek <polacek@redhat.com>
7531 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7534 2014-01-17 Richard Biener <rguenther@suse.de>
7536 PR tree-optimization/46590
7537 * opts.c (default_options_table): Add entries for
7538 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7539 all enabled at -O1 but not for -Og.
7540 * common.opt (fbranch-count-reg): Remove Init(1).
7541 (fmove-loop-invariants): Likewise.
7542 (ftree-pta): Likewise.
7544 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7546 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7547 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7548 decls to at least the GCC 4.8 used alignments.
7551 * tree-nested.c (convert_nonlocal_reference_stmt,
7552 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7553 of GIMPLE_BIND stmts, adjust associated decls.
7555 2014-01-17 Richard Biener <rguenther@suse.de>
7557 PR tree-optimization/46590
7558 * vec.h (vec<>::bseach): New member function implementing
7559 binary search according to C89 bsearch.
7560 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7561 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7562 bitmap pointer again. Make accesses_in_loop a flat array.
7563 (mem_ref_obstack): New global.
7564 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7565 (mark_ref_stored): Likewise.
7566 (ref_indep_loop_p_2): Likewise.
7567 (set_ref_stored_in_loop): New helper function.
7568 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7569 (memref_free): Adjust.
7570 (record_mem_ref_loc): Simplify.
7571 (gather_mem_refs_stmt): Adjust.
7572 (sort_locs_in_loop_postorder_cmp): New function.
7573 (analyze_memory_references): Sort accesses_in_loop after
7574 loop postorder number.
7575 (find_ref_loc_in_loop_cmp): New function.
7576 (for_all_locs_in_loop): Find relevant cluster of locs in
7577 accesses_in_loop and iterate without recursion.
7578 (execute_sm): Avoid uninit warning.
7579 (struct ref_always_accessed): Simplify.
7580 (ref_always_accessed::operator ()): Likewise.
7581 (ref_always_accessed_p): Likewise.
7582 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7583 loop postorder numbers here.
7584 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7587 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7590 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7591 on decls for which assemble_alias has been called.
7593 2014-01-17 Nick Clifton <nickc@redhat.com>
7595 * config/msp430/msp430.opt: (mcpu): New option.
7596 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7597 (msp430_option_override): Parse target_cpu. If the MCU name
7598 matches a generic string, clear target_mcu.
7599 (msp430_attr): Allow numeric interrupt values up to 63.
7600 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7601 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7603 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7605 * config/msp430/msp430.md (popm): Use %J rather than %I.
7606 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7607 (addhi_cy_i): Use immediate_operand for operand 2.
7608 * doc/invoke.texi: Document -mcpu option.
7610 2014-01-17 Richard Biener <rguenther@suse.de>
7612 PR rtl-optimization/38518
7613 * df.h (df_analyze_loop): Declare.
7614 * df-core.c: Include cfgloop.h.
7615 (df_analyze_1): Split out main part of df_analyze.
7616 (df_analyze): Adjust.
7617 (loop_inverted_post_order_compute): New function.
7618 (loop_post_order_compute): Likewise.
7619 (df_analyze_loop): New function avoiding whole-function
7621 * loop-invariant.c (find_defs): Use df_analyze_loop.
7622 (find_invariants): Adjust.
7623 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7625 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7627 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7628 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7630 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7632 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7633 traversal when removing references.
7635 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7638 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7640 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7643 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7644 pushing a reload for an autoinc when we had previously reloaded an
7645 inner part of the address.
7647 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7649 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7651 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7652 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7653 when not giving up or versioning for alias only because of
7655 (vect_analyze_data_ref_dependences): Set to true.
7656 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7658 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7659 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7663 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7666 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7667 operand 0 predicate for gathers, use a new pseudo as subtarget.
7669 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7672 * lra-constraints.c (process_alt_operands): Add printing debug info.
7673 Check absence of input/output reloads for matched operands too.
7675 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7677 PR rtl-optimization/59835
7678 * ira.c (ira_init_register_move_cost): Increase cost for
7681 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7683 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7685 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7688 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7689 non-register objects. Use gen_(high/low)part more consistently.
7692 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7695 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7696 endian support, remove tests for WORDS_BIG_ENDIAN.
7697 (p8_mfvsrd_3_<mode>): Likewise.
7698 (reload_gpr_from_vsx<mode>): Likewise.
7699 (reload_gpr_from_vsxsf): Likewise.
7700 (p8_mfvsrd_4_disf): Likewise.
7702 2014-01-16 Richard Biener <rguenther@suse.de>
7704 PR rtl-optimization/46590
7705 * lcm.c (compute_antinout_edge): Use postorder iteration.
7706 (compute_laterin): Use inverted postorder iteration.
7708 2014-01-16 Nick Clifton <nickc@redhat.com>
7711 * varasm.c (output_constant): Return the number of bytes actually
7713 (output_constructor_array_range): Update the field size with the
7714 number of bytes emitted by output_constant.
7715 (output_constructor_regular_field): Likewise. Also do not
7716 complain if the total number of bytes emitted is now greater
7717 than the expected fieldpos.
7718 * output.h (output_constant): Update prototype and descriptive comment.
7720 2014-01-16 Marek Polacek <polacek@redhat.com>
7723 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7724 it is error_mark_node.
7726 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7728 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7729 VALID_AVX256_REG_OR_OI_MODE.
7731 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7733 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7734 current procedure should be profiled.
7736 2014-01-15 Andrew Pinski <apinski@cavium.com>
7738 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7739 of moving from/to the STACK_REG register class.
7741 2014-01-15 Richard Henderson <rth@redhat.com>
7744 * reginfo.c (global_regs_decl): Globalize.
7745 * rtl.h (global_regs_decl): Declare.
7746 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7747 reserved via global_regs.
7749 2014-01-15 Teresa Johnson <tejohnson@google.com>
7751 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7753 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7755 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7756 and vmulosh rather than call gen_vec_widen_smult_*.
7757 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7758 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7759 (vec_widen_smult_even_v16qi): Likewise.
7760 (vec_widen_umult_even_v8hi): Likewise.
7761 (vec_widen_smult_even_v8hi): Likewise.
7762 (vec_widen_umult_odd_v16qi): Likewise.
7763 (vec_widen_smult_odd_v16qi): Likewise.
7764 (vec_widen_umult_odd_v8hi): Likewise.
7765 (vec_widen_smult_odd_v8hi): Likewise.
7766 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7767 vmuloub rather than call gen_vec_widen_umult_*.
7768 (vec_widen_umult_lo_v16qi): Likewise.
7769 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7770 vmulosb rather than call gen_vec_widen_smult_*.
7771 (vec_widen_smult_lo_v16qi): Likewise.
7772 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7773 rather than call gen_vec_widen_umult_*.
7774 (vec_widen_umult_lo_v8hi): Likewise.
7775 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7776 rather than call gen_vec_widen_smult_*.
7777 (vec_widen_smult_lo_v8hi): Likewise.
7779 2014-01-15 Jeff Law <law@redhat.com>
7781 PR tree-optimization/59747
7782 * ree.c (find_and_remove_re): Properly handle case where a second
7783 eliminated extension requires widening a copy created for elimination
7784 of a prior extension.
7785 (combine_set_extension): Ensure that the number of hard regs needed
7786 for a destination register does not change when we widen it.
7788 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7790 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7791 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7792 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7793 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7794 (avr-*-rtems*): Likewise.
7795 (bfin*-rtems*): Likewise.
7796 (moxie-*-rtems*): Likewise.
7797 (h8300-*-rtems*): Likewise.
7798 (i[34567]86-*-rtems*): Likewise.
7799 (lm32-*-rtems*): Likewise.
7800 (m32r-*-rtems*): Likewise.
7801 (m68k-*-rtems*): Likewise.
7802 (microblaze*-*-rtems*): Likewise.
7803 (mips*-*-rtems*): Likewise.
7804 (powerpc-*-rtems*): Likewise.
7805 (sh-*-rtems*): Likewise.
7806 (sparc-*-rtems*): Likewise.
7807 (sparc64-*-rtems*): Likewise.
7808 (v850-*-rtems*): Likewise.
7809 (m32c-*-rtems*): Likewise.
7811 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7813 PR rtl-optimization/59511
7814 * ira.c (ira_init_register_move_cost): Use memory costs for some
7815 cases of register move cost calculations.
7816 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7817 instead of BB frequency.
7818 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7819 * lra-assigns.c (find_hard_regno_for): Ditto.
7821 2014-01-15 Richard Biener <rguenther@suse.de>
7823 PR tree-optimization/59822
7824 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
7825 (vectorizable_load): Use it to hoist defs of uses of invariant
7826 loads out of the loop.
7828 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7829 Kugan Vivekanandarajah <kuganv@linaro.org>
7832 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
7835 2014-01-15 Richard Biener <rguenther@suse.de>
7837 PR rtl-optimization/59802
7838 * lcm.c (compute_available): Use inverted postorder to seed
7839 the initial worklist.
7841 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7844 * config/s390/s390.c (s390_preferred_reload_class): Don't return
7845 ADDR_REGS for invalid symrefs in non-PIC code.
7847 2014-01-15 Jakub Jelinek <jakub@redhat.com>
7850 * builtins.c (determine_block_size): Initialize *probable_max_size
7851 even if len_rtx is CONST_INT.
7853 2014-01-14 Andrew Pinski <apinski@cavium.com>
7855 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
7856 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
7857 (cortexa53_tunings): Likewise.
7858 (aarch64_sched_issue_rate): New function.
7859 (TARGET_SCHED_ISSUE_RATE): Define.
7861 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7863 * ira-costs.c (find_costs_and_classes): Add missed
7864 ira_init_register_move_cost_if_necessary.
7866 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7869 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
7871 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
7874 * config/i386/i386.c (type_natural_mode): Add a bool parameter
7875 to indicate if type is used for function return value. Warn ABI
7876 change if the vector mode isn't available for function return value.
7877 (ix86_function_arg_advance): Pass false to type_natural_mode.
7878 (ix86_function_arg): Likewise.
7879 (ix86_gimplify_va_arg): Likewise.
7880 (function_arg_32): Don't warn ABI change.
7881 (ix86_function_value): Pass true to type_natural_mode.
7882 (ix86_return_in_memory): Likewise.
7883 (ix86_struct_value_rtx): Removed.
7884 (TARGET_STRUCT_VALUE_RTX): Likewise.
7886 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7888 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
7889 converting a conditional jump into a conditional return.
7891 2014-01-14 Richard Biener <rguenther@suse.de>
7893 PR tree-optimization/58921
7894 PR tree-optimization/59006
7895 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
7896 hoisting invariant stmts.
7897 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
7898 invariant loads on the preheader edge if possible.
7900 2014-01-14 Joey Ye <joey.ye@arm.com>
7902 * doc/plugin.texi (Building GCC plugins): Update to C++.
7904 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
7906 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
7907 (_mm_rcp28_round_ss): Ditto.
7908 (_mm_rsqrt28_round_sd): Ditto.
7909 (_mm_rsqrt28_round_ss): Ditto.
7910 (_mm_rcp28_sd): Ditto.
7911 (_mm_rcp28_ss): Ditto.
7912 (_mm_rsqrt28_sd): Ditto.
7913 (_mm_rsqrt28_ss): Ditto.
7914 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
7915 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
7916 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
7917 (IX86_BUILTIN_RCP28SD): Ditto.
7918 (IX86_BUILTIN_RCP28SS): Ditto.
7919 (IX86_BUILTIN_RSQRT28SD): Ditto.
7920 (IX86_BUILTIN_RSQRT28SS): Ditto.
7921 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
7922 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
7923 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
7924 (ix86_expand_special_args_builtin): Expand new FTYPE.
7925 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
7926 (srcp14<mode>): Make insn unary.
7927 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
7928 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
7929 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
7930 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
7931 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
7932 Fix rounding: make it SAE only.
7933 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7935 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7937 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
7938 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
7939 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
7940 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
7941 (round_saeonly_mask_scalar_operand4): Ditto.
7942 (round_saeonly_mask_scalar_op3): Ditto.
7943 (round_saeonly_mask_scalar_op4): Ditto.
7945 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7947 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7948 Implement -maltivec=be for vec_insert and vec_extract.
7950 2014-01-10 DJ Delorie <dj@redhat.com>
7952 * config/msp430/msp430.md (call_internal): Don't allow memory
7953 references with SP as the base register.
7954 (call_value_internal): Likewise.
7955 * config/msp430/constraints.md (Yc): New. For memory references
7956 that don't use SP as a base register.
7958 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
7959 "an integer without a # prefix"
7960 * config/msp430/msp430.md (epilogue_helper): Use it.
7962 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7965 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
7966 AVX512F gather builtins.
7967 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
7968 on gather decls with INTEGER_TYPE masktype.
7969 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
7970 directly into the builtin rather than hoisting it before loop.
7972 PR tree-optimization/59387
7973 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
7974 (scev_const_prop): If folded_casts and type has undefined overflow,
7975 use force_gimple_operand instead of force_gimple_operand_gsi and
7976 for each added stmt if it is assign with
7977 arith_code_with_undefined_signed_overflow, call
7978 rewrite_to_defined_overflow.
7979 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
7980 gimple-fold.h instead.
7981 (arith_code_with_undefined_signed_overflow,
7982 rewrite_to_defined_overflow): Moved to ...
7983 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
7984 rewrite_to_defined_overflow): ... here. No longer static.
7985 Include gimplify-me.h.
7986 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
7987 rewrite_to_defined_overflow): New prototypes.
7989 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7991 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
7993 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
7995 * builtins.c (get_object_alignment_2): Minor tweak.
7996 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
7998 2014-01-13 Christian Bruel <christian.bruel@st.com>
8000 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
8001 optimized non constant lengths.
8003 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8006 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
8007 load as __atomic_load_N if possible.
8009 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8011 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
8013 (rs6000_expand_builtin): Adjust call.
8015 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8018 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
8019 * config/rs6000/rs6000.c: Include target-globals.h.
8020 (rs6000_set_current_function): Instead of doing target_reinit
8021 unconditionally, use save_target_globals_default_opts and
8022 restore_target_globals.
8024 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
8026 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
8027 (rs6000_expand_builtin): Handle mffs and mtfsf.
8028 (rs6000_init_builtins): Define mffs and mtfsf.
8029 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
8030 (rs6000_mffs): New pattern.
8031 (rs6000_mtfsf): New pattern.
8033 2014-01-11 Bin Cheng <bin.cheng@arm.com>
8035 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
8036 Start narrowing with START. Apply candidate-use pair
8037 and check overall cost in narrowing.
8038 (iv_ca_prune): Pass new argument.
8040 2014-01-10 Jeff Law <law@redhat.com>
8043 * ree.c (combine_reaching_defs): Ensure the defining statement
8044 occurs before the extension when optimizing extensions with
8045 different source and destination hard registers.
8047 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8050 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
8051 vtables into the type inheritance graph.
8053 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8055 PR rtl-optimization/59754
8056 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
8057 modes in the REGNO != REGNO case.
8059 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8061 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
8063 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8065 PR tree-optimization/59745
8066 * tree-predcom.c (tree_predictive_commoning_loop): Call
8067 free_affine_expand_cache if giving up because components is NULL.
8069 * target-globals.c (save_target_globals): Allocate < 4KB structs using
8070 GC in payload of target_globals struct instead of allocating them on
8071 the heap and the larger structs separately using GC.
8072 * target-globals.h (struct target_globals): Make regs, hard_regs,
8073 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
8074 of GTY((skip)) and change type to void *.
8075 (reset_target_globals): Cast loads from those fields to corresponding
8078 2014-01-10 Steve Ellcey <sellcey@mips.com>
8081 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
8082 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
8083 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
8085 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
8088 * aarch64-modes.def (CC_Zmode): New flags mode.
8089 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
8090 represents an equality.
8091 (aarch64_get_condition_code): Handle CC_Zmode.
8092 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
8094 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8096 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
8097 extraction in good case.
8099 2014-01-10 Richard Biener <rguenther@suse.de>
8101 PR tree-optimization/59374
8102 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
8103 checking after SLP discovery. Mark stmts not participating
8104 in any SLP instance properly.
8106 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8108 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
8109 when handling a SET rtx.
8111 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8113 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
8114 (cortex-a57): Likewise.
8115 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
8117 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8119 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
8120 non-iwmmxt builtins.
8122 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8126 * ipa-devirt.c record_target_from_binfo): Take as argument
8127 stack of binfos and lookup matching one for virtual inheritance.
8128 (possible_polymorphic_call_targets_1): Update.
8130 2014-01-10 Huacai Chen <chenhc@lemote.com>
8132 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
8133 kernel strings for Loongson-2E/2F/3A.
8135 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8138 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
8139 is_gimple_call before calling gimple_call_internal_p.
8141 2014-01-09 Steve Ellcey <sellcey@mips.com>
8143 * Makefile.in (TREE_FLOW_H): Remove.
8144 (TREE_SSA_H): Add file names from tree-flow.h.
8145 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
8146 * tree.h: Remove tree-flow.h reference.
8147 * hash-table.h: Remove tree-flow.h reference.
8148 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
8149 reference with tree-ssa-loop.h.
8151 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8153 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8154 default element-order behavior for -maltivec.
8155 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8156 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8157 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8158 when targeting big endian, at least for now.
8159 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8161 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8164 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8165 var satisfies use_register_for_decl, just take into account type
8166 alignment, rather than decl alignment.
8168 PR tree-optimization/59622
8169 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8170 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8171 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8172 Don't devirtualize for inplace at all. For targets.length () == 1,
8173 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8175 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8177 * config/i386/i386.md (cpu): Remove the unused btver1.
8179 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8181 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8183 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8186 * tree-core.h (struct target_globals): New forward declaration.
8187 (struct tree_target_option): Add globals field.
8188 * tree.h (TREE_TARGET_GLOBALS): Define.
8189 (prepare_target_option_nodes_for_pch): New prototype.
8190 * target-globals.h (struct target_globals): Define even if
8192 * tree.c (prepare_target_option_node_for_pch,
8193 prepare_target_option_nodes_for_pch): New functions.
8194 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8195 * config/i386/i386.c: Include target-globals.h.
8196 (ix86_set_current_function): Instead of doing target_reinit
8197 unconditionally, use save_target_globals_default_opts and
8198 restore_target_globals.
8200 2014-01-09 Richard Biener <rguenther@suse.de>
8202 PR tree-optimization/59715
8203 * tree-cfg.h (split_critical_edges): Declare.
8204 * tree-cfg.c (split_critical_edges): Export.
8205 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8207 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8209 * cfgexpand.c (expand_stack_vars): Optionally disable
8210 asan stack protection.
8211 (expand_used_vars): Likewise.
8212 (partition_stack_vars): Likewise.
8213 * asan.c (asan_emit_stack_protection): Optionally disable
8214 after return stack usage.
8215 (instrument_derefs): Optionally disable memory access instrumentation.
8216 (instrument_builtin_call): Likewise.
8217 (instrument_strlen_call): Likewise.
8218 (asan_protect_global): Optionally disable global variables protection.
8219 * doc/invoke.texi: Added doc for new options.
8220 * params.def: Added new options.
8221 * params.h: Likewise.
8223 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8225 PR rtl-optimization/59724
8226 * ifcvt.c (cond_exec_process_if_block): Don't call
8227 flow_find_head_matching_sequence with 0 longest_match.
8228 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8229 non-active insns if !stop_after.
8230 (try_head_merge_bb): Revert 2014-01-07 changes.
8232 2014-01-08 Jeff Law <law@redhat.com>
8234 * ree.c (get_sub_rtx): New function, extracted from...
8235 (merge_def_and_ext): Here.
8236 (combine_reaching_defs): Use get_sub_rtx.
8238 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8240 * cgraph.h (varpool_variable_node): Do not choke on null node.
8242 2014-01-08 Catherine Moore <clm@codesourcery.com>
8244 * config/mips/mips.md (simple_return): Attempt to use JRC
8246 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8248 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8250 PR rtl-optimization/59137
8251 * reorg.c (steal_delay_list_from_target): Call update_block for
8253 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8255 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8257 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8258 two duplicate entries.
8260 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8263 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8265 * config/mips/mips.c (mips_truncated_op_cost): New function.
8266 (mips_rtx_costs): Adjust test for BADDU.
8267 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8269 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8271 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8272 (*baddu_si): ...this new pattern.
8274 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8277 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8279 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8282 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8284 (expand_expr, expand_normal): Adjust.
8285 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8286 inner_reference_p. Use inner_reference_p to expand inner references.
8287 (store_expr): Adjust.
8288 * cfgexpand.c (expand_call_stmt): Adjust.
8290 2014-01-08 Rong Xu <xur@google.com>
8292 * gcov-io.c (gcov_var): Move from gcov-io.h.
8293 (gcov_position): Ditto.
8294 (gcov_is_error): Ditto.
8295 (gcov_rewrite): Ditto.
8296 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8297 only part to libgcc/libgcov.h.
8299 2014-01-08 Marek Polacek <polacek@redhat.com>
8302 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8304 2014-01-08 Marek Polacek <polacek@redhat.com>
8307 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8309 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8311 PR rtl-optimization/59649
8312 * stor-layout.c (get_mode_bounds): For BImode return
8313 0 and STORE_FLAG_VALUE.
8315 2014-01-08 Richard Biener <rguenther@suse.de>
8318 * gimple.h (is_gimple_builtin_call): Remove.
8319 (gimple_builtin_call_types_compatible_p): New.
8320 (gimple_call_builtin_p): New overload.
8321 * gimple.c (is_gimple_builtin_call): Remove.
8322 (validate_call): Rename to ...
8323 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8325 (validate_type): New static function.
8326 (gimple_call_builtin_p): New overload and adjust.
8327 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8328 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8329 (gimple_fold_stmt_to_constant_1): Likewise.
8330 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8332 2014-01-08 Richard Biener <rguenther@suse.de>
8335 * gimplify.c (gimplify_expr): Gimplify register-register type
8336 VIEW_CONVERT_EXPRs to separate stmts.
8338 2014-01-07 Jeff Law <law@redhat.com>
8341 * ree.c (combine_set_extension): Handle case where source
8342 and destination registers in an extension insn are different.
8343 (combine_reaching_defs): Allow source and destination registers
8344 in extension to be different under limited circumstances.
8345 (add_removable_extension): Remove restriction that the
8346 source and destination registers in the extension are the same.
8347 (find_and_remove_re): Emit a copy from the extension's
8348 destination to its source after the defining insn if
8349 the source and destination registers are different.
8352 * ifcvt.c (merge_if_block): If we are merging a block with more than
8353 one successor with a block with no successors, remove any BARRIER
8354 after the second block.
8356 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8358 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8360 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8363 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8364 for 14-bit register offsets when INT14_OK_STRICT is false.
8366 2014-01-07 Roland Stigge <stigge@antcom.de>
8367 Michael Meissner <meissner@linux.vnet.ibm.com>
8370 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8371 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8373 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8375 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8378 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8380 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8381 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8382 rtx is const0_rtx or not.
8384 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8387 * target-globals.c (save_target_globals): Remove this_fn_optab
8389 * toplev.c: Include optabs.h.
8390 (target_reinit): Temporarily restore the global options if another
8391 set of options are in force.
8393 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8395 PR rtl-optimization/58668
8396 * cfgcleanup.c (flow_find_cross_jump): Don't count
8397 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8398 to determine what is counted.
8399 (flow_find_head_matching_sequence): Use active_insn_p to determine
8401 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8403 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8404 determine what is counted.
8406 PR tree-optimization/59643
8407 * tree-predcom.c (split_data_refs_to_components): If one dr is
8408 read and one write, determine_offset fails and the write isn't
8409 in the bad component, just put the read into the bad component.
8411 2014-01-07 Mike Stump <mikestump@comcast.net>
8412 Jakub Jelinek <jakub@redhat.com>
8415 * tree-core.h (struct tree_optimization_option): Change optabs
8416 type from unsigned char * to void *.
8417 * optabs.c (init_tree_optimization_optabs): Adjust
8418 TREE_OPTIMIZATION_OPTABS initialization.
8420 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8423 * config/i386/i386.h (struct machine_function): Add
8424 no_drap_save_restore field.
8425 * config/i386/i386.c (ix86_save_reg): Use
8426 !cfun->machine->no_drap_save_restore instead of
8427 crtl->stack_realign_needed.
8428 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8429 this function clears frame_pointer_needed. Set
8430 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8431 and DRAP reg is needed.
8433 2014-01-06 Marek Polacek <polacek@redhat.com>
8436 * doc/implement-c.texi: Mention that other integer types are
8437 permitted as bit-field types in strictly conforming mode.
8439 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8441 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8444 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8446 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8448 2014-01-06 Martin Jambor <mjambor@suse.cz>
8451 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8453 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8455 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8459 * var-tracking.c (add_stores): Preserve the value of the source even if
8460 we don't record the store.
8462 2014-01-06 Terry Guo <terry.guo@arm.com>
8464 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8466 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8469 * config/darwin.c (darwin_function_section): Adjust return values to
8470 correspond to optimisation changes made in r206070.
8472 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8474 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8475 from prefetch_block tune setting.
8476 (nocona_cost): Correct size of prefetch block to 64.
8478 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8480 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8481 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8482 used to save the static chain register in the computation of the offset
8483 from which the FP registers need to be restored.
8485 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8487 PR tree-optimization/59519
8488 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8489 ICE if get_current_def (current_new_name) is already non-NULL, as long
8490 as it is a phi result of some other phi in *new_exit_bb that has
8493 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8494 or vmovdqu* for misaligned_operand.
8495 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8496 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8497 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8498 aligned_mem for AVX512F masked aligned load and store builtins and for
8501 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8503 PR tree-optimization/59651
8504 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8505 Address range for negative step should be added by TYPE_SIZE_UNIT.
8507 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8509 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8510 overlapping registers also for registers other than the stack pointer.
8512 2014-01-03 Marek Polacek <polacek@redhat.com>
8515 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8518 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8521 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8524 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8525 (push splitter): Use <P:MODE_SIZE> instead of
8526 GET_MODE_SIZE (<P:MODE>mode).
8527 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8528 (mov -1, reg peephole2): Likewise.
8529 * config/i386/sse.md (*mov<mode>_internal,
8530 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8531 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8532 *<code><mode>3, *andnot<mode>3<mask_name>,
8533 <mask_codefor><code><mode>3<mask_name>): Likewise.
8534 * config/i386/subst.md (mask_mode512bit_condition,
8535 sd_mask_mode512bit_condition): Likewise.
8537 2014-01-02 Xinliang David Li <davidxl@google.com>
8539 PR tree-optimization/59303
8540 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8541 (dump_predicates): Better output format.
8542 (pred_equal_p): New function.
8543 (is_neq_relop_p): Ditto.
8544 (is_neq_zero_form_p): Ditto.
8545 (pred_expr_equal_p): Ditto.
8546 (pred_neg_p): Ditto.
8547 (simplify_pred): Ditto.
8548 (simplify_preds_2): Ditto.
8549 (simplify_preds_3): Ditto.
8550 (simplify_preds_4): Ditto.
8551 (simplify_preds): Ditto.
8553 (push_to_worklist): Ditto.
8554 (get_pred_info_from_cmp): Ditto.
8555 (is_degenerated_phi): Ditto.
8556 (normalize_one_pred_1): Ditto.
8557 (normalize_one_pred): Ditto.
8558 (normalize_one_pred_chain): Ditto.
8559 (normalize_preds): Ditto.
8560 (normalize_cond_1): Remove function.
8561 (normalize_cond): Ditto.
8562 (is_gcond_subset_of): Ditto.
8563 (is_subset_of_any): Ditto.
8564 (is_or_set_subset_of): Ditto.
8565 (is_and_set_subset_of): Ditto.
8566 (is_norm_cond_subset_of): Ditto.
8567 (pred_chain_length_cmp): Ditto.
8568 (convert_control_dep_chain_into_preds): Type change.
8569 (find_predicates): Ditto.
8570 (find_def_preds): Ditto.
8571 (destroy_predicates_vecs): Ditto.
8572 (find_matching_predicates_in_rest_chains): Ditto.
8573 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8574 (is_pred_expr_subset): Ditto.
8575 (is_pred_chain_subset_of): Ditto.
8576 (is_included_in): Ditto.
8577 (is_superset_of): Ditto.
8579 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8581 Update copyright years.
8583 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8585 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8586 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8587 config/arc/arc.md, config/arc/arc.opt,
8588 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8589 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8590 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8591 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8592 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8593 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8594 vtable-verify.c, vtable-verify.h: Use the standard form for the
8597 2014-01-02 Tobias Burnus <burnus@net-b.de>
8599 * gcc.c (process_command): Update copyright notice dates.
8600 * gcov-dump.c: Ditto.
8602 * doc/cpp.texi: Bump @copying's copyright year.
8603 * doc/cppinternals.texi: Ditto.
8604 * doc/gcc.texi: Ditto.
8605 * doc/gccint.texi: Ditto.
8606 * doc/gcov.texi: Ditto.
8607 * doc/install.texi: Ditto.
8608 * doc/invoke.texi: Ditto.
8610 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8612 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8614 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8616 * config/i386/sse.md (*mov<mode>_internal): Guard
8617 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8619 PR rtl-optimization/59647
8620 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8621 new_rtx into UNSIGNED_FLOAT rtxes.
8623 Copyright (C) 2014 Free Software Foundation, Inc.
8625 Copying and distribution of this file, with or without modification,
8626 are permitted in any medium without royalty provided the copyright
8627 notice and this notice are preserved.